Table 3 GRADE evidence quality for each outcome
| Quality assessment | Summary of findings | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Number of studies | Design | Quality | Consistency | Directness | Other modifying factors | No. of patients | Summary | Quality | |
| Delayed | Early | ||||||||
| Concomitant meniscus tear | |||||||||
| 18 |
RCT: 7 Non-RCT: 11 | Very serious limitations (−2) |
Important inconsistency (−1) | Some uncertainty (−1) | Evidence of a dose-response gradient (+ 1) | 1308 | 1062 | The incidence of meniscus tear in delayed group was higher than in early group. Only 3 studies reported higher incidence of meniscus tear in early group | Low |
| Lysholm score | |||||||||
| 17 |
RCT: 4 Non-RCT: 13 | Very serious limitations (−2) | No important inconsistency | Some uncertainty (−1) |
Imprecise or sparse data (− 1). Evidence of a dose-response gradient (+ 1) | 402 | 455 | The Lysholm scores decreased in delayed surgery group. Only 2 studies reported higher scores in delayed surgery group | Low |
| Tegner score | |||||||||
| 15 |
RCT: 8 Non-RCT: 7 | Very serious limitations (−2) | No important inconsistency | Some uncertainty (−1) | None | 496 | 524 | The Tegner scores decreased in delayed surgery group, but those for delayed surgery group in MLKI were marginal | Low |
| Concomitant cartilage injury | |||||||||
| 10 |
RCT: 5 Non-RCT: 5 | Very serious limitations (−2) |
Important inconsistency (−1) | Some uncertainty (− 1) |
Imprecise or sparse data (− 1) | 673 | 1008 | The incidence of cartilage injury in delayed group was higher than in early group. Only 1 study of MLKI was included for this meta-analysis. | Very low |