1. Objectives

To assess the safety and efficacy of micronutrient supplementation for reducing mortality and morbidity in adults and children with HIV infection

2. How studies were identified

The following databases were searched to January 2010:

• CENTRAL (The Cochrane Library)
• EMBASE
• PUBMED
• GATEWAY

No unpublished data were included in this review

3. Criteria for including studies in the review

3.1 Study type

Randomized controlled trials

3.2 Study participants

Adults and children with confirmed HIV infection

(Both HIV-exposed and HIV-infected infants were included)

(HIV-infected pregnant women receiving vitamin A supplements were excluded as these trials are reviewed elsewhere)

3.3 Interventions

Micronutrient supplements compared with other micronutrient supplements, to placebo, or to no treatment. Supplements could include the following vitamins: A, D, E, C, B1, B2, niacin, B6, B12, K, folic acid, beta-carotene; or trace elements: zinc, selenium, magnesium, iron, iodine, copper, manganese, chromium, cobalt, molybdenum; or different combinations of these micronutrients

3.4 Primary outcomes

• Mortality
• Morbidity
• Hospitalizations
• Pregnancy outcomes

Secondary outcomes were indicators of HIV disease progression (viral load, T cell counts), anthropometric measures, and adverse outcomes

4. Main results

4.1 Included studies

Thirty trials involving 22,120 participants were included in this review; 20 trials investigated the effects of single supplements (vitamin A, vitamin D, zinc, selenium), while 10 investigated multiple micronutrient supplements:

• Vitamin A: six trials in adults with 14,763 participants; five trials in children with 1120 participants
• Vitamin D: one trial in adults with 365 participants; one trial in children with 59 participants
• Zinc: two trials in adults with 559 participants; two trials in children with 128 participants
• Selenium: three trials in adults with 1361 participants
• Multiple micronutrient supplements: nine trials in adults with 2687 participants (including three trials in pulmonary tuberculosis (TB) patients; one trial in female injection drug users; and one trial in pregnant and lactating women)
• Use of antiretroviral agents varied across trials

4.2 Study settings

• Canada (2 trials), Guinea-Bissau, Kenya, Malawi, South Africa (4 trials), the United Republic of Tanzania (6 trials), Thailand, Uganda, the United States of America (10 trials), Zambia, and Zimbabwe
• Study settings were varied, including rural and urban primary care and community care clinics, hospital HIV and TB outpatient clinics, antenatal clinics and maternity hospitals, university-based clinics and study clinics, and tertiary hospitals

4.3 Study settings

How the data were analysed
Micronutrient supplements were compared to no treatment, other supplements or placebo in separate analyses for each single micronutrient (vitamin A, vitamin D, zinc, selenium), and also for multiple micronutrients, with analyses further subgrouped for adults and children. Random effects meta-analysis was used to estimate pooled risk ratios (RR) for dichotomous data and mean differences (MD) for continuous data, with corresponding 95% confidence intervals (CI). Statistical heterogeneity in meta-analyses was evaluated using the I² statistic

Results
Vitamin A supplementation in HIV-infected adults
Mortality
Using a factorial design, a large Zimbabwean trial of 14,110 mother-infant pairs, including 4495 HIV-infected mothers, compared the effect a of single bolus dose of vitamin A among postpartum women (400,000 IU) and/or their infants (50,000 IU) versus placebo. In HIV-exposed infants who were HIV-negative at baseline and HIV-positive at six weeks, neonatal supplementation reduced mortality by 28% (p=0.01), while maternal supplementation had no effect. In HIV-exposed infants who were HIV-negative at six weeks, supplementation with vitamin A was associated with a two-fold increase in mortality (p≤0.05).

Morbidity and other outcomes
There was no effect of vitamin A supplementation on total morbidity among the HIV-positive women in the factorial Zimbabwean trial as measured by the total number of sick clinic visits. However, cause-specific visits for malaria, pelvic inflammatory disease, cracked and bleeding nipples, and vaginal infection were statistically significantly reduced with vitamin A supplementation by 16% to 40%. High dose beta-carotene (600,000 IU/day) administered for one month raised CD4 counts in a crossover trial in 21 American outpatients, but not when given with multivitamins at three months follow-up. In a factorial trial in Tanzania among HIV-infected pregnant and lactating women (n=1078), children born to women in the vitamin A arms experienced a reduced risk of cough with rapid respiratory rate (RR 0.69, 95% CI [0.49 to 0.96]) and a reduced risk of developing clinical malaria (RR 0.29, 95% CI [0.09 to 0.89]). No further trials reported any statistically significant differences between treatment groups.

Vitamin A supplementation in HIV-infected children
Mortality
Meta-analysis of three studies in 262 HIV-infected infants and children revealed a 45% reduction in the risk of all-cause mortality with vitamin A supplementation, which ranged from periodic boluses of 50,000 IU to 200,000 IU (RR 0.55, 95% CI [0.31 to 0.82], p=0.0031). Periodic doses (100 000 IU to 200 000 IU) of vitamin A given to 58 HIV-infected Tanzanian children admitted with pneumonia reduced AIDS-related deaths by two-thirds (RR 0.32, 95% CI [0.10 to 0.99]).

Morbidity and other outcomes
In 181 HIV-infected Ugandan children supplemented with 200,000 IU of vitamin A quarterly, the likelihood of persistent cough was more than halved (odds ratio (OR) 0.47, 95% CI [0.23 to 0.96]), and the duration of ear discharge was reduced (p=0.03) in comparison to controls. In 118 South African children born to HIV-infected mothers, periodic vitamin A supplementation (50 000 IU to 100 000 IU) reduced the likelihood of all-cause morbidity by 31% (OR 0.69, 95% CI [0.48 to 0.99]) over 18 months follow-up. Among the 28 children born HIV-positive in this trial, episodes of diarrhoea were approximately halved (OR 0.51, 95% CI [0.27 to 0.99]). In a North American trial of 59 HIV-infected children receiving influenza vaccine, 200,000 IU of vitamin A for two days decreased HIV viral load at two weeks follow-up in comparison to the placebo group (p=0.02), and in an identical regimen in 75 South African children, CD4 counts were raised at four weeks (p=0.03). In the trial of 58 HIV-infected Tanzanian children admitted with pneumonia, treatment with vitamin A at four months follow-up had improved linear growth in comparison to placebo by 2.8 cm (95% CI [1.00 to 4.60 cm]) among children <18 months of age.

Adverse effects
While no statistically significant adverse effects of vitamin A supplementation in adults were reported, in both HIV-infected and -uninfected children, adverse effects included vomiting and a bulging fontanelle, and among children hospitalized with pneumonia, acute diarrhoea. Among HIV-exposed infants who were HIV-negative at six weeks, supplementation with vitamin A was associated an increased risk in mortality (p≤0.05). In addition, vitamin A reduced the benefits of multivitamin supplementation in pregnant and lactating women in Tanzania (outlined below).

Vitamin D supplementation in HIV-infected adults
In a trial of 365 adults with TB in Guinea-Bissau, of whom 131 were also HIV-infected, vitamin D (100,000 IU) given periodically did not reduce the TB clinical severity score and had no effect on 12 month mortality overall, or in the HIV-infected subgroup, in comparison to placebo.

Vitamin D supplementation in HIV-infected children
Fifty-nine HIV-infected children and adolescents were administered 100,000 IU of vitamin D or placebo bimonthly in a North American trial, in which significant increases in serum 25-hydroxyvitamin D (p<0.0001) were observed in the supplemented group. No differences in HIV disease progression, as measured by CD4 counts (p=0.18) and viral load (p=0.66) were found between treatment groups.

Zinc supplementation in HIV-infected adults
In a factorial trial among 213 pulmonary TB and HIV co-infected Tanzanian patients, those in the zinc arms had a non-statistically significant reduced risk of mortality (RR 0·63, 95% CI [0·37 to 1·08]). In a trial of 400 HIV-infected pregnant Tanzanian women, 25 mg of zinc per day given until six weeks postpartum had no effect on fetal and neonatal mortality, duration of pregnancy, birth weight, maternal T cell counts, or viral load in comparison to placebo. However, among zinc-treated women, increases in haemoglobin, packed cell volume, and red blood cell count were lower (all p≤0.03), and a three-fold increase in the risk of maternal wasting (mid-upper arm circumference (MUAC) <22 cm) was observed (p=0.03). Among 159 Peruvian HIV-infected adults with diarrhoea lasting for one week or more, 50 mg zinc twice daily for 14 days did not reduce the persistence or severity of diarrhoea.

Adverse effects
Daily zinc supplements (25 mg) given to pregnant Tanzanian women reduced the increase in haematological indicators and raised the risk of maternal wasting in comparison to controls.

Zinc supplementation in HIV-infected children
Daily zinc (10 mg) given to 96 HIV-infected children in South African for six months did not increase viral load and significantly reduced diarrhoeal morbidity (p=0.001). In another study of both HIV-infected (n=32) and -uninfected (n=341) children in rural South Africa, zinc supplementation with or without vitamin A and other micronutrients had no effect on diarrhoeal and respiratory morbidity.

Selenium supplementation in HIV-infected adults
Selenium supplementation (200 mg/day) for 12 months in a trial of 186 North American drug users reduced the number of hospitalizations for opportunistic infections and HIV-related conditions by 60% (RR 0.40, 95% CI [0.21 to 0.75]), and reduced the risk of a CD4 decline >50 cells/mm³ in comparison to placebo. Using an identical regimen in a trial of 262 HIV-infected adults, selenium supplements were found to suppress viral load (p<0.02), and increase CD4 counts (p<0.04). In a trial of 913 pregnant HIV-infected women in Tanzania, 200 µg/day of selenium given until six months post- partum had no significant effect on overall neonatal or child mortality. However, a reduction in child mortality at six weeks of age was observed with maternal selenium supplementation (RR 0.43, 95% CI [0.19 to 0.99]), and maternal diarrhoeal morbidity was reduced by 40% (RR 0.60, 95% CI [0.42 to 0.84]).

Multiple micronutrient supplementation in HIV-infected adults
Mortality
Micronutrients plus natural carotenoids were trialed in 331 Canadian adults with advanced AIDS, and were found to have no effect on time to new or recurrent AIDS-defining event or death (hazard ratio (HR) 1.81, 95% CI [0.95 to 3.42], p=0.07). However, a reduction in time to all-cause mortality was observed with supplementation (HR 3.15, 95% CI [1.10 to 8.98], p=0.03) and there was improved survival in those with higher CD4 counts at baseline (p=0.005). Conversely, in a trial of 481 HIV-infected Thai adults, supplementation for almost one year with 18 vitamins and minerals at doses of up to 20 times the recommended daily allowance (RDA) had no significant effect on overall mortality, while among those with low CD4 counts at baseline (CD4 <100), mortality was reduced by 74% (HR 0.26, 95% CI [0.07 to 0.97]). Supplementation with multivitamins and zinc was investigated in a factorial trial in 499 pulmonary TB patients in Tanzania, 213 of whom were also HIV-positive. While zinc and multivitamins had no effect on mortality when administered separately to HIV co-infected patients, when given in combination, a 71% reduction in mortality was observed (RR 0.29, 95% CI [0.10 to 0.80], p=0.016). In another factorial study in Tanzania, multivitamins (up to 22 times the RDA of vitamins B, C, E) with or without vitamin A were compared to vitamin A alone or placebo in 1078 HIV-infected pregnant and lactating women. Delays in the progression to stage 4 disease or AIDS-related mortality (HR 0.71, 95% CI [0.51 to 0.98]) and in progression to stage 4 disease alone (HR 0.50, 95% CI [0.28 to 0.90]) were observed among women receiving multivitamin supplements in comparison to placebo. The children of multivitamin-supplemented women with low immunological or nutritional status also experienced a 70% reduction in the risk of mortality by two years of age (RR 0.30, 95% CI [0.10 to 0.92]. In a Malawian trial in 1402 pulmonary TB patients, of whom 829 were HIV-infected, daily micronutrients did not reduce mortality in the HIV-infected subgroup in comparison to placebo over two years. Among 887 Tanzanian adults with pulmonary TB, of whom 471 were also HIV-positive, supplementation with micronutrients for eight months had no effect on overall mortality. Micronutrient supplementation (vitamins A, C, E; selenium, zinc) had no effect on mortality in a trial of 135 Zambian adults with persistent diarrhoea.

Morbidity and other outcomes
In the factorial Tanzanian trial in pregnant and lactating women, maternal viral load and HIV-related complications were decreased relative to placebo, and CD4 and CD8 counts were increased in comparison to vitamin A or placebo. In addition, multivitamin supplementation improved maternal weight gain during the third trimester (MD 304 g, 95% CI [17 to 590 g]) and reduced the risk of wasting (RR 0.66, 95% CI [0.47 to 0.94]). The risk of adverse pregnancy outcomes, including fetal death (RR 0.61, 95% CI [0.39 to 0.94]), low birth weight (RR 0.55, 95% CI [0.38 to 0.81]), severe preterm birth (<34 weeks’ gestation) (RR 0.61, 95% CI [0.38 to 0.95]), and small for gestational age (RR 0.57, 95% CI [0.40 to 0.83]), was also reduced with multivitamin supplementation, and birth weight was significantly higher (MD 94 g; p=0.02). Overall, maternal multivitamin supplementation improved growth to two years of age, with significant differences in attained weight (MD 459 g, 95% CI [35 to 882), weight-for-age Z-score (MD 0.42, 95% CI [0.07 to 0.77]), and weight-for-length Z-score (MD 0.38, 95% CI [0.07 to 0.68]). Child diarrhoea at two years was also reduced with maternal supplementation (RR 0.83, 95% CI [0.71 to 0.98]), and during the first six months of life, micronutrient status and CD4 counts were improved (MD 153 cells/μL, 95% CI [67.6 to 238.4). The beneficial effects for almost all maternal and infant outcomes were abolished when vitamin A was given in conjunction with multiple micronutrients in this trial. In 471 Tanzanian adults with pulmonary TB and HIV co-infection, supplementation with micronutrients reduced the risk of TB recurrence (RR 0.37, 95% CI [0.15 to 0.92). Twelve weeks of daily micronutrient supplementation increased CD4 counts in comparison to placebo in a trial of 40 North American adults (p=0.029). No other statistically significant effects on morbidity were reported in the included trials.

5. Additional author observations*

Overall, the majority of trials were judged to be at low risk of selection bias due to random sequence generation, but at unclear risk of selection bias due to inadequate reporting of allocation concealment. In general, risk of performance and detection bias was low because of adequate blinding methodology, but fewer than half of the included studies were judged to be of low risk of attrition and reporting bias.

While evidence of the benefit of micronutrient supplements in children and adults with HIV is limited, the findings of this review support periodic vitamin A supplementation of HIV-infected children over six months of age in resource-limited settings, which is consistent with findings in HIV-uninfected children. Zinc supplements reduced diarrhoeal morbidity and had no adverse effects on disease progression among South African HIV-infected children, and thus children with HIV should receive the same treatment with zinc supplements for diarrhoea and severe acute malnutrition as HIV-uninfected children. Daily multivitamin supplementation for HIV-infected pregnant and lactating women was found to provide many benefits, although further research into the optimal dosage and duration of supplementation is warranted. In addition, the effect of specific micronutrients requires further investigation, namely selenium, vitamin D and zinc. While research participants should be diverse with regard to stage of disease, use of antiretroviral therapy, immune status, and nutritional status, it must be recognized that antiretroviral treatment consistently reduces morbidity and mortality and improves the nutritional status of adults and children infected with HIV/AIDS, and thus future research should not be to the detriment of this crucial intervention.