Outcome | Illustrative comparative risks^{*} (95% CI) | Relative effect (95% CI) | No. of participants (studies) | GRADE quality of evidence^{†} | |
---|---|---|---|---|---|

Assumed risk for control group | Corresponding risk with zinc | ||||

Duration of symptoms | 1.65 days lower (2.5 to 0.81 days lower) | 934 (8) | Moderate^{a},^{b},^{c} | ||

Severity of symptoms | 0.27 SDs lower (0.58 lower to 0.05 higher) | 412 (4) | Low^{d},^{e} | ||

No. of symptomatic patients after 3 d of treatment | 858 per 1000 | 789 per 1000 (712 to 875 per 1000) | RR 0.92 (0.83 to 1.02) | 1252 (8) | Low^{f},^{g} |

No. of symptomatic patients after 7 d of treatment | 471 per 1000 | 297 per 1000 (207 to 424 per 1000) | RR 0.63 (0.44 to 0.9) | 1325 (9) | Low^{h},^{i} |

Adverse events leading to stopping treatment | 0 per 1000 | 0 per 1000 (0 to 0 per 1000) | RR 11 (0.62 to 193.8) | 230 (2) | Low^{j},^{k} |

Adverse event | |||||

Any | 385 per 1000 | 477 per 1000 (404 to 562 per 1000) | RR 1.24 (1.05 to 1.46) | 1487 (9) | Moderate^{l},^{m} |

Bad taste | 204 per 1000 | 337 per 1000 (259 to 441 per 1000) | RR 1.65 (1.27 to 2.16) | 961 (8) | Moderate^{n},^{o} |

Nausea | 102 per 1000 | 167 per 1000 (121 to 232 per 1000) | RR 1.64 (1.19 to 2.27) | 973 (9) | Moderate^{p},^{q} |

Abdominal pain | 94 per 1000 | 112 per 1000 (78 to 162 per 1000) | RR 1.19 (0.83 to 1.72) | 876 (7) | Moderate^{r} |

Diarrhea | 29 per 1000 | 55 per 1000 (28 to 108 per 1000) | RR 1.88 (0.95 to 3.72) | 831 (7) | Moderate^{s} |

Constipation | 23 per 1000 | 33 per 1000 (15 to 72 per 1000) | RR 1.42 (0.64 to 3.12) | 876 (7) | Moderate^{t} |

Note: CI = confidence interval, GRADE = Grading of Recommendations Assessment, Development and Evaluation, RR = risk ratio, SD = standard deviation.

↵* The basis for the assumed risk (e.g., the median risk for the control group across studies) is provided in footnotes. The corresponding risk (and its 95% CI) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI).

↵† GRADE Working Group grades of evidence: high quality: further research is very unlikely to change our confidence in the estimate of effect; moderate quality: further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate; low quality: further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate; very low quality: we are very uncertain about the estimate.

↵a No serious design limitations: Blinding was adequate in all trials. Two trials had unclear allocation concealment.31,35 One trial did not describe random sequence generation,35 and two trials had unclear selective reporting bias.30,35 One trial had incomplete outcome data.30 Other bias was unclear in three trials.30,34,35 Sensitivity analysis excluding these trials did not change the results, so the evidence was not downgraded.

↵b Serious inconsistency: Very high statistical heterogeneity (

*I*^{2}= 95%). Age, ionic zinc dose and zinc formulation partially accounted for between-study variation.↵c No serious imprecision: Cumulative sample size was appropriate. The optimal information size to detect a one-day difference in duration of symptoms (α = 0.05, 90% power) assuming a mean of 7 days (SD 3 days) was 190 participants per arm. The 95% CI (2.50 to 0.81) crossed the minimally important difference of one day. However, the CI was narrow and did not include “no treatment effect.”

↵d Serious inconsistency: Substantial heterogeneity (

*I*^{2}= 55%).↵e Serious imprecision: Total sample size 412. The optimal information size to detect a 1-point difference in score for severity of symptoms (α = 0.05, 80% power) assuming a mean score of 3 (SD 4) was 252 participants per arm.

↵f Serious design limitations: Five trials had significant design limitations (all had incomplete outcome data, unclear allocation concealment and did not report the method of randomization).28,38,24

^{[A,B]},39 The remaining three trials had no significant limitations.↵g Serious Inconsistency: High statistical heterogeneity (

*I*^{2}= 80%) not explained by subgroup analyses.↵h Serious design limitations: Six trials had significant design limitations.28,30,38,24

^{[A,B]},39 All had incomplete outcome data. Allocation concealment and method of randomization were unclear in all but one of the six.30 Other bias was present in one trial.28↵i Serious inconsistency: High statistical heterogeneity (

*I*^{2}= 78%).↵j Serious design limitations: One trial had serious design limitations.39 This trial had incomplete outcome data and was high risk for selective reporting. It also had unclear allocation concealment and did not report the method of randomization.

↵k Serious imprecision: Only two trials reported on this outcome, and one of these trials had no outcomes to report in either group;39 this resulted in a large 95% CI and small sample size.

↵l No serious limitations: Five trials had serious design problems.28,30,24

^{[A,B]},39 All five had incomplete outcome data and unclear allocation concealment, and all but one trial30 did not report the randomization method. Sensitivity analysis excluding these trials did not change the results, so the evidence was not downgraded.↵m Serious imprecision: Estimated range of adverse events from 19 more to 177 more per 1000 versus placebo.

↵n No serious design limitations: Two trials had serious design limitations.28,39 Sensitivity analysis excluding these trials did not change the results, so the evidence was not downgraded.

↵o Serious imprecision: Estimated range of “bad taste” events from 55 to 237 more per 1000 versus placebo.

↵p No serious design limitations: Three trials had significant design concerns.28,29

^{[B]},38 All had incomplete outcome data and unclear allocation concealment. The method of randomization was not reported in two trials.28,38 The remaining six trials had low risk of bias. Sensitivity analysis excluding the trials with high risk of bias did not change the results, so the evidence was not downgraded.↵q Serious imprecision: Estimated range of nausea events from 19 more to 130 more per 1000 versus placebo.

↵r Serious imprecision: Low number of events (102) and 95% CI crosses no treatment effect (1.0) and the threshold for appreciable harm (1.25).

↵s Serious imprecision: Low number of events (36) and 95% CI crosses no treatment effect (1.0) and the threshold for appreciable harm (1.25).

↵t Serious imprecision: Low number of events (28) and 95% CI crosses no treatment effect (1.0) and threshold for appreciable harm (1.25).