508postpartum-vas.pdf

TECHNICAL BRIEF Postpartum Vitamin A Supplementation: Evaluating the Evidence for Action Amy L. Rice, PhD BACKGROUND GLOBAL PRE...

TECHNICAL BRIEF Postpartum Vitamin A Supplementation: Evaluating the Evidence for Action Amy L. Rice, PhD BACKGROUND GLOBAL PREVALENCE OF VITAMIN A Scientific and policy-making groups have issued somewhat DEFICIENCY conflicting statements about the benefits of postpartum vitamin Vitamin A deficiency remains a widespread public health A supplementation programs over the past few years (de Benoist problem among women and children. Over 20 percent of all et al, 2001; IVACG, 2002; Ross, 2002; Sommer, 2005). This preschool age children (~130 million) and nearly six percent has led some public health professionals to question the value of of all pregnant women (~7 million) suffer from vitamin A providing vitamin A supplements to women immediately after deficiency and its adverse health consequences (West, 2002; delivery as a strategy for improving maternal and child health. Rice et al, 2004). The majority of vitamin A deficient women and children currently live in South Asia or Sub-Saharan PURPOSE OF BRIEF Africa, in the same countries where vitamin A supplementation programs for children over six months of age have helped This brief reviews the scientific rationale for postpartum reduce high child mortality rates. But while more than 60 supplementation and examines the available evidence on countries now have supplementation programs for preschool whether or not this intervention can improve the nutritional age children, fewer have launched large-scale program status, breast milk quality, and functional health outcomes initiatives to address the problem among younger infants or in in mothers and children. The results suggest that program women of reproductive age. planners may need to balance the evidence that vitamin A supplementation improves the health status in women and children against programmatic costs. Thus far, there is ADVERSE EFFECTS OF VITAMIN A evidence that postpartum vitamin A supplementation does DEFICIENCY not completely correct deficiency in severely deficient women and children, little evidence that it improves functional health HIV-negative Populations outcomes, and recent evidence for potential risks and benefits unique to HIV-positive populations. Vitamin A deficiency (often defined using low serum retinol concentrations as an indicator of deficiency) is known to increase the risk of blindness and death among children in METHODOLOGY HIV-negative populations (Sommer and West, 1996) and to Published studies evaluating the impact of postpartum vitamin adversely affect maternal health. Night-blindness, a well-known A supplementation were reviewed. These included trials in sign of vitamin A deficiency, has been associated with higher which women were given vitamin A supplements (200,000 rates of morbidity among pregnant Nepali women (Christian IU–400,000 IU orally) within eight weeks of delivery. Several et al, 1998) as well as higher rates of mortality during the early of these trials also included vitamin A supplementation for postpartum period (Christian et al, 2000). the infants. Additional background documents about vitamin A deficiency, breastfeeding, and the evolution of vitamin HIV-positive Populations A supplementation guidelines were reviewed and selected The interpretation of low serum retinol concentrations in HIV- publications about these issues are cited. positive individuals is complicated by the body’s response to infections and trauma. During the acute phase of an infection, For more information about the A2Z project: Call: 202-884-8970 Email: [email protected] Visit: www.a2zproject.org 2 Postpartum Vitamin A Supplementation: Evaluating the Evidence for Action retinol-binding protein, which is part of the chemical complex supplementation, and dietary vitamin A intake may affect that ferries retinol through the bloodstream on its way from women’s and children’s vitamin A status and potentially lead to the liver to other tissues, is called into action elsewhere and better functional health outcomes, including improved growth temporarily disappears from the circulation. As a result, serum and immune function, reduced morbidity, and improved retinol concentrations decrease. More severe infections cause survival rates. more dramatic declines, but serum retinol concentrations Studies among populations presumed to be HIV negative generally rebound after the acute event has passed. However, clearly show that maternal dietary intake is an important in HIV-positive individuals with chronically activated immune determinant of vitamin A status and breast milk vitamin A systems, low serum retinol concentrations may indicate a concentrations (Haskell and Brown, 1999; Newman, 1993). more active viral infection, true vitamin A deficiency, or a Intervention trials have demonstrated that it is possible to combination of both. Reliable methods for correctly identifying improve maternal vitamin A status and/or breast milk vitamin the underlying condition(s) have yet to be developed. A concentrations by providing low-dose vitamin A or beta- Along similar lines, the interpretation of population level carotene supplements to women before, during, and after studies that assess vitamin A status based on serum retinol pregnancy (Baylin et al, 2005; Tanumihardjo et al, 1996; concentrations is more complicated in the context of HIV. For Yamini et al, 2001), through food fortification programs example, although observational studies of pregnant women (Arroyave et al, 1974; Muhilal et al, 1988) and in the form showed that low serum retinol concentrations were associated of food-based interventions (de Pee et al, 1998; de Pee et al, with higher rates of maternal-to-child transmission of HIV, the 1995; de Pee et al, 1998; Villard and Bates, 1987). However, results of the subsequent supplementation trials which provided the effectiveness of these approaches varies and appears to be varying amounts of vitamin A and/or beta-carotene to women influenced by women’s underlying vitamin A status. Women before, during, and/or after pregnancy have been disappointing. who are more severely deficient tend to be more responsive to Trials conducted in Malawi, South Africa, and Zimbabwe intervention. found no protective effect, while another in Tanzania observed Postpartum vitamin A supplementation (i.e. giving women a an increased risk of transmission (Wiysonge et al, 2005). Did large dose of vitamin A that is 200,000 IU or more within the vitamin A fail to reduce HIV transmission rates because: the first six weeks after delivery) is another strategy for improving women were not vitamin A deficient; they did not respond to maternal and infant vitamin A status and health outcomes. supplementation; the risk of maternal-to-child transmission is Postpartum supplementation is designed to improve women’s not increased by inadequate maternal vitamin A status; or by vitamin A status and to increase the vitamin A content of some combination of these and/or other factors? The reason breast milk (Stolzfus and Humphrey, 2002). This is meant remains unclear. to protect the mother’s vitamin A reserves, while addressing On the other hand, the trial in Zimbabwe also found that low one of the fundamental reasons that children become vitamin maternal serum retinol concentrations were associated with A deficient—low dietary vitamin A intake from breast milk a ten-fold greater risk of sexual acquisition of HIV among (Miller et al, 2002). HIV-negative women, and preliminary data suggest that Breast milk represents the single most important source of supplementation of vitamin A deficient, HIV-negative women vitamin A for very young infants. All infants are naturally born may lower their own risk of acquiring HIV (Humphrey et with low body stores of vitamin A and depend upon vitamin al, 2006). These examples illustrate that more work is needed A-rich colostrum and breast milk to meet their physiological to identify the unique health risks associated with vitamin A need for vitamin A and other nutrients needed for proper deficiency in the context of HIV. growth and development. For well-nourished women and infants, nearly 60 times as much vitamin A will be transferred SCIENTIFIC RATIONALE FOR from mother to infant during breastfeeding as compared to POSTPARTUM SUPPLEMENTATION pregnancy (Stolzfus, 1994). AMONG HIV-NEGATIVE WOMEN Unfortunately, mothers with compromised nutritional status produce breast milk with levels of vitamin A that are too low Several proven strategies exist for improving the vitamin to meet their infants’ immediate physiological needs and for A status of HIV-negative women and their breastfeeding infants to build liver stores of vitamin A for the future. Without infants. Figure 1 provides a conceptual framework that shows adequate stores, infants are at greater risk of developing vitamin how maternal vitamin A supplementation, infant vitamin A A deficiency and dying during their first few years of life. For more information about the A2Z project: Call: 202-884-8970 Email: [email protected] Visit: www.a2zproject.org  Postpartum Vitamin A Supplementation: Evaluating the Evidence for Action 3 Figure 1. Pathway to improved vitamin A status and functional health outcomes in mothers, infants, and children VA for pre-school children twice a year (100,000 IU for 6–11 mo and 200,000 IU for 12–59 mo) Infant dietary VA intake (complementary foods after 6 mo of age) Child vitamin A status Functional health (6–59 mo) outcomes in children (6–59 mo old) Low dose VA before 6 mo* (50,000 IU x 3) Infant vitamin A status Functional health (< 6 mo) outcomes in infants (< 6 mo old) Newborn vitamin A dosing* (50,000 IU) Breast milk vitamin A content Prenatal VA supplementation* Functional health Maternal vitamin A status (daily RDA) outcomes in women Maternal dietary VA intake from food Postpartum vitamin A supplementation (200,000 IU or 400,000 IU) * Under consideration in current revision of WHO Guidelines For more information about the A2Z project: Call: 202-884-8970 Email: [email protected] Visit: www.a2zproject.org 4 Postpartum Vitamin A Supplementation: Evaluating the Evidence for Action EVIDENCE REGARDING POSTPARTUM underway to investigate this last issue (J Humphrey, personal communication). VITAMIN A SUPPLEMENTATION Among HIV-positive women in Zimbabwe, postpartum Table 1 (see pages 6–9) summarizes the results of 17 different vitamin A supplementation had no overall effect on publications from trials designed to evaluate the impact of maternal-to-child transmission of HIV, on child mortality postpartum vitamin A supplementation. The studies are rates between birth and 24 months of age, or on maternal presented by the dose of vitamin A the women received morbidity or mortality rates. However, the effect of vitamin A (200,000 IU, 300,000 IU, or 400,000 IU) and then by the supplementation differed for certain subgroups. Infants infected date of publication, with earlier studies listed first. The only with HIV during delivery had prolonged survival if they study that compared two different doses of vitamin A (200,000 themselves received vitamin A at birth (irrespective of whether IU vs. 400,000 IU), rather than vitamin A versus a non-dosed their mothers received vitamin A), while infants who were HIV group, is listed at the end of the table. negative at six weeks of age (but potentially infected after this The trials from Bangladesh, Ghana, India, Indonesia, Peru, and point in time) experienced higher mortality at 24 months of Thailand were conducted among women whose HIV status was age if either they and/or their mothers had received vitamin A. unknown, but is presumed to be HIV negative. Only the recent Among women, postpartum supplementation had no overall study in Zimbabwe was designed to address questions related to effect on the rate of hospitalization or sick clinic visits, but vitamin A and HIV and tested the HIV status of women and supplementation was associated with reduced clinic visits for children throughout the follow-up period of the trial. malaria, cracked and bleeding nipples, pelvic inflammatory disease, and vaginal infections. The majority of these studies were designed to examine the relationship between maternal and/or infant vitamin A In summary, the available evidence suggests that postpartum supplementation and biochemical outcomes (serum retinol or supplementation results in modest, short-term improvements liver stores of vitamin A in women and/or children, or breast in maternal and child vitamin A status (measured as higher milk retinol). A handful of trials included other measures serum retinol concentrations, higher liver vitamin A stores, of health or nutritional status, including infant growth and or higher breast milk vitamin A concentrations) and only immune function, and infant or maternal morbidity. The only minor improvements, if any, in functional health outcomes trial conducted in an HIV-endemic population also assessed for mothers and their children. These potential benefits are the incidence of maternal-to-child transmission of HIV, the summarized in Table 2 (see page 9). acquisition of HIV among women, as well as maternal and child morbidity and mortality rates. CONFUSION AROUND THE VALUE OF Nearly all of the trials conducted among the presumably HIV- POSTPARTUM SUPPLEMENTATION AND negative populations in Bangladesh, Ghana, India, Indonesia, WHAT TO DO Peru, and Thailand found short-term beneficial effects on biochemical outcomes (i.e. serum retinol, breast milk retinol, or Two different factors deserve consideration. liver stores of vitamin A), but only one documented adequate There is no universally recognized benchmark for vitamin A status at the end of the study period among all of measuring “success” the women who received vitamin A (Tchum et al, 2006). None of the studies that assessed infant vitamin A status found that The most recent recommendations were not widely vitamin A deficiency was completely eliminated by dosing disseminated their mothers. In most cases vitamin A supplementation had either no overall effect on the functional health outcomes No Universally Recognized Definition of measured (infant morbidity or infant growth) or a slightly Success positive effect (infant morbidity). In Zimbabwe, where the There is no universally recognized indicator for measuring the vitamin A status of women was generally adequate at delivery, “success” of this intervention, at least in terms of biological postpartum supplementation of HIV-negative women had efficacy. Opinions seem to differ about what this intervention no overall effect on infant mortality rates, on women’s risk of should accomplish when it is being implemented properly acquiring HIV, or on mortality rates among women. However, and new mothers are receiving vitamin A supplements in a low maternal serum retinol concentrations were associated timely fashion after delivery. For some, improving biochemical with a ten-fold greater risk of sexual acquisition of HIV, and indicators of vitamin A status without changing functional preliminary data from a small subset of women suggest that health outcomes among women and children is of questionable supplementing HIV-negative women with low serum retinol value, while others feel that improving maternal vitamin A concentrations may lower this risk. Additional work is currently status, increasing breast milk vitamin A concentrations, or For more information about the A2Z project: Call: 202-884-8970 Email: [email protected] Visit: www.a2zproject.org  Postpartum Vitamin A Supplementation: Evaluating the Evidence for Action 5 increasing vitamin A stores among infants is sufficient alone, given the low cost of the intervention and the available justification for launching a postpartum vitamin A platforms for its delivery, a strong argument can be made for its supplementation program. inclusion as one component in a comprehensive micronutrient program in vitamin A deficient populations (see Figure 1 on Thus, some of the confusion about the value of postpartum page 3). dosing programs may simply be a lack of clarity about how “success” should be defined. Improving vitamin A status, breast milk vitamin A concentrations, and functional health outcomes CONCLUSIONS AND may all be goals of maternal and child health programs. Policy RECOMMENDATIONS makers and program managers need be sure they are clear about what they want to accomplish and what postpartum Country-level decision-makers always need to balance the costs, supplementation programs may be able to deliver. benefits, and potential risks (if any exist) of an intervention against the available resources and their program goals. These decision-makers rely heavily on policy recommendations Most Recent Recommendations Were Not from WHO for assistance in guiding these decisions. The Widely Disseminated new WHO policy recommendations about vitamin A The last set of vitamin A supplementation guidelines jointly supplementation will need to be clear about what postpartum published by the World Health Organization (WHO), supplementation can accomplish. Country-level decision UNICEF, and the International Vitamin A Consultative Group makers and program planners should consider their own (IVACG) appeared in 1997 and recommended a single 200,000 context and determine whether implementing a postpartum IU dose of vitamin A for postpartum women soon after vitamin A supplementation program can help them reach their delivery (WHO/UNICEF/IVACG Task Force, 1997). Since particular goals in a cost-effective manner. that time, many of the countries that initiated supplementation At the time this brief was finalized, December 2006, WHO programs for postpartum women based their policies on these had initiated a comprehensive review of its recommendations guidelines. for vitamin A supplementation. Therefore, it would be In the meantime, updated guidelines have been formulated and sensible for country decision-makers to await the updated agreed on by experts convened during an informal technical recommendations before making changes in their own policies consultation organized by the WHO in 2000 and an expert and programs. panel meeting convened by IVACG in 2001. The updated guidelines recommend a 400,000 IU dose of vitamin A for postpartum women (given in the form of 2 doses of 200,000 IU at least 24 hours apart—see Table 3 on page 10) based on the findings of previous studies that documented limited benefits of dosing women with 200,000 IU or 300,000 IU after delivery. These recommendations were presented to the public health nutrition community at the XX IVACG meeting in Vietnam in 2001 and have been published in policy statements (IVACG, 2002) and in the scientific literature (de Benoist et al, 2001; Ross, 2002). However, many health professionals may not be aware of these newer recommendations because they have not appeared in the same format or been as widely disseminated to an international audience as the joint guidelines previously published by WHO, UNICEF, and IVACG. WHO is currently in the process of revising these guidelines. The scientific rationale for postpartum supplementation remains unchanged—that is, finding a safe and effective way to improve the vitamin A status of women and children. The best available evidence suggests that this intervention provides short-term, but measurable, benefits for many women and their infants. Although the evidence for improvements in functional health outcomes is not strong for postpartum supplements For more information about the A2Z project: Call: 202-884-8970 Email: [email protected] Visit: www.a2zproject.org 6 Postpartum Vitamin A Supplementation: Evaluating the Evidence for Action Table 1. Impact of postpartum vitamin A supplementation on vitamin A status and functional health outcomes in women, infants, and children Comparisons between Vitamin A Total number of study Follow- Primary supplementation (vitamin A or beta- dose, study site, participants and up author and year outcomes carotene) and control (placebo or treatment groups1 period nothing) groups2 200,000 IU Bangladesh, Roy 50 mother-infant pairs 9 mo Biochemical Maternal serum retinol and breast et al, 1997 indicators milk retinol higher through 3 and 6 mo 200,000 IU vitamin A or (women) postpartum, respectively Nothing given to women (within 24 hr after delivery) Morbidity Lower incidence of infant morbidity and (infants and pedal edema in women women) India, Bhaskaram 100 mother-infant pairs 3 mo Biochemical Breast milk retinol higher through 45 days and Balakrishna, indicators postpartum 200,000 IU vitamin A or 1998 Placebo given to women Immune No effect on infant serum retinol (within 24 hr after delivery) response No effect on infant immune response to polio to polio Plus oral polio vaccine given vaccine immunization to all infants between 24 and (infants) 72 hours of birth Bangladesh, Rice 222 mother-infant pairs 9 mo Biochemical Maternal liver stores of VA and breast milk et al, 1999 indicators retinol improved among women receiving 200,000 IU VA given to women at 1-3 wk No functional VA (at 3 mo) and women receiving beta- postpartum, then daily health carotene (Semba et al, 1995) placebos until 9 mo outcomes Infant liver stores of VA improved among the postpartum or VA treatment group (Christian et al, 2001) Placebos from 1-3 wk to 9 mo postpartum or Beta-carotene (7.8 mg) from 1-3 wk to 9 mo postpartum Ghana, 9424 mother-infant pairs 12 mo Biochemical Breast milk retinol increased at 2 mo India, Peru, indicators 200,000 IU vitamin A to Infant serum retinol concentrations and liver Anonymous, women (18-42 days after Side effects stores of VA increased at 6 mo of age 1998 and Bahl et delivery) and 25,000 IU within 48 hrs of al, 2002 Slight increase in the rate of bulging to infants with the first administration fontanelle (but all groups