Table 1 Randomized controlled trials included in the systematic review
Quality Assessment | Summary of Findings | GRADE Rating | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|
Treatment Groups (sample size) | Effect | |||||||||
Ref | Design | Indirectness | Inconsistency | Imprecision | Study Limitations (risk of bias) | Intervention Cases | Control Comparison | Relative Risk (95% CI) | Absolute | |
Sattar, 2012 [19] | RCT, double-blind | Pop: children 6-59 mo with SAM and diarrhea and/or ALRI | SAM def: WHZ < -3 or bipedal edema | No serious imprecision | No differences in baseline characteristics | High-dose and daily low-dose VA (n = 130) | Daily low-dose VA (n = 130) | MOD-HIGH | ||
Aim: test the efficacy and safety of WHO protocol for VAS in treatment of SAM among children 6-59 mo | VA dose: day 1: high-dose group, 200,000 IU (> 12 mo) or 100,000 IU (< 12 mo); low-dose group, placebo; day 2-15, 5000 IU daily for 15 days (both groups) | External validity: all children had diarrhea and low mean baseline serum retinol 13.15 ± 9.28 ug/dl | Resolution by 48 hr of acute diarrhea, dysentery, and ALRI | NS | ||||||
Secondary outcome examined and found no increased risk of adverse events VA toxicity and morbidities. | Duration of acute diarrhea, invasive diarrhea, cough, fever, rales, edema, and skin changes | NS | ||||||||
Changes in nutritional status, weight, length, MUAC, head circumference | NS | |||||||||
Incidence of nosocomial diarrhea | aOR 1.25 (0.67, 2.34) | NS | ||||||||
and ALRI infections | aOR 0.63 (0.36-1.09) | NS | ||||||||
Donnen, 2007 [33] | RCT, double-blind | Pop: 44-48% WHZ < -2; 30-37% MUAC < 125 mm | SAM def: WHZ < -2; MUAC < 125 mm | No serious imprecision | Age confounding: high-dose group younger | Daily low-dose VA (n = 610) | Single high-dose VA and placebo daily (n = 604) | MOD | ||
Aim: assess effect of single high-dose vs. daily low-dose VA supplements on hospitalized malnourished Senegalese children’s (0 - 14 yr) morbidity | VA dose: high-dose: single 200,000 IU (> 12 mo) or 100,000 IU (< 12 mo); low-dose: 5000 IU daily until discharge | Baseline nutrition: MUAC lower in high-dose group | Incidence of respiratory disease | HR 0.26 (0.07 - 0.92) | p < 0.05 | |||||
Sub-group: children with edema mortality lower in low-dose (AOR 0.21, 0.05-0.99) | Duration: respiratory disease | HR of cure 1.41 (1.05 - 1.89) | p = 0.019 | |||||||
Incidence & duration of diarrhea | HR 1.02 (0.68 - 1.52) | NS | ||||||||
Mahalan-abis, 2004 [32] | RCT, factorial | Pop: children with marasmus or edema excluded | No serious inconsistency | Small sample | Baseline nutrition: serum retinol was 0.387 μmol/L higher (P = 0.001) in VA treated group | zinc + VA (n = 38) zinc (n = 39) VA (n = 38) | Placebo (n = 38) | LOW-MOD | ||
Aim: evaluate effect of zinc and VA on clinical recovery of Indian children (2 - 24 mo) with severe acute lower respiratory infections | VA dose: 10,000 μg RE 2x/d for 4d; zinc 10 mg 2x/d for 5d | Recovery rates from very ill status zinc (boys) | 2.63 (1.35 - 5.10) | |||||||
Recovery rate from fever zinc | 3.12 (1.47 - 6.6) | NS | ||||||||
Recovery rates for VA | ||||||||||
Adverse event risk increased risk of diarrhea in VA group | p = 0.028 | |||||||||
increased risk of any adverse events in VA group | RR 3.8 (1.32-10.93) | |||||||||
Fawzi, 2000 [31] | RCT | VA dose: given on days 0 and 2, and again at 4 and 8 mo; dose after discharge, 200,000 IU (> 1 yr) or 100,000 IU (infants) | SAM def: NCHS references; WHZ < -2 | Statistics: few variables studied in multivariate analyses | Potential seasonal bias: enrollment during 1993-1997 | VA (n = 289) | Placebo (n = 285) | LOW-MOD | ||
Aim: determine effect of VAS on risk of diarrhea and acute respiratory infections in hospitalized Tanzanian children (6 - 60 mo) | Sub-group: adverse event risk normally nourished children: VA increased risk of acute diarrhea; wasted children protective against diarrhea | Diarrhea def: mother’s perception; 3+ days elapsed between episode; persistent diarrhea defined as lasting 14+ days | Severe watery diarrhea | AOR 0.56 (0.32 - 0.99) | ||||||
Hospitalization | NS | |||||||||
Adverse event risk cough and rapid respiratory rate | AOR 1.67 (1.17 - 2.36) | |||||||||
Faruque, 1999 [29] | RCT, factorial | Pop: 34-35% WAZ < -2; acute diarrhea ≤ 3 days | Diarrhea def: mother’s perception with picture | No serious imprecision | No differences observed by stunting at baseline | Group A: VA 15d (n = 170) Group B: zinc (n = 172) Group C: VA + zinc (n = 171) | Placebo (n = 171) | MOD | ||
Aim: to assess efficacy of zinc or VAS in Bangladeshi children (6 mo - 2 yr) with acute diarrhea | VA dose: 4500 μg RE for 15d | Children observed in hospital 24 h and followed at home 15d | Zinc groups Diarrhea duration reduced by 13 % | p = 0.03 | ||||||
Zinc dose: 14.2 - 40 mg zinc for 15d | Prolonged diarrhea reduced by 43% | p = 0.017 | ||||||||
VA groups Prolonged diarrhea | NS | |||||||||
Overall diarrhea duration | NS | |||||||||
Donnen, AJCN 1998 [26] | RCT, double-blind | Pop: 26-28% WHZ < -2; 20-29% MUAC < 125 mm | SAM def: WHZ < -2, MUAC < 125 mm | Small effect size | Baseline CRP elevated in high dose group | Daily low-dose VA (n = 298) | High-dose VA (n = 300) | MOD | ||
Aim: assess effect of single high-dose vs. daily low-dose VA supplements on hospitalized pre-school aged children in Democratic Republic of Congo | VA dose: high-dose: single 200,000 IU (> 12 mo) or 100,000 IU (< 12 mo) at admission; low-dose: 5000 IU daily until discharge | Sub-group: SAM (low dose VA vs. placebo) Incidence of severe diarrhea in SAM | Placebo (n = 302) | RR 0.21(0.07 - 0.62) | ||||||
Adverse event risk Sub-group: children with no edema increased risk of diarrhea with high-dose VA compared to placebo (AOR 2.42, 1.15 - 5.11) | Duration of: severe diarrhea | NS | ||||||||
Acute lower respiratory infection | NS | |||||||||
Fever | NS | |||||||||
Donnen, J of Nutr. 1998 [20] | RCT | VA dose: 60 mg oily solution of retinyl palmitate (30 mg for children < 12 mo) | SAM def: WHZ < -2, MUAC < 125 mm; NCHS standards; discharged children | Failure to blind: no placebo used | Group 1: VA baseline and at 6 mo (n = 118) | Group 3: placebo (n = 117) | MOD | |||
Aim: assess effect of single high-dose VA supplements and regular antiparasitic therapy on growth of moderately malnourished children (0 - 72 mo) from eastern Zaire | Pop: 4-9% WHZ < -2 | Unclear why changes in Z-scores not compared | Group 2: Mebendazole every 3 mo for 1 yr (n = 123) | |||||||
Sub-group: VA boys gain more weight and height | Group 1: Annual weight gain 2.09 kg | Group 3: Annual weight gain 1.18 kg | p = 0.029 | |||||||
MUAC gain 2.24 cm | MUAC gain 0.95 cm | p = 0.012 | ||||||||
Group 2 vs. 3 Growth | NS | |||||||||
Adverse risk: VA girls gained less height than controls. De-wormed boys and girls gained less weight than control boys and girls | ||||||||||
Stephen-sen, 1998 [28] | RCT, double-blinded | VA dose: children ≤ 1 yr: 100,000 IU on day 1 and 50,000 IU on day 2; children > 1 yr: 200,000 IU on day 1 and 100,000 IU on day 2 | SAM def: US reference population | Small sample | Potential confounding: WHZ lower in VA group at baseline | VA (n = 48) | Placebo (n = 49) | LOW-MOD | ||
Aim: test hypothesis that high-dose VA supplements will enhance recovery of Peruvian children (3 mo - 10 yrs) hospitalized with pneumonia | Statistics: univariate comparisons reported primarily without adjusting for covariates | Selective reporting | Adverse events: ↓ blood oxygen saturation | Adverse events: ↑blood oxygen saturation | p < 0.05 | |||||
Prevalence of retractions 37% | Prevalence of retractions 15% | |||||||||
Auscultatory evidence of consolidation 28% | Auscultatory evidence of consolidation 17% | |||||||||
supplemental oxygen need 21% | supplemental oxygen need 8% | |||||||||
Duration of hospitalization | NS | |||||||||
Chest x-rays | NS | |||||||||
Nacul, 1997 [24] | RCT, double-blind placebo | VA dose: 200,000 IU (< 12 mo); 400 000 IU (> 12 mo) | Pop: some hospitalized while others outpatient | No serious imprecision | Low serum retinol at baseline in both groups likely due to infection; serum retinol in VA group higher on day 11 | VA (n = 239) | Placebo (n = 233) | MOD-HIGH | ||
Aim: evaluate impact of VAS on clinical recovery and severity of pneumonia among Brazilian children (6 - 59 mo) | Subgroup analysis showed beneficial effect on those most severely affected with VA deficiency | Fever by day 3: 16 % | Fever by day 3: 26.4 % | p = 0.008 | ||||||
Transitory bulging fontanelle (4%) | Failure to respond to first line antibiotic | rate ratio 0.71 (0.50 - 1.01) | p = 0.054 | |||||||
No effect attributable to etiological agent | Duration of pneumonia | NS | ||||||||
Incidence of adverse outcomes | NS | |||||||||
Julien, 1999 [30] | RCT, double-blind placebo | VA dose: 200,000 IU (>12 mo); 100,000 IU (< 12 mo) | Small sample size to detect morbidity differences | VA deficiency high at baseline in both groups: 68.9% had serum retinol <10 μg/dL and 93.2% serum retinol < 20 μg/dL | VA (n = 71) | Placebo (n = 93) | LOW-MOD | |||
Aim: test the potential of VAS at admission to speed up recovery during hospitalization for lower respiratory infection and decrease morbidities at 6 wk in Mozambican children (6 - 72 mo) | Pop: kwashiorkor and marasmus excluded | Rate of clinical discharge on day 5: 88.4% | Rate of clinical discharge on day 5: 73.9% | RR 1.9 (1.01 - 3.05) | NS | |||||
No adjustment with covariates | 6 wk Health care use Illness (fever, cough, or diarrhea) | NS | ||||||||
Hossain, 1998 [27] | RCT, double-blind | Pop: excluded if WAZ ≤75% NCHS | Diarrhea def: liquid stools that can be poured or contain blood or mucus; clinical cure: ≤3 formed stools without visible blood or mucus | Small sample size | Baseline VA status not determined, but population prevalence high | VA (n = 42) | Control (n = 41) | LOW-MOD | ||
Aim: evaluate efficacy of single oral dose VA in treating shigellosis among Bangladeshi children (1 - 7 yr) | No adjustment with covariates | Clinical cure 19/24 (45%) | Clinical cure 8/14 (20%) | risk ratio 0.68 (0.50 - 0.93) | p = 0.02 | |||||
Bacteriological cure 16/42 (38%) | ||||||||||
Bacteriological cure 16/41 (39%) | risk ratio 0.98 (0.70 - 1.39) | NS | ||||||||
Si, 1997 [25] | RCT, double-blind | Pop: SAM excluded; 45% moderately underweight | Nutrition def: moderate malnutrition defined as WAZ 60-80% of NCHS median | No confidence intervals provided | Potential selection bias: trend for greater pneumonia severity in VA group (p = 0.06) | VA (n = 280) | Placebo (n = 312) | NS | LOW-MOD | |
Aim: evaluate effect of high dose VA supplement on morbidity in hospitalized Vietnamese children (1 - 59 mo) with pneumonia | VA dose: 200,000 in peanut oil (< 1 yr); 400,000 IU in peanut oil (1 - 4 yr) | No baseline VA status | Mean time to normal: fever respiratory rate | NS | ||||||
Sub-group analysis: moderately malnourished girls showed shorter duration of hospitalization | No adjustment for covariates | Duration of hospitalization | NS | |||||||
Dibley, 1995 [23] | RCT, double-blind | Pop: 3.4% wasted; 37% stunted | Morbidity def: 3+ loose stools per day; time between episodes 2+ diarrhea-free days | Small effect size for acute respiratory infection; large CI for acute lower respiratory infection | Baseline difference in immunization status: ever vaccinated for polio and measles higher in VA group | VA (n = 396) | Placebo (n = 386) | MOD-HIGH | ||
Aim: test efficacy of high-dose VA supplement on acute respiratory and diarrheal illnesses in Indonesian children (6 - 47 mo) | VA dose: 103,000 IU (< 1 yr); 206,000 IU (≥ 1 yr) | Acute respiratory episode with 2+ adjoining days for cough; time between episodes of 3+ symptom-free days | Adverse event risks Acute respiratory infection | rate ratio 1.08 (1.01 - 1.19) | ||||||
Adverse event risk: Subgroup analysis showed VA increased diarrhea incidence in children < 30 mo | Acute lower respiratory infection | rate ratio 1.39 (1.00 - 1.93) | ||||||||
Diarrhea | NS | |||||||||
Coutsou-dis, 1991 [22] | RCT, double blind | VA dose: 54.5 mg (< 12 mo); 109 mg (>12 mo) | Nut def: WAZ by NCHS | Small sample size | Losses to follow-up were 20% at 6 weeks and 40% at 6 months; no characteristics of this group provided | VA (n = 24) | Placebo (n = 24) | MOD | ||
Aim: test efficacy of VAS on measles morbidity in South African children (4 - 24 mo) | Morbidity def: diarrhea: 3+ loose or watery stools per day; pneumonia: presence of tachypnea with retractions, crackles, wheezes | Statistics: β set to 0.5 only to detect 30% increase in absolute recovery | No adjustment for covariates; only univariate analyses | IMS score 0.24 ± 0.15 | IMS score 1.37 ± 0.40 | p = 0.037 | ||||
Integrated morbidity score (IMS) used | Weight gain at 6 wk 1.29 ± 0.17 kg | Weight gain at 6 wk 0.90 ± 0.14 kg | P = 0.04 | |||||||
Pneumonia recovery 3.8 ± 0.40 d | Pneumonia recovery 5.7 ± 0.79 d | P < 0.05 | ||||||||
Hussey, 1990 [21] | RCT, double-blind | Pop: 70/189 (37%) with WHZ < 5th% | Nutrition def: percentiles of NCHS standards | Rare events: mortality; pneumonia / diarrhea duration ≥10 days | Differences in baseline characteristics except rash duration and lower levels of total protein and albumin in placebo | VA (n = 92) | Placebo (n = 97) | MOD-HIGH | ||
Aim: test efficacy of high dose VAS on measles complications among hospitalized South African children | VA dose: 400,000 IU at admission | High prevalence of “hyporetinemia” (< 7.0 μmol /L) 76% | Pneumonia recovery 6.3 d | Pneumonia recovery 12.4 d | P < 0.001 | |||||
No adjustment for covariates | Diarrhea recovery 5.6 d | Diarrhea recovery 8.5 d | P < 0.001 | |||||||
Croup 13 cases | Croup 27 cases | 0.51 (0.28 - 0.92) | P = 0.01 | |||||||
Days in hospital 10.6 d | Days in hospital 14.8 d | 0.21 (0.05 - 0.94) | ||||||||
Mortality 2 deaths | Mortality 10 deaths | |||||||||
Risk of death due to major complication | RR 0.51 (0.35, 0.74) | |||||||||