Intervention Review

You have free access to this content

Micronutrient supplementation for children with HIV infection

  1. James H Irlam1,*,
  2. Nandi Siegfried2,
  3. Marianne E Visser3,
  4. Nigel C Rollins4,5

Editorial Group: Cochrane HIV/AIDS Group

Published Online: 11 OCT 2013

Assessed as up-to-date: 27 NOV 2011

DOI: 10.1002/14651858.CD010666


How to Cite

Irlam JH, Siegfried N, Visser ME, Rollins NC. Micronutrient supplementation for children with HIV infection. Cochrane Database of Systematic Reviews 2013, Issue 10. Art. No.: CD010666. DOI: 10.1002/14651858.CD010666.

Author Information

  1. 1

    University of Cape Town, Primary Health Care Directorate, Cape Town, Western Cape, South Africa

  2. 2

    University of Cape Town, Department of Psychiatry and Mental Health, Cape Town, South Africa

  3. 3

    Cape town, South Africa

  4. 4

    World Health Organization, Department of Maternal, Newborn, Child and Adolescent Health (MCA), 1211 Geneva 27, Switzerland

  5. 5

    University of Kwazulu-Natal, Department of Paediatrics and Child Health, Durban, South Africa

*James H Irlam, Primary Health Care Directorate, University of Cape Town, E47 OMB, Groote Schuur Hospital, Cape Town, Western Cape, 7925, South Africa. James.Irlam@uct.ac.za.

Publication History

  1. Publication Status: New
  2. Published Online: 11 OCT 2013

SEARCH

 
Characteristics of included studies [ordered by study ID]
Arpadi 2009

MethodsCountry: US

Setting: 4 hospital-based HIV treatment programmes

Duration of recruitment: 2004-2005

Design: Placebo-controlled, parallel group


ParticipantsINCLUSION CRITERIA:

Perinatally-infected children and adolescents, aged 6-16 years.

EXCLUSION CRITERIA:
severe vitamin D deficiency

Participants randomised: 59

- 26 M and 33 F

- mean age = 10.4 yrs

Participants analysed: 56

Loss to follow-up/ withdrawal: 6

Exclusions post-randomisation: 0


InterventionsVitamin D 100 000 IU bimonthly, and 1000 mg calcium (2 chews) per day for 12 months

CONTROL: double placebo


OutcomesPRIMARY OUTCOMES:

Total Body Bone Mineral Content (listed in registered protocol)

SECONDARY OUTCOMES:

Serum 25 hydroxyvitamin D (25-OHD) concentrations

Serum and urine calcium

HIV disease progression (CD4 count, viral load, ARV failure)


NotesNo adverse events reported


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputerised randomisation

Allocation concealment (selection bias)Low riskCentral allocation by study statistician

Blinding of participants and personnel (performance bias)
All outcomes
Low riskStudy personnel and participants were blinded

Blinding of outcome assessment (detection bias)
All outcomes
Low riskStudy personnel were blinded

Incomplete outcome data (attrition bias)
All outcomes
Low riskLow attrition (3 of 59 failed to complete the study)

Selective reporting (reporting bias)Low riskStudy protocol is available and all pre-specified outcomes of interest are reported on

Other biasLow riskDeclared no conflict of interest

Bobat 2005

MethodsCountry: South Africa

Setting: hospital outpatient clinic

Duration of recruitment: Mar - Dec 2003

Design: Placebo-controlled, parallel group


ParticipantsINCLUSION CRITERIA:

Children aged 6-60 months with HIV-1 infection attending hospital clinic in Pietermaritzburg, South Africa

EXCLUSION CRITERIA: Children receiving ARVs

Participants randomised: 96

- 49 F and 46 M

- median age (zinc group): 40.1 months (27.4 to 48.4)

- median age (placebo group): 36.6 months (25 to 49.4)

Participants analysed: 85

Losses to follow-up/ withdrawal: 11

Exclusions post-randomisation: 0


InterventionsZinc sulphate 10 mg daily for 6 months

CONTROL: placebo


OutcomesPRIMARY OUTCOME:
Viral load

SECONDARY OUTCOMES:

% CD4 cells

Haemoglobin concentrations

Mortality

Morbidity (Watery diarrhoea; Pneumonia; URTI; Ear infection)


NotesNo adverse events reported


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputerised randomisation in blocks of size 8

Allocation concealment (selection bias)Low riskAllocation by investigator at hospital who were unaware of the allocation until follow up was completed

Blinding of participants and personnel (performance bias)
All outcomes
Low riskStudy personnel and participants were blinded

Blinding of outcome assessment (detection bias)
All outcomes
Low riskStudy personnel were blinded

Incomplete outcome data (attrition bias)
All outcomes
Low riskSmall loss to follow up and reasons were given

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement; study protocol not available

Other biasLow riskDeclared no conflict of interest

Coutsoudis 1995

MethodsCountry: South Africa

Setting: Tertiary hospital study clinic

Duration of recruitment: April 1991-November 1993

Design: Placebo-controlled, parallel group


ParticipantsINCLUSION CRITERIA:

Infants of women with HIV infection who had attended the antenatal clinic, delivered in hospital, and who lived within 10 miles of the hospital.

EXCLUSION CRITERIA:

Preterm infants

Participants randomised: 118

- 28 with HIV infection (13 in vitamin A group)

- 66 M and 52 F

- mean maternal age = 25 yrs (vitamin A) vs. 24.8 yrs (placebo)

Loss to follow-up:

17% (vitamin A) vs. 25% (placebo) at 6 months

36% (vitamin A) vs. 33% (placebo) at 12 months

58% (vitamin A) vs. 63% (placebo) at 18 months

Exclusions post-randomisation: 0


InterventionsVitamin A 50 000 IU retinyl palmitate at 1 and 3 months and 100 000 IU at 6, 9,12 and 15 months

CONTROL: placebo


OutcomesPRIMARY OUTCOMES:

Overall morbidity

SECONDARY OUTCOMES:

Acute diarrhoea
Persistent diarrhoea (> or = 7 days)
Hospitalised for diarrhoea
Thrush
Upper respiratory tract infection
Lower respiratory tract infection (LRTI)
Hospitalised for LRTI
Rash


NotesSide effects (vomiting and bulging fontanelle) were monitored by history and, when possible, by clinical examination.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandom number table

Allocation concealment (selection bias)Unclear riskThe capsules looked identical and were placed in number-coded envelopes from which they were removed when appropriate.

Blinding of participants and personnel (performance bias)
All outcomes
Low riskStudy personnel and participants were blinded

Blinding of outcome assessment (detection bias)
All outcomes
Low riskStudy personnel were blinded

Incomplete outcome data (attrition bias)
All outcomes
High riskConsiderable losses to follow up were insufficiently explained

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement; study protocol not available

Other biasUnclear riskDid not declare on conflict of interest

Fawzi 1999

MethodsCountry: Tanzania

Setting: Hospital inpatient and outpatient follow-up

Duration of recruitment: April 1993 - March 1997.

Design: Placebo-controlled, parallel group


ParticipantsINCLUSION CRITERIA:

Admitted to hospital for pneumonia; aged 6 - 60 months; no eye signs or symptoms of vitamin A deficiency

EXCLUSION CRITERIA:

treatment with vitamin A for 4 months prior to study entry; severe malnutrition; measles; pulmonary tuberculosis; diphtheria; whooping cough; xerophthalmia

Participants randomised:687

- 58 with HIV infection of 648 with known status (9%)

- 353 M and 295 F

- mean maternal age = 25.6 yrs (vitamin A group) vs. 26.2 (placebo)

Participants analysed: 648

Loss to follow-up/ withdrawal:76

Exclusions post-randomisation: 0


InterventionsVitamin A single dose on hospital admission, on day 2 and at 4, and at 8 months after discharge (100,000 IU dose for infants; 200,000 IU for children)

CONTROL: placebo


OutcomesPRIMARY OUTCOMES:

All-cause mortality

Cause-specific mortality (AIDS, diarrhoea, pneumonia, malaria, anaemia and other infections (measles, meningitis, dysentery, fever of unknown origin, malnutrition))

Diarrhoea
Acute respiratory infection

SECONDARY OUTCOMES:

Hospitalisation

Visits to health center


NotesSera from children were tested for HIV antibodies by enzyme- linked immunosorbent assay and Western blot tests. For positive children <15 months of age, HIV infection was confirmed by amplified heat-denatured HIV-p24 antigen assays with confirmatory neutralization assays.

Cause of death ascertained by review of hospital records and home verbal autopsy questionnaire by two physicians. Discrepancies resolved by third physician.

Bi-weekly morbidity recall. No data provided on episodes of persistent diarrhoea or hospitalisation of children with HIV infection.

Source of funding: Thrasher Research Fund; International Development Research Center


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskRandomisation in blocks of 20

Allocation concealment (selection bias)Low riskVitamin A and placebo were dispensed out of a dropper from identical 25-ml opaque bottles that were labelled with one of four batch numbers. The batch number code was retained by the manufacturer until the end of the study

Blinding of participants and personnel (performance bias)
All outcomes
Low riskStudy personnel and participants were blinded

Blinding of outcome assessment (detection bias)
All outcomes
Low riskStudy personnel were blinded

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskNo reasons for losses to follow up were given

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement; study protocol not available

Other biasUnclear riskDid not declare on conflict of interest

Hanekom 2000

MethodsCountry: USA

Setting: Hospital HIV clinic

Duration of recruitment: Not stated

Design: Randomised placebo-controlled trial


ParticipantsINCLUSION CRITERIA:

Clinic patients who were eligible for non-primary annual influenza vaccination

EXCLUSION CRITERIA:

Hypersensitivity to eggs, acute febrile illness, recent receipt of intravenous gammaglobulin or vaccination.

Participants randomised: 59 children

- M : F ratio = 0.6 (vitamin A) and 1.3 (placebo)

- median (range) age = 84 (31-209) months in vitamin A group; 77 months (25-142) in placebo group

Participants analysed: 59

Loss to follow-up/ withdrawal: 1

Exclusions post-randomisation: 0


InterventionsVitamin A 200 000 IU retinyl palmitate daily for 2 days

CONTROL: placebo


OutcomesPRIMARY OUTCOMES:

Viral load changes after vaccination

Antibody levels (H1N1, H3N2) after vaccination

SECONDARY OUTCOMES:

T-cell counts

Vitamin A levels


NotesNo adverse events reported


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated by pharmacy

Allocation concealment (selection bias)Unclear riskNo method of concealment was described

Blinding of participants and personnel (performance bias)
All outcomes
Low riskStudy personnel and participants were blinded

Blinding of outcome assessment (detection bias)
All outcomes
Low riskStudy personnel were blinded

Incomplete outcome data (attrition bias)
All outcomes
Low riskOne exclusion due to incomplete follow up

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement; study protocol not available

Other biasUnclear riskDid not declare on funding sources or conflicts of interest

Hussey 1996

MethodsCountry: South Africa

Setting: HIV clinic at a children's hospital

Duration of recruitment:1994-1995.

Design: Randomised placebo-controlled trial.

Lost to follow-up at 2 months:
Total sample: 1 (3%)

Intention-to-treat: not performed.


ParticipantsINCLUSION CRITERIA:

Child attendees at HIV clinic

EXCLUSION CRITERIA:

Acute infections, fever

Participants randomised: 75

- mean age = 17 mo.

Participants analysed: 75

Loss to follow-up/ withdrawal: 0

Exclusions post-randomisation: 0.


InterventionsVitamin A 200 000 IU daily for 2 days

CONTROL: placebo


OutcomesPRIMARY OUTCOMES:

T-Cell counts (absolute; CD4; CD56; CD29)

SECONDARY OUTCOMES:

Vitamin A levels


NotesConference abstract only


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Unclear riskSequence generation method not described

Allocation concealment (selection bias)Unclear riskNo method of concealment was described

Blinding of participants and personnel (performance bias)
All outcomes
Low riskStudy personnel and participants were blinded

Blinding of outcome assessment (detection bias)
All outcomes
Low riskStudy personnel were blinded

Incomplete outcome data (attrition bias)
All outcomes
Unclear riskInsufficient information on losses to follow up

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement; study protocol and full report not available

Other biasUnclear riskDid not declare on funding sources or conflicts of interest

Luabeya 2007

MethodsCountry: South Africa

Setting: Five rural primary care clinics

Duration of recruitment: June 2003 - Oct 2004

Design: Randomised controlled trial


ParticipantsINCLUSION CRITERIA:

children 4-6 months old

EXCLUSION CRITERIA:

underweight for age; nutritional oedema; persistent diarrhoea; taking vitamin or micronutrient supplements in past month

Participants randomised: 373

- 32 with HIV infection; 154 born to mothers with HIV infection; 187 born to mothers without HIV

- 173 M and 162 F

- mean age = 5.5 months

Participants analysed: 335

Loss to follow-up/ withdrawal: 88

Exclusions post-randomisation: 0


InterventionsZinc (10mg), vitamin A (1250 IU) and multiple micronutrients (B vitamins; vitamins C, D, E, K; copper, iron, iodine) vs. zinc and vitamin A; daily for median duration of 14.9 months

CONTROL: Vitamin A


OutcomesPRIMARY OUTCOMES:

Incidence of diarrhoea per child (% of days)

SECONDARY OUTCOMES:

Distribution of diarrhoeal morbidity

Severity of diarrhoea

Prevalence (% of weeks) of upper respiratory symptoms

Percentage of children who ever had pneumonia (maternal and field worker reports)

Longitudinal growth (length-for-age Z-scores or LAZ)

Anaemia


NotesVomiting reported post-supplementation in 9 children by fieldworkers


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputerised randomisation in blocks of 6

Allocation concealment (selection bias)Low riskPre-packed sequentially numbered study supplements

Blinding of participants and personnel (performance bias)
All outcomes
Low riskInvestigators, study staff and participants were blinded

Blinding of outcome assessment (detection bias)
All outcomes
Low riskStudy personnel were blinded

Incomplete outcome data (attrition bias)
All outcomes
Low riskMissing data addressed used appropriate statistical methods

Selective reporting (reporting bias)Low riskStudy protocol is available and all pre-specified outcomes of interest are reported on

Other biasUnclear riskDelay in shipment of supplements prevented 243 children from receiving supplements for 11 weeks

Mda 2010a

MethodsCountry: South Africa

Setting: academic hospital

Duration of recruitment: November 2005 and May 2007

Design: Placebo-controlled trial: individual randomised stratified by diarhoea or pneumonia


ParticipantsINCLUSION CRITERIA: children with HIV infection aged between 4 mo and 2 y; admitted with diarrhoea or pneumonia to the paediatric wards of an academic hospital

EXCLUSION CRITERIA: diarrheal episode longer than 72 h on admission; pneumonia complicated by respiratory failure; children on ART or those who had received vitamin or micronutrient supplementation; children with chronic illness unrelated to HIV

Participants randomised:118

Exclusions post-randomisation: 0

Deaths: 12

Participants analysed:106


InterventionsMulti-micronutrient supplement until hospital discharge with RDAs for one-year old: vitamin A 300 mcg, vitamin B1 0.6 mg, vitamin B2 0.6 mg, vitamin B3 8 mg, vitamin B6 0.6 mg, vitamin B12 1 mcg, folic acid 70 mcg, vitamin C 25 mg, vitamin D 5 mcg, vitamin E 7 mg, copper 700 mcg, iron 8 mg, selenium 30 mcg and zinc 8 mg.

CONTROL: Placebo


OutcomesPRIMARY OUTCOMES:

Duration of hospitalization in children with HIV infection admitted with acute diarrhea and pneumonia who are not yet receiving ART

SECONDARY OUTCOMES:

Appetite and levels of appetite regulating hormones


NotesNo adverse events reported


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated randomization by independent person

Allocation concealment (selection bias)Unclear riskConcealment of allocation insufficiently described

Blinding of participants and personnel (performance bias)
All outcomes
Low riskStudy personnel and participants were blinded

Blinding of outcome assessment (detection bias)
All outcomes
Low riskStudy personnel were blinded

Incomplete outcome data (attrition bias)
All outcomes
Low riskAll survivors were included in the analysis

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement; study protocol not available

Other biasLow riskAll the authors declare that they have no conflict of interests.

Mda 2010b

MethodsCountry: South Africa

Setting: paediatric outpatient department of academic hospital

Duration of recruitment: November 2005 until November 2006

Design: Placebo-controlled trial


ParticipantsINCLUSION CRITERIA: children with HIV infection aged between 6 and 24 months; previously admitted with diarrhoea or pneumonia to the paediatric wards of an academic hospital

EXCLUSION CRITERIA: Children on ARV or who had received micronutrient supplementation in the previous two months; children diagnosed with a chronic illness unrelated to HIV infection.

Participants randomised:140

Exclusions post-randomisation: 0

Deaths: 22

Participants analysed: 99


InterventionsMulti-micronutrient supplement for 6 months with RDAs for one-year old: vitamin A 300 mcg, vitamin B1 0.6 mg, vitamin B2 0.6 mg, vitamin B3 8 mg, vitamin B6 0.6 mg, vitamin B12 1 mcg, folic acid 70 mcg, vitamin C 25 mg, vitamin D 5 mcg, vitamin E 7 mg, copper 700 mcg, iron 8 mg, selenium 30 mcg and zinc 8 mg.

CONTROL: Placebo


OutcomesPRIMARY OUTCOMES:

Appetite of children

SECONDARY OUTCOMES:

Anthropometric measures (WAZ, HAZ, WHZ) at enrolment, 3 and 6 months

Micronutrient status (Zinc, iron, ferritiin) at enrolment, 3 and 6 months

Levels of appetite-regulating hormones (insulin, leptin) at enrolment, 3 and 6 months


NotesNo adverse events reported


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated randomization by independent person

Allocation concealment (selection bias)Unclear riskConcealment of allocation insufficiently described

Blinding of participants and personnel (performance bias)
All outcomes
Low riskStudy personnel and participants were blinded

Blinding of outcome assessment (detection bias)
All outcomes
Low riskStudy personnel were blinded for the outcome of appetite testing

Incomplete outcome data (attrition bias)
All outcomes
High risk29% dropout overall, due to death, taking ARVs, relocation, or other reasons. Differences in deaths and relocations were higher in the supplemented group, but did not reach statistical significance.

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement; study protocol not available

Other biasLow riskThe authors declare that they have no conflict of interests.

Ndeezi 2010

MethodsCountry: Uganda

Setting:Paediatric HIV clinics of the national referral hospital

Duration of trial:June 2005-June 2008

Design: Placebo-controlled trial


ParticipantsINCLUSION CRITERIA: children aged 1 - 5 years whose mothers had attended the clinic at least once, and who adhered to a regular study follow-up schedule for one year.

EXCLUSION CRITERIA: Children enrolled in other studies, those residing more than 15 kilometres from the clinic and those whose parents or caretakers were anticipating moving from the study area

Participants randomised:847

- 56% were less than 36 months

Participants analysed:695

Loss to follow-up/ withdrawal:152

Exclusions post-randomisation: 0


InterventionsMultiple micronutrient supplements (2 x RDA) for 6 months: vitamin A 800 mcg, vitamin B1 1.2 mg, vitamin B2 1.2 mg, vitamin B3 1.6 mg, vitamin B6 1.2 mg, vitamin B12 2.4 mcg, folic acid 400 mcg, vitamin C 50 mg, vitamin D 400 IU, vitamin E 14 mg, copper 800 mcg, selenium 60 mcg, zinc 10 mg, iodine 180 mcg.

CONTROL: RDA of six multivitamins: vitamin A 400 mcg, vitamin B1 0.6 mg, vitamin B2 0.6 mg, vitamin B3 0.8 mg, vitamin C 25 mg, vitamin D 200 IU


OutcomesPRIMARY OUTCOMES:Mortality at 12 months

SECONDARY OUTCOMES:Growth (weight-for-height at 12 months); CD4 counts


NotesMinor adverse effects were reported in 16 children (1.9%) and these included vomiting in 12 children and diarrhoea in four children; there was no difference between study arms.


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskThe eligible participants were block randomized to either the intervention or "standard-of-care" in two strata (ART; non-ART) using a computer-generated code

Allocation concealment (selection bias)Low riskThe randomization code was not generated by a study team member and was made available to the investigators only upon completion of data collection

Blinding of participants and personnel (performance bias)
All outcomes
Low riskAll investigators, staff and parents or caretakers were blinded to treatment assignment

Blinding of outcome assessment (detection bias)
All outcomes
Low riskStudy personnel were blinded to treatment assignment

Incomplete outcome data (attrition bias)
All outcomes
High riskBy 12 months, 124 (14.6%) children were lost to follow up: 15.7% in the intervention arm and 13.5% in the comparative arm. Those lost to follow up were more likely to be ill (underweight and CRP-positive) than those who remained.

Selective reporting (reporting bias)Low riskStudy protocol is available and all pre-specified outcomes of interest are reported on

Other biasLow riskThe authors declare that they have no competing interests.

Semba 2005

MethodsCountry: Uganda

Setting:Hospital clinic

Duration of recruitment:Jan. 1995 - June 1998

Design: Placebo-controlled trial


ParticipantsINCLUSION CRITERIA:

Children aged 6 months; resident near hospital for duration of trial

EXCLUSION CRITERIA:

Evidence of vitamin A deficiency

Participants randomised: 181 at age 15 months

- 90 M and 91 F

- mean age = 15 months

Participants analysed: 168

Loss to follow-up/ withdrawal: 0

Exclusions post-randomisation: 0


InterventionsVitamin A 200 000 IU every 3 months from 15 to 36 months

CONTROL: placebo


OutcomesPRIMARY OUTCOMES:

Mortality

SECONDARY OUTCOMES:
Morbidity (diarrhoea, cough, fever, ear discharge, hospitalisation)


NotesNo adverse events reported


Risk of bias

BiasAuthors' judgementSupport for judgement

Random sequence generation (selection bias)Low riskComputer-generated, random codes

Allocation concealment (selection bias)Low riskSequentially numbered pill cards were used

Blinding of participants and personnel (performance bias)
All outcomes
Low riskParticipants, paediatrician and clinic staff were blinded

Blinding of outcome assessment (detection bias)
All outcomes
Low riskStudy personnel were blinded to treatment assignment

Incomplete outcome data (attrition bias)
All outcomes
Low riskSurvival analysis used to account for differences in duration of follow up due to early termination of the trial

Selective reporting (reporting bias)Unclear riskInsufficient information to permit judgement; study protocol not available

Other biasHigh riskTrial stopped early due to change in national guideline on vitamin A supplementation

 
Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion

Chatterjee 2010Prospective study of children born to Tanzanian women with HIV infection within a RCT setting (Fawzi et al 1998)

CIGNIS 2010Outcomes analysed by maternal HIV status, not child HIV status

 
Comparison 1. Vitamin A

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 All-cause mortality3262Risk Ratio (M-H, Random, 95% CI)0.55 [0.37, 0.82]

 2 Child growth at 12 months1Risk Ratio (M-H, Fixed, 95% CI)Subtotals only

    2.1 Wasting
141Risk Ratio (M-H, Fixed, 95% CI)0.25 [0.06, 1.05]

    2.2 Stunting
129Risk Ratio (M-H, Fixed, 95% CI)0.67 [0.34, 1.32]

 
Comparison 2. Vitamin D

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 CD4 counts at 12 months156Mean Difference (IV, Random, 95% CI)115.0 [-74.26, 304.26]

 2 Viral load at 12 months156Mean Difference (IV, Random, 95% CI)-0.10 [-0.63, 0.43]

 
Comparison 3. Zinc

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Viral load at 9 months185Mean Difference (IV, Random, 95% CI)-0.10 [-0.40, 0.20]

 2 Mortality at 9 months196Risk Ratio (M-H, Random, 95% CI)0.31 [0.07, 1.42]

 3 CD4 % at 9 months185Mean Difference (IV, Random, 95% CI)1.0 [-2.87, 4.87]

 4 Clinic visits1Risk Ratio (M-H, Random, 95% CI)Subtotals only

    4.1 Scheduled visits; watery diarrhoea
1730Risk Ratio (M-H, Random, 95% CI)0.63 [0.39, 1.03]

    4.2 Scheduled visits; pneumonia
1730Risk Ratio (M-H, Random, 95% CI)0.85 [0.57, 1.27]

    4.3 Scheduled visits; URI
1730Risk Ratio (M-H, Random, 95% CI)0.88 [0.74, 1.04]

    4.4 Scheduled visits; ear infection
1730Risk Ratio (M-H, Random, 95% CI)0.77 [0.52, 1.14]

    4.5 All visits; watery diarrhoea
1854Risk Ratio (M-H, Random, 95% CI)0.51 [0.34, 0.76]

    4.6 All visits; URI
1854Risk Ratio (M-H, Random, 95% CI)0.75 [0.55, 1.03]

    4.7 All visits; pneumonia
1854Risk Ratio (M-H, Random, 95% CI)0.89 [0.76, 1.04]

    4.8 All visits; ear infection
1854Risk Ratio (M-H, Random, 95% CI)0.78 [0.55, 1.11]

 5 Diarrhoea incidence in HIV+ children with at least one follow-up visitOther dataNo numeric data

    5.1 Any diarrhoea
Other dataNo numeric data

    5.2 Persistent diarrhoea
Other dataNo numeric data

    5.3 Severe diarrhoea
Other dataNo numeric data

 
Analysis 3.5 Comparison 3 Zinc, Outcome 5 Diarrhoea incidence in HIV+ children with at least one follow-up visit.
Diarrhoea incidence in HIV+ children with at least one follow-up visit

StudyVitamin A (n=8)Vitamin A + Zinc (n=9)Vitamin A + Zinc + multiple supplements (n=11)

Any diarrhoea

Luabeya 20074.16 [2.40; 7.25]5.15 [3.05, 8.71]3.93 [2.36, 6.56]

Persistent diarrhoea

Luabeya 20070.050.27 (0.10, 0.71)0.39 (0.19, 0.78)

Severe diarrhoea

Luabeya 20070.050.27 (0.11, 0.66)0.39 (0.21, 0.74)

 
Comparison 4. Multiple micronutrients

Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size

 1 Mortality at 12 months1847Risk Ratio (M-H, Fixed, 95% CI)0.88 [0.52, 1.49]

 2 CD4 count1399Mean Difference (IV, Fixed, 95% CI)-36.0 [-148.53, 76.53]

 3 Duration of hospitalisation1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    3.1 All admissions
1106Mean Difference (IV, Fixed, 95% CI)-1.70 [-3.39, -0.01]

    3.2 Diarrhoea admissions
146Mean Difference (IV, Fixed, 95% CI)-1.60 [-4.35, 1.15]

    3.3 Pneumonia admissions
160Mean Difference (IV, Fixed, 95% CI)-1.90 [-4.08, 0.28]

 4 Change in appetite at 6 months199Mean Difference (IV, Fixed, 95% CI)6.1 [0.23, 11.97]

 5 Anthropometric Z scores1Mean Difference (IV, Fixed, 95% CI)Subtotals only

    5.1 WAZ at 6 months
199Mean Difference (IV, Fixed, 95% CI)0.30 [0.04, 0.56]

    5.2 HAZ at 6 months
199Mean Difference (IV, Fixed, 95% CI)-0.04 [-0.39, 0.31]

    5.3 WHZ at 6 months
199Mean Difference (IV, Fixed, 95% CI)0.45 [-0.02, 0.92]

 
Summary of findings for the main comparison. Vitamin A for children with HIV infection

Vitamin A for children with HIV infection

Patient or population: Children with HIV infection
Settings: South Africa, Tanzania, Uganda
Intervention: Vitamin A

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

ControlVitamin A

All-cause mortality
Death reports
Study populationRR 0.55
(0.37 to 0.82)
262
(3 studies)
⊕⊕⊕⊝
moderate1
Coutsoudis 1995 (SA); Fawzi 1999 (Tanzania); Semba 2005 (Uganda)

381 per 1000210 per 1000
(141 to 312)

Medium risk population

400 per 1000220 per 1000
(148 to 328)

Child growth at 12 months - Wasting
Anthropometric measurements at monthly clinic visits
Follow-up: median 351 days
Study populationRR 0.25
(0.06 to 1.05)
41
(1 study)
⊕⊕⊝⊝
low2
Fawzi 1999 (Tanzania)

368 per 100092 per 1000
(22 to 386)

Medium risk population

368 per 100092 per 1000
(22 to 386)

Child growth at 12 months - Stunting
Anthropometric measurements at monthly clinic visits
Follow-up: median 351 days
Study populationRR 0.67
(0.34 to 1.32)
29
(1 study)
⊕⊕⊝⊝
low2
Fawzi 1999 (Tanzania)

688 per 1000461 per 1000
(234 to 908)

Medium risk population

688 per 1000461 per 1000
(234 to 908)

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval; RR: Risk ratio;

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: 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.
Very low quality: We are very uncertain about the estimate.

 1 Number of events was low
2 Number of participants was very low
 
Summary of findings 2. Vitamin D for children with HIV infection

Vitamin D for children with HIV infection

Patient or population: Children with HIV infection
Settings: US hospital-based HIV treatment programmes
Intervention: Vitamin D

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

ControlVitamin D

CD4 counts at 12 months
Follow-up: mean 12 months
The mean in the control group was 661 cells/mLThe mean in the intervention group was 115 higher
(74.26 lower to 304.26 higher)
56
(1 study)
⊕⊕⊝⊝
low1
Arpadi 2009

Viral load at 12 monthsThe mean in the control group was
2.5 RNA copies/mL
The mean in the intervention group was 0.1 lower
(0.63 lower to 0.43 higher)
56
(1 study)
⊕⊕⊝⊝
low1
Arpadi 2009<BR/>

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval;

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: 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.
Very low quality: We are very uncertain about the estimate.

 1 Number of participants was very low
 
Summary of findings 3. Zinc for children with HIV infection

Zinc for children with HIV infection

Patient or population: Children with HIV infection
Settings: South African hospital outpatient clinic
Intervention: Zinc

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

ControlZinc

Viral load at 9 months
Follow-up: mean 9 months
The mean viral load at 9 months in the control groups was
5.5 RNA copies/mL
The mean Viral load at 9 months in the intervention groups was
0.1 lower

(0.4 lower to 0.2 higher)
85
(1 study)
⊕⊕⊕⊝
moderate1
Bobat 2005

Mortality at 9 months
Hospital records and verbal autopsies
Study populationRR 0.31
(0.07 to 1.42)
96
(1 study)
⊕⊕⊝⊝
low2
Bobat 2005

140 per 100043 per 1000
(10 to 199)

Medium risk population

140 per 100043 per 1000
(10 to 199)

CD4 % at 9 months

Follow-up: mean 9 months
The mean in the control group was 19 % The mean in the intervention group was 1% higher

(2.87 lower to 4.87 higher)
85
(1 study)
⊕⊕⊕⊝
moderate1
Bobat 2005

All clinic visits - watery diarrhoeaStudy populationRR 0.51
(0.34 to 0.76)
854
(1 study)
⊕⊕⊕⊝
moderate3
Bobat 2005

145 per 100074 per 1000
(49 to 110)

Medium risk population

145 per 100074 per 1000
(49 to 110)

All clinic visits - URIStudy populationRR 0.75
(0.55 to 1.03)
854
(1 study)
⊕⊕⊕⊝
moderate3
Bobat 2005

186 per 1000140 per 1000
(102 to 192)

Medium risk population

186 per 1000140 per 1000
(102 to 192)

All clinic visits - pneumoniaStudy populationRR 0.89
(0.76 to 1.04)
854
(1 study)
⊕⊕⊕⊝
moderate3
Bobat 2005

452 per 1000402 per 1000
(344 to 470)

Medium risk population

452 per 1000402 per 1000
(344 to 470)

All clinic visits - ear infectionStudy populationRR 0.78
(0.55 to 1.11)
854
(1 study)
⊕⊕⊕⊝
moderate3
Bobat 2005

145 per 1000113 per 1000
(80 to 161)

Medium risk population

145 per 1000113 per 1000
(80 to 161)

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval; RR: Risk ratio;

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: 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.
Very low quality: We are very uncertain about the estimate.

 1 Number of participants was low
2 Number of events and participants was very low
3 Number of events was low
 
Summary of findings 4. Multiple micronutrients for children with HIV infection

Multiple micronutrients for children with HIV infection

Patient or population: Children with HIV infection
Settings: Hospitals in South Africa and Uganda
Intervention: Multiple micronutrients

OutcomesIllustrative comparative risks* (95% CI)Relative effect
(95% CI)
No of Participants
(studies)
Quality of the evidence
(GRADE)
Comments

Assumed riskCorresponding risk

ControlMultiple micronutrients

Mortality at 12 monthsStudy populationRR 0.88
(0.52 to 1.49)
847
(1 study)
⊕⊕⊕⊝
moderate1
Ndeezi 2010 (Uganda)

67 per 100059 per 1000 (35 to 100)

Medium risk population

67 per 100059 per 1000 (35 to 100)

CD4 countThe mean CD4 count in the control group was 1060 cells/mLThe mean CD4 count in the intervention group was 36 lower (148.53 lower to 76.53 higher)399
(1 study)
⊕⊕⊕⊝
moderate2
Ndeezi 2010 (Uganda); subgroup of survivors at one year

Duration of hospitalisation - All admissionsThe mean duration in the control group was 9 days The mean duration in the intervention group was 1.7 lower (3.39 to 0.01 lower)106
(1 study)
⊕⊕⊕⊝
moderate2
Mda 2010a (SA)

Change in appetite at 6 monthsThe mean change in the control group was
-1.4 g per kg bodyweight eaten
The mean change in the intervention group was 6.1 higher (0.23 to 11.97 higher)99
(1 study)
⊕⊕⊕⊝
moderate2
Mda 2010b (SA)

Anthropometric Z scores - WAZ at 6 monthsThe mean WAZ in the control group was 5.7 The mean WAZ in the intervention group was 0.3 higher (0.04 to 0.56 higher)99
(1 study)
⊕⊕⊕⊝
moderate2
Mda 2010b (SA)

Anthropometric Z scores - HAZ at 6 monthsThe mean HAZ in the control group was 0.18The mean Anthropometric Z scores - HAZ at 6 months in the intervention groups was 0.04 lower (0.39 lower to 0.31 higher)99
(1 study)
⊕⊕⊕⊝
moderate2
Mda 2010b (SA)

Anthropometric Z scores - WHZ at 6 monthsThe mean WHZ in the control group was 0.08The mean Anthropometric Z scores - WHZ at 6 months in the intervention groups was 0.45 higher (0.02 lower to 0.92 higher)99
(1 study)
⊕⊕⊕⊝
moderate2
Mda 2010b (SA)

*The basis for the assumed risk (e.g. the median control group risk across studies) is provided in footnotes. The corresponding risk (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).
CI: Confidence interval; RR: Risk ratio;

GRADE Working Group grades of evidence
High quality: Further research is very unlikely to change our confidence in the estimate of effect.
Moderate quality: Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Low quality: 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.
Very low quality: We are very uncertain about the estimate.

 1 Number of events was low
2 Number of participants was low