1. Objectives

To evaluate the effectiveness and safety of vitamin A in children with diagnosed non-measles pneumonia

2. How studies were identified

The following databases were searched between June and August 2010:

• Cochrane Acute Respiratory Infections Group’s Specialized Register
• CENTRAL (The Cochrane Library 2010, Issue 3)
• MEDLINE
• EMBASE
• LILACS
• CINAHL
• Biological Abstracts
• Current Contents
• Chinese Biomedicine Database
• WHO ICTRP

The review authors directly contacted authors of identified Chinese articles. The updated search in 2010 identified no new trials

3. Criteria for including studies in the review

3.1 Study type

Randomized controlled trials, including quasi-randomized trials

3.2 Study participants

Children less than 15 years of age with diagnosed non-measles pneumonia

(Pneumonia was defined as a clinical case definition, radiological confirmation, or both)

3.3 Interventions

Vitamin A plus standard care versus standard care with or without placebo

3.4 Primary outcomes

• Mortality

Secondary outcomes included signs of pneumonia (fever, tachypnoea, dyspnoea, chest X-ray findings), clinical severity (oxygen saturation, requirement for mechanical ventilation or supplemental oxygen, crepitations, bronchial breathing, duration of hospitalization, failure of first line treatment or change of antibiotic required), and adverse events associated with vitamin A, such as diarrhoea and signs of toxicity (vomiting, desquamation)

4. Main results

4.1 Included studies

Six randomized controlled trials, enrolling 1740 infants and children, were included in this review

• Participants’ ages ranged from one month to 14 years
• One trial was a multi-armed selenium and vitamin A trial with 150,000 IU of vitamin A administered in the vitamin A arm (only data from the vitamin A and placebo arms were used in analyses)
• In two trials, 100,000 IU of vitamin A was provided to infants <1 year and 200,000 IU was provided to children ≥1 year for two days
• In one trial, infants <1 year received 100,000 IU of vitamin A on admission to hospital and 50,000 IU on the second day, while children ≥1 year received 200,000 IU on the first day and 100,000 IU on the second day
• One trial administered 50,000 IU of vitamin A to infants <1 year and 100,000 IU to children ≥1 year
• In one trial in 80 children with recurrent bronchopneumonia, 10,000 IU of vitamin A was given twice daily for six days followed by 1500 IU for 20 days
• All trials used placebo controls

4.2 Study settings

• China (2 trials), Brazil, Ecuador, Peru, and the United Republic of Tanzania
• All trials were conducted in hospital settings, and all participants were hospitalized

4.3 Study settings

How the data were analysed
One comparison was made: vitamin A versus placebo for non-measles pneumonia. Continuous data were summarized with mean differences (MD). Dichotomous data were summarized using odds ratios (OR), with the exception of mortality data, for which Peto OR were used due to the low number of events. Results were presented with 95% confidence intervals (CI). Potential sources of heterogeneity were to be assessed using subgroup analysis or meta-regression by dose and duration of treatment, but insufficient data were available. Sensitivity analyses by trial quality, publication status, and using random versus mixed effect models were also planned but not conducted.

Results
Vitamin A versus placebo
Mortality
No statistically significant difference between vitamin A and placebo groups was found in pooled analysis of three trials involving 1446 children (OR 1.29, 95% CI [0.63 to 2.66]).

Additional outcomes
Time with cough was statistically significantly reduced among children receiving vitamin A in one trial involving 38 infants (MD -2.00 [units not specified], 95% CI [-3.51 to -0.49], p=0.0094). In a trial in 80 children with recurrent bronchopneumonia, children receiving low-dose vitamin A had a lower rate of no improvement at short-term follow-up (OR 0.06, 95% CI [0.01 to 0.30], p=0.00053) and a lower rate or recurrence at long-term follow-up (OR 0.12, 95% CI [0.03 to 0.46], p=0.002) than those in the placebo group. The odds of requiring a change of antibiotic were borderline statistically significantly reduced with vitamin A supplementation in one trial including 472 children (OR 0.65, 95% CI [0.42 to 1.01], p=0.054). In a single trial, among 52 children whose baseline serum retinol concentrations were >200 μg/L, time with fever, tachypnoea and hypoxaemia of pneumonia was significantly reduced with vitamin A supplementation (MD -61.4 hours, 95% CI [-119.1 to -3.7], p=0.037). In one study reporting on clinical severity in which children received up to 300,000 IU of vitamin A over two days, those in the vitamin A group had lower mean blood oxygen saturations and were more likely to require supplemental oxygen, had higher mean respiratory rates and heart rates, and had a higher prevalence of consolidation on auscultation over the treatment period (all p≤0.001). No statistically significant difference between treatment and control groups was found for the outcomes time with fever, time with respiratory rate >40 breaths/minute/day, time with positive finding on auscultation, time with positive findings on X-ray, duration of hospitalization, and time with hypoxia. No trials reported on the requirement for mechanical ventilation.

Adverse effects
No statistically significant difference between treatment and control groups was found for the adverse effects vomiting, diarrhoea, irritability, or nausea. In one trial, a bulging fontanelle occurred in four out of 99 children receiving vitamin A and in none of those receiving placebo; however, this difference did not reach statistical significance (OR 8.25, 95% CI [0.44 to 155.37], 186 children).

5. Additional author observations*

One of the six indentified trials was considered to be at low risk of bias and one was considered to be at high risk of bias due to quasi-randomization, with children being assigned to vitamin A or placebo based on order of hospital admission. Using GRADE quality of evidence assessment criteria, evidence for the outcomes mortality and the adverse events diarrhoea, bulging fontanelle, and irritability was rated as moderate quality, while evidence for other outcomes was rated as low to very low quality.

Current evidence suggests that vitamin A supplementation in children with non-measles pneumonia has no effect on mortality, and has no clear effect on measures of morbidity or the clinical course of pneumonia.

Few trials reported on each outcome, and thus meta-analyses may have been underpowered to detect statistically significant differences between vitamin A and placebo groups. Therefore, further large randomized controlled trials are needed in which clinically important outcomes are reported on. Different doses of vitamin A in populations stratified by risk of vitamin A deficiency could also be investigated.