Fig. 2
a. Trial sequential analysis for a relative risk reduction of all-cause mortality of 5.8 % of hypothermia after cardiac arrest in five trials with 1363 patients reporting mortality. A required diversity-adjusted information size of 16,287 patients was calculated based on a control event proportion of 51.0 %, a hypothermia-induced relative risk reduction of mortality of 5.8 % suggested by all trials, α = 0.05 two-sided, β = 0.20 (power = 80 %), and diversity D2 = 60 %. The cumulated Z-curve (blue) crosses the traditional boundary (p = 0.05) but not the trial sequential monitoring boundary, indicating lack of firm evidence for a beneficial effect of 5.8 % relative risk reduction of the intervention when the analysis is adjusted for repetitive testing on accumulating data. There is insufficient information to reject or detect an intervention effect of 5.8 % relative risk reduction of all-cause mortality as the required information size is not yet reached. b. Trial sequential analysis (TSA) for a relative risk reduction of 7.29 % of hypothermia after cardiac arrest in six trials with 1409 patients reporting neurological function. A required diversity-adjusted information size of 15,568 patients was calculated based on a control event proportion of 56.9 %, a hypothermia-induced relative risk reduction of poor neurological function of 7.29 % suggested by all trials, α = 0.05 two-sided, β = 0.20 (power = 80 %), and diversity D2 = 79 %. The cumulated Z-curve (blue) crosses the traditional boundary (p = 0.05) but not the trial sequential monitoring boundary, indicating lack of firm evidence for a beneficial effect of 7.29 % relative risk reduction of the intervention when the analysis is adjusted for repetitive testing on accumulating data. There is insufficient information to reject or detect an intervention effect of 7.29 % relative risk reduction of poor neurological outcome as the required information size is not yet reached