Table 2

GRADE assessment

Certainty assessmentEffect estimate
(95% CI)
Certainty
Number of studiesStudy designRisk of biasInconsistencyIndirectnessImprecisionOther considerations
Mortality
8Observational studiesNot seriousNot seriousNot seriousNot seriousPublication bias strongly suspected* all plausible residual confounding would reduce the demonstrated effectRR 0.72
(0.55 to 0.95)
LOW
  • Meta-regression analysis.

  • *Asymmetrical funnel plot.

  • RR, relative risk.