TABLE 3.
GRADE evaluation for body weight (kg) and all comparisons1
| Direct evidence | Indirect evidence | Network meta-analysis | |||||
|---|---|---|---|---|---|---|---|
| Comparison (meals/d) | No. of studies | MD (95% CI) | Certainty of evidence | MD (95% CI) | Certainty of evidence | MD (95% CI) | Certainty of evidence |
| 1 vs. 2 | 0 | — | — | −0.93 (−3.80, 1.94) | ⊕⊕◯◯2 | −0.93 (−3.80, 1.94) | ⊕◯◯◯3 |
| 1 vs. 3 | 2 | −3.35 (−6.60, −0.09) | ⊕⊕⊕◯4 | 2.28 (−3.40, 7.96) | ⊕⊕◯◯ | −1.95 (−4.78, 0.87) | ⊕⊕◯◯ 3, 5, 6 |
| 1 vs. 4 | 0 | — | — | −0.33 (−12.39, 11.73) | ⊕⊕◯◯2 | −0.33 (−12.39, 11.73) | ⊕◯◯◯3(↓↓) |
| 1 vs. 6 | 1 | 0.01 (−3.94, 3.96) | ⊕⊕⊕◯4 | −4.61 (−8.68, −0.53) | ⊕⊕◯◯2 | −2.22 (−5.06, 0.61) | ⊕⊕◯◯3, 5, 6 |
| 1 vs. ≥8 | 0 | — | — | −2.25 (−5.13, 0.63) | ⊕⊕◯◯2 | −2.25 (−5.13, 0.63) | ⊕◯◯◯3 |
| 2 vs. 3 | 1 | −1.13 (−3.36, 1.10) | ⊕⊕⊕⊕ | −1.01 (−1.72, −0.31) | ⊕⊕⊕◯2 | −1.02 (−1.70, −0.35) | ⊕⊕⊕◯ |
| 2 vs. 4 | 1 | 0.60 (−11.11, 12.31) | ⊕⊕⊕⊕ | — | — | 0.60 (−11.11, 12.31) | ⊕⊕◯◯3(↓↓) |
| 2 vs. 6 | 3 | −1.29 (−1.75, −0.83) | ⊕⊕⊕◯4 | −1.41 (−3.70, 0.89) | ⊕⊕⊕◯2 | −1.29 (−1.74, −0.84) | ⊕⊕⊕◯5 |
| 2 vs. ≥8 | 0 | — | — | −1.32 (−2.19, −0.45) | ⊕⊕◯◯2 | −1.32 (−2.19, −0.45) | ⊕◯◯◯3 |
| 3 vs. 4 | 0 | — | — | 1.62 (−10.11, 13.35) | ⊕⊕◯◯2 | 1.62 (−10.11, 13.35) | ⊕◯◯◯3(↓↓) |
| 3 vs. 6 | 7 | −0.28 (−0.81, 0.26) | ⊕⊕⊕◯4 | −0.18 (−2.44, 2.09) | ⊕⊕⊕◯2 | −0.27 (−0.79, 0.25) | ⊕⊕⊕◯5 |
| 3 vs. ≥8 | 5 | −0.30(−0.85, 0.26) | ⊕⊕⊕⊕ | — | — | −0.30(−0.85, 0.26) | ⊕⊕⊕◯7 |
| 4 vs. 6 | 0 | — | — | −1.89 (−13.61, 9.83) | ⊕⊕◯◯2 | −1.89 (−13.61, 9.83) | ⊕◯◯◯3(↓↓) |
| 4 vs. ≥8 | 0 | — | — | −1.92 (−13.66, 9.82) | ⊕⊕◯◯2 | −1.92 (−13.66, 9.82) | ⊕◯◯◯3(↓↓) |
| 6 vs. ≥8 | 0 | — | — | −0.02 (−0.78, 0.74) | ⊕⊕◯◯2 | −0.02 (−0.78, 0.74) | ⊕⊕◯◯ |
Direct estimates were evaluated with the following GRADE criteria: risk of bias, indirectness, inconsistency, and publication bias. As suggested recently by the GRADE working group, consideration of imprecision is not necessary when rating the direct and indirect estimates to inform the rating of NMA estimates. GRADE, Grading of Recommendations Assessment, Development, and Evaluation; MD, mean difference; NMA, network meta-analysis; ⊕⊕⊕⊕, high; ⊕⊕⊕, moderate; ⊕⊕◯◯, low; ⊕◯◯◯, very low; ↓↓, downgraded twice due to very serious imprecision.
2Downgraded due to intransitivity (i.e., patients with obesity and healthy participants included).
Downgraded due to imprecision (95% CI overlaps important benefit: −2 kg; or important harm: +2 kg).
4Downgraded due to risk of bias (≥1 RCT with high risk of bias).
5Direct evidence contributing more to the NMA estimate (>50%).
6Not downgraded due to incoherence (dominant estimate similar to network estimate).
7Downgraded due to imprecision (sample size: <400).