Table 11

Summary of findings for PICO 13

Certainty assessmentNo of patientsEffectQuality of evidenceImportance
No of studiesStudy designRisk of biasInconsistencyIndirectnessImprecisionOther considerationsGACSRelative
(95% CI)
(95% CI)
mRS score 0–2 (dedicated RCTs)
3Randomized trialsNot seriousNot seriousSerious*Serious†None90/183 (49.2%) 65/185 (35.1%) RR=0.74
(0.54 to 1.01)
91 fewer per 1 000
(from 4 more to 162 fewer)
mRS score 0–2 (HERMES)
7Observational studiesSerious‡Not seriousNot seriousNot seriousNone94/234 (40.2%)282/561 (50.3%) OR=1.65
(1.14 to 2.38)
122 more per 1 000
(from 33 more to 204 more)
  • *We believe that the setting in which the interventions in the three RCTs were performed cannot easily be implemented with the same sophistication in daily practice in many centers.

  • †Clinical action would differ if the upper versus the lower boundary of the 95% CI of the risk ratio represents the truth: GA would be recommended if the true RR is 0.54; GA would not be recommended if the true RR is 1.01.

  • ‡Probable indication bias (confounding by indication): it is likely that patients who underwent GA had more frequently a medically required GA rather than an ’elective' GA.

  • CI, confidence interval; CS, conscious sedation; GA, general anesthesia; OR, odds ratio; RCTs, randomized controlled trials; RR, risk ratio.