MAJOR ARTICLE

Effect of Perinatal Antiretroviral Drug Exposure

on Hematologic Values in HIV-Uninfected
Children: An Analysis of the Women
and Infants Transmission Study
Susan E. Pacheco,1 Kenneth McIntosh,2 Ming Lu,3 Lynne M. Mofenson,4 Clemente Diaz,5 Marc Foca,6
Margaret Frederick,3 Edward Handelsman,7 Karen Hayani,8 and William T. Shearer,1 for the Women
and Infants Transmission Studya
1
Baylor College of Medicine, Houston, Texas; 2Childrens Hospital Boston, Boston, Massachusetts; 3Clinical Trials and Surveys Corporation,
Baltimore, and 4National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland; 5University
of Puerto Rico, San Juan; 6Columbia Presbyterian Hospital, New York, and 7State University of New York at Downstate, Brooklyn, New York;
8
University of Illinois at Chicago, Chicago

Background. With the increasing use of antiretroviral (ARV) drugs to prevent mother-to-child transmission
of human immunodeficienc virus (HIV), large numbers of infants are exposed, with possible consequent toxicity.
Methods. Hematologic values in 1820 uninfected HIV- and ARV-exposed children were compared with those
in 351 ARV-unexposed children from the Women and Infants Transmission Study. Hemoglobin concentrations
and platelet, neutrophil, lymphocyte, and CD4+ and CD8+ cell counts were analyzed at birth and ages 2, 6, 12,
18, and 24 months. Multivariate analysis was conducted age 02 and 624 months, with adjustment for multiple
cofactors.
Results. Hemoglobin concentrations and neutrophil, lymphocyte, and CD4+ cell counts were significantl lower
at age 02 months in infants exposed to ARV drugs than in those who were not. At 624 months, differences in
hemoglobin concentrations and neutrophil counts were no longer significant whereas differences in platelet,
lymphocyte, and CD4+ cell counts persisted and CD8+ cell counts became significantl lower. In comparison with
ARV monotherapy, combination therapy was associated with larger decreases in neutrophil, lymphocyte, and CD8+
cell counts at age 02 months but with only differences in CD8+ cell counts at age 624 months. Clinically
significan abnormalities were rare and did not differ by exposure to ARV drugs.
Conclusion. Infants exposed to ARV drugs have small but significan differences in several hematologic pa-
rameters for the firs 24 months of life. These results indicate the need for long-term follow-up of uninfected
infants with ARV exposure.

There has been a dramatic decrease in perinatal HIV
infection in the United States and other resource-rich
Received 2 February 2006; accepted 17 May 2006; electronically published 11
September 2006.
Presented in part: XV International AIDS Conference, Bangkok, Thailand, 11
16 July 2004 (abstract ThPeB7024).
Financial support to local Clinical Research Centers: National Institutes of Health
(grants GCRC RR00188 to Baylor College of Medicine, GCRC RR00645 to Columbia
University, and GCRC RR02172 to Childrens Hospital Boston).
Potential conflicts of interest: none reported.
a
Study group members are listed after the text.
Reprints or correspondence: Dr. Kenneth McIntosh, Div. of Infectious Diseases,
Childrens Hospital Boston, 300 Longwood Ave., Boston, MA 02115 (Kenneth
.mcintosh@childrens.harvard.edu).
The Journal of Infectious Diseases 2006; 194:108997
 2006 by the Infectious Diseases Society of America. All rights reserved.
0022-1899/2006/19408-0009$15.00
countries since 1994, when Pediatric AIDS Clinical Tri-
als Group protocol 076 showed that administration of
zidovudine to the HIV-infected woman during preg-
nancy and labor and to her newborn reduced the risk
of mother-to-child HIV transmission by nearly 70%
[1]. Subsequently, it was recognized that the use of
combination antiretroviral (ARV) drug regimens dur-
ing pregnancy could further reduce transmission [2].
Current recommendations for the prevention of moth-
er-to-child transmission in the United States include
the use of ARV combination therapy during pregnancy
for women with HIV RNA loads 11000 copies/mL, with
the use of zidovudine prophylaxis alone restricted to
women with HIV RNA loads !1000 copies/mL [3].
Although mother-to-child transmission in the United

Hematologic Values in ARV-Exposed Infants JID 2006:194 (15 October) 1089

States has dramatically decreased, from 25% to !2% [2], this
significan advance entails the in utero and neonatal exposure
of many infants to 1 drug of unknown toxicity.
Although data regarding the short-term safety of exposure
to ARV drugs for uninfected infants have been reassuring [4
6], there is limited information about long-term effects. Pre-
clinical data have indicated that some ARV drugs, particularly
those in the nucleoside-analogue reverse-transcriptase inhibitor
(NRTI) class, may be associated with mitochondrial dysfunc-
tion [7, 8], and it has been suggested that some infants with
in utero or neonatal exposure to these drugs may have labo-
ratory or clinical signs of mitochondrial dysfunction. The
French Perinatal Cohort Study Group studied 2644 uninfected
HIV- and ARV-exposed children, and they reported an 18-
month incidence of clinical symptoms (primarily neurologic)
of mitochondrial dysfunction of 0.26% and a mortality inci-
dence of 0.07% [9]. The same group also reported an increased
risk of simple febrile seizures during the firs 18 months of life
and persistently lower (but clinically insignificant neutrophil,
lymphocyte, and platelet counts in infants associated with in
utero NRTI exposure [10, 11]. Although these clinical abnor-
malities and mortality finding have not been duplicated to
date in other cohorts in the United States and Europe [12, 13],
the European Collaborative Study reported similar hematologic
abnormalities in uninfected infants with in utero or neonatal
ARV exposure [14, 15].
We report here an analysis of data from the Women Infant
Transmission Study (WITS), a longitudinal cohort of HIV-in-
fected mothers and their infants from the United States and
Puerto Rico that explores the effect of perinatal exposure to ARV
drugs on hematologic and lymphocytic parameters in HIV-un-
infected infants and children during the firs 2 years of life.
SUBJECTS AND METHODS
WITS. The present analysis was conducted using retrospec-
tive data obtained from mother-infant pairs enrolled in the
WITS cohort during 19892004. WITS is a prospective obser-
vational, multicenter study that was established in 1989. The
study population is recruited from 6 sites in the United States
and Puerto Rico (Boston and Worcester, Massachusetts; Brook-
lyn and Manhattan, New York; Chicago, Illinois; Houston,
Texas; and San Juan, Puerto Rico). HIV-infected women were
recruited during pregnancy and monitored at regular intervals
throughout pregnancy, delivery, and the postpartum period.
Data obtained from mothers include the use of ARV drugs
before and during pregnancy, hard drug use, and maternal
medical complications. Their children are monitored at regular
intervals from birth, and clinical and laboratory parameters are
recorded at regular intervals. Flow-cytometric analysis was per-
formed and controlled for quality as described elsewhere [16,
17]. Informed consent was obtained from all women before
1090 JID 2006:194 (15 October) Pacheco et al.
their participation in the study. All research activities were ap-
proved by institutional review boards at each local site, in ac-
cordance with federal guidelines and regulations for the con-
duct of research involving human subjects.
Study population. Data from singleton pregnancies re-
sulting in HIV-exposed but uninfected infants were analyzed.
An infant was define as HIV-uninfected if 2 HIV virologic
tests were negative at or after age 1 month and at or after age
4 months or if 1 HIV serologic test was negative after age 18
months. Perinatal maternal and infant ARV use was categorized
as no ARV (no ARV use during pregnancy or by the infant
after birth), and any ARV (ARV use during pregnancy and/
or by the infant). The any-ARV group consisted of 2 large
groups: ARV monotherapy (1 antiretroviral drug used during
pregnancy and/or by the infant) and ARV combination therapy
(2 ARVs or highly active ARV therapy used during pregnancy
and/or by the infant). The ARV combination therapy group
was further divided into 2 subgroups: persons who received
protease inhibitors (PIs) and persons who did not. In general,
as in previous WITS analyses, mother-infant pairs in the no-
ARV group were from before 1994, those in the ARV mono-
therapy group were from 1994 to 1997, and those in the ARV
combination therapy group were from subsequent years [2].
Hard drug use during pregnancy was define as the use of
cocaine, heroin/opiates, methadone, and/or injection drugs as
ascertained by self-report and/or positive urine toxicologic re-
sults at prenatal or delivery visits [18]. Infants were evaluated
and blood was collected at age 07 days (usually within the
firs 48 h of life), at ages 610 days and 1 and 2 months, and
at 26-month intervals thereafter. For most analyses, laboratory
data were divided into 2 groups: data from blood collected
during the firs 2 months of life (07 days and 2 months) and
data from blood collected from age 624 months (6, 12, 18,
and 24 months). Infants who had laboratory values for at least
1 study visit were included in the analysis. The variables eval-
uated included hemoglobin concentration and platelet, neu-
trophil, total lymphocyte, and CD4+ and CD8+ cell counts.
Statistical analysis. Descriptive statistics were used to char-
acterize the study population. For continuous variables, means
and ranges were reported; for discrete variables, frequencies
and percentages were reported. Multivariate regression analysis
with robust variance (general estimating equations) was used
to assess the association between perinatal exposure to ARV
drugs and the define hematologic variables. The models were
adjusted for age, preterm birth (!37 weeks of gestation), ma-
ternal hard drug use [18], race/ethnicity, mode of delivery,
maternal Centers for Disease Control and Prevention (CDC)
clinical classification infants birth weight, infants sex, and
maternal CD4+ cell count at delivery. The associations were
examined during 2 periods of life: from birth until age 2 months
and from age 624 months. SAS software (version 8.2; SAS
Institute) was used to perform statistical analysis. Two-tailed P
values are reported.
RESULTS
Population characteristics. A total of 2171 HIV-infected
mothers (table 1) and their HIV-exposed, uninfected children
(table 2) were included in the analysis. Among participant
mother-infant pairs, 351 were classifie as having received no
ARV drugs, and 1820 were classifie as having received any
ARV drugs. The latter group was divided into 803 pairs for
whom exposure was to ARV monotherapy (91% zidovudine)
and 1017 pairs for whom exposure was to ARV combination
therapy. Within the any-ARV group, there were 18 mothers
who were not themselves treated but whose infants received 4
6 weeks of zidovudine alone (n p 14) or ARV combination
therapy (n p 4) after birth. The group that received ARV com-
bination therapy was further divided into ARV combination
therapy with (n p 676) and without (n p 341 ) PIs. Of women
who received ARV combination therapy that included PIs, 90
also received a nonnucleoside reverse-transcriptase inhibitor
(NNRTI); of women who received ARV combination therapy
that did not include PIs, 129 received an NNRTI. The racial
and ethnic distribution of HIV-infected women was similar in
the no-ARV and any-ARV groups. There were significan dif-
ferences between the no-ARV group and the any-ARV group
in the proportion of women using hard drugs during preg-
nancy, CDC clinical classification mean CD4+ cell count at
delivery, and proportion of women with CD4+ cell counts !200
cells/mm3 at delivery (4.6% vs. 14.2%, respectively).
As shown in table 2, the gestational age of ARV-exposed
children was slightly but significantl lower than that of ARV-
unexposed infants (38.1 vs. 38.5 weeks), possibly because of
more-frequent delivery via cesarean section after 1994. As men-
tioned above, all but a few of the ARV-unexposed infants were
born before 1994. Of the 1687 infants who received ARV pro-
phylaxis, 1536 (91.05%) received zidovudine alone.
Overall results. In univariate analyses, a number of vari-
ables were significantl associated with decreased hematologic
values, including African American race/ethnicity (for hemo-
globin concentration and neutrophil, lymphocyte, CD4+, and
CD8+ cell counts), maternal antenatal hard drug use (for he-
moglobin concentration and lymphocyte and CD4+ cell counts),
maternal CD4+ cell count !200 cells/mm3 (for hemoglobin con-
centration and platelet, lymphocyte, and CD4+ cell counts), pre-

Table 1. Women and Infants Transmission Study population: mothers.

No ARV drugs Any ARV drugs

Mothers (n p 351) (n p 1820) P
Age at delivery, mean (range), years 27.8 (14.545.1) 27.6 (15.144.3) .31
CD4+ cell count at delivery
Mean (range), cells/mm3 670.9 (502600) 511.4 (02709) .0005
a
!200 cells/mm3, no. (%) 13 (4.6) 210 (14.2) !.0001
b a
Race/ethnicity, no. (%) .34
White/non-Hispanic 44 (12.7) 190 (10.7)
Black/non-Hispanic 161 (46.5) 885 (49.6)
Hispanic 130 (37.6) 628 (35.2)
Other 11 (3.2) 80 (4.5)
NA 5 37
c a
Hard drug use, no. (%) 161 (45.7) 440 (24.3) !.0001
a
CDC classification during pregnancy, no. (%) .0003
Class B 73 (20.8) 497 (27.3)
Class C 10 (2.8) 120 (6.6)
Neither 268 (76.4) 1203 (66.1)
a
Mode of delivery, no. (%) !.0001
Scheduled cesarean 14 (4.2) 223 (15.6)
Nonscheduled cesarean 41 (12.2) 193 (13.6)
Vaginal birth 280 (83.6) 1000 (70.6)
NA 16 404

NOTE. Missing data were not included for calculation of the percentages. ARV, antiretroviral; CDC, Centers
for Disease Control and Prevention; NA, not available.
a
x2 test; all other comparisons were by t test.
b
Hispanic includes blacks and whites. Other includes Asian/Pacific Islander, American Indian, Alaskan
Native, and others.
c
Cocaine, heroin/opiates, methadone, and/or injection drug use, as ascertained by self-report and/or positive
urine toxicology at prenatal or delivery visits.

Hematologic Values in ARV-Exposed Infants JID 2006:194 (15 October) 1091

 Table 2. Women and Infants Transmission Study population: infants.
Infants
Birth weight, mean (range), kg
Gestational age, mean (range), weeks
Preterm delivery (!37 weeks), no. (%)
Sex, no. (%)
Male
Female
ARV prophylaxis during the first 6 weeks life, no. (%)
Duration of follow-up, mean (range), months
Samples/child, median (range), no.
NOTE. ARV, antiretroviral.
a
x2 test; all other comparisons were by 2-tailed Students t test.
maturity (for hemoglobin concentration and platelet and neu-
trophil counts), birth weight (for hemoglobin concentration and
CD8+ cell count); mode of delivery (for hemoglobin concentra-
tion), maternal CDC clinical classificatio (for neutrophil, lym-
phocyte, and CD4+ cell counts), and sex of the infant (for platelet,
lymphocyte, CD4+, and CD8+ cell counts) (data not shown).
Further analyses controlled for these variables.
Table 3 summarizes the multivariate analysis of various he-
matologic parameters in infants exposed and not exposed to
ARV drugs, with separate comparisons of any, single, and a
combination of ARV drugs with no ARV drugs. Hemoglobin
concentration and platelet, neutrophil, total lymphocyte, and
CD4+ cell counts were significantl lower in the any-ARV group
than in the no-ARV group during the firs 2 months of life.
The CD8+ cell count was borderline significantl lower in in-
fants exposed to ARV drugs. By age 624 months, there were
no longer significan differences in hemoglobin concentration
and neutrophil count between ARV exposure groups, although
the neutrophil count in infants exposed to ARV drugs remained
lower than that in infants not exposed to ARV drugs (regression
coefficient 152 cells/mm3; P p .05 ). However, platelet, total
lymphocyte, CD4+, and CD8+ cell counts all remained signif-
icantly lower in infants exposed to ARV drugs than in infants
not exposed to ARV drugs through age 24 months.
In infants through age 2 months, exposure to ARV combi-
nation therapy was associated with greater differences in neu-
trophil, lymphocyte, CD4+, and CD8+ counts than exposure to
a single ARV drug. For example, the difference in neutrophil
count between the no-ARV and ARV monotherapy groups was
321 cells/mm3, whereas the difference between the no-ARV
and ARV combination therapy groups was 548 cells/mm3;
the corresponding values for lymphocytes were 418 and 568
cells/mm3; those for CD4+ cell counts were 136 and 183
cells/mm3; and those for CD8+ cell counts were 33 and 115
cells/mm3. However, in infants age 624 months, with the ex-
ception of CD8+ cell count, there was little difference between
1092 JID 2006:194 (15 October) Pacheco et al.
No ARV drugs Any ARV drugs
(n p 351) (n p 1820) P
3.1 (0.84.6) 3.1 (0.86.1) .71
38.5 (1743) 38.1 (2543) .02
65 (18.5) 269 (14.8) .08a
.37a
185 (52.7) 912 (50.1)
166 (47.3) 908 (49.9)
1687 (92.7)
21.2 (!127) 15.8 (!134) !.0001
5 (16) 4 (16) !.0001
decreases in infants in the ARV monotherapy groups, compared
with those in the ARV combination therapy group.
We also analyzed the impact of exposure to ARV combinations
with PIs, compared with ARV combinations without PIs. There
was remarkably little difference between these groups, with the
sole exception of an apparent minor platelet-sparing effect of PI
exposure that was evident both in infants 02 months old and
those 624 months old (P p .006 and .02, respectively; data not
shown).
The effect of maternal immune status (CD4+ cell count at
delivery !200, 200500, or 1500 cells/mm3) on infant hema-
tologic parameters was also examined. In the multivariate mod-
els, a maternal CD4+ cell count !200 cells/mm3 was significantl
associated with a lower infant CD4+ cell count in both age groups:
02 months (169 cells/mm3; P p .03) and 624 months (183
cells/mm3; P p .02).
Progression of hematologic parameters over time. Figure
1A1F shows the changes in hematologic parameters and lym-
phocyte subsets between birth and age 24 months. The figu e
illustrates several points. First, the differences in hemoglobin
concentration and neutrophil count, although statistically sig-
nificant were minimal. Second, the differences in platelet, total
lymphocyte, CD4+, and CD8+ cell counts between ARV-unex-
posed and -exposed infants were significan at most visits
throughout the period of study.
Number of infants with 1 event of clinically relevant he-
matologic abnormalities, by ARV exposure status. We eval-
uated the incidence of clinically significan laboratory abnor-
malities among the various ARV exposure groups. Clinically
significan abnormalities were define as grade 3 toxicities in
the Division of AIDS pediatric toxicity tables. These thresholds
included hemoglobin concentration !7 g/dL, platelet count
!50,000/mm3, neutrophil count !400 cells/mm3, and CD4+ cell
count !750 cells/mm3 in infants age !12 months or !500 cells/
mm3 in infants age 1224 months. Less than 3% of children
had 1 hematologic or lymphocytic value in this abnormal
Table 3. Hematologic variables in infants according to antiretroviral (ARV) drug use and age of infant, adjusted for each factor in
multivariate analysis.

Hemoglobin Platelet Neutrophil Lymphocyte CD4+ CD8+

concentration, count, 103 count, count, cell count, cell count,
g/dL cells/mm3 cells/mm3 cells/mm3 cells/mm3 cells/mm3
Factor CE P CE P CE P CE P CE P CE P
Birth2 months
Intercept 14.9 288.1 7057.6 4938.1 2438.1 1160.2

Any ARV vs. no ARV 0.4 !.0001 21.7 .005 434.3 .01 492.8 !.0001 159.1 .03 73.7 .06
Monotherapy vs. no ARV 0.4 !.0001 18.4 .02 320.6 .07 418.1 .001 135.5 .08 32.7 .4
Combination therapy vs. no ARV 0.4 !.0001 24.7 .004 548.1 .003 567.6 !.0001 182.7 .02 114.7 .007
624 months
Intercept 11.8 463.9 3304.5 8307.9 3697.0 1599.5

Any ARV vs. no ARV 0.03 .6 25.7 !.0001 153.3 .05 530.8 !.0001 224.3 .0002 161.8 !.0001
Monotherapy vs. no ARV 0.07 .2 24.4 !.0001 152.3 .06 490.3 !.0001 204.2 .0008 103.4 .002
Combination therapy vs. no ARV 0.1 .04 27.0 !.0001 154.3 .08 570.2 !.0001 244.4 .0005 220.1 !.0001

NOTE. Use of ARV drugs is as defined in Subjects and Methods. Multivariate analysis was adjusted for maternal antenatal use of hard drugs, maternal CD4+
cell count at delivery (!200 or 200 cells /mm3), infant gestational age (!37 weeks), infant birth weight, race/ethnicity, mode of delivery, maternal Centers for
Disease Control and Prevention clinical classification, and infant sex and age. CE, coefficient; combination therapy, 2 ARVs or highly active antiretroviral therapy
used during pregnancy; monotherapy, 1 ARV used during pregnancy.

range, and the proportion of infants with 1 occurrence of
these values did not differ significantl among the various ARV
exposure groups.
DISCUSSION
These data from WITS, a large cohort study in the United
States, demonstrate that several hematologic parameters are
lower in HIV-exposed, uninfected infants with in utero and/
or neonatal exposure to ARV drugs, although these differences
are small and appear to be clinically insignificant The hema-
tologic effects of exposure to ARV drugs observed during the
firs 2 months of life could be secondary to the ongoing ex-
posure to ARV drugs given to the infant during the firs 6
weeks of life to prevent the transmission of HIV. However,
although the difference in hemoglobin concentration resolved
and that in neutrophil count became insignifican after age 2
months, significan differences persisted for platelet, lympho-
cyte, CD4+, and CD8+ cell counts through age 24 months.
Although exposure to ARV drugs in WITS was not randomized
and there were changes over time in certain aspects of the study
population and in the use of ARV drugs, the finding persisted
even after we controlled for a number of parameters that have
been shown to be associated with hematologic variables in other
studies, such as maternal race/ethnicity, drug use, maternal
CD4+ cell count at delivery, and infant gestational age, birth
weight, sex, and age. Despite some differences in methodology
and considerable differences in race/ethnicity, similar finding
were reported from the French Perinatal Cohort Study Group
and the European Collaborative Study [11, 14, 15].
In the French Perinatal Cohort Study, HIV-uninfected in-
fants with perinatal exposure to ARV drugs had significantl
lower hemoglobin concentrations and platelet, neutrophil, total
lymphocyte, CD4+, and CD8+ cell counts than HIV-exposed
infants without perinatal exposure to ARV drugs [10]. Although
differences in hemoglobin concentrations resolved by age 2
months, the differences in the other hematologic parameters
persisted through age 18 months. Those researchers reported
that exposure to combination ARV drugs was associated with
larger decreases than exposure to ARV monotherapy up to age
15 months. In addition, there was a negative relationship be-
tween the duration of treatment and neutrophil and lympho-
cyte counts, and a lower maternal CD4+ cell count at delivery
was associated with lower infant lymphocyte and CD4+ cell
counts.
In the European Collaborative Study, perinatal ARV exposure
of HIV-exposed but uninfected children was associated with a
reduced neutrophil count through age 8 years, regardless of
race or sex; maternal immune status was only marginally as-
sociated with infant neutrophil counts in uninfected children
[14]. In a separate analysis, the total lymphocyte count was
significantl lower in uninfected infants with perinatal exposure
to ARV drugs than in those without exposure to ARV drugs;
the CD4+ cell count was lower only during the firs year of life,
but the CD8+ count was reduced through age 8 years [15].
However, there was no significan difference between infants
who were exposed to ARV drugs, compared with those who
were not, in the incidence of lymphopenia, define as CDC
immune stage 2. A lower maternal CD4+ cell count (!200
cells/mm3) was associated with lower CD4+ cell counts but not
with CD8+ cell counts in infants. Exposure to ARV combination

Hematologic Values in ARV-Exposed Infants JID 2006:194 (15 October) 1093

Figure 1. AF, Hematologic and lymphocyte parameters over time, shown by age and perinatal exposure to antiretroviral (ARV) drugs for infants
born to HIV-infected women. Infants are grouped as those with no exposure to ARV drugs (no ARV), those with exposure (any ARV), and those with
exposure to single agents (ARV monotherapy) or to multiple agents (ARV combination therapy). Means at each age are shown, with bars spanning
1 SE above or below the mean. A, Hemoglobin concentration; B, platelet count; C, neutrophil count; D, lymphocyte count; E, CD4+ cell count; F, CD8+
count.
therapy had a stronger effect on total lymphocyte and CD8+
cell counts than exposure to ARV monotherapy, but the type
of drug in the combination (e.g., PIs or NNRTIs) had no effect.
Similar to the 2 other reports, our analysis of the WITS
cohort found significantl lower platelet, total lymphocyte,
CD4+, and CD8+ cell counts in uninfected infants with exposure
to ARV drugs than in those without exposure, and these dif-
ferences persisted through age 24 months. Our data also dem-
onstrate that the maternal immune status is associated with
infant CD4+ cell counts. The mechanism by which maternal
immune status might affect infant lymphocyte parameters is
not clear. There have been suggestions that an immunodeficien
mother may have defective transplacental transfer of hemato-
poietic cytokines, resulting in long-term consequences for the
thymic output of CD4+ cells in the child [10, 15]. Additional
studies are needed to elucidate this effect further.
However, unlike the European studies, although exposure to
ARV combination was associated in WITS with a stronger effect
on these parameters during the firs 2 months of life, this
potentiating effect of combination, compared with exposure to
ARV monotherapy, persisted through age 24 months only for
CD8+ cell count and not for platelet, total lymphocyte, or CD4+
cell counts. Additionally, in contrast to the French and Euro-
pean cohorts, in WITS, the effect of exposure to ARV drugs on
neutrophil count did not remain significan after age 2 months.
A recent study from Amsterdam compared hematologic pa-
rameters in 92 uninfected, HIV-exposed children, all but a few
of whom had mothers who had received 3 ARV drugs during
pregnancy and all of whom received combination zidovudine/
lamivudine therapy for the firs 4 weeks of life, with those in
75 children matched for gestational age, sex, and ethnicity who
were born to HIV-uninfected women [19]. Like WITS and the
other studies, differences in hemoglobin concentration were
transient. Those researchers reported significantl lower neu-
trophil counts in the HIV- and ARV-exposed infants than in
infants born to uninfected women through age 8 months. How-
ever, similar to the WITS analyses, these differences did not
persist through age 20 months. By contrast, significan differ-
ences in platelet and absolute lymphocyte counts were not
found in this considerably smaller cohort.
In vitro studies have demonstrated that ARV drugs can sup-
press the erythroid and myeloid cell lineages in bone marrow
[2024]. Cells cultured in the presence of zidovudine show a
concentration-dependent inhibition of CD34+ progenitor pro-
liferation in both myeloid and erythroid lineages, a decrease in
mtDNA, and an increase in lactic acid [22]. In vitro studies of
primary peripheral-blood lymphocyte cultures from individuals
without HIV infection who are exposed to zalcitabine, didan-
osine, or stavudine have shown dose- and time-dependent
mtDNA depletion accompanied by decreased cell proliferation,
Hematologic Values in ARV-Exposed Infants JID 2006:194 (15 October) 1095
increased lactate production, and morphologic changes in mi-
tochondria [23, 24].
It is known that NRTIs cross the placenta and achieve cord-
blood levels 50%100% of those in maternal blood [3]. He-
matopoietic progenitor cells from the fetus and neonate may
be more sensitive to the myelotoxic effects of drugs. In a study
that assessed the effect of zidovudine exposure in vitro, fetal
erythroid progenitors were more severely inhibited than those
from women of childbearing age, as manifested by a diminution
in clone generation and in the number of normoblasts per ery-
throid clone and neutrophils per granulocyte clone [25]. This
effect was secondary to an action on CD34+ progenitor cells,
given that no effect was observed on the production of gran-
ulocyte colony-stimulating factor or erythropoietin. The precise
mechanism of zidovudines action on CD34+ cells is unclear.
Additionally, HIV infection of the mother may by itself affect
fetal hematologic precursor cells. Poirier et al. [26] measured
mtDNA in cord blood and, at ages 1 and 2 years, peripheral-
blood mononuclear cells in children born to HIV-uninfected
mothers, HIV-infected mothers who did not receive ARV drugs,
and HIV-infected mothers who had received zidovudine during
pregnancy. Infants born to HIV-infected women who did not
receive ARV drugs had significan reductions in mtDNA that
persisted up to age 2 years; mtDNA was even further depleted
in infants born to HIV-infected mothers who had received
zidovudine. Similarly, a study that compared HIV-exposed, un-
infected infants born before 1994 who did not have perinatal
exposure to ARV drugs with 72 infants born to HIV-uninfected
women reported that CD4+ cell counts were persistently lower
in HIV-exposed infants [27].
ARV drugs are routinely prescribed during the second tri-
mester, a time in the fetus when hepatic hematopoiesis is active,
spleen development is occurring, and active lymphopoiesis, thy-
mic education, and bone-marrow development are ongoing.
The administration of ARVs during the critical window of he-
matopoiesis and lymphopoiesis may affect the generation of
these precursors. The differences in hemoglobin concentrations
in our patient population were no longer found in children 6
24 months old, which suggests a reversible insult in a system
with ample recovery potential. Likewise, significan differences
in neutrophil counts did not persist, which possibly reflect the
fact that granulocytes are the last population to appear during
development, are not formed in large numbers until after birth,
and have brief life spans [28, 29].
In conclusion, studies in different populations from Europe
[11, 14, 15] and, now, the United States have indicated that
perinatal exposure to ARV drugs is associated with minor de-
creases in hematologic parameters, most of which persist for
at least the firs 2 years of life in uninfected, HIV-exposed
infants. The studies vary in terms of which parameters remain
persistently low. In all studies, hemoglobin concentrations re-
turned to normal after ARV prophylaxis in infants was com-
pleted. However, in 2 of 3 European studies (but not WITS),
significantl lower neutrophil counts persisted in children ex-
posed to ARV drugs. Both of these European studies and WITS,
but not the Amsterdam cohort, found small but persistent dif-
ferences in lymphocyte parameters with exposure to ARV drugs,
with a suggestion that exposure to ARV combination therapy
may be associated with more-pronounced differences. Although
these effects do not appear to have clinical significance they
provide further evidence of the importance of the long-term
follow-up of uninfected, HIV- and ARV-exposed children, as
is recommended in the US Public Health Service Task Force
guidelines for treatment of pregnant women with ARV drugs
and the prevention of mother-to-child transmission [3]. Fur-
ther studies are needed to better defin the mechanism of these
effects and to evaluate whether they have clinical significanc
in the long term.
WITS MEMBERS
Principal investigators, study coordinators, program officers
and funding include Clemente Diaz and Edna Pacheco-Acosta
(University of Puerto Rico, San Juan; U01-AI-34858 [National
Institute of Allergy and Infectious Diseases]); Ruth Tuomala,
Ellen Cooper, and Donna Mesthene (Boston/Worcester Site,
Boston, Massachusetts; U01-DA-15054 [National Institute on
Drug Abuse]); Phil La Russa and Alice Higgins (Columbia
Presbyterian Hospital, New York, New York; U01-DA-15053
[National Institute on Drug Abuse]); Sheldon Landesman, Ed-
ward Handelsman, and Ava Dennie (State University of New
York, Brooklyn; U01-HD-36117 [National Institute of Child
Health and Human Development]); Kenneth Rich (Primary
Investigator) and Delmyra Turpin (Study Coordinator; Uni-
versity of Illinois at Chicago, Chicago; U01-AI-34841 and 1-
U01-AI-50274-01 [National Institute of Allergy and Infectious
Diseases]); William Shearer, Susan Pacheco, and Norma Cooper
(Baylor College of Medicine, Houston, Texas; U01-HD- 41983
[National Institute of Child Health and Human Develop-
ment]); Joana Rosario (National Institute of Allergy and In-
fectious Diseases, Bethesda, Maryland); Kevin Ryan (National
Institute of Child Health and Human Development, Bethesda,
Maryland); Vincent Smeriglio and Katherine Davenny (Na-
tional Institute on Drug Abuse, Bethesda, Maryland); and Bruce
Thompson (Clinical Trials & Surveys Corp., Baltimore, Mary-
land; N01-AI-85339 [National Institute of Allergy and Infec-
tious Diseases]).
Acknowledgments
We thank the children and their families who have participated in WITS
and the dedicated study personnel at all sites throughout the study, whose
efforts have made this analysis possible.
1096 JID 2006:194 (15 October) Pacheco et al.
References
1. Connor EM, Sperling RS, Gelber R, et al. Reduction of maternal-infant
transmission of human immunodeficienc virus type 1 with zidovudine
treatment. Pediatric AIDS Clinical Trials Group Protocol 076 Study
Group. N Engl J Med 1994; 331:117380.
2. Cooper ER, Charurat M, Mofenson L, et al. Combination antiretroviral
strategies for the treatment of pregnant HIV-1-infected women and
prevention of perinatal HIV-1 transmission. J Acquir Immune Defi
Syndr 2002; 29:48494.
3. Centers for Disease Control. Recommendations for use of antiretroviral
drugs in pregnant HIV-1-infected women for maternal health and
interventions to reduce perinatal HIV-1 transmission in the United
States. Available at: http://AIDSInfo.nih.gov. Accessed 25 November
2005.
4. Chotpitayasunondh T, Vanprapar N, Simonds RJ, et al. Safety of late
in utero exposure to zidovudine in infants born to human immuno-
deficienc virus-infected mothers: Bangkok. Bangkok Collaborative
Perinatal HIV Transmission Study Group. Pediatrics 2001; 107:E5.
5. Culnane M, Fowler M, Lee SS, et al. Lack of long-term effects of in
utero exposure to zidovudine among uninfected children born to HIV-
infected women. Pediatric AIDS Clinical Trials Group Protocol 219/
076 Teams. JAMA 1999; 281:1517.
6. Mofenson LM, Munderi P. Safety of antiretroviral prophylaxis of peri-
natal transmission for HIV-infected pregnant women and their infants.
J Acquir Immune Defi Syndr 2002; 30:20015.
7. Dagan T, Sable C, Bray J, Gerschenson M. Mitochondrial dysfunction
and antiretroviral nucleoside analog toxicities: what is the evidence?
Mitochondrion 2002; 1:397412.
8. Kamemoto LE, Shiramizu B, Gerschenson M. HIV-associated mito-
chondrial toxicity in pregnancy. Mitochondrion 2004; 4:15362.
9. Barret B, Tardieu M, Rustin P, et al. Persistent mitochondrial dys-
function in HIV-1-exposed but uninfected infants: clinical screening
in a large prospective cohort. AIDS 2003; 17:176985.
10. Landreau-Mascaro A, Barret B, Mayaux MJ, Tardieu M, Blanche S.
Risk of early febrile seizure with perinatal exposure to nucleoside an-
alogues. Lancet 2002; 359:5834.
11. Le Chenadec J, Mayaux MJ, Guihenneuc-Jouyaux C, Blanche S. Peri-
natal antiretroviral treatment and hematopoiesis in HIV-uninfected
infants. AIDS 2003; 17:205361.
12. The Perinatal Safety Review Working Group. Nucleoside exposure in
the children of HIV-infected women receiving antiretroviral drugs: ab-
sence of clear evidence for mitochondrial disease in children who died
before 5 years of age in fiv United States cohorts. J Acquir Immune
Defi Syndr 2000; 25:2618.
13. European Collaborative Study. Exposure to antiretroviral therapy in
utero or early life: the health of uninfected children born to HIV-
infected women. J Acquir Immune Defi Syndr 2003; 32:3807.
14. European Collaborative Study. Levels and patterns of neutrophil cell
counts over the firs 8 years of life in children of HIV-1-infected moth-
ers. AIDS 2004; 18:200917.
15. Bunders M, Thorne C, Newell ML. Maternal and infant factors and
lymphocyte, CD4 and CD8 cell counts in uninfected children of HIV-
1-infected mothers. AIDS 2005; 19:10719.
16. Calvelli T, Denny TN, Paxton H, Gelman R, Kagan J. Guideline for
flo cytometric immunophenotyping: a report from the National In-
stitute of Allergy and Infectious Diseases, Division of AIDS. Cytometry
1993; 14:70215.
17. Garcia PM, Kalish LA, Pitt J, et al. Maternal levels of plasma human
immunodeficienc virus type 1 RNA and the risk of perinatal trans-
mission. Women and Infants Transmission Study Group. N Engl J Med
1999; 341:394402.
18. Rodriguez EM, Mofenson LM, Chang BH, et al. Association of maternal
drug use during pregnancy with maternal HIV culture positivity and
perinatal HIV transmission. AIDS 1996; 10:27382.
19. Bunders MJ, Bekker V, Scherpbier HJ, Boer K, Godfried M, Kuijpers
 TW. Haematological parameters of HIV-1-uninfected infants born to
HIV-1-infected mothers. Acta Paediatr 2005; 94:15717.
20. Chitnis S, Mondal D, Agrawal KC. Zidovudine (AZT) treatment sup-
presses granulocyte-monocyte colony stimulating factor receptor type
alpha (GM-CSFR alpha) gene expression in murine bone marrow cells.
Life Sci 2002; 71:96778.
21. Dainiak N, Worthington M, Riordan MA, Kreczko S, Goldman L. 3
-
Azido-3
-deoxythymidine (AZT) inhibits proliferation in vitro of hu-
man haematopoietic progenitor cells. Br J Haematol 1988; 69:299304.
22. Lewis LD, Amin S, Civin CI, Lietman PS. Ex vivo zidovudine (AZT)
treatment of CD34+ bone marrow progenitors causes decreased steady
state mitochondrial DNA (mtDNA) and increased lactate production.
Hum Exp Toxicol 2004; 23:17385.
23. Setzer B, Schlesier M, Thomas AK, Walker UA. Mitochondrial toxicity
of nucleoside analogues in primary human lymphocytes. Antivir Ther
2005; 10:32734.
Hematologic Values in ARV-Exposed Infants JID 2006:194 (15 October) 1097
24. Setzer B, Schlesier M, Walker UA. Effects of didanosine-related de-
pletion of mtDNA in human T lymphocytes. J Infect Dis 2005; 191:
84855.
25. Shah MM, Li Y, Christensen RD. Effects of perinatal zidovudine on
hematopoiesis: a comparison of effects on progenitors from human
fetuses versus mothers. AIDS 1996; 10:123947.
26. Poirier MC, Divi RL, Al-Harthi L, et al. Long-term mitochondrial
toxicity in HIV-uninfected infants born to HIV-infected mothers. J
Acquir Immune Defi Syndr 2003; 33:17583.
27. Gesner M, Papaevangelou V, Kim M, et al. Alteration in the proportion
of CD4 T lymphocytes in a subgroup of human immunodeficienc
virus-exposed-uninfected children. Pediatrics 1994; 93:62430.
28. Holt PG, Jones CA. The development of the immune system during
pregnancy and early life. Allergy 2000; 55:68897.
29. West LJ. Definin critical windows in the development of the human
immune system. Hum Exp Toxicol 2002; 21:499505.