HIV/AIDS • CID 2007:44 (15 June) • 1647
H I V / A I D S M A J O R A R T I C L E
Time to Undetectable Viral Load after Highly Active
Antiretroviral Therapy Initiation among HIV-Infected
European Collaborative Studya
which highly active antiretroviral therapy (HAART) regimens are more effective for optimal viral response in
antiretroviral-naive, human immunodeficiency virus (HIV)–infected pregnant women.
Data on 240 HIV-1–infected women starting HAART during pregnancy who were enrolled in the
prospective European Collaborative Study from 1997 through 2004 were analyzed. An interval-censored survival
model was used to assess whether factors, including type of HAART regimen, race, region of birth, and baseline
immunological and virological status, were associated with the duration of time necessary to suppress viral load
below undetectable levels before delivery of a newborn.
Protease inhibitor–based HAART was initiated in 156 women (65%), 125 (80%) of whom received
nelfinavir, and a nevirapine-based regimen was initiated in the remaining 84 women (35%). Undetectable viral
loads were achieved by 73% of the women by the time of delivery. Relative hazards of time to achieving viral
suppression were 1.54 (95% confidence interval, 1.05–2.26) for nevirapine-based HAART versus PI-basedregimens
and 1.90 (95% confidence interval, 1.16–3.12) for western African versus non-African women.Themedianduration
of time from HAART initiation to achievement of an undetectable viral load was estimated to be 1.4 times greater
in women receiving PI-based HAART, compared with women receiving nevirapine-based HAART. Baseline HIV
RNA load was also a significant predictor of the rapidity of achieving viral suppression by delivery, but baseline
immune status was not.
In this study, nevirapine-based HAART (compared with PI [mainly nelfinavir]-based HAART),
western African origin, and lower baseline viral load were associated withshortertimetoachievingviralsuppression.
There have been no clinical trials in resource-rich regions that have addressed the question of
Plasma HIV RNA load is the preeminent risk factor for
mother-to-child transmission (MTCT) of HIV infec-
tion [1, 2]. In resource-rich regions, HAART (typically
composed of 3 antiretroviral agentsfrom2drugclasses)
has substantially reduced MTCT rates through suc-
cessful suppression of HIV RNA load [2, 3]. Although
an increasing proportion of HIV-infected pregnant
women in these regions are identifiedandtreatedbefore
pregnancy, a substantial minority receive a diagnosis
Received 13 December 2006; accepted 18 February 2007; electronically
published 2 May 2007.
aThis article was prepared by Deven Patel, Mario Cortina-Borja, Claire Thorne,
and Marie-Louise Newell (Centre for Paediatric Epidemiology and Biostatistics,
Institute of Child Health, University College London, London, United Kingdom).
Members of the study group are listed at the end of the text.
Reprints or correspondence: Deven Patel, Centre for Paediatric Epidemiology
and Biostatistics, Institute of Child Health, University College London, 30 Guilford
St., London WC1N 1EH, United Kingdom (email@example.com).
Clinical Infectious Diseases 2007;44:1647–56
? 2007 by the Infectious Diseases Society of America. All rights reserved.
antenatally and start antiretroviral therapy (ART) for
the first time during pregnancy to delay disease pro-
gression and/or to prevent MTCT . In manyWestern
European countries, thesewomenareincreasinglylikely
to have acquired HIV infection through heterosexual
contact and to be from countries with generalized ep-
idemics (mainly countries in sub-Saharan Africa) .
No clinical trials in resource-rich regions have ad-
dressed the question of which regimens are more ef-
fective for optimal viral response in ART-naive, HIV-
multicenter prospective cohort, this study was con-
ducted to determine whether choice of initial HAART
regimen for HIV-infected pregnantwomenisassociated
with the duration of time that is necessary to achieve
undetectable viral load by delivery.
The European Collaborative Study is an ongoing ob-
servational cohort study that was established in 1985,
1648 • CID 2007:44 (15 June) • HIV/AIDS
in which HIV-1–infected pregnant women are enrolled and
their infants are prospectively observed according to standard
protocols . Informed consent was obtained, and local ethics
committee approval was granted. Information collected in-
cluded timing and type of ART, maternal CD4 cell count, HIV
RNA load, and sociodemographic characteristics.
Laboratory tests were performed locally in laboratories lo-
cated in tertiary care centers and university hospitals. For HIV
RNA load quantification, Amplicor HIV-1 Monitor Test, ver-
sions 1.5 and ultrasensitive (Roche Diagnostic Systems);Quan-
tiplex HIV-1 RNA (b-DNA) assay, version 3.0 (Chiron Diag-
nostics); ornucleicacid sequence–based amplification/
nuclisens (Organon Teknika) were used. Classification of
undetectable viral load (viral suppression) was based on the
lower limit of quantification of the assay. Of the 759 antenatal
HIV RNA load measurements available, 561 (74%) were mea-
sured with ultrasensitive assays (quantification limit, ?50 cop-
ies/mL). HAART was defined as a regimen of ?3 antiretroviral
drugs, including a nucleoside reverse-transcriptase inhibitor
backbone and nevirapine (NVP)—a nonnucleoside reverse-
transcriptase inhibitor (NNRTI)—or a protease inhibitor (PI).
We restricted the analysis to HIV-infected women who were
ART-naive at conception; 153 women were first identified as
having HIV-infection during pregnancy, and 87 women were
known to have HIV infection before pregnancy (46 of these
women were known notto havepreviouslyreceivedprophylaxis
for prevention of MTCT, either because of no prior pregnancies
or documented nonreceipt of ART during previouspregnancies
that occurred while enrolled in European CollaborativeStudies,
and 41 were not receiving ART at conception and had no
documented prior ART use). Other eligibility criteria were a
detectable HIV RNA load measurement within 6 weeks before
initiation in the study and at least 1 subsequent measurement.
To determine the association of ma-
ternal factors and initial HAART regimen with viral suppres-
sion, we examined the duration of time from treatment initi-
ation to the achievement of undetectable HIV RNA load until
the time of delivery. For women not achieving an undetectable
HIV RNA load, the time was right-censored at the last mea-
surement at or before delivery. For those achieving an unde-
tectable HIV RNA load, the end point was known only to have
occurred between initiation and the first measurement (left-
censoring) or between any 2 measurements followinginitiation
A parametric survival model based on the Weibull distri-
bution, incorporating left-, right-, and interval-censoring, was
used . Estimates from Weibull models can be represented
either as relative hazards (RHs) or as the acceleration factor
between 2 levels of a covariate. Because most women reached
viral suppression, the RH reflects the rapidity with (rate at)
which viral suppression occurred, rather than solely the prob-
ability of the event occurring; larger RHs are associated with
more rapid attainment of viral suppression.
After adjusting for baseline viral load, race, type of HAART,
baseline CD4 cell count, maternal age, trimester at initiation
of HAART, timing of HIV diagnosis, history of injection drug
use, and year of delivery were considered in the Weibull model.
A stepwise modelselectionprocedurewasusedtochoosesalient
prognostic variables, with a variable retained if its inclusion
resulted in a significantly improved log likelihood. The pro-
pensity of being treated with NVP-based HAARTwasestimated
by a logistic regression model that included the covariatesmen-
tioned above; the propensity score was then included in all
adjusted models after stratification into quintiles .
Stratified survival curves and associated survivalprobabilities
were obtained with Turnbull’s generalization of the Kaplan
Meier estimate, allowing for interval-censored data [5, 7], with
95% CIs calculated using the adjusted bootstrap percentile
method with 1000 replications. Turnbull estimates of the pro-
portion of women achieving undetectable viral loads, by treat-
ment group and race, were calculated separately for women
with baseline viral loads ?4 log10copies/mL or !4 log10copies/
mL. Skewed continuous variables were compared using the
Mann-Whitney U test, and categorical variableswerecompared
using the x2test. Analyses were performed using Statasoftware,
version 9.1 (StataCorp), and R, version 2.2.0 (R Development
Of 1346 women receiving antenatal HAART who delivered a
newborn from 1997 through 2004, 240 pregnant, HIV-infected
women met our inclusion criteria. The characteristics of these
women are shown in table 1. Most women (59%) were black,
90% of whom were born in sub-Saharan Africa (table 1). PI-
based HAART was initiated in 156 women (65%), with the
remaining 84 women (35%) receiving an NNRTI-based regi-
men (all including NVP). Most regimens had a zidovudine and
lamivudine combination nucleoside reverse-transcriptase in-
hibitor backbone (table 2). One hundred twenty-five PI-based
regimens (80%) included nelfinavir (NFV), with the remaining
containing a lopinavir and ritonavir combination (4 regimens),
ritonavir (13 regimens), indinavir (8 regimens), or saquinavir
(6 regimens). The proportion of women receiving NVP-based
HAART increased from 16 (25%) of 64 women during 1997–
2000 to 37 (35%) of 105 women and 31 (44%) of 71 women
during 2001–2002 and 2003–2004, respectively (
There were no differences between black and non-blackwomen
with respect to type of HAART received (48 [34%] of 141 black
women vs. 34 [35%] of 96 non-black women received an NVP-
based HAART regimen;
P p .94
median HIV RNA load (4.13 log10copies/mL [interquartile
range (IQR), 3.54–4.60] vs. 4.20 log10copies/mL [IQR, 3.76–
P p .02
), the distribution of baseline
women at initiation of HAART and the number of women acheiving the end
point of an undetectable HIV RNA load by delivery.
Characteristics of 240 treatment-naive, HIV-infected pregnant
Region of birtha
Age at delivery
Median years (IQR)
History of IDU
Timing of diagnosis of HIV infection
Stage of pregnancy at initiation of HAART
Median weeks of gestation (IQR)
HIV RNA viral load, log10copies/mL
CD4 cell count
Median cells/mL (IQR)
aRegions of Africa were defined according to the United Nations groupings.
Data are no. (%) of patients, unless otherwise indicated. IDU, injection drug
1650 • CID 2007:44 (15 June) • HIV/AIDS
Table 2. Characteristics of study patients by HAART category.
(n p 156)
(n p 84)Pa
Received NRTI backbone
No. of patients receiving zidovudine and lamivudine (%)
No. of patients receiving another dual combination (%)
Median time of initiation of HAART, weeks of gestation (IQR)
Median baseline HIV RNA load, log10copies/mL (IQR)
Median baseline CD4 cell count, cells/mm3(IQR)
Median no. of viral load measurements (IQR)
Median interval between successive HIV RNA load tests, weeks (IQR)
Median duration of gestation at delivery, weeks (range)
aP values were calculated with the x2test or Mann-Whitney U test, as appropriate.
NRTI, nucleoside reverse-trascriptase inhibitor; NVP , nevirapine; PI, protease inhibitor.
mm3[IQR, 210–449] vs. 370 cells/mm3[IQR, 231–528]; P p
). However, black women tended to start treatment later.12
than non-black women (median of 24 weeks of gestation [IQR,
20–28] vs. median of 20.5 weeks of gestation [IQR, 15–27];
).P ? .01
The median number of virologicalmeasurementsperwoman
was 3 (range, 2–7 measurements), with a similar interval be-
tween successive tests for the 2 treatment groups (table 2).
Figure 1 shows the distribution of HIV RNA load measure-
ments at initiation, by weeks of gestation,togetherwithbaseline
median viral load and baseline median CD4 cell count. Thirty-
nine (24%) of 165 women starting HAART during the first or
second trimester had a CD4 cell count !200 cells/mm3, com-
pared with 9 (16%) of 55 women starting HAART during the
third trimester ().P p .35
Although time of HAART initiation and baseline HIV RNA
loads were similar between treatment groups, the PI-based
group had significantly lower baseline CD4 cell counts (table
2). Among the NVP group, 61 women (73%) had a CD4 cell
count 1250 cells/mm3at initiation; of these women, 57 (93%)
delivered a newborn before February 2004 (when NVP pre-
scribing information changed ).
The median gestational age at delivery was 38 weeks (range,
23–42 weeks), and 175 (73%) of 240 women had undetectable
viral loads by this time; table 1 shows the number of women
reaching this outcome, by maternal characteristics. The pro-
portion of women achieving an undetectable viral load did not
differ between treatment groups ( 111 [71%] of 156 women
receiving PI-based regimens and 64 [76%] of 84 women re-
ceiving NVP-based HAART regimens;
plays estimated proportions of women achieving undetectable
viral loads—beginning at the time of initiation of therapy—
stratified by HAART category and baseline viral load. For
women with a baseline HIV RNA load ?4 log10copies/mL,
), and baseline median CD4 cell count (313 cells/P p .18
). Figure 2 dis-P p .49
35.2% (95% CI, 20%–51.1%) of the PI group and 53.0% (95%
CI, 20%–94%) of the NVP group achieved an undetectable
HIV RNA load by 5 weeks, 56.4% (95% CI, 45%–71%) of the
PI group and 76.4% (95% CI, 56%–94%) of the NVP group
achieved an undetectable viral load by 10 weeks, and 59.4%
(95% CI, 43%–755) in the PI group and 93.4% (95% CI, 66%–
99.0%) of the NVP group achieved an undetectable viral load
by 15 weeks, indicating a differing response by treatment cat-
egory (figure 2B). Stratifying by race and baseline viral load,
58.7% (95% CI, 44%–75%) of non-black women and 66.3%
(95% CI, 53%–79%) of black women with baseline HIV RNA
loads ?4 log10copies/mL achieved an undetectable viral load
at 10 weeks, and 64.1% (95% CI, 38%–85%) of non-black
women and 79.3% (95% CI, 65%–93%) of black women
achieved an undetectable viral load at 15 weeks, suggesting
possible race-associated differences.
RHs from univariable survival analyses adjusting for baseline
viral load confirmed the race differences described above (RH
for black women vs. non-black women, 1.40; 95% CI, 1.00–
1.97) and were explored infurthermodelsthroughexamination
of region of birth. Table 3 shows the RHs for time from ini-
tiation of HAART to achievement of an undetectable viral load
for the full-adjusted model. Twenty-three women with infor-
mation missing on race or CD4 cell count were excluded from
the model; these women had characteristics that were similar
to those of women who were included (data not shown). In-
cluding 2-way interaction terms between any of the variables
in the final model did not significantly improve the fit. The
rate of women achieving an undetectable viral load in the NVP
group was estimated to be almost 1.5 times than that in the
PI group (table 3). Baseline viral load and being of Western
African origin were also significant factors affecting the rate of
achieving undetectable viral load by delivery. A sensitivityanal-
ysis including women receiving a PI-based regimen including
NFV only revealed similar RHs among women receiving NVP-
HIV/AIDS • CID 2007:44 (15 June) • 1651
of HAART during pregnancy. IQR, interquartile range.
Distribution of baseline HIV RNA load measurements, median baseline HIV RNA load, and median CD4 cell count, by timing of initiation
based HAART (RH, 1.56; 95% CI, 1.05–2.32), as did including
only women initiating therapy in the first and secondtrimesters
(RH, 1.82; 95% CI, 1.18–2.84).
Using the acceleration factor format, the median time to
achievement of an undetectable viral load for a woman re-
ceiving PI-containing HAART was estimated to be 1.38 times
(95% CI, 1.04–1.83 times) that for a woman receiving NVP-
containing HAART. The predicted median time to achievement
of an undetectable viral load for non-African womenwithbase-
line CD4 cell counts of 200–499 cells/mm3and HIV RNA viral
loads of 3.81–4.39 log10copies/mL who initiated treatmentdur-
ing the second trimester was 7.1 weeks (95% CI, 3.60–10.53
weeks) for the NVP group and 9.8 weeks (95% CI, 5.38–14.16
weeks) for the PI group; for Western African women with sim-
ilar characteristics, these times were 4.4 weeks (95% CI, 2.1–
6.7 weeks) and 6 weeks (95% CI, 3.2–8.9 weeks), respectively.
Viral response in 70 women eligible for NVP, according to
current prescribing advice  (i.e., with baseline CD4 cell
counts !250 cells/mm3) was explored; the median baselineviral
load was 4.35 log10copies/mL (IQR, 4.05–4.69), the median
duration of gestation at initiation of HAART was 22 weeks,
and 16 women (24%) received NVP. The percentage of women
reaching an undetectable viral load at 5 weeks was 34.8% (95%
CI, 20.4%–46.7%) for those receiving PI-based regimens and
52.9% (95% CI, 22.1%–71.6%) for those receiving NVP reg-
imens; the percentages increased to 50.4% (95% CI, 34.0%–
62.8%) and 82.4% (95% CI, 50.7%–93.7%), respectively, at 8.5
weeks. In adjusted analyses, these treatment group differences
were not statistically significant (RH for NVP-containing
HAART, 1.89; 95% CI, 0.87–4.12).
The 65 women (27%) who delivered with a detectable viral
load were similar to those achieving undetectable viral loads
with regard to regard to race and type and timing of HAART
(data not shown). However, more of thewomenwithdetectable
as a CD4 cell count !200 cells/mm3; 19 [33%] of 58 women
with a detectable viral load vs. 29 [18%] of 162 women with
an undetectable viral load;P p .03
had baseline viral loads 15 log10copies/mL (10 [15%] of 65
women with a detectable viral load vs. 10 [6%] of 175 with an
undetectable viral load;P p .03
livery for women with a detectable viral load was 2.48 log10
had a viral load 11000 copies/mL.
), and more of these women
); the median viral load at de-
1652 • CID 2007:44 (15 June) • HIV/AIDS
women with a baseline viral load !4 log10copies/mL, and (B), women with a baseline viral load ?4 log10copies/mL. NVP, nevirapine; PI, protease
Survival curves for the time from initiation of HAART to achievement of an undetectable viral load, by initial treatment category: (A),
Suppressing plasma HIV RNA load below detectable limits is
one of the goals for effective treatment of HIV-infected women
during pregnancy and for prevention of MTCT [9, 10]. In our
study, most pregnant women (73%) initiating HAART anten-
atally delivered with an undetectable viral load, and the re-
maining women delivered with a detectable but generally very
low viral load. Less than one-quarter of the women had im-
munological indications for treatment [9, 10]. Most women
were prescribed PI-containing HAART, with a highly homog-
enous approach, with 76% of these women receiving a com-
bination of NFV, zidovudine, and lamiduvine. NFV has the
most extensive data on pharmacokinetics and safety during
pregnancy among all of the PIs and is currently preferred for
use in antenatal HAART, especially in patients in whom there
are no maternal indications for treatment . One-third of
women received NVP-containing HAART, with increasing use
over time; this trend is unlikely to continue because of updated
NVP prescribing advice . The predominance of zidovudine-
and lamiduvine-containing regimens in our study reflects cur-
rent recommendations for this nucleoside reverse-transcriptase
inhibitor combination to be the backbone treatment for preg-
nant women [10, 11] and is consistent withprescribingpatterns
in Europe for nonpregnant individuals .
To our knowledge, this study is the first to suggestthatchoice
of initial HAART regimen has implications for timely achieve-
ment of undetectable viral load during pregnancy. Adjusting
for baseline prognostic factors, the hazard of achieving an un-
detectable viral load was greater for women receiving NVP-
containing HAART, with women in the PI group requiring an
average of 1.4 times longer to achieve viral suppression. Find-
ings conflict with regard to the relative effectiveness of PI-
containing versus NNRTI-containingHAARTregimensinnon-
pregnant adults.A recent
“head-to-head” randomized trials suggested that NNRTI-based
HAART (predominated by efavirenz) was 60% more effective
for virological suppression than was PI-based HAART (50%
boosted PIs), although no difference in clinical outcomes was
reported . However, an indirect meta-analysis yielded con-
tradictory results (i.e., NNRTI-based HAART was less effective
than PI-based HAART for virological suppression ); these
discordant results may be a result of differences in population,
study design, or type of nucleoside reverse-transcriptaseinhib-
itor backbone, highlighting the difficulties of translating trial
findings into clinical recommendations and the importance of
direct comparisons .
Therapeutic decision-making during pregnancy is compli-
cated by unique factors, including the need to consider pre-
vention of MTCT, safety and toxicity, and physiological
changes, which may affect pharmacokinetics [8, 14]. Accu-
mulating data on NFV pharmacokinetics suggest that drug lev-
els during the third trimester may frequently be subtherapeutic
[15–17]. This may explain why we found a superior virologic
response with NVP-containing versus PI-containing (mostly
HIV/AIDS • CID 2007:44 (15 June) • 1653
during pregnancy among 217 study women.
Factors associated with time to achieving undetectable HIV RNA load after initiation of HAART
Univariate analysis Multivariable analysis
RH (95% CI)P RH (95% CI)P
Region of birth
Northern or southern Africa
Baseline HIV RNA load, log10copies/mLa
Baseline CD4 cell count, cells/mm3
for all covariates listed in the table, with the addition of the treatment propensity score and trimester of initiation during
pregnancy. NVP , nevirapine; PI, protease inhibitor; RH, relative hazard.
aBaseline HIV RNA load was categorized according to their tertiles.
Univariate and multivariable estimates were adjusted for baseline viral load. Multivariable estimates were adjusted
NFV) HAART—in contrast to the Combine Study, in which
an equivalent response was reported among ART-naive non-
pregnant individuals randomized to a zidovudine and lami-
duvine backbone with NVP or NFV .
African and non-African pregnant women had similar base-
line immune and virological status—in contrast to previous
findings based on the whole cohort —probably reflecting
eligibility criteria for this analysis. The median baseline CD4
cell count among black women in our study was marginally
lower than those reported in African prevention of MTCT trials
(335–363 cells/mm3among ART-naive pregnant women) [19,
20]. Univariably, black women in our study responded to
HAART more favorably than did non-black women; further
investigation, stratifying by region of birth, revealed that this
effect was limited to women of western African origin. Limited
information is available regarding response to HAART among
African populations, and even less is available for pregnant
African women. In the Drug Resource Enhancement against
AIDS and Malnutrition pilot in Mozambique, 26 (65%) of 40
pregnant women starting HAART, with a median baseline HIV
RNA load of 4.2 log10copies/mL, achieved viral suppression
(viral load, !400 copies/mL) by delivery after an average of 12
weeks ; these data are consistentwith ourresults.Animpact
of race on disease progression or response to HAART has been
suggested by several studies of pregnant and nonpregnant in-
dividuals, which generally revealed poorer virologicalresponses
among black and/or African groups; these findings were sug-
gested to be a result of coinfections or adherence [22–24]. We
did not have adherence data available, but it seems unlikely
that differing adherence levels could explain our findings, be-
cause the better virological response to HAART was limited to
the western African group only. Little is known about the im-
pact of different HIV subtypes on the effectiveness of HAART
[25, 26]. Differences in underlying maternal subtype may pos-
sibly explain our findings, although host biological and genetic
differences may also play a part .
We found no significant difference in time to attaining un-
detectable viral load between severely immunosuppressed
women and those with greater immunocompetence. This is
consistent with other studies that have found a significant as-
sociation between baseline viral load and subsequentvirological
response after HAART initiation butnotanassociationbetween
baseline CD4 cell count and virological response [28, 29]. Few
studies have examined the latter association among ART-naive
between baseline CD4 cell count and viral load may differ from
that in nonpregnant adults [30, 31].
Our data are limited by their observational nature;how-
ever, we allowed for interval-censoring , adjusted for timing
of initiation of therapy during pregnancy, and minimized con-
founding by ART experience through our selection criteria.
Additionally, we used a treatment propensity score to reduce
1654 • CID 2007:44 (15 June) • HIV/AIDS
bias in the comparison of a treatment group to a nonrandom-
ized control group . We could not account for additional
factors potentially influencing response to HAART, such as ad-
herence, biological differences in drug activity arising from var-
iations in body weight and pharmacokinetics between groups,
HIV subtypes, and other genetic factors [14, 33, 34]. A dis-
advantage of cohort data is their limited contemporary rele-
vance when therapeutic practices have changed over time. A
case-in-point is the NVP prescribing changes after the asso-
ciation of the drug with hepatotoxicity in women with mod-
erate to high CD4 cell counts . If clinicians comply with
prescribing advice, one would expect the future group of ART-
naive women starting NVP-containing HAART to have lower
CD4 cell counts than the women in our study. However, our
subanalysis of women with CD4 cell counts !250 cells/mm3
indicates that our results may be generalizable but had limited
Although guidelines state thatpregnancyshouldnotpreclude
use of optimal ART regimens , in reality, there are limited
options. For nonpregnant adults, NNRTI-containing HAART
is recommended as a first-line regimen, preserving PI-contain-
ing HAART for later treatment, with efavirenz as the preferred
agent. Because efavirenz is contraindicated during the first tri-
mester of pregnancy, NVP-containing HAART has been in-
creasingly used for ART-naive pregnant women in Europe, but
this is no longer recommended for women with relativelygood
immune functioning. The potential option of initiating efavi-
renz-based HAART during the second or third trimester, if
contraception can be assured after delivery, has been suggested
in current World Health Organization recommendations for
resource-limited countries , but whether this approach will
be used in Europe is uncertain.
To date, NFV has been the overwhelming choice to accom-
pany zidovudine and lamivudine in the treatmentofART-naive
women in our study, but was less effective with regard to vi-
rological suppression than NVP-containing HAART. Boosted
PI regimens appear to offer superior virological suppression in
ART-naive adults, compared with PI alone , and lopinavir
and ritonavir combination therapy is identified as a preferred
PI regimen for initial HAART during pregnancy in current US
guidelines—albeit, with limited pharmacokinetic and safety
data [11, 37]. As more information becomes available, includ-
ing boosted PIs in initial HAART regimens during pregnancy
may become increasingly common, and the question of the
equivalence of such regimens to other HAART regimens war-
rants investigation (in addition to research of new agents, such
as integrase inhibitors). In the absence of clinical trials of
HAART among pregnant women, our findings add to the evi-
dence base to assist therapeuticdecision-makingforART-naive,
HIV-infected pregnant women. Our results strongly suggest
that an ART-naive pregnant woman with a CD4 cell count
HAART rather than NFV-containing HAART. In addition, our
results highlight the urgent need for further research of the
should begin receiving NVP-containing
EUROPEAN COLLABORATIVE STUDY
Dr. C. Giaquinto, Dr. O. Rampon, Dr. R. D’Elia, and Prof. A.
De Rossi (Universita degli Studi di Padova, Padua, Italy); Prof.
I. Grosch-Wo ¨rner (Charite Virchow-Klinikum, Berlin, Ger-
many); Dr. J. Mok (Royal Hospital for Sick Children, Edin-
burgh, Scotland); Dr. I. de Jose ´, Dra. Laaru ´, Dr. I. Bates, Dr.
A. Salas, Dr. J. Ma Pen ˜a, Dr. J. Gonzalez Garcia, Dr. J. R.Arribas
Lopez, and Dr. M. C. Garcia-Rodriguez (Hospital Infantil La
Paz, Madrid, Spain); Prof. F. Asensi-Botet, Dr. M. C. Otero,
and Dr. D. Pe ´rez-Tamarit (Hospital La Fe, Valencia, Spain); Dr.
H. J. Scherpbier, M. Kreyenbroek, Dr. M. H. Godfried, Dr. F.
J. B. Nellen, and Dr. K. Boer (Academisch Medisch Centrum,
Amsterdam, The Netherlands); Dr. A. B. Bohlin, Dr. S. Lind-
gren, Dr. B. Anze ´n, Dr. K. Lidman, Dr. K. Elfgren, Dr. K.
Gyllensten, and Dr. P. O. Pehrson (Karolinska University Hos-
pital, Huddinge and Solna, Sweden); Prof. J. Levy,Dr.P.Barlow,
Dr. Y. Manigart, Dr. M. Hainaut, Dr. A. Peltier, and Dr. T.
Goetghebuer (Hospital St. Pierre, Brussels, Belgium); Dr. A.
Ferrazin (Infectious Diseases Clinic, University of Genoa,
Genoa, Italy); Prof. A. De Maria (Department of Internal Med-
icine, University of Genoa, Genoa, Italy) Prof. G. Bentivoglio,
Dr. S. Ferrero, and Dr. C. Gotta (Department of Obstetrics and
Gynecology-Neonatology Unit, University of Genoa, Genoa,
Italy); Prof. A. Mu ˆr, Dr. A. Paya `, Dr. M. A. Lo ´pez-Vilchez, and
Dr. R. Carreras (Hospital del Mar, UniversidadAutonoma,Bar-
celona, Spain); Dr. N. H. Valerius and Dr. V. Rosenfeldt (Hvi-
dovre Hospital, Hvidovre, Denmark); Dr. J. Jimenez (Hospital
12 De Octubre, Madrid, Spain); Dr. O. Coll, Dr. A. Suy, and
Dr. J. M. Perez ( Hospital Clı ´nic, Barcelona, Spain); Dr. C.
Fortuny and Dr. J. Bogun ˜a (Hospital Sant Joan de Deu, Bar-
celona, Spain); Dr. M. Casellas Caro (Hospital Vall D’Hebron,
Barcelona, Spain); Dr. Y. Canet (HospitalParcTaulideSabadell,
Barcelona, Spain); Prof. G. Pardi and Dr. M. Ravizza (Ospedale
San Paolo, Milan, Italy); Dr. B. Guerra, Dr. M. Lanari, Dr. S.
Bianchi, and Dr. L. Bovicelli (Policlinico S Orsola, Bologna,
Italy); Dr. E. Prati and Prof. M. Duse (Universita di Brescia,
Brescia, Italy); Dr. G. Scaravelliand Dr.M.Stegagno(Universita
La Sapienza, Rome, Italy); Dr. M. De Santis (Universita Cat-
tolica, Rome, Italy); Dr. V. Savasi, Dr. S. Fiore, Dr. M. Crivelli,
and Prof. E. Ferrazzi (Ospedale L. Sacco, Milan, Italy); Dr. A.
Vigano `, Dr. V. Giacomet, Dr. D. Frasca, and Prof. G. Zuccotti
(Department of Pediatrics, L. Sacco Hospital, University of
Milan, Milan, Italy); Dr. F. Ravagni Probizer and Prof. A. Mac-
cabruni (Policlinico S. Matteo, Pavia, Italy); Dr. A. Bucceri and
Dr. L. Rancilio (Clinica Mangiagalli and Clinica De Marchi,
HIV/AIDS • CID 2007:44 (15 June) • 1655
Milan, Italy); Dr. S. Alberico, Dr. M. Rabusin, and M. Ber-
nardon (IRCCS Burlo Garofolo, Trieste, Italy); Dr. G. P. Taylor
and Dr. E. G. H. Lyall (St. Mary’s Hospital, London,United
Kingdom); Z. Penn (Chelsea and Westminster Hospital, Lon-
don, United Kingdom); Drssa. W. Buffolano and Dr. R. Tiseo
(Pediatric Dept, Federico II University, Naples, Italy), Prof. P.
Martinelli, Drssa. M. Sansone, Dr. G. Maruotti, and Dr. A.
Agangi (Obstetric Dept, Federico II University, Naples, Italy);
Dr. C. Tibaldi, Dr. S. Marini, Dr. G. Masuelli, and Prof. C.
Benedetto (University di Torino, Turin, Italy); Dr. T. Niemiec ¸
(National Research Institute of Mother & Child, Warsaw, Po-
land), Prof. M. Marczynska, Dr. S. Dobosz, Dr. J. Popielska,
and Dr. A. Oldakowska (Medical University of Warsaw, Infec-
tious Diseases Hospital, Warsaw, Poland); Dr. R. Malyuta, Dr.
I. Semenenko, and T. Pilipenko (ECS Ukraine coordinating
centre, Odessa, Ukraine); Dr. S. Posokhova and Dr. T. Kaleeva
(Odessa, Ukraine); and Dr. A. Stelmah and Dr. G. Kiseleva
We thank Kirsty England at the European Collaborative Study coordi-
nating centre (London, United Kingdom); Prof. L. Chieco-Bianchi, Prof.
F. Zacchello, Dr. E. Ruga, Dr. A. M. Laverda, Dr. A. Mazza, and S. Oletto
(Padua, Italy); Dr. T. Schmitz, Dr. R. Weogel, Dr. Karen Seel, and Dr. S.
Casteleyn (Berlin, Germany); Dr. S. Burns, Dr. N. Hallam, Dr. P. L. Yap,
and Dr. J. Whitelaw (Edinburgh, Scotland); Dra. B. Sancho and Dr. G.
Fontan-Casanego (Madrid, Spain); Dr. A. Gonzalez Molina, Dr. M. Gob-
ernado, Dr. J. L. Lopez, and Dr. J. Cordoba (Valencia, Spain); A. van der
Plas and E. M. Lepoole (Amsterdam, The Netherlands); Dr. E. Belfrage,
Dr. L. Nave ´r, Dr. A. Kaldma, and Dr. A. C. Lindholm (Sweden); Dr. G.
Di Siena, G. Mantero, Prof. S. Trasino, and Dr. J. Nicoletti (Genoa, Italy);
Dr. A. Mur (Barcelona, Spain); Dr. B. Martinez de Tejada, Dr. L. Zamora,
and Dr. R. Vidal (Barcelona, Spain); Dr. G. Zucotti (Milan, Italy); Dr. M.
Carla Re (Bologna, Italy); Prof. P. A. Tovo and Dr. C. Gabiano (Turin,
Italy); Dr. A. Maccabruni (Pavia, Italy); Dr. G. Ferraris(ClinicaMangiagalli,
Milan, Italy); Dr. T. Bruno (Naples, Italy); The Regional Health Office and
RePuNaRC (Naples, Italy); G. Mantero, Dr. A. Nicoletti, Dr. B. Bruzzone,
Dr. R. Rosso, and Dr. M. Setti (Genoa, Italy); and M. Kaflik (Medical
University of Warsaw, Warsaw, Poland). We also thank the anonymous
referees for their helpful comments.
The European Collaborative Study (ECS) is a co-
ordination action of the European Commission (PENTA/ECS 018865),and
the coordinating center receives support from the UK MRC Sexual Health
and HIV Strategy Committee.
Potential conflicts of interest.
All authors: no conflicts.
1. Mofenson LM, Lambert JS, Stiehm ER, et al. Risk factors for perinatal
transmission of human immunodeficiency virus type 1 in women
treated with zidovudine. Pediatric AIDS Clinical Trials Group Study
185 Team. N Engl J Med 1999;341:385–93.
2. European Collaborative Study. Mother-to-child transmission of HIV
infection in the era of highly active antiretroviral therapy. Clin Infect
3. Cooper ER, Charurat M, Mofenson L, et al. Combinationantiretroviral
strategies for the treatment of pregnant HIV-1–infected women and
prevention of perinatal HIV-1 transmission. J Acquir Immune Defic
4. European Collaborative Study. The mother-to-child HIV transmission
epidemic in Europe: evolving in the East and established in the West.
5. Lindsey JC, Ryan LM. Tutorial in biostatistics methods for interval-
censored data. Stat Med 1998;17:219–38.
6. D’Agostino RB Jr. Propensity score methods for bias reduction in the
comparison of a treatment to a non-randomized control group. Stat
7. Giolo SR. Turnbull’s nonparametric estimator for interval-censored
data. Technical report 2004. Available at: http://www.est.ufpr.br/rt/
suely04a.pdf. Accessed 9 October 2006.
8. Public Health Service Taskforce. Safety and toxicity of individual an-
tiretroviral agents in pregnancy. Available at: http://aidsinfo.nih.gov/.
Accessed 9 October 2006.
9. British HIV Association. Guidelines for the management of HIV in-
fection in pregnant women and the prevention of mother-to-child
transmission of HIV. http://www.bhiva.org/guidelines/2005/pregnancy
/. Accessed 9 October 2006.
10. Public Health Service Taskforce. Recommendations for use of anti-
and interventions to reduce perinatal HIV-1 transmissionintheUnited
States. Available at: http://aidsinfo.nih.gov/. Accessed 9 October 2006.
11. Department of Health and Human Services Panel on Antiretroviral
Guidelines for Adults and Adolescents. Guidelines for the use of an-
tiretroviral agents in HIV-1–infected adults and adolescents. Available
at: http://aidsinfo.nih.gov/. Accessed 9 October 2006.
12. Mocroft A, Phillips AN, Ledergerber B, et al. Relationship between
antiretrovirals used as part of a cART regimen and CD4 cell count
increases in patients with suppressed viremia. AIDS 2006;20:1141–50.
13. Chou R, Fu R, Huffman LH, Korthuis PT. Initial highly-active anti-
retroviral therapy with a protease inhibitor versus a non-nucleoside
reverse transcriptase inhibitor: discrepancies between direct and in-
direct meta-analyses. Lancet 2006;368:1503–15.
14. Mirochnick M, Capparelli E. Pharmacokinetics of antiretrovirals in
pregnant women. Clin Pharmacokinet 2004;43:1071–87.
15. Villani P, Floridia M, Pirillo MF, et al. Pharmacokinetics of nelfinavir
in HIV-1–infected pregnant and nonpregnant women. Br J Clin Phar-
16. Nellen JF, Schillevoort I, Wit FW, et al. Nelfinavir plasma concentra-
tions are low during pregnancy. Clin Infect Dis 2004;39:736–40.
17. van Heeswijk RP, Khaliq Y, Gallicano KD, et al. The pharmacokinetics
of nelfinavir and M8 during pregnancy and post partum. Clin Phar-
macol Ther 2004;76:588–97.
18. Podzamczer D, Ferrer E, Consiglio E, et al. A randomized clinical trial
comparing nelfinavir or nevirapine associated to zidovudine/lamivu-
dine in HIV-infected naive patients (the Combine Study). Antivir Ther
19. Taha TE, Brown ER, Hoffman IF, et al. A phase III clinical trial of
antibiotics to reduce chorioamnionitis-related perinatal HIV-1 trans-
mission. AIDS 2006;20:1313–21.
20. Shapiro RL, Thior I, Gilbert PB, et al. Maternal single-dose nevirapine
versus placebo as part of an antiretroviral strategy to prevent mother-
to-child HIV transmission in Botswana. AIDS 2006;20:1281–8.
21. Giuliano M, Guidotti G, AndreottiM,et al.Tripleantiretroviraltherapy
administered during pregnancy and after delivery significantly reduces
breast milk viral load [abstract 727]. In: Program and abstracts of the
13th Conference on Retroviruses and Opportunistic Infections (Den-
22. Frater AJ, Dunn DT, Beardall AJ, et al. ComparativeresponseofAfrican
HIV-1–infected individuals tohighlyactiveantiretroviraltherapy.AIDS
23. Cunningham CK, Balasubramanian R, Delke I, et al. The impact of
race/ethnicity on mother-to-child HIV transmission in the United
States in Pediatric AIDS Clinical Trials Group Protocol 316. J Acquir
Immune Defic Syndr 2004;36:800–7.
24. Anastos K, Gange SJ, Lau B, et al. Association of race and gender with
HIV-1 RNA levels and immunologic progression. J Acquir Immune
Defic Syndr 2000;24:218–26.
1656 • CID 2007:44 (15 June) • HIV/AIDS Download full-text
25. Bocket L, Cheret A, Deuffic-Burban S, et al. Impact of human im-
munodeficiency virus type 1 subtype on first-line antiretroviraltherapy
effectiveness. Antivir Ther 2005;10:247–54.
26. Geretti AM. HIV-1 subtypes: epidemiology and significance for HIV
management. Curr Opin Infect Dis 2006;19:1–7.
27. Brumme ZL, Harrigan PR. The impact of human genetic variation on
HIV disease in the era of HAART. AIDS Rev 2006;8:78–87.
28. Phillips AN, Staszewski S, Weber R, et al. HIV viral load response to
antiretroviral therapy according to the baseline CD4 cell count and
viral load. JAMA 2001;286:2560–7.
29. Matthews GV, Sabin CA, Mandalia S, et al. Virological suppression at
6 months is related to choice of initial regimen in antiretroviral-naive
patients: a cohort study. AIDS 2002;16:53–61.
30. Melvin AJ, Burchett SK, Watts DH, et al. Effect of pregnancy and
zidovudine therapy on viral load in HIV-1–infected women. J Acquir
Immune Defic Syndr Hum Retrovirol 1997;14:232–6.
31. Burns DN, Landesman S, Minkoff H, et al. The influence of pregnancy
on human immunodeficiency virus type 1 infection: antepartum and
postpartum changes in human immunodeficiency virus type 1 viral
load. Am J Obstet Gynecol 1998;178:355–9.
32. Phillips AN, Grabar S, Tassie JM, Costagliola D, Lundgren JD, Egger
M. Use of observational databases to evaluate the effectiveness of an-
tiretroviral therapy for HIV infection: comparison of cohort studies
with randomized trials. EuroSIDA, the French Hospital Database on
HIV and the Swiss HIV Cohort Study Groups. AIDS 1999;13:2075–82.
33. Stebbing J, Bower M. Comparative pharmacogenomicsofantiretroviral
and cytotoxic treatments. Lancet Oncol 2006;7:61–8.
34. Lowe SH, Prins JM, Lange JM. Antiretroviral therapy in previously
untreated adults infected with the human immunodeficiencyvirustype
I: established and potential determinants of virological outcome. Neth
J Med 2004;62:424–40.
35. World Health Organization. Anitretroviral therapy for HIV infection
in adults and adolescents in resource-limitedsettings:towardsuniversal
access. Geneva, Switzerland: World Health Organization, 2006.
36. Bartlett JA, Fath MJ, Demasi R, et al. An updated systematic overview
of triple combination therapy in antiretroviral-naive HIV-infected
adults. AIDS 2006;20:2051–64.
37. Stek AM, Mirochnick M, Capparelli E, et al. Reduced lopinavir ex-
posure during pregnancy. AIDS 2006;20:1931–9.