INTRODUCTION
Highly active antiretroviral therapy (HAART) during pregnancy and breastfeeding for the prevention of mother-to-child transmission of HIV is a highly efficacious public health intervention.1-5 6-8 9-12 growth of HIV -exposed uninfected infants over time by in utero antiretroviral (ARV) exposure, but this study grouped dual maternal therapy with triple therapy and did not evaluate HAART in isolation.13 14,15 growth implications of in utero HAART exposure on HIV -exposed uninfected infants represents an important public health issue.
We investigated growth patterns through 6 months of life by in utero ARV exposure among breast-fed infants enrolled in two clinical trials in Botswana.
METHODS
Sources of Data
The Mashi (meaning “milk” in Setswana) and Mma Bana (meaning “mother of the baby” in Setswana) PMTCT studies were conducted at the same four study sites in southern Botswana. Both studies enrolled HIV -1-infected pregnant women regardless of baseline CD4+ cell count. The Botswana Health Research Development Committee and the Harvard School of Public Health Human Subjects Committee approved both studies, and independent Data and Safety Monitoring Boards reviewed study safety and efficacy data approximately every 6 months. Participants in both studies provided written informed consent.
The Mashi Study, described previously,16,17 HIV -1-infected pregnant women between March 2001 and October 2003. Women received a short course of 300 mg zidovudine (ZDV) twice daily initiated at 34 weeks of gestation and continued through labor. Dosing frequency was increased to every 3 hours during labor. By randomized design, half of the women participating in the Mashi study also received a single dose of nevirapine (NVP) as did more than half of infants (all infants received NVP after a design modification 17 months into the study). The feeding intervention randomized half of the mother-infant pairs to 6 months of exclusive breastfeeding with 6 months of prophylactic ZDV for the infant versus formula feeding with 4 weeks of prophylactic ZDV to the infant. HAART was offered as part of the Botswana National ARV program midway through the study. A total of 71 women started HAART antenatally in the Mashi study. Infants born to Mashi study participants who took HAART during pregnancy were included in the HAART exposure group of this infant longitudinal growth analysis.
The Mma Bana Study enrolled a total of 730 HIV -1-infected pregnant women between July 2006 and May 2008 and has been described in detail previously.1 + cell counts 200 cells/mm3 or greater were randomized to receive either abacavir/ZDV/lamivudine coformulated as Trizivir (GlaxoSmithKline, Greenford, United Kingdom) twice daily or lopinavir/ritonavir with ZDV/lamivudine coformulated as Kaletra (Abbott Virology, Abbott Park, IL)/combivir (GlaxoSmithKline, Greenford, United Kingdom) twice daily. A total of 170 women with CD4+ cell counts less than 200 cells/mm3 or with an AIDS-defining illness were enrolled in an observational arm and received nevirapine/ZDV/lamivudine (NVP/combivir) twice daily (after 2 weeks of 200 mg once-daily NVP) in accordance with Botswana National PMTCT Guidelines. During labor, Mma Bana participants took ZDV every 3 hours regardless of their assigned treatment arm. Women initiated HAART between 18 and 34 weeks of gestation and continued through scheduled weaning by 6 months postpartum; HAART was continued for maternal health if indicated. Infants received single-dose NVP at birth and ZDV from birth through 4 weeks of age in keeping with the Botswana National PMTCT Guidelines.
Statistical Methods
We performed a retrospective analysis comparing infant growth data during the first 6 months of life from qualifying infants in the Mashi and Mma Bana studies. We restricted this analysis to singleton infants carried to term (37 weeks or greater gestational age) who were breast-fed. Gestational age was calculated from an algorithm using maternal reported last menstrual period and a second-trimester ultrasound. Infants born to mothers who took HAART for less than 2 weeks before delivery were excluded from the analysis. Any infant with a positive HIV -1 DNA polymerase chain reaction result within the first 7 months of life was excluded from the analysis.
During evaluations conducted monthly in the first 6 months of life for both studies, each infant was weighed and his or her length was measured. These data were used to calculate z-scores for an infant's weight-for-age (WAZ), length-for-age (LAZ), and weight-for-length (WLZ) using the 2006 World Health Organization's (WHO) Child Growth Standards.18 growth that correlate with increased infant mortality include wasting and stunting.19-21
Statistical analyses were performed using SAS, Version 9.2 (SAS Institute, Cary, NC). P values for statistical tests involving a continuous normally distributed variable were derived from a two-sided Student t test; otherwise, a Wilcoxon rank sum test was used. P values for comparisons of categorical variables were derived from Fisher exact or χ2 testing. A two-sided Student t test was used to compare mean weight, length, WAZ, LAZ, and WLZ at birth for the two exposure groups. Analysis of response profiles was used to evaluate the sequence of mean WAZ, LAZ, and WLZ by exposure group over time. Linear mixed effects models were fitted separately for WAZ, WLZ, and LAZ using a piecewise linear spline with a knot at 2 months. The models included random effects for intercept and both (early and late phase) slope, and an unstructured correlated covariance matrix was assumed for the random effects. Maternal enrollment CD4+ cell count, maternal body mass index at 1 month postpartum as a marker of maternal nutritional status, and site of enrollment were all controlled for in the linear mixed effects model, because these attributes were found to be significant predictors of infant growth (P < 0.05). In addition, use of the WHO's Child Growth Standards z-scores ensured that infant gender was also controlled for in the linear mixed-effects model. All testing used a significance level of 0.05 with two-sided hypothesis testing and no corrections for multiple testing.
RESULTS
Baseline Characteristics of Comparison Groups
The study schema for the Mashi and Mma Bana studies are shown in Figure 1 . There were 1877 live infants born to the 1930 HIV -infected pregnant women enrolled in the Mashi and Mma Bana studies (Fig. 2 ). Of live-born infants, 1856 were the product of a singleton birth. Two hundred twenty-five infants were excluded because delivery took place before 37 weeks gestational age (204 [10.9%]) or gestational age was lacking (23 [1.2%]). Another 517 (27.9%) Mashi study participants were excluded from the current growth analysis, because they were randomized to formula feeding. Additional exclusions included 47 infants with positive HIV -1 DNA polymerase chain reaction results within the first 7 months of life (42 [8.7%] in the ZDV exposure group and five [0.8%] in the HAART exposure group) and six (1.0%) infants born less than 14 days after maternal HAART was initiated. In total, 619 infants with in utero HAART exposure and 440 infants with in utero ZDV exposure were included in the study.
FIGURE 1: Mashi and Mma Bana mother-infant pair antiretroviral exposures for breast-fed infants. 1 Beginning in October 2002, highly active antiretroviral therapy (HAART) became available through the national program. Women enrolled in the Mashi study with CD4+ counts less than 200 cells/mm3 were offered HAART. 2 In August 2002, as a result of efficacy data from a trial in Thailand, the peripartum intervention was revised to eliminate infant placebo and provide open-label nevirapine to all infants to be given as a single dose within 72 hours of birth. 3 Infants in the ZDV exposure group remained on oral prophylactic ZDV up to a maximum of 6 months throughout the period of breastfeeding. sdNVP, single-dose nevirapine; ZDV, zidovudine; BF, breastfeeding; 3TC, lamivudine; rtv, ritonavir.
FIGURE 2: Study eligibility.
Baseline characteristics of mothers and infants included in this analysis are presented in Table 1 . Many maternal characteristics, including age, number of pregnancies, marital status, education level, and hemoglobin level at the time of enrollment, did not differ significantly between groups. Mothers taking HAART during pregnancy had more personal income, even after adjustment for inflation, compared with women who took ZDV during pregnancy and were more likely to have electricity at home. Over 50% of the women in each exposure group reported the current pregnancy to be either their first or second pregnancy.
TABLE 1: Maternal and Infant Characteristics
The maternal enrollment CD4+ counts for women in the ZDV exposure group were higher compared with the HAART group (median 392 cells/mm3 versus 331 cells/mm3 , P < 0.001); 24% of infants exposed to HAART in utero were born to mothers with enrollment CD4+ counts less than 200 cells/mm3 , whereas only 11% of infants exposed to ZDV in utero were born to mothers with enrollment CD4+ counts less than 200 cells/mm3 . Therefore, a stratified analysis of CD4+ cell count was undertaken based on maternal enrollment CD4+ cell count either less than 200 cells/mm3 or 200 cells/mm3 or greater (Table 2 ). Baseline plasma HIV -1 RNA viral load was higher for women in the ZDV exposure group compared with women in the HAART exposure group (median 4.34 log10 copies/mL versus 4.18 log10 copies/mL, P = 0.02).
TABLE 2: Infant Birth Measurements by Maternal Enrollment CD4+
There were no statistically significant differences in gender or distribution of gestational age of infants between the two groups nor was there a difference in the percent of infants born small for gestational age, defined as a weight of less than 2500 g for an infant at 37 weeks or greater gestational age. Median duration of in utero exposure to HAART (12.1 weeks; range, 2.6-22.3 weeks) was longer than median duration in utero exposure to ZDV (5.7 weeks; range, 2.0-10.9 weeks) as a result of study design differences relative to gestational age at enrollment between the Mma Bana and Mashi studies.
Weight and Length at Birth
Infants exposed in utero to HAART had a mean birth weight z-score of -0.64, whereas the ZDV-exposed group had a birth weight z-score of -0.34 (P < 0.001) (Table 2 ). These differences were similar for both male and female subgroups, but mean birth weight z-scores were lower for females. In a stratified analysis of normalized weight for age based on maternal enrollment. CD4+ cell count of either less than 200 cells/mm3 or 200 cells/mm3 or greater; mean birth weight z-score was lower if maternal CD4+ cell count was less than 200 cells/mm3 compared with 200 cells/mm3 or greater regardless of exposure group (Table 2 ).
Overall, the birth length z-score was also higher in the ZDV-exposed group. However, in stratified analysis based on maternal enrollment CD4+ count, this difference was only statistically significant for infants born to mothers with CD4+ counts 200 cells/mm3 or greater in which the ZDV group had a mean birth length z-score of 0.08 versus a mean birth length z-score of -0.15 (P = 0.018). This same trend was observed in birth WLZ.
Growth Over 6 MonthsWeight for Age
Controlling for enrollment site, maternal enrollment CD4+ as a binary variable (less than 200 cells/mm3 or 200 cells/mm3 or greater), maternal body mass index at 1 month postpartum, and infant gender, using a liner mixed-effects model, infants exposed to HAART in utero had a lower birth weight z-score but a more rapid increase in WAZ during the first 2 months of life than the ZDV-exposed infant group (P = 0.03). From 3 months through 6 months of life, the two groups experienced similar rates of weight gain for age (P = 0.26) (Fig. 3 ).
FIGURE 3: Response profile analysis modeling for mean infant z-scores by antiretroviral exposure status.
Length for Age
In adjusted analyses, the mean change in age-adjusted LAZ differed significantly between the two exposure groups from birth through the 2 months of life (P = 0.002) (Fig. 3 ). From the third through sixth months of life, the mean change in the LAZ for both the HAART- and ZDV-exposed infants was no longer statistically different (P = 0.08). When comparing absolute differences in mean recumbent length at 6 months of life between exposure groups by gender, mean recumbent length for HAART-exposed males was 0.4 cm shorter than ZDV-exposed males. The difference in mean recumbent length for females by exposure group at 6 months of life was 1.0 cm with HAART-exposed female infants on average having a lower recumbent length.
Weight for Length
In adjusted analyses using a linear mixed-effects model, the HAART-exposed group had a more rapid increase in WLZ during the first 2 months of life compared with the ZDV-exposed group (P < 0.0001) such that the mean for this group was higher than the mean score of the ZDV-exposed group at 2 months of life (Fig. 3 ). From the third through the sixth months of life, the mean WLZ z-score in the HAART-exposed group declined, reflecting a more rapid increase in LAZ than WAZ. Over the same period, the ZDV-exposed group had a modest increase in mean WLZ z-score. These distinctively different growth patterns by exposure group from the third through the sixth month of life were statistically different (P = 0.04).
Wasting and Stunting
There was no significant difference in the proportion of infants between exposure groups meeting criteria for wasting (2 or less standard deviations below norm for WLZ z-score) or stunting (2 or less standard deviations below norm for LAZ z-score) at 6 months of life. Wasting was present in 6.02% of the ZDV-exposed infants and 6.09% of the HAART-exposed infants (P = 0.96) at 6 months of life. Stunting was present in 4.76% of the ZDV-exposed group and 4.70% of the HAART-exposed infants (P = 0.96) at 6 months of life.
DISCUSSION
This is the first study to compare early growth of infants over time after in utero HAART or ZDV exposure. At birth, infants exposed to HAART in utero had significantly lower weight, length, and weight-for-length normalized scores compared with those exposed to ZDV in utero. However, weight differences were not apparent by 3 months of age and the incidence of wasting (defined as WLZ falling more than 2 standard deviations below norm) did not differ significantly between the two groups at 6 months of life.
It is reassuring that mean infant weights in the HAART-exposed group improved rapidly over the first 3 months of life with the two exposure groups experiencing an overlapping mean normalized weight pattern from 3 months through 6 months of life. In an exploratory analysis, the normalized longitudinal growth pattern was similar for infants in the HAART exposure group regardless of the maternal regimen. However, the lower mean birth weight for HAART-exposed HIV -uninfected infants represents a potential for higher rates of early infant mortality and/or morbidity. Survival rates for low-birth-weight infants remains a challenge in many resource-limited settings, including those where HAART is now becoming increasingly available for maternal treatment and for PMTCT. In the Mashi study as well as a cohort from the Zambia Exclusive Breastfeeding Study, low birth weight was associated with increased infant mortality.22,23 HIV -exposed infants observed that lower birth weight, but not transmission of HIV , was associated with higher mortality in the first 28 days of life; for infants who remained HIV -negative, lower birth weight continued to be associated with a higher risk of mortality after the first month of life and through the first year of life.24 growth in the first several months of life in an effort to mitigate morbidity and mortality.
Although HAART-exposed infants were observed to have shorter normalized LAZ throughout the first 6 months of life, their mean normalized WLZ was more age-appropriate by the second month of life than the ZDV-exposed cohort and there was no statistically significant difference between exposure groups in the proportion of infants with growth stunting according to WHO guidelines. Of note, we were unable to identify a biologically plausible explanation for shorter mean LAZ among HAART-exposed infants, and this presents an opportunity for further research. The clinical significance of lower normalized LAZ for HAART-exposed infants at 6 months of life is uncertain and should be interpreted with caution. However, as Mma Bana infants continue to be followed beyond 6 months of life, the pattern in normalized mean length can be re-evaluated and correlated with morbidity and/or mortality outcomes.
Maternal HAART provided a possible advantage to infants in terms of normalized WLZ in comparison to infants exposed to in utero ZDV and ongoing ZDV prophylaxis during breastfeeding. Overall, the cohort of infants exposed to HAART in utero were observed to have normalized WLZ that approached or exceeded norms from 2 months of life through 6 months of life, whereas the ZDV exposure group had lower normalized WLZ over this same period. In resource-limited settings where infants are at higher risk of infant morbidity and mortality from diarrheal disease and respiratory illnesses, infants with WLZ that are near normal may have a survival advantage. This hypothesis is consistent with reports of lower mortality among infants whose mothers received HAART in Malawi.25
Findings of this study should be considered in the context of protocol differences, which represent the major limitation of this study. First, Mashi ZDV-exposed infants received prophylactic ZDV throughout the period of breastfeeding up to 6 months of life, whereas Mma Bana infants only received 4 weeks of ZDV. Although the reported differences at birth by exposure group were not impacted by this issue, if ZDV prophylaxis during breastfeeding affected infant growth , it would represent a confounder with respect to longitudinal results. Second, mean duration of in utero ARV exposure differed significantly between the groups as a result of each study's enrollment protocol, and the potential impact of different exposure durations could not be evaluated. However, no differences were observed when we included or excluded HAART-exposed infants from the Mashi cohort, and we found no trend for lower z-scores among infants with longer HAART exposure in the Mma Bana cohort (data not shown). Third, it is possible that if lower-weight infants died during the period of the study at a higher rate than larger infants, then observed mean weight gain within either group could be a function of informative censoring by death. However, only 13 (2.1%) infants in the HAART exposure group and 12 (2.7%) infants in the ZDV exposure group died during the first 6 months of life. Therefore, this small number of deaths would be unlikely to influence the overall findings.
Finally, although the Mashi and Mma Bana studies were conducted at the same four sites in Botswana, the Mashi study took place before the Mma Bana study and temporal differences (including limited HAART availability and lower inflation-adjusted income in the Mashi study period) may have influenced study findings. Of note, there were no policy changes calling for nutritional, vitamin, or mineral supplementation between these studies and the fact that both cohorts of infants were exclusively breast-fed controlled for differences in the infants' nutritional status. We controlled for baseline CD4+ cell count differences in the linear mixed-effects model and by stratifying infant growth outcomes by maternal enrollment CD4+ cell counts as a binary variable, either less than 200 cells/mm3 or 200 cells/mm3 or greater (Table 2 ). Although maternal plasma HIV -1 RNA level also differed by exposure groups, it was neither a significant predictor nor an effect modifier of growth outcomes.
In summary, HIV -uninfected infants exposed in utero to HAART had lower birth weight than comparable infants exposed to short-course ZDV, but subsequent weight gain was rapid and approached the norm for age and gender by 3 months of life. HAART exposure was associated with lower mean infant length throughout the first 6 months of life. Although statistically significant, the lower mean length coupled with weight gain resulted in more age- and gender-appropriate normalized WLZ for HAART-exposed infants. Higher WAZ have been noted to have a survival benefit.21 HIV -uninfected infants.
ACKNOWLEDGMENTS
We are indebted to the women and infants who participated in the Mashi and Mma Bana studies, Mashi and Mma Bana study teams as well as the administration and staff at Scottish Livingstone, Deborah Retief Memorial, Athlone and Princess Marina Hospitals, and the staff at the referring health clinics.
REFERENCES
1. Shapiro R, Hughes M, Ogwu A, Kitch D, et al. Antiretroviral regimens in pregnancy and breast-feeding in Botswana.
N Engl J Med . 2010;362:2282-2294.
2. Marazzi MC, Nielsen-Saines K, Buonomo E, et al. Increased infant human immunodeficiency virus-type one free survival at one year of age in sub-Saharan Africa with maternal use of highly active antiretroviral therapy during breast-feeding.
Pediatr Infect Dis J . 2009;28:483-487.
3. Awino J, Zeh C, Bondo P, et al.
CD4 and Viral Load Response and Adherence Among ART-naive Women in a Trial of HAART for PMTCT in Kisumu, Kenya [Abstract 760]. 14th Conference on Retroviruses and Opportunistic Infections; 2007; Los Angeles, CA.
4. Kilewo C, Karlsson K, Ngarina M, et al.
Prevention of mother-to-child transmission of
HIV -1 through breastfeeding by treating mothers with triple antiretroviral therapy in Dar es Salaam, Tanzania: the Mitra Plus Study.
J Acquir Immune Defic Syndr . 2009;52:406-416.
5. Arendt V, Ndimubanzi P, Vyankandondera J, et al.
AMATA Study: Effectiveness of Antiretroviral Therapy in breastfeeding mothers to Prevent Post-natal Vertical Transmission in Rwanda [TuAX102]. 4th IAS Conference on
HIV Pathogenesis, Treatment and Prevention; 2007; Sydney, Australia.
6. Briand N, Mandelbrot L, Le Chenadec Jerome, et al. No relationship between in-utero exposure to HAART and intrauterine
growth retardation.
AIDS . 2009;23:1235-1243.
7. Townsend CL, Cortina-Borja M, Peckham CS, et al. Antiretroviral therapy and premature delivery in diagnosed
HIV -infected women in the United Kingdom and Ireland.
AIDS . 2007;21:1019-1026.
8. Ekouevi DK, Coffee P, Becquet R, et al. Antiretroviral therapy in pregnant women with advanced
HIV disease and pregnancy outcomes in Abidjan, Cote d'Ivoire.
AIDS . 2008;22:1815-1820.
9. 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 Defic Syndr . 2003;32:380-387.
10. Tuomala RE, Shapiro DE, Mofenson LM, et al. Antiretroviral therapy during pregnancy and the risk of an adverse outcome.
N Engl J Med . 2002;346:1863-1870.
11. Szyld E, Warley E, Freimanis L, et al. Maternal antiretroviral drugs during pregnancy and infant low birth weight and preterm birth.
AIDS . 2006;20:2345-2353.
12. Cotter AM, Garcia AG, Duthely ML, et al. Is antiretroviral therapy during pregnancy associated with an increased risk of preterm delivery, low birth weight, or stillbirth?
J Infect Dis . 2006;193:1195-201.
13. Hankin C, Thorne C, Newell ML. Does exposure to antiretroviral therapy affect
growth in the first 18 months of life in uninfected children born to
HIV -infected women? European Collaborative Study.
J Acquir Immune Defic Syndr . 2005;40:364-370.
14. New WHO Recommendations: Preventing mother-to-child transmission, November 30, 2009. Available at:
http://www.who.int/hiv /pub/mtct/mtct_key_mess.pdf . Accessed December 16, 2009.
15. World Health Organization Rapid Advice: Use of antiretroviral drugs for treating pregnant women and preventing
HIV infection in infants November 2009. Available at:
http://www.who.int/hiv /pub/mtct/rapid_advice_mtct.pdf . Accessed January 11, 2010.
16. 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;20:1281-1288.
17. Thior I, Lockman S, Smeaton LM, et al. Breastfeeding plus zidovudine prophylaxis for 6 months vs formula feeding plus infant zidovudine for 1 month to reduce mother-to-child transmission in Botswana: a randomized trial: the Mashi Study.
JAMA . 2006;296:794-805.
18. World Health Organization Child
Growth Standards. Available at:
http://www.who.int/childgrowth/1_what.pdf . Accessed December 18, 2009.
19. Chatterjee A, Bosch RJ, Hunter DJ, et al. Maternal disease stage and child undernutrition in relation to mortality among children born to
HIV -infected women in Tanzania.
J Acquir Immune Defic Syndr . 2007;46:599-606.
20. Black RE, Allen LH, Bhutta ZA, et al. Maternal and child undernutrition: global and regional exposures and health consequences.
Lancet . 2008;371:243-260.
21. Fawzi WW, Guillermo Herrera M, Spiegelman DL, et al. A prospective study of malnutrition in relation to child mortality in the Sudan.
Am J Clin Nutr . 1997;65:1062-1069.
22. Lockman S, Smeaton L, Shapiro R, et al.
Risk Factors and Timing of Infant Mortality Among HIV -Exposed Children in a Randomized Infant Feeding Trail in Botswana . [Abstract 643] 15th Conference on Retroviruses and Opportunistic Infections; 2008; Boston, MA.
23. Kuhn L, Sinkala M, Semrau K, et al. Elevations in mortality associated with weaning persist into the second year of life among uninfected children born to
HIV -infected mothers.
Clin Infect Dis . 2010;50:437-444.
24. Wei R, Msamanga GI, Spiegelman D, et al. Association between low birth weight and infant mortality in children born to human immunodeficiency virus 1-infected mothers in Tanzania.
Pediatr Infect Dis J . 2004;23:530-535.
25. Taha T, Kumwenda J, Kafulafula G, et al.
Maternal Highly Active Antiretroviral Treatment (HAART): Does It Improve Child Survival? [Abstract No. MOPEB083] IAS Conf
HIV Pathog Treatment; 2009; Cape Town.
