Table 3. Exclusion of possible overlaps.

		All studies		Excluding possible overlap
Exposure	Outcome	Number of studies	Pooled estimate (RR)	Number of studies	Pooled estimate (RR)
Sex (male vs female)	Infection	41	1.14	28	1.18
	Hospitalization	22	1.33	18	1.35
	Severe disease	77	1.22	17	1.28
	ICU	48	1.41	37	1.39
	Death	91	1.35	40	1.27

Studies with possible overlap of patients were excluded from the analysis, results presented in bold. Possible overlap was assumed when studies were from the same region, recruitment period and hospital. In a group of studies with possible overlap only the largest study was included in the analysis. The estimates from the meta-analysis were very robust and did not demonstrate any substantial shifts in pooled effect estimates for any of the five study outcomes after exclusion of studies with possible overlap.

RR = relative risk; ICU=intensive care unit.