GRADE summary of findings table
| Certainty assessment | No of patients | Effect | Certainty | |||||||||
| Outcome | No of studies | Study design | Risk of bias | Inconsistency | Indirectness | Imprecision | Other considerations | Surgery | Endovascular | Relative (95% CI) | Absolute (95% CI) | |
| Initial failure | 32 | rCS | Serious | Not serious | Not serious | Not serious | Possible publication bias | 627 | 753 | 0.15 (0.09–0.24) | 95 fewer per 1000 (83-107) | Low |
| Late recurrence | 22 | rCS | Serious | Not serious | Not serious | Not serious | Possible publication bias | 423 | 525 | 0.19 (0.09–0.39) | 54 fewer per 1000 (48-62) | Low |
| Neurological improvement | 13 | rCS | Serious | Serious | Not serious | Serious | Possible publication bias | 387 | 404 | 2.73 (1.67–4.48) | 130 more per 1000 (63-197) | Very low |
| Complications | 22 | rCS | Serious | Not serious | Not serious | Serious | Possible publication bias | 213 | 202 | 1.78 (0.97–3.26) | three more per 1000 (10 fewer to five more) | Very low |
| Initial failure/late recurrence (Onyx vs NBCA) | 9 | rCS | Serious | Not serious | Not serious | Serious | Possible publication bias | 58 | 75 | 3.87 (1.73–8.68) | 113 more per 1000 (77-148) | Very low |
NBCA, n-butyl 2-cyanoacrylate; rCS, retrospective case series.