Research Letters
AIDS 2010, 24:28812889
Is Kaposis sarcoma occurring at higher CD4 cell
counts over the course of the HIV epidemic?
Nancy F. Crum-Cianflone a,b, Katherine Huppler
Hullsieka,c, Anuradha Ganesana,d, Amy Weintroba,e,
Jason F. Okulicza,f, Brian K. Agana, the Infectious
Disease Clinical Research Program HIV Working
Group
We evaluated longitudinal rates of Kaposis sarcoma
and trends in CD4 cell counts at the time of Kapo-
sis sarcoma diagnosis during the HIV epidemic
(19852008). Although rates of Kaposis sarcoma
have decreased, cases are now occurring at higher
CD4 cell counts over time, with more than one-
third of cases diagnosed in 20022008 occurring at
CD4 cell counts of at least 350 cells/ml. These data
support future studies evaluating the impact of
highly active antiretroviral therapy initiation at
higher CD4 cell counts to further reduce Kaposis
sarcoma.
During the HIV epidemic, the types and presentations of
cancers have dramatically changed [14]. As an AIDS-
defining cancer, most Kaposis sarcoma cases have
traditionally occurred at low CD4 cell counts
(<200 cells/ml) [5,6]. Although Kaposis sarcoma rates
have decreased [7], it is unknown whether Kaposis
sarcoma will now be observed at higher CD4 cell counts.
We evaluated Kaposis sarcoma rates and trends in CD4
cell counts at Kaposis sarcoma diagnosis among HIV-
infected persons using the US Military HIV Natural
History Study [3,8]. The diagnosis of Kaposis sarcoma
was based on medical record review using standardized
criteria [3]. Rates and rate ratios (overall and for time
spent with CD4 cell count <350 and 350 cells/ml) with
95% confidence intervals (CI) were calculated with
Poisson regression models for four a priori defined
calendar periods (19851990, 19911995, 19962001, Among the 3422 participants with 20 263 person-years of
and 20022008). Participants contributed follow-up
time to all possible calendar periods from baseline
(6 months prior to HIV diagnosis) to the event or
censoring time (last study visit). Among those with
Kaposis sarcoma and a proximal CD4 cell count (within 1
year prior to Kaposis sarcoma diagnosis), participants
were compared by proximal CD4 cell count category
(<350 versus 350 cells/ml) with descriptive statistics
(chi-squared and Wilcoxon tests) as appropriate. Medians
are presented with interquartile ranges (IQR). We also
evaluated factors associated with Kaposis sarcoma during
the highly active antiretroviral therapy (HAART) era (the
ISSN 0269-9370 Q 2010 Wolters Kluwer Health | Lippincott Williams & Wilkins
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
latest of 1 January 1996 or HIV diagnosis date) with time-
updated proportional hazards models.
There were 5067 participants with 39 522 person-years of
follow-up between 1985 and 2008. At HIV diagnosis, the
median age was 28 (IQR 2434) years; 92% were men;
45% were African American and 43% were white. Median
CD4 cell count was 504 (IQR 350672) cells/ml) and
median HIV RNA level (available for 38% of the cohort)
was 4.4 (IQR 3.74.9) log10 copies/ml.
Of the 247 Kaposis sarcoma events during the study
period, there were 52, 138, 38, and 19 during the four
calendar periods, respectively. The rates of Kaposis
sarcoma decreased over time (Table 1). Compared with
19851990, HIV-infected persons in 20022008 had a
72% lower rate of Kaposis sarcoma (relative risk 0.28,
95% CI 0.160.47, P < 0.001). Within each calendar
period, the rates were higher for time spent with CD4 cell
count of less than 350 versus at least 350 cells/ml, although
the rate ratios for those comparisons fell from 9.1 (95% CI
3.722.0) in 19851990 to 6.2 (95% CI 2.316.6) in
20022008.
Among the 247 Kaposis sarcoma patients, 179 (72%) had
a proximal CD4 cell count available. For the four calendar
periods, the proximal CD4 cell count at Kaposis sarcoma
diagnosis was at least 350 cells/ml for 18, 7, 14, and 35%,
respectively (P 0.01, Fig. 1). Participants with proximal
CD4 cell count of less than 350 compared with at least
350 cells/ml at Kaposis sarcoma diagnosis were more
likely to have a prior non-Kaposis sarcoma AIDS event
(47 versus 9%, P < 0.001), diagnosed with HIV in the
pre-HAART era (97 versus 83%, P 0.001), and spent
a smaller percentage of time on antiretroviral therapy
(median of 49 versus 62%, P 0.09); the two groups did
not differ by demographics or HIV duration at the time of
Kaposis sarcoma diagnosis.
follow-up since availability of HAART in 1996, 45 had
Kaposis sarcoma and a proximal CD4 cell count. From
a proportional hazards model considering only time-
updated CD4 cell count, each incremental increase of
50 cells/ml decreased the risk of Kaposis sarcoma by 30%
(hazard ratio 0.70, 95% CI 0.640.76, P < 0.001). In a
model with both CD4 cell count category and HAART
use as time-updated covariates, compared with those with
CD4 cell count of at least 350 cells/ml and on HAART,
those with CD4 cell count of at least 350 cells/ml but not
on HAART had an increased risk of Kaposis sarcoma
(hazard ratio 2.0, 95% CI 0.76.30, P 0.22) that did not
2881
2882 AIDS 2010, Vol 24 No 18

Table 1. Ratesa (overall and by time spent in CD4 cell count categories) of Kaposis sarcoma by calendar period.

By time spent in CD4 cell count categories

Overall CD4 <350 CD4 350

Calendar period Rate (95% CI) Rate (95% CI) Rate (95% CI) Rate ratiob (95% CI) P

19851990 6.5 (5.08.5) 18.6 (11.625.6) 2.0 (0.43.7) 9.1 (3.722.0) <0.001
19911995 12.6 (10.714.9) 21.5 (17.125.9) 1.4 (0.42.4) 15.4 (7.133.1) <0.001
19962001 3.8 (2.85.3) 5.3 (2.77.7) 0.7 (0.01.3) 7.9 (2.723.5) <0.001
20022008 1.8 (1.12.8) 4.6 (1.97.4) 0.8 (0.21.3) 6.2 (2.316.6) <0.001
a
Rates are per 1000 person-years of follow-up and are given with 95% confidence intervals (CI).
b
Comparing rates for time spent with CD4 cell count <350 versus 350 cells/ml.

reach statistical significance, whereas those with CD4 cell
count of less than 350 cells/ml (regardless of HAARTuse)
had an increased risk (hazard ratio 8.3, 95% CI 3.420.2,
P < 0.001).
Our study demonstrates that although the Kaposis
sarcoma rates have declined during the HAART era
and lower CD4 cell counts remain an important risk
factor, a greater proportion of Kaposis sarcoma cases are
now occurring at higher CD4 cell counts. During the late
HAART period, over one-third of Kaposis sarcoma cases
occurred at CD4 cell counts of at least 350 cells/ml.
Clinicians should be aware of these trends and watchful
for the occurrence of Kaposis sarcoma despite robust
CD4 cell counts.
The occurrence of Kaposis sarcoma at higher than
expected CD4 cell counts has been previously reported
[913]. However, our study is unique in that we describe
the changing trends of CD4 cell counts at Kaposis
sarcoma diagnosis over the entire HIV epidemic and
demonstrate a rising proportion of cases at higher CD4
cell counts. To our knowledge, only one other study
examined CD4 cell count trends at Kaposis sarcoma
diagnosis over time, but found no change in CD4 cell
counts between the pre-HAART and post-HAART eras;
however, their population had high rates of drug use and
poor antiretroviral adherence [12], whereas our popu-
lation had free medical care, excellent reported medi-
cation adherence, and low rates of drug use (<1%) [14].
Similar to other studies in the HAARTera [9,10,15], 35%
of our cases were on HAART and 9% had an HIV RNA
level of less than 400 copies/ml at Kaposis sarcoma
diagnosis. Such cases are somewhat surprising because
HAART has reduced the number of Kaposis sarcoma
cases by its effects on HIV suppression and potential
antiangiogenic effects [10,16]. Some Kaposis sarcoma
cases in the setting of HAART may be related to the
immune reconstitution inflammatory syndrome [17,18];
however, most of our cases were not associated with the
introduction of HAART.
Given these trends, determining whether HAART use at
higher CD4 cell counts will reduce the impact of Kaposis
sarcoma is of clinical importance. We found a suggestion
of increased risk of Kaposis sarcoma among those not on
HAART compared with those on HAART with CD4
cell counts of at least 350 cells/ml. Prior studies have
shown that Kaposis sarcoma in the setting of HAART
results in less aggressive and more localized disease [19].
In summary, Kaposis sarcoma remains an important
disease among HIV-infected persons, despite achieve-
ment of higher CD4 cell counts. Among patients with
access to HAART, the proportion of Kaposis sarcoma
cases occurring at high CD4 cell counts appears to be
rising. Future studies are needed to determine whether
earlier HAART initiation will further decrease the
burden of Kaposis sarcoma among HIV-infected persons.

50
CD4 500 Acknowledgements
CD4 350
CD4 200
40 Support for this work (IDCRP-000-04) was provided by
the Infectious Disease Clinical Research Program
Percentage

30 (IDCRP), a Department of Defense (DoD) program

executed through the Uniformed Services University of
20 the Health Sciences. This project has been funded in
whole, or in part, with federal funds from the National
10 Institute of Allergy and Infectious Diseases, National
Institutes of Health (NIH), under Inter-Agency Agree-
0
19851990 19911995 19962001 20022008
ment Y1-AI-5072. The content of this publication is the
sole responsibility of the authors and does not necessarily
Fig. 1. CD4 cell count at diagnosis of Kaposis sarcoma reflect the views or policies of the NIH or the
during the course of the HIV epidemic (19852008). Department of Health and Human Services, the DoD,

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

the Departments of the Army, Navy or Air Force, nor the
US Government. Mention of trade names, commercial
products, or organizations does not imply endorsement
by the US Government. The authors acknowledge that
the research protocol (Incidence and Risk Factors for
AIDS-Defining and Non-AIDS-Defining Cancers in an
HIV-Infected Cohort) received applicable Institutional
Board review and approval.
a
Infectious Disease Clinical Research Program, Uni-
formed Services University of the Health Sciences,
Bethesda, Maryland, bInfectious Disease Clinic, Naval
Medical Center San Diego, San Diego, California,
c
Division of Biostatistics, University of Minnesota,
Minneapolis, Minnesota, dInfectious Disease Clinic,
National Naval Medical Center, Bethesda, Maryland,
e et al. Prognostic significance of immune subset measurement in
Infectious Disease Clinic, Walter Reed Army Medical
Center, Washington, District of Columbia, and fInfec-
tious Disease Service, San Antonio Military Medical
Center, San Antonio, Texas, USA.
Correspondence to Dr Nancy Crum-Cianflone, MD,
MPH, Clinical Investigation Department (KCA), Naval
Medical Center San Diego, 34800 Bob Wilson Drive,
Ste. 5, San Diego, CA 92134-1005, USA.
Tel: +1 619 532 6189 40; fax: +1 619 532 8137;
e-mail: nancy.crum@med.navy.mil
Part of these data will be presented at the 48th Annual
Meeting of the Infectious Disease Society of America;
2124 October 2010; Vancouver, Canada.
Received: 25 June 2010; revised: 12 August 2010;
accepted: 18 August 2010.
References
1. Patel P, Hanson DL, Sullivan PS, Novak RM, Moorman AC,
Tong TC, et al., Adult and Adolescent Spectrum of Disease
Project and HIV Outpatient Study Investigators. Incidence of
types of cancer among HIV-infected persons compared with
the general population in the United States, 19922003. Ann
Intern Med 2008; 148:728736.
2. Engels EA, Pfeiffer RM, Goedert JJ, Virgo P, McNeel TS, Scoppa
SM, et al., for the HIV/AIDS Cancer Match Study. Trends in
cancer risk among people with AIDS in the United States
19802002. AIDS 2006; 20:16451654.
3. Crum-Cianflone N, Hullsiek KH, Marconi V, Weintrob A,
Ganesan A, Barthel RV, et al. Trends in the incidence of cancers
among HIV-infected persons and the impact of antiretroviral
therapy: a 20-year cohort study. AIDS 2009; 23:4150.
4. Shiels MS, Cole SR, Kirk GD, Poole C. A meta-analysis of the
incidence of non-AIDS cancers in HIV-infected individuals.
J Acquir Immune Defic Syndr 2009; 52:611622.
5. Farizo KM, Buehler JW, Chamberland ME, Whyte BM, Froeli-
cher ES, Hopkins SG, et al. Spectrum of disease in persons with
human immunodeficiency virus infection in the United States.
JAMA 1992; 267:17981805.
6. Centers for Disease Control and Prevention (CDC). 1993 revised
classification system for HIV infection and expanded surveil-
lance case definition for AIDS among adolescents and adults.
MMWR Recomm Rep 1992; 41:119.
7. Mocroft A, Kirk O, Clumeck N, Gargalianos-Kakolyris P, Trocha
H, Chentsova N, et al. The changing pattern of Kaposi sarcoma
in patients with HIV, 19942003: the EuroSIDA Study. Cancer
2004; 100:26442654.
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Research Letters 2883
8. Weintrob AC, Fieberg AM, Agan BK, Ganesan A, Crum-
Cianflone NF, Marconi VC, et al. Increasing age at HIV sero-
conversion from 18 to 40 years is associated with favorable
virologic and immunologic responses to HAART. J Acquir
Immune Defic Syndr 2008; 49:4047.
9. Maurer T, Ponte M, Leslie K. HIV-associated Kaposis sarcoma
with a high CD4 count and a low viral load. N Engl J Med 2007;
357:13521353.
10. Mani D, Neil N, Israel R, Aboulafia DM. A retrospective
analysis of AIDS-associated Kaposis sarcoma in patients with
undetectable HIV viral loads and CD4 counts greater than
300 cells/mm(3). J Int Assoc Physicians AIDS Care (Chic Ill)
2009; 8:279285.
11. Krown SE, Lee JY, Dittmer DP, AIDS Malignancy Consortium.
More on HIV-associated Kaposis sarcoma. N Engl J Med 2008;
358:535536.
12. Gallafent JH, Buskin SE, De Turk PB, Aboulafia DM. Profile of
patients with Kaposis sarcoma in the era of highly active
antiretroviral therapy. J Clin Oncol 2005; 23:12531260.
13. Stebbing J, Sanitt A, Teague A, Powles T, Nelson M, Gazzard B,
individuals with AIDS-associated Kaposis sarcoma. J Clin On-
col 2007; 25:22302235.
14. Brodine SK, Shaffer RA, Starkey MJ, Tasker SA, Gilcrest JL,
Louder MK, et al. Drug resistance patterns, genetic subtypes,
clinical features, and risk factors in military personnel with
HIV-1 seroconversion. Ann Intern Med 1999; 131:502506.
15. Chan J, Kravcik S, Angel JB. Development of Kaposis sarcoma
despite sustained suppression of HIV plasma viremia. J Acquir
Immune Defic Syndr 1999; 22:209210.
16. Lebbe C, Blum L, Pellet C, Blanchard G, Verola O, Morel P, et al.
Clinical and biological impact of antiretroviral therapy with
protease inhibitors on HIV-related Kaposis sarcoma. AIDS
1998; 12:F45F49.
17. Connick E, Kane MA, White IE, Ryder J, Campbell TB. Immune
reconstitution inflammatory syndrome associated with Kaposi
sarcoma during potent antiretroviral therapy. Clin Infect Dis
2004; 39:18521855.
18. Nathan RV. Suspected immune reconstitution inflammatory
syndrome associated with the proliferation of Kaposis sarcoma
during HAART. AIDS 2007; 21:775.
19. Nasti G, Martellotta F, Berretta M, Mena M, Fasan M, Di Perri G,
et al. Impact of highly active antiretroviral therapy on the
presenting features and outcome of patients with acquired
immunodeficiency syndrome-related Kaposi sarcoma. Cancer
2003; 98:24402446.
DOI:10.1097/QAD.0b013e32833f9fb8
Erythropoiesis in HIV-infected and uninfected
Malawian children with severe anemia
Job C.J. Calisa,b, Kamija S. Phirib, Raymond J.W.M.
Vetc, Rob J. de Haand, Francis Munthalib, Robert J.
Kraaijenhagen e, Paul J.M. Hulshoff, Malcolm E.
Molyneuxb,g, Bernard J. Brabina,g, Michael Boele van
Hensbroeka,b,g and Imelda Batesg
Anemia is common in HIV infection, but the
pathophysiology is poorly understood. Bone
marrow analysis in 329 severely anemic (hemo-
globin <5 g/dl) Malawian children with (n U 40)
and without (n U 289) HIV infection showed that
HIV-infected children had fewer CD34R hemato-
poietic progenitors (median 10 vs. 15%, P U 0.04)
and erythroid progenitors (2.2 vs. 3.4%, P U 0.05),
but there were no differences in erythrocyte viabi-
lity and maturation in later stages of erythropoiesis.
Despite an HIV-associated reduction in early red
2884 AIDS 2010, Vol 24 No 18
cell precursors, subsequent erythropoiesis appears
to proceed similarly in HIV-infected and HIV-
uninfected children with severe anemia.
Anemia is the most common hematological complication
in HIV-infected adults [1,2] and is positively associated
with disease progression [35]. In adults, anemia results
primarily from reduced erythropoiesis [69]. Infor-
mation about anemia mechanisms in HIV-infected
children is scarce [1013] and there have been no
pediatric studies from sub-Saharan Africa.
We have previously reported that HIV infection was more
common among severely anemic Malawian children than
in a carefully selected control population (13 vs. 6%,
P < 0.001) [14]. The aim of the present study was to
determine whether HIV infection was associated with
reduced erythroid precursor cells, or increased rates of
apoptosis and dyserythropoiesis, and to investigate the
role of cytokines, erythropoietin and plasma vitamin A in
reducing apoptosis.
This study was part of a large casecontrol study
investigating the causes of severe anemia in southern
Malawi [14]. All children aged 660 months with a
primary diagnosis of severe anemia (hemoglobin con-
centration <5 g/dl), and no blood transfusion within the
previous month, were recruited prospectively between
2002 and 2004. HIV-uninfected children aged 6
60 months with no obvious signs of infection and
undergoing elective operations were recruited as controls.
An automated full blood count, including reticulocytes,
was performed on peripheral blood samples (Coulter
counter, Beckman Coulter, Durban, South Africa).
Malaria slides were read by two independent micro-
scopists. Stained bone marrow aspirate smears from all
children were used to determine the myeloid : erythroid
ratio [15] and assess dyserythropoiesis, which was defined
and scored according to a published protocol [16].
C-reactive protein and erythropoietin were determined
using a Roche p800/e170 system (Roche, Basel,
Switzerland). Inflammatory cytokine profiles were
measured by Cytometric Bead Array flow cytometry
(FACS-Calibur, BD Biosciences, San Jose, California,
USA). Serum vitamin A (retinol) was measured using
HPLC [17]. HIV testing was performed using two rapid
tests (Determine, Abbott-Laboratories, Tokyo, Japan;
Unigold, Trinity-Biotech, Dublin, Ireland). Reactive
results in children less than 18 months of age and
discordant results were resolved by PCR [18].
Fresh bone marrow aspirates underwent automated cell
count (Coulter counter, Beckman Coulter) and four
color flow cytometry (FACS-Calibur, BD Biosciences).
Bone marrow cells were separated and incubated with
different combinations of CD14-PE-Cy5 (Tuk4), CD34-
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
FITC/PE (QBEND/10), CD36-PE (CLB-IVC7),
CD235a-FITC (CLB-AME-1) (all from Sanquin
Reagents, Amsterdam, The Netherlands), laser dye
styril-751 (LDS, Applied Laser Technology, Maarheeze,
The Netherlands) and Annexin V and propidium iodide
(IQ-products, Groningen, The Netherlands) [19].
Patient characteristics and hematological variables were
compared using the x2-test, Fishers exact test, Students
t-test and the MannWhitney U-test. Correlations were
assessed using the Pearson productmoment correlation
coefficient or Spearmans rank correlation coefficient. A
two-sided significance level was set at P 0.05.
Complete data (bone marrow samples and HIV tests) for
this study were available for 329 of 381 children enrolled
in our original casecontrol study. The original study
had shown that bacteremia, malaria, hookworm, HIV,
glucose-6-phosphate dehydrogenase (G6PD ) deficiency
and vitamin A and B12 deficiency were associated with
severe anemia. Iron deficiency was negatively associated
with severe anemia. Folate deficiency and sickle cell
disease were uncommon [14].
Forty of the 329 children (12%) were infected with HIV.
Their median age was 25 months compared with
16 months for HIV-uninfected children (P < 0.01). No
significant differences were found between HIV-infected
and uninfected children with regard to other baseline
characteristics, mean hemoglobin levels (P 0.67) or
other erythrocytic indices (Table 1) [20].
HIV-infected children had fewer bone marrow CD34
hematopoietic progenitors, erythroid progenitor cells and
erythroid precursor cells than HIV-uninfected children,
but numbers of bone marrow proerythroblasts, basophilic
and polychromatic erythroblasts and peripheral blood
reticulocytes were similar (Table 1). Correction for age or
malaria did not alter the results (data not displayed).
Dyserythropoiesis occurred in 2.8% and 3.8% of erythroid
precursors in HIV-infected and uninfected children,
respectively (P 0.12, Table 1). The proportions of viable
erythroid precursor cells and those at various stages of
apoptosis were similar between the two groups (Table 1).
The proportions of dyserythropoietic cells and red cells
undergoing early apoptosis were positively correlated
(range r 0.34, P 0.01). There were no correlations
(r 0.140.15) between the proportion of either
dyserythropoietic or apoptotic cells and the peripheral
blood concentrations of cytokines tumor necrosis factor-a
(P 0.90 and 0.28), interferon-g (P 0.15 and 0.36),
interleukin-10 (P 0.74 and 0.19), erythropoietin
(P 0.22 and 0.83) or vitamin A (P 0.83 and 0.22).
This study is the first detailed prospective analysis of
erythropoiesis using bone marrow samples and flow
cytometry in HIV-infected children. HIV-infected
 Table 1. Characteristics and hematological parameters in HIV-infected and uninfected children with severe anemia and a control population of children without HIV infection or severe anemia.

HIV-infected children n 40 HIV-uninfected children n 289 PM Control n 18

Characteristics
Age median, IQR (months) 24.9 (15.638.4) 15.8 (10.225.5) <0.01 24.0 (11.332.8)
Boys 17/40 (43%) 144/289 (50%) 0.39 15/18 (83%)
Prior transfusion 7/40 (18%) 41/287 (14%) 0.59 0/18 (0%)
Wasting 6/33 (18%) 37/261 (14%) 0.54 3/12 (25%)
Iron deficiency 5/21 (24%) 34/155 (22%) 0.85 4/12 (33%)
Malaria parasitemia 23/39 (59%) 170/289 (59%) 0.99 2/17 (12%)
CRP, median and IQR (mg/l) 117 (47193) n 38 95 (42153) n 269 0.73 2.5 (1.84.4) n 18
Automated count
Hemoglobin concentration, mean  SD (g/dl) 3.6  0.7 n 40 3.6  0.8 n 289 0.67 9.7  1.8 n 18
MCV, mean  SD (fl) 81.1  13.7 n 35 83.3  15.6 n 247 0.27 71.6  7.1 n 16
MCHC, mean  SD (g/dl) 32.4  3.1 n 35 32.5  7.2 n 245 0.68 33.0  3.6 n 16
RDW, mean  SD (%) 25.2  8.7 n 35 24.4  7.4 n 246 0.50 18.0  3.8 n 16
Reticulocytes, median and IQR (109/l) 58.6 (30.388.2) n 32 52.7 (30.291.7) n 209 0.85 70.7 (54.3117.9) n 13
Light microscopy
Myeloid : erythroid ratio
Decreased (<2.0 : 1) 29/34 (85%) 194/261 (74%) 0.38 4/18 (22%)
Normal (2.04.9 : 1) 4/34 (12%) 54/261 (21%) 10/18 (56%)
Increased (>5.0 : 1) 1/34 (3%) 13/261 (5%) 4/18 (22%)
Erythroid cells
Proerythroblasts, median and 0.8 (0.01.6) n 34 0.4 (0.01.5) n 261 0.33 0.0 (0.00.8) n 18
IQR (% of nucleated cells)
Basophilic erythroblast, median 0.8 (0.02.4) n 34 0.8 (0.01.6) n 261 0.55 0.8 (0.01.6) n 18
and IQR (% of nucleated cells)
Ortho and polychromatic erythroblast, 37  15 n 34 36  16 n 261 0.69 18.8  12.3 n 18
mean  SD (% of nucleated cells)
Dyserythropoiesis
Dyserythropoietic cells, mean  SD 2.8  2.2 n 25 3.8  3.0 n 213 0.12 1.2  1.6 n 13
(% of erythrocytic precursors)
Coulter counter
Cellularity
Nucleated bone marrow cells, 62.2 (42.6108.3) n 32 76.6 (45.6119.6) n 246 0.37 91.6 (61.2114.0) n 16
median and range (109/l)
Flow cytometry
Cells
All CD34 hematopoietic progenitors, CD34 10% (520%) n 34 15% (730%) n 242 0.04 4.4% (2.88.1%) n 17
median and IQR (% of
mononucleated fraction)
Erythroid progenitor cells, median CD34; CD36; CD14- 2.2% (0.84.4%) n 27 3.4% (1.56.6%) n 210 0.05 0.5% (0.40.7%) n 16
and IQR (% of mononucleated fraction)
Erythroid precursor cells, median CD235, LDS 17.9% (13.030.8%) n 35 25.6% (14.938.3%) n 248 0.06 12.4% (10.918.6%) n 17
and IQR (% of mononucleated fraction)
Apoptosis of erythroid precursors
Viable cells, median and IQR CD235, Annexin, PI 87% (7394%) n 15 85% (6891%) n 78 0.25 81.7% (6596%) n 15
Early apoptotic, median and IQR CD235; Annexin, PI 9.3% (4.419.8%) n 15 12.1% (6.322.6%) n 78 0.23 14.7% (3.430.1%) n 15
Late apoptotic, median and IQR CD235, Annexin, PI 2.1% (1.24.9%) n 15 2.6% (1.05.5%) n 78 0.67 1.8% (0.32.7%) n 15

Wasting was defined as a weight for height Z-score of less than 2 [20]. Dyserythropoiesis was defined as multinuclearity, karyorrhexis, intercellular chromatin bridging or incomplete mitoses. Early
apoptosis refers to the expression of phosphatidylserine only, whereas in late apoptosis also propidium iodide was detected. In viable cells neither of these dyes were detected [19]. LDS was used to
stain DNA. CRP, C-reactive protein; IQR, inter quartile range; LDS, laser dye styril-751; MCHC, mean corpuscular hemoglobin concentration; MCV, mean corpuscular volume; PI, propidium iodide;
RDW, red cell distribution width.
Research Letters

M
P-value refers to the difference between HIV-infected and uninfected children.

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
2885
2886 AIDS 2010, Vol 24 No 18
children with severe anemia had 33% fewer CD34
hematopoietic progenitors and 35% less erythroid pro-
genitors in their bone marrow than uninfected children.
This supports the hypothesis that red cell production
failure is an important cause of severe anemia in HIV-
infected children and may be caused by a reduced stem
cell capacity [21]. However, the proportion of more
mature erythroid precursor cells in bone marrow or
peripheral blood (reticulocytes) did not differ between
the two groups, suggesting that HIV-uninfected children
had less efficient later stages of erythropoiesis than HIV-
infected children. This is supported by the trend toward
less dyserythropoiesis and apoptosis in HIV-infected
children, but is in contrast to previous reports suggesting
that anemia due to dyserythropoiesis is more common in
later stages of HIV disease [2,10]. Alternatively, the lost
CD34 cells in HIV-infected children may have been
precursors that were not committed to erythropoiesis.
HIV infection affects hematopoietic processes possibly
through abnormal expression of cellular genes and cyto-
kines [22]. The African HIV-1 subtype C can directly
infect CD34 hematopoietic progenitors [23]. Unlike
previous studies [24,25], we found no association between
dyserythropoiesis or apoptosis and altered cytokine levels
or vitamin A deficiency [26,27], despite 90% of children
having vitamin A deficiency [14]. More intensive
investigations might identify cytokines that affect regulat-
ory signals and could potentially be therapeutic targets to
reduce hemopoietic inhibition in HIV patients.
In common with previous studies, we did not find any
differences in peripheral blood erythrocytic indices or
bone marrow microscopy in HIV-infected compared
with uninfected children [68,10,28], possibly because of
the multifactorial cause of anemia in African children
[14,29].
None of the children were on antiretroviral therapy,
which can exacerbate blood and bone marrow abnorm-
alities [30]. Although not all tests were done on all
children, the large sample size increases confidence that
the study sample was representative.
The findings in these severely anemic Malawian children
indicate that despite an HIV-associated reduction in early
red cell precursors, subsequent erythropoiesis appears to
proceed similarly in HIV-infected and HIV-uninfected
children with severe anemia.
Acknowledgements
This study was funded by the Wellcome Trust and
supported by independent grants of the Nutricia
Research Foundation and the Ter Meulen Fund, Royal
Netherlands Academy of Arts and Sciences.
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
We thank the parents and guardians of the children
admitted to the study, the SevAna study team, the staff
of the Queen Elizabeth Central Hospital, Chikwawa
District Hospital and Wellcome Trust Research Labora-
tories, and particularly, S.M. Graham, E.M. Molyneux,
M. Cornelissen, M. Beld, L. van Lieshout, F.A. Wijnberg
and W.J. van Luling for their contributions to the study.
a
Global Child Health Group, Emma Childrens Hospi-
tal, Academic Medical Center, Amsterdam, The
Netherlands, bMalawiLiverpool Wellcome Trust Clin-
ical Research Programme, College of Medicine, Uni-
versity of Malawi, Blantyre, Malawi, cDepartment of
Specialized Hematology, dDepartment of Clinical
Epidemiology and Biostatistics, Academic Medical
Center, Amsterdam, eDepartment of Clinical Chemis-
try, Meander Medical Center, Amersfoort, fDivision of
Human Nutrition, Wageningen University, Wagenin-
gen, The Netherlands, and gLiverpool School of
Tropical Medicine, Liverpool, UK.
Correspondence to Job C.J. Calis, MD, PhD, Emma
Childrens Hospital, Academic Medical Center, Mei-
bergdreef 9, 1105 AZ Amsterdam, The Netherlands.
Tel: +31 20 5667150; fax: +31 20 6917735;
e-mail: job.calis@gmail.com
Received: 27 November 2009; revised: 17 August
2010; accepted: 26 August 2010.
References
1. Spivak JL, Bender BS, Quinn TC. Hematologic abnormalities in
the acquired immune deficiency syndrome. Am J Med 1984;
77:224228.
2. Zon LI, Arkin C, Groopman JE. Haematologic manifestations of
the human immune deficiency virus (HIV). Br J Haematol 1987;
66:251256.
3. Mocroft A, Kirk O, Barton SE, Dietrich M, Proenca R, Colebun-
ders R, et al. Anaemia is an independent predictive marker for
clinical prognosis in HIV-infected patients from across Europe.
EuroSIDA study group. AIDS 1999; 13:943950.
4. Moore RD, Keruly JC, Chaisson RE. Anemia and survival in HIV
infection. J Acquir Immune Defic Syndr Hum Retrovirol 1998;
19:2933.
5. Sullivan PS, Hanson DL, Chu SY, Jones JL, Ward JW. Epide-
miology of anemia in human immunodeficiency virus (HIV)-
infected persons: results from the multistate adult and adoles-
cent spectrum of HIV disease surveillance project. Blood 1998;
91:301308.
6. Bain BJ. The haematological features of HIV infection. Br J
Haematol 1997; 99:18.
7. Bain BJ. Pathogenesis and pathophysiology of anemia in HIV
infection. Curr Opin Hematol 1999; 6:8993.
8. Moses A, Nelson J, Bagby GC Jr. The influence of human
immunodeficiency virus-1 on hematopoiesis. Blood 1998;
91:14791495.
9. Volberding PA, Baker KR, Levine AM. Human immunodefi-
ciency virus hematology. Hematology Am Soc Hematol Educ
Program 2003:294313.
10. Ellaurie M, Burns ER, Rubinstein A. Hematologic manifestations
in pediatric HIV infection: severe anemia as a prognostic
factor. Am J Pediatr Hematol Oncol 1990; 12:449453.
11. Meira DG, Lorand-Metze I, Toro ADC, Silva MTN, Vilela
MMDS. Bone marrow features in children with HIV infection
and peripheral blood cytopenias. J Trop Pediatr 2005; 51:114
119.

12. Mueller BU, Tannenbaum S, Pizzo PA. Bone marrow aspirates
and biopsies in children with human immunodeficiency virus
infection. J Pediatr Hematol Oncol 1996; 18:266271.
13. Sandhaus LM, Scudder R. Hematologic and bone marrow
abnormalities in pediatric patients with human immunodefi-
ciency virus (HIV) infection. Pediatr Pathol 1989; 9:277288.
14. Calis JCJ, Phiri KS, Faragher EB, Brabin BJ, Bates I, Cuevas LE,
et al. Factors associated with severe anemia in Malawian
children. N Engl J Med 2008; 358:888899.
15. Bain BJ, Clark DM, Lampert IA, Wilkins ES. Bone Marrow
Pathology. Oxford: Blackwell Science; 2007.
16. Newton CR, Warn PA, Winstanley PA, Peshu N, Snow RW,
Pasvol G, et al. Severe anaemia in children living in a malaria
endemic area of Kenya. Trop Med Int Health 1997; 2:165178.
17. Bieri JG, Tolliver TJ, Catignani GL. Simultaneous determination
of alpha-tocopherol and retinol in plasma or red cells by high
pressure liquid chromatography. Am J Clin Nutr 1979; 32:
21432149.
18. Molyneux EM, Walsh AL, Malenga G, Rogerson S, Molyneux
ME. Salmonella meningitis in children in Blantyre, Malawi,
19961999. Ann Trop Paediatr 2000; 20:4144.
19. Koopman G, Reutelingsperger CP, Kuijten GA, Keehnen RM,
Pals ST, van Oers MH. Annexin V for flow cytometric detection
of phosphatidylserine expression on B cells undergoing apop-
tosis. Blood 1994; 84:14151420.
20. Dibley MJ, Goldsby JB, Staehling NW, Trowbridge FL. Devel-
opment of normalized curves for the international growth
reference: historical and technical considerations. Am J Clin
Nutr 1987; 46:736748.
21. Redd AD, Avalos A, Phiri K, Essex M. Effects of HIV type 1
infection on hematopoiesis in Botswana. AIDS Res Hum Retro-
viruses 2007; 23:9961003.
22. Koka PS, Reddy ST. Cytopenias in HIV infection: mechanisms
and alleviation of hematopoietic inhibition. Curr HIV Res 2004;
2:275282.
23. Redd AD, Avalos A, Essex M. Infection of hematopoietic
progenitor cells by HIV-1 subtype C, and its association with
anemia in southern Africa. Blood 2007; 110:31433149.
24. Ellaurie M, Rubinstein A. Elevated tumor necrosis factor-alpha
in association with severe anemia in human immunodeficiency
virus infection and Mycobacterium avium intracellulare infec-
tion. Pediatr Hematol Oncol 1995; 12:221230.
25. Testa U. Apoptotic mechanisms in the control of erythropoi-
esis. Leukemia 2004; 18:11761199.
26. Herault O, Domenech J, Georget MT, Clement N, Colombat P,
Binet C. All-trans retinoic acid prevents apoptosis of human
marrow CD34R cells deprived of haematopoietic growth
factors. Br J Haematol 2002; 118:289295.
27. Zauli G, Visani G, Vitale M, Gibellini D, Bertolaso L, Capitani S.
All-trans retinoic acid shows multiple effects on the survival,
proliferation and differentiation of human fetal CD34R hae-
mopoietic progenitor cells. Br J Haematol 1995; 90:274282.
28. Galli L, de Martino M, Rossi ME, Panza B, Farina S, Vierucci A.
Hemochrome parameters during the first two years of life in
children with perinatal HIV-1 infection. Pediatr AIDS HIV
Infect 1995; 6:340345.
29. Brabin BJ, Premji Z, Verhoeff F. An analysis of anemia and child
mortality. J Nutr 2001; 131:636S645S.
30. Harris CE, Biggs JC, Concanon AJ, Dodds A. Peripheral blood
and bone marrow findings in patients with acquired immune
deficiency syndrome. Pathology 1990; 22:206211.
DOI:10.1097/QAD.0b013e32833fed27
Maraviroc does not affect humoral response to the
pandemic influenza A-H1N1v 2009 adjuvanted
vaccine in HIV-1-infected patients
Ana Canestria,c, Anne Krivineb, Lambert Assoumouc,
Monique Le Corre d , Flore Rozenberg b , Anne-
Genevieve Marcelinc,e,f, Luminita Schneidera,c, Assia
Samrid, Guislaine Carcelaina,e,g, Brigitte Autrana,e,g,
Christine Katlamaa,c,e and Amelie Guihota,e,g
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Research Letters 2887
Immune changes induced by the CCR5 antagonist
maraviroc raise the question of an impairment of
responses to vaccines. We evaluated the immuno-
genicity of the adjuvanted pandemic influenza
A-H1N1v 2009 vaccine in HIV-1-infected patients
with suppressed HIV viremia with or without
a maraviroc-containing regimen. Seroprotection,
seroconversion, and geometric mean titer ratio of
specific antibody titers did not differ between
groups. These results suggest that maraviroc does
not significantly affect the immune response to this
adjuvanted vaccine.
Maraviroc (MVC) blocks HIV-1 entry into the CD4 host
cells [1] by binding to the chemokine receptor CCR5.
Several studies have shown an increased severity of West
Nile virus infection and tick-borne encephalitis virus in
patients with the CCR5D32 deletion [2,3], suggesting a
key role of CCR5 in immune control of viruses. In mice,
a CCR5 gene deletion is associated with a more severe
influenza infection [4]. Because CD4CCR5 T cells of
memory phenotype help with differentiation of B cells
into antibody producing plasma cells and help CD8
antiviral effector T cells, MVC may influence immune
response to live or attenuated viruses. These observations
led some authors to recommend avoiding yellow fever
vaccination in HIV-1-infected patients treated with
MVC [5]. However, the MVC immunomodulatory
effects on responses to vaccines remain unknown.
Although one dose of the 2009 pandemic influenza A-
H1N1v adjuvanted vaccine had been shown to trigger a
protective antibody response to influenza in HIV-1
patients [6], concerns were raised about a possible
interference of MVC treatment with responses to this
highly immunogenic vaccine.
To test such hypothesis, we conducted a prospective
casecontrol study to compare the humoral immuno-
genicity of the adjuvanted pandemic influenza A-H1N1v
2009 vaccine. HIV-1 patients on combined antiretroviral
therapy (cART) containing MVC (MVC group) or not
(control group) were studied after providing informed
consent according to ethical recommendations.
All patients had to be on cART with a viral load less than
50 copies/ml. Control patients were matched for sex, age
(5 years), and CD4 cell counts (75 cells/ml). All
patients received one dose (3.75 mg hemagglutinin) of the
2009 Pandemrix influenza A-H1N1v adjuvanted vaccine
at day 0 (D0) between 23 November 2009 and 4 January
2010. Influenza-specific antibody titers were measured at
D0 and D21 by using a hemagglutination inhibition assay
(HIA) modified from Kendal and Skehel [7] by the use of
human O Rh- red blood cells and the nonadjuvanted
influenza A-H1N1v 2009 vaccine Panenza as antigen.
Immunogenicity was evaluated upon percentages of
seroprotection (antibody titers 1/40), seroconversion
2888 AIDS 2010, Vol 24 No 18

(antibody titers <1/10 to 1/40) or four-fold increase in in the MVC group and 12/ml (282591) in the
antibody titers, and geometric mean titer (GMT) ratio control group. All patients remained aviremic with a viral
(D21/D0). Any clinical events, changes in CD4 cell load less than 50 copies/ml at D21 except for two patients
counts, and HIV viral load were recorded. who experienced a blip at week 4 that was subsequently
suppressed.
We included 22 patients in the MVC group and 29 in
the control group. Baseline characteristics were similar Influenza immunogenicity analysis showed seroprotec-
between both groups and showed a vast majority of men tion at D0 in four of 51 patients (two in the MVC and two
(86 and 93%), a median (minmax) age of 50 (4074) in the control group, Fig. 1a). Patients who received the
and 53 (3778) years, and a median CD4 cell count of seasonal influenza vaccine before the pandemic one had a
459/ml (411002) and 520/ml (1151065) in the MVC higher baseline antibody GMT for influenza A-H1N1v
and the control groups, respectively. Patients had been virus than nonseasonal vaccinated patients: 17.72 (13.05
infected for a median duration of 20 (925) and 15 (2 24.07) vs. 8.42 (4.8410.87), respectively (P < 0.001).
25) years, with a nadir of CD4 cell count of 70/ml (3 After vaccination, seroprotection was obtained at D21 for
274) and 109/ml (1571) in the MVC and control 20 of 22 (91%) patients in the MVC group and 27 of
groups, respectively. Only nine of 22 (41%) patients in the 29 (93%) in the control group (P 1.00; Fig. 1b).
MVC group and 11 of 29 (38%) in the control group had Seroconversion was observed at D21 in 17 of 22 patients
previously received the seasonal 2009/2010 influenza (77%) in the MVC group and 25 of 29 (86%) in the
vaccine (P 1.00) with a median delay of 40 (2179) control group (P 0.47; Fig. 1c). The GMT ratios (D21/
days before the pandemic vaccine. Five patients in each D0) were 8.3 [95% confidence interval (CI) 5.313] for
group experienced side effects: three patients in the MVC the MVC group and 11.6 (95% CI 7.418.2) for the
group and five in the control group experienced systemic control group (P 0.35; Fig. 1d). The prior seasonal
reactions (fever and fatigue). Two patients in the MVC influenza vaccine did not influence the GMT at D21:
group and three in the control group complained for a 113.2 (72.6176.5) vs. 113.8 (83.8154.5), respectively
local reaction (pain and redness). Median CD4 cell count (P 0.53). Neither the duration of HIV infection nor the
changes between D21 and D0 were 11/ml (248107) CD4 nadir or the CD4 cell counts at D0 were associated

(a)
(I) MVC (II) Controls

100 100

80 80
Participants
Participants

D0 D0
60 60
D21 D21
40 40

20 20

0 0
5

10

20

40

80

80

60

10

20

40

80

80

60
16

32

64

16

32

64
12

25

12

25

Titer Titer

(b) Seroprotection rate (D21) (c) Seroconversion rate (D21) (d) Increase in GMT (D21/D0)
20
P = 1,000 P = 0,351
P = 0,474
Fold increase in GMT

100 n = 20/22 n = 27/29 100 16

n = 25/29
n = 17/22
80 80 n = 29
12
Participants

Participants

60 60 n = 22
8
40 40

20 20 4

0 0 0
MVC Controls MVC Controls MVC Controls

Fig. 1. Antibody response to influenza after pandemic A-H1N1v 2009 vaccination. (a) Reverse cumulative distribution curves
of antibody titers in serum samples obtained at day 0 (D0) of vaccination ( ) and at day 21 (D21) after vaccination ( ) in
22 maraviroc (MVC)-treated patients (I) and 29 controls with other antiretroviral therapy (II). (b) Seroprotection rate at D21
after vaccination: percentage of vaccinated participants with influenza A-H1N1v antibody titers >1/40 (McNemar test).
(c) Seroconversion rate at D21 after vaccination: percentage of vaccinated participants with a four-fold increase in influenza-
specific A-H1N1v antibody titers from D0 to D21 after vaccination, or a titer <1/10 at D0 and a titer 1/40 at D21 after vaccination
(McNemar test). (d) Fold increase in geometric mean titer (GMT) of antibody titers between D21 and D0 of vaccination (Wilcoxon
test).

Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.

to the antibody response to the pandemic vaccine. Finally,
the duration of exposure to MVC did not significantly
influence the response to this adjuvanted vaccine.
Our study showed that MVC did not significantly
affect the antibody response to the influenza A-H1N1v
adjuvanted vaccine in HIV-1-infected patients despite a
slightly lower antibody levels in patients receiving MVC.
In addition, this influenza A-H1N1v adjuvanted vaccine
showed a good immunogenicity and a good tolerance
in HIV-1 patients, with a persisting undetectable viral
load while on cART. Noteworthy, the seasonal influenza
vaccine may have conferred some protection to the
pandemic A-H1N1v virus, suggesting a cross-reactivity
between the pandemic A-H1N1v and the A-H1N1/
Brisbane/59/07 strain contained in the seasonal vaccine.
Altogether these results, observed, however, in a small
series of patients, suggest that the MVC CCR5 antagonist
does not impair the response to influenza vaccines in
stably suppressed HIV-1-infected patients with sub-
normal CD4 cell counts. It remains to be determined
whether such a lack of negative immunomodulatory
effect of CCR5 antagonists reflects the strong immuno-
genicity of this adjuvanted vaccine or can be extrapolated
to any vaccine preparation.
Acknowledgements
The authors want to thank the Institut de Microbiologie
et Maladies Infectieuses for his organizational support.
A.C., C.K., G.C., B.A., and A.G. designed the study and
wrote the article. A.C. and L.S. selected patients. M.L.C.
and A.S. organized reception of blood samples. A.K. and
F.R. performed HIA analysis. A.G.M. performed viro-
logical studies. L.A. performed statistical analysis. L.A.
and M.L.C. contributed equally to this study.
a Services, CDC. Concepts and procedures for laboratory-based
AP-HP, Hopital Pitie-Salpetriere, Maladies Infectieuses
et Tropicales, bAP-HP, Hopital Saint Vincent Paul,
Laboratoire de Virologie, c INSERM, UMR 943,
d
INSERM, UMR S945, eUniversite Pierre et Marie
Curie- Paris 6, fAP-HP, Hopital Pitie Salpetriere,
Copyright Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.
Research Letters 2889
Laboratoire de Virologie, and gAP-HP, Hopital Pitie
Salpetriere, Laboratoire dImmunologie cellulaire et
tissulaire, Paris, France.
Correspondence to Dr Ana Canestri, Service des
Maladies Infectieuses et Tropicales, Hopital Pitie-
Salpetriere, 47 bd de lHopital, 75013 Paris, France.
Tel: +33 1 42 16 01 03; fax: +33 1 42 16 01 65;
e-mail: ana.canestri@psl.aphp.fr
The present work will be presented to the AIDS
Vaccine 2010 International Conference in Atlanta,
Georgia (abstract #188141).
Received: 2 August 2010; revised: 1 September 2010;
accepted: 2 September 2010.
References
1. Dorr P, Westby M, Dobbs S, Griffin P, Irvine B, Macartney M,
et al. Maraviroc (UK-427,857), a potent, orally bioavailable, and
selective small-molecule inhibitor of chemokine receptor CCR5
with broad-spectrum antihuman immunodeficiency virus type 1
activity. Antimicrob Agents Chemother 2005; 49:47214732.
2. Lim JK, Louie CY, Glaser C, Jean C, Johnson B, Johnson H, et al.
Genetic deficiency of chemokine receptor CCR5 is a strong
risk factor for symptomatic West Nile virus infection: a meta-
analysis of 4 cohorts in the US epidemic. J Infect Dis 2008;
197:262265.
3. Kindberg E, Mickiene A, Ax C, Akerlind B, Vene S, Lindquist L,
et al. A deletion in the chemokine receptor 5 (CCR5) gene is
associated with tickborne encephalitis. J Infect Dis 2008; 197:
266269.
4. Dawson TC, Beck MA, Kuziel WA, Henderson F, Maeda N.
Contrasting effects of CCR5 and CCR2 deficiency in the pul-
monary inflammatory response to influenza A virus. Am J Pathol
2000; 156:19511959.
5. Roukens AH, Visser LG, Kroon FP. A note of caution on yellow
fever vaccination during maraviroc treatment: a hypothesis on a
potential dangerous interaction. AIDS 2009; 23:542543.
6. Launay O, Desaint C, Durier C, Loulergue P, Duval X, Pialoux G,
et al., for the Natl Network of Clin Investigation in Vaccinology
and ANRS. Immunogenicity of one dose of influenza A H1N1v
2009 vaccine formulated with and without AS03A-adjuvant in
HIVR adults: preliminary report of the ANRS 151 randomized
HIFLUVAC trial. In: 17th Conference on Retroviruses and
Opportunistic Infections; San Francisco; 2010.
7. Kendal AP, Skehel JJ. US Department of Health and Human
influenza surveillance. Kendal AP, Skehel JJ, editors. Atlanta, GA:
US Department of Health and Human Services, CDC; 1982.
DOI:10.1097/QAD.0b013e3283402bc1