NOTABLE

ARTICLES
OF 2016
A collection of important studies
from the past year as selected by NEJM editors
 December 2016

Dear Reader,

In 2016, the Journal published trials that sought to answer complicated questions. One such study
looked at whether men with early prostate cancer should undergo prostatectomy, radiation, or
watchful waiting to achieve the best outcome at 10 years. This study found that men with low-risk or
intermediate-risk prostate cancer had low prostate-cancerspecific mortality after 10 years, irrespective
of the treatment assigned. Importantly, these data helped with the conundrum of treating prostate cancer.
Since this is a disease of older men, the study balanced the competing issues of aggressive treatment
of a redolent disease with the reality that other factors may claim the life of the patient before he suc-
cumbs to prostate cancer. It provided solid landmarks for men wrestling with what to do when they
were diagnosed with low-intermediate risk prostate cancer.

Another study examined whether inducing labor at 39 weeks in pregnant women 35 years of age or
older, compared to expectant management, reduced stillbirth. While the study was underpowered
to assess differences in perinatal outcomes, it found no effect between the two groups on the rate of
caesarean section. This trial makes an important contribution to our current medical knowledge, and
helps build the foundation for larger, forthcoming studies. And even without larger studies, the data
presented helped pregnant women and their physicians visualize the risks and benefits of inducing
labor.

As the medical information published in NEJM is regularly used in daily practice, we ensure each paper
published meets exacting standards for editorial quality, clinical relevance, and impact on patient out-
comes. Among all papers published in 2016, this most notable collection was selected by the editors
as being the most meaningful in improving medical practice and patient care. We hope that you will
take valuable insights from these articles as you continue along your path of lifelong learning.

Jeffrey M. Drazen, M.D.

Editor-In-Chief, The New England Journal of Medicine
Distinguished Parker B. Francis Professor of Medicine
Harvard Medical School

800.843.6356 | f: 781.891.1995 | nejmgroup@mms.org

860 winter street, waltham, ma 02451-1413
nejmgroup.org
contents
original article
Incidence of Dementia over Three Decades in the Framingham Heart Study............................................. 1
perspective
Is Dementia in Decline? Historical Trends and Future Trajectories........................................ 2

original article
Effects of Testosterone Treatment in Older Men........................................................................................... 5
editorial
Establishing a Framework Does Testosterone Supplementation Help Older Men? 6

original article
National Cluster-Randomized Trial of Duty-Hour Flexibility in Surgical Training........................................ 8
editorial
Surgical Resident Duty-Hour Rules Weighing the New Evidence....................................... 9
perspective
Leaping without Looking Duty Hours, Autonomy,and the Risks of Research
and Practice.............................................................................................................................. 11

original article
Randomized Trial of Labor Induction in Women 35 Years of Age or Older............................................... 14
editorial
Induction of Labor and Cesarean Delivery.............................................................................. 15

original articles
Randomized Trial of Stent versus Surgery for Asymptomatic Carotid Stenosis......................................... 17
Long-Term Results of Stenting versus Endarterectomy for Carotid-Artery Stenosis.................................. 18
editorial
Endarterectomy, Stenting, or Neither for Asymptomatic Carotid-Artery Stenosis................ 19

original articles
Blood-Pressure Lowering in Intermediate-Risk Persons without Cardiovascular Disease......................... 21
Cholesterol Lowering in Intermediate-Risk Persons without Cardiovascular Disease............................... 22
Blood-Pressure and Cholesterol Lowering in Persons without Cardiovascular Disease............................ 23
editorial
More HOPE for Prevention with Statins.................................................................................. 24

original article
Efficacy of the Herpes Zoster Subunit Vaccine in Adults 70 Years of Age or Older................................... 27
editorial
Preventing Shingles and Its Complications in Older Persons................................................ 28

(continued on next page)

The New England Journal of Medicine is a publication of NEJM Group, a division of the Massachusetts Medical Society.
2016 Massachusetts Medical Society, All rights reserved.
contents (continued from previous page)

original articles
10-Year Outcomes after Monitoring, Surgery, or Radiotherapy for Localized Prostate Cancer................. 30
original article
Patient-Reported Outcomes after Monitoring, Surgery, or Radiotherapy for Prostate Cancer.................. 31
editorial
Treatment or Monitoring for Early Prostate Cancer................................................................ 32

original article
GuillainBarr Syndrome Associated with Zika Virus Infection in Colombia............................................. 34
editorial
Zika Getting on Your Nerves? The Association with the GuillainBarr Syndrome.............. 35

original article
A Randomized Trial of Long-term Oxygen for COPD with Moderate Desaturation .................................. 37
editorial
Clinical Usefulness of Long-Term Oxygen Therapy in Adults................................................. 38

perspective
Zika Virus in the Americas Yet Another Arbovirus Threat ..................................................................... 40

perspective
Reducing the Risks of Relief The CDC Opioid-Prescribing Guideline ................................................... 44

perspective
Rethinking the Primary Care Workforce An Expanded Role for Nurses................................................. 48
1 Notable Articles of 2016 nejm.org
The n e w e ng l a n d j o u r na l of m e dic i n e

Original Article

Incidence of Dementia over Three Decades

in the Framingham Heart Study
Claudia L. Satizabal, Ph.D., Alexa S. Beiser, Ph.D., Vincent Chouraki, M.D., Ph.D.,
Genevive Chne, M.D., Ph.D., Carole Dufouil, Ph.D., and Sudha Seshadri, M.D.

A BS T R AC T

BACKGROUND
The prevalence of dementia is expected to soar as the average life expectancy in- From the Boston University Schools of
creases, but recent estimates suggest that the age-specific incidence of dementia Medicine (C.L.S., A.S.B., V.C., S.S.) and
Public Health (A.S.B.), Boston, and the
is declining in high-income countries. Temporal trends are best derived through Framingham Heart Study, Framingham
continuous monitoring of a population over a long period with the use of consis- (C.L.S., A.S.B., V.C., S.S.) all in Mas-
tent diagnostic criteria. We describe temporal trends in the incidence of dementia sachusetts; and Inserm Unit 1219 and
CIC 1401-EC (Clinical Epidemiology) and
over three decades among participants in the Framingham Heart Study. University of Bordeaux, ISPED (Bordeaux
School of Public Health) both in Bor-
METHODS deaux, France (G.C., C.D.). Address re-
Participants in the Framingham Heart Study have been under surveillance for in- print requests to Dr. Seshadri at the Bos-
ton University School of Medicine,
cident dementia since 1975. In this analysis, which included 5205 persons 60 years Department of Neurology, 72 E. Concord
of age or older, we used Cox proportional-hazards models adjusted for age and sex St., B602, Boston, MA 02118, or at
to determine the 5-year incidence of dementia during each of four epochs. We also suseshad@bu.edu.
explored the interactions between epoch and age, sex, apolipoprotein E 4 status, N Engl J Med 2016;374:523-32.
and educational level, and we examined the effects of these interactions, as well DOI: 10.1056/NEJMoa1504327
Copyright 2016 Massachusetts Medical Society.
as the effects of vascular risk factors and cardiovascular disease, on temporal
trends.
Read Full Article at NEJM.org
RESULTS
The 5-year age- and sex-adjusted cumulative hazard rates for dementia were 3.6 per
100 persons during the first epoch (late 1970s and early 1980s), 2.8 per 100 persons
during the second epoch (late 1980s and early 1990s), 2.2 per 100 persons during
the third epoch (late 1990s and early 2000s), and 2.0 per 100 persons during the
fourth epoch (late 2000s and early 2010s). Relative to the incidence during the first
epoch, the incidence declined by 22%, 38%, and 44% during the second, third,
and fourth epochs, respectively. This risk reduction was observed only among
persons who had at least a high school diploma (hazard ratio, 0.77; 95% confi-
dence interval, 0.67 to 0.88). The prevalence of most vascular risk factors (except
obesity and diabetes) and the risk of dementia associated with stroke, atrial fibril-
lation, or heart failure have decreased over time, but none of these trends com-
pletely explain the decrease in the incidence of dementia.
CONCLUSIONS
Among participants in the Framingham Heart Study, the incidence of dementia
has declined over the course of three decades. The factors contributing to this
decline have not been completely identified. (Funded by the National Institutes of
Health.)

n engl j med 374;6 nejm.org February 11, 2016 523

Back to Table of Contents

 2 Notable Articles of 2016
PE R S PE C T IV E
Is Dementia in Decline?
His tory of Medicine
Is Dementia in Decline? Historical Trends and Future
Trajectories
David S. Jones, M.D., Ph.D., and Jeremy A. Greene, M.D., Ph.D.
I n 2005, researchers from the
Duke Center for Demographic
Studies reported a surprising
trend: data from the National
Long-Term Care Surveys showed
that the prevalence of severe cog-
nitive impairment in the Medi-
care population had decreased
significantly between 1982 and
1999.1 At a time when baby-
boomer demographics led to pre-
dictions of a looming dementia
crisis, this finding offered hope.
Since that time, other reports
have similarly shown that the in-
cidence or prevalence of demen-
tia is decreasing in various popu-
lations. Researchers have offered
many possible explanations, in-
cluding increased wealth, better
education, control of vascular risk
factors, and use of statins, anti-
hypertensive agents, and nonste-
roidal antiinflammatory drugs.1,2
However, even as researchers de-
scribe their cautious optimism
about specific populations, they
still project a quadrupling of
global prevalence over the com-
ing decades.3
In this issue of the Journal,
Satizabal and colleagues report
more robust evidence of demen-
tias decline (pages 52332). Using
surveillance data collected from
the Framingham Heart Study
from 1975 to the present, they
found a 20% decrease in demen-
tia incidence each decade, even
as average body-mass index, dia-
betes prevalence, and population
age have increased. Can we now
conclude that the tide has turned
n engl j med 374;6
in the dementia epidemic? The
potential decline of dementia,
seen in light of the rise and fall
of other major diseases, raises an
even more tantalizing prospect:
Can we control our burden of
disease?
This is not the first time that
the medical profession and the
public health community have
struggled to interpret reports of an
unexpected reversal of a chronic-
disease epidemic.4 In 1964, Cali-
fornia health officials reported
that rates of coronary artery dis-
ease (CAD) had begun to de-
crease. This finding, which defied
the widespread belief that the
CAD epidemic would only worsen
as life expectancy grew, garnered
scant attention. Even a decade
later, most health officials as-
sumed that CAD was still on the
rise. It was only in 1974 that re-
searchers began taking the pros-
pect of decline seriously. By 1978,
they had accepted that CADs na-
tional decline had begun in the
mid-1960s. Similar decreases were
soon reported in many other high-
income countries, from Australia
to Finland. This recognition trig-
gered debate over the contribu-
tion of medical and public health
interventions, in hopes that knowl-
edge of the causes of decline
would guide policies and resource
allocation and ensure continua-
tion of these health benefits.
The history of the debate on
CAD decline carries important
lessons for emerging reports of
dementias decline. First, it can
nejm.org February 11, 2016
Back to Table of Contents
nejm.org
Is Dementia in Decline?
Related article, p. 523
be extremely difficult to produce
timely and convincing data about
the trajectories of chronic dis-
eases.4 When physicians began to
debate CAD trends in 1974, they
had to rely on government data
that were 5 years out of date. It
took 4 years of concerted effort
to reach consensus about an in-
flection that had occurred more
than a decade earlier. Even though
better and timelier data are now
available, dementia researchers
must still be resourceful in seek-
ing convincing data. As Satizabal
et al. indicate, each existing re-
port has limitations. Their new
data, which overcome many of
these limitations, demonstrate the
value of investments in long-term,
longitudinal epidemiologic re-
search such as the Framingham
Heart Study. But the data still re-
flect only one population sample.
Whether they are accepted as
conclusive evidence of a broad-
based reduction in dementia in-
cidence will become clear only
over time.
Second, since trajectories of
chronic-disease incidence reflect
complex interactions of many
causal factors, it will almost al-
ways be uncertain whether de-
creases will continue or reverse.
Even as consensus about interna-
tional CAD reduction consolidat-
ed between the 1970s and the
1990s, worrisome evidence about
countervailing trends also ap-
peared.4 Enthusiasm for anti-CAD
public health campaigns has been
fragile, even in countries like
507
 3 Notable Articles of 2016
PE R S PE C T IV E
Finland that demonstrated their
promise so well. The widespread
increases in obesity and diabetes
could fuel CAD resurgences.
Many researchers have warned
that CADs decline could stall or
even reverse something that
has happened among young
adults and other subpopulations
in Europe, Australia, and the
United States. Other countries,
such as China, continue to see
increases in CAD with no evi-
dence of plateau or reversal.
History offers reasons for hope.
Evidence of dementias decline shows
once again that our burden of disease
is malleable.
All these countervailing trends
could affect dementia as well.
Rocca and colleagues have warned
that increases in obesity, diabetes,
and hypertension could under-
mine the gains achieved through
improved education, wealth, and
control of vascular risk factors.
Even if a dementia decline has
begun, it might not last: the out-
come depends on the balance of
diverging trends.2,3
Third, these ambiguities open
up a battleground for conflicting
interpretations by interested par-
ties. Policymakers can use the
same data to tell vastly different
stories about public health. Fore-
casts of CADs future continue to
swing between narratives of tri-
umph and catastrophe.4 The good
news is that more and more
countries are reporting evidence
of decline. The bad news is the
evidence of the fragility of these
gains.
Narratives of dementia remain
similarly malleable. In the early
508 n engl j med 374;6
1980s, even after CADs decline
had been accepted and despite
knowledge that dementia shares
many risk factors with CAD, phy-
sicians began to warn about an
exploding dementia epidemic.5
The decrease in prevalence that
surprised Manton and colleagues
in 2005 could have been predict-
ed decades earlier. But dementia
will remain a problem despite
these decreases. The prevalence
of dementia can increase, even if
the incidence falls, if global pop-
ulations live longer. The absolute
number of people with dementia
can increase, even if both inci-
dence and prevalence fall, if the
size of the elderly population
grows. That explains why, 10 years
into the era of reports of decreas-
ing dementia in selected popula-
tions, Satizabal and colleagues
still write that the prevalence of
dementia is expected to soar as
our societies age. Even research-
ers rigorously examining the evi-
dence of decreases continue to
worry about what the future will
bring.
History offers reasons for
hope. Evidence of dementias de-
cline shows once again that our
burden of disease is malleable.
This lesson has been hard won.
Mid-19th-century physicians saw
cholera and tuberculosis as in-
evitable scourges of urban envi-
ronments. But those epidemics
yielded to sanitary reform, im-
proved standards of living, and
eventually medical care. As con-
nejm.org February 11, 2016
Back to Table of Contents
nejm.org
Is Dementia in Decline?
trol of infectious disease led to
dramatic gains in life expectancy,
physicians in the early 20th cen-
tury came to see CAD and cancer
as the inevitable scourges of long
lives. Over recent decades, that
pessimism has largely given way
as well: CAD and many forms of
cancer are increasingly prevent-
able and curable. The burden of
disease of the 20th century, like
that of the 19th, was not an in-
evitable fact of life, but a product
of lives lived amid specific
and malleable conditions.
What should we expect as can-
cer and heart disease come under
control? Many people think that
we can live even longer lives
but lives compromised by demen-
tia, vision loss, and hearing loss.
Whether that fate is inevitable or
whether these, too, are malleable
scourges remains to be seen.
Such questions are better left to
futurists and geriatricians than
to historians. Yet Satizabal et al.
believe theres cause for cautious
hope. Primary and secondary pre-
vention might diminish the mag-
nitude of the long-feared dementia
epidemic. Something else might
save our vision and hearing.
Faced with choices between
equally defensible epidemiologic
projections, physicians and re-
searchers must think carefully
about what stories they empha-
size to patients and policymak-
ers. The implications, especially
for investment in long-term care
facilities, are enormous. Our ex-
planations of decline are equally
important, since they guide in-
vestments in behavior change,
medications, and other treat-
ments.
With this latest contribution,
optimism about dementia is more
justified than ever before. Even if
death and taxes remain inevita-
 4 Notable Articles of 2016
PER S PE C T IV E
ble, cancer, CAD, and dementia
may not. But cautious optimism
should not become complacency.
If we can elucidate the changes
that have contributed to these
improvements, perhaps we can ex-
tend them. Today, the dramatic
reductions in CAD-related mor-
tality are under threat. The incipi-
ent improvements in dementia
are presumably even more fragile.
The burden of disease, ever mal-
leable, can easily relapse.
n engl j med 374;6
Disclosure forms provided by the authors
are available with the full text of this article
at NEJM.org.
From the Department of Global Health and
Social Medicine, Harvard Medical School,
Boston (D.S.J); the Department of the His-
tory of Science, Harvard University, Cam-
bridge, MA (D.S.J.); and the Division of
General Internal Medicine and the Depart-
ment of the History of Medicine, Johns
Hopkins University School of Medicine,
Baltimore (J.A.G.).
1. Manton KC, Gu XL, Ukraintseva SV. De-
clining prevalence of dementia in the U.S. el-
derly population. Adv Gerontol 2005;16:30-7.
2. Larson EB, Yaffe K, Langa KM. New in-
nejm.org February 11, 2016
Back to Table of Contents
nejm.org
Is Dementia in Decline?
sights into the dementia epidemic. N Engl J
Med 2013;369:2275-7.
3. Rocca WA, Petersen RC, Knopman DS,
et al. Trends in the incidence and prevalence
of Alzheimers disease, dementia, and cog-
nitive impairment in the United States. Al-
zheimers Dement 2011;7:80-93.
4. Jones DS, Greene JA. The decline and
rise of coronary heart disease: understand-
ing public health catastrophism. Am J Pub-
lic Health 2013;103:1207-18.
5. Beck JC, Benson DF, Scheibel AB, Spar
JE, Rubenstein LZ. Dementia in the elderly:
the silent epidemic. Ann Intern Med 1982;
97:231-41.
DOI: 10.1056/NEJMp1514434
Copyright 2016 Massachusetts Medical Society.
Is Dementia in Decline?
509
5 Notable Articles of 2016 nejm.org

new england
The
journal of medicine
established in 1812 February 18, 2016 vol. 374 no. 7

Effects of Testosterone Treatment in Older Men

P.J. Snyder, S. Bhasin, G.R. Cunningham, A.M. Matsumoto, A.J. Stephens-Shields, J.A. Cauley, T.M. Gill,
E. Barrett-Connor, R.S. Swerdloff, C. Wang, K.E. Ensrud, C.E. Lewis, J.T. Farrar, D. Cella, R.C. Rosen, M. Pahor,
J.P. Crandall, M.E. Molitch, D. Cifelli, D. Dougar, L. Fluharty, S.M. Resnick, T.W. Storer, S. Anton, S. Basaria,
S.J. Diem, X. Hou, E.R. Mohler III, J.K. Parsons, N.K. Wenger, B. Zeldow, J.R. Landis, and S.S. Ellenberg,
for the Testosterone Trials Investigators*

a bs t r ac t

BACKGROUND
Serum testosterone concentrations decrease as men age, but benefits of raising testos- The authors full names, academic de-
terone levels in older men have not been established. grees, and affiliations are listed in the Ap-
pendix. Address reprint requests to Dr.
METHODS Snyder at pjs@mail.med.upenn.edu.
We assigned 790 men 65 years of age or older with a serum testosterone concentration * A complete list of investigators in the
of less than 275 ng per deciliter and symptoms suggesting hypoandrogenism to receive Testosterone Trials is provided in the
either testosterone gel or placebo gel for 1 year. Each man participated in one or more Supplementary Appendix, available at
NEJM.org.
of three trials the Sexual Function Trial, the Physical Function Trial, and the Vital-
ity Trial. The primary outcome of each of the individual trials was also evaluated in all Drs. Bhasin, Cunningham, Matsumoto,
Stephens-Shields, and Ellenberg contrib-
participants. uted equally to this article.
RESULTS N Engl J Med 2016;374:611-24.
Testosterone treatment increased serum testosterone levels to the mid-normal range for DOI: 10.1056/NEJMoa1506119
men 19 to 40 years of age. The increase in testosterone levels was associated with sig- Copyright 2016 Massachusetts Medical Society.

nificantly increased sexual activity, as assessed by the Psychosexual Daily Questionnaire

(P<0.001), as well as significantly increased sexual desire and erectile function. The Read Full Article at NEJM.org
percentage of men who had an increase of at least 50 m in the 6-minute walking dis-
tance did not differ significantly between the two study groups in the Physical Function
Trial but did differ significantly when men in all three trials were included (20.5% of
men who received testosterone vs. 12.6% of men who received placebo, P = 0.003). Tes-
tosterone had no significant benefit with respect to vitality, as assessed by the Func-
tional Assessment of Chronic Illness TherapyFatigue scale, but men who received tes-
tosterone reported slightly better mood and lower severity of depressive symptoms than
those who received placebo. The rates of adverse events were similar in the two groups.
CONCLUSIONS
In symptomatic men 65 years of age or older, raising testosterone concentrations for
1 year from moderately low to the mid-normal range for men 19 to 40 years of age had
a moderate benefit with respect to sexual function and some benefit with respect to
mood and depressive symptoms but no benefit with respect to vitality or walking dis-
tance. The number of participants was too few to draw conclusions about the risks of
testosterone treatment. (Funded by the National Institutes of Health and others;
ClinicalTrials.gov number, NCT00799617.)

n engl j med 374;7 nejm.org February 18, 2016 611

Back to Table of Contents
 6 Notable Articles of 2016 nejm.org
The n e w e ng l a n d j o u r na l of m e dic i n e

Edi t or i a l

Establishing a Framework Does Testosterone

Supplementation Help Older Men?
Eric S. Orwoll, M.D.

Aging is variably but inevitably accompanied by
declines in health; concomitantly, in men, circu-
lating sex-steroid levels fall with age.1 To what
extent these two processes are causally linked
and whether testosterone therapy can prevent or
ameliorate important age-related problems have
been major issues in mens health. In 2003, a com-
mittee assembled by the Institute of Medicine
(IOM) found a paucity of randomized, placebo-
controlled clinical trials involving older men and
noted a lack of definite evidence that testoster-
one therapy conferred benefits.2 The committee
recommended that clinical trials be initiated,
first to evaluate the efficacy of testosterone
supplementation in older men and then to assess
long-term benefits and risks through large-scale
trials.
Little has changed to alter the conclusions of
that report; if anything, the issue of testosterone
supplementation has become more controversial.3
However, in this issue of the Journal, Snyder et al.4
describe the long-awaited initial results of the
National Institutes of Healthsponsored Testos-
terone Trials, which were designed to address
the key issues identified by the IOM. Their re-
port is important, not only because it deals with
an essential public health issue but also because
the investigators have succeeded in conducting
the kind of generally well-conceived studies that
are sorely needed in the field. The findings be-
gin to provide a basis for more rational clinical
decisions about testosterone use as well as for
additional research.
The overall design of the Testosterone Trials is
complex.5 It includes seven independent, double-
blind, placebo-controlled trials intended to ad-
dress specific outcomes that are postulated to be
related to testosterone deficiency (sexual function,
vitality, physical function, cognitive function,
anemia, bone density, and cardiovascular status).
The trials are knitted together by common
methods and some shared measures, thus maxi-
mizing the power of the overall investigation.
This inaugural report describes the findings of
the three main studies (with primary outcomes
related to sexual function, physical function,
and vitality).
The results show that testosterone therapy did
yield certain benefits, but at this point their
clinical importance is uncertain. Therapy was
not a panacea, and the findings alone might be
insufficient to support a decision to initiate tes-
tosterone therapy in symptomatic older men. The
study confirmed that testosterone supplementa-
tion can yield improvements in sexual function,
but the benefits were modest, tended to wane in
the latter months of the treatment period, and,
as the authors note, were not as robust as those
of phosphodiesterase type 5 inhibitors.6 There
were only small gains in physical performance
and in indexes of mood and depression; overall
vitality was no better with testosterone therapy
than with placebo. For each of the outcomes,
some older men may have a more vigorous re-
sponse to testosterone therapy and thus could be
more attractive candidates for supplementation;
however, it was not possible to confidently iden-
tify them by the testosterone levels achieved
with therapy. As expected, estradiol levels also
increased; those levels have been linked to key
health variables in men (e.g., sexual function).7
Its not yet clear whether responses (or the lack
thereof) in the Testosterone Trials may be due to
changes in estradiol levels.
There is considerable controversy about pos-
sible adverse effects of testosterone therapy in
older men, and these studies do not resolve this
controversy. Although there were minor effects

682 n engl j med 374;7 nejm.org February 18, 2016

Back to Table of Contents

7 Notable Articles of 2016
Editorial
on hemoglobin and prostate-specific antigen
levels, and, reassuringly, no apparent major
toxic effects, larger and more extended trials
would be needed to determine whether therapy
has negative effects on outcomes such as pros-
tate or cardiovascular health.
Importantly, the study participants were re-
cruited on the basis of stringent criteria (age 65
years, total testosterone levels below the normal
range in men 19 to 40 years of age [<275 ng per
deciliter], symptoms related to predetermined
outcomes, and no contraindications to partici-
pation). Only 1.5% of those screened (790 of
51,085 men) were eligible and enrolled. The aver-
age participant was 72 years of age; almost 90%
of participants were white, most were obese,
most had hypertension, more than one third had
diabetes, and almost 20% had sleep apnea. The
select nature of the participants reflects the sci-
entific rigor of the trials (and the causes of low
testosterone levels) but also clearly limits the
generalizability of the conclusions. We should
not assume that the benefits, lack of benefits, or
adverse-event profile observed in these studies
would be similar in younger men (most testos-
terone prescriptions are written for middle-aged
men3), men with higher testosterone levels, or
those with different demographic or clinical
characteristics.
The report by Snyder et al. is likely to stimu-
n engl j med 374;7
Back to Table of Contents
nejm.org
late controversy and to engender additional re-
search questions as did the Womens Health
Initiative with respect to estrogen-replacement
therapy. Nevertheless, it is a landmark study in
the field of mens health and no doubt a bell-
wether for additional important contributions
from the Testosterone Trials.
Disclosure forms provided by the author are available with the
full text of this article at NEJM.org.
From the Department of Medicine, Oregon Health and Science
University, Portland.
1. Harman SM, Metter EJ, Tobin JD, Pearson J, Blackman MR.
Longitudinal effects of aging on serum total and free testoster-
one levels in healthy men. J Clin Endocrinol Metab 2001;86:
724-31.
2. Liverman C, Blazer D, eds. Testosterone and aging: clinical
research directions. Washington, DC: National Academies
Press, 2004.
3. Nguyen CP, Hirsch MS, Moeny D, Kaul S, Mohamoud M,
Joffe HV. Testosterone and age-related hypogonadism FDA
concerns. N Engl J Med 2015;373:689-91.
4. Snyder PJ, Bhasin S, Cunningham GR, et al. Effects of testos-
terone treatment in older men. N Engl J Med 2016;374:611-24.
5. Snyder PJ, Ellenberg SS, Cunningham GR, et al. The Testos-
terone Trials: seven coordinated trials of testosterone treatment
in elderly men. Clin Trials 2014;11:362-75.
6. Spitzer M, Basaria S, Travison TG, et al. Effect of testoster-
one replacement on response to sildenafil citrate in men with
erectile dysfunction: a parallel, randomized trial. Ann Intern
Med 2012;157:681-91.
7. Finkelstein JS, Yu EW, Burnett-Bowie SA. Gonadal steroids
and body composition, strength, and sexual function in men.
N Engl J Med 2013;369:2455-7.
DOI: 10.1056/NEJMe1600196
Copyright 2016 Massachusetts Medical Society.
nejm.org February 18, 2016 683
8 Notable Articles of 2016 nejm.org

new england
The
journal of medicine
established in 1812 February 25, 2016 vol. 374 no. 8

National Cluster-Randomized Trial of Duty-Hour Flexibility

in Surgical Training
Karl Y. Bilimoria, M.D., M.S.C.I., Jeanette W. Chung, Ph.D., Larry V. Hedges, Ph.D., Allison R. Dahlke, M.P.H.,
Remi Love, B.S., Mark E. Cohen, Ph.D., David B. Hoyt, M.D., Anthony D. Yang, M.D., John L. Tarpley, M.D.,
John D. Mellinger, M.D., David M. Mahvi, M.D., Rachel R. Kelz, M.D., M.S.C.E., Clifford Y. Ko, M.D., M.S.H.S.,
David D. Odell, M.D., M.M.Sc., Jonah J. Stulberg, M.D., Ph.D., M.P.H., and Frank R. Lewis, M.D.

a bs t r ac t

BACKGROUND
Concerns persist regarding the effect of current surgical resident duty-hour policies on From the Surgical Outcomes and Quality
patient outcomes, resident education, and resident well-being. Improvement Center (SOQIC), Depart-
ment of Surgery and Center for Health-
METHODS care Studies, Feinberg School of Medicine
and Northwestern Medicine, Northwest-
We conducted a national, cluster-randomized, pragmatic, noninferiority trial involving ern University (K.Y.B., J.W.C., A.R.D., R.L.,
117 general surgery residency programs in the United States (20142015 academic year). A.D.Y., D.M.M., D.D.O., J.J.S.), and the
Programs were randomly assigned to current Accreditation Council for Graduate Medical American College of Surgeons (K.Y.B.,
M.E.C., D.B.H., C.Y.K.), Chicago, the De-
Education (ACGME) duty-hour policies (standard-policy group) or more flexible policies partment of Statistics, Northwestern Uni-
that waived rules on maximum shift lengths and time off between shifts (flexible-policy versity, Evanston (L.V.H.), and the De-
group). Outcomes included the 30-day rate of postoperative death or serious complica- partment of Surgery, Southern Illinois
University, Springfield (J.D.M.) all in
tions (primary outcome), other postoperative complications, and resident perceptions Illinois; the Department of Surgery,
and satisfaction regarding their well-being, education, and patient care. Vanderbilt University, Nashville (J.L.T.);
the Department of Surgery and the Cen-
RESULTS ter for Surgery and Health Economics,
In an analysis of data from 138,691 patients, flexible, less-restrictive duty-hour policies Perelman School of Medicine, University
were not associated with an increased rate of death or serious complications (9.1% in the of Pennsylvania (R.R.K.), and the Ameri-
can Board of Surgery (F.R.L.) both in
flexible-policy group and 9.0% in the standard-policy group, P = 0.92; unadjusted odds Philadelphia; and the Department of Sur-
ratio for the flexible-policy group, 0.96; 92% confidence interval, 0.87 to 1.06; P = 0.44; gery, University of California, Los Ange-
noninferiority criteria satisfied) or of any secondary postoperative outcomes studied. les, School of Medicine, Los Angeles
(C.Y.K.). Address reprint requests to Dr.
Among 4330 residents, those in programs assigned to flexible policies did not report Bilimoria at the Surgical Outcomes and
significantly greater dissatisfaction with overall education quality (11.0% in the flexible- Quality Improvement Center (SOQIC),
policy group and 10.7% in the standard-policy group, P = 0.86) or well-being (14.9% and Department of Surgery, Feinberg School
of Medicine and Northwestern Medicine,
12.0%, respectively; P = 0.10). Residents under flexible policies were less likely than those Northwestern University, 633 N. St. Clair
under standard policies to perceive negative effects of duty-hour policies on multiple St., 20th Fl., Chicago, IL 60611, or at
aspects of patient safety, continuity of care, professionalism, and resident education but k-bilimoria@northwestern.edu.

were more likely to perceive negative effects on personal activities. There were no sig- This article was published on February 2,
nificant differences between study groups in resident-reported perception of the effect of 2016, at NEJM.org.

fatigue on personal or patient safety. Residents in the flexible-policy group were less N Engl J Med 2016;374:713-27.
likely than those in the standard-policy group to report leaving during an operation (7.0% DOI: 10.1056/NEJMoa1515724
Copyright 2016 Massachusetts Medical Society.
vs. 13.2%, P<0.001) or handing off active patient issues (32.0% vs. 46.3%, P<0.001).
CONCLUSIONS Read Full Article at NEJM.org
As compared with standard duty-hour policies, flexible, less-restrictive duty-hour policies
for surgical residents were associated with noninferior patient outcomes and no signifi-
cant difference in residents satisfaction with overall well-being and education quality.
(FIRST ClinicalTrials.gov number, NCT02050789.)
n engl j med 374;8 nejm.org February 25, 2016 713

Back to Table of Contents

9 Notable Articles of 2016 nejm.org
The n e w e ng l a n d j o u r na l of m e dic i n e

Edi t or i a l

Surgical Resident Duty-Hour Rules Weighing the New

Evidence
John D. Birkmeyer, M.D.

Surgical training has always been hard on resi-
dents. During my own residency more than 20
years ago, 100-hour workweeks and in-house call
every other night were routine. A residents life
outside the hospital was simply not a priority.
Residency may be even harder on patients. A large
body of research has linked sleep deprivation in
resident physicians to poor performance in neu-
robehavioral testing and, more alarmingly, to
higher rates of attention failure in patient care.1,2
Reacting to concerns about both resident well-
being and patient safety, the Accreditation Coun-
cil for Graduate Medical Education (ACGME)
implemented duty-hour reforms in 2003 that
constrained resident workweeks to 80 hours,
among other changes. A 2011 update added new
limits to the length of individual shifts (24 hours
plus 4 hours for transition) and guaranteed a
minimum amount of time off between 24-hour
shifts (14 hours). Although they are not nearly as
stringent as standards set in other occupations
in which performance has implications for pub-
lic safety (e.g., airline pilots), the ACGME rules
were nonetheless criticized by many in the medi-
cal community. Surgeons in particular were
concerned that the new duty-hour rules would
paradoxically increase medical errors as a result
of increased handoffs residents signing out
their sickest patients to providers who are not
familiar with their cases. In other words, the
safety benefits of reducing resident fatigue would
be offset by harms associated with disrupting
continuity of care.
Extending the results of a previous national
study based on Medicare claims data,3 a very
ambitious, scientifically robust study by Bilimoria
et al. now published in the Journal should help
allay these concerns.4 By random assignment, 59
general-surgery training programs were required
to adhere to the ACGME rules about maximum
shift length and minimum time off between 24-
hour shifts. Another 59 programs were granted
flexibility and did not have to adhere to those
rules. Both groups adhered to ACGME require-
ments for total workweek hours. Residents who
were not required to adhere to the duty-hour rules
were less likely to report dissatisfaction with
continuity of care and hand-offs. After 1 year,
however, the two groups of teaching hospitals
had virtually indistinguishable rates of death,
overall complications, and specific types of com-
plications, on the basis of data on risk-adjusted
clinical outcomes from the American College of
Surgeons National Surgical Quality Improvement
Program.
It is not surprising that outcomes did not
vary according to whether programs adhered to
ACGME requirements on maximum shift length
and time off between shifts. The patients most
likely to be affected by resident handoffs
those with acute or deteriorating clinical condi-
tions represent only a small percentage of
surgical patients at teaching hospitals. More
important, teaching hospitals have become far
less reliant on surgical residents than they used
to be. In earlier eras, surgical residents had con-
siderable autonomy. During my own residency,
surgical residents often operated independently,
particularly at night and on weekends. Today,
they operate almost exclusively in the presence
of an attending surgeon. Intensive care units,
which house the sickest surgical patients, are
increasingly closed and staffed by board-certi-
fied intensivists. Postoperative care is delivered
by multidisciplinary teams staffed with associate
providers as well as residents.

n engl j med 374;8 nejm.org February 25, 2016 783

Back to Table of Contents

 10 Notable Articles of 2016
Editorial
The Flexibility in Duty Hour Requirements for
Surgical Trainees (FIRST) Trial also assessed the
effects of ACGME duty-hour restrictions on resi-
dent perceptions of educational quality and well-
being, with the use of survey data collected
annually by the American Board of Surgery.
Residents in the two groups of teaching hospi-
tals had similarly high rates of satisfaction with
the quality of their training. Although residents in
programs not required to adhere to the ACGME
duty-hour rules were more likely to be dissatis-
fied with time for rest, there were no significant
differences in overall resident well-being and
morale between the two groups.
What do the results of the FIRST Trial mean
for ACGME policy on resident duty hours? The
authors conclude, as will many surgeons, that
surgical training programs should be afforded
more flexibility in applying work-hour rules.
This interpretation implicitly places the burden
of proof on the ACGME. Thus, because the
FIRST Trial found no evidence that removing
restrictions on resident shift length and time
off between shifts was harmful to patients,
programs should have more autonomy to train
residents as they choose.
I reach a different conclusion. The FIRST Trial
effectively debunks concerns that patients will
suffer as a result of increased handoffs and
breaks in the continuity of care. Rather than
backtrack on the ACGME duty-hour rules, surgi-
784 n engl j med 374;8
Back to Table of Contents
nejm.org
cal leaders should focus on developing safe, re-
silient health systems that do not depend on
overworked resident physicians. They should also
recognize the changing expectations of post-
millennial learners. To many current residents
and medical students, 80-hour (or even 72-hour)
workweeks and 24-hour shifts probably seem
long enough. Although few surgical residents
would ever acknowledge this publicly, Im sure
that many love to hear, We can take care of this
case without you. Go home, see your family, and
come in fresh tomorrow.
Disclosure forms provided by the author are available with the
full text of this article at NEJM.org.
From the DartmouthHitchcock Medical Center and the Dart-
mouth Institute for Health Policy and Clinical Practice both
in Lebanon, NH.
This article was published on February 2, 2016, at NEJM.org.
1. Veasey S, Rosen R, Barzansky B, Rosen I, Owens J. Sleep loss
and fatigue in residency training: a reappraisal. JAMA 2002;288:
1116-24.
2. Lockley SW, Cronin JW, Evans EE, et al. Effect of reducing
interns weekly work hours on sleep and attentional failures.
N Engl J Med 2004;351:1829-37.
3. Volpp KG, Rosen AK, Rosenbaum PR, et al. Mortality among
hospitalized Medicare beneficiaries in the first 2 years following
ACGME resident duty hour reform. JAMA 2007;298:975-83.
4. Bilimoria KY, Chung JW, Hedges LV, et al. National cluster-
randomized trial of duty-hour flexibility in surgical training.
N Engl J Med 2016;374:713-27.
DOI: 10.1056/NEJMe1516572
Copyright 2016 Massachusetts Medical Society.
nejm.org February 25, 2016
 11 Notable Articles of 2016 nejm.org

The NEW ENGLA ND JOURNAL of MEDICINE

Perspective February 25, 2016

Leaping without Looking Duty Hours, Autonomy,

and the Risks of Research and Practice
Lisa Rosenbaum, M.D.
Related article, p. 713

I
Leaping without Looking
n 2014, Facebook users were furious to discover randomization at the residency-
that theyd unwittingly been experimented on.1 program level, and neither required
consent of residents or patients.
Researchers had randomly assigned users to news That consent waiver has drawn
feeds with reduced positive content or reduced criticism from Public Citizen and
the American Medical Student
negative content and found that leveled at investigators who are Association, which in open letters
happy posts beget happy posts comparing the 2011 duty-hour to the Office for Human Research
and that grim ones beget grim restrictions imposed by the Ac- Protections (OHRP) accuse the in-
ones.2 Although that may now creditation Council for Graduate vestigators of egregious ethical
seem obvious, previous evidence Medical Education (ACGME) with and regulatory violations.3,4
had suggested that because we more flexible shift lengths for The allegations, focused pri-
tend to compare ourselves to oth- residents. The Flexibility in Duty marily on serious health risks
ers, exposure to positive content Hour Requirements for Surgical to residents from long shifts, are
compromises users well-being. Trainees (FIRST) trial, whose re- dizzyingly tautological. The critics
There was thus no reason to be- sults are now reported by Bilimo- claim its unethical not to obtain
lieve that the status quo news ria et al. in the Journal, compared residents consent; but because
feeds curated by an algorithm 59 surgical training programs pressure on residents to conform
tailored to users viewing habits randomly assigned to an ACGME- makes seeking their consent akin
was any safer than the experi- compliant schedule with 58 grant- to coercion, thats unethical too.
mental interventions. And given ed flexibility in designing shift Thus, theres no ethical way to
Facebooks reach, there were com- lengths (still within an 80-hour study the duty-hour rules in a
pelling reasons to find out. Never- workweek). The ongoing Individ- randomized fashion. But thats
theless, the results triggered out- ualized Comparative Effectiveness fine, because we already know
rage that 700,000 users had been of Models Optimizing Safety and theyre beneficial; we know that
exposed to potential emotional Resident Education (iCOMPARE) because the ACGME made the
damage without their consent. trial involves internal medicine rules in the first place. And if the
Similar accusations have been programs. Both used cluster trials found otherwise, their re-

n engl j med 374;8 nejm.org February 25, 2016 701

Back to Table of Contents

 12 Notable Articles of 2016 nejm.org
Leaping without Looking
PER S PE C T IV E

sults challenging the status quo
would be suspect because the in-
vestigators, who have publicly ac-
knowledged the need for data to
inform policy, are consequently
too biased to generate those data.
To unpack these allegations,
its important to understand that
even if the trials are considered
No drug would be approved solely
on the basis of laboratory evidence. Yet we
require neither consideration of complexity
nor rigorous studies before implementing
policies with broad implications. Why?
human-subjects research, there
are circumstances under which
federal rules deem it ethical to
waive consent. The key one here
is that the incremental risk posed
by the research should be, at
most, minimal. For trials like
these that evaluate a standard
practice, the question becomes:
Is there equipoise between the
status quo and investigational
groups in terms of possible risks?
Though the letters to OHRP
claim otherwise, the answer is un-
equivocally yes. The complaints
ignore a considerable body of re-
search suggesting, as Bilimoria
et al. point out, that duty-hour re-
forms have not improved patient
safety; some trials have even
raised concerns that theyve actu-
ally worsened quality of care and
patient outcomes.
As for risks to residents, the
letters cite data suggesting that
fatigue causes harms such as in-
creased motor vehicle accidents,
needlesticks, and burnout. Yet
theres little evidence to suggest
that shorter hours have reduced
occupational hazards or burnout
rates. Though I suspect that these
findings partly reflect the emo-
tional toll of work compression
and the reality that many trainees
dont actually sleep more, they
also speak to a fundamental
challenge in improving care: the
factors affecting physicians per-
formance are so numerous and
interdependent that no single vari-
able, such as sleep, can be under-
stood or targeted in isolation. Be-
cause of the unknown real-life
consequences of such myriad in-
teractions, no drug would be ap-
proved solely on the basis of lab-
oratory evidence. Yet we require
neither consideration of complex-
ity nor rigorous studies before
implementing policies with simi-
larly broad implications. Why?
Bioethicist and legal scholar
Michelle Meyer has described our
tendency to view a field experi-
ment designed to study the ef-
fects of an existing or proposed
practice as more morally suspi-
cious than an immediate, univer-
sal implementation of an untest-
ed practice. She argues that
people in power often rely on in-
tuition in creating and imple-
menting wide-reaching policies.
Indeed, neither residents nor pa-
tients consented to the ACGME
rules, yet no one finds this omis-
sion ethically suspect. Moreover,
intuition seems particularly sa-
lient to debates over duty hours,
since everyone knows how it
feels to be tired. Unfortunately,
few people know how it feels to
see a patient through illness,
spend a fifth of your time en-
gaged in hand-offs, leave halfway
through an operation because
your shifts up, or perceive resent-
ment in your supervisors who
think you have it easier than they
did. Given such trade-offs and
uncertainties, its not just ethical
but laudable to comparatively
evaluate duty-hours policies. The
question then becomes: Can the
research be accomplished if con-
sent is required?
The Facebook experiments re-
sults would have been invalid had
consent been sought, since we
couldnt determine how much
users adjusted their emotional
content because they knew it was
being monitored. Similarly, requir-
ing residents consent in duty-
hour trials would render the re-
sults uninterpretable, given the
selection bias that would be intro-
duced if those preferring longer
hours were more likely to par-
ticipate.
The challenges with regard to
patients are more pragmatic. Con-
sider, for instance, caring for a
man with a myocardial infarc-
tion. After obtaining his consent
for percutaneous coronary inter-
vention, youd have to add, I also
need your consent to be cared for
by residents who are working
longer hours. If he said no,
would you have to transfer him,
as heart muscle continued to die,
to a nonteaching hospital? Surely
here the risk posed by seeking
consent is greater than that from
the research itself.
Moreover, as we examine the
implications for efforts to develop
learning health systems, a corol-
lary of this hypothetical situation
is worth considering. Imagine tell-
ing a patient, I need your per-
mission to care for you at a hos-
pital where were using a new
electronic health record, are bas-
ing your doctors reimbursement

702 n engl j med 374;8 nejm.org February 25, 2016

Back to Table of Contents

 13 Notable Articles of 2016
PER S PE C T IV E
on whether you stay healthy, and
are under pressure to discharge
you quickly and make sure you
dont come back. We dont really
know how all this will affect your
health, but we believe its for the
better. Can you sign here?
The point is that our approach
to human-subjects research per-
petuates a misleading distinction
between risks posed by research
and those posed by practice, de-
manding greater scrutiny for in-
vestigative efforts while assum-
ing that untested practice is safe.
In describing this phenomenon,
Meyer cites the moratorium that
the OHRP imposed on a study
assessing a checklist designed
to reduce catheter-related blood-
stream infections because re-
searchers hadnt obtained physi-
cians or patients consent. The
OHRP explained that its regula-
tions dont apply when institu-
tions are merely implementing
practices aiming to improve care,
but if theyre planning research
activities examining the effective-
ness of interventions to improve
the quality of care, then the reg-
ulatory protections are important
to protect the rights and welfare
of human research subjects. This
double standard leaves us, para-
doxically, with unregulated prac-
tices that may be ineffective and
unsafe because we cant surmount
the regulatory hurdles to conduct-
ing research to improve them.
Pharmaceutical Policy Reform
To address this problem, we
n engl j med 374;8
must understand the values of the
people were professing to pro-
tect. In one relevant study, Hal-
pern and colleagues asked patients
undergoing dialysis to imagine
two hypothetical scenarios.5 In
the research scenario, patients
in a trial are randomly assigned
to a prespecified dialysis dura-
tion of 4.5 hours or a duration at
the physicians discretion (both
approaches are within the stan-
dard of care). In the clinical
care scenario, patients receive di-
alysis for a duration determined
by a protocol (also common prac-
tice). Participants were more will-
ing in the research than the
practice setting to give up their
own decision-making autonomy,
including written informed con-
sent. They recognized the value
of research and didnt perceive
the hypothetical study as posing
higher risk than ordinary care.
But they expressed deep reserva-
tions about compromising physi-
cians autonomy to individualize
treatment absent compelling rea-
sons for doing so.
This last finding highlights
the ultimate irony of both duty-
hour restrictions and objections
to studying them: weve created
an educational system that com-
promises trainees freedom to
judge for themselves when their
patients need them. The value
that physicians and patients place
on such autonomy is not measur-
able in mortality rates or hours
nejm.org February 25, 2016
Back to Table of Contents
nejm.org
Leaping without Looking
slept but should remain foremost
in our discussions. An essential
contribution of the duty-hour
trials is that, in assessing flexi-
bility itself, they remind us that
autonomy is an ethical concept
that matters to both doctors and
patients in research and in
practice.
Disclosure forms provided by the author
are available with the full text of this article
at NEJM.org.
Dr. Rosenbaum is a national correspondent
for the Journal.
This article was published on February 2,
2016, and updated on February 4, 2016, at
NEJM.org.
1. Meyer MN. Two cheers for corporate ex-
perimentation: the A/B illusion and the vir-
tues of data-driven innovation. Colo Tech L J
2015;13(2):273.
2. Kramer ADI, Guillory JE, Hancock JT.
Experimental evidence of massive-scale emo-
tional contagion through social networks.
Proc Natl Acad Sci U S A 2014;111:8788-90.
3. Carome MA, Wolfe SM, Almashat S,
Hall DV. Letter to Jerry Menikoff, director,
and Kristina Borror, director, Division of
Compliance Oversight, Office for Human
Research Protections, Department of Health
and Human Services, regarding iCOMPARE
trial. November 19, 2015 (http://www.citizen
.org/documents/2283.pdf).
4. Carome MA, Wolfe SM, Almashat S,
Hall DV. Letter to Jerry Menikoff, director,
and Kristina Borror, director, Division of
Compliance Oversight, Office for Human
Research Protections, Department of Health
and Human Services, regarding FIRST trial.
November 19, 2015 (http://www.citizen.org/
documents/2284.pdf).
5. Kraybill A, Dember LM, Joffe S, et al.
Patient and physician views about proto-
colized dialysis treatment in randomized tri-
als and clinical care. AJOB Empirical Bioeth-
ics 2015 October 23 (Epub ahead of print).
DOI: 10.1056/NEJMp1600233
Copyright 2016 Massachusetts Medical Society.
Leaping without Looking
703
14 Notable Articles of 2016 nejm.org

new england
The
journal of medicine
established in 1812 March 3, 2016 vol. 374 no. 9

Randomized Trial of Labor Induction in Women

35 Years of Age or Older
Kate F. Walker, M.R.C.O.G., George J. Bugg, M.D., Marion Macpherson, M.D., Carol McCormick, M.Sc.,
Nicky Grace, M.A., Chris Wildsmith, B.A., Lucy Bradshaw, M.Sc., Gordon C.S. Smith, D.Sc.,
and James G. Thornton, M.D., for the 35/39 Trial Group*

a bs t r ac t

BACKGROUND
The risk of antepartum stillbirth at term is higher among women 35 years of age or From the Division of Child Health, Ob-
older than among younger women. Labor induction may reduce the risk of stillbirth, stetrics and Gynaecology, School of Clin-
ical Sciences (K.F.W., M.M., C.M., J.G.T.),
but it also may increase the risk of cesarean delivery, which already is common in and Nottingham Clinical Trials Unit (L.B.),
this older age group. and the University of Nottingham, the
Division of Obstetrics and Gynaecology,
METHODS Nottingham University Hospitals NHS
Trust (G.J.B., N.G.), Nottingham, Still-
We conducted a randomized, controlled trial involving primigravid women who were birth and Neonatal Death Charity, Lon-
35 years of age or older. Women were randomly assigned to labor induction between don (C.W.), and the Department of Ob-
39 weeks 0 days and 39 weeks 6 days of gestation or to expectant management (i.e., stetrics and Gynaecology and National
Institute for Health Research Biomedical
waiting until the spontaneous onset of labor or until the development of a medical Research Centre, Cambridge University,
problem that mandated induction). The primary outcome was cesarean delivery. The Cambridge (G.C.S.S.) all in the United
trial was not designed or powered to assess the effects of labor induction on stillbirth. Kingdom. Address reprint requests to Dr.
Thornton at the Division of Child Health,
RESULTS Obstetrics and Gynaecology, School of
Medicine, University of Nottingham,
A total of 619 women underwent randomization. In an intention-to-treat analysis, Hucknall Rd., Nottingham NG5 1PB,
there were no significant between-group differences in the percentage of women who United Kingdom, or at jim.thornton@
underwent a cesarean section (98 of 304 women in the induction group [32%] and nottingham.ac.uk.

103 of 314 women in the expectant-management group [33%]; relative risk, 0.99; 95% * A complete list of investigators in the
confidence interval [CI], 0.87 to 1.14) or in the percentage of women who had a 35/39 Trial Group is provided in the
Supplementary Appendix, available with
vaginal delivery with the use of forceps or vacuum (115 of 304 women [38%] and 104 the full text of this article at NEJM.org.
of 314 women [33%], respectively; relative risk, 1.30; 95% CI, 0.96 to 1.77). There were
N Engl J Med 2016;374:813-22.
no maternal or infant deaths and no significant between-group differences in the DOI: 10.1056/NEJMoa1509117
womens experience of childbirth or in the frequency of adverse maternal or neonatal Copyright 2016 Massachusetts Medical Society.

outcomes.
CONCLUSIONS Read Full Article at NEJM.org
Among women of advanced maternal age, induction of labor at 39 weeks of gestation,
as compared with expectant management, had no significant effect on the rate of
cesarean section and no adverse short-term effects on maternal or neonatal out-
comes. (Funded by the Research for Patient Benefit Programme of the National Insti-
tute for Health Research; Current Controlled Trials number, ISRCTN11517275.)

n engl j med 374;9 nejm.org March 3, 2016 813

Back to Table of Contents

 15 Notable Articles of 2016 nejm.org
The n e w e ng l a n d j o u r na l of m e dic i n e

Edi t or i a l

Induction of Labor and Cesarean Delivery

William A. Grobman, M.D.

At the heart of obstetrical care is a seemingly
simple calculus: when are the benefits of deliv-
ery greater than the benefits of continued preg-
nancy? However, making this determination is
anything but straightforward, given the poten-
tially conflicting needs of the mother and the
needs of her offspring, which must both be
taken into account to maximize maternal and
perinatal health.
In the absence of maternal or fetal complica-
tions, current consensus favors the consideration
of delivery between 41 weeks 0 days and 42 weeks
0 days of gestation. In addition, for these women,
delivery is recommended after 42 weeks 0 days
and no later than 42 weeks 6 days of gestation,
given the increase in perinatal morbidity and
mortality at these gestational ages.1 Thus, induc-
tion before 41 weeks 0 days of gestation in the
absence of complications is considered not to be
medically indicated.
One consideration that traditionally has tipped
the balance toward continuing pregnancy is the
concern that labor induction may increase the
risk of cesarean delivery, particularly among
nulliparous women. This belief is based on the
findings of multiple observational studies in
which outcomes in women who underwent in-
duction were compared with those of women
who had spontaneous labor.2 However, sponta-
neous labor is not a clinical strategy, and thus
it is not the appropriate comparison.
Observational studies in which outcomes in
women who underwent induction were compared
with those in women who received expectant
management generally have not shown an in-
creased risk of cesarean delivery among women
who underwent induction.3-6 However, trials that
have explored whether, in the absence of compli-
cations, labor induction before 41 weeks 0 days
of gestation is associated with adverse maternal
or perinatal outcomes have been too small to
guide clinical practice.7
In this issue of the Journal, Walker et al.8 have
attempted to rectify this gap in evidence. They
report the results of a trial in which more than
600 women who were at least 35 years of age
were randomly assigned to labor induction be-
tween 39 weeks 0 days and 39 weeks 6 days of
gestation or to expectant management. This
study was powered to detect at least a 36% rela-
tive difference between the two groups in the
frequency of cesarean delivery. A total of 32%
of the women assigned to the induction group,
as compared with 33% of the women assigned to
the expectant-management group, underwent a
cesarean delivery (relative risk, 0.99; 95% confi-
dence interval, 0.87 to 1.14). There were no sig-
nificant differences between the groups in other
adverse maternal or perinatal outcomes, but such
outcomes were uncommon.
On the basis of the results of this trial, it
would be premature to alter recommendations
regarding the timing of delivery in uncomplicated
pregnancies. Although the study did not show
evidence of harm from induction at 39 weeks
of gestation, it also did not show evidence of
benefit, and one could argue that medical inter-
ventions in general, and intervention in the
natural progress of gestation specifically, should
be performed only when benefit has been shown.
Because this trial was not designed or ade-
quately powered to assess differences in perinatal
outcomes, whether labor induction at 39 weeks of
gestation affects these outcomes remains un-
known. We do not know whether the findings of
this trial are generalizable to women younger
than 35 years of age or whether the results
would differ according to whether or not women
require cervical ripening. Finally, women in this
trial received care in the United Kingdom, which

880 n engl j med 374;9 nejm.org March 3, 2016

Back to Table of Contents

16 Notable Articles of 2016
Editorial
has a health delivery system that differs from
that of the United States in ways that could af-
fect the relationship between labor induction
and cesarean delivery; these differences include
a higher rate of operative vaginal delivery in the
United Kingdom.9
The authors note the need for a larger trial
to test the effects of induction on stillbirth and
uncommon adverse neonatal outcomes. I am the
principal investigator of such a trial (Clinical-
Trials.gov number, NCT01990612), which is
currently under way within the MaternalFetal
Medicine Units Network of the Eunice Kennedy
Shriver National Institute of Child Health and
Human Development. This trial, which has a
targeted enrollment of 6000 women, is designed
to identify differences in perinatal outcomes
among nulliparous women with uncomplicated
singleton pregnancies who are randomly as-
signed to induction between 39 weeks 0 days
and 39 weeks 4 days of gestation or to expectant
management. The trial is more than halfway
complete.
Although the trial by Walker et al. was not
designed to assess the effect of labor induction
on stillbirth and adverse neonatal outcomes, it
makes an important contribution to medical
knowledge. It is the largest trial of its type to be
completed, and it suggests that a belief that
guides decisions about the timing of delivery
namely, that induction of labor at term increases
the risk of cesarean delivery may not be true
after all.
n engl j med 374;9
Back to Table of Contents
nejm.org
Disclosure forms provided by the author are available with the
full text of this article at NEJM.org.
From the Northwestern University Feinberg School of Medi-
cine, Chicago.
1. American College of Obstetricians and Gynecologists. Prac-
tice bulletin no. 146: Management of late-term and postterm
pregnancies. Obstet Gynecol 2014;124:390-6.
2. Maternal and neonatal outcomes of elective induction of
labor. Rockville, MD: Agency for Healthcare Research and
Quality, October 2014 (http://www.ahrq.gov/research/findings/
evidence-based-reports/eiltp.html).
3. Osmundson S, Ou-Yang RJ, Grobman WA. Elective induction
compared with expectant management in nulliparous women
with an unfavorable cervix. Obstet Gynecol 2011;117:583-7.
4. Gibson KS, Waters TP, Bailit JL. Maternal and neonatal out-
comes in electively induced low-risk term pregnancies. Am J
Obstet Gynecol 2014;211(3):249.e1-249.e16.
5. Stock SJ, Ferguson E, Duffy A, Ford I, Chalmers J, Norman
JE. Outcomes of elective induction of labour compared with ex-
pectant management: population based study. BMJ 2012;344:
e2838.
6. Cheng YW, Kaimal AJ, Snowden JM, Nicholson JM, Caughey
AB. Induction of labor compared to expectant management in
low-risk women and associated perinatal outcomes. Am J Obstet
Gynecol 2012;207(6):502.e1-8.
7. Saccone G, Berghella V. Induction of labor at full term in
uncomplicated singleton gestations: a systematic review and
metaanalysis of randomized controlled trials. Am J Obstet Gy-
necol 2015;213:629-36.
8. Walker KF, Bugg GJ, Macpherson M, et al. Randomized trial
of labor induction in women 35 years of age or older. N Engl J
Med 2016;374:813-22.
9. Kyser KL, Lu X, Santillan D, et al. Forceps delivery volumes
in teaching and nonteaching hospitals: are volumes sufficient
for physicians to acquire and maintain competence? Acad Med
2014;89:71-6.
DOI: 10.1056/NEJMe1516461
Copyright 2016 Massachusetts Medical Society.
nejm.org March 3, 2016 881
17 Notable Articles of 2016 nejm.org

new england
The
journal of medicine
established in 1812 March 17, 2016 vol. 374 no. 11

Randomized Trial of Stent versus Surgery for Asymptomatic

Carotid Stenosis
Kenneth Rosenfield, M.D., M.H.C.D.S., Jon S. Matsumura, M.D., Seemant Chaturvedi, M.D., Tom Riles, M.D.,
Gary M. Ansel, M.D., D. Chris Metzger, M.D., Lawrence Wechsler, M.D., Michael R. Jaff, D.O.,
and William Gray, M.D., for the ACT I Investigators*

a bs t r ac t

BACKGROUND
Previous clinical trials have suggested that carotid-artery stenting with a device to From Massachusetts General Hospital,
capture and remove emboli (embolic protection) is an effective alternative to Boston (K.R., M.R.J.); the University of
Wisconsin, Madison (J.S.M.); the Univer
carotid endarterectomy in patients at average or high risk for surgical complications. sity of Miami, Miami (S.C.); NYU Lan
gone School of Medicine, New York
METHODS (T.R.); Ohio Health System, Columbus
In this trial, we compared carotid-artery stenting with embolic protection and ca- (G.M.A.); Wellmont Cardiovascular As
rotid endarterectomy in patients 79 years of age or younger who had severe carotid sociates Heart Institute, Kingsport, TN
(D.C.M.); the University of Pittsburgh
stenosis and were asymptomatic (i.e., had not had a stroke, transient ischemic attack, Medical Center, Pittsburgh (L.W.); and
or amaurosis fugax in the 180 days before enrollment) and were not considered to the Main Line Health System, Philadel
be at high risk for surgical complications. The trial was designed to enroll 1658 pa- phia (W.G.). Address reprint requests to
Dr. Rosenfield at the Division of Cardi
tients but was halted early, after 1453 patients underwent randomization, because of ology, Massachusetts General Hospital,
slow enrollment. Patients were followed for up to 5 years. The primary composite 55 Fruit St., Boston, MA 02114.
end point of death, stroke, or myocardial infarction within 30 days after the pro- * A complete list of the Asymptomatic
cedure or ipsilateral stroke within 1 year was tested at a noninferiority margin of Carotid Trial (ACT) I Investigators is pro
3 percentage points. vided in the Supplementary Appendix,
available at NEJM.org.
RESULTS Drs. Rosenfield and Matsumura contrib
Stenting was noninferior to endarterectomy with regard to the primary composite end uted equally to this article.
point (event rate, 3.8% and 3.4%, respectively; P = 0.01 for noninferiority). The rate of This article was published on February 17,
stroke or death within 30 days was 2.9% in the stenting group and 1.7% in the endar- 2016, at NEJM.org.
terectomy group (P = 0.33). From 30 days to 5 years after the procedure, the rate of N Engl J Med 2016;374:1011-20.
freedom from ipsilateral stroke was 97.8% in the stenting group and 97.3% in the DOI: 10.1056/NEJMoa1515706
endarterectomy group (P = 0.51), and the overall survival rates were 87.1% and 89.4%, Copyright 2016 Massachusetts Medical Society.

respectively (P = 0.21). The cumulative 5-year rate of stroke-free survival was 93.1% in
the stenting group and 94.7% in the endarterectomy group (P = 0.44). Read Full Article at NEJM.org
CONCLUSIONS
In this trial involving asymptomatic patients with severe carotid stenosis who were not
at high risk for surgical complications, stenting was noninferior to endarterectomy
with regard to the rate of the primary composite end point at 1 year. In analyses that
included up to 5 years of follow-up, there were no significant differences between the
study groups in the rates of nonprocedure-related stroke, all stroke, and survival.
(Funded by Abbott Vascular; ACT I ClinicalTrials.gov number, NCT00106938.)

n engl j med 374;11 nejm.org March 17, 2016 1011

Back to Table of Contents

18 Notable Articles of 2016 nejm.org
The n e w e ng l a n d j o u r na l of m e dic i n e

Original Article

Long-Term Results of Stenting versus

Endarterectomy for Carotid-Artery Stenosis
Thomas G. Brott, M.D., George Howard, Dr.P.H., Gary S. Roubin, M.D., Ph.D.,
James F. Meschia, M.D., Ariane Mackey, M.D., William Brooks, M.D.,
Wesley S. Moore, M.D., Michael D. Hill, M.D., Vito A. Mantese, M.D.,
Wayne M. Clark, M.D., Carlos H. Timaran, M.D., Donald Heck, M.D.,
Pierre P. Leimgruber, M.D., Alice J. Sheffet, Ph.D., Virginia J. Howard, Ph.D.,
Seemant Chaturvedi, M.D., Brajesh K. Lal, M.D., Jenifer H. Voeks, Ph.D.,
and Robert W. Hobson II, M.D.,* for the CREST Investigators

A BS T R AC T

BACKGROUND
In the Carotid Revascularization Endarterectomy versus Stenting Trial, we found From the Mayo Clinic, Jacksonville, FL
no significant difference between the stenting group and the endarterectomy (T.G.B., J.F.M.); the University of Alabama
at Birmingham (G.H., V.J.H.) and Cardiovas-
group with respect to the primary composite end point of stroke, myocardial in- cular Associates of the Southeast (G.S.R.)
farction, or death during the periprocedural period or any subsequent ipsilateral both in Birmingham; Centre Hospitalier
stroke during 4 years of follow-up. We now extend the results to 10 years. Universitaire de QubecUniversit Laval,
Quebec, QC (A.M.), and the University of
Calgary, Calgary, AB (M.D.H.) both in
METHODS Canada; Baptist Health Lexington, KY (W.B.);
Among patients with carotid-artery stenosis who had been randomly assigned to the University of California, Los Angeles, Los
stenting or endarterectomy, we evaluated outcomes every 6 months for up to 10 years Angeles (W.S.M.); Mercy Hospital St. Louis,
St. Louis (V.A.M.); Oregon Health and Sci-
at 117 centers. In addition to assessing the primary composite end point, we ence University, Portland (W.M.C.); the Uni-
assessed the primary end point for the long-term extension study, which was versity of Texas Southwestern Medical Cen-
ipsilateral stroke after the periprocedural period. ter, Dallas (C.H.T); Novant Health Clinical
Research, Winston-Salem, NC (D.H.); the
RESULTS Providence Sacred Heart Medical Center and
Childrens Hospital, Spokane, WA (P.P.L.);
Among 2502 patients, there was no significant difference in the rate of the pri- Rutgers New Jersey Medical School, Newark
mary composite end point between the stenting group (11.8%; 95% confidence (A.J.S., R.W.H.); the University of Miami
interval [CI], 9.1 to 14.8) and the endarterectomy group (9.9%; 95% CI, 7.9 to 12.2) Miller School of Medicine, Miami (S.C.); the
University of Maryland Medical Center, Balti-
over 10 years of follow-up (hazard ratio, 1.10; 95% CI, 0.83 to 1.44). With respect more (B.K.L.); and the Medical University of
to the primary long-term end point, postprocedural ipsilateral stroke over the South Carolina, Charleston (J.H.V.). Address
10-year follow-up occurred in 6.9% (95% CI, 4.4 to 9.7) of the patients in the stent- reprint requests to Dr. Brott at the Mayo
Clinic, Griffin Bldg. 1st Fl., Rm. 170, 4500
ing group and in 5.6% (95% CI, 3.7 to 7.6) of those in the endarterectomy group; San Pablo Rd. S, Jacksonville, FL 32224, or at
the rates did not differ significantly between the groups (hazard ratio, 0.99; 95% brott.thomas@mayo.edu.
CI, 0.64 to 1.52). No significant between-group differences with respect to either * Deceased.
end point were detected when symptomatic patients and asymptomatic patients
A complete list of investigators in the
were analyzed separately. Carotid Revascularization Endarterec-
tomy versus Stenting Trial (CREST) is
CONCLUSIONS provided in the Supplementary Appen-
Over 10 years of follow-up, we did not find a significant difference between patients dix, available at NEJM.org.
who underwent stenting and those who underwent endarterectomy with respect to This article was published on February 18,
the risk of periprocedural stroke, myocardial infarction, or death and subsequent 2016, at NEJM.org.
ipsilateral stroke. The rate of postprocedural ipsilateral stroke also did not differ N Engl J Med 2016;374:1021-31.
between groups. (Funded by the National Institutes of Health and Abbott Vascular DOI: 10.1056/NEJMoa1505215
Copyright 2016 Massachusetts Medical Society.
Solutions; CREST ClinicalTrials.gov number, NCT00004732.)

Read Full Article at NEJM.org

n engl j med 374;11 nejm.org March 17, 2016 1021

Back to Table of Contents

19 Notable Articles of 2016 nejm.org
The n e w e ng l a n d j o u r na l of m e dic i n e

Edi t or i a l

Endarterectomy, Stenting, or Neither for Asymptomatic

Carotid-Artery Stenosis
J. David Spence, M.D., and A. Ross Naylor, M.D.

Important data from two large, randomized
trials comparing early and late outcomes after
carotid endarterectomy and carotid-artery stent-
ing have now been published in the Journal.1,2 In
common with every other large, multicenter, ran-
domized trial to date, the Asymptomatic Carotid
Trial (ACT I) and the Carotid Revascularization
Endarterectomy versus Stenting Trial (CREST)
showed that after the perioperative period, there
was no difference in the rate of late ipsilateral
stroke after endarterectomy or stenting. In ACT I,
which included asymptomatic patients who were
deemed to be at average risk, the 5-year rate of
ipsilateral stroke (excluding the perioperative
period) was 2.2% after stenting (i.e., 0.4% per year)
and 2.7% after endarterectomy (0.5% per year).1
In CREST, which included symptomatic and
asymptomatic patients who were deemed to be
at average risk, the estimated 10-year rate of
ipsilateral stroke (excluding the perioperative
period) was 6.9% after stenting (i.e., 0.7% per year)
and 5.6% (0.6% per year) after endarterectomy.2
The fact that there is near-unanimous con-
sensus within randomized trials that after the
perioperative period the rates of late ipsilateral
stroke after stenting do not differ significantly
from those after endarterectomy should dispel
any lingering concerns about the durability of
stenting. That issue has now surely been resolved.
What has not been resolved, however, is the issue
of the generalizability of randomized-trial find-
ings into routine clinical practice, and, more
importantly, the vexed question of how best to
treat the asymptomatic patient. No one should
harbor any illusions that ACT I and CREST have
resolved the latter issue.
CREST and ACT I both used credentialing to
ensure that only the best interventionists and
surgeons performed stenting or endarterectomy
within the trials. The commendably low rates of
death and stroke during the procedure in ACT I
and CREST attest to this. It therefore remains to
be seen whether these findings can be translated
into routine clinical practice, if guidelines are
changed to further liberalize indications for
stenting, especially in asymptomatic patients.
This is an important point, because a recent
systematic review showed that 9 of 21 large ad-
ministrative data-set registries (43%) reported
rates of death and stroke in excess of the 3% risk
threshold that is recommended by the American
Heart Association in asymptomatic patients
undergoing stenting, as compared with 1 of 21
registries (5%) after endarterectomy.3 Further-
more, the 3% risk threshold is clearly too high,
given the reduction of risk with intensive medi-
cal therapy. Discrepancies between randomized-
trial data (i.e., from ACT I and CREST) and real-
world practice are nothing new and, in this case,
are probably attributable to the fact that many
real-world practitioners in the United States are
performing two or fewer procedures annually in
asymptomatic patients, with poorer outcomes
than their more experienced colleagues.4
The magnitude of the initial procedural risk
will ultimately determine whether endarterectomy
or stenting is preferable in recently symptomatic
patients, and this will be determined by recency
of symptoms, age of the patient, and coexisting
conditions. However, there is a major concern
that the data from these two trials will be un-
critically interpreted to mean that stenting is
equivalent to endarterectomy and so further ex-
acerbate the situation in the United States,
where more than 90% of carotid-artery interven-
tions are performed in asymptomatic patients,
even though evidence suggests that up to 90% of
them will undergo an ultimately unnecessary
and potentially harmful procedure.5,6 By contrast,
the percentage of interventions that are per-
formed for asymptomatic stenoses is approxi-
mately 60% in Germany and Italy, 15% in Canada

n engl j med 374;11 nejm.org March 17, 2016 1087

Back to Table of Contents

 20 Notable Articles of 2016
Editorial
and Australia, and 0% in Denmark.7 Such dis-
crepancies call into question the appropriateness
of advocating routine interventions for asymp-
tomatic carotid-artery stenosis.
The ACT I authors conceded that in hindsight
it would have been preferable to have included a
medical group in their trial.1 However, the de-
bate about how improvements in modern medi-
cal therapy may have lowered the annual risk of
stroke had not reached its zenith when ACT I
was conceived. It is certainly a highly topical and
controversial issue in the current era, because
data from both randomized trials and nonran-
domized studies suggest that the annual rate of
stroke among medically treated asymptomatic
patients has declined over the past two decades,
regardless of the severity of stenosis at baseline.8
Evidence now suggests that the annual rate of
ipsilateral stroke may be as low as 0.5 to 1%8
a rate that is very similar to that observed in
ACT I and CREST after successful stenting or
endarterectomy.1,2
Accordingly, contemporary guidelines, which
recommend that interventions may be appropri-
ate if they can be performed with a risk of less
than 3%, are based on historical data from ran-
domized trials that were completed decades ago
and that should now be considered obsolete.
Outside clinical trials, endarterectomy and stent-
ing should be reserved for patients with symp-
tomatic severe stenosis or for asymptomatic pa-
tients who are shown to be at higher risk for
stroke with medical therapy than with interven-
tion. Such patients (approximately 10 to 15% of
patients with asymptomatic stenosis of 70 to
99%) may be identified by an algorithm that
incorporates information about microemboli de-
tected by means of transcranial Doppler,6,9,10 and
in the future by imaging strategies that identify
the vulnerable plaque.11
It is hoped that the Carotid Revascularization
and Medical Management for Asymptomatic Ca-
rotid Stenosis Trial (CREST-2; ClinicalTrials.gov
number, NCT02089217), which includes a medi-
cal group, will help settle this issue. Unfortu-
nately, the Stent-Protected Angioplasty in Asymp-
tomatic Carotid Artery Stenosis vs. Endarterectomy
(SPACE-2; Current Controlled Trials number,
ISRCTN78592017) trial (which also had a third
group receiving medical therapy) has now been
1088 n engl j med 374;11
Back to Table of Contents
nejm.org
abandoned because of poor recruitment. Pend-
ing the completion of CREST-2, we think that it
would be desirable for interventionists and sur-
geons to forgo stenting and endarterectomy in
low-risk asymptomatic patients outside that trial.
This restraint would not only spare patients from
procedures that may be unnecessary, but it
should also facilitate early completion of the
trial (and so avoid the fate of SPACE-2), so that
it may be possible to identify which patients will
benefit from an intervention rather than medical
therapy alone in an evidence-based rather than
an eminence-based manner.
Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
From the Stroke Prevention and Atherosclerosis Research Cen-
tre, Robarts Research Institute, Western University, London, ON,
Canada (J.D.S.); and the Vascular Surgery Group, Division of
Cardiovascular Sciences, Leicester Royal Infirmary, Leicester,
United Kingdom (A.R.N.).
This article was published on February 18, 2016, at NEJM.org.
1. Rosenfield K, Matsumura JS, Chaturvedi S, et al. Random-
ized trial of stent versus surgery for asymptomatic carotid steno-
sis. N Engl J Med 2016;374:1011-20.
2. Brott TG, Howard G, Roubin GS, et al. Long-term results
of stenting versus endarterectomy for carotid-artery stenosis.
N Engl J Med 2016;374:1021-31.
3. Paraskevas KI, Kalmykov EL, Naylor AR. Stroke/death rates
following carotid artery stenting and carotid endarterectomy in
contemporary administrative dataset registries: a systematic re-
view. Eur J Vasc Endovasc Surg 2016;51:3-12.
4. Choi JC, Johnston SC, Kim AS. Early outcomes after carotid
artery stenting compared with endarterectomy for asymptom-
atic carotid stenosis. Stroke 2015;46:120-5.
5. Bogiatzi C, Cocker MS, Beanlands R, Spence JD. Identifying
high-risk asymptomatic carotid stenosis. Expert Opin Med Diagn
2012;6:139-51.
6. Spence JD, Tamayo A, Lownie SP, Ng WP, Ferguson GG.
Absence of microemboli on transcranial Doppler identifies low-
risk patients with asymptomatic carotid stenosis. Stroke 2005;
36:2373-8.
7. Vikatmaa P, Mitchell D, Jensen LP, et al. Variation in clinical
practice in carotid surgery in nine countries 2005-2010: lessons
from VASCUNET and recommendations for the future of na-
tional clinical audit. Eur J Vasc Endovasc Surg 2012;44:11-7.
8. Naylor AR. Why is the management of asymptomatic carotid
disease so controversial? Surgeon 2015;13:34-43.
9. Spence JD, Coates V, Li H, et al. Effects of intensive medical
therapy on microemboli and cardiovascular risk in asymptom-
atic carotid stenosis. Arch Neurol 2010;67:180-6.
10. Markus HS, King A, Shipley M, et al. Asymptomatic emboli-
sation for prediction of stroke in the Asymptomatic Carotid Em-
boli Study (ACES): a prospective observational study. Lancet
Neurol 2010;9:663-71.
11. Naylor AR, Schroeder TV, Sillesen H. Clinical and imaging
features associated with an increased risk of late stroke in pa-
tients with asymptomatic carotid disease. Eur J Vasc Endovasc
Surg 2014;48:633-40.
DOI: 10.1056/NEJMe1600123
Copyright 2016 Massachusetts Medical Society.
nejm.org March 17, 2016
21 Notable Articles of 2016 nejm.org

new england
The
journal of medicine
established in 1812 May 26, 2016 vol. 374 no. 21

Blood-Pressure Lowering in Intermediate-Risk Persons

without Cardiovascular Disease
Eva M. Lonn, M.D., Jackie Bosch, Ph.D., Patricio LpezJaramillo, M.D., Ph.D., Jun Zhu, M.D., Lisheng Liu, M.D.,
Prem Pais, M.D., Rafael Diaz, M.D., Denis Xavier, M.D., Karen Sliwa, M.D., Ph.D., Antonio Dans, M.D.,
Alvaro Avezum, M.D., Ph.D., Leopoldo S. Piegas, M.D., Ph.D., Katalin Keltai, M.D., Ph.D., Matyas Keltai, M.D., Ph.D.,
Irina Chazova, M.D., Ph.D., Ron J.G. Peters, M.D., Ph.D., Claes Held, M.D., Ph.D., Khalid Yusoff, M.D.,
Basil S. Lewis, M.D., Petr Jansky, M.D., Alexander Parkhomenko, M.D., Ph.D., Kamlesh Khunti, M.D., Ph.D.,
William D. Toff, M.D., Christopher M. Reid, Ph.D., John Varigos, B.Sc., Lawrence A. Leiter, M.D.,
Dora I. Molina, M.D., Robert McKelvie, M.D., Ph.D., Janice Pogue, Ph.D.,* Joanne Wilkinson, B.A.,
Hyejung Jung, M.Sc., Gilles Dagenais, M.D., and Salim Yusuf, M.B., B.S., D.Phil., for the HOPE3 Investigators

a bs t r ac t

BACKGROUND
Antihypertensive therapy reduces the risk of cardiovascular events among high-risk The authors affiliations are listed in the
persons and among those with a systolic blood pressure of 160 mm Hg or higher, Appendix. Address reprint requests to Dr.
Lonn at the Population Health Research
but its role in persons at intermediate risk and with lower blood pressure is unclear. Institute and Hamilton Health Sciences,
METHODS 237 Barton St. E., Hamilton ON L8L 2X2,
Canada, or at eva.lonn@phri.ca.
In one comparison from a 2-by-2 factorial trial, we randomly assigned 12,705 par-
ticipants at intermediate risk who did not have cardiovascular disease to receive * Deceased.
either candesartan at a dose of 16 mg per day plus hydrochlorothiazide at a dose A complete list of the Heart Outcomes
of 12.5 mg per day or placebo. The first coprimary outcome was the composite of Prevention Evaluation (HOPE)3 trial
investigators is provided in the Supple
death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke; mentary Appendix, available at NEJM.org.
the second coprimary outcome additionally included resuscitated cardiac arrest,
This article was published on April 2, 2016,
heart failure, and revascularization. The median follow-up was 5.6 years. at NEJM.org.
RESULTS
N Engl J Med 2016;374:2009-20.
The mean blood pressure of the participants at baseline was 138.1/81.9 mm Hg; DOI: 10.1056/NEJMoa1600175
the decrease in blood pressure was 6.0/3.0 mm Hg greater in the active-treatment Copyright 2016 Massachusetts Medical Society.
group than in the placebo group. The first coprimary outcome occurred in 260 par-
ticipants (4.1%) in the active-treatment group and in 279 (4.4%) in the placebo Read Full Article at NEJM.org
group (hazard ratio, 0.93; 95% confidence interval [CI], 0.79 to 1.10; P = 0.40); the
second coprimary outcome occurred in 312 participants (4.9%) and 328 partici-
pants (5.2%), respectively (hazard ratio, 0.95; 95% CI, 0.81 to 1.11; P = 0.51). In one
of the three prespecified hypothesis-based subgroups, participants in the subgroup
for the upper third of systolic blood pressure (>143.5 mm Hg) who were in the
active-treatment group had significantly lower rates of the first and second copri-
mary outcomes than those in the placebo group; effects were neutral in the middle
and lower thirds (P = 0.02 and P = 0.009, respectively, for trend in the two outcomes).
CONCLUSIONS
Therapy with candesartan at a dose of 16 mg per day plus hydrochlorothiazide at
a dose of 12.5 mg per day was not associated with a lower rate of major cardio-
vascular events than placebo among persons at intermediate risk who did not have
cardiovascular disease. (Funded by the Canadian Institutes of Health Research and
AstraZeneca; ClinicalTrials.gov number, NCT00468923.)

n engl j med 374;21 nejm.org May 26, 2016 2009

Back to Table of Contents

22 Notable Articles of 2016 nejm.org
The n e w e ng l a n d j o u r na l of m e dic i n e

Original Article

Cholesterol Lowering in Intermediate-Risk

Persons without Cardiovascular Disease
S. Yusuf, J. Bosch, G. Dagenais, J. Zhu, D. Xavier, L. Liu, P. Pais,
P. LpezJaramillo, L.A. Leiter, A. Dans, A. Avezum, L.S. Piegas, A. Parkhomenko,
K. Keltai, M. Keltai, K. Sliwa, R.J.G. Peters, C. Held, I. Chazova, K. Yusoff,
B.S. Lewis, P. Jansky, K. Khunti, W.D. Toff, C.M. Reid, J. Varigos,
G. SanchezVallejo, R. McKelvie, J. Pogue,* H. Jung, P. Gao, R. Diaz,
and E. Lonn, for the HOPE3 Investigators

A BS T R AC T

BACKGROUND
Previous trials have shown that the use of statins to lower cholesterol reduces the risk The authors full names, academic degrees,
of cardiovascular events among persons without cardiovascular disease. Those trials and affiliations are listed in the Appendix.
Address reprint requests to Dr. Yusuf at
have involved persons with elevated lipid levels or inflammatory markers and involved the Population Health Research Institute,
mainly white persons. It is unclear whether the benefits of statins can be extended to 237 Barton St. E., Hamilton, ON L8L 2X2,
an intermediate-risk, ethnically diverse population without cardiovascular disease. Canada, or at yusufs@mcmaster.ca.

*Deceased.
METHODS
In one comparison from a 2-by-2 factorial trial, we randomly assigned 12,705 partici- A complete list of the Heart Outcomes
Prevention Evaluation (HOPE)3 trial in
pants in 21 countries who did not have cardiovascular disease and were at intermedi- vestigators is provided in the Supplemen
ate risk to receive rosuvastatin at a dose of 10 mg per day or placebo. The first tary Appendix, available at NEJM.org.
coprimary outcome was the composite of death from cardiovascular causes, nonfatal This article was published on April 2, 2016,
myocardial infarction, or nonfatal stroke, and the second coprimary outcome addi- at NEJM.org.
tionally included revascularization, heart failure, and resuscitated cardiac arrest. The N Engl J Med 2016;374:2021-31.
median follow-up was 5.6 years. DOI: 10.1056/NEJMoa1600176
Copyright 2016 Massachusetts Medical Society.
RESULTS
The overall mean low-density lipoprotein cholesterol level was 26.5% lower in the
Read Full Article at NEJM.org
rosuvastatin group than in the placebo group. The first coprimary outcome occurred
in 235 participants (3.7%) in the rosuvastatin group and in 304 participants (4.8%) in
the placebo group (hazard ratio, 0.76; 95% confidence interval [CI], 0.64 to 0.91;
P = 0.002). The results for the second coprimary outcome were consistent with the re-
sults for the first (occurring in 277 participants [4.4%] in the rosuvastatin group and
in 363 participants [5.7%] in the placebo group; hazard ratio, 0.75; 95% CI, 0.64 to
0.88; P<0.001). The results were also consistent in subgroups defined according to
cardiovascular risk at baseline, lipid level, C-reactive protein level, blood pressure, and
race or ethnic group. In the rosuvastatin group, there was no excess of diabetes or
cancers, but there was an excess of cataract surgery (in 3.8% of the participants, vs.
3.1% in the placebo group; P = 0.02) and muscle symptoms (in 5.8% of the participants,
vs. 4.7% in the placebo group; P = 0.005).
CONCLUSIONS
Treatment with rosuvastatin at a dose of 10 mg per day resulted in a significantly
lower risk of cardiovascular events than placebo in an intermediate-risk, ethnically
diverse population without cardiovascular disease. (Funded by the Canadian Institutes
of Health Research and AstraZeneca; HOPE-3 ClinicalTrials.gov number, NCT00468923.)

n engl j med 374;21 nejm.org May 26, 2016 2021

Back to Table of Contents

23 Notable Articles of 2016 nejm.org
The n e w e ng l a n d j o u r na l of m e dic i n e

Original Article

Blood-Pressure and Cholesterol Lowering

in Persons without Cardiovascular Disease
Salim Yusuf, M.B., B.S., D.Phil., Eva Lonn, M.D., Prem Pais, M.D.,
Jackie Bosch, Ph.D., Patricio LpezJaramillo, M.D., Ph.D., Jun Zhu, M.D.,
Denis Xavier, M.D., Alvaro Avezum, M.D., Ph.D., Lawrence A. Leiter, M.D.,
Leopoldo S. Piegas, M.D., Ph.D., Alexander Parkhomenko, M.D., Ph.D.,
Matyas Keltai, M.D., Ph.D., Katalin Keltai, M.D., Ph.D., Karen Sliwa, M.D., Ph.D.,
Irina Chazova, M.D., Ph.D., Ron J.G. Peters, M.D., Ph.D., Claes Held, M.D., Ph.D.,
Khalid Yusoff, M.D., Basil S. Lewis, M.D., Petr Jansky, M.D.,
Kamlesh Khunti, M.D., Ph.D., William D. Toff, M.D., Christopher M. Reid, Ph.D.,
John Varigos, B.Sc., Jose L. Accini, M.D., Robert McKelvie, M.D., Ph.D.,
Janice Pogue, Ph.D.,* Hyejung Jung, M.Sc., Lisheng Liu, M.D., Rafael Diaz, M.D.,
Antonio Dans, M.D., and Gilles Dagenais, M.D., for the HOPE3 Investigators

A BS T R AC T

BACKGROUND
Elevated blood pressure and elevated low-density lipoprotein (LDL) cholesterol increase the The authors affiliations are listed in the
risk of cardiovascular disease. Lowering both should reduce the risk of cardiovascular Appendix. Address reprint requests to
Dr. Yusuf at the Population Health Re
events substantially. search Institute, 237 Barton St. E., Hamil
METHODS ton, ON L8L 2X2, Canada, or at yusufs@
mcmaster.ca.
In a trial with 2-by-2 factorial design, we randomly assigned 12,705 participants at inter-
mediate risk who did not have cardiovascular disease to rosuvastatin (10 mg per day) or *Deceased.
placebo and to candesartan (16 mg per day) plus hydrochlorothiazide (12.5 mg per day) A complete list of the Heart Outcomes
or placebo. In the analyses reported here, we compared the 3180 participants assigned to Prevention Evaluation (HOPE)3 trial in
vestigators is provided in the Supplemen
combined therapy (with rosuvastatin and the two antihypertensive agents) with the 3168 tary Appendix, available at NEJM.org.
participants assigned to dual placebo. The first coprimary outcome was the composite of
This article was published on April 2, 2016,
death from cardiovascular causes, nonfatal myocardial infarction, or nonfatal stroke, and at NEJM.org.
the second coprimary outcome additionally included heart failure, cardiac arrest, or revas-
N Engl J Med 2016;374:2032-43.
cularization. The median follow-up was 5.6 years. DOI: 10.1056/NEJMoa1600177
RESULTS Copyright 2016 Massachusetts Medical Society.

The decrease in the LDL cholesterol level was 33.7 mg per deciliter (0.87 mmol per liter)
greater in the combined-therapy group than in the dual-placebo group, and the decrease
in systolic blood pressure was 6.2 mm Hg greater with combined therapy than with dual Read Full Article at NEJM.org
placebo. The first coprimary outcome occurred in 113 participants (3.6%) in the com-
bined-therapy group and in 157 (5.0%) in the dual-placebo group (hazard ratio, 0.71; 95%
confidence interval [CI], 0.56 to 0.90; P = 0.005). The second coprimary outcome occurred
in 136 participants (4.3%) and 187 participants (5.9%), respectively (hazard ratio, 0.72;
95% CI, 0.57 to 0.89; P = 0.003). Muscle weakness and dizziness were more common in
the combined-therapy group than in the dual-placebo group, but the overall rate of dis-
continuation of the trial regimen was similar in the two groups.
CONCLUSIONS
The combination of rosuvastatin (10 mg per day), candesartan (16 mg per day), and hy-
drochlorothiazide (12.5 mg per day) was associated with a significantly lower rate of
cardiovascular events than dual placebo among persons at intermediate risk who did not
have cardiovascular disease. (Funded by the Canadian Institutes of Health Research and
AstraZeneca; ClinicalTrials.gov number, NCT00468923.)

n engl j med 374;21 nejm.org May 26, 2016 2032

Back to Table of Contents

24 Notable Articles of 2016 nejm.org
The n e w e ng l a n d j o u r na l of m e dic i n e

Edi t or i a l

More HOPE for Prevention with Statins

William C. Cushman, M.D., and David C. Goff, Jr., M.D., Ph.D.

In view of the worldwide burden of cardiovascu-
lar disease and the high cost of and poor adher-
ence to medication regimens for the prevention of
cardiovascular disease, the concept of a polypill
a single pill that combines several medica-
tions is an attractive public health approach.
However, evidence that each component of a
polypill would independently reduce the risk of
cardiovascular events and that the combination
of agents would be safe is lacking. The primary
results of the Heart Outcomes Prevention Evalu-
ation (HOPE)3 trial are now reported in three
articles in the Journal.1-3 HOPE-3 was a double-
blind, randomized, placebo-controlled trial with
a 2-by-2 factorial design, in which 12,705 inter-
mediate-risk men (55 years of age) and women
(60 years of age) who did not have cardiovascu-
lar disease were randomly assigned to receive
cholesterol-lowering treatment with rosuvastatin
at a dose of 10 mg per day or placebo and were
also randomly assigned to receive blood-pressure
lowering treatment with candesartan at a dose
of 16 mg per day plus hydrochlorothiazide at a
dose of 12.5 mg per day or placebo for a median
of 5.6 years. Treatment with rosuvastatin resulted
in a 24% lower risk of cardiovascular events than
that with placebo (absolute difference, 1.1 per-
centage points), but the antihypertensive therapy
did not result in a significantly lower risk of
cardiovascular events. The HOPE-3 trial provides
evidence to reinforce some current guideline
recommendations and to influence future guide-
lines.
The cholesterol-lowering component of the
trial1 produced results consistent with a meta-
analysis of randomized trials of statin therapy,
which showed that a reduction of 1 mmol per
liter in the low-density lipoprotein (LDL) choles-
terol level was associated with a 25% lower risk
of cardiovascular events in a primary-prevention
population.4 Furthermore, the rate of cardiovas-
cular events that was observed in the placebo
group (4.8% over a period of 5.6 years) was
within the range of the rates that were observed
among the lowest-risk groups shown to have a
benefit from statin therapy in the meta-analysis.
The trial participants who had high-sensitivity
C-reactive protein (CRP) levels higher than 2 mg
per deciliter and those who had levels lower than
2 mg per deciliter had similar rates of cardiovas-
cular events and a similar benefit from rosuva-
statin. Hence, these results support a risk-based
approach to statin use, which has been recom-
mended in recent guidelines,5 rather than an
approach that is based primarily on LDL choles-
terol levels, and the results add to the evidence
supporting statin use for primary prevention.
The blood-pressurelowering component of
the trial2 showed no significant benefit of anti-
hypertensive therapy in reducing the risk of
cardiovascular events. The observed difference
between the active-treatment group and the pla-
cebo group in the decrease in blood pressure
over the course of the trial (6.0/3.0 mm Hg) was
small, and the 95% confidence interval for the
estimated hazard ratio did not exclude the bene-
fit one might expect (on the basis of the results
from the meta-analysis) from this degree of
blood-pressure lowering.6 Neither of the drugs
for blood-pressure lowering that were used in
the trial have been shown to reduce the risk of
cardiovascular events at such low doses. If higher
doses had been used, the risk of cardiovascular
events might have been significantly reduced,
whether from greater blood-pressure lowering,
additional effects of the antihypertensive drugs,
or both. Hydrochlorothiazide, even at a dose of
25 mg per day, has been less effective in reduc-
ing the risk of cardiovascular events than has a
full dose of amlodipine,7 whereas chlorthalidone

n engl j med 374;21 nejm.org May 26, 2016 2085

Back to Table of Contents

 25 Notable Articles of 2016
The n e w e ng l a n d j o u r na l
at a dose of 25 mg per day has been effective in
reducing the risk of cardiovascular events in a
placebo-controlled trial8 and has been at least
as effective as amlodipine.9 These observations
suggest that the use of chlorthalidone could
have been more effective than the use of hydro-
chlorothiazide in HOPE-3.
The trial population was at a lower cardiovas-
cular risk than the populations in previous hyper-
tension trials. The observed rate of cardiovascu-
lar events in the dual-placebo group was 5.0%
over a period of 5.6 years. Since most previous
trials of blood-pressure lowering have used in-
clusion criteria that are designed to increase the
level of cardiovascular risk in order to increase
trial efficiency, those trials have included few
low-risk adults. Meta-analyses of such trials pro-
vide evidence of cardiovascular benefit from the
use of blood-pressurelowering medications in
adults with an average systolic blood pressure
higher than 130 mm Hg and either clinical car-
diovascular disease or a high cardiovascular risk
(defined as a 5-year risk of cardiovascular events
of 6.5%).6 In addition, the Systolic Blood Pres-
sure Intervention Trial (SPRINT) provides support
for the use of blood-pressurelowering medica-
tions in patients who do not have cardiovascular
disease but who have a systolic blood pressure
higher than 130 mm Hg; in SPRINT, the risks of
cardiovascular events and death from any cause
were significantly reduced with the use of regi-
mens for blood-pressure lowering that were more
intensive than the regimen used in this trial.10
However, the SPRINT participants who did not
have clinical cardiovascular disease at baseline
were required to have subclinical cardiovascular
disease or a 10-year cardiovascular risk (on the
basis of the Framingham risk score) that was
higher than 15%. The difference in systolic
blood pressure between the active-treatment and
control groups that was seen in SPRINT was
twice the difference seen in HOPE-3 because the
treatment regimen was more intensive.
The overall null results of the blood-pressure
lowering component of HOPE-3 could be due to
insufficient dosing of antihypertensive medica-
tions, treatment of a relatively low-risk group, or
chance. Setting aside the play of chance, we may
take from these results new insight regarding the
initiation threshold and treatment targets for
blood-pressurelowering medications. Although
no benefit of blood-pressure lowering was ob-
2086 n engl j med 374;21
Back to Table of Contents
nejm.org
of m e dic i n e
served overall, a prespecified subgroup analysis
showed a 27% lower risk of cardiovascular
events with blood-pressurelowering therapy in
the subgroup of participants who were in the
upper third of systolic blood pressure levels
(>143.5 mm Hg). Among the patients in that
subgroup who received placebo, the rate of car-
diovascular events was 6.5% over a period of
5.6 years. This rate is within the range of rates
reported in the previously mentioned meta-
analysis. However, the rates of cardiovascular
events in the subgroups of participants in the
lower and middle thirds of systolic blood pres-
sure levels who received placebo were lower than
the rates among the lowest-risk groups shown to
have benefit from blood-pressure lowering in
previous trials. Blood-pressurelowering treat-
ment with low doses of the two drugs used in
HOPE-3 may not be effective over the period
studied in this trial among patients with low
levels of systolic blood pressure and low levels of
cardiovascular risk. These results may help to
define the combined threshold of systolic blood
pressure (<140 mm Hg) and cardiovascular risk
(<5.0%) below which the use of blood-pressure
lowering medications may not be useful in the
short term. However, these results do not rule
out the possibility of a benefit with longer-term
treatment in a portion of this relatively low-risk
population.
The results of the comparison of the effects
of the combined intervention (rosuvastatin and
candesartan plus hydrochlorothiazide) with pla-
cebo3 generally agreed with the results for the
separate interventions. There was no evidence of
harm or synergy between the two interventions.
Although the addition of blood-pressure lower-
ing to rosuvastatin therapy appeared to provide
more benefit than that observed with rosuva-
statin alone in the subgroup of participants who
were in the upper third of systolic blood pres-
sure levels, the P value for interaction was not
significant.
The results of the HOPE-3 trial suggest that
rosuvastatin at a dose of 10 mg per day is more
effective in preventing cardiovascular events than
is candesartan at a dose of 16 mg per day plus
hydrochlorothiazide at a dose of 12.5 mg per day
in this relatively low-risk population. Although
these results do not exclude the possibility that
more effective therapy for blood-pressure lower-
ing might be beneficial in a relatively low-risk,
nejm.org May 26, 2016
26 Notable Articles of 2016
Editorials
older population, they provide support for the use
of statins as a safe and effective intervention to
prevent cardiovascular events in such patients.
The opinions expressed in this article do not necessarily repre-
sent the official views of the Department of Veterans Affairs or
the U.S. government.
Disclosure forms provided by the authors are available with
the full text of this article at NEJM.org.
From the Preventive Medicine Section, Veterans Affairs Medi-
cal Center, Memphis, TN (W.C.C.); and the Colorado School of
Public Health, University of Colorado Anschutz Medical Campus,
Aurora (D.C.G.).
This article was published on April 2, 2016, at NEJM.org.
1. Yusuf S, Bosch J, Dagenais G, et al. Cholesterol lowering in
intermediate-risk persons without cardiovascular disease. N Engl
J Med 2016;374:2021-31.
2. Lonn EM, Bosch J, Lpez-Jaramillo P, et al. Blood-pressure
lowering in intermediate-risk persons without cardiovascular
disease. N Engl J Med 2016;374:2009-20.
3. Yusuf S, Lonn E, Pais P, et al. Blood-pressure and cholesterol
lowering in persons without cardiovascular disease. N Engl J
Med 2016;374:2032-43.
4. Cholesterol Treatment Trialists (CTT) Collaborators. The
effects of lowering LDL cholesterol with statin therapy in people
n engl j med 374;21
Back to Table of Contents
nejm.org
at low risk of vascular disease: meta-analysis of individual data
from 27 randomised trials. Lancet 2012;380:581-90.
5. Stone NJ, Robinson JG, Lichtenstein AH, et al. 2013 ACC/
AHA guideline on the treatment of blood cholesterol to reduce
atherosclerotic cardiovascular risk in adults: a report of the Amer-
ican College of Cardiology/American Heart Association Task
Force on Practice Guidelines. Circulation 2014;129:Suppl 2:S1-45.
6. The Blood Pressure Lowering Treatment Trialists Collabo-
ration. Blood pressure-lowering treatment based on cardiovas-
cular risk: a meta-analysis of individual patient data. Lancet 2014;
384:591-8.
7. Jamerson K, Weber MA, Bakris GL, et al. Benazepril plus
amlodipine or hydrochlorothiazide for hypertension in high-risk
patients. N Engl J Med 2008;359:2417-28.
8. Prevention of stroke by antihypertensive drug treatment in
older persons with isolated systolic hypertension: final results
of the Systolic Hypertension in the Elderly Program (SHEP).
JAMA 1991;265:3255-64.
9. The ALLHAT Officers and Coordinators for the ALLHAT
Collaborative Research Group. The Antihypertensive and Lipid-
Lowering Treatment to Prevent Heart Attack Trial: major out-
comes in high-risk hypertensive patients randomized to angio-
tensin-converting enzyme inhibitor or calcium channel blocker
vs diuretic. JAMA 2002;288:2981-97.
10. The SPRINT Research Group. A randomized trial of inten-
sive versus standard blood-pressure control. N Engl J Med 2015;
373:2103-16.
DOI: 10.1056/NEJMe1603504
Copyright 2016 Massachusetts Medical Society.
nejm.org May 26, 2016 2087
27 Notable Articles of 2016 nejm.org

new england
The
journal of medicine
established in 1812 September 15, 2016 vol. 375 no. 11

Efficacy of the Herpes Zoster Subunit Vaccine in Adults

70 Years of Age or Older
A.L. Cunningham, H. Lal, M. Kovac, R. Chlibek, S.-J. Hwang, J. Dez-Domingo, O. Godeaux, M.J. Levin,
J.E. McElhaney, J. Puig-Barber, C. Vanden Abeele, T. Vesikari, D. Watanabe, T. Zahaf, A. Ahonen, E. Athan,
J.F. Barba-Gomez, L. Campora, F. de Looze, H.J. Downey, W. Ghesquiere, I. Gorfinkel, T. Korhonen, E. Leung,
S.A. McNeil, L. Oostvogels, L. Rombo, J. Smetana, L. Weckx, W. Yeo, and T.C. Heineman, for the ZOE-70 Study Group*

a bs t r ac t

BACKGROUND
A trial involving adults 50 years of age or older (ZOE-50) showed that the herpes zoster The authors full names, academic de-
subunit vaccine (HZ/su) containing recombinant varicellazoster virus glycoprotein E and grees, and affiliations are listed in the Ap-
pendix. Address reprint requests to Dr.
the AS01B adjuvant system was associated with a risk of herpes zoster that was 97.2% Heineman at Genocea Biosciences, Cam-
lower than that associated with placebo. A second trial was performed concurrently at bridge Discovery Park, 100 Acorn Park
the same sites and examined the safety and efficacy of HZ/su in adults 70 years of age Dr., 5th Floor, Cambridge, MA 02140, or
at thomas.heineman@genocea.com.
or older (ZOE-70).
* A complete list of investigators in the
METHODS Zoster Efficacy Study in Adults 70 Years
This randomized, placebo-controlled, phase 3 trial was conducted in 18 countries and of Age or Older (ZOE-70) Study Group is
provided in the Supplementary Appen-
involved adults 70 years of age or older. Participants received two doses of HZ/su or dix, available at NEJM.org.
placebo (assigned in a 1:1 ratio) administered intramuscularly 2 months apart. Vaccine
Drs. Cunningham and Lal contributed
efficacy against herpes zoster and postherpetic neuralgia was assessed in participants equally to this article. Authors from Dr.
from ZOE-70 and in participants pooled from ZOE-70 and ZOE-50. Godeaux to Dr. Zahaf (listed alphabeti-
cally) also contributed equally, as did au-
RESULTS thors from Dr. Ahonen to Dr. Yeo (listed
In ZOE-70, 13,900 participants who could be evaluated (mean age, 75.6 years) received alphabetically).
either HZ/su (6950 participants) or placebo (6950 participants). During a mean follow-up N Engl J Med 2016;375:1019-32.
period of 3.7 years, herpes zoster occurred in 23 HZ/su recipients and in 223 placebo DOI: 10.1056/NEJMoa1603800
recipients (0.9 vs. 9.2 per 1000 person-years). Vaccine efficacy against herpes zoster was Copyright 2016 Massachusetts Medical Society.

89.8% (95% confidence interval [CI], 84.2 to 93.7; P<0.001) and was similar in partici-
pants 70 to 79 years of age (90.0%) and participants 80 years of age or older (89.1%). In Read Full Article at NEJM.org
pooled analyses of data from participants 70 years of age or older in ZOE-50 and ZOE-70
(16,596 participants), vaccine efficacy against herpes zoster was 91.3% (95% CI, 86.8 to
94.5; P<0.001), and vaccine efficacy against postherpetic neuralgia was 88.8% (95% CI,
68.7 to 97.1; P<0.001). Solicited reports of injection-site and systemic reactions within
7 days after injection were more frequent among HZ/su recipients than among placebo
recipients (79.0% vs. 29.5%). Serious adverse events, potential immune-mediated dis-
eases, and deaths occurred with similar frequencies in the two study groups.
CONCLUSIONS
In our trial, HZ/su was found to reduce the risks of herpes zoster and postherpetic neu-
ralgia among adults 70 years of age or older. (Funded by GlaxoSmithKline Biologicals;
ZOE-50 and ZOE-70 ClinicalTrials.gov numbers, NCT01165177 and NCT01165229.)

n engl j med 375;11 nejm.org September 15, 2016 1019

Back to Table of Contents

28 Notable Articles of 2016 nejm.org
The n e w e ng l a n d j o u r na l of m e dic i n e

Edi t or i a l

Preventing Shingles and Its Complications in Older Persons

Kathleen M. Neuzil, M.D., M.P.H., and Marie R. Griffin, M.D., M.P.H.

In the United States each year, herpes zoster, or
shingles, develops in half a million people 60
years of age or older. Although the symptoms
are often mild in younger persons, the risk for
serious complications of herpes zoster, including
postherpetic neuralgia, ocular involvement, and
central nervous system disease, increases with
advancing age.1 The rising age-specific incidence
of shingles and the aging population in the
United States are likely to contribute to additional
shingles-associated morbidity in coming years.2
The prevention of herpes zoster and its compli-
cations in older persons will improve quality of
life and should be a public health priority.
Since 2008, the U.S. Advisory Committee on
Immunization Practices has recommended that
all immunocompetent persons 60 years of age
or older receive a single dose of a live attenu-
ated herpes zoster vaccine (Zostavax).1 In a large,
placebo-controlled trial, the efficacy of this
vaccine against herpes zoster was 51.3%, and
the efficacy against postherpetic neuralgia was
66.5%.3 Further follow-up of participants and
postmarketing studies have confirmed the effec-
tiveness of the vaccine and have documented
declines in the efficacy of the vaccine over time.4
Among 176,078 members of Kaiser Permanente
who were 60 years of age or older and matched
controls, the effectiveness of the live attenuated
vaccine against herpes zoster decreased from
68.7% (95% confidence interval [CI], 66.3 to
70.9) in the first year after vaccination to 4.2%
(95% CI, 24.0 to 25.9) in the eighth year.5 These
data, coupled with information on the immuno-
genicity of booster doses of the vaccine, will in-
form recommendations regarding the need for a
subsequent dose or doses.6
In this issue of the Journal, Cunningham et al.
report on the efficacy of two doses of an investi-
gational, adjuvanted herpes zoster subunit vaccine
(HZ/su) in immunocompetent persons 70 years
of age or older.7 This trial, involving 13,900 per-
sons, was conducted concurrently with a previ-
ously reported trial involving persons 50 years
or age or older in which the same vaccine and
schedule were used.8 The vaccine contains a
recombinant varicellazoster virus (VZV) glyco-
protein E with a novel adjuvant (AS01B) designed
to improve CD4+ T-cellmediated immune re-
sponses, which are thought to be important in
preventing the reactivation of latent VZV. A lower
dose of this adjuvant is used in a malaria vaccine
that was approved in 2015 by the European
Medicines Agency for children living in areas in
which malaria is endemic.
In 2015, the efficacy of HZ/su against herpes
zoster was reported as 97.2% (95% CI, 93.7 to
99.0) among participants 50 years of age or
older and as 97.9% (95% CI, 87.9 to 100.0)
among participants 70 years of age or older dur-
ing a mean follow-up period of 3.2 years.8 In the
current trial, during a mean follow-up period of
3.7 years, the efficacy against herpes zoster was
89.8% (95% CI, 84.2 to 93.7) in persons 70 years
of age or older. Efficacy was similar among par-
ticipants who were 70 to 79 years of age and those
who were 80 years of age or older, and it was
maintained for the duration of the trial. For the
outcome of postherpetic neuralgia, the investiga-
tors included the participants who were 70 years
of age or older from both trials, and they report
an efficacy of 88.8% (95% CI, 68.7 to 97.1).
Given the limited efficacy and duration of
Zostavax, newer vaccine formulations with im-
proved efficacy are welcome. Although the higher
point estimates of efficacy with the HZ/su vaccine
are encouraging, the direct comparison of results
from different trials is problematic. For example,
in the pivotal trial evaluating Zostavax, the inci-
dence of postherpetic neuralgia in the control
group was higher than that in the control group
in the HZ/su trial, which may indicate that the

n engl j med 375;11 nejm.org September 15, 2016 1079

Back to Table of Contents

 29 Notable Articles of 2016
Editorial
Zostavax efficacy trial included a more frail
population, more active surveillance, or the use
of a more sensitive case definition. A major benefit
of the HZ/su vaccine as compared with Zostavax
appears to be retention of high efficacy against
herpes zoster and postherpetic neuralgia in the
oldest age groups and over time. Continued fol-
low-up of the vaccinated cohorts is warranted.
Although the safety profile regarding serious
adverse events reported in the trials of HZ/su
was reassuring, a full understanding of less
common serious side effects will be known only
as larger and more diverse populations are vac-
cinated. This is particularly pertinent given the
new adjuvant included in this vaccine. It is worth
noting that the short-term reactogenicity with
this adjuvanted vaccine is higher than with other
adult vaccines. In the first 7 days after vaccina-
tion, 79.0% of vaccine recipients, versus 29.5% of
placebo recipients, reported local or systemic re-
actions, and 11.9% of vaccine recipients, versus
2.0% of placebo recipients, reported that their
reactions were severe enough to prevent normal
activity. It is remarkable that few participants
declined the second injection, but whether ad-
herence would be similar in a different popula-
tion, especially one that included more frail
older adults, is unknown.
Policy deliberations regarding the HZ/su vac-
cine will need to include consideration of how
these trial data will translate into routine condi-
tions of use. The HZ/su trials reported data on
participants who received two doses of vaccine
therefore, the efficacy of a single dose, or of
two doses given on a different schedule, is not
known. Persons with a history of herpes zoster
or of herpes zoster vaccination were excluded
from these trials, so the benefit of the vaccine in
those populations is uncertain. Ultimately, HZ/su
may provide an option for immunocompromised
persons who are at high risk for herpes zoster
and its complications and are unable to receive
the live attenuated vaccine. This would be a ma-
jor advance in efforts to prevent herpes zoster.
Despite the 2008 recommendations for the
zoster vaccine, by 2014 only 27.9% of adults 60
years of age or older reported being vaccinated.9
In the early years after vaccine approval, supply
1080 n engl j med 375;11
Back to Table of Contents
nejm.org
constraints limited uptake. In more recent years,
the supply has been sufficient, and the reasons
for the continued poor uptake include provider
challenges (e.g., cost, storage of the frozen formu-
lation, and complex Medicare reimbursement),
limited public awareness of the disease and vac-
cine, a lack of requirements for adult vaccina-
tion, and the focus on acute medical care over
prevention among practitioners caring for adult
patients.10 Although HZ/su may address some of
these issues, such as easier storage requirements
for a nonreplicating product, it will have its own
challenges, including the two-dose schedule and
the higher reactogenicity. Thus, the full public
health value of herpes zoster vaccines will not be
realized unless we identify and address barriers
to delivery and uptake.
Disclosure forms provided by the authors are available with
the full text of this editorial at NEJM.org.
From the Center for Vaccine Development, University of Mary-
land School of Medicine, Baltimore (K.M.N.); and the Depart-
ments of Health Policy and Medicine, Vanderbilt University
Medical Center, Nashville (M.R.G.).
1. Harpaz R, Ortega-Sanchez IR, Seward JF. Prevention of herpes
zoster: recommendations of the Advisory Committee on Immuni-
zation Practices (ACIP). MMWR Recomm Rep 2008;57(RR-5):1-30.
2. Kawai K, Yawn BP, Wollan P, Harpaz R. Increasing incidence
of herpes zoster over a 60-year period from a population-based
study. Clin Infect Dis 2016;63:221-6.
3. Oxman MN, Levin MJ, Johnson GR, et al. A vaccine to pre-
vent herpes zoster and postherpetic neuralgia in older adults.
N Engl J Med 2005;352:2271-84.
4. Morrison VA, Johnson GR, Schmader KE, et al. Long-term
persistence of zoster vaccine efficacy. Clin Infect Dis 2015;60:
900-9.
5. Tseng HF, Harpaz R, Luo Y, et al. Declining effectiveness of
herpes zoster vaccine in adults aged 60 years. J Infect Dis 2016;
213:1872-5.
6. Levin MJ, Schmader KE, Pang L, et al. Cellular and humoral
responses to a second dose of herpes zoster vaccine adminis-
tered 10 years after the first dose among older adults. J Infect Dis
2016;213:14-22.
7. Cunningham AL, Lal H, Kovac M, et al. Efficacy of the her-
pes zoster subunit vaccine in adults 70 years of age or older.
N Engl J Med 2016;375:1019-32.
8. Lal H, Cunningham AL, Godeaux O, et al. Efficacy of an
adjuvanted herpes zoster subunit vaccine in older adults. N Engl
J Med 2015;372:2087-96.
9. Williams WW, Lu PJ, OHalloran A, et al. Surveillance of
vaccination coverage among adult populations United States,
2014. MMWR Surveill Summ 2016;65:1-36.
10. Hurley LP, Bridges CB, Harpaz R, et al. Physician attitudes
toward adult vaccines and other preventive practices, United
States, 2012. Public Health Rep 2016;131:320-30.
DOI: 10.1056/NEJMe1610652
Copyright 2016 Massachusetts Medical Society.
nejm.org September 15, 2016
30 Notable Articles of 2016 nejm.org

new england
The
journal of medicine
established in 1812 October 13, 2016 vol. 375 no. 15

10-Year Outcomes after Monitoring, Surgery, or Radiotherapy

for Localized Prostate Cancer
F.C. Hamdy, J.L. Donovan, J.A. Lane, M. Mason, C. Metcalfe, P. Holding, M. Davis, T.J. Peters, E.L. Turner,
R.M. Martin, J. Oxley, M. Robinson, J. Staffurth, E. Walsh, P. Bollina, J. Catto, A. Doble, A. Doherty, D. Gillatt,
R. Kockelbergh, H. Kynaston, A. Paul, P. Powell, S. Prescott, D.J. Rosario, E. Rowe, and D.E. Neal,
for the ProtecT Study Group*

a bs t r ac t

BACKGROUND
The comparative effectiveness of treatments for prostate cancer that is detected by prostate- The authors full names, academic de
specific antigen (PSA) testing remains uncertain. grees, and affiliations are listed in the
Appendix. Address reprint requests to
METHODS Dr. Hamdy at the Nuffield Department of
We compared active monitoring, radical prostatectomy, and external-beam radiotherapy for Surgical Sciences, University of Oxford,
Old Road Campus Research Bldg., Roose
the treatment of clinically localized prostate cancer. Between 1999 and 2009, a total of velt Dr., Headington, Oxford OX3 7DQ,
82,429 men 50 to 69 years of age received a PSA test; 2664 received a diagnosis of localized United Kingdom, or at freddie.hamdy@
prostate cancer, and 1643 agreed to undergo randomization to active monitoring (545 men), nds.ox.ac.uk.

surgery (553), or radiotherapy (545). The primary outcome was prostate-cancer mortality * A complete list of investigators in the
at a median of 10 years of follow-up. Secondary outcomes included the rates of disease Prostate Testing for Cancer and Treat
ment (ProtecT) trial is provided in the
progression, metastases, and all-cause deaths. Supplementary Appendix, available at
NEJM.org.
RESULTS
There were 17 prostate-cancerspecific deaths overall: 8 in the active-monitoring group (1.5 Drs. Hamdy, Donovan, Lane, and Neal
contributed equally to this article.
deaths per 1000 person-years; 95% confidence interval [CI], 0.7 to 3.0), 5 in the surgery
group (0.9 per 1000 person-years; 95% CI, 0.4 to 2.2), and 4 in the radiotherapy group This article was published on September
14, 2016, at NEJM.org.
(0.7 per 1000 person-years; 95% CI, 0.3 to 2.0); the difference among the groups was not
significant (P = 0.48 for the overall comparison). In addition, no significant difference was N Engl J Med 2016;375:1415-24.
DOI: 10.1056/NEJMoa1606220
seen among the groups in the number of deaths from any cause (169 deaths overall; P = 0.87 Copyright 2016 Massachusetts Medical Society.
for the comparison among the three groups). Metastases developed in more men in the
active-monitoring group (33 men; 6.3 events per 1000 person-years; 95% CI, 4.5 to 8.8)
than in the surgery group (13 men; 2.4 per 1000 person-years; 95% CI, 1.4 to 4.2) or the Read Full Article at NEJM.org
radiotherapy group (16 men; 3.0 per 1000 person-years; 95% CI, 1.9 to 4.9) (P = 0.004 for
the overall comparison). Higher rates of disease progression were seen in the active-mon-
itoring group (112 men; 22.9 events per 1000 person-years; 95% CI, 19.0 to 27.5) than in
the surgery group (46 men; 8.9 events per 1000 person-years; 95% CI, 6.7 to 11.9) or the
radiotherapy group (46 men; 9.0 events per 1000 person-years; 95% CI, 6.7 to 12.0) (P<0.001
for the overall comparison).
CONCLUSIONS
At a median of 10 years, prostate-cancerspecific mortality was low irrespective of the treat-
ment assigned, with no significant difference among treatments. Surgery and radiotherapy
were associated with lower incidences of disease progression and metastases than was
active monitoring. (Funded by the National Institute for Health Research; ProtecT Current
Controlled Trials number, ISRCTN20141297; ClinicalTrials.gov number, NCT02044172.)

n engl j med 375;15 nejm.org October 13, 2016 1415

Back to Table of Contents

31 Notable Articles of 2016 nejm.org
The n e w e ng l a n d j o u r na l of m e dic i n e

Original Article

Patient-Reported Outcomes after Monitoring,

Surgery, or Radiotherapy for Prostate Cancer
J.L. Donovan, F.C. Hamdy, J.A. Lane, M. Mason, C. Metcalfe, E. Walsh,
J.M. Blazeby, T.J. Peters, P. Holding, S. Bonnington, T. Lennon, L. Bradshaw,
D. Cooper, P. Herbert, J. Howson, A. Jones, N. Lyons, E. Salter, P. Thompson,
S. Tidball, J. Blaikie, C. Gray, P. Bollina, J. Catto, A. Doble, A. Doherty, D. Gillatt,
R. Kockelbergh, H. Kynaston, A. Paul, P. Powell, S. Prescott, D.J. Rosario, E. Rowe,
M. Davis, E.L. Turner, R.M. Martin, and D.E. Neal, for the ProtecT Study Group*

A BS T R AC T

BACKGROUND
Robust data on patient-reported outcome measures comparing treatments for clinically The authors full names, academic de-
localized prostate cancer are lacking. We investigated the effects of active monitoring, grees, and affiliations are listed in the
Appendix. Address reprint requests to
radical prostatectomy, and radical radiotherapy with hormones on patient-reported outcomes. Dr. Donovan at the School of Social and
Community Medicine, University of Bris-
METHODS tol, Canynge Hall, 39 Whatley Rd., Bristol
We compared patient-reported outcomes among 1643 men in the Prostate Testing for BS8 2PS, United Kingdom, or at jenny
Cancer and Treatment (ProtecT) trial who completed questionnaires before diagnosis, .donovan@bristol.ac.uk.
at 6 and 12 months after randomization, and annually thereafter. Patients completed * A complete list of investigators in the
validated measures that assessed urinary, bowel, and sexual function and specific ef- Prostate Testing for Cancer and Treat-
ment (ProtecT) Study Group is provided
fects on quality of life, anxiety and depression, and general health. Cancer-related qual- in the Supplementary Appendix, avail-
ity of life was assessed at 5 years. Complete 6-year data were analyzed according to the able at NEJM.org.
intention-to-treat principle. Drs. Donovan, Hamdy, Lane, and Neal
contributed equally to this article.
RESULTS
The rate of questionnaire completion during follow-up was higher than 85% for most This article was published on September
14, 2016, at NEJM.org.
measures. Of the three treatments, prostatectomy had the greatest negative effect on
sexual function and urinary continence, and although there was some recovery, these N Engl J Med 2016;375:1425-37.
DOI: 10.1056/NEJMoa1606221
outcomes remained worse in the prostatectomy group than in the other groups Copyright 2016 Massachusetts Medical Society.
throughout the trial. The negative effect of radiotherapy on sexual function was greatest
at 6 months, but sexual function then recovered somewhat and was stable thereafter;
Read Full Article at NEJM.org
radiotherapy had little effect on urinary continence. Sexual and urinary function de-
clined gradually in the active-monitoring group. Bowel function was worse in the radio-
therapy group at 6 months than in the other groups but then recovered somewhat,
except for the increasing frequency of bloody stools; bowel function was unchanged
in the other groups. Urinary voiding and nocturia were worse in the radiotherapy
group at 6 months but then mostly recovered and were similar to the other groups
after 12 months. Effects on quality of life mirrored the reported changes in function.
No significant differences were observed among the groups in measures of anxiety,
depression, or general health-related or cancer-related quality of life.
CONCLUSIONS
In this analysis of patient-reported outcomes after treatment for localized prostate cancer,
patterns of severity, recovery, and decline in urinary, bowel, and sexual function and as-
sociated quality of life differed among the three groups. (Funded by the U.K. National
Institute for Health Research Health Technology Assessment Program; ProtecT Current
Controlled Trials number, ISRCTN20141297; ClinicalTrials.gov number, NCT02044172.)

n engl j med 375;15 nejm.org October 13, 2016 1425

Back to Table of Contents

 32 Notable Articles of 2016
The n e w e ng l a n d j o u r na l
Edi t or i a l s
Treatment or Monitoring for Early Prostate Cancer
Anthony V. DAmico, M.D., Ph.D.
The best initial approach to early (low-risk or
intermediate-risk)1 prostate cancer remains un-
known. Specifically, does active monitoring with
the use of prostate-specific antigen (PSA) testing
as opposed to treatment lead to increased metas-
tasis and death from prostate cancer? If yes, then
which treatment, radical prostatectomy or radia-
tion with or without short-term (3 to 6 months)
androgen-suppression therapy, minimizes metas-
tasis and death from prostate cancer?
Hamdy and colleagues now report in the Jour-
nal the results of a randomized comparison of
three of these four approaches after a median
follow-up of 10 years,2 and Donovan and col-
leagues present data on patient-reported health-
related quality of life at 6 years of follow-up.3
Men were screened with PSA testing and pre-
sented at a median age of 62 years with favorable
clinical characteristics: 76% had stage T1c (PSA-
detected) disease, 77% and 21% had tumors with
Gleason scores of 6 and 7, respectively (on a
scale from 6 to 10, with higher scores indicating
a worse prognosis), and the median PSA level
was 4.6 ng per milliliter. Although a median
follow-up of 10 years was too short to evaluate
the primary outcome of prostate-cancer mortal-
ity in this favorable cohort (death from prostate
cancer occurred in 8 of the 545 men assigned to
active monitoring, 5 of the 553 men assigned
to surgery, and 4 of the 545 men assigned to ra-
diotherapy), it was adequate to evaluate the sec-
ondary outcome of the incidence of metastatic
disease, defined as bony, visceral, or lymph-node
metastasis on imaging or a PSA level above 100 ng
per milliliter.
Several important observations were made.
First, men assigned to active monitoring were
significantly more likely to have metastatic dis-
ease than those assigned to treatment (P = 0.004
for the overall comparison), with an incidence
that was more than twice as high (6.3 per 1000
person-years vs. 2.4 to 3.0 per 1000 person-years).
There was also a trend toward decreased death
1482 n engl j med 375;15
Back to Table of Contents
nejm.org
of m e dic i n e
from prostate cancer among men assigned to
surgery (hazard ratio, 0.63; 95% confidence inter-
val [CI], 0.21 to 1.93) or radiation and androgen-
deprivation therapy (hazard ratio, 0.51; 95% CI,
0.15 to 1.69) versus active monitoring. Although
further follow-up will determine whether these
trends become significant, causality between an
increase in metastatic disease and the use of
active monitoring versus treatment was estab-
lished. The clinical significance of this finding
is that with the use of active monitoring, more
men will have metastasis and the side effects of
salvage treatment (meaning at least lifelong in-
termittent androgen-deprivation therapy), which
are not inconsequential.4
Second, within the prerandomization stratum
of age, a near-significant interaction (P = 0.09 for
interaction) was observed given that men 65 years
of age or older were more likely to die from
prostate cancer if assigned to active monitoring
than if assigned to treatment. This finding prob-
ably reflects the fact that advancing age is associ-
ated with higher-grade disease than disease iden-
tified at an initial biopsy5 owing to sampling
error, resulting in undergrading, the risk of which
rises with the increasing prostate-gland volume6
that occurs with advancing age.7 Should the in-
teraction between age and death from prostate
cancer among men assigned to treatment versus
monitoring become significant, it would support
recommending treatment as opposed to moni-
toring to otherwise healthy men 65 years of age
or older with early prostate cancer who today are
increasingly being placed on active surveillance,8
given that the reduction in death from prostate
cancer (hazard ratio, 0.63; 95% CI, 0.36 to 1.09)
in the Prostate Cancer Intervention versus Ob-
servation Trial (PIVOT) only trended toward
significance (P = 0.09).9 However, the increasing
use of surveillance is already of potential con-
cern, considering that men enrolled in PIVOT had
a shorter life expectancy owing to coexisting
disease than men of similar age entered into
nejm.org October 13, 2016
33 Notable Articles of 2016
Editorials
the Surveillance, Epidemiology, and End Results
database.10
Finally, a trend favoring radiation and short-
course androgen-deprivation therapy over surgery
was observed. Specifically, the point estimate for
the hazard ratio for death from prostate cancer
when comparing these two treatments was 0.80
(95% CI, 0.22 to 2.99). If this trend becomes
significant, then one may consider radiation and
androgen-deprivation therapy as a preferred op-
tion for otherwise healthy men 65 years of age
or older with early prostate cancer for whom
treatment as compared with monitoring may be
more effective (P = 0.09 for interaction) in reduc-
ing death from prostate cancer.
For today, we can conclude on the basis of
level 1 evidence2 that PSA monitoring, as com-
pared with treatment of early prostate cancer,
leads to increased metastasis. Therefore, if a
man wishes to avoid metastatic prostate cancer
and the side effects of its treatment,3 monitoring
should be considered only if he has life-shorten-
ing coexisting disease such that his life expec-
tancy is less than the 10-year median follow-up
of the current study.2 In addition, given no sig-
nificant difference in death due to prostate can-
cer with surgery versus radiation and short-course
androgen-deprivation therapy, men with low-risk
or intermediate-risk1 prostate cancer should feel
free to select a treatment approach using the
data on health-related quality of life3 and with-
out fear of possibly selecting a less effective can-
cer therapy.
n engl j med 375;15
Back to Table of Contents
nejm.org
Disclosure forms provided by the author are available with the
full text of this editorial at NEJM.org.
From the Department of Radiation Oncology, Brigham and
Womens Hospital and DanaFarber Cancer Institute, Boston.
This editorial was published on September 14, 2016, at NEJM.org.
1. Mohler JL, Armstrong AJ, Bahnson RR, et al. Prostate can-
cer, version 1.2016. J Natl Compr Canc Netw 2016;14:19-30.
2. Hamdy FC, Donovan JL, Lane JA, et al. 10-year outcomes
after monitoring, surgery, or radiotherapy for localized prostate
cancer. N Engl J Med 2016;375:1415-24.
3. Donovan JL, Hamdy FC, Lane JA, et al. Patient-reported out-
comes after monitoring, surgery, or radiotherapy for prostate
cancer. N Engl J Med 2016;375:1425-37.
4. Sharifi N, Gulley JL, Dahut WL. Androgen deprivation ther-
apy for prostate cancer. JAMA 2005;294:238-44.
5. Anderson CB, Sternberg IA, Karen-Paz G, et al. Age is asso-
ciated with upgrading at confirmatory biopsy among men with
prostate cancer treated with active surveillance. J Urol 2015;194:
1607-11.
6. Moussa AS, Kattan MW, Berglund R, Yu C, Fareed K, Jones
JS. A nomogram for predicting upgrading in patients with low-
and intermediate-grade prostate cancer in the era of extended
prostate sampling. BJU Int 2010;105:352-8.
7. Berry SJ, Coffey DS, Walsh PC, Ewing LL. The development
of human benign prostatic hyperplasia with age. J Urol 1984;
132:474-9.
8. Cooperberg MR, Carroll PR. Trends in management for pa-
tients with localized prostate cancer, 1990-2013. JAMA 2015;
314:80-2.
9. Wilt TJ, Brawer MK, Jones KM, et al. Radical prostatectomy
versus observation for localized prostate cancer. N Engl J Med
2012;367:203-13.
10. Aizer AA, Chen MH, Hattangadi J, DAmico AV. Initial man-
agement of prostate-specific antigen-detected, low-risk prostate
cancer and the risk of death from prostate cancer. BJU Int 2014;
113:43-50.
DOI: 10.1056/NEJMe1610395
Copyright 2016 Massachusetts Medical Society.
nejm.org October 13, 2016 1483
34 Notable Articles of 2016 nejm.org

new england
The
journal of medicine
established in 1812 October 20, 2016 vol. 375 no. 16

GuillainBarr Syndrome Associated with Zika Virus Infection

in Colombia
Beatriz Parra, Ph.D., Jairo Lizarazo, M.D., Jorge A. Jimnez-Arango, M.D., Andrs F. Zea-Vera, M.D., Ph.D.,
Guillermo Gonzlez-Manrique, M.D., Jos Vargas, M.D., Jorge A. Angarita, M.D., Gonzalo Zuiga, M.D.,
Reydmar Lopez-Gonzalez, M.D., Cindy L. Beltran, M.D., Karen H. Rizcala, M.D., Maria T. Morales, M.D.,
Oscar Pacheco, M.D., Martha L. Ospina, M.D., Anupama Kumar, M.B., B.S., David R. Cornblath, M.D.,
Laura S. Muoz, M.D., Lyda Osorio, M.D., Ph.D., Paula Barreras, M.D., and Carlos A. Pardo, M.D.

a bs t r ac t

BACKGROUND
Zika virus (ZIKV) infection has been linked to the GuillainBarr syndrome. From No- From the Department of Microbiology
vember 2015 through March 2016, clusters of cases of the GuillainBarr syndrome were (B.P., A.F.Z.-V.), the Department of Inter-
nal Medicine, Hospital Universitario del
observed during the outbreak of ZIKV infection in Colombia. We characterized the clini- Valle (A.F.Z.-V., G.Z.), and Escuela de
cal features of cases of GuillainBarr syndrome in the context of this ZIKV infection Salud Publica (L.O.), Universidad del Valle,
outbreak and investigated their relationship with ZIKV infection. Cali, Hospital Universitario Erasmo Meoz,
Universidad de Pamplona, Cucuta (J.L.),
METHODS Universidad de Antioquia, Clinica Leon
XIII, Neuroclinica, Medellin (J.A.J.-A.,
A total of 68 patients with the GuillainBarr syndrome at six Colombian hospitals were R.L.-G.), Universidad Surcolombiana, Hos-
evaluated clinically, and virologic studies were completed for 42 of the patients. We per- pital Universitario de Neiva (G.G.-M.,
formed reverse-transcriptasepolymerase-chain-reaction (RT-PCR) assays for ZIKV in blood, C.L.B.), and Clinica Medilaser (J.A.A.),
Neiva, Clinica La Misericordia Internacio-
cerebrospinal fluid, and urine, as well as antiflavivirus antibody assays. nal, Barranquilla (J.V., K.H.R., M.T.M.),
and Instituto Nacional de Salud, Bogota
RESULTS (O.P., M.L.O.) all in Colombia; and the
A total of 66 patients (97%) had symptoms compatible with ZIKV infection before the Departments of Neurology (A.K., D.R.C.,
onset of the GuillainBarr syndrome. The median period between the onset of symp- L.S.M., P.B., C.A.P.) and Pathology (C.A.P.),
Johns Hopkins University School of Med-
toms of ZIKV infection and symptoms of the GuillainBarr syndrome was 7 days (inter- icine, Baltimore. Address reprint requests
quartile range, 3 to 10). Among the 68 patients with the GuillainBarr syndrome, 50% to Dr. Pardo at Johns Hopkins University
were found to have bilateral facial paralysis on examination. Among 46 patients in whom School of Medicine, 600 N. Wolfe St., 627
Pathology Bldg., Baltimore, MD 21287, or
nerve-conduction studies and electromyography were performed, the results in 36 pa- at cpardov1@jhmi.edu.
tients (78%) were consistent with the acute inflammatory demyelinating polyneuropathy
This article was published on October 5,
subtype of the GuillainBarr syndrome. Among the 42 patients who had samples tested 2016, at NEJM.org.
for ZIKV by RT-PCR, the results were positive in 17 patients (40%). Most of the positive
N Engl J Med 2016;375:1513-23.
RT-PCR results were in urine samples (in 16 of the 17 patients with positive RT-PCR re- DOI: 10.1056/NEJMoa1605564
sults), although 3 samples of cerebrospinal fluid were also positive. In 18 of 42 patients Copyright 2016 Massachusetts Medical Society.
(43%) with the GuillainBarr syndrome who underwent laboratory testing, the presence
of ZIKV infection was supported by clinical and immunologic findings. In 20 of these 42
Read Full Article at NEJM.org
patients (48%), the GuillainBarr syndrome had a parainfectious onset. All patients
tested were negative for dengue virus infection as assessed by RT-PCR.
CONCLUSIONS
The evidence of ZIKV infection documented by RT-PCR among patients with the Guillain
Barr syndrome during the outbreak of ZIKV infection in Colombia lends support to the
role of the infection in the development of the GuillainBarr syndrome. (Funded by the
Bart McLean Fund for Neuroimmunology Research and others.)
n engl j med 375;16 nejm.org October 20, 2016 1513

Back to Table of Contents

35 Notable Articles of 2016 nejm.org
The n e w e ng l a n d j o u r na l of m e dic i n e

Edi t or i a l s

Zika Getting on Your Nerves? The Association

with the GuillainBarr Syndrome
Jennifer A. Frontera, M.D., and Ivan R.F. da Silva, M.D., Ph.D.

Parra and colleagues1 report in the Journal the
results of a prospective study of 68 Colombian
patients who had a syndrome consistent with
the GuillainBarr syndrome, 66 of whom had
previously had symptoms of Zika virus (ZIKV)
infection. Major strengths of this study include
the documentation of a temporal relationship
between the GuillainBarr syndrome and ZIKV
infection (marked by a substantial increase in
the incidence of the GuillainBarr syndrome
after the introduction of ZIKV, from 20 to 90
cases per month throughout Colombia), the cri-
teria applied for the diagnosis of the Guillain
Barr syndrome, and the molecular and sero-
logic flavivirus data from analyses of serum,
cerebrospinal fluid (CSF), and urine.
However, the difficulties related to diagnos-
ing ZIKV infection are multifold. First, the symp-
toms associated with ZIKV infection are similar
to those caused by dengue virus (DENV) and
chikungunya virus, both of which are endemic in
Colombia. Second, the serologic cross-reactivity
among flaviviruses (including yellow fever virus,
West Nile virus, DENV, and Japanese encephalitis
virus) have been well described.2 Although the
Centers for Disease Control and Prevention
(CDC) recommends neutralizing antibody testing
with a plaque-reduction neutralization test to
distinguish among flaviviruses,3 this testing is
expensive, requires cell culture, and is also sus-
ceptible to cross-reactivity.4 Polymerase-chain-
reaction (PCR) testing can definitively identify
ZIKV, but molecular studies of serum are usu-
ally sensitive only during the first week after
infection. Because the GuillainBarr syndrome
has been linked to microbial pathogens through
a molecular mimicry mechanism, it is typically
diagnosed 1 week or longer after an infection.
Indeed, Parra et al. observed that the median
time to onset of the GuillainBarr syndrome
was 7 days after ZIKV infection.
The authors deal with these diagnostic dilem-
mas by showing that ZIKV PCR testing of other
body fluids (particularly urine) may remain sen-
sitive for a longer duration than does testing of
serum. Indeed, in 13 patients, ZIKV PCR results
were positive only in urine, whereas serum, CSF,
or both were PCR-negative when tested in a
similar time frame. IgM antibody testing of CSF
for both ZIKV and DENV may be another diag-
nostic strategy, since the IgM pentamer is too
large to cross the bloodbrain barrier.5 There-
fore, CSF that is positive for ZIKV IgM and
negative for DENV IgM would be suggestive of a
primary central nervous system ZIKV infection. Of
the patients who tested positive for ZIKV by PCR
and underwent CSF IgM testing, 8 were PCR-
positive but ZIKV IgMnegative in CSF, which
suggested that ZIKV PCR testing of urine may be
more sensitive than serologic testing of CSF.
The difficulties in diagnosing ZIKV infection
are borne out in this study, as only 17 patients
had definitive laboratory evidence of recent ZIKV
infection. On the basis of Table S5 in the Parra
et al. Supplementary Appendix, of these 17 pa-
tients, only 14 had electrophysiological data con-
sistent with the GuillainBarr syndrome and
therefore could have met Brighton level 1 diag-
nostic criteria for the syndrome, although the
actual number of patients meeting level 1 criteria
may have been smaller because we do not know
the corresponding results of CSF testing for
these patients.6 Because of these limitations in
diagnostic certainty for both ZIKV infection and
the GuillainBarr syndrome, a strong associa-
tion was identified in approximately 20% of pa-

n engl j med 375;16 nejm.org October 20, 2016 1581

Back to Table of Contents

 36 Notable Articles of 2016
The n e w e ng l a n d j o u r na l
tients in this cohort (14 of 68). Among the 25
ZIKV PCRnegative patients, DENV IgG antibod-
ies were present in the CSF of 12 patients and in
the serum of 10 patients, and serum DENV IgM
test results were positive in 1. These data raise
the possibility of primary DENV infection and
false positive ZIKV serologic test results due to
cross-reactivity. In addition, data on yellow fever
vaccination or infection were not provided; yel-
low fever is also endemic in much of Colombia
and may complicate the interpretation of the
ZIKV serologic results.
As is true with most clinical studies, proving
a causal relationship between ZIKV infection and
the GuillainBarr syndrome is challenging. In
keeping with Hill criteria for causality,7 the au-
thors show a consistent, specific, temporal rela-
tionship, which is analogous to relationships
between ZIKV infection and the GuillainBarr
syndrome observed in other countries.8,9 What is
more difficult to demonstrate is pathophysiolog-
ical plausibility. The authors point out that 20 pa-
tients had neurologic symptoms immediately
after the viral syndrome (only 9 of 20 had defi-
nite laboratory-proven ZIKV) and speculate that
other mechanisms, including a hyperacute im-
mune response or direct viral neuropathic mech-
anisms, may be in effect, rather than postinfec-
tious molecular mimicry. Although studies using
human neural progenitor cells have shown that
ZIKV infection increases cell death and dysregu-
lates cell-cycle progression,10 evidence of direct
neurotropism in adult neuronal cells is still lack-
ing. A recent study showed that there is a high
peptide overlap between the ZIKV polyprotein
and human proteins related to myelin and axons,
which suggests that an immune-mediated mech-
anism may be more likely.11 Although protein
epitopes and antibodies that are normally in-
volved in the genesis of the GuillainBarr syn-
drome seem not to be highly involved in one
cohort with ZIKV-associated acute motor axonal
neuropathy,8 it is possible that differences in sub-
types of the GuillainBarr syndrome and host
genetic factors may lead to varying immune-
mediated mechanisms in different populations.
1582 n engl j med 375;16
Back to Table of Contents
nejm.org
of m e dic i n e
Overall, the study by Parra and colleagues
supports the association between ZIKV and the
GuillainBarr syndrome, although confirmation
in another cohort would strengthen this asser-
tion. Although high rates of seropositivity may
prove protective against further waves of ZIKV-
related GuillainBarr syndrome in Central and
South America, the ZIKV pandemic is just begin-
ning in North America and Africa, and an in-
crease in the incidence of the GuillainBarr
syndrome may follow.
Disclosure forms provided by the authors are available with
the full text of this editorial at NEJM.org.
From the Cerebrovascular Center of the Neurological Institute,
Cleveland Clinic, Cleveland (J.A.F.); and the Neurology Depart-
ment, Universidade Federal Fluminense, Niteroi (I.R.F.S.), and
the Neurocritical Care Department, Americas Medical City, Rio
de Janeiro (I.R.F.S.) both in Brazil.
This article was published on October 5, 2016, at NEJM.org.
1. Parra B, Lizarazo J, Jimnez-Arango JA, et al. GuillainBarr
syndrome associated with Zika virus infection in Colombia.
N Engl J Med 2016;375:1513-23.
2. Mansfield KL, Horton DL, Johnson N, et al. Flavivirus-
induced antibody cross-reactivity. J Gen Virol 2011;92:2821-9.
3. Update on interim Zika virus clinical guidance and recom-
mendations: Clinician Outreach and Communication Activity
(COCA) call. Atlanta: Centers for Disease Control and Preven-
tion, February 25, 2016 (http://emergency.cdc.gov/coca/ppt/2016/
03_3_16_zika_update.pdf).
4. Laboratory testing for Zika virus infection. Geneva: World
Health Organization, March 23, 2016 (http://apps.who.int/iris/
bitstream/10665/204671/1/WHO_ZIKV_LAB_16.1_eng.pdf?ua=1).
5. Tunkel AR, Glaser CA, Bloch KC, et al. The management of
encephalitis: clinical practice guidelines by the Infectious Dis-
eases Society of America. Clin Infect Dis 2008;47:303-27.
6. Sejvar JJ, Kohl KS, Gidudu J, et al. Guillain-Barr syndrome
and Fisher syndrome: case definitions and guidelines for collec-
tion, analysis, and presentation of immunization safety data.
Vaccine 2011;29:599-612.
7. Hill AB. The environment and disease: association or causa-
tion? Proc R Soc Med 1965;58:295-300.
8. Cao-Lormeau VM, Blake A, Mons S, et al. Guillain-Barr syn-
drome outbreak associated with Zika virus infection in French
Polynesia: a case-control study. Lancet 2016;387:1531-9.
9. da Silva IR, Frontera JA, Moreira do Nascimento OJ. News
from the battlefront: Zika virus-associated Guillain-Barr syn-
drome in Brazil. Neurology 2016 July 15 (Epub ahead of print).
10. Tang H, Hammack C, Ogden SC, et al. Zika virus infects
human cortical neural progenitors and attenuates their growth.
Cell Stem Cell 2016;18:587-90.
11. Lucchese G, Kanduc D. Zika virus and autoimmunity: from
microcephaly to Guillain-Barr syndrome, and beyond. Autoim-
mun Rev 2016;15:801-8.
DOI: 10.1056/NEJMe1611840
Copyright 2016 Massachusetts Medical Society.
nejm.org October 20, 2016
37 Notable Articles of 2016 nejm.org

new england
The
journal of medicine
established in 1812 October 27, 2016 vol. 375 no. 17

A Randomized Trial of Long-Term Oxygen for COPD

with Moderate Desaturation
The Long-Term Oxygen Treatment Trial Research Group*

a bs t r ac t

BACKGROUND
Long-term treatment with supplemental oxygen has unknown efficacy in patients with The members of the writing committee
stable chronic obstructive pulmonary disease (COPD) and resting or exercise-induced (Richard K. Albert, M.D., David H. Au,
M.D., Amanda L. Blackford, Sc.M., Richard
moderate desaturation. Casaburi, M.D., Ph.D., J. Allen Cooper,
Jr., M.D., Gerard J. Criner, M.D., Philip
METHODS Diaz, M.D., Anne L. Fuhlbrigge, M.D.,
We originally designed the trial to test whether long-term treatment with supplemental Steven E. Gay, M.D., Richard E. Kanner,
oxygen would result in a longer time to death than no use of supplemental oxygen among M.D., Neil MacIntyre, M.D., Fernando J.
Martinez, M.D., Ralph J. Panos, M.D.,
patients who had stable COPD with moderate resting desaturation (oxyhemoglobin satu- Steven Piantadosi, M.D., Ph.D., Frank
ration as measured by pulse oximetry [Spo2], 89 to 93%). After 7 months and the ran- Sciurba, M.D., David Shade, J.D., Thomas
domization of 34 patients, the trial was redesigned to also include patients who had Stibolt, M.D., James K. Stoller, M.D.,
Robert Wise, M.D., Roger D. Yusen, M.D.,
stable COPD with moderate exercise-induced desaturation (during the 6-minute walk test, James Tonascia, Ph.D., Alice L. Stern-
Spo2 80% for 5 minutes and <90% for 10 seconds) and to incorporate the time to the berg, Sc.M., and William Bailey, M.D.) as-
first hospitalization for any cause into the new composite primary outcome. Patients were sume responsibility for this article. The
affiliations of the members of the writing
randomly assigned, in a 1:1 ratio, to receive long-term supplemental oxygen (supplemental- committee are listed in the Appendix. Ad-
oxygen group) or no long-term supplemental oxygen (no-supplemental-oxygen group). In the dress reprint requests to Dr. Wise at the
supplemental-oxygen group, patients with resting desaturation were prescribed 24-hour Johns Hopkins Asthma and Allergy Cen-
ter, 4B.72, Division of Pulmonary and
oxygen, and those with desaturation only during exercise were prescribed oxygen during Critical Care, 5501 Hopkins Bayview Cir-
exercise and sleep. The trial-group assignment was not masked. cle, Baltimore, MD 21224, or at rwise@
jhmi.edu.
RESULTS
A total of 738 patients at 42 centers were followed for 1 to 6 years. In a time-to-event * A complete list of investigators in the
Long-Term Oxygen Treatment Trial
analysis, we found no significant difference between the supplemental-oxygen group and (LOTT) Research Group is provided in
the no-supplemental-oxygen group in the time to death or first hospitalization (hazard the Supplementary Appendix, available
ratio, 0.94; 95% confidence interval [CI], 0.79 to 1.12; P = 0.52), nor in the rates of all at NEJM.org.

hospitalizations (rate ratio, 1.01; 95% CI, 0.91 to 1.13), COPD exacerbations (rate ratio, N Engl J Med 2016;375:1617-27.
1.08; 95% CI, 0.98 to 1.19), and COPD-related hospitalizations (rate ratio, 0.99; 95% CI, DOI: 10.1056/NEJMoa1604344
Copyright 2016 Massachusetts Medical Society.
0.83 to 1.17). We found no consistent between-group differences in measures of quality
of life, lung function, and the distance walked in 6 minutes.
Read Full Article at NEJM.org
CONCLUSIONS
In patients with stable COPD and resting or exercise-induced moderate desaturation, the
prescription of long-term supplemental oxygen did not result in a longer time to death or
first hospitalization than no long-term supplemental oxygen, nor did it provide sustained
benefit with regard to any of the other measured outcomes. (Funded by the National
Heart, Lung, and Blood Institute and the Centers for Medicare and Medicaid Services;
LOTT ClinicalTrials.gov number, NCT00692198.)

n engl j med 375;17 nejm.org October 27, 2016 1617

Back to Table of Contents

38 Notable Articles of 2016 nejm.org
The n e w e ng l a n d j o u r na l of m e dic i n e

Edi t or i a l s

Clinical Usefulness of Long-Term Oxygen Therapy in Adults

Magnus Ekstrm, M.D., Ph.D.

The fact that we all need oxygen to survive might
make the benefit of supplemental oxygen in hypox-
emia seem obvious. It is not. Long-term oxygen
therapy was the first treatment to improve prog-
nosis in patients with chronic obstructive pulmo-
nary disease (COPD) and chronic severe hypox-
emia.1,2 However, the question of whether
long-term oxygen therapy is beneficial in moder-
ate hypoxemia has been floating in the air.
The literature on the efficacy of long-term
oxygen therapy requires no librarian. The current
indications for its use are based on two un-
blinded, randomized trials that were conducted in
the 1970s and involved a total of 290 patients.1,2
Long-term oxygen therapy given for 15 hours or
more per day prolonged survival, as compared
with only nocturnal use or no such therapy.1,2 Par-
ticipants had COPD and chronic severe hypox-
emia (a partial pressure of arterial oxygen [Pao2]
of 55 mm Hg or an oxyhemoglobin saturation
level as measured by pulse oximetry [Spo2] of
approximately 88%) or moderate hypoxemia
(Pao2 of 56 to 59 mm Hg or Spo2 between 88%
and 90%) with signs of heart failure on the right
side or polycythemia.1,2 These characteristics
have since been the clinical eligibility criteria for
long-term oxygen therapy.3 Of note, patients in
the trials were younger and had fewer coexisting
conditions than patients starting long-term oxy-
gen therapy in current practice.4 A survival benefit
was not seen in two smaller trials in the 1990s
of nocturnal oxygen (in 76 patients)5 or long-
term oxygen therapy (in 135)6 among patients
with mild-to-moderate hypoxemia.
But does long-term oxygen therapy reduce
breathlessness or improve quality of life? With
regard to patients who have COPD with mild-to-
moderate hypoxemia, some data suggest that
supplemental oxygen decreases breathlessness
during exercise testing or structured training,7
but evidence of efficacy with regard to symp-
toms and quality of life in the home setting is
lacking.3,8,9 In patients with severe hypoxemia,
the effect of long-term oxygen therapy on patient-
reported outcomes has not been studied.9
Thus, despite frequent prescription, relatively
high costs, and the potential burden on patients
of long-term oxygen therapy, quality evidence
regarding its clinical usefulness in patients with
COPD and moderate hypoxemia has been lack-
ing.3 Until now.
As published in this issue of the Journal, the
Long-Term Oxygen Treatment Trial (LOTT)10 ran-
domly assigned 738 participants with COPD (73%
of whom were men) and mild-to-moderate hypox-
emia at rest or during a 6-minute walk test to
receive either long-term supplemental oxygen or
no long-term supplemental oxygen. The supple-
mental oxygen was prescribed as 2 liters of oxy-
gen per minute continuously in participants with
resting hypoxemia (57% of the participants) and
as an adjusted oxygen dose during exercise and
2 liters of oxygen per minute during sleep in
participants with exertional hypoxemia only (43%).
During a median follow-up of 18.4 months,
there was no significant between-group differ-
ence in the rate of death or first hospitalization
in the time-to-event analysis (primary outcome)
or in mortality and the rate of hospitalizations
separately, COPD exacerbations, quality of life,
anxiety, depression, or functional status.
This landmark study is the largest to date
with regard to long-term oxygen therapy. It evalu-
ated clinically relevant patient outcomes in daily
life. Owing to slow recruitment, there were early
changes in the trial; from two separate trials,
one involving patients with resting desaturation
and one involving patients with exercise desatu-

n engl j med 375;17 nejm.org October 27, 2016 1683

Back to Table of Contents

 39 Notable Articles of 2016
The n e w e ng l a n d j o u r na l
ration, the investigators created a composite trial
that included both types of patients. Although
this change added complexity, the trial met its
target sample size and included clinically rele-
vant subgroups of patients and treatment strate-
gies for which evidence is needed. Similar to
previous trials,1,2,5,6 the LOTT was unblinded,
which may have confounded the effect of long-
term oxygen therapy to seem more beneficial
(e.g., because of a placebo effect or more clinical
contacts in the supplemental-oxygen group, es-
pecially for patient-reported outcomes) or less
beneficial (e.g., a lower threshold for seeking or
providing care in people in the no-supplemental-
oxygen group). The validity of the findings is
supported by the consistent lack of effect across
outcomes, which was not modified by type of
oxygen prescription, desaturation profile, oxygen
use, sex, smoking status, and lung function.
I believe that on the basis of all available cur-
rent data, long-term oxygen therapy should be
prescribed to prolong survival among patients
with COPD who have chronic (>3 weeks) severe
resting hypoxemia (Pao2 of 55 mm Hg or Spo2
of <88%) while they are breathing ambient air.
Since a lack of evidence of effect is not evidence
of a lack of any clinical effectiveness, a trial of
oxygen use might still be appropriate in selected
patients with moderate exertional hypoxemia
and intractable breathlessness despite appropriate
evidence-based treatment. I think that the oxy-
gen treatment should be evaluated by means of
blinded exercise tests while the patient is breath-
ing ambient air or oxygen and discontinued if the
patient perceives no benefit during the test or
within a day or two after it.8 If there is benefit,
these selected patients should be prescribed oxy-
gen, and I think that this treatment should be
1684 n engl j med 375;17
Back to Table of Contents
nejm.org
of m e dic i n e
covered by insurance payers. However, long-term
oxygen therapy should not be routinely prescribed
in patients with mild or moderate hypoxemia at
rest or during exercise.
Disclosure forms provided by the author are available with the
full text of this editorial at NEJM.org.
From the Department of Clinical Sciences, Division of Respira-
tory Medicine and Allergology, Lund University, Lund, and the
Department of Medicine, Blekinge Hospital, Karlskrona both
in Sweden.
1. Nocturnal Oxygen Therapy Trial Group. Continuous or noc-
turnal oxygen therapy in hypoxemic chronic obstructive lung
disease: a clinical trial. Ann Intern Med 1980;93:391-8.
2. Long term domiciliary oxygen therapy in chronic hypoxic
cor pulmonale complicating chronic bronchitis and emphysema:
report of the Medical Research Council Working Party. Lancet
1981;1:681-6.
3. Hardinge M, Annandale J, Bourne S, et al. British Thoracic
Society guidelines for home oxygen use in adults. Thorax 2015;
70:Suppl 1:i1-43.
4. Ekstrm MP, Wagner P, Strm KE. Trends in cause-specific
mortality in oxygen-dependent chronic obstructive pulmonary
disease. Am J Respir Crit Care Med 2011;183:1032-6.
5. Chaouat A, Weitzenblum E, Kessler R, et al. A randomized
trial of nocturnal oxygen therapy in chronic obstructive pulmo-
nary disease patients. Eur Respir J 1999;14:1002-8.
6. Grecka D, Gorzelak K, Sliwiski P, Tobiasz M, Zieliski J.
Effect of long-term oxygen therapy on survival in patients with
chronic obstructive pulmonary disease with moderate hypoxae-
mia. Thorax 1997;52:674-9.
7. Uronis HE, Ekstrm MP, Currow DC, McCrory DC, Samsa
GP, Abernethy AP. Oxygen for relief of dyspnoea in people with
chronic obstructive pulmonary disease who would not qualify
for home oxygen: a systematic review and meta-analysis. Thorax
2015;70:492-4.
8. Abernethy AP, McDonald CF, Frith PA, et al. Effect of pallia-
tive oxygen versus room air in relief of breathlessness in patients
with refractory dyspnoea: a double-blind, randomised controlled
trial. Lancet 2010;376:784-93.
9. Cranston JM, Crockett AJ, Moss JR, Alpers JH. Domiciliary
oxygen for chronic obstructive pulmonary disease. Cochrane
Database Syst Rev 2005;4:CD001744.
10. The Long-Term Oxygen Treatment Trial Research Group.
A randomized trial of long-term oxygen for COPD with moderate
desaturation. N Engl J Med 2016;375:1617-27.
DOI: 10.1056/NEJMe1611742
Copyright 2016 Massachusetts Medical Society.
nejm.org October 27, 2016
 40 Notable Articles of 2016 nejm.org

The NEW ENGLA ND JOURNAL of MEDICINE

Perspective February 18, 2016

Zika Virus in the Americas Yet Another Arbovirus Threat

Anthony S. Fauci, M.D., and David M. Morens, M.D.

T
Zika Virus in the Americas

he explosive pandemic of Zika virus infection Asia. The virus circulated pre-
occurring throughout South America, Central dominantly in wild primates and
arboreal mosquitoes such as Aedes
America, and the Caribbean (see map) and africanus and rarely caused recog-
potentially threatening the United States is the most nized spillover infections in hu-
mans, even in highly enzootic
recent of four unexpected arrivals mosquitoes and ticks. Arboviruses areas.2 Its current explosive pan-
of important arthropod-borne vi- are often maintained in complex demic reemergence is therefore
ral diseases in the Western Hemi- cycles involving vertebrates such truly remarkable.3 Decades ago,
sphere over the past 20 years. It as mammals or birds and blood- African researchers noted that
follows dengue, which entered feeding vectors. Until recently, aedes-transmitted Zika epizootics
this hemisphere stealthily over only a few arboviruses had caused inexplicably tended to follow aedes-
decades and then more aggres- clinically significant human dis- transmitted chikungunya epizo-
sively in the 1990s; West Nile vi- eases, including mosquito-borne otics and epidemics. An analogous
rus, which emerged in 1999; and alphaviruses such as chikungunya pattern began in 2013, when chi-
chikungunya, which emerged in and flaviviruses such as dengue kungunya spread pandemically
2013. Are the successive migra- and West Nile. The most histori- from west to east, and Zika later
tions of these viruses unrelated, cally important of these is yellow followed. Zika has now circled
or do they reflect important new fever virus, the first recognized the globe, arriving not only in the
patterns of disease emergence? viral cause of deadly epidemic Americas but also, in September,
Furthermore, are there secondary hemorrhagic fever. in the country of Cape Verde in
health consequences of this arbo- Zika, which was discovered in- West Africa, near its presumed
virus pandemic that set it apart cidentally in Uganda in 1947 in ancient ancestral home.
from others? the course of mosquito and pri- With the exception of West
Arbovirus is a descriptive mate surveillance,1 had until now Nile virus, which is predominantly
term applied to hundreds of pre- remained an obscure virus con- spread by culex-species mosqui-
dominantly RNA viruses that are fined to a narrow equatorial belt toes, the arboviruses that recently
transmitted by arthropods, notably running across Africa and into reached the Western Hemisphere

n engl j med 374;7 nejm.org February 18, 2016 601

Back to Table of Contents

 41 Notable Articles of 2016 nejm.org
Zika Virus in the Americas
PER S PE C T IV E

Locally acquired cases

or virus isolation
Serosurvey data only

Countries with Past or Current Evidence of Zika Virus Transmission (as of December 2015).
For countries with serosurvey data only, evidence of Zika virus transmission is derived from studies that detected Zika virus anti-
bodies in healthy people. Outlined areas, all with locally acquired cases or virus isolation, include Cape Verde, Cook Islands, Easter
Island, Federated States of Micronesia, French Polynesia, Martinique, New Caledonia, Puerto Rico, Solomon Islands, and Vanuatu.
Data are from the Centers for Disease Control and Prevention (http://www.cdc.gov/zika).

have been transmitted by aedes
mosquitoes, especially the yellow
fever vector mosquito A. aegypti.
These viruses started to emerge
millennia ago, when North Afri-
can villagers began to store wa-
ter in their dwellings. Arboreal
A. aegypti then adapted to deposit
their eggs in domestic water-
containing vessels and to feed on
humans, which led to adaptation
of arboreal viruses to infect hu-
mans. The yellow fever, dengue,
and chikungunya viruses evolved
entirely new maintenance cycles
of humanA. aegyptihuman trans-
mission.4 Now, 5000 years later,
the worst effects of this evolu-
tionary cascade are being seen in
the repeated emergence of arbo-
viruses into new ecosystems in-
volving humans. Moreover, arbo-
viruses transmitted by different
mosquitoes have, in parallel,
adapted to humans domestic an-
imals, such as horses in the case
of Venezuelan equine encephalitis
and pigs in the case of Japanese
encephalitis virus, or to verte-
brate hosts and non-aedes mos-
quitoes found in areas of human
habitation, as West Nile virus did.
The possibility that Zika may yet
adapt to transmission by A. albop-
ictus, a much more widely distrib-
uted mosquito found in at least
32 states in the United States, is
cause for concern.
Through early epidemiologic
surveillance and human challenge
studies, Zika was characterized
as a mild or inapparent dengue-
like disease with fever, muscle
aches, eye pain, prostration, and
maculopapular rash. In more than
60 years of observation, Zika has
not been noted to cause hemor-
rhagic fever or death. There is in
vitro evidence that Zika virus
mediates antibody-dependent en-
hancement of infection, a phe-
nomenon observed in dengue
hemorrhagic fever; however, the
clinical significance of that find-
ing is uncertain.
The ongoing pandemic con-
firms that Zika is predominantly
a mild or asymptomatic dengue-
like disease. However, data from
French Polynesia documented a
concomitant epidemic of 73 cases
of GuillainBarr syndrome and
other neurologic conditions in
a population of approximately
270,000, which may represent
complications of Zika. Of greater
concern is the explosive Brazilian
epidemic of microcephaly, mani-
fested by an apparent 20-fold in-
crease in incidence from 2014 to
2015, which some public health
officials believe is caused by Zika
virus infections in pregnant
women. Although no other flavi-
virus is known to have terato-
genic effects, the microcephaly
epidemic has not yet been linked
to any other cause, such as in-
creased diagnosis or reporting,
increased ultrasound examina-
tions of pregnant women, or other
infectious or environmental
agents. Despite the lack of defini-
tive proof of any causal relation-
ship,5 some health authorities in
afflicted regions are recommend-
ing that pregnant women take
meticulous precautions to avoid
mosquito bites and even to delay
pregnancy. It is critically impor-

602 n engl j med 374;7 nejm.org February 18, 2016

Back to Table of Contents

 42 Notable Articles of 2016
P ER S PE C T IV E
tant to confirm or dispel a causal
link between Zika infection of
pregnant women and the occur-
rence of microcephaly by doing
intensive investigative research,
including careful casecontrol and
other epidemiologic studies as
well as attempts to duplicate this
phenomenon in animal models.
In a pure Zika epidemic, a
diagnosis can be made reliably
on clinical grounds. Unfortunate-
ly, the fact that dengue and chi-
kungunya, which result in simi-
lar clinical pictures, have both
been epidemic in the Americas
confounds clinical diagnoses. Spe-
cific tests for dengue and chi-
kungunya are not always avail-
able, and commercial tests for
Zika have not yet been developed.
Moreover, because Zika is closely
related to dengue, serologic sam-
ples may cross-react in tests for
either virus. Gene-detection tests
such as the polymerase-chain-
reaction assay can reliably distin-
guish the three viruses, but Zika-
specific tests are not yet widely
available.
The mainstays of management
are bed rest and supportive care.
When multiple arboviruses are co-
circulating, specific viral diagno-
sis, if available, can be important
in anticipating, preventing, and
managing complications. For ex-
ample, in dengue, aspirin use
should be avoided and patients
should be monitored for a rising
hematocrit predictive of impend-
ing hemorrhagic fever, so that
potentially lifesaving treatment
can be instituted promptly. Pa-
tients with chikungunya virus in-
fection should be monitored and
treated for acute arthralgias and
postinfectious chronic arthritis.
There are no Zika vaccines in
advanced development, although
a number of existing flavivirus
vaccine platforms could presum-
n engl j med 374;7
ably be adapted, including flavi-
virus chimera or glycoprotein sub-
unit technologies. Zika vaccines
would, however, face the same
problem as vaccines for chikun-
gunya,4 West Nile, St. Louis en-
cephalitis, and other arboviruses:
since epidemics appear sporadi-
cally and unpredictably, preemp-
tively vaccinating large popula-
tions in anticipation of outbreaks
may be prohibitively expensive
and not cost-effective, yet vaccine
stockpiling followed by rapid
deployment may be too slow to
counter sudden explosive epi-
demics. Although yellow fever
has historically been prevented
entirely by aggressive mosquito
control, in the modern era vector
control has been problematic be-
cause of expense, logistics, pub-
lic resistance, and problems posed
by inner-city crowding and poor
sanitation. Among the best pre-
ventive measures against Zika vi-
rus are house screens, air-condi-
tioning, and removal of yard and
household debris and containers
that provide mosquito-breeding
sites, luxuries often unavailable to
impoverished residents of crowd-
ed urban locales where such epi-
demics hit hardest.
With its recent appearance in
Puerto Rico, Zika virus forces us to
confront a potential new disease-
emergence phenomenon: pandem-
ic expansion of multiple, heretofore
relatively unimportant arboviruses
previously restricted to remote
ecologic niches. To respond, we
urgently need research on these
viruses and the ecologic, entomo-
logic, and host determinants of
viral maintenance and emergence.
Also needed are better public
health strategies to control arbo-
viral spread, including vaccine
platforms for flaviviruses, alpha-
viruses, and other arbovirus
groups that can be quickly modi-
nejm.org February 18, 2016
Back to Table of Contents
nejm.org
Zika Virus in the Americas
fied to express immunogenic an-
tigens of newly emerging viruses.
With respect to treatment, the
arbovirus pandemics suggest that
the one-bugone-drug approach
is inadequate; broad-spectrum
antiviral drugs effective against
whole classes of viruses are ur-
gently needed.
As was realized more than 50
years ago, when enzootic Zika vi-
rus spread was linked to human
activity, arboviruses continually
evolve and adapt within ecologic
niches that are increasingly be-
ing perturbed by humans. Zika
is still a pandemic in progress,
and many important questions
about it, such as that of terato-
genicity, remain to be answered.
Yet it has already reinforced one
important lesson: in our human-
dominated world, urban crowd-
ing, constant international travel,
and other human behaviors com-
bined with human-caused micro-
perturbations in ecologic balance
can cause innumerable slumber-
ing infectious agents to emerge
unexpectedly. In response, we
clearly need to up our game with
broad and integrated research
that expands understanding of
the complex ecosystems in which
agents of future pandemics are
aggressively evolving.
Disclosure forms provided by the authors
are available with the full text of this article
at NEJM.org.
From the National Institute of Allergy and
Infectious Diseases, Bethesda, MD.
This article was published on January 13,
2016, at NEJM.org.
1. Dick GW, Kitchen SF, Haddow AJ. Zika
virus. I. Isolations and serological specificity.
Trans R Soc Trop Med Hyg 1952;46:509-20.
2. Pierson TC, Diamond MS. Flaviviruses.
In: Knipe DM, Howley PM, Cohen IC, et al.,
eds. Fields virology. 6th ed. Vol. 1. Philadel-
phia: Wolters Kluwer, 2014:746-94.
3. Marcondes CB, Ximenes MF. Zika virus
in Brazil and the danger of infestation by
Aedes (Stegomyia) mosquitoes. Rev Soc Bras
603
 43 Notable Articles of 2016
PE R S PE C T IV E
Med Trop 2015 December 22 (Epub ahead of
print).
4. Morens DM, Fauci AS. Chikungunya at
the door dj vu all over again? N Engl J
Med 2014;371:885-7.
5. European Centre for Disease Prevention
How Employers Are Responding to the ACA
604
nejm.org
Zika Virus in the Americas
and Control. Microcephaly in Brazil poten- News_DispForm.aspx?ID=1329&List=
tially linked to the Zika virus epidemic: 8db7286c-fe2d-476c-9133-18ff4cb1b568
ECDC assesses the risk. Solna, Sweden: &Source=http%3A%2F%2Fecdc.europa
European Centre for Disease Prevention and .eu%2Fen%2FPages%2Fhome.aspx).
Control, November 25, 2015 (http://ecdc DOI: 10.1056/NEJMp1600297
.europa.eu/en/press/news/_layouts/forms/ Copyright 2016 Massachusetts Medical Society.
Zika Virus in the Americas
n engl j med 374;7 nejm.org February 18, 2016
Back to Table of Contents
44 Notable Articles of 2016 nejm.org

The NEW ENGLA ND JOURNAL of MEDICINE

Perspective April 21, 2016

Reducing the Risks of Relief The CDC Opioid-Prescribing

Guideline
Thomas R. Frieden, M.D., M.P.H., and Debra Houry, M.D., M.P.H.

D
eaths from prescription-opioid overdose have given what we know about the
increased dramatically in the United States, risks associated with long-term
opioid therapy and the availability
quadrupling in the past 15 years. Efforts to of effective nonpharmacologic and
improve pain management resulted in quadrupled nonopioid pharmacologic treat-
ment options, the guideline uses
rates of opioid prescribing, which long-acting opioids to physicians.1 the best available scientific data
propelled a tightly correlated epi- It has become increasingly clear to provide information and recom-
demic of addiction, overdose, and that opioids carry substantial risks mendations to support patients
death from prescription opioids and uncertain benefits, especially and clinicians in balancing the
that is now further evolving to in- as compared with other treat- risks of addiction and overdose
clude increasing use and overdoses ments for chronic pain. with the limited evidence of ben-
of heroin and illicitly produced On March 15, 2016, the Cen- efits of opioids for the treatment
fentanyl. ters for Disease Control and Pre- of chronic pain.
The pendulum of opioid use in vention (CDC) released a Guide- Most placebo-controlled, ran-
pain management has swung back line for Prescribing Opioids for domized trials of opioids have
and forth several times over the Chronic Pain to chart a safer, lasted 6 weeks or less, and we are
past 100 years. Beginning in the more effective course.2 The guide- aware of no study that has com-
1990s, efforts to improve treat- line is designed to support clini- pared opioid therapy with other
ment of pain failed to adequately cians caring for patients outside treatments in terms of long-term
take into account opioids addic- the context of active cancer treat- (more than 1 year) outcomes re-
tiveness, low therapeutic ratio, and ment or palliative or end-of-life lated to pain, function, or quality
lack of documented effectiveness care. More research is needed to of life.2 The few randomized tri-
in the treatment of chronic pain. fill in critical evidence gaps re- als to evaluate opioid efficacy for
Increased prescribing was also garding the effectiveness, safety, longer than 6 weeks had consis-
fueled by aggressive and some- and economic efficiency of long- tently poor results. In fact, several
times misleading marketing of term opioid therapy. However, studies have showed that use of

n engl j med 374;16 nejm.org April 21, 2016 1501

Back to Table of Contents

 45 Notable Articles of 2016 nejm.org
PER S PE C T IV E the cdc opioid-prescribing guideline

ceived opioid therapy were less

The CDC Opioid-Prescribing Guideline.
likely to have improvement in
1. Nonpharmacologic therapy and nonopioid pharmacologic therapy are preferred pain (odds ratio, 0.42; 95% con-
for chronic pain. Clinicians should consider opioid therapy only if expected ben- fidence interval [CI], 0.36 to
efits for both pain and function are anticipated to outweigh risks to the patient.
If opioids are used, they should be combined with nonpharmacologic therapy 0.49) and had worsened function
and nonopioid pharmacologic therapy, as appropriate. (odds ratio, 1.25; 95% CI, 1.04
2. Before starting opioid therapy for chronic pain, clinicians should establish treat- to 1.51).3 An observational case
ment goals with all patients, including realistic goals for pain and function, and control study of patients under-
should consider how therapy will be discontinued if benefits do not outweigh
risks. Clinicians should continue opioid therapy only if there is clinically mean- going orthopedic surgery showed
ingful improvement in pain and function that outweighs risks to patient safety. that those receiving long-term
3. Before starting and periodically during opioid therapy, clinicians should discuss opioid therapy had significantly
with patients known risks and realistic benefits of opioid therapy and patient
and provider responsibilities for managing therapy.
higher levels of preoperative hy-
4. When starting opioid therapy for chronic pain, clinicians should prescribe imme- peralgesia.4 After surgery, patients
diate-release opioids instead of extended-release/long-acting (ER/LA) opioids. who had received long-term opi-
5. When opioids are started, clinicians should prescribe the lowest effective dosage. oid therapy reported higher pain
Clinicians should use caution when prescribing opioids at any dosage, should intensity (a rating of 7.6 vs. 5.5 out
carefully reassess evidence of individual benefits and risks when increasing dos-
age to 50 morphine milligram equivalents (MME) per day, and should avoid of 10) in the recovery room than
increasing dosage to 90 MME per day or carefully justify a decision to titrate patients who had not been tak-
dosage to 90 MME per day. ing opioids.
6. Long-term opioid use often begins with treatment of acute pain. When opioids
are used for acute pain, clinicians should prescribe the lowest effective dose of
Whereas the benefits of opi-
immediate-release opioids and should prescribe no greater quantity than need- oids for chronic pain remain un-
ed for the expected duration of pain severe enough to require opioids. Three certain, the risks of addiction
days or less will often be sufficient; more than 7 days will rarely be needed.
and overdose are clear. Although
7. Clinicians should evaluate benefits and harms with patients within 14 weeks of
starting opioid therapy for chronic pain or of dose escalation. Clinicians should partial agonists such as buprenor-
evaluate benefits and harms of continued therapy with patients every 3 months phine may carry a lower risk of
or more frequently. If benefits do not outweigh harms of continued opioid ther- dependence, prescription opioids
apy, clinicians should optimize other therapies and work with patients to taper
opioids to lower dosages or to taper and discontinue opioids. that are full mu-opioidreceptor
8. Before starting and periodically during continuation of opioid therapy, clinicians agonists nearly all the prod-
should evaluate risk factors for opioid-related harms. Clinicians should incorpo- ucts on the market are no less
rate into the management plan strategies to mitigate risk, including considering addictive than heroin. Although
offering naloxone when factors that increase risk for opioid overdose, such as
history of overdose, history of substance use disorder, higher opioid dosages abuse-deterrent formulations may
(50 MME/day), or concurrent benzodiazepine use are present. reduce the likelihood that pa-
9. Clinicians should review the patients history of controlled substance prescrip- tients will inject melted pills,
tions using state prescription drug monitoring program (PDMP) data to deter-
mine whether the patient is receiving opioid dosages or dangerous combina-
these formulations are no less
tions that put him or her at high risk for overdose. Clinicians should review addictive and do not prevent
PDMP data when starting opioid therapy for chronic pain and periodically dur- opioid abuse or fatal overdose
ing opioid therapy for chronic pain, ranging from every prescription to every
3 months.
through oral intake.
10. When prescribing opioids for chronic pain, clinicians should use urine drug test- The prevalence of opioid de-
ing before starting opioid therapy and consider urine drug testing at least annu- pendence may be as high as 26%
ally to assess for prescribed medications as well as other controlled prescription among patients in primary care
drugs and illicit drugs.
receiving opioids for chronic non
11. Clinicians should avoid prescribing opioid pain medication and benzodiazepines
concurrently whenever possible. cancer-related pain.2 Risk-stratifi-
12. Clinicians should offer or arrange evidence-based treatment (usually medication- cation tools do not allow clinicians
assisted treatment with buprenorphine or methadone in combination with be- to predict accurately whether a
havioral therapies) for patients with opioid-use disorder.
patient will become addicted to
opioids, although persons with
a history of mental illness or
opioids for chronic pain may ac- observational study of more than addiction are at higher risk.2
tually worsen pain and function- 69,000 postmenopausal women Overdose risk increases in a dose
ing, possibly by potentiating pain with recurrent pain conditions response manner, at least dou-
perception. A 3-year prospective showed that patients who had re- bling at 50 to 99 morphine milli-

1502 n engl j med 374;16 nejm.org April 21, 2016

Back to Table of Contents

 46 Notable Articles of 2016
PER S PE C T IV E
gram equivalents (MME) per day
and increasing by a factor of up
to 9 at 100 or more MME per
day, as compared with doses of
less than 20 MME per day.2 Over-
all, 1 of every 550 patients start-
ed on opioid therapy died of
opioid-related causes a median
of 2.6 years after the first opioid
prescription; the proportion was
as high as 1 in 32 among pa-
tients receiving doses of 200 MME
or higher.5 We know of no other
medication routinely used for a
nonfatal condition that kills pa-
tients so frequently.
The new CDC guideline em-
phasizes both patient care and
safety. We developed the guide-
line using a rigorous process that
included a systematic review of
the scientific evidence and input
from hundreds of leading experts
and practitioners, other federal
agencies, more than 150 profes-
sional and advocacy organiza-
tions, a wide range of key patient
and provider groups, a federal
advisory committee, peer review-
ers, and more than 4000 public
comments.
Three key principles underlie
the guidelines 12 recommenda-
tions (see box). First, nonopioid
therapy is preferred for chronic
pain outside the context of active
cancer, palliative, or end-of-life
care. Opioids should be added
An audio interview
to other treatments
with Dr. Houry is for chronic pain
available at NEJM.org only when their ex-
pected benefits for
both pain and function are likely
to outweigh the substantial risks
inherent in this class of medi-
cation.
Nonpharmacologic therapies
can ameliorate chronic pain while
posing substantially less risk to
patients. In some instances, other
therapies result in better out-
n engl j med 374;16
the cdc opioid-prescribing guideline
comes than opioids. These ther-
apies include exercise therapy,
weight loss, psychological thera-
pies such as cognitive behavioral
therapy, interventions to improve
sleep, and certain procedures.
The evidence review conducted
in developing the guideline re-
vealed that exercise therapy
helped improve, and sustain im-
provements in, pain and function
in patients with osteoarthritis. It
did not find evidence that opi-
oids were more effective for pain
reduction than nonopioid treat-
ments such as nonsteroidal anti-
inflammatory drugs for low back
pain or antidepressants for neu-
ropathic pain, but it did find
that nonopioid treatments could
be better tolerated and superior
for improving physical function
while conferring little or no risk
of addiction and substantially
lower risks of overdose and
death.2
Second, when opioids are used,
the lowest possible effective dose
should be prescribed to reduce
the risks of opioid use disorder
and overdose. Clinicians should
carefully reassess individual ben-
efits and risks when increasing a
dose to 50 MME or more per day.
Doses of 90 MME or more should
be avoided, or the decision to ti-
trate above this level should be
carefully considered and justified.
When prescribing opioids, the rule
of thumb is to start low and
go slow.
Third, clinicians should exer-
cise caution when prescribing
opioids and should monitor all
patients closely. Prescribers should
mitigate risk by, for example,
avoiding concurrent use of ben-
zodiazepines if possible, review-
ing data from a prescription-
drug monitoring program when
deciding whether to start or con-
nejm.org April 21, 2016
Back to Table of Contents
nejm.org
tinue opioid therapy, offering
naloxone at least to patients who
are at greater risk for overdose,
having a clear off-ramp plan to
taper and discontinue therapy,
reevaluating the dosage and ne-
cessity of opioid treatment regu-
larly, and obtaining urine toxi-
cology screening at the initiation
of treatment and, for some pa-
tients, periodically thereafter. For
patients who become addicted to
opioids, treatment with metha-
done, buprenorphine, or naltrex-
one improves outcomes.
Initiation of treatment with
opioids is a momentous decision
and should be undertaken only
with full understanding by both
the physician and the patient of
the substantial risks involved.
Clinicians need to recognize the
risk associated with any treat-
ment with opioids and should
prescribe only the shortest course
needed. Although the guideline
addresses chronic pain, many pa-
tients become addicted to opioids
after being treated for acute pain.
Three days of treatment or less
will often be sufficient; more
than 7 days will rarely be re-
quired. Some trauma and surgery
may require longer courses; treat-
ment of postsurgical pain is be-
yond the scope of this guideline.
Furthermore, it is important to
discuss storage of opioids in a
secure location to prevent diver-
sion, as well as to counsel pa-
tients regarding the overdose risk
posed to household members and
other persons.
Management of chronic pain
is an art and a science. The sci-
ence of opioids for chronic pain
is clear: for the vast majority of
patients, the known, serious, and
too-often-fatal risks far outweigh
the unproven and transient ben-
efits.
1503
 47 Notable Articles of 2016
PER S PE C T IV E
Disclosure forms provided by the authors
are available with the full text of this article
at NEJM.org.
From the Centers for Disease Control and
Prevention, Atlanta.
This article was published on March 15,
2016, at NEJM.org.
1. Van Zee A. The promotion and market-
ing of oxycontin: commercial triumph, pub-
1504
the cdc opioid-prescribing guideline
lic health tragedy. Am J Public Health 2009;
99:221-7.
2. CDC guideline for prescribing opioids
for chronic pain United States, 2016.
MMWR Recomm Rep 2016;65(RR-1):1-49.
3. Braden JB, Young A, Sullivan MD, Walitt
B, Lacroix AZ, Martin L. Predictors of change
in pain and physical functioning among
post-menopausal women with recurrent pain
conditions in the Womens Health Initiative
observational cohort. J Pain 2012;13:64-72.
4. Hina N, Fletcher D, Poindessous-Jazat F,
n engl j med 374;16 nejm.org April 21, 2016
Back to Table of Contents
nejm.org
Martinez V. Hyperalgesia induced by low-
dose opioid treatment before orthopaedic
surgery: an observational case-control study.
Eur J Anaesthesiol 2015;32:255-61.
5. Kaplovitch E, Gomes T, Camacho X,
Dhalla IA, Mamdani MM, Juurlink DN. Sex
differences in dose escalation and overdose
death during chronic opioid therapy: a pop-
ulation-based cohort study. PLoS One 2015;
10(8):e0134550.
DOI: 10.1056/NEJMp1515917
Copyright 2016 Massachusetts Medical Society.
48 Notable Articles of 2016
The NEW ENGLA ND JOURNAL
Rethinking the Primary Care Workforce An Expanded Role
for Nurses
Thomas Bodenheimer, M.D., and Laurie Bauer, R.N., M.S.P.H.
T he adult population of the
United States will soon have
a different primary care experi-
ence than weve been used to. In
the primary care practice of the
future, the physicians role will
increasingly be played by nurse
practitioners (NPs). In addition,
the 150 million adults with one
or more chronic conditions will
receive some of their care from
registered nurses (RNs) function-
ing as care managers.
Workforce experts agree on the
growing gap between the popula-
tions demand for primary care
and the number of primary care
physicians available to meet that
demand. About 8000 primary care
physicians (including doctors of
osteopathy and international med-
ical graduates) entered the work-
force in 2015, up only slightly
from 7500 in 2005. And in fact,
the number of yearly entrants is
expected to plateau at around
8000. But the number of primary
care physicians who retire each
year is projected to reach 8500 in
n engl j med 375;11
Perspective
2020 in other words, the num-
ber of retirees may exceed that of
new entrants. And the size of the
primary care physician workforce
will be declining even as the U.S.
population grows, ages, and be-
comes more adequately insured.1
In contrast, the number of NPs
entering the workforce each year
has mushroomed from 6600 in
2003 to 18,000 in 2014. The
number of primary care NPs is
projected to increase by 84% be-
tween 2010 and 2025. The num-
ber of physician assistants (PAs)
entering the workforce is also
growing, though not as rapidly. If
these trends continue, the pro-
portion of primary care practi-
tioners who are physicians will
drop from 71% in 2010 to 60%
in 2025 and will continue to de-
cline thereafter. The proportion
of practitioners who are NPs will
jump from 19% to 29% during
those years and will continue to
rise.2 In rural communities, this
trend is even more pronounced,
since NPs are considerably more
nejm.org September 15, 2016
Back to Table of Contents
nejm.org
of MEDICINE
September 15, 2016
likely than physicians to settle in
rural America.
Clearly, more and more pa-
tients will see an NP or a PA as
their primary care practitioner.
Physicians will probably focus on
diagnostic conundrums and lead
teams caring for patients with
complex health care needs. A large
and growing body of research
demonstrates that care delivered
by NPs is at least as high quality
as that delivered by physicians.
In addition, patient-satisfaction
scores are similar for NPs and
physicians.3 Moreover, care may
cost less when its provided by
NPs rather than physicians: Medi-
care beneficiaries assigned to an
NP had primary care costs that
were 29% lower and office-visit
and inpatient costs that were 11 to
18% lower than those of benefi-
ciaries assigned to a primary care
physician.
Even with the increased num-
bers of NP and PA graduates, the
ratio of primary care practitioners
to population will decline, because
1015
 49 Notable Articles of 2016
P ER S PE C T IV E
only 50% of NPs and 32% of PAs
choose primary care careers. Thus,
other professionals will be needed
to care for the growing number
of U.S. adults with chronic condi-
tions and geriatric syndromes.
Enter the enhanced role of the RN.
While the NP role begins to
approximate that of the physician,
RNs are assuming three impor-
tant emerging primary care func-
tions: managing the care of pa-
tients with chronic disease by
helping them with behavior change
and adjusting their medications
(e.g., for hypertension and diabe-
tes) according to physician-written
protocols; leading complex care
management teams to help im-
prove care and reduce the cost of
care for patients with multiple
diagnoses who are high users of
health care services; and coordi-
nating care between the primary
care home and providers of other
health care services in par-
ticular, assisting with transitions
among hospital, primary care
settings, and home.4
RNs are well on their way to
An audio interview filling the gap. In
with Dr. Bodenheimer 2015, a total of
is available at NEJM.org
43% of U.S. physi-
cians worked with nurse care
managers for patients with
chronic conditions. The 3.1 mil-
lion RNs in the United States rep-
resent the countrys largest health
profession, and its numbers are
projected to grow by an aston-
ishing 33% between 2012 and
2025. Government data show
that the number of RN graduates
per year has increased from 69,000
in 2001 to 155,000 in 2013 (see
graph); a separate analysis put
the number of RN graduates at
200,000 in 2014. Thus, primary
care practices are likely to bene-
fit from a pool of RNs who could
be hired to serve as chronic care
managers.
Several studies indicate that
1016 n engl j med 375;11
160,000
150,000
140,000
No. of Nursing Graduates
130,000
120,000
110,000
100,000
90,000
80,000
70,000
69,000
60,000
0
2001
Numbers of Nursing Graduates, 20012013.
Data are from the U.S. Bureau of Health Workforce.
RNs are qualified to perform
these enhanced roles. For exam-
ple, in a randomized, controlled
trial, patients with diabetes and
elevated blood pressure who re-
ceived care from RN care man-
agers (including initiation of medi-
cations and titration of doses)
were more likely to reach their
blood-pressure goals than patients
whose care was managed by phy-
sicians alone.5 Some state boards
of registered nursing have created
a mechanism by which RNs can
change medication doses using
standardized procedures author-
ized by their physician leader-
ship.4 Using these procedures,
RNs whove been trained as health
coaches could provide most of the
care for patients with uncompli-
cated diabetes, hypertension, and
hyperlipidemia, thereby adding
considerable primary care capac-
ity. And RN coordination of tran-
sitions from hospital to home has
resulted in improved patient self-
management and reduced hospi-
tal readmissions.
Although NPs and RNs are in-
creasingly central to primary care,
there are still obstacles to their
performing these roles that need
to be overcome. Physicians report
that new NP graduates are not
nejm.org
Back to Table of Contents
nejm.org
the 20th anniversary of the hospitalist
155,000
145,000
135,000
120,000
99,000
77,000
2003 2005 2007 2009 2011 2013
initially comfortable taking re-
sponsibility for a panel of pa-
tients. To address this problem,
intensive 1-year primary care NP
residencies are springing up. Thus
far, 37 such programs exist. Doc-
tor of Nursing Practice degree
programs were designed to sup-
plant masters level NP programs,
but they are growing more slowly
than expected.
As for an enhanced role for
RNs, one barrier is that public
and private insurers rarely pay
for RN services, but that barrier
is beginning to crumble. Even
under the fee-for-service payment
model, practices can receive pay-
ment for Medicare wellness visits
and chronic care management
encounters, both of which can be
conducted entirely by RNs. As
alternative payment models grad-
ually expand, primary care pay-
ment will become less visit-based,
which will allow practices to re-
allocate more and more responsi-
bilities to RNs and other team
members.
The inadequacy of primary care
training in nursing schools pre-
sents another obstacle to RNs be-
coming chronic care managers.
The focus of nursing education
on inpatient care skills has left
September 15, 2016
 50 Notable Articles of 2016
P ER S PE C T IV E
some primary care RNs unpre-
pared for the care manager role.
The American Academy of Am-
bulatory Care Nursing and nurs-
ing leaders are addressing this
problem with new curricula and
training programs.
Finally, although RNs may be
attracted to primary cares regu-
lar work hours, its focus on pre-
vention, and long-term relation-
ships with patients, the fact that
salaries are lower in primary care
than in hospitals could also be a
barrier.
Despite these challenges, the
shortage of primary care physi-
cians and the increasing preva-
n engl j med 375;11
Rethinking the Primary Care Workforce
lence of chronic diseases are
powerful forces pushing primary
care toward stronger NP and RN
participation. Its fortunate that
the growth in the supply of NPs
and RNs enables us to rethink
who does what in primary care.
Disclosure forms provided by the authors
are available at NEJM.org.
From the Center for Excellence in Primary
Care (T.B.) and the School of Nursing
(L.B.), University of California, San Francis-
co, San Francisco.
1. Petterson SM, Liaw WR, Tran C, Baze-
more AW. Estimating the residency expan-
sion required to avoid projected primary
care physician shortages by 2035. Ann Fam
Med 2015;13:107-14.
nejm.org September 15, 2016
Back to Table of Contents
nejm.org
2. Auerbach DI, Chen PG, Friedberg MW,
et al. Nurse-managed health centers and
patient-centered medical homes could miti-
gate expected primary care physician short-
age. Health Aff (Millwood) 2013;32:1933-41.
3. Stanik-Hutt J, Newhouse RP, White KM,
et al. The quality and effectiveness of care
provided by nurse practitioners. J Nurse
Pract 2013;9:492-500.
4. Bodenheimer T, Bauer L, Olayiwola JN,
Syer S. RN role reimagined: how empower-
ing registered nurses can improve primary
care. California Health Care Foundation,
2015 (http://www.chcf.org/publications/2015/
08/rn-role-reimagined).
5. Denver EA, Barnard M, Woolfson RG,
Earle KA. Management of uncontrolled hy-
pertension in a nurse-led clinic compared
with conventional care for patients with type
2 diabetes. Diabetes Care 2003;26:2256-60.
DOI: 10.1056/NEJMp1606869
Copyright 2016 Massachusetts Medical Society.
Rethinking the Primary Care Workforce
1017
 practice-
what-you-
practice.org

Test your diagnostic and therapeutic skills through quizzes and

peer-reviewed interactive cases with real clinical images and patients.
Hone your medical knowledge and stay sharp.
Discover the many ways to test your skills at
my.nejm.org/challenge