Table A3.
Post Hoc Subgroup Analysis of the Association of Study-Level Factors With Treatment Effects With HA Injection vs Oral NSAIDsa
| Variable | SMD (95% CI)b | Risk Ratio (95% CI)c | ||||
|---|---|---|---|---|---|---|
| Knee Pain | Knee Function | AE | Serious AE | Study Withdrawal | Study Withdrawal Because of AE | |
| Double-blinding | ||||||
| Yes | 0.14 (–0.04, 0.32) | 0.11 (−0.22, 0.45) | 0.75 (0.60, 0.95) | 1.31 (0.29, 5.97) | 1.11 (0.85, 1.47) | 0.66 (0.32, 1.36) |
| No | 0.17 (–0.06, 0.39) | 0.29 (0.06, 0.51) | 0.34 (0.06, 1.84) | 0.99 (0.06, 15.6) | 0.80 (0.43, 1.48) | 0.61 (0.08, 4.90) |
| Industry funding | ||||||
| Yes | 0.17 (0.00, 0.34) | 0.13 (−0.15, 0.40) | 0.34 (0.06, 1.84) | 0.99 (0.06, 15.6) | 0.60 (0.31, 1.19) | 0.61 (0.08, 4.90) |
| No | 0.11 (–0.14, 0.36) | 0.32 (0.07, 0.58) | 0.75 (0.60, 0.95) | 1.31 (0.29, 5.97) | 1.15 (0.88, 1.51) | 0.66 (0.32, 1.36) |
| Sample size | ||||||
| ≥100/group | 0.15 (–0.02, 0.33) | 0.15 (−0.14, 0.44) | 0.67 (0.50, 0.91) | 1.93 (0.16, 22.8) | 1.03 (0.77, 1.37) | 0.64 (0.31, 1.32) |
| <100/group | 0.15 (–0.10, 0.39) | 0.29 (0.05, 0.54) | 0.86 (0.60, 1.25) | 1.02 (0.21, 4.94) | 1.16 (0.67, 2.00) | 0.80 (0.10, 6.45) |
| No. of HA injections | ||||||
| 1 | d | d | d | d | d | d |
| 3 | 0.24 (−0.04, 0.52) | 0.13 (−0.15, 0.40) | 0.84 (0.58, 1.22) | 1.03 (0.18, 5.78) | 1.09 (0.61, 1.94) | 0.14 (0.01, 2.65) |
| 5 | 0.15 (−0.02, 0.33) | 0.15 (−0.14, 0.44) | 0.67 (0.50, 0.91) | 1.93 (0.16, 22.8) | 1.03 (0.77, 1.37) | 0.64 (0.31, 1.32) |
| Oral NSAID type | ||||||
| Cox-selective | d | d | d | d | d | d |
| Nonselectiveb | 0.18 (0.03, 0.33) | 0.14 (−0.06, 0.34) | 0.73 (0.58, 0.93) | 1.26 (0.31, 5.20) | 1.04 (0.80, 1.34) | 0.59 (0.29, 1.18) |
aAE, adverse event; HA, hyaluronic acid; NSAID, nonsteroidal anti-inflammatory drug; SMD, standardized mean difference.
bBolded values indicate a statistically significant difference between groups because the 95% CI of the SMD does not include 0.
cBolded values indicate a statistically significant difference between groups because the 95% CI of the risk ratio does not include 1.
dData reported in a single trial; therefore, pooled estimates are not available.