Influence of zinc supplementation in acute diarrhea differs by the isolated organism

Abstract

Zinc supplementation is recommended in all acute diarrheas in children from developing countries. We aimed to assess whether zinc supplementation would be equally effective against all the common organisms associated with acute diarrheas. We used data on 801 children with acute diarrhea recruited in a randomized, double blind controlled trial (ISRCTN85071383) of zinc and copper supplementation. Using prespecified subgroup analyses, multidimensionality reduction analyses, tests of heterogeneity, and stepwise logistic regression for tests of interactions, we found that the influence of zinc on the risk of diarrhea for more than 3 days depended on the isolated organism-beneficial in Klebsiella, neutral in Esherichia coli and parasitic infections, and detrimental in rotavirus coinfections. Although we found similar results for the outcome of high stool volume, the results did not reach statistical significance. Our findings suggest that the current strategy of zinc supplementation in all cases of acute diarrheas in children may need appropriate fine tuning to optimize the therapeutic benefit based on the causative organism, but further studies need to confirm and extend our findings.

Figure 1

Investigation into the interaction of isolated organisms with the effect of zinc on two diarrhea-related outcomes—risk of continued diarrhea for 3 or more days (left column) and high stool volume (right column) based on the results from multifactor dimensionality reduction (MDR) analyses. Results show the proportion of subjects with (red and magenta bars) and without (blue and green bars) the indicated outcome for a given combination of zinc supplementation and isolated organisms. The background of each cell representing a specific combination is gradient, coded as shown in the key at the bottom of the MDR grid. Numbers at the top of the bars are proportions and numbers in yellow are cell identifiers.

Figure 2

Statistical evaluation of the bivariate interactions between zinc supplementation and each of the four major isolates for two diarrhea-related outcomes—risk of continued diarrhea for 3 or more days (left column) and high stool volume (right column). The interaction was quantified using the interaction contrast statistic [14], the significance was tested using the Breslow-Day (B-D P) test for heterogeneity, and statistical power of interpretation in the present study was determined using the θ statistic described by Gönen [14]. Interaction contrast exceeding or below zero indicates a detrimental or protective interaction, respectively.

Figure 3

Multivariate association of main and interactive effects of zinc supplementation and diarrheal isolates using stepwise logistic regression models for two diarrhea-related outcomes—risk of continued diarrhea for 3 or more days (left column) and high stool volume (right column). In this grid, yellow background indicates the main effects, black circle indicates the included covariate, the blue background indicates interactive effects, and black circles in the blue part of the grid indicate the interactions among the circled covariates. Z: zinc supplementation; E: Escherichia coli; K: Klebsiella; and R: rotavirus. Diamonds and error bars indicate the point and 95% confidence interval estimates for the log odds ratio (Log OR). Results are from full model (on the left) and final model (on the right) from the stepwise logistic regression analyses. The full model included all the 14 covariates indicated in the grid while the stepwise regression analyses used a backward elimination strategy with a probability retention criterion of 0.2. Numbers at the right-hand side in the final model plots are significance values for the covariates that were retained in the final model.