Body Composition Changes During Lactation in HIV-Infected and HIV-Uninfected South African Women

Papathakis, Peggy C PhD, MS, RD*†; Loan, Marta D. Van PhD, FACSM†§; Rollins, Nigel C MD, MRCP*∥; Chantry, Caroline J MD*¶; Bennish, Michael L MD‡#; Brown, Kenneth H MD†

JAIDS Journal of Acquired Immune Deficiency Syndromes: 1 December 2006 - Volume 43 - Issue 4 - pp 467-474
doi: 10.1097/01.qai.0000243094.42276.92
Epidemiology and Social Science

Background: The nutritional consequences of HIV infection in lactating women are unknown.

Objective: To measure the body composition of South African lactating women in relation to HIV status.

Methods: Fat-free mass (FFM) and fat mass (FM) using bioimpedance spectrometry (BIS) and anthropometric measurements were obtained at 8 and 24 weeks postpartum in 92 HIV-infected (HIVpos) and 50 HIV-uninfected (HIVneg) lactating mothers.

Results: At 8 weeks, HIVpos and HIVneg mothers were not significantly different in height (159.7 vs. 158.9 cm), weight (62.7 vs. 63.9 kg), body mass index (BMI; 24.6 vs. 25.3 kg/m2), FFM (40.7 vs. 42.8 kg), or FM (21.6 vs. 22.0 kg), respectively. In HIVpos women, the median CD4+ count was 621 (range: 101-1585) cells/μL; 95% had CD4+ counts >200 cells/μL. Between 8 and 24 weeks, HIVpos mothers had a mean weight loss of 1.4 kg in contrast to a 0.4-kg weight gain in HIVneg mothers (P < 0.01). There were no significant group differences with regard to change in FFM (0.3 vs. 0.1 kg; P = 0.9) and FM (−1.5 vs. −0.3 kg; P = 0.2).

Conclusion: HIVpos South African breast-feeding mothers without severe immune suppression lost weight and subcutaneous fat between 8 and 24 weeks postpartum, whereas HIVneg mothers gained weight. FFM was maintained postpartum in HIVpos and HIVneg mothers.

From the *Africa Center for Health and Population Studies KwaZulu-Natal Province, South Africa.; †University of California, Davis, Program in International Nutrition, Davis, CA; ‡Division of Geographic Medicine and Infectious Diseases, Tufts-New England Medical Center, Boston, MA; §Western Human Nutrition Research Center, United States Department of Agriculture, Davis, CA; ∥Department of Paediatrics, Nelson R. Mandela School of Medicine, University of KwaZulu Natal, Durban, South Africa; ¶Department of Pediatrics, University of California, Davis, Medical Center, Sacramento, CA; and #Nuffield Department of Clinical Medicine, University of Oxford, Oxford, England.

Received for publication July 16, 2005; accepted July 25, 2006.

Supported by the Wellcome Trust (grant 063957). P. Papathakis was supported by a Fulbright scholarship. M. Bennish was supported by a midcareer grant (1 K24 AI/HDO1671-01) from the National Institute of Allergy and Infectious Diseases of the US National Institutes of Health and by the Wellcome Trust (grant 62925). N. Rollins was supported by Wellcome Trust (grant 050524). M. Van Loan was supported by the US Department of Agriculture, Agricultural Research Service, Western Human Nutrition Research Center.

Reprints: Peggy C. Papathakis, PhD, MS, RD, University of California, Davis, Program in International Nutrition, One Shields Avenue, Davis, CA 95616 (e-mail:

Article Outline

In southern Africa, HIV prevalence is 35% or higher among women attending antenatal clinics.1 Despite concerns about the risk of transmitting HIV to the infant via breast-feeding2 and the suggestion that breast-feeding can have a deleterious effect on HIV-positive (HIVpos) women,3 African mothers often consider the use of replacement milks for infant feeding to be unacceptable, unaffordable, or unsafe.4-7 Because breast-feeding remains the most common form of infant feeding for all mothers in sub-Saharan Africa, it is important to understand the effect of HIV infection on the nutritional status of breast-feeding mothers.

Body composition is an important measure of nutritional status. Accurate measurement of body composition in lactating women is challenging, especially in poorer rural communities in developing countries, where HIV is often highly prevalent. Because hydration and density of fat-free mass (FFM) do not return to prepregnancy values in breast-feeding women until 2 or more weeks postpartum,8 body composition measurements are more accurate when taken at least 4 to 6 weeks after delivery.9 We previously validated the use of bioimpedance spectrometry (BIS) to measure FFM and fat mass (FM) in HIVpos and HIV-uninfected (HIVneg) breast-feeding mothers in rural South Africa by comparing it with the deuterium dilution method. We found that the methods correlated well with each other in both sets of women.10,11 There is little information on longitudinal changes in body composition in HIVpos individuals in developing countries. Most studies of body composition have been conducted amongst HIVpos men living in industrialized countries. These studies found that wasting was related to a loss of lean body mass in those with AIDS and other concurrent illness.12,13 By contrast, in stable HIVpos individuals with weight loss, FFM, FM, or both are lost, depending on the relative size of these compartments before the weight loss.14,15 Clinically stable HIVpos women with and without AIDS and with weight loss have been reported to have a greater relative loss of FM than lean tissue.15-17

HIVpos breast-feeding mothers may be particularly vulnerable to nutrient deficiencies because of inadequate dietary intake and the combined impact of increased nutrient requirements associated with breast-feeding, HIV,18 and other infections. We studied the change in body composition of breast-feeding mothers living in a poor rural community with high HIV prevalence in relation to their HIV status. We hypothesized that HIVpos breast-feeding mothers would lose more weight, FFM, and FM than HIVneg breast-feeding mothers between 8 and 24 weeks postpartum.

Back to Top | Article Outline


This longitudinal observational prospective study measured the body composition of HIVpos and HIVneg breast-feeding mothers between 8 and 24 weeks postpartum.

Back to Top | Article Outline


The study was conducted at the Africa Center for Health and Population Studies in northern KwaZulu-Natal province, South Africa. The study site is largely rural, with 1 periurban township; the inhabitants are mostly of Zulu ethnic origin. The population is characterized by a high prevalence of HIV (36.5% of women attending antenatal clinics in 2002),19 high unemployment (54%), poor access to clean water (87%), and high infant mortality (79 per 1000 live births).20

Back to Top | Article Outline


Mothers were consecutively enrolled at a single clinic in the periurban township and in 2 clinics in more rural areas. Mothers who were participating in an ongoing Africa Center for Health and Population Studies study to investigate the relationship between exclusive breast-feeding and HIV transmission (Vertical Transmission Study [VTS]) were offered entry into this study when they brought their infant to the clinic for routine 6-week immunizations. The HIVpos women were offered a single dose of nevirapine during labor or delivery, and none were receiving antiretroviral treatment for their HIV infection. Because most of the women enrolled in the study of HIV transmission were HIVpos, additional HIVneg mothers had to be recruited among women not participating in that study. Enrollment for this study began in May 2002, and the last 6-month follow-up visit was completed in February 2004. The field and laboratory staff members were blinded to the participants' HIV status.

Back to Top | Article Outline


Study participants were enrolled at 6 weeks postpartum and had subsequent study visits at 14 and 24 weeks. At these visits, maternal weight, height, and middle upper arm circumference (MUAC) were measured, and a venous blood sample was obtained. To standardize the time of the BIS measurements before the mother performed any physical activity, a morning home visit was made shortly after each clinic visit; skinfold thicknesses were also measured at this time.

Back to Top | Article Outline

Demographic and Health Information

Demographic and health information data were collected through the VTS of the Africa Center for Health and Population Studies. At each clinic visit, a nurse obtained a history of illnesses since the last visit, focusing on respiratory symptoms, diarrhea, and the presence of fever. Breast-feeding status was self-reported at each home visit.

Back to Top | Article Outline

HIV Viral Load and CD4+ T-Cell Counts

HIV viral load and CD4+ T-cell counts were determined at the Africa Center Virology laboratory in Durban. Total HIV RNA was isolated from plasma using guanidinium-silica methods (NucliSens Isolation Kit; Organon Teknika, Boxtel, The Netherlands) and an automated extractor (Organon Teknika). The NucliSens HIV-1 QT (bioMerieux, Inc., Durham, NC) assay has a quantitative range of 40 to >500,000 copies/mL of plasma. CD4+ T-cell cell counts were measured on venous blood within 24 hours of sampling using an Epics XL cell counter (Beckman Coulter, Fullerton, CA) and a 4-color protocol, with the primary gating being on CD45.

Back to Top | Article Outline

C-Reactive Protein and α-1-Acid Glycoprotein

C-reactive protein (CRP) and α-1-acid glycoprotein (AGP), measures of the acute-phase reaction, were used as proxy indicators of inflammation. Venous blood samples were labeled, protected from light, and kept cool for transport to Durban. Serum was separated, frozen, and stored at −70°C until tested in the Biochemistry Department at the University of KwaZulu-Natal in Durban. CRP and AGP were measured by immunologic agglutination using a Roche Modular P800 (Basel, Switzerland). The quantitative range of detection of CRP and AGP were 0.003 to 0.24 g/L and 0.25 to 3.0 g/L, respectively.

Back to Top | Article Outline


Height without shoes was measured to the nearest 0.1 cm using a stadiometer (Scales 2000, Durban, South Africa), and weight was measured to the nearest 100 g using an electronic digital scale (Scales 2000). Body mass index (BMI) was calculated as weight in kilograms per height in meters.2 MUAC was measured with a steel tape, and skinfold thicknesses were measured with Lange calipers (Lange, Cambridge, MA) according to standardized methods at 4 body sites: triceps skinfold (TSF), biceps skinfold (BSF), subscapular skinfold, and midline suprailiac skinfold. The sum of skinfold thicknesses was calculated as the sum of the measurements taken at all 4 body sites.

Back to Top | Article Outline

Bioimpedance Spectrometry

BIS measurements (BIS 4000 Analyzer, version 1.00d; Xitron Technologies, San Diego, CA) were taken on a morning home visit as soon as possible after the mother breast-fed. Mothers were requested not to eat or perform heavy physical labor on the morning of the visit. The impedance electrodes were placed in the standard tetrapolar positions. Multifrequency measurements were completed, and the impedance data were analyzed using the preprogrammed Cole-Cole plot for determination of FFM, FM. and percent of body fat (%BF).

Back to Top | Article Outline


To minimize the probability of type I error attributable to testing many outcome variables, we performed multivariate analysis of variance (MANOVA) on the 11 body composition measurements taken at each time point, and multivariate analysis of covariance (MANCOVA) on the change in all anthropometric variables (controlling for baseline values). HIV status was significant for the 6-month value and the change value; univariate results are presented.

For demographic and body composition variables, differences between HIVpos and HIVneg mothers were tested for significance using the χ2 test for categoric variables and the Student t test for continuous variables. For longitudinal measures, the groups were compared using repeated-measures analysis of covariance (ANCOVA; covariates included age, number of previous pregnancies, years of education, season of enrollment, clinic location, and illness). Variables not normally distributed were transformed logarithmically.

Descriptive univariate analysis was performed to identify outliers for FFM, FM, %BF, and changes over time in these measurements. Four of 251 measurements (3 in HIVpos women and 1 in an HIVneg woman) were excluded because of implausible FM and %BF measurement results. To understand the factors associated with weight change separately for HIVpos and HIVneg mothers, we performed stepwise regression analyses with the following independent variables: fever, upper or lower respiratory illness, CRP, AGP, CD4+ count, viral load, change in CD4+ count, change in viral load, years of education, and number of previous pregnancies.

Back to Top | Article Outline


This study was approved by the Research Ethics Committee of the Nelson R. Mandela School of Medicine of the University of KwaZulu-Natal, Durban, South Africa, and the Human Subjects Committee of the University of California, Davis, California. All participants provided written informed consent. No incentive or compensation was provided to participants.

Back to Top | Article Outline


One hundred forty-two mothers were enrolled: 92 were HIVpos and 50 were HIVneg. Twenty-four mothers (21.7% HIVpos women vs. 8.0% of HIVneg women; P = 0.03) withdrew or were lost to follow-up before the 24-week observations (Fig. 1). When compared with those who completed the study, those who did not differed only in that they were younger (23.3 ± 5.0 years vs. 26.0 ± 7.3 years; P = 0.03) and more likely to reside near to the periurban township clinic location (46.2% vs. 27.6%; P = 0.04).

Back to Top | Article Outline


The time of the first planned 6-week postpartum measurement at baseline occurred at 7.9 weeks (8 weeks) postpartum. Similarly, the planned 14- and 22-week measurements occurred closer to 16 and 24 weeks postpartum, respectively. Because the results at 16 weeks and 24 weeks were not statistically significantly different from each other, only the 24-week results are shown. The season of enrollment was not different between groups (P = 0.953).

Back to Top | Article Outline

Demographics and Health Information

Mothers were between 14 and 50 years of age, with 7 mothers <18 years of age and 3 > 40 years of age. Thirty-one percent of mothers were primiparous. The HIVpos and HIVneg mothers did not differ in demographic characteristics (Table 1). Reported illness symptoms were not different between groups and were generally low (point prevalence per group: <15% with respiratory symptoms and <5% with diarrhea symptoms), although the study was not powered to detect small differences between groups in symptoms. The more severe illnesses noted during clinic visits were 2 women with tuberculosis (1 HIVpos and 1 HIVneg), 8 with clinically diagnosed lower respiratory infection (7 HIVpos and 1 HIVneg), and 4 with septic wounds (3 HIVpos and 1 HIVneg). No malarial infections were identified.

Back to Top | Article Outline

Laboratory Results

In HIVpos mothers, only 1.3% at 8 weeks and 3.3% at 24 weeks had CD4+ cell counts <200 cells/μL. The median CD4+ cell counts at 8 and 24 weeks were 658 and 590 cells/μL, respectively. Median log10 HIV viral loads at 8 and 24 weeks were 4.02 and 3.99, respectively. For measures of acute inflammation, at 8 weeks, the mean log10 AGP was higher in HIVpos women than in HIVneg women, and the percent of HIVpos versus HIVneg mothers with an acute-phase response as measured by an AGP >1.2 g/L approached statistical significance (P = 0.056) (Table 2). At 24 weeks, there was no significant difference between groups for the mean log10 CRP or AGP, although the HIVpos mothers tended to have higher CRP values (P = 0.056), and more HIVpos mothers were having an acute-phase reaction as measured by CRP >0.01 g/L (Table 2).

Back to Top | Article Outline


At 8 weeks, none of the mothers were considered underweight by the World Health Organization (WHO) cutoff of BMI <18.5 kg/m2. Seven mothers (3 HIVpos) had a BMI <20 kg/m2. At 24 weeks, 2 mothers (1 HIVpos) had a BMI of <18.5 kg/m2 and 7 mothers (2 HIVpos) had a BMI <20 kg/m2. At 8 weeks postpartum, 6 mothers (3 HIVpos) had MUAC measurements <23 cm (fifth percentile of US National Heath and Nutrition Examination Survey). At 24 weeks, only 3 had low MUAC measures, 1 of whom was HIVpos. Conversely, 26.4% of the HIVpos mothers and 29.8% of the HIVneg mothers were classified as mildly to moderately overweight (BMI >26 kg/m2) at 8 weeks. At 24 weeks, the proportion of overweight mothers in the HIVpos group decreased slightly, whereas it increased among the HIVneg mothers (22.9% vs. 37.7%, respectively; P = 0.176).

At 8 weeks postpartum, there were no differences between HIVpos and HIVneg mothers in the mean values for any of the anthropometric measurements. At 24 weeks, mean TSF and BSF thicknesses were significantly less in the HIVpos mothers (Table 3). More HIVpos mothers than HIVneg mothers (70.0% vs. 46.7%, respectively; P = 0.05) lost weight between 8 and 24 weeks postpartum. When comparing the mean change during this time, the HIVpos mothers lost weight, whereas the HIVneg mothers gained slightly (−1.4 ± 3.1 kg vs. 0.4 ± 3.3 kg, respectively, P = 0.004). Likewise, HIVpos mothers' TSF thicknesses decreased, on average, whereas HIVneg mothers' TSF thicknesses increased. These differences between groups remained after controlling for baseline characteristics, CRP, AGP, and illness. For all mothers, those who ever reported illness during the period of follow-up tended to lose more weight compared with those who were never ill (−1.4 kg ever ill and −0.3 kg never ill; P = 0.07). There were no significant differences by HIV status in the change in weight in mothers ever ill versus never ill (HIVpos mothers' weight change if ever ill [−1.75 ± 3.62 kg] vs. those never ill [−1.13 ± 2.59 kg], P = 0.410; HIVneg mothers' weight change if ever ill [−0.57 ± 3.62 kg] vs. never ill [0.71 ± 3.23 kg], P = 0.272).

In the HIVpos mothers, the change in weight was not correlated to the change in CD4+ cell count or viral load and was not different when controlling for initial CD4+ cell count or viral load. When comparing HIVpos mothers with a lower CD4+ cell count (using a cutoff of <350 or <500 cells/μL) with those with higher values (≥350 or ≥500 cells/μL), there was no significant difference in mean change in weight (57 HIVpos mothers with CD4 count >350 cells/μL [−1.4 ± 3.3 kg] vs. 7 with <350 cells/μL [−1.7 ± 2.6 kg]; P = 0.822). When comparing HIVpos mothers with a normal BMI (18.5-25.9) with those with a BMI ≥26, there was no difference between groups in CD4+ count or viral load at any time point.

Back to Top | Article Outline

Bioimpedance Spectrometry

There were no significant differences between HIVpos and HIVneg mothers in the mean FFM, FM, and %BF results at any time point (Table 4). Although not statistically significant, the HIVpos mothers tended to lose absolute FM during the 4 months of follow-up. Controlling for age, years of education, number of previous pregnancies, initial weight, CRP, AGP, and recent illness did not affect these results. For HIVpos and HIVneg mothers, there was a small increase in mean FFM.

In HIVpos mothers, the change in weight was significantly correlated with the change in MUAC (r = 0.60, P < 0.001), TSF (r = 0.59, P < 0.001), BSF (r = 0.42, P = 0.001), subscapular skinfold (r = 0.55, P < 0.001), suprailiac skinfold (r = 0.60, P = 0.001), and FM measurements as determined by BIS (r = 0.26, P = 0.037). Using regression to determine which variables were the strongest predictors of weight change in HIVpos mothers, a history of fever at 8 weeks alone (r = 0.337, P = 0.041) and a history of fever at 8 weeks and 14 weeks as well as respiratory symptoms (r = 0.578, P = 0.014) were the strongest predictors. For the HIVneg mothers, 8-week weight and any respiratory symptoms were the strongest predictors (r = 0.514, P = 0.025); women whose weight at 8 weeks was above the median weight gained more than those whose weight was below the median weight.

Back to Top | Article Outline


In this study, we found that HIVpos South African breast-feeding mothers, 95% of whom had CD4+ counts >200 cells/μL, on average, lost weight between 8 and 24 weeks postpartum, whereas the HIVneg mothers gained weight. The weight loss in the HIVpos mothers was attributable primarily to loss of FM. Although HIVpos and HIVneg mothers did not differ in anthropometric measurements at 8 weeks, at 14 and 24 weeks postpartum, the HIVpos mothers had less subcutaneous fat as measured by TSF and BSF thicknesses. Reported illness (fever and respiratory illness) was the strongest predictor of weight loss in HIVpos mothers, whereas baseline weight was the strongest predictor of weight change in HIVneg mothers. FFM, FM, and %BF as measured by BIS were not different by HIV status at any time point; both groups of women tended to lose FM, whereas their FFM increased slightly between 8 and 24 weeks postpartum.

Our study found little evidence of wasting in HIVpos or HIVneg breast-feeding mothers living in this region. At 24 weeks postpartum, only 3% of HIVpos mothers and 10% of HIVneg mothers were classified as slightly underweight by a BMI <20. Despite weight loss in HIVpos mothers, most continued to have normal BMIs. Indeed, the mean %BF (32%) in all mothers at 24 weeks postpartum is generally higher than that of lactating women in other countries.21-24

There is no consistent pattern of weight change during lactation in developing countries. Butte and Hopkinson25 reviewed studies of weight change in affluent and low-income populations in developing countries during the first 6 months of lactation and found the results to be highly variable, ranging from a mean loss of 0.9 kg/mo to a mean gain of 0.8 kg/mo. The mean weight loss of HIVpos mothers in the current study was 0.34 kg/mo, with a mean gain of 0.10 kg/mo in their HIVneg peers (P = 0.005), both of which were within the weight change range found in other African breast-feeding mothers of unknown HIV status.25

There is little information on the body composition of HIVpos African women or of HIVpos lactating mothers available for comparison. Compared with a study of body composition in 46 HIVpos women (breast-feeding status unknown) in Zaire, the HIVpos mothers in our study weighed approximately 15 kg more, with approximately 2 kg more FFM and 13 kg more FM, as determined by bio-impedance analysis.26 This may be partially explained by the relatively low occurrence of acute illness and HIV symptomatology among the mothers in our study. Only 3% of HIVpos mothers in this study had advanced immunosuppression as defined by a CD4 count <200 cells/μL, in contrast to most reports from mother-to-child transmission prevention trials, which have reported 8% to 15% with a CD4 count <200 cells/μL.3,27,28

A mean 1.4-kg weight loss between 8 and 24 weeks postpartum in HIVpos mothers in the present study represents a theoretic total deficit of approximately 10,780 kcal over 112 days (or a daily deficit of approximately 96 kcal). This theoretic deficit represents 5.6% of the calculated resting energy expenditure (1717 kcal)29 based on the mean height, weight, and age of the HIVpos mothers in our study. In a study of 33 HIVpos presumably non-breast-feeding women, Grinspoon et al18 found that the resting energy expenditure was higher (119% ± 23%) than in 26 HIVneg weight-matched control women. We are unable to determine if the weight loss observed in this population was attributable to decreased energy intake, malabsorption, increased physical activity, energy expenditure associated with HIV or other illness, or a combination of these factors.

Our finding of weight loss in HIVpos mothers was somewhat different from a study of body weight in HIVpos and HIVneg Rwandan mothers,30 which found that HIVpos and HIVneg mothers gained weight between 10 days and 5 months postpartum (0.7 ± 3.8 kg in HIVpos mothers and 1.9 ± 4.7 kg in HIVneg mothers; P = 0.003), even though the measurement at 10 days most likely still reflects some of the fluid gained during pregnancy. Although this period of follow-up was slightly different, in our study, the HIVpos mothers lost weight, whereas the HIVneg mothers gained 75% less than their Rwandan peers. The Rwandan study did not comment on whether the mothers were breast-feeding, and body composition was not mentioned.

Nduati et al3 suggested that weight loss associated with lactation might be more severe in HIVpos women than in healthy women. In a randomized clinical trial of formula-feeding versus breast-feeding in Kenyan HIVpos mothers, baseline weight measured between 0.5 and 3 months postpartum was compared with weight measured between 5 and 9 months postpartum. Formula-feeding mothers lost no weight, whereas breast-feeding mothers lost 0.17 kg/mo (P = 0.03). The wide time span for each measurement makes interpretation of this finding difficult.

Nearly 25% of the HIVpos mothers in our study had a BMI ≥25 and were thus overweight or obese using this measure. Anyone who is overweight or obese is generally advised to lose weight to avoid obesity-related health consequences, such as diabetes and heart disease. Our data raise the dilemma regarding whether HIVpos women should try to lose weight to achieve a more healthy body composition with less body fat. Several studies have found that HIVpos women with higher BMIs were less likely to have a CD4+ cell count <200 cells/μL and more likely to have slower disease progression than their thinner counterparts,31-33 although the causal direction of this association is unknown. It is difficult to know whether weight loss would be in the best long-term interest for these overweight and obese HIVpos mothers.

There are several limitations of our study that should be noted. First, to determine the effect of breast-feeding on HIVpos mothers' nutritional status, ideally, we would have preferred to compare their body composition with that of HIVpos non-breast-feeding mothers. Few HIVpos mothers in this population elect to formula-feed, however. Thus, we were unable to determine whether the weight loss in HIVpos women was primarily an effect of HIV or a combination of breast-feeding and the presence of HIV infection. During the 17 months of subject enrollment, only 5 HIVpos non-breast-feeding mothers were followed. The non-breast-feeding mothers were no different in body composition or changes than the HIVpos breast-feeding mothers, but because of the small sample size, the power to detect differences was small. Second, weight change postpartum can be influenced by many factors, including prepregnancy weight, weight gain during pregnancy, and exclusivity of breast-feeding.23,34 We attempted to obtain prepregnancy weight information, but most women did not know their prepregnancy weight. In addition, we attempted to find information on the mothers' weight at the time of their first antenatal visit, but few records could be located at the clinics. Information on exclusivity of breast-feeding and breast health collected by the VTS was not yet available for analysis; therefore, we were unable to adjust for these parameters. Third, 6 of the presumed HIVneg mothers enrolled were subsequently determined to be HIVpos and were recategorized to that group. Therefore, the number of HIVneg mothers with follow-up until 24 weeks postpartum is less than that originally estimated to be necessary to detect a difference between groups of 0.5 SD for the body composition variables; this decreased the power to detect the differences we set out to find. Based on our actual sample size, we have the ability to detect a difference of 0.58 SD between groups. Fourth, because <5% of the mothers had advanced immunosuppression as defined by CD4+ counts <200 cells/μL, we are unable to assess the impact of more advanced disease stage on body composition changes. Therefore, our results may only be generalizable to the “well” population of HIVpos breast-feeding women. Finally, although we previously validated BIS against deuterium dilution in this population, we found likely errors in approximately 3.0% of the BIS results, especially during extremely hot and humid climate conditions. These were equally distributed, however, between HIVpos and HIVneg mothers; thus, they should not have biased the results.

In summary, between 8 and 24 weeks postpartum, HIVpos breast-feeding South African mothers, on average, lost weight and subcutaneous fat, whereas the HIVneg mothers gained weight. Although most HIVpos mothers were not severely immune suppressed, there were no significant changes in CD4+ cell counts or HIV viral loads. FFM did not change significantly in either group of mothers. The prognostic importance of maintaining versus losing excess weight and FM in this group of women is unknown. Further study is needed to determine whether weight and fat loss in normal and overweight breast-feeding HIVpos mothers is associated with long-term consequences for their health.

Back to Top | Article Outline


P. C. Papathakis was responsible for the study concept and design, field supervision, analysis and interpretation of results, and drafting the manuscript. K. H. Brown, M. D. Van Loan, N. C. Rollins, C. J. Chantry, and M. L. Bennish contributed to the study design, interpretation of results, and manuscript revisions. The manuscript has been reviewed and approved by all authors. This paper is written on behalf of the Child Health Group of the Africa Center for Health and Population studies: H. M. Coovadia, R. M. Bland, A. Coutsoudis, M. L. Newell, and J. van den Broeck. The authors gratefully thank the Maternal Nutrition Study staff for subject recruitment; Carina Herbst and Ruth Bland for study management support; the Africa Center VTS HIV counselors, nurses, clinic assistants, and staff for study organization and data collection; the mothers who generously gave of their time and participated in this study; and the Hlabisa district clinical nursing staff at Madwaleni, Nkundusi, and KwaMsane clinics. The insightful computer programming and data management support of Arjan van Bentem and patient statistical guidance of Jan Peerson are thankfully acknowledged. The Western Human Nutrition Research Center, United States Department of Agriculture, Davis, CA, provided bioimpedance equipment.

Back to Top | Article Outline


1. (UNAIDS) JUNPoHA. 2004 Report on the Global HIV/AIDS Epidemic: Fourth Global Report. Geneva, Switzerland: UNAIDS; 2004.
2. WHO. New Data on the Prevention of Mother-to-Child Transmission of HIV and Their Policy Implications: Conclusions and Recommendations. Technical Consultation on Behalf of the UNFPA/WHO/UNAIDS Inter-Agency Task Team on Mother-to-Child Transmission of HIV. Geneva, Switzerland: WHO; 2001.
3. Nduati R, Richardson BA, John G, et al. Effect of breastfeeding on mortality among HIV-1 infected women: a randomised trial. Lancet. 2001;357:1651-1655.
4. Coutsoudis A, Goga AE, Rollins N, et al. Free formula milk for infants of HIV-infected women: blessing or curse? Health Policy Plan. 2002;17:154-160.
5. Papathakis P, Rollins N. Are WHO/UNAIDS/UNICEF-recommended replacement milks for infants of HIV-infected mothers appropriate in the South African context? Bull World Health Organ. 2004;82:164-171.
6. Latham MC, Preble EA. Appropriate feeding methods for infants of HIV infected mothers in sub-Saharan Africa. BMJ. 2000;320:1656-1660.
7. Phadke MA, Gadgil B, Bharucha KE, et al. Replacement-fed infants born to HIV-infected mothers in India have a high early postpartum rate of hospitalization. J Nutr. 2003;133:3153-3157.
8. Hopkinson JM, Butte NF, Ellis KJ, et al. Body fat estimation in late pregnancy and early postpartum: comparison of two-, three-, and four-component models. Am J Clin Nutr. 1997;65:432-438.
9. Butte NF, Hopkinson JM, Ellis KJ, et al. Changes in fat-free mass and fat mass in postpartum women: a comparison of body composition models. Int J Obes Relat Metab Disord. 1997;21:874-880.
10. Papathakis PC, Rollins NC, Brown KH, et al. Comparison of isotope dilution with bioimpedance spectroscopy and anthropometry for assessment of body composition in asymptomatic HIV-infected and HIV-uninfected breastfeeding mothers. Am J Clin Nutr. 2005;82:538-546.
11. Papathakis P, Rollins N, Van Loan M, et al. Body composition changes in HIV-infected South African breastfeeding mothers [abstract ThPeC7290]. Presented at: XV International AIDS Conference; 2004; Bangkok.
12. Kotler DP, Wang J, Pierson RN. Body composition studies in patients with the acquired immunodeficiency syndrome. Am J Clin Nutr. 1985;42:1255-1265.
13. Kotler DP, Tierney AR, Wang J, et al. Magnitude of body-cell-mass depletion and the timing of death from wasting in AIDS. Am J Clin Nutr. 1989;50:444-447.
14. Mulligan K, Tai VW, Schambelan M. Cross-sectional and longitudinal evaluation of body composition in men with HIV infection. J Acquir Immune Defic Syndr Hum Retrovirol. 1997;15:43-48.
15. Forrester JE, Spiegelman D, Tchetgen E, et al. Weight loss and body-composition changes in men and women infected with HIV. Am J Clin Nutr. 2002;76:1428-1434.
16. Swanson B, Hershow RC, Sha BE, et al. Body composition in HIV-infected women. Nutrition. 2000;16:1064-1068.
17. Grinspoon S, Corcoran C, Miller K, et al. Body composition and endocrine function in women with acquired immunodeficiency syndrome wasting. J Clin Endocrinol Metab. 1997;82:1332-1337.
18. Grinspoon S, Corcoran C, Miller K, et al. Determinants of increased energy expenditure in HIV-infected women. Am J Clin Nutr. 1998;68:720-725.
19. Department of Health RoSA. National HIV and Syphilis Antenatal Sero-Prevalence Survey in South Africa: 2002. Directorate: Health Systems Research, Research Coordination and Epidemiology. Available at: 2004.
20. Hosegood V. Africa Centre Demographic Information System Household Socio-Economic Data, 2001. Mtubatuba, South Africa: Africa Centre for Health and Population Studies; 2002.
21. Motil KJ, Sheng HP, Kertz BL, et al. Lean body mass of well-nourished women is preserved during lactation. Am J Clin Nutr. 1998;67:292-300.
22. Lovelady CA, Garner KE, Moreno KL, et al. The effect of weight loss in overweight, lactating women on the growth of their infants. N Engl J Med. 2000;342:449-453.
23. Butte NF, Garza C, Stuff JE, et al. Effect of maternal diet and body composition on lactational performance. Am J Clin Nutr. 1984;39:296-306.
24. Kac G, Benicio MH, Velasquez-Melendez G, et al. Gestational weight gain and prepregnancy weight influence postpartum weight retention in a cohort of Brazilian women. J Nutr. 2004;134:661-666.
25. Butte NF, Hopkinson JM. Body composition changes during lactation are highly variable among women. J Nutr. 1998;128(Suppl):381S-385S.
26. Kotler DP, Thea DM, Heo M, et al. Relative influences of sex, race, environment, and HIV infection on body composition in adults. Am J Clin Nutr. 1999;69:432-439.
27. John GC, Nduati RW, Mbori-Ngacha DA, et al. Correlates of mother-to-child human immunodeficiency virus type 1 (HIV-1) transmission: association with maternal plasma HIV-1 RNA load, genital HIV-1 DNA shedding, and breast infections. J Infect Dis. 2001;183:206-212.
28. Bailey A, Newell M-L, Peckham C, et al. Maternal viral load and vertical transmission of HIV-1: an important factor but not the only one. The European Collaborative Study. AIDS. 1999;13:1377-1385.
29. Harris JBF. A Biometric Study of Basal Metabolism in Man. Washington, DC: Carnegie Institute of Washington; 1919.
30. Ladner J, Castetbon K, Leroy V, et al. Pregnancy, body weight and human immunodeficiency virus infection in African women: a prospective cohort study in Kigali (Rwanda), 1992-1994. Pregnancy and HIV Study Group (EGE). Int J Epidemiol. 1998;27:1072-1077.
31. Jones CY, Hogan JW, Snyder B, et al. Overweight and human immunodeficiency virus (HIV) progression in women: associations HIV disease progression and changes in body mass index in women in the HIV epidemiology research study cohort. Clin Infect Dis. 2003;37 (Suppl 2):S69-S80.
32. Shor-Posner G, Campa A, Zhang G, et al. When obesity is desirable: a longitudinal study of the Miami HIV-1-infected drug abusers (MIDAS) cohort. J Acquir Immune Defic Syndr. 2000;23:81-88.
33. Shuter J, Chang CJ, Klein RS. Prevalence and predictive value of overweight in an urban HIV care clinic. J Acquir Immune Defic Syndr. 2001;26:291-297.
34. Brewer MM, Bates MR, Vannoy LP. Postpartum changes in maternal weight and body fat depots in lactating vs nonlactating women. Am J Clin Nutr. 1989;49:259-265.

breast-feeding; lactation; body composition; fat-free mass; fat mass; HIV

© 2006 Lippincott Williams & Wilkins, Inc.