Oral zinc for treating diarrhoea in children

Randomized controlled trials in children aged 1 month to 5 years with acute or persistent diarrhoea (including dysentery); settings included hospital and community; many trials double-blind; total 33 trials with 10,841 children.

Oral zinc supplementation (zinc sulfate, zinc acetate, or zinc gluconate) taken by mouth; typical daily dose around 20 mg (range from 5 mg/day to >20 mg/day in some trials); duration commonly about 14 days (half of trials), with some shorter (4 days) or longer (up to 14 days); dosing mostly once daily (some trials used twice or thrice daily).

Zinc supplementation reduces diarrhoea duration and persistence in several subgroups. In acute diarrhoea, zinc reduces mean duration by about 13.4 hours overall (MD −13.42 hours; 95% CI −20.52 to −6.31; 5096 children, 20 trials), with substantial heterogeneity (I2 = 84%). In older than six months, duration shortened by around half a day; in malnourished children, duration reduced by about a day; in infants under six months, little or no effect on duration and possible increased persistence to day seven. Vomiting was more common with zinc in both age groups (moderate certainty). In persistent diarrhoea, zinc probably shortens duration by about 16 hours (MD −15.84 hours; 95% CI −25.43 to −6.24; 529 children, 5 trials). There is no clear evidence that zinc reduces death or hospitalization. Conclusion: In areas with high zinc deficiency or malnutrition, zinc may benefit children aged six months or older; in well-nourished or low-deficiency settings, benefits are unclear. More research is needed to clarify mortality/hospitalization effects and to explain heterogeneity.

High heterogeneity across trials with varying populations, settings, dosing regimens, and co-supplementation; risk-of-bias concerns in some trials; most trials conducted in hospital settings and in populations with higher zinc deficiency; limited data on mortality/hospitalization; potential publication bias suggested by funnel plots; results not fully generalizable to low-deficiency populations.