Vitamin A supplementation had no effect on mortality of all women combined or for either HIV-status group (Fig. 3).
At 12 and 24 months, respectively, mortality rates among HIV-negative mothers were 0.27% (0.18-0.40, here and elsewhere ranges give 95% confidence limits; 24 deaths) and 0.38% (0.24.59; 3 deaths) by K-M survival analysis, and 2.7 per 1000 person-years (p-y) and 2.3 per 1000 p-y by person-time analysis. During the same periods, mortality rates for HIV-positive women were 3.6% (3.0-4.2; 149 deaths) and 7.7% (6.7-8.7; 88 deaths) by K-M survival analysis, and 36.5 per 1000 p-y and 38.3 per 1000 p-y by person-time analysis. Compared with HIV-negative women, HIV-positive women were 14.2 (8.9-22.6) and 54.1 (13.3-220) times more likely to die during the first and second postpartum years, respectively, after adjusting for age. During the first and second years, respectively, HIV-positive women made up 88% (143/163) and 98% (88/90) of all deaths (P = 0.007, χ 2, df = 1 for the difference in the percentage between years), and the PAF of death caused by HIV was 82% and 96%. Among HIV-negative women, the conditional probability of death during the second 12 months, given survival to 12 months postpartum, was 100 × [(0.38 − 0.27)/(100 − 0.27)] = 0.11% (ie, less than half that of the first year). Conversely, among HIV-positive women, the corresponding probability was 100 × [(7.7 − 3.6)/(100 − 3.6)] = 4.3% (i.e., slightly higher in the second compared with the first year). In fully adjusted Cox analyses applied to data in 5-year age strata, the increased hazard of death associated with HIV was slightly lower among teenagers (12.9), rising to 17.1 among women aged 25 to 30 years, and then declining to 9.1 among women older than 40 years. However the 95% confidence intervals (CIs) of the estimates for each age group were wide (primarily due to the very small number of deaths among HIV-negative women), and there was no statistically significant effect of age (P = 0.06 and 0.08 for the linear and quadratic effects, respectively) (Fig. 4).
Risk Factors for Mortality
Among HIV-negative women, age, widowhood, and husband's education were independently predictive of mortality (data not presented). The adjusted hazard ratios of mortality associated with vitamin A supplementation at 6, 12, and 24 months were 1.4 (0.4-5.0), 1.11 (0.81, 1.51, and 1.0 (0.4-2.3), respectively.
Among HIV-positive women, age, parity, CD4 and Hb concentrations, MUAC, marital status, maternal education, and the survival of the previous child were independently predictive of mortality (Table 3). Parity and loss of the previous child entered jointly into the model: primiparous mothers and those who had lost their previous child were at higher risk than those with a living previous child. The adjusted hazard ratios of mortality associated with vitamin A treatment at 6, 12, and 24 months were 1.1 (0.7-1.7), 1.1 (0.8-1.6), and 1.0 (0.8-1.3), respectively.
Cause of Death
There was no significant effect of the vitamin A supplementation on any cause-specific risk of death among HIV-positive or HIV-negative women (data not presented).
Cause-specific mortality during the postpartum year was significantly higher among HIV-positive compared with HIV-negative women for all causes except hypertension and accident (Table 4). The risk ratios were highest for causes associated with HIV infection, but HIV-positive women were also significantly more likely to die due to obstetric causes: 88% (7/8) and 90% (9/10) of deaths associated with direct maternal causes within 42 days and 1 year postpartum, respectively, were among HIV-positive women.
Sick Clinic Visits and Hospitalizations Among HIV-Positive Women
HIV-positive women made a total of 8322 sick clinic visits during the postpartum year (Table 5). After correction for overdispersion, vitamin A supplementation had no overall effect on sick clinic visits (IRR, 0.95 [0.86-1.04]). However, vitamin A supplementation was associated with significant reductions in cause-specific clinic visits for malaria, vaginal infection, pelvic inflammatory disease, and cracked or bleeding nipples. Women with baseline CD4 count of less than 200 cells × 106/L had 1.8 (1.6-2.0) times more sick clinic visits compared with infected women with higher CD4 counts, but adjustment for CD4 did not alter the vitamin A effect (data not presented).
During the postpartum year, 259 HIV-positive women were hospitalized at least once (126 and 133 in the vitamin A and placebo groups, respectively), for 339 hospitalizations (165 [vitamin A] and 174 [placebo]), and 880 total hospital days (390 [vitamin A] and 490 [placebo]). Only 28 HIV-positive women in the vitamin A group and 34 in the placebo group had a hospitalization of more than 1 day's duration. After correction for overdispersion, there was no effect of vitamin A on the proportion of women ever hospitalized (IRR, 0.98 [0.76-1.26]), the number of hospitalizations (0.96 [0.66-1.40]), or the risk of having a hospitalization of more than 1 day in duration (0.84 [0.45-1.58]). HIV-positive women with CD4 of less than 200 cells were 3.0 (2.28-4.00) times more likely to be hospitalized and had 3.6 (2.4-5.3) times more hospitalizations than HIV-positive women with higher CD4 counts, but adjusting for CD4 had no effect on hazard ratios associated with vitamin A treatment.
A single 400,000-IU dose of vitamin A given during the immediate postpartum period had no effect on mortality among HIV-negative or HIV-positive women. Lack of effect among HIV-negative women might be because few were apparently vitamin A deficient: the distribution of serum retinol concentration at 6 weeks, even among those who received placebo, lay midway between those of similarly-aged white and black women in the United States National Health and Nutrition Examination Survey III.31
Among HIV-positive women, vitamin A deficiency was apparently common: 30% had a serum retinol concentration of less than 1.05 μmol/L at 6 weeks. Despite this, serum retinol increased in response to vitamin A supplementation only among the 15% of HIV-positive women who had CD4 count of less than 200 cells × 106/L. Among all other HIV-positive women, 30% remained with a concentration of less than 1.05 μmol/L regardless of their baseline CD4 count or whether they had received vitamin A, and the entire distribution was shifted substantially lower than that of HIV-negative women 6 weeks after supplementation. In short, the primary factor predicting serum retinol concentration 6 weeks after dosing was HIV status, not receipt of vitamin A or placebo.
The programmatic implications of this depend on whether these low serum retinol concentrations reflect deficiency that might have responded to higher or more continuously administered doses of vitamin A or an HIV-induced metabolic alteration intractable to even more aggressive supplementation. We believe that our data are more consistent with the second possibility. Retinol circulates in complex with 2 protein carriers, retinol-binding protein (RBP) and transthyretin (TTR),32 which are both "negative acute-phase proteins" (APPs). During the acute-phase response (APR) to infection, RBP, TTR, and retinol concentrations decline.33 However, among asymptomatic, HIV-positive men in the United Kingdom, plasma concentrations of RBP and TTR were not reduced compared with HIV-negative healthy controls, leading the authors to conclude that asymptomatic HIV elicits an "atypical" APR in which positive APP concentrations increase, but negative APP concentrations do not decline.34 Citing this work, authors of a meta-analysis on the effect of subclinical infection on serum retinol concentration suggested that HIV is unlikely to affect retinol concentration in population-based surveys, so that "…even where [HIV] infection is endemic, it can probably be overlooked, and the sample size needed to establish the prevalence of vitamin A deficiency can be calculated with the same disease and environmental factors as for other population groups."35 Because serum retinol among HIV-negative women responded to supplementation, whereas it did not among asymptomatic HIV-positive women in this population, we suggest the possibility that the capacity to maintain negative APP concentrations during asymptomatic HIV may be sensitive to dietary protein intake. Estimated per capita dietary protein intake in Zimbabwe is lower (54 g/d per capita, 10% of calories)36 compared with the HIV-positive subjects in the UK study (122 g/d, 20% of calories), and the percentage of protein from animal sources (i.e., protein quality) is 12% in Zimbabwe36 compared with 70% in western diets.37 Indeed, among asymptomatic HIV-positive men in Cote d'Ivoire, RBP and TTR concentrations were 20% significantly lower, whereas concentrations of 2 positive APPs (C-reactive protein and α-1 glycoprotein) were not higher compared with HIV-negative healthy controls.38 Our observation that supplementation increased, but did not normalize, serum retinol among HIV-positive women with CD4 count of less than 200 cells could be explained if dietary vitamin A was limiting relative to circulating RBP and TTR concentrations among these women who were more poorly nourished based on several measures of nutritional status (i.e., MUAC, body mass index, and Hb).
Simultaneous measurement of APPs during serum retinol surveys is recommended to adjust for the effects of acute infection.39-41 Further research is required to better understand vitamin A metabolism and the APR during asymptomatic HIV in populations ingesting marginal quantity and quality of dietary protein. Until this is better understood, we suggest that simultaneous HIV testing would also be informative in interpreting serum retinol in these populations and that sample sizes for surveys should be sufficient to estimate serum retinol for HIV status strata.
Vitamin A supplementation was associated with a decrease in sick clinic visits for malaria, cracked and bleeding nipples, pelvic inflammatory disease, and vaginal infection. In Nepal, weekly vitamin A supplementation reduced urinary and genital tract infections during late pregnancy,42 and in Papua New Guinea, periodic high-dose vitamin A supplementation reduced malarial morbidity among children.43
Our data underscore the enormous toll that HIV is taking on maternal health in Zimbabwe: among the women who were HIV positive at delivery, 7.5% died (38/1000 p-y) during the following 2 years. This is similar to rates observed in 2 studies of Rwandan antenatal clinic attendees (7% during 2 years29 and 50.1/1000 p-y).44 Our data corroborate the recently highlighted substantial mortality effect of coinfection with tuberculosis and HIV among women:9 in ZVITAMBO, tuberculosis was the most common cause of death, and nearly all tuberculosis-associated deaths were among HIV-positive women.
The risk of death among infected compared with uninfected women did not vary significantly with age. In contrast, the mortality rate ratio of infected to uninfected people in several large African population-based cohorts was 12 in teenagers and in people older than 30 years (similar to the excess risk we observed among all age groups of ZVITAMBO women) but was much higher (20-30) among people aged 25 to 30 years.45 The "flatter" relationship of mortality rate ratio to age among women in ZVITAMBO might be because they were immediate postpartum and so were selectively healthier than similarly aged HIV-positive people in cohorts representative of general populations. This relative health would be greatest among the 20- to 30-year age group because teenagers in both the population-based cohorts and ZVITAMBO would have been infected for a similarly short period, and underlying mortality rates among older HIV-negative adults are higher in Uganda, Tanzania, and Malawi compared with Zimbabwe.46
In another article, we report that, in ZVITAMBO, maternal vitamin A supplementation had no beneficial effects on infant HIV infection or overall mortality and that it had an adverse effect among a subgroup of infants: among HIV-exposed infants who escaped becoming infected by their mothers during pregnancy and delivery, maternal vitamin A supplementation increased their risk of death by age 2 years.18 The explanation offered was that priming with vitamin A may have increased viral load among babies who subsequently became infected during breast-feeding. Taken together with the lack of benefit on maternal mortality and limited benefit on maternal morbidity reported in the current article, this finding raises concern about universal postpartum maternal vitamin A supplementation in areas of HIV endemicity. However, at least 2 issues require further research to fully inform policy decisions about postpartum maternal supplementation in HIV endemic areas. First, if ongoing studies confirm the beneficial effect of frequent supplementation on maternal mortality which was observed in Nepal,4 then the effect of maternal vitamin A supplementation (perhaps given as smaller, more frequent doses) on maternal mortality should be reexamined in an HIV-endemic population that has poorer vitamin A status than the women who participated in ZVITAMBO. Second, among a small subsample of HIV-negative women in the ZVITAMBO trial for whom serum retinol was measured, those with low concentrations (<0.7 μmol/L) were 10.4 (95% CI, 3.0-36.3) times more likely to acquire HIV during the postpartum period, and supplementation of these apparently deficient women tended to reduce HIV incidence, although not significantly so (perhaps due to limited statistical power).47 Should this protection be confirmed and reach statistical significance, this benefit may override other issues related to postpartum maternal supplementation in populations in which vitamin A deficiency is widespread and HIV incidence is high.48
1. West KP Jr. Extent of vitamin A deficiency among preschool children and women of reproductive age. J Nutr
2. Christian P, West KP Jr, Khatry SK, et al. Night blindness of pregnancy in rural Nepal-nutritional and health risks. Int J Epidemiol
3. Christian P, West KP Jr, Khatry SK, et al. Night blindness during pregnancy and subsequent mortality among women in Nepal: effects of vitamin A and β-carotene supplementation. Am J Epidemiol
4. West KP Jr, Katz J, Khatry SK, et al. Double blind, cluster randomised trial of low dose supplementation with vitamin A or β-carotene on mortality related to pregnancy in Nepal. BMJ
5. Semba RD, Caiaffa WT, Graham NMH, et al. Vitamin A deficiency and wasting as predictors of mortality in human immunodeficiency virus-infected injection drug users. J Infect Dis
6. Semba RD, Graham NMH, Caiaffa WT, et al. Increased mortality associated with vitamin A deficiency during human immunodeficiency virus type 1 infection. Arch Intern Med
7. WHO World Health Report. Geneva; 1999.
8. Ahmed Y, Mwaba P, Chintu C, et al. A study of maternal mortality at the university teaching hospital, Lusaka, Zambia: the emergence of tuberculosis a major non-obstetric cause of maternal death. Int J Tuberc Lung Dis
9. Khan M, Pillay T, Moodley JM, et al. Maternal mortality associated with tuberculosis-HIV-1 co-infection in Durban, South Africa. AIDS
10. McIntryre J. Mothers infected with HIV. Br Med Bull
11. Bicego G, Boerma JT, Ronsmans C. The effects of AIDS on maternal mortality in Malawi and Zimbabwe. AIDS
12. WHO/UNICEF/IVACG Task Force. Vitamin A Supplements: A Guide to Their Use in the Treatment and Prevention of Vitamin A Deficiency and Xeropthalmia
. Geneva, Switzerland: World Health Organization; 1998.
13. Rice AL, Stoltzfus RJ, de Francisco A, et al. Maternal vitamin A or β-carotene supplementation in lactating Bangladeshi women benefits mothers and infants but does not prevent subclinical deficiency. J Nutr
14. Stoltzfus RJ, Hakimi M, Miller KW, et al. High dose vitamin A supplementation of breast-feeding Indonesian mothers: effects on the vitamin A status of mother and infant. J Nutr
15. de Benoist B, Martines J, Goodman T. Vitamin A supplementation and the control of vitamin A deficiency: conclusions. Food Nutr Bull
16. Ross DA. Recommendations for vitamin A supplementation. J Nutr
17. Ministry of Health and Child Welfare. Report of the National Survey on vitamin A, Iron, and Iodine; October 11-31, 1999. Harare, Zimbabwe; 1999.
18. Humphrey JH, Iliff PJ, Marinda E, et al. Effects of single large dose of vitamin A given during postpartum period to HIV-positive women and their infants, on child HIV infection, HIV-free survival, and mortality. J Infect Dis
19. Iliff PJ, Piwoz EG, Tavengwa NV, et al. Early exclusive breastfeeding reduces the risk of postnatal HIV-1 transmission and increases HIV-free survival. AIDS
20. Malaba LC, Iliff PJ, Nathoo KJ, et al. Effect of postpartum maternal or neonatal vitamin A supplementation on infant mortality among infants born to HIV-negative mothers in Zimbabwe. Am J Clin Nutr
21. Miller MF, Stoltzfus RJ, Mbuya NV, et al. Total body iron in HIV-positive and HIV-negative Zimbabwean newborns strongly predicts anemia throughout infancy and is predicted by maternal hemoglobin concentration. J Nutr
22. Gibson RS. Nutrition Assessment, A Laboratory Manual
. New York: Oxford University Press; 1993:68.
23. Piwoz EG, Iliff PJ, Tavengwa NV, et al. An education and counselling program for preventing breastfeeding-associated HIV transmission in Zimbabwe: design and impact on maternal knowledge and behaviour. J Nutr
24. Gavin L, Tavengwa N, Iliff P, et al. The development of an intervention to inform women about infant feeding in the context of HIV. Poster presentation at the second conference on Global Strategies for the Prevention of HIV Transmission from Mothers to Infants; Montreal, Canada; September 1-5, 1999.
25. Ghana VAST Study Team. Vitamin A supplementation in northern Ghana: effects on clinic attendances, hospital admissions, and child mortality. Lancet
26. World Health Organization. The measurement of overall and cause-specific mortality in infants and children: report of a Joint WHO/UNICEF consultation; December 15-17; Geneva, Switzerland; 1992.
27. WHO. International Statistical Classification of Disease and Related Health Problems: Tenth Revision
. Geneva: World Health Organization; 1992.
28. Rockhill B, Newman B, Weinberg C. Use and misuse of population attributable fractions. Am J Public Health
29. Lindan CP, Allen S, Serufilira A, et al. Predictors of mortality among HIV-infected women in Kigali, Rwanda. Ann Intern Med
30. Goff JP, Kimura K, Horst RL. Effect of mastectomy on milk fever, energy, and vitamin A, E, and β-carotene status at parturition. J Dairy Sci
31. Pilch SM. Analysis of vitamin A data from the health and nutrition examination surveys. J Nutr
32. Fleck A. Clinical and nutritional aspects of changes in acute-phase proteins during inflammation. Proc Nutr Soc
33. Olson JA. Serum levels of vitamin A and carotenoids as reflectors of nutritional status. J Natl Cancer Inst
34. Jahoor F, Gazzard B, Phillips G, et al. The acute phase protein response to human immunodeficiency virus infection in human subjects. Am J Physiol
35. Thurnham DI, McCabe GP, Northrop-Clewes CA, et al. Effects of subclinical infection on plasma retinol concentrations and assessment of prevalence of vitamin A deficiency: meta-analysis. Lancet
36. Tagwireyi J, Greiner T. Nutrition in Zimbabwe: An Update
. Washington, DC: The World Bank; 1994.
37. USDA, Nationwide food consumption survey. Nutrient intakes: individuals in 48 states, year 1977-78. Report no. 1-2. Hyattsville, MD: Consumer Nutrition Division, Human Nutrition Information Service; 1984 [439 pp].
38. Monnet D, Kacou E, Damet G, et al. Marquers de la reaction inflammatoire et marqueurs nutritionnels au cours de I'infection a VIH. Ann Biol Clin
39. Friis H, Gome E, Kastel P, et al. HIV and other predictors of serum β-carotene and retinol in pregnancy: a cross-sectional study in Zimbabwe. Am J Clin Nutr
40. Filteau SM, Morris SS, Abbott RA, et al. Influence of morbidity on serum retinol of children in a community-based study in northern Ghana. Am J Clin Nutr
41. Christian P, Schulze K, Stoltzfus RJ, et al. Hyporetinolemia, illness symptoms, and acute phase protein response in pregnant women with and without night blindness. Am J Clin Nutr
42. Christian P, West KP Jr, Khatry SK. Vitamin A or β-carotene supplementation reduces symptoms of illness in pregnant and lactating Nepali Women. J Nutr
43. Shankar AH, Genton B, Semba RD, et al. Effect of vitamin A supplementation on morbidity due to Plasmodium falciparum
in young children in Papua New Guinea: a randomized trial. Lancet
44. Leroy V, Msellati P, Lepage P, et al. Four years of natural history of HIV-1 infection in African women: a prospective cohort study in Kigali (Rwanda), 1988-1993. J Acquir Immune Defic Syndr
45. Porter K, Zaba B. The empirical evidence for the impact of HIV on adult mortality in the developing world: data from serological studies. AIDS
46. Bicego G. Estimating adult mortality rates in the context of the AIDS epidemic in sub-Saharan Africa: analysis of DHS sibling histories. Health Transition Review
47. Humphrey JH, Iliff PJ, Hargrove JW, et al. HIV incidence among post-natal women in Zimbabwe: impact of post partum vitamin A supplementation and associated risk factors. AIDS
Keywords:© 2006 Lippincott Williams & Wilkins, Inc.
HIV; vitamin A; mortality; morbidity; mothers; tuberculosis; serum retinol; Zimbabwe