点击显示 收起
1 From the Africa Centre for Health and Population Studies, Somekele, South Africa (PCP, NCR, and MLB); the Program in International Nutrition, University of California, Davis, CA (PCP and KHB); the Department of Pediatrics, University of California Davis Medical Center, Sacramento, CA (CJC); the University of KwaZulu-Natal, Nelson R Mandela School of Medicine, Department of Paediatrics (NCR); Tufts–New England Medical Center, Division of Geographic Medicine and Infectious Diseases, Boston, MA (MLB); the Nuffield Department of Clinical Medicine, University of Oxford, Oxford, United Kingdom (MLB)
2 Supported by a grant from the Wellcome Trust (grant 063957). PCP was supported by a Fulbright scholarship. MLB 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). NCR was supported by the Wellcome Trust (grant 050524). 3 Reprints not available. Address correspondence to PC Papathakis, Department of Food Science and Nutrition, Cal Poly, San Luis Obispo, CA 93407. E-mail: ppatha{at}calpoly.edu.
ABSTRACT
Background:Little information on the micronutrient status of HIV-infected (HIV-positive) breastfeeding women is available.
Objective:The objective was to compare the protein and micronutrient status of South African breastfeeding women by HIV status.
Design:Serum albumin, prealbumin, vitamin B-12, folate, retinol, -tocopherol, hemoglobin, ferritin, and zinc concentrations were compared between 92 HIV-positive and 52 HIV-uninfected (HIV-negative) mothers 6, 14, and 24 wk after delivery. C-reactive protein and 1-acid glycoprotein were used as proxy indicators of an inflammatory process.
Results:Mean albumin and prealbumin were significantly lower in HIV-positive mothers, and a higher proportion of HIV-positive mothers had low albumin concentrations (<35 g/L). Less than 45% of the mothers were vitamin B-12 or folate sufficient. Significantly more HIV-positive (70.5%) than HIV-negative (46.2%) mothers had marginal vitamin B-12 status (P < 0.05), and mean folate concentrations were lower in HIV-positive mothers (P = 0.05). Mean serum retinol was significantly lower in HIV-positive mothers, even after control for the acute phase response. At 24 wk, 70% of both groups had an -tocopherol deficiency (<11.6 µmol/L), but no significant difference by HIV status was observed. More HIV-positive (33.3%) than HIV-negative (8.7%) mothers had anemia (P = 0.018), whereas 25% of all mothers had low serum ferritin concentrations. After the acute phase response was controlled for, zinc deficiency was more common in HIV-positive (45.0%) than in HIV-negative (25.0%) mothers (P = 0.05).
Conclusions:Deficiencies in vitamins B-12, folate, -tocopherol, ferritin, and zinc are common in South African breastfeeding mothers. HIV-positive mothers had lower mean serum concentrations of albumin, prealbumin, folate, retinol, and hemoglobin than did HIV-negative mothers.
Key Words: HIV infection • breastfeeding women • South Africa • micronutrient status
INTRODUCTION
An estimated 17 million women aged 15–49 y worldwide are HIV-infected (HIV-positive), 77% of whom reside in sub-Saharan Africa (1, 2). In many areas of southern Africa, the prevalence of HIV is 40% in women attending antenatal clinics (2). Despite concerns about the risk of transmitting HIV to infants via breastfeeding (3) and the suggestion that breastfeeding may have an adverse effect on the nutritional status of HIV-positive women (4), the use of replacement milks for infant feeding is often unacceptable, unaffordable, or unsafe (5-8). Because breastfeeding remains the most common form of infant feeding for both HIV-positive and HIV-uninfected (HIV-negative) mothers in this setting, it is important to understand the effect of HIV infection on the nutritional status of breastfeeding mothers.
Little is known about the effect of lactation on maternal micronutrient status. The mammary gland exerts metabolic priority for many micronutrients, even at the expense of maternal stores (9). The maternal recommendations for protein, riboflavin, pantothenic acid, zinc, iodine, and vitamins A, B-6, and C during lactation are 40–90% higher than those before pregnancy; the recommendations for thiamine, niacin, folate, vitamin E, and selenium are 25% higher (10).
Complex bidirectional relations exist between HIV infection, nutrition, and immune function (11). In HIV-negative adults, inadequate micronutrient intake and mild deficiency are associated with altered immune function. Zinc, selenium, iron, copper, folic acid, and vitamins A, C, E, and B-6 all have important roles in immune function (12). Several studies in industrialized countries have found an association between HIV disease progression and low blood concentrations of copper, zinc, riboflavin, and vitamins A, E, B-6, and B-12, although the causal direction of the relation is uncertain (13-17). Nevertheless, there is concern that inadequate diets may compound the susceptibility to secondary infection and thereby cause a further worsening of immune function in HIV-positive persons.
The aim of this study was to compare the serum protein and micronutrient status of HIV-positive and HIV-negative breastfeeding South African mothers during the first 6 mo after delivery. We hypothesized that HIV-positive mothers would have lower serum concentrations of vitamins A and B-12, selenium and zinc than HIV-negative lactating women during the first 6 mo post partum.
SUBJECTS AND METHODS
Study design
This prospective longitudinal observation study measured select serum proteins, vitamins, and minerals in clinic-attending HIV-positive and HIV-negative breastfeeding South African mothers between 6 and 24 wk after delivery. The Maternal Nutrition (MN) study discussed herein was conducted at the Africa Centre for Health and Population Studies in northern KwaZulu-Natal Province, South Africa. The study site is largely rural and has only one periurban township. The population is mostly of Zulu ethnic origin and is characterized by a high prevalence of HIV infection (36.5% of women attending antenatal clinics in 2002) (18), unemployment (54%), poor access to clean water (87%), and a high infant mortality rate (79/1000 live births) (19).
Subjects
Most of the mothers enrolled in the MN study were simultaneously participating in the Africa Centre Vertical Transmission (VT) study to investigate the relation between exclusive breastfeeding and HIV transmission from mother-to-child; however, some HIV-negative mothers in the MN study were recruited from outside the VT study. The MN study was a substudy of the larger VT study. The subjects included HIV-positive pregnant women living and attending antenatal clinics in the region and a time-related random sample of HIV-negative pregnant women attending the same clinics. Mothers who intended to return to school or work within 2 mo of delivery or to leave the area within 3 mo of delivery were excluded. At the time of the MN study, the provincial government's prevention of mother-to child transmission program had only just begun to offer free infant formula, and few women elected to use the formula. Mothers were consecutively enrolled at one clinic in the periurban township and at 2 clinics in rural areas. Enrollment for the MN study began in May 2002, and the last 6-mo follow-up visit was completed in February 2004. The field and laboratory staff members were blinded to the participants' HIV status. Demographic data were collected either through the VT study or directly from the women enrolled only in the MN study. Demographic information included years of schooling, employment status, household structure, type of cooking fuel used, and source of water. Breastfeeding status was self-reported at each study visit.
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, CA. All participants provided written informed consent. There was no incentive or compensation provided to participants.
Measurements
The study participants were enrolled 6 wk after delivery and had subsequent study visits at 14 and 24 wk after delivery; the visits were timed to coincide with their infants' routine immunization schedule and clinic visits.
HIV viral load and CD4+ T cell count
HIV viral load and CD4+ T cell counts were determined in the Africa Centre Virology laboratory in Durban. Total HIV RNA was isolated from plasma by using guanidinium-silica methods (NucliSens Isolation Kit; Organon Teknika, Boxtel, Netherlands) and an automated extractor (Organon Teknika). The NucliSens HIV-1 QT assay has a quantitative range of 40 to >500 000 copies/mL plasma. CD4+ T cell counts were measured in fresh venous blood within 24 h of sampling with the use of an Epics XL cell counter (Beckman Coulter, Fullerton, CA) and a 4-color protocol, with the primary gating being on CD45.
Serum and plasma proteins and micronutrients
Venous blood samples were drawn in the morning, but fasting status was not ascertained. Samples were collected into mineral-free gel separator tubes for serum protein, minerals, vitamins B-12, and folate measurements. EDTA-coated tubes were used to collect blood for the measurement of vitamins A and E. Vials were labeled, protected from light, and kept cool for transport to the Africa Centre Laboratory in KwaMsane or Somkhele, where they were centrifuged within 3 h of collection. The serum or plasma was separated, frozen, and stored at –70 °C until tested in the Biochemistry Department at the University of KwaZulu-Natal in Durban. All proteins were measured with an automated clinical chemistry analyzer (Modular P800; Roche, Basel, Switzerland).
C-reactive protein (CRP) and 1-acid glycoprotein (AGP), measures of the acute phase response, were used as proxy indicators of an inflammatory process. AGP and CRP were measured by immunologic agglutination; the quantitative range of detection for AGP and CRP concentrations were 0.25–3.00 and 0.002–0.240 g/L, respectively. If either AGP or CRP were elevated, the subject was considered to be having an acute phase response; the cutoff for AGP was 1.2 g/L and for CRP was >5.0 mg/L (20). The continuous variables were converted to categorical variables on the basis of the above values for the acute phase response.
Albumin and prealbumin were used as proxy measures of longer- and shorter-term visceral protein status (half-lives of 20 and 2 d, respectively). Albumin was measured with a colorimetric assay; the quantitative range of detection was 2.0–70.0 g/L, and a low albumin concentration was defined as <35.0 g/L (21). Prealbumin was measured by immunologic agglutination; the quantitative range of detection was 0.015 –0.80 g/L, and a low prealbumin concentration was defined as <0.14 g/L.
Vitamin B-12 and folate in serum were measured with a paramagnetic particle, chemiluminescent immunoassay (Beckman Coulter). Vitamin B-12 deficiency was defined as <150 pmol/L and marginal vitamin B-12 status as <210 pmol/L. Folate deficiency was defined as <6.8 nmol/L and marginal folate status as <14.0 nmol/L (22).
Retinol and -tocopherol were measured by HPLC (Series 200; Perkin Elmer, Wellesley, MA). The plasma sample was deproteinized by using ethanol, which was followed by liquid-liquid extraction with hexane. The hexane phase was evaporated, and the precipitate was reconstituted in methanol. Vitamin A deficiency was defined as a plasma retinol concentration <0.35 µmol/L and marginal vitamin A status as <0.70 µmol/L. Vitamin E deficiency was defined as an -tocopherol concentration <11.6 µmol/L.
Serum ferritin was measured by immunoassay with a Modular E170 (Roche); the quantitative range of detection was 0.5–2000 µg/L, and iron deficiency was defined as a serum ferritin concentration <13 µg/L. Anemia was defined as a hemoglobin concentration <12.0 g/dL.
Zinc, selenium, and copper were measured by inductively coupled plasma mass spectrometry (Elan DRC II; Perkin Elmer). Deficiency was defined as a zinc concentration <10.2 µmol/L (nonfasting cutoff) (23), a selenium concentration <0.59 µmol/L (24), and a copper concentration <12.56 µmol/L.
Statistical analysis
The sample size estimate was based on existing data for those nutrients that had been previously studied in HIV-positive women (13, 14, 17, 25-27). Fifty-one women per group were required to detect a 0.5-SD unit difference between groups for selected nutrients, with a 0.05 level of significance and 80% power. We anticipated 20% attrition, with a target sample size of 61 per group.
For demographic and biochemical variables at each time point, differences between HIV-positive and HIV-negative mothers were tested for significance with the use of a chi-square test for categorical variables and Student's t test for continuous variables. For longitudinal measures, the groups were compared by using repeated-measures mixed-model analysis of variance. Covariates included age, number of previous pregnancies, years of education, and clinic location. Variables that were not normally distributed were transformed logarithmically. Retinol distributions could not be normalized with logarithmic transformation, so a square root transformation was used.
Initially, the analysis for a nutrient using AGP as a continuous variable was completed and compared with the same analysis using AGP as a categorical variable, and likewise for CRP. We found it more useful to use the categorical treatment of the acute phase protein variable for the analysis reported here. For each nutrient (albumin, prealbumin, and all micronutrients), AGP or CRP from each time point was entered into a mixed model individually to explore interactions and significance; in addition, AGP or CRP from all 3 time points was entered together. Because the results are consistent between methods and for simplicity, we report that the significant difference between groups for individual nutrients remained after the control for measures of acute phase response.
RESULTS
One hundred forty-four mothers were enrolled from clinics, of whom 92 were HIV-positive and 52 were HIV-negative. The number of HIV-positive mothers was higher than anticipated for 2 main reasons. First, because of the masking procedure, the staff members were unaware of the HIV status of subjects in the VT study, and a greater than expected proportion of HIV-positive women were enrolled by chance. Second, for the 19 presumed HIV-negative mothers who were enrolled outside of the VT study, 6 were found to be HIV-positive. Twenty-four mothers (21.7% of the HIV-positive women and 7.6% of the HIV-negative mothers; P = 0.03) withdrew or were lost to follow-up before the 6-mo observations (Figure 1). When compared with those who completed the study, those who did not finish were younger (23.3 ± 5.0 compared with 26.0 ± 7.3 y; P = 0.03) and were more likely to attend the periurban township clinic (46.2% compared with 27.6%; P = 0.04). Those who failed to complete the study did not differ from the others with regard to socioeconomic variables, the number of previous pregnancies, micronutrient status, or viral load or CD4+ measurements.
FIGURE 1.. Details of subject enrollment and follow-up.
Nine mothers (n = 6 HIV-positive and 3 HIV-negative) stopped breastfeeding between the second and third follow-up visits. Their demographic, anthropometric, and biochemical measures were not different from those who continued breastfeeding, so they were not removed from analysis.
Although the provision of multivitamins is reported to be the standard of care for HIV-positive mothers, 20 HIV-positive and 7 HIV-negative mothers reported at only one postnatal clinic visit that they had taken multivitamins during the previous 4 wk. Five HIV-positive and 1 HIV-negative mother reported taking both folate and iron supplements at one visit only. 1 HIV-positive mother reported taking folate and iron supplements at 2 postnatal visits, and 1 HIV-positive and 2 HIV-negative mothers reported taking iron alone at one visit.
The actual mean number of weeks after delivery for the visit scheduled at 6, 14, and 24 wk were 6.7 ± 1.6 and 7.1 ± 2.1 wk (P = 0.24), 15.1 ± 2.1 and 14.9 ± 2.1 wk (P = 0.57), and 23.7 ± 3.6 and 23.0 ± 3.0 wk (P = 0.32) in the HIV-positive and HIV-negative mothers, respectively. The mothers were between 14 and 50 y of age; 8 mothers were <18 y and 3 were >40 y of age. Thirty-five percent of the mothers were primiparous. The demographic characteristics did not differ significantly between the HIV-positive and HIV-negative mothers (Table 1).
View this table:
TABLE 1. Baseline characteristics of the breastfeeding mothers by HIV status
Laboratory results
Of the HIV-positive mothers, only 2 had CD4+ cell counts <200 cells/µL at 6 (2.6%) and 24 wk (3.3%) after delivery. The median (25th–75th percentile range) CD4+ cell counts at 6, 14, and 24 wk after delivery were 622 (433–788), 673 (429–770), and 590 (466–793) cells/µL, respectively. The median HIV log10 (25th–75th percentile range) viral load at 6, 14, and 24 wk after delivery were 4.04 (3.11–4.82), 4.15 (3.48–4.61), and 3.97 (3.21–4.57), respectively.
Proteins
The mean AGP concentration was higher in the HIV-positive women than in the HIV-negative women (Table 2), and the percentage of HIV-positive mothers with an acute phase response (ie, AGP 1.2 g/L) was significantly higher than that of the HIV-negative mothers (Table 3) . Mean AGP concentrations decreased significantly over time after delivery in both groups. Mean CRP concentrations tended to be higher in the HIV-positive than in the HIV-negative mothers, but the difference was not significant (P = 0.075), and no significant difference in the proportion of HIV-positive and HIV-negative mothers with an acute phase response (ie, CRP > 5 mg/L) was observed (Table 3). Surprisingly, at each time point, 50% of both HIV-positive and HIV-negative mothers were having an acute phase response, as measured by CRP. Serum albumin and prealbumin concentrations were significantly lower in the HIV-positive mothers (Table 2). This significant difference between groups remained after measures of the acute phase response and demographic variables were controlled for. Serum albumin increased significantly in both groups between 6 and 24 wk after delivery. No significant interactions between HIV group and time were observed for any of the proteins, nor were any 3-factor interactions observed. For albumin, however, there was a significant interaction between HIV group and AGP. The nature of the interaction was explored by comparing those with an acute phase response with those without one, by HIV status. Those in the HIV-positive group with an elevated AGP concentration had a significantly lower mean albumin concentration than did those with a normal AGP concentration; no significant difference was observed in the HIV-negative group (Table 4). Significantly more HIV-positive than HIV-negative mothers had low serum albumin and prealbumin concentrations (Table 3).
View this table:
TABLE 2. Serum protein concentrations in South African breastfeeding women by HIV status and time after delivery1
View this table:
TABLE 3. Percentage of South African breastfeeding women with abnormal serum protein concentrations by HIV status and time after delivery1
View this table:
TABLE 4. Mean serum nutrients and interactions with 1-acid glycoprotein (AGP) in South African breastfeeding HIV-infected (HIV+) and HIV-uninfected (HIV–) mothers by acute phase response status1
Vitamins and minerals
No significant differences in mean micronutrient status indicators were observed when mothers who reported ever taking supplements were compared with those who reported never taking supplements, regardless of HIV status; however, the power to detect differences was small.
Vitamin B-12
No significant differences in mean serum vitamin B-12 concentrations were observed between groups. (Table 5). The results did not change after demographic variables, measures of acute phase response, or CD4+ cell counts and viral load in the HIV-positive mothers were controlled for. No significant differences between groups were observed in the proportion of women who were vitamin B-12 deficient (Table 6); however, 45% of all mothers had marginal vitamin B-12 concentrations (210 pmol/L). A significantly greater proportion of HIV-positive mothers had a marginal B-12 status, whereas a greater proportion of HIV-negative mothers were vitamin B-12 sufficient (Table 6).
View this table:
TABLE 5. Serum and plasma micronutrient concentrations in South African breastfeeding women by HIV status and time after delivery
View this table:
TABLE 6. Percentage of South African breastfeeding women with abnormal serum micronutrients by HIV status and time after delivery1
Folate
The mean serum folate concentration was low (<14.0 nmol/L) in both groups and was significantly lower in the HIV-positive mothers (Table 5). The results of the 2-group comparisons did not change after demographic variables, measures of acute phase response, or CD4+ counts and viral load in the HIV-positive mothers were controlled for. More than 70% of the mothers in both groups had marginal folate status at any time point, but no significant differences were observed in the proportion of mothers with low folate status by HIV group (Table 6). With time, mean serum folate concentrations decreased and the proportion of women with a deficiency increased.
Vitamin A
HIV-positive mothers had significantly lower mean plasma retinol concentrations than did HIV-negative mothers (Table 5). This significant difference between groups remained after measures of the acute phase response and demographic variables were controlled for. A significant interaction between HIV group and AGP was observed. To further explore this interaction, we compared the mean retinol concentrations of those with a positive acute phase response with those without a response, by HIV status. The mean retinol concentration was significantly lower in the HIV-positive mothers with an acute phase response (ie, AGP 1.2 g/L), but was not significantly different in the HIV-negative group (Table 4). More than 85% of all mothers were vitamin A sufficient (0.70 µmol/L) and <6% of the mothers were vitamin A deficient.
Vitamin E
Mean serum -tocopherol concentrations were low (<11.6 µmol/L) at 14 and 24 wk and did not differ significantly by HIV status (Table 5). The results of the groupwise comparisons did not change after demographic variables, measures of the acute phase response, or CD4+ cell counts and viral load in the HIV-positive mothers were controlled for. Many of the women had a deficiency at all time points, which reached 70% at 24 wk, but no significant differences in the percentage of women with low values were observed between groups (Table 6). Mean -tocopherol concentrations decreased significantly over time after delivery in both groups.
Hemoglobin
The mean hemoglobin concentrations was significantly lower in the HIV-positive mothers and the differences remained after measures of the acute phase response and demographic variables were controlled for (Table 5). The proportion of subjects who were anemic (hemoglobin < 12.0 g/dL) was higher in the HIV-positive group than in the HIV-negative group (Table 6).
Ferritin
The mean serum ferritin concentration was not significantly different by HIV status, either before or after measures of the acute phase response were controlled for, and decreased significantly over time after delivery in both groups (Table 5). Depending on the time point, between 15% and 34% of the mothers were considered iron depleted, but no significant between-group differences were observed (Table 6).
Zinc
The mean serum zinc concentrations were not significantly different by HIV status, either before or after measures of the acute phase response were controlled for (Table 5). Because the fasting status of the mothers at the time blood was drawn was not recorded, the more conservative nonfasting cutoff of serum zinc, <10.2 µmol/L, was used. No significant group effect was observed on the proportion of mothers who were deficient (HIV-positive: 45%; HIV-negative: 25%; P = 0.11); however, the group effect became significant when the presence of an acute phase response was controlled for (Table 6). When the fasting cutoff of 10.7 µmol/L was used instead, 41.9%, 23.8%, and 41.0% of all mothers at 6, 14 and 24 wk were deficient, respectively; however, no significant differences were observed between groups.
Selenium
The mean serum selenium concentration and the proportion of subjects who were deficient was not significantly different by HIV status. Less than 8.5% of the mothers were considered deficient at any single time point. The serum selenium concentration decreased significantly over time after delivery in both groups (Table 5).
Copper
Serum copper concentrations (Table 5) and the proportion of mothers who were deficient (Table 6) were not significantly different by HIV status. The results did not change after measures of the acute phase response were controlled for. Serum copper concentrations decreased significantly over time in both groups.
DISCUSSION
In this study we found biochemical evidence that multiple micronutrient deficiencies were common in clinic-attending South African breastfeeding women between 6 and 24 wk after delivery, regardless of HIV status. More than 50% of both HIV-positive and HIV-negative mothers had marginal or deficient concentrations of vitamin B-12, folate, and vitamin E; nearly 25% of the mothers were deficient in iron, zinc or copper. HIV-positive women had lower mean concentrations of serum albumin, prealbumin, folate, retinol, and hemoglobin than did HIV-negative women, and the HIV-positive women had greater rates of deficiency or a marginal status of serum albumin, prealbumin, vitamin B-12, zinc, and anemia. Because serum vitamin B-12 and folate concentrations are not believed to be affected by an acute phase response, the concentrations of these B vitamins seemed to be related directly to HIV infection.
Micronutrient status is difficult to assess in the presence of infection because biochemical indicators of several micronutrients are affected by the acute phase response. Serum ferritin and copper are "positive" acute phase nutrients that increase during an acute phase response, whereas albumin, retinol, and zinc are "negative" responders. Thus, some of the results of the present study may be confounded by increased rates of acute phase response measures in the HIV-positive women. In persons with infections, the use of serum ferritin and copper can lead to an overestimation of adequacy of nutrient status, whereas the use of serum retinol and zinc can lead to an underestimation of adequacy of nutrient status. Accordingly, in the present study we controlled for the acute phase response by using 2 proteins in the analysis: AGP and CRP. AGP appeared to be more closely associated with the differences in the micronutrient concentrations than did CRP; AGP and CRP were not correlated at any time point. In general, the mean concentrations of serum ferritin or copper were not significantly different by group when acute inflammation was controlled for, but at 24 wk, a larger proportion of HIV-negative mothers than HIV-positive mothers had low ferritin concentrations. Although not statistically significant, there was also a trend for a higher proportion of HIV-negative mothers to be copper deficient. The data suggest that the systemic inflammation in the HIV-positive mothers may be masking some deficiency of iron and copper.
Even after measures of the acute phase response were controlled for, the mean serum retinol concentrations of the HIV-positive mothers remained lower than those of the HIV-negative mothers. For zinc, however, the significance of the relation with HIV status increased from P = 0.127 to P = 0.095 after measures of the acute phase response were controlled for. The data suggest that HIV infection may have been the cause of the observed nutrient depletion; however, it is possible that the lower adjusted concentrations did not represent true deficiency.
We are unaware of other studies concerning the protein and micronutrient status of HIV-positive breastfeeding mothers; thus, we are unable to compare our findings with those in a similar population. In a large study of pregnant Zimbabwean women (28), Friis et al (29) found that a large proportion (75%) of pregnant women had marginal folate status or were deficient (folate < 13.5 nmol/L) and that HIV-positive women had lower mean hemoglobin and serum retinol concentrations than did HIV-negative women, even after measures of the acute phase response were controlled for. Compared with HIV-negative women, a larger proportion of HIV-positive Zimbabwean women were folate deficient (ie, folate < 6.7nmol/L; HIV-positive: 16%; HIV-negative: 10%; P < 0.01), vitamin A deficient, (ie, retinol < 0.70 µmol/L; HIV-positive: 34%; HIV-negative: 24%; P < 0.01), or anemic. No significant differences in the mean serum ferritin concentration or in the proportion of subjects with depleted iron stores were observed between groups, similar to the findings of the present study.
The lower mean serum concentrations of albumin, prealbumin, folate, retinol, and hemoglobin in the HIV-positive mothers and the greater proportion of HIV-positive mothers with low serum albumin, vitamin B-12, hemoglobin, and zinc concentrations suggest that the HIV-positive mothers consumed less of these nutrients, had a greater requirement for these nutrients because of malabsorption, or had both of these conditions. This study was not designed to investigate the causes of nutrient deficiencies.
Women with high BMIs (>29) were not protected from low concentrations of vitamins B-12, folate, -tocopherol, or hemoglobin. Apparently, a high BMI does not ensure an improved nutrition status.
Several limitations of our study should be noted. First, to determine the effect of breastfeeding on the nutritional status of HIV-positive mothers, we would have preferred to compare their nutrient status with that of HIV-positive nonbreastfeeding mothers. At the time of the study, however, very few HIV-positive mothers in the study community had chosen not to breastfeed, so we were unable to determine whether the protein and micronutrient differences observed in HIV-positive women were primarily an effect of HIV or of a combination of breastfeeding and the presence of HIV infection. The fact that the prevalence of some deficiencies was greater in the HIV-positive than in the HIV-negative breastfeeding women suggests that HIV status is partially responsible, at least for those nutrients not affected by an acute phase response. During the 17 mo of subject enrollment, 7 nonbreastfeeding mothers were followed. The protein and micronutrient status of the nonbreastfeeding mothers (n = 6 HIV-positive and 1 HIV-negative) was not significantly different from that of their breastfeeding counterparts, but the power to detect differences was small. Second, 6 of the presumed HIV-negative mothers enrolled were subsequently determined to be HIV-positive. These subjects were reassigned to the HIV-positive group. Therefore, the number of HIV-negative mothers followed up until 6 mo after delivery was less than that originally estimated to be necessary to detect a difference between groups of 0.50 SD. On the basis of our actual sample size, we had the ability to detect a difference of 0.52 SD between groups at 6 wk and of 0.58 SD between groups at 24 wk. Finally, because <4% of the mothers had advanced immunosuppression on the basis of CD4+ cell counts <200 cells/µL, we are unable to assess the effect of more advanced disease stage on micronutrient status. Therefore, our results may only be generalizable to the "well" population of HIV-positive breastfeeding women
The findings of this study are important for several reasons. First, we found that a large proportion of clinic-attending breastfeeding women in rural South Africa had multiple nutrient deficiencies, which can affect both their health and that of their infants. Given the known effects of malnutrition, particularly of folate, zinc, copper, iron, and vitamin A, B-6, C, and E deficiencies, on immune function (30-33), these nutrient deficiencies may increase the mothers' susceptibility to diseases commonly found in the region, including HIV. Second, and perhaps more important, deficiencies of protein, zinc, iron, and vitamins A and B-12 and were more common and severe in HIV-positive mothers, although reverse causality was a possibility. All of these nutrients have been associated with progression to advanced stages of HIV disease, reduced CD4+ T cell counts, and increased morbidity and mortality in HIV-positive persons (14, 17, 34-36). These data suggest that a multiple micronutrient supplement is currently indicated for HIV-positive persons, at least until an improved and diverse dietary intake is achievable. Because antiretroviral medication is not yet commonly available to most HIV-positive persons in South Africa, it seems particularly important to address and prevent malnutrition when it may be the only intervention available. Potentially preventable nutritional inadequacies are likely contributing to both greater morbidity and mortality rates in these mothers. Additional research is needed to determine whether these nutrient deficiencies are common in HIV-positive and HIV-negative breastfeeding women in other regions, the appropriate means to correct them, and the implications for nutrition to improve the overall health and well being of mothers and their infant.
ACKNOWLEDGMENTS
We thank the MNS staff for subject recruitment and follow-up; Carina Herbst and Ruth Bland for study management support; the HIV counselors, nurses, clinic assistants, and staff of the Africa Centre for Vertical Transmission Study for study organization and data collection; the mothers, who generously gave of their time and participated in this study; the Hlabisa district clinical nursing staff at the Madwaleni, Nkundusi, and KwaMsane clinics; Arjan van Bentem for insightful computer programming and data management support; and Jan Peerson for help with the statistical analysis.
PCP was responsible for the study concept and design, for obtaining the funding, for the field supervision, for the analysis and interpretation of the results, and for drafting the manuscript. KHB, NCR, CJC, and MLB contributed to the study design, interpretation of the results, and manuscript revisions. The manuscript was reviewed and approved by all authors. This article was written on behalf of the Child Health Group of the Africa Centre for Health and Population studies: HM Coovadia, RM Bland, A Coutsoudis, and ML Newell. None of the authors had a conflict of interest.
REFERENCES