1. Objectives

To evaluate vitamin D supplementation for preventing pneumonia, tuberculosis, diarrhoea, and malaria in children less than five years of age

2. How studies were identified

The following databases were searched in June 2016:

• Cochrane Infectious Diseases Group Specialized Register
• CENTRAL (The Cochrane Library 2016, Issue 6)
• MEDLINE
• EMBASE
• LILACS
• WHO International Clinical Trials Registry Platform
• ClinicalTrials.gov
• ISRCTN Registry

Reference lists were also hand-searched and the authors directly contacted researchers in the field

3. Criteria for including studies in the review

3.1 Study type

Randomized controlled trials, including cluster-randomized trials

3.2 Study participants

Children less than five years of age

(Studies that exclusively enrolled children with pre-existing illnesses, such as HIV, rickets, severe malnutrition or sickle cell disease, were excluded, as were studies in preterm and low birth weight infants)

3.3 Interventions

Vitamin D, with or without other micronutrients, compared with placebo or no treatment, or the same micronutrients without vitamin D

3.4 Primary outcomes

• Pneumonia incidence rate (the number of cases or episodes per total child-years)
• Tuberculosis incidence rate (the number of cases or episodes per total child-years)

Secondary outcomes included incidence rates of diarrhoea, malaria, and febrile illness; duration (mean number of days of all episodes) of pneumonia and diarrhoea; severity of infection: moderate or severe pneumonia (as defined by trialists), moderate of severe diarrhoea (history of loose stools >3 times/day and history of oral dehydration or intravenous fluid therapy); all-cause mortality, cause-specific mortality (pneumonia, tuberculosis, diarrhoea, malaria); hospital admission rate due to infection; change in mean serum vitamin D concentrations; and adverse outcomes including nausea, vomiting, headache, constipation and kidney stones

4. Main results

4.1 Included studies

Four randomized controlled trials, enrolling 3198 infants, were included in this review

• Three trials supplemented infants with 400 IU vitamin D/day from birth or within 15 days of birth until 12 weeks (n=18), six months (n=46), or 12 months (n=88) of age
• One trial enrolled 3046 infants aged one to 11 months, with the intervention group receiving quarterly 100,000 IU oral vitamin D supplements

4.2 Study settings

• Afghanistan, Spain, and the United States of America (2 trials)
• Baseline vitamin D deficiency ranged from 6.2% in Spain, to 10 to 12% in the United States of America, and 73.1% in Afghanistan
• Trials were conducted in private paediatric clinics, primary healthcare centres, and a teaching hospital

4.3 Study settings

How the data were analysed
One comparison was made: vitamin D versus control for preventing infections in children under five years of age. Rate ratios with 95% confidence intervals (CI) were employed to summarize count data using the generic inverse variance method. For dichotomous data, risk ratios (RR) and corresponding 95% CI were employed, using intention-to-treat analysis results from the included trials. Continuous data were summarized using mean differences (MD) and 95% CI. Where substantial heterogeneity was detected (I²>50%), random effects models were used, in other cases, fixed effects models were employed. Subgroup and sensitivity analyses were planned, but were not conducted due to insufficient data.

Results
Vitamin D versus control
Incidence rates of pneumonia
In pooled analysis of two trials, the incidence rate of the first or only episode of pneumonia, confirmed by chest X-ray, was not different between vitamin D and control groups (rate ratio 1.06, 95% CI [0.89 to 1.26], 3134 participants). The incidence rate of the first or only episode of pneumonia, confirmed by chest X-ray or without radiological confirmation, was also similar (RR 0.95, 95% CI [0.87 to 1.04], 1 trial/3046 participants). The incidence rate of repeat episodes of pneumonia was statistically significantly increased with vitamin D supplementation in analysis of cases determined by chest X-ray (rate ratio 1.69, 95% CI [1.28 to 2.21], 1 trial/3046 participants), but not in analysis of both confirmed and unconfirmed cases (rate ratio 1.06. 95% CI [1.00 to 1.13], 1 trial/3046 participants). Tuberculosis incidence rates were not reported on in the included trials.

Additional outcomes
The incidence rate for the first or only episode of diarrhoea did not statistically significantly differ between treatment and control groups in the two trials reporting on this outcome (rate ratio 0.14, 95% CI [0.01 to 2.59], 1 trial/88 participants; hazard ratio (HR) 1.02, 95% CI [0.95 to 1.11], 1 trial/3046 participants), and nor was there a difference in the incidence rate of repeat episodes of diarrhoea (HR 1.05, 95% CI [0.98 to 1.17], 1 trial/3046 participants). The risk of all-cause mortality was not different between groups in one trial involving 3046 infants, although only 17 deaths occurred overall (RR 1.43, 95% CI [0.54 to 3.74]). Cause-specific deaths due to pneumonia and septicaemia were also not different between vitamin D and control groups (RR 1.50, 95% CI [0.42 to 5.30], 1 trial/3046 participants). No statistically significant difference in hospital admissions was found between treatment groups in either of the two trials reporting on this outcome (RR 0.86, 95% CI [0.20 to 3.62], 1 trial/88 participants; data from one trial not available). In pooled analysis of four trials, mean concentrations of serum 25-hydroxyvitamin D at the end of the intervention period were statistically significantly greater among infants randomized to vitamin D supplementation (MD 7.72 ng/mL, 95% CI [0.50 to 14.93], p=0.04; 266 participants). One child receiving 400 IU of vitamin D per day had serum 25-hydroxyvitamin D concentrations that were too high (almost 100 ng/mL) and two children receiving 100,000 IU of vitamin D quarterly had toxic concentrations (150 ng/mL).

5. Additional author observations*

Few trials were indentified for inclusion in this review, the largest of which was judged to be at low risk of bias in all areas except attrition, with approximately 30% of all participants being lost to follow-up. The methodological quality of the other three trials was mixed, with inadequate reporting of the methods of allocation concealment in two trials. Using the GRADE approach, evidence for the outcome incidence rate of first or only episode of radiologically-confirmed pneumonia was judged to be of moderate quality, while evidence for the outcomes all-cause mortality, cause-specific mortality, and mean serum vitamin D concentrations was rated as low quality, and evidence for the outcome any hospital admissions was rated as very low quality. The largest trial was conducted in an area with a high prevalence of childhood vitamin D deficiency and infection, making the findings of this review applicable to similar populations.

Vitamin D supplementation had little effect on the incidence rate of the first or only episode on pneumonia. The finding that vitamin D supplementation increased the rate of repeat episodes of pneumonia may be due to chance, but warrants further investigation. Vitamin D supplementation elevated mean serum concentrations of serum 25-hydroxyvitamin D at the end of the intervention period overall, although the effect was negligible in infants receiving a quarterly dosing regimen. No clear differences were found between treatment groups for other outcomes, although the number of deaths was too low to confirm or refute any effect of vitamin D supplementation.

While further trials in this area are warranted, the largest trial was conducted in a vitamin D-deficient area with the greatest potential for improvements in infection rates, and yet no benefit of supplementation was found.