GRADE evidence profile for PICO 2
| Certainty assessment | № of patients | Effect | Certainty | Importance | ||||||||
| № of studies | Study design | Risk of bias | Inconsistency | Indirectness | Imprecision | Other considerations | EVT plus BMT | BMT alone | Relative (95% CI) | Absolute (95% CI) | ||
| mRS 0–3 at 90 days: RCT | ||||||||||||
| 3 | Randomized trials | Serious* | Serious† | Serious‡§¶ | Serious** | None | 200/446 (44.8%) | 102/325 (31.4%) | RR 1.45 (1.03 to 2.04) | 141 more per 1000 (from 9 more to 326 more) | ⨁◯◯◯ Very low†† | CRITICAL |
| mRS 0–2 at 90 days: RCT | ||||||||||||
| 3 | Randomized trials | Serious* | Serious‡‡ | Serious‡§¶ | Serious** | None | 151/446 (33.9%) | 74/325 (22.8%) | RR 1.59 (0.89 to 2.86) | 134 more per 1000 (from 25 fewer to 424 more) | ⨁◯◯◯ Very low§§ | CRITICAL |
| Shift (ordinal) mRS at 90 days: RCT | ||||||||||||
| 3 | Randomized trials | Serious* | Serious¶¶ | Serious‡§¶ | Serious** | None | – | – | OR 1.81 (1.06 to 3.08) | Two fewer per 1000 (from 3 fewer to one fewer) | ⨁◯◯◯ Very low*** | CRITICAL |
| Mortality at 90 days: RCT | ||||||||||||
| 3 | Randomized trials | Serious* | Serious* | Serious‡§¶ | Serious** | None | 164/446 (36.8%) | 151/325 (46.5%) | RR 0.78 (0.63 to 0.95) | 102 fewer per 1000 (from 172 fewer to 23 fewer) | ⨁◯◯◯ Very low | IMPORTANT |
| Symptomatic Intracranial Hemorrhage (sICH): RCT | ||||||||||||
| 3 | Randomized trials | Serious* | Not serious | Serious‡§¶ | Serious** | Very strong association | 24/446 (5.4%) | 1/325 (0.3%) | RR 8.91 (2.10 to 37.89) | 24 more per 1000 (from 3 more to 114 more) | ⨁⨁⨁◯ Moderate | IMPORTANT |
| mRS 0–3 NRSI | ||||||||||||
| 3 | Non-randomized studies | Serious‡‡‡ | Serious§§§ | Serious¶¶¶ | Serious**** | None | – | – | RR 0.86 (0.47 to 1.59) | One fewer per 1000 (from 2 fewer to 0 fewer) | ⨁◯◯◯ Very low††† | CRITICAL |
*Serious risk of bias arising from the deviations from intended intervention in all RCTs, high risk of performance bias. Some concerns in other domains.
†I2 statistic, which quantifies the proportion of the variation in point estimates due to among-study differences was 64%, assessed as substantially high.
‡Enrolled patients had severe/very severe symptoms. Patients with mild-to-moderate symptoms were missing or underrepresented.
§Comparator not the same in the trials; it differs by proportion of IVT in the BMT arms and by timing of IVT administration.
¶Time window 6 h only in 1 trial, whereas 8 h and 12 h in the other 2 trials.
**Serious imprecision due to low optimal information size. The total number of patients included is less than the number of patients generated by a conventional size sample calculation for a single adequately powered clinical trial.
††P-value for interaction between trials with high (European trial) and low (Asian trials) proportion of IVT in BMT arms (0.02).
‡‡I2 statistic, which quantifies the proportion of the variation in point estimates due to among-study differences was 79%, assessed as substantially high.
§§P-value for interaction between trials with high (European trial) and low (Asian trials) proportion of IVT in BMT arms (0.003).
¶¶I2 statistic, which quantifies the proportion of the variation in point estimates due to among-study differences was 65%, assessed as substantially high.
***P-value for interaction between trials with high (European trial) and low (Asian trials) proportion of IVT in BMT arms (0.03).
†††P-value for interaction between registry studies with high (European) and low (Asian) proportion of IVT in BMT arms (<0.00001).
‡‡‡Serious risk of bias due to serious confounding reported in one of these studies implemented for this outcome according to ROBINS-I tool for observational studies.
§§§Presence of heterogeneity.
¶¶¶Raw data was not available in one study; generic inverse variance meta-analysis of the reported RRs in the studies was performed.
****Inconclusive confidence interval.
.BMT, give details; CI, confidence interval; EVT, give details; mRS, modified Rankin Scale; OR, odds ratio; RCT, randomized controlled trial; RR, risk ratio.