Table 2.
Results of meta-regression analyses. Analyzed using a random-effects restricted maximum of likelihood (REML)-based meta-regression model.
| Variable | Total Trials, k | SMD for Pain | 95% CI | τ2 | I2 | p-Value for Interaction |
|---|---|---|---|---|---|---|
| All trials | 30 | −0.24 | (−0.42 to −0.07) | 0.150 | 63% | n.a. |
| Diagnoses | 0.147 | 60% | 0.024 a | |||
| RA | 22 | −0.21 | (−0.42 to −0.00) a | |||
| OA | 5 | −0.16 | (−0.57 to 0.24) | |||
| Other | 3 | −0.63 | (−1.20 to −0.06) a | |||
| Supplementation type | 0.137 | 60% | 0.009 a | |||
| Capsule | 23 | −0.25 | (−0.44 to −0.06) a | |||
| Bottle | 3 | 0.19 | (−0.37 to 0.75) | |||
| Unclear | 4 | −0.61 | (−1.16 to −0.05) a | |||
| Type of control | 0.148 | 60% | 0.051 | |||
| PUFA w/o EPA and DHA | 11 | −0.12 | (−0.41 to 0.17) | |||
| Non-PUFA oils | 10 | −0.30 | (−0.59 to −0.01) a | |||
| Non-oils | 3 | −0.63 | (−1.13 to −0.12) a | |||
| Unclear | 3 | −0.01 | (−0.59 to 0.56) | |||
| None | 3 | −0.21 | (−0.87 to 0.44) | |||
| Duration | 0.165 | 63% | 0.074 | |||
| <12 weeks | 3 | −0.30 | (−0.89 to 0.30) | |||
| ≥12 weeks and <24 weeks | 12 | −0.25 | (−0.53 to 0.03) | |||
| ≥24 weeks | 15 | −0.23 | (−0.48 to 0.03) | |||
| Unspecified | 0 | |||||
| Ratio of EPA/DHA | 0.153 | 61% | 0.031 a | |||
| Ratio of EPA/DHA ≤1.5 | 12 | −0.12 | (−0.40 to 0.15) | |||
| Ratio of EPA/DHA of >1.5 | 11 | −0.38 | (−0.67 to −0.10) a | |||
| Unspecified | 7 | −0.23 | (−0.62 to 0.17) | |||
| Dosage of EPA plus DHA | 0.138 | 58% | 0.016 a | |||
| <2.6 g/day | 8 | −0.45 | (−0.75 to −0.15) a | |||
| ≥2.6 g/day and <3.6 g/day | 11 | −0.21 | (−0.51 to 0.09) | |||
| ≥3.6 g/day | 4 | 0.13 | (−0.32 to 0.57) | |||
| Unspecified | 7 | −0.22 | (−0.60 to 0.16) | |||
| Source of marine oil | 0.130 | 57% | 0.011 a | |||
| Whole fish | 19 | −0.17 | (−0.38 to 0.05) | |||
| Mussel | 2 | −0.95 | (−1.60 to −0.31) a | |||
| Other | 6 | −0.19 | (−1.53 to 0.16) | |||
| Unspecified | 3 | −0.35 | (−0.97 to 0.26) | |||
| Bias domains | ||||||
| Random sequence generation (selection bias) | 0.160 | 61% | 0.043 a | |||
| Adequate | 4 | −0.33 | (−0.78 to 0.13) | |||
| Unclear | 22 | −0.19 | (−0.40 to 0.02) | |||
| Inadequate | 4 | −0.48 | (−1.00 to 0.05) | |||
| Allocation concealment (selection bias) | 0.155 | 61% | 0.039 a | |||
| Adequate | 3 | −0.52 | (−1.09 to 0.05) | |||
| Unclear | 25 | −0.22 | (−0.42 to −0.03) a | |||
| Inadequate | 2 | −0.03 | (−0.76 to 0.70) | |||
| Blinding of participants (performance bias) | 0.149 | 61% | 0.026 a | |||
| Adequate | 9 | −0.10 | (−0.42 to 0.22) | |||
| Unclear | 10 | −0.41 | (−0.70 to −0.11) a | |||
| Inadequate | 11 | −0.20 | (−0.49 to 0.10) | |||
| Blinding of personnel (performance bias) | 0.163 | 63% | 0.060 | |||
| Adequate | 5 | −0.25 | (−0.70 to 0.20) | |||
| Unclear | 17 | −0.29 | (−0.52 to −0.05) a | |||
| Inadequate | 8 | −0.13 | (−0.50 to 0.24) | |||
| Incomplete outcome data (attrition bias) | 0.152 | 60% | 0.020 a | |||
| Adequate | 2 | 0.06 | (−0.61 to 0.73) | |||
| Unclear | 10 | −0.44 | (−0.77 to −0.12) a | |||
| Inadequate | 18 | −0.18 | (−0.40 to 0.04) | |||
| Outcome reporting (outcome reporting bias for pain) | 0.145 | 61% | 0.016 a | |||
| Adequate | 27 | −0.20 | (−0.38 to −0.02) a | |||
| Unclear | 1 | −0.94 | (−2.10 to 0.21) | |||
| Inadequate | 2 | −0.72 | (−1.52 to 0.08) | |||
| Funding source | 0.149 | 61% | 0.031 a | |||
| Industry only | 8 | −0.10 | (−0.41 to 0.21) | |||
| Mixed | 8 | −0.10 | (−0.44 to 0.25) | |||
| Nonprofit only | 4 | −0.41 | (−0.87 to 0.05) | |||
| Not reported | 7 | −0.31 | (−0.70 to 0.08) | |||
| Unclear | 3 | −0.80 | (−1.42 to −0.16) a | |||
95% CI: 95% confidence interval, I2: heterogeneity, SMD: standardized mean difference, k: number of trials, τ2: estimated between-study variance; a p < 0.05.