• zinc supplementation;
  • HIV infection;
  • systematic review
  • supplémentation en zinc;
  • infection VIH;
  • revue systématique
  • suplementación con Zinc;
  • infección por VIH;
  • revisión sistemática


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Objectives  To determine the efficacy and safety of zinc supplementary in children, adults and pregnant women with HIV infection.

Methods  We conducted a comprehensive search in Medline, Embase, the Cochrane Library, CBM, VIP and CNKI. Only randomized controlled trials conducted subsequent to the introduction of zinc supplementation were included in this systematic review. Two reviewers assessed and extracted data for analysis.

Results  Six trials with a total of 1009 participants were included. The findings in this review suggested a benefit of zinc supplementation in reducing opportunistic infection for both adults and children with HIV infection. In terms of increase in zinc level and CD4 counts, however, only adults with HIV infection benefited. For other outcomes, such as viral load, mortality, mother-to-child transmission of HIV and foetal outcomes, zinc supplementation conferred no benefit over placebo. No adverse event related to zinc supplementation was found in all the included trials.

Conclusion  Based on the current evidence, zinc supplementation seems to be beneficial in adult patients with HIV infection in some aspects. More research is needed in children and pregnant women. The influence of zinc dose, duration and usage of antiretroviral medicine also requires further investigation.

Objectifs:  Déterminer l’efficacité et la sécurité de la supplémentation en zinc chez les enfants, les adultes et les femmes enceintes infectés par le VIH.

Méthodes:  Nous avons mené une recherche exhaustive sur MEDLINE, Embase, Cochrane Library, CBM, VIP et CNKI. Seuls les essais randomisés contrôlés menés à la suite de l’introduction de la supplémentation en zinc ont été inclus dans cette revue systématique. Deux auteurs ont évalué et extrait les données pour analyse.

Résultats:  Six essais avec un total de 1009 participants ont été inclus. Les conclusions de cette revue suggèrent un bénéfice de la supplémentation en zinc dans la réduction des infections opportunistes, à la fois pour les adultes et les enfants infectés par le VIH. Cependant, en termes d’augmentation du taux de zinc et de celui des CD4, seuls les adultes infectés par le VIH en ont bénéficié. Pour d’autres résultats, tels que la charge virale, la mortalité, la transmission mère-enfant du VIH et l’issue du fœtus, la supplémentation en zinc ne confère aucun avantage par rapport au placebo. Aucun effet indésirable liéà la supplémentation en zinc n’a été observé dans tous les essais inclus.

Conclusion:  Sur la base des données actuelles, la supplémentation en zinc semble être bénéfique chez les patients adultes atteints par l’infection VIH dans certains aspects. Des recherches supplémentaires sont nécessaires chez les enfants et les femmes enceintes. L’influence de la dose de zinc, la durée et l’utilisation des médicaments antirétroviraux requièrent également une investigation plus approfondie.

Objetivos:  Determinar la eficacia y seguridad de la suplementación con zinc en niños, adultos y mujeres embarazadas infectados con VIH.

Métodos:  Hemos realizado una búsqueda integral en Medline, Embase, la Cochrane Library, CBM, VIP y CNKI. Solo se incluyeron los ensayos aleatorizados y controlados realizados tras la introducción de la suplementación con zinc. Dos revisores evaluaron y extrajeron datos para el análisis.

Resultados:  Se incluyeron seis ensayos con un total de 1009 participantes. Los hallazgos de esta revisión sugieren que existe un beneficio de suplementar con zinc en la reducción de infecciones oportunistas, tanto en adultos como en niños infectados con VIH. En términos del aumento de los niveles de zinc y los conteos de CD4, sin embargo, solo se beneficiaban los adultos con VIH. Para otros resultados, tales como la carga viral, la mortalidad, la transmisión madre-hijo del VIH, o los resultados fetales, la suplementación con zinc no confería ningún beneficio frente al placebo. No se encontraron efectos adversos relacionados con la suplementación con zinc en ninguno de los ensayos incluidos.

Conclusión:  Basándose en la evidencia actualmente disponible sobre la suplementación con zinc, esta parece ser beneficiosa en algunos aspectos en pacientes adultos infectados con VIH. Se requiere realizar más estudios en niños y mujeres embarazadas. La influencia de la dosis de zinc, la duración y el uso de antirretrovirales también requieren de más estudios.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

According to estimates by World Health Organization (WHO) and the United Nations Joint Programme on HIV/AIDS (UNAIDS), 33.4 million people were living with HIV worldwide at the end of 2008, of whom, 2.1 million were children and 15.7 million were women (WHO 2010). Sub-Saharan Africa is hardest hit; nearly 30 million adults and children there had HIV/AIDS in 2004 (WHO 2005). While antiretroviral drugs are important for those with advanced HIV infection, nutrition is of fundamental importance for all people with HIV infection (Semba & Tang 1999; Baeten et al. 2002). Adequate zinc status is critical for immune function (Wellinghausen et al. 1997; Shankar & Prasad 1998). Zinc acts as an antioxidant and may protect cells from oxygen radicals produced during the non-specific immune response (Coovadia & Bobat 2002). Zinc can bind to thymulin (which enhances T cell maturation, interleukin-2 production and cytotoxicity) and produce conformational changes essential for specific immunity (Coovadia & Bobat 2002). Zinc deficiency in HIV-infected participants was linked to declining CD4 cell counts (Baum et al. 2003; Fufa et al. 2009; Ndeezi et al. 2010) and reduced survival (Miller & Strittmatter 1992). Thus, zinc supplementation as a supporting therapeutic intervention seems reasonable for patients with HIV/AIDS (Zazzo et al. 1989; Mocchegiani et al. 1995). However, in other instances, zinc supplementation had no effect on immune response, vaccination, CD4/CD8 ratio or viral load (Deloria-Knoll et al. 2006). Two nutritional studies showed that increased intake of zinc in HIV-1-infected patients led to an augmented risk of progression to AIDS (Tang et al. 1993) and lower survival (Tang et al. 1996). The San Francisco Men’s Health Study (Abrams et al. 1993) found no association between increased dietary intakes of zinc (from food and supplements) and time to progression to AIDS. It seems that current studies cannot come to a consensus on the benefit of zinc supplementation. We therefore performed a systematic review to evaluate the efficacy and safety of zinc supplement on patients with HIV infection.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Only randomized controlled clinical trials with children, adults or pregnant women who had confirmed HIV infection were included in this systematic review. Zinc had to be supplemented as an oral drug intervention; trials that supplied zinc by dietary intake were excluded. Other basic care supplied could include other trace elements (selenium, magnesium, iron, iodine, copper, manganese, chromium, cobalt and molybdenum), vitamins (A, D, E, C, B1, B2, niacin, B6, B12, K, folate and beta-carotene). Outcome measures were changes in zinc level; CD4 counts; viral load; incidence of opportunistic infections (diarrhoea, pneumonia, upper respiratory tract infection, ear infection); mortality; incidence of adverse events; mother-to-child transmission (MTCT) of HIV; and foetal outcomes (miscarriage, stillbirth, foetal death and neonatal death).

Search methods

Searches were performed in three English databases, Cochrane Central Register of Controlled Trials (2010.09), Ovid Medline (1947-2010.09) and Embase (1966-2010.09), and three Chinese databases, Chinese Biomedical Literature Database (CBM) (1978-2010.09), China National Knowledge Infrastructure (CNKI) (1979-2010.09) and VIP database for Chinese Technical Periodicals (VIP) (1989-2010.09), using the following combination of search terms:

  • #1 HIV infections

  • #2 Human immunodeficiency virus

  • #3 Acquired immunodeficiency syndrome

  • #4 HIV

  • #5 #1 OR #2 OR #3 OR #4

  • #6 Zinc

  • #7 Trace elements

  • #8 Micronutrient

  • #9 #6 OR #7 OR #8

  • #10 #5 AND #9

We also checked the reference lists of relevant publications returned by the above searches.

Data collection and analysis

Study details (citation, country, founding and study design), participant details (inclusion and exclusion criteria, diagnosis, age, sex, sample size and follow-up), interventions details (dose, frequency and duration, whether multivitamins or antiretroviral therapy was given) and outcome details (change in zinc level, change in CD4 counts, change in viral load, opportunistic infection, mortality, adverse events, MTCT and foetal outcomes) were extracted by reviewers. For each outcome, the number of participants randomized and the number of participants analysed were recorded for each treatment group. For dichotomous data, the number experiencing events was recorded. For continuous data, arithmetic mean and standard deviation or medians and ranges were recorded. Relative risks were calculated for dichotomous data, and standardized mean differences for continuous data or change in continuous data. Risk ratios and standardized mean differences were presented with 95% confidence intervals (CI). All P-values reported were two-sided. Statistical significance in this study was defined as P ≤ 0.05. All statistical analyses were conducted using RevMan 5.0 (Nordic Cochrane Centre 2008). Heterogeneity between trials was assessed using the chi-square test (P < 0.1), which was quantified by the Higgins I2 statistic. An I2 value of 50% or greater was an indication of substantial heterogeneity. Subgroup analysis was conducted in children, adults and pregnant women.

Quality assessment

The quality of evidence was assessed using GRADE approach (GRADEpro 3.5.1) (Brozek et al. 2008), defining the quality of evidence for each outcome as ‘the extent to which one can be confident that an estimate of effect or association is close to the quantity of specific interest’ (Higgins & Green 2008). There are four levels of quality: high (further research is very unlikely to change our confidence in the estimate of effect), moderate (further research is likely to have an impact on our confidence in the estimate of effect and may change the estimate), low (further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate) and very low (we are very uncertain about the estimate). Factors that may lower the quality of evidence are limitation in design, inconsistency of results, indirectness of evidence, imprecision and high probability of publication bias. Factors that can raise the quality of evidence are as follows: a large magnitude of effect, if all plausible confounding would reduce a demonstrated effect; and a dose-response gradient (Guyatt et al. 2008).


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Description of studies

Database searches yielded 1329 titles, 1324 of which were excluded for not meeting inclusion criteria. Two more references were identified through searching reference lists of relevant publications. Seven publications were retained (Mocchegiani et al. 1995; Bobat et al. 2005; Fawzi et al. 2005; Green et al. 2005; Carcamo et al. 2006; Villamor et al. 2006; Baum et al. 2010). Two reported the same trial of zinc supplementation for pregnant women in Tanzania (Fawzi et al. 2005; Villamor et al. 2006). Of the six trials with a total of 1009 participants, four were in HIV-infected adults (Mocchegiani et al. 1995; Green et al. 2005; Carcamo et al. 2006; Baum et al. 2010), one was in HIV-infected children (Bobat et al. 2005) and one in pregnant women with HIV infection (Fawzi et al. 2005; Villamor et al. 2006). All participants in Carcamo et al. (2006) had persistent diarrhoea lasting for more than 7 days. All of the studies were single-centred clinical trials, three in developed countries and three in developing countries (Table 1).

Table 1.   Characteristics of included studies
Study IDCountryNo. of centresFounding sourceStudy designParticipantsIntervention
Type of trialITTDurationDiagnosisSample sizeFollow-up rateSex (male %)Age (mean ± SD or median and range)Dose and duration of zincART useMultivitamin use
  1. Y, Yes; N, No; U, Unclear; △, Intention to treat.

  2. *Body weight, plasma zinc, CD4 cells were detected 3 months after the beginning of zinc or placebo treatment; opportunistic infection were detected during the 2 years of follow-up.

Baum et al. (2010)America1YRandomized, double-blind, placebo-controlled trialY18 monthsAdults with HIV infection231200/231169/23142.7 ± 7.0Women:12 mg/day144/231U
Men:15 mg/day, for 18 months
Carcamo et al. (2006)Peru1YRandomized, double-blind, placebo-controlled trialY2 weeksHIV-infected adults with persistent diarrhoea159108/159110/159Zinc: 30 (19–57)100 mg elemental zinc/day, for 2 weeksUU
Placebo: 31 (19–64)
Green et al. (2005)Singapore1YRandomized, double-blind, placebo-controlled trialN28 daysHIV-infected adults with a CD4 count of <200/mm36665/6661/66Zinc: 40 ± 7.850 mg element zinc/day, for 28 days63/64U
Placebo: 40 ± 8.3
Mocchegiani et al. (1995)Italy1YRandomized controlled trialU2 years*HIV-infected adults at stage III or stage IV57UStage III: 19/35Stage III: 26.5 ± 545.5 mg element zinc/day, for1 month57/57U
stage IV: 16/22stage IV: 28 ± 4.3
Bobat et al. (2005)South Africa1YRandomized, double-blind, placebo-controlled trialN9 monthsHIV-infected children, aging between 6 and 60 months9685/96Zinc: 40.1 (27.4–48.4)Zinc: 40.1 (27.4–48.4)10 mg element zinc/day, for 6 months96/9691/96
Placebo: 36.6 (25–49.4)Placebo: 36.6 (25–49.4)
Fawzi et al. (2005) and Villamor et al. (2006)Tanzania1URandomized, double-blind, placebo-controlled trialYFrom 12–27 gestational age to 6 weeks post-partumHIV-infectedpregnant women with an estimated gestational age at randomization of between 12 and 27 weeks400397/400Zinc: 26.7 ± 4.9Zinc: 26.7 ± 4.925 mg element zinc/day, from enrolment until 6 weeks post-partum400/400400/400
Placebo: 27.0 ± 5.0Placebo: 27.0 ± 5.0

Effects of interventions

Four trials conducted in HIV-infected adults analysed the change in zinc concentration (Mocchegiani et al. 1995; Green et al. 2005; Carcamo et al. 2006; Baum et al. 2010). The zinc status at baseline was almost the same among trials (Table 2). Only two studies that reported the mean and SD could be combined (Mocchegiani et al. 1995; Green et al. 2005). However, heterogeneity between two trials was too great to perform a meta-analysis (χ2 = 10.22, I2 = 80%). Participants in these two trials received moderate-dose, short-duration zinc supplementation. In Green et al. (2005), the zinc level rose in both zinc and placebo arm without statistical difference (P = 0.67); Mocchegiani et al. (1995) only observed this in the intervention arm (Table 3). Participants received low-dose, long-duration zinc supplementation had significantly higher zinc levels than those in the placebo group (β = 0.04; P = 0.047) (Baum et al. 2010). In addition, participants who received high-dose, short-duration zinc supplementation had a better outcome (Carcamo et al. 2006). Studies conducted in children or pregnant women did not report the change in zinc level during trials.

Table 2.   Zinc status at baseline
Study ID Zinc status at baseline
  1. U: Unclear

Baum et al. (2010)Serum zinc level: mean (SD)All patient: 0.6 (0.1) mg/l
Zinc: 0.6 (0.1)
Placebo: 0.7 (0.2)
Carcamo et al. (2006)N (%) ZincPlacebo
Normal:6/8 (75)5/9 (55.6)
Marginal deficiency:5/8 (62.5)3/5 (60)
Overt deficiency:19/34 (55.9)23/40 (57.5)
Green et al. (2005)Serum zinc: mean (SD)Zinc: 14.1 (2.8) umol/l
Placebo: 13.3 (2.3)
Mocchegiani et al. (1995)Serum zinc: mean (SD) (ug/dl) FourthThird
Zinc78.0 (4.3)76.8 (2.7)
Placebo80.0 (3.1)79.8 (3.5)
Bobat et al. (2005) U
Fawzi et al. (2005) and Villamor et al. (2006) U
Table 3.   Outcomes, Zinc vs. Placebo (continuous data)
Study or SubgroupZincPlaceboMean difference IV, random, 95% CI
Change of zinc level, zinc vs. placebo
  Green et al. (2005)0.00430.0096320.00380.0053340.00 [−0.00, 0.00]
  Mocchegiani et al. (1995) (stage 3)0.00080.000418−0.00120.0006190.00 [0.00, 0.00]
  Mocchegiani et al. (1995) (stage 4)0.00060.000610−0.00230.0006110.00 [0.00, 0.00]
Change of CD4 counts, zinc vs. placebo
  Mocchegiani et al. (1995) (stage 3)492017−1112118160.00 [146.42, 173.58]
  Mocchegiani et al. (1995) (stage 4)401012−30101070.00 [61.61, 78.39]
  Bobat et al. (2005) (3 months)09.844−18.3411.00 [−2.85, 4.85]
  Bobat et al. (2005) (6 months)19.844−18.7412.00 [−1.93, 5.93]
  Bobat et al. (2005) (9 months)110.14409.1411.00 [−3.08, 5.08]
 Pregnant women
  Fawzi et al. (2005)95126200101137200−6.00 [−31.80, 19.80]
Change of viral loads
  Green et al. (2005)2.91241.83231−1.73543.453324647.00 [−16412.72, 25706.72]
  Bobat et al. (2005) (3 months)0.10.664400.73410.10 [−0.20, 0.40]
  Bobat et al. (2005) (6 months)0.20.674400.7410.20 [−0.09, 0.49]
  Bobat et al. (2005) (9 months)0.20.64440.10.79410.10 [−0.21, 0.41]
 Pregnant women
  Villamor et al. (2006)−0.180.65500.010.5650−0.19 [−0.43, 0.05]

CD4 counts rose in the zinc group for both adult and children participants, but fell in the placebo or no treatment group (Mocchegiani et al. 1995; Bobat et al. 2005). Zinc supplementation was effective in preventing immunological failure in adults, which was defined as a drop in CD4 counts to <200 cells/mm3 (Baum et al. 2010). However, CD4 counts increased in HIV-infected mothers at 6 weeks post-partum in both zinc and placebo groups, without a statistically significant difference (P = 0.97) (Fawzi et al. 2005) (Table 3).

There was no significant change in HIV viral load in either zinc or placebo group in adult participants during a 28-day intervention (Green et al. 2005). Likewise, Log10 viral loads were similar between the zinc supplementation and placebo groups during a 6-month treatment in children (Bobat et al. 2005). Villamor et al. (2006) selected 100 women for viral load analysis. The baseline sample was collected at gestation week 22 on average and the follow-up measurement at 6 weeks post-partum. Log10 viral loads decreased from baseline to the follow-up measurement in the zinc but not in the placebo group. The risk of having a high viral load (>100 000 copies/ml) at the post-partum visit was not significantly lower in women who received zinc (RR = 0.77, 95% CI: 0.42–1.40, P = 0.38) (Table 3).

Three studies reported opportunistic infections in HIV-infected adults, of which only two could be combined in a meta-analysis (Mocchegiani et al. 1995; Green et al. 2005). Zinc supplementation seemed to help reduce the risk of infection in HIV-infected adults. An intent-to-treat analysis in Baum et al. (2010) showed that zinc supplementation significantly reduced the rate of diarrhoea over time by more than half in adult patients (OR = 0.4, 95% CI: 0.183–0.981, P = 0.019). Diarrhoea was significantly associated with lower mean plasma levels of zinc (0.59 ± 0.11 vs. 0.68 ± 0.21 mg/l). In children, the proportion of scheduled and illness visits at which children were diagnosed with watery diarrhoea (P = 0.001) or pneumonia (P = 0.07) was smaller for the zinc group than for the placebo group; the number of illness visits per month was slightly lower in the zinc arm (0.11 illness visits/month) than in the placebo arm (0.16 illness visits/month, P = 0.05) (Bobat et al. 2005) (Table 4).

Table 4.   Outcomes, Zinc vs. Placebo (dichotomous data)
Study or SubgroupZincPlaceboWeight (%)Risk ratio M-H, random, 95% CI
Opportunistic infection
  Green et al. (2005)43233418.41.42 [0.34, 5.84]
  Mocchegiani et al. (1995)929132881.60.67 [0.34, 1.31]
  Subtotal (95% CI)13611662100.00.77 [0.42, 1.41]
Heterogeneity: τ2 = 0.00; χ2 = 0.90, df = 1 (P = 0.34); I2 = 0%
Test for overall effect: Z = 0.85 (P = 0.39)
  Baum et al. (2010)111158116351.39 [0.58, 3.32]
  Carcamo et al. (2006)081078 Not estimable
  Green et al. (2005)032034 Not estimable
  Mocchegiani et al. (1995)529112833.90.44 [0.17, 1.10]
  Subtotal (95% CI)162571925669.00.79 [0.25, 2.44]
Heterogeneity: τ2 = 0.45; χ2 = 3.17, df = 1 (P = 0.08); I2 = 68%
Test for overall effect: Z = 0.41 (P = 0.68)
  Bobat et al. (2005)24675022.20.31 [0.07, 1.42]
Test for overall effect: Z = 1.51 (P = 0.13)
 Pregnant women
  Fawzi et al. (2005)320002008.87.00 [0.36, 134.64]
Test for overall effect: Z = 1.29 (P = 0.20)
  Total (95% CI)2150326506100.00.78 [0.29, 2.05]
Heterogeneity: τ2 = 0.50; χ2 = 6.74,df = 3 (P = 0.08); I2 = 56%
Test for overall effect: Z = 0.51 (P = 0.61)
Mother-to-child transmission of HIV
 At birth (0–21 days)
  Villamor et al. (2006)81416145100.01.37 [0.49, 3.85]
Test for overall effect: Z = 0.60 (P = 0.55)
 By 6 weeks
 Villamor et al. (2006)1514111145100.01.40 [0.67, 2.95]
Test for overall effect: Z = 0.89 (P = 0.37)
Foetal outcomes
  Villamor et al. (2006)51983199100.01.68 [0.41, 6.91]
Test for overall effect: Z = 0.71 (P = 0.48)
 Still birth
  Villamor et al. (2006)1319810199100.01.31 [0.59, 2.91]
Test for overall effect: Z = 0.65 (P = 0.51)
 Foetal death
  Villamor et al. (2006)1819813198100.01.38 [0.70, 2.75]
Test for overall effect: Z = 0.93 (P = 0.35)
 Neonatal death
  Villamor et al. (2006)101807186100.01.48 [0.57, 3.79]
Test for overall effect: Z = 0.81 (P = 0.42)

Mortality of adults and children in the zinc group was slightly lower than that of placebo group, but without statistical significance (RR = 0.79, 95% CI: 0.25–2.44, P = 0.6; RR = 0.31, 95% CI: 0.07–1.42, P = 0.13, respectively). However, three mothers in a zinc group died after delivery while none died in the placebo group for reasons that remain unclear (Fawzi et al. 2005) (Table 4).

Zinc supplementation had no significant effect on MTCT of HIV at birth (0–21 days) (RR = 1.37, 95% CI: 0.49–3.85, P = 0.55) or at 6 weeks post-partum (RR = 1.40, 95% CI: 0.67–2.95, P = 0.37). This was determined by testing whole blood samples from babies using the Amplicor HIV-1 detection kit version 1.5 (Table 4).

Compared with placebo, zinc had no significant effect on risk of death during the foetal (RR = 1.38, 95% CI: 0.70–2.75, P = 0.35) and early post-partum periods (RR = 1.48, 95% CI: 0.57–3.79, P = 0.42) (Table 4).

No adverse events attributable to zinc supplementation were observed in the included studies. However, indicators of zinc toxicity, such as copper status, were not monitored except in one trial, in which zinc supplementation did not induce copper deficiency (Carcamo et al. 2006).

Quality of included studies

There was low-quality evidence for a benefit from zinc supplementation for change in zinc level, change in CD4 counts and for opportunistic infection among HIV-infected patients. The quality of evidence was downgraded for risk of bias (lack of allocation concealment and lack of intent-to-treat analysis) and imprecision (small number of participants or events and thus wide confidence intervals).There was moderate-quality evidence for effect on viral load and mortality. The quality of evidence was downgraded owing to the lack of allocation concealment and intent-to-treat analysis. Despite low risk of bias from the trial conducted in pregnant women with HIV infection, the overall body of evidence for the effects of zinc supplement was downgraded owing to the lack of allocation concealment (Table 5).

Table 5.   Assessment of quality of evidence across studies by GRADE
Quality assessmentQualityImportance
No of studiesDesignLimitationsInconsistencyIndirectnessImprecisionOther considerations
  1. *The method of sequence generation and allocation concealment was not reported in Green et al. (2005) or Mocchegiani et al. (1995). And intent-to-treat analysis was not performed in Green et al. (2005).

  2. †The method of sequence generation and allocation concealment was not reported in Green et al. (2005). And intent-to-treat analysis was not performed in Green et al. (2005) or Fawzi et al. (2005).

  3. Green et al. (2005) and Mocchegiani et al. (1995) include few participants and few events and thus have wide confidence intervals.

  4. §The confidence interval in Fawzi et al. (2005) is wide.

  5. ¶Lack of allocation concealment in Fawzi et al. (2005) and Villamor et al. (2006).

Change of zinc level (follow-up mean 213 weeks; better indicated by higher values)
2Randomized trialsSerious*No serious inconsistency†No serious indirectnessSerious‡None⊕⊕ΟΟ LowNot important
Change of CD4 counts (better indicated by lower values)
3Randomized trialsSerious†No serious inconsistencyNo serious indirectnessSerious§None⊕⊕ΟΟ LowImportant
Change of viral load (better indicated by lower values)
3Randomized trialsSerious*No serious inconsistencyNo serious indirectnessNo serious imprecisionNone⊕⊕⊕Ο ModerateCritical
Opportunistic infection
4Randomized trialsSeriousNo serious inconsistencyNo serious indirectnessSerious‡None⊕⊕ΟΟ LowCritical
6Randomized trialsSerious*No serious inconsistencyNo serious indirectnessNo serious imprecisionNone⊕⊕⊕Ο ModerateImportant
Mother-to-child transmission of HIV
1Randomized trialsSerious¶No serious inconsistencyNo serious indirectnessNo serious imprecisionNone⊕⊕⊕Ο ModerateCritical
Foetal outcomes
1Randomized trialsSerious¶No serious inconsistencyNo serious indirectnessNo serious imprecisionNone⊕⊕⊕Ο ModerateImportant


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References

Potential bias in this review

This review is limited by the available literature on zinc supplementation for treating HIV infection patients. Half of the studies included in this review were conducted in developed countries, and thus, results from such studies might not well reflect the conditions in resource-constrained countries. Confounding factors were not well balanced across trials. Only three of six trials reported the HIV stage of participants, and the CD4 counts of participants at baseline varied across trials. In adult studies, the percentage of participants with CD 4 counts below 200 cells/mm3 in Carcamo’s study (2006) was much higher than that of Baum et al. (2010). Difference in other characteristics, such as multivitamin treatment, drug use and sexual behaviour, may also have caused the heterogeneity in this review. Some intermediate outcomes may not well indicate the efficacy of zinc, as they might be influenced by other factors. Kupka and Fawzi (2002) demonstrated that serum and plasma zinc levels can vary widely as a result of circadian rhythm, meals and stress (Kupka & Fawzi 2002). Zinc levels may also be depressed by opportunistic infections (Graham et al. 1991; Fauci et al. 1996). For those outcome measurements, we gave limited level of importance in GRADE approach.

Implication for practice

In terms of rising zinc levels and CD4 counts and reducing opportunistic infections, zinc supplementation is beneficial for both adults and children with HIV infection, regardless of dose and duration. In terms of reducing viral load, preventing MTCT of HIV or foetal outcomes, zinc supplementation made no difference. In terms of mortality, its effect remains unclear as the numbers involved were too small. Zinc supplementation caused no adverse events regardless of dosage and duration of the intervention; hence, it seems safe for patients with HIV infection. Therefore, we conclude that zinc supplementation seems to be beneficial in terms of change in zinc level, CD4 count and prevention of opportunistic infection for patients with HIV infection, but it is still uncertain that whether zinc has any effect on viral load, mortality, MTCT of HIV and foetal outcomes. However, as most of results were from adult trials, the safety and efficacy of zinc supplementation in children and pregnant women remain to be ascertained.

Implications for research

Most current studies mainly focused on the effect and safety of micronutrient supplementation or dietary supplementation in patients with HIV infection (Kelly et al. 1999; Jiamton et al. 2003; Irlam et al. 2005). Micronutrients, however, often interact with each other (Wieringa et al. 2003), and thus, the efficacy of a particular trace element should be evaluated in single-interventional trials. Studies evaluating zinc supplementation as single interventions may have been underpowered to assess efficacy outcomes. There is a tremendous need for clinical research on this issue by intervention trials and the influence of treatment dose and duration on the efficacy. More trials are needed in resource-limited countries where the burden of disease is greatest, as the initial nutritional status and other characteristics of participants may influence the efficacy of zinc supplementation (Friis 2006); studies should consider investigating its role in patients at particular HIV stages.


  1. Top of page
  2. Summary
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. References