Table 2.
Ratings of methodological quality for included studies
| Study | Inclusion of consecutive patients | Representativeness (multicentre adequate) | Percentage follow-up (>80% adequate) | Minimisation of potential confounding (multivariable analysis adequate) |
| Crosbie et al38 | + | − | + | + |
| Edwards et al39 | − | − | + | |
| Elson and Brenkel40 | − | − | − | |
| Grosu et al41 | − | − | − | |
| Núñez et al42 | + | − | − | + |
| Núñez et al43 | + | − | − | + |
| Phillips et al44 | + | − | + | − |
| Pinto et al*45 | + | − | − | + |
| Riis et al46 | + | − | + | + |
| Sayers et al47 | +† | − | + | + |
| Stephens et al48 | + | + | ||
| Thomazeau et al49 | + | − | + | + |
| Kocic et al50 | − | + | − | |
| Veal et al*51 | − | + | − |
*For studies where authors provided data on patients with total knee replacement, ratings are based on the study as reported in the article.
†Information obtained through personal contact.
+, adequate; −, inadequate; blank, not reported.