Sensitivity analysis of chondroitin for pain at short term
| Supplement | Variable | Number of studies (patients) | SMD (95% CI) | I², % | Meta-regression | |
| ES difference | p Value | |||||
| Chondroitin* | ||||||
| Randomisation | ||||||
| Low risk | 5 (1434) | −0.2 (−0.3 to −0.1) | 0 | |||
| Unclear | 4 (388) | −0.6 (−0.9 to −0.3) | 40 | −0.4 (−0.6 to −0.1) | 0.005 | |
| Allocation concealment | ||||||
| Low risk | 2 (417) | −0.1 (−0.3 to 0.1) | 0 | |||
| Unclear | 7 (1405) | −0.4 (−0.6 to −0.2) | 51 | −0.3 (−0.6 to 0.1) | 0.11 | |
| Incomplete outcome | ||||||
| Low risk | 7 (1073) | −0.3 (−0.4 to −0.1) | 20 | |||
| Unclear | 2 (749) | −0.6 (−1.2 to −0.1) | 83 | −0.2 (−0.6 to 0.1) | 0.20 | |
| Other bias (funding) | ||||||
| Unclear | 6 (850) | −0.5 (−0.6 to −0.3) | 34 | |||
| High risk | 3 (972) | −0.2 (−0.3 to −0.0) | 0 | 0.3 (0.1 to 0.6) | 0.01 | |
| Study sample size | ||||||
| Large | 5 (1521) | −0.2 (−0.3 to −0.1) | 0 | |||
| Small | 4 (301) | −0.6 (−0.9 to −0.2) | 55 | −0.4 (−0.7 to −0.1) | 0.01 | |
*Blinding was all unclear, selective reporting was all low risk, OA site investigated was all knee OA.
ES, effect size.; I², measures of heterogeneity; OA, osteoarthritis; SMD, standardised mean difference.