The NEW ENGLA ND JOURNAL of MEDICINE

Perspective january 17, 2013

Fever of Unknown Origin or Fever of Too Many Origins?

Harold W. Horowitz, M.D.

P etersdorf and Beeson’s classic articles cataloguing

the causes of fever of unknown origin (FUO) have
framed the way generations of physicians think about
fevers whose source is not readily explainable.1 FUO
as they define it — a tempera­
ture rising above 38.3°C (101°F)
on several occasions over a peri­
od of more than 3 weeks, for
which no diagnosis has been
reached despite 1 week of inpa­
tient investigation — is con­
sidered classic FUO. In the past
60 years, clinician-scientists have
tracked the changing causes of
these problematic fevers, as dis­
ease patterns and definitions
have changed and as improved
serologic and imaging technolo­
gies have begun revealing diag­
noses more quickly. The standard
definition of FUO no longer in­
cludes the requirement for a week
of inpatient evaluation. And in the
early 1990s, Durack and Street
proposed dividing FUOs into four
groups: classic, nosocomial, neu­
tropenic, and HIV-associated.2
According to Petersdorf and
Beeson’s original report, FUOs
were caused by infection (in 36%
of patients), malignancy (19%),
collagen vascular diseases (19%),
and miscellaneous other causes
(19%), such as drug fever.1 No
cause was determined in 7% of
patients. It is paradoxical that
despite the introduction of com­
puted tomography (CT), magnetic
resonance imaging, improved cul­
ture techniques, numerous new
serologic assays, and polymerase-
chain-reaction studies, in recent
years more FUOs have actually
eluded diagnosis. In 2003, Vander­
schueren and colleagues reported
that in nearly a third of 290 im­
munocompetent patients in Bel­
gium, no diagnosis was made,3
and in 2007, Bleeker-Rovers et al.
reported that among 73 immu­
nocompetent patients from five
hospitals in the Netherlands, no
cause of FUO was identified in
51% of cases.4
As an infectious-disease physi­
cian who has practiced at academ­
ic, tertiary care facilities in the
metropolitan New York area for
nearly three decades, I’ve been
struck by the fact that traditionally
caused FUOs are now rarer than
the FUOs that I’m increasingly
asked to evaluate. The new FUOs
are often found in patients in the
intensive care unit (ICU) who have
traumatic brain injury, other neu­
rologic events, or dementia; are
mechanically ventilated; have some
combination of urethral, central,
and peripheral catheters placed;
have recently undergone surgery;
and are already receiving multiple
broad-spectrum antibiotics. How­
ever, they continue to spike multi­
ple fevers daily for weeks and
sometimes months on end, usually
without other signs or symptoms
of sepsis.
Physical examination often re­
veals edema (if not anasarca),
early decubital ulcers in the
sacral region at minimum, cuta­
neous eruptions that do not ap­
pear to be drug-related, mild ab­
dominal distention, wounds that

n engl j med 368;3  nejm.org  january 17, 2013º 197

The New England Journal of Medicine
Downloaded from nejm.org at BIBLIOTHEQUE UNIV LAVAL SEC ACQ on July 14, 2014. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
PERSPE C T I V E Fever of Unknown Origin or Too Many Origins?

plexity of decisions regarding anti­

N=100 N=105 N=133 N=199 N=153 N=167 N=73 biotics are also affected by the
10036.0
100
30.5 30.8 28.7 25.8 10.8 16.0 dissonant messages bombarding
Diagnoses in Febrile Patients (%) 90 90
80 80
9.7
7.0
physicians: the mantra that anti­
70 70 12.6
18.4 22.0 biotics must be used sparingly to
31.4 19.6 14.4 avoid creating antibiotic-resistant
60 6019.0
22.7 8.1 bacteria versus the urgency to
30.7 4.0
50 50 53.0 Infection
15.0 51.0
Neoplasm start antibiotics earlier while en­
40 4017.0
12.4 7.8 Noninfectious suring they are the “appropriate”
30 30 12.7
19.0 14.4
31.3 inflammatory choices (translated as “broad,”
20 20 9.5 disease
16.2
21.8
Miscellaneous
given the resistance patterns in
10 10
9.0 11.8
No diagnosis many ICUs). When patients have
0 0
1961 1982 1992 1994 1997 2002 2007 been hospitalized for many months
Publication Year
and have received numerous anti­
biotics but have persistent fevers,
Distributions of Diagnoses (and Lack of Diagnosis) among Patients with Fever. it can be unclear whether appro­
Data for studies published in 1961 through 2002 are from Vanderschueren et al.,3 priate antibiotics exist or are
and 2007 data are from Bleeker-Rovers et al.4 warranted. Although some phy­
AUTHOR: Horowitz Revised sicians sing CRP’s praises, the
have minimal erythema
FIGURE: 1 of 1 and pler studies are negative or reveal near-daily variation in this mea­
some serous drainage without nonocclusive thrombosis. SIZE C-reac­ sure and its nonspecificity make
ARTIST: ts
purulence or obvious infection, tive protein (CRP) 2 col levels fluctuate it difficult to use to guide treat­
TYPE: of
no signs suggestive Linedeep
Combo
ve­ 4-CwildlyH/T from day to day. Clostridium ment decisions. Certainly, neither
nous thrombosis, and AUTHOR, coarse PLEASE
difficile
NOTE:assays performed because CRP levels nor procalcitonin levels
Figure has been redrawn and type has been reset.
breath sounds on respiratory of chronic
Please check carefully. loose stools are invari­ help determine which cultures
exam. And their lines have been ably negative. The transthoracic should be addressed with treat­
JOB: 36802 ISSUE: 01-10-13
recently changed. echocardiogram is normal, and ment. Moreover, if one chooses
Laboratory results include nor­ there is a debate about the safety to use antibiotics, the question
mal or mildly elevated white-cell of and need for a transesopha­ of which of the multiple bacterial
counts; intermittent coagulase- geal echocardiographic study. isolates need to be covered is
negative, staphylococcus-positive Generally, before I evaluate the complex.
blood cultures; urinalysis with patient, many diagnostic studies As the keeper of the antibiot­
intermittent pyuria and cultures have been done. Nevertheless, de­ ics, should I be a conservative or
revealing, sequentially, various termining the cause of a fever a cowboy? Should the current anti­
gram-negative organisms with and which antibiotics to prescribe biotics be continued, changed, or
counts of 10,000 to 20,000 colony- is frequently daunting. Although stopped? If there are no pre­
forming units interspersed with these fevers would be considered scribed antibiotics, should I rec­
negative cultures; sputum sam­ nosocomial by Durack and Street ommend some? These are inter­
ples with few or moderate num­ and may be of infectious origin, esting questions in the abstract,
bers of white cells; and chest im­ the differential diagnosis extends but there is a real patient suffer­
ages revealing bilateral basilar well beyond the usual infectious ing, a family with questions, and
congestion with atelectasis, whose suspects. In fact, I wonder medical teams awaiting my opin­
readers say they cannot rule out whether these are FUOs or fevers ion. There are no evidence-based
infection. Wound cultures reveal of too many origins (FTMOs). studies and there is no guidance
several bacteria but few, if any, Decisions about which other on which potential source of fe­
white cells, and CT scans show or repeat diagnostic evaluations ver is the single appropriate one
“postsurgical” changes or small and procedures to undertake, to treat. Frequently, the treat­
fluid collections not particularly whether to treat empirically for ment approach is like playing
suspicious for infection. Sinus C. difficile (if that isn’t already be­ Whac-A-Mole: positive cultures
films invariably demonstrate thick­ ing done), and whether to expand are treated sequentially — pneu­
ened sinus membranes without the antibiotic potpourri or per­ monia, then catheter cultures,
air–fluid levels or other clear-cut haps discontinue antibiotics are then urine cultures. When the fever
findings of sinusitis. Venous Dop­ not easy. The nuances and com­ persists, the cycle begins again.

198 n engl j med 368;3  nejm.org  january 17, 2013

The New England Journal of Medicine

Downloaded from nejm.org at BIBLIOTHEQUE UNIV LAVAL SEC ACQ on July 14, 2014. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
 PERSPECTIVE
The medical team members
may be frustrated or believe
they’ve exhausted the workup
studies, and they may prefer not
to order any more. They may not
be too keen on continuing the
same antibiotics. The ICU team
hungers for something new and
preferably simple. As I review the
differential diagnosis, with dis­
claimers as to why any given di­
agnosis does or does not ade­
quately explain the fever, I get a
feeling of déjà vu. The team has
heard these ruminations from me
and my colleagues many times,
and I suspect that by now the
discussion is minimally compel­
ling or interesting academically.
This is not the multidimen­
sional “great case” that FUOs
were once advertised to be — the
cases presented on chief-of-ser­
vice rounds in which an expert
diagnostician pontificates about
the differential diagnosis of rare
Should Blood Be an Essential Medicine?
Harvey G. Klein, M.D.
A ccording to the World Health
Organization (WHO), approx­
imately 92 million units of blood
are collected worldwide each year.
Given that transfusions are gener­
ally credited with saving millions
of lives, it may surprise clinicians
to know that blood and blood
components are not included on
the WHO Model List of Essential
Medicines.
The Model List, established in
1977, originally included about 200
active substances. It was meant
to guide countries in providing
access to cost-effective medicines
that are vital for public health.1
The list is revised every 2 years
by a WHO expert committee.
Medicines are designated as es­
sential on the basis of their effi­
n engl j med 368;3  nejm.org  january 17, 2013
The New England Journal of Medicine
Downloaded from nejm.org at BIBLIOTHEQUE UNIV LAVAL SEC ACQ on July 14, 2014. For personal use only. No other uses without permission.
Copyright © 2013 Massachusetts Medical Society. All rights reserved.
Fever of Unknown Origin or Too Many Origins?
or subtle disease complexes and
their presentations. Given the na­
ture of the illness in many of
these patients, the conferences
are more likely to be family con­
ferences that include plans for
palliative care. If the old FUOs
were sometimes exhilarating, the
FTMOs can be debilitating. Al­
though some patients will recover
and be discharged to lead full
and active lives, many will either
die or be sent to a long-term care
facility.
We debate whether using anti­
biotics in apparently futile situa­
tions is ethical. After all, we may
“create” some extremely resistant
bacteria in one patient that could
be transmitted to others. Alter­
natively, antibiotics may be life­
saving. There are few directives,
ethical guidelines, or clinical path­
ways to follow in these cases. As
I mull over the options, I am dis­
heartened by the knowledge that
cacy and safety, availability, ease
of use in various settings, compar­
ative cost-effectiveness, and pub­
lic health need. In many coun­
tries, the list forms the basis of
national drug policies. Govern­
ments and health ministries often
refer to it when making decisions
regarding resource allocation and
health care spending. The list
does not include all efficacious
medicines, the latest medicines,
or even all medicines needed in a
country. Rather, it helps to define
the minimum medicine needs for
a basic health system.
Although some protein con­
centrates (factors VIII and IX and
immunoglobulins) are listed, no
labile blood components are on
the Model List. The reason for
whether I use or withdraw anti­
biotics (asking the team to ob­
serve the patient closely) or re­
quest more testing, I may simply
be deferring the tough decisions
for another day.
Disclosure forms provided by the author
are available with the full text of this arti­
cle at NEJM.org.
From the Division of Infectious Diseases
and Immunology, New York University
School of Medicine, New York.
1. Petersdorf RG, Beeson PB. Fever of unex-
plained origin: report on 100 cases. Medi-
cine (Baltimore) 1961;40:1-30.
2. Durack DT, Street AC. Fever of unknown
origin — reexamined and redefined. Curr
Clin Top Infect Dis 1991;11:35-51.
3. Vanderschueren S, Knockaert D, Adri-
aenssens T, et al. From prolonged febrile ill-
ness to fever of unknown origin: the chal-
lenge continues. Arch Intern Med 2003;163:
1033-41.
4. Bleeker-Rovers CP, Vos FJ, de Kleijn
EMHA, et al. A prospective multicenter
study on fever of unknown origin: the yield of
a structured diagnostic protocol. Medicine
(Baltimore) 2007;86:26-38.
DOI: 10.1056/NEJMp1212725
Copyright © 2013 Massachusetts Medical Society.
their absence is unclear. Certain­
ly, the lengthy, exhaustive process
for applying for a listing can be
discouraging: each component re­
quires a separate detailed, complex
application. Most medicines are
proposed by manufacturers with
a commercial interest in having
their products listed. There has
been no similar advocacy for blood
components that are collected and
prepared by not-for-profit organi­
zations, until now.
There are compelling reasons
to add whole blood and red-cell
concentrates to the list. Blood
transfusion originated as a medi­
cal practice requiring either sur­
gical intervention to join donor to
recipient or a licensed practitioner
to draw and immediately infuse
199