Table 5.
GRADE evidence profile of RCTs for effect of vertebroplasty for osteoporotic vertebral fractures
| Outcome | Summary of findings | Quality assessment | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Time (months) | Number (treated/control) | WMD (95% CI, g/cm2) | Risk of bias* | Inconsistency† | Indirectness | Imprecision‡ | Others§ | Quality | |
| Pain score | 1–29 | 8 (449/285) | −2.06 (−3.39, −0.74) | No serious | Not serious | Not serious | Not serious | None | High |
| 90 | 8 463/289) | −1.47 (−2.20, −0.75) | Not serious | Not serious | Not serious | Not serious | None | High | |
| Greater than 180 | 6 (379/237) | −1.25 (−1.64, −0.86) | Not serious | Not serious | Not serious | Not serious | None | High | |
| RDQ score | 30 | 4 (191/161) | −1.75 (−2.71, −0.78) | Not serious | Not serious | Not serious | Serious | None | Moderate |
| 90 | 3 (173/145) | −1.44 (−2.45, −0.44) | Not serious | Not serious | Not serious | Serious | None | Moderate | |
| EQ-5D score | 30–90 | 3 (103/85) | −0.11 (−0.13, −0.08) | Not serious | Not serious | Not serious | Serious | None | Moderate |
| QUALEFFO | 1–29 | 3 (139/129) | −5.68 (−9.10, −2.26) | Not serious | Not serious | Not serious | Serious | None | Moderate |
| Refracture | 7 (81/45) | −1.44 (−2.45, −0.44) | Not serious | Not serious | Not serious | Not serious | None | High | |
GRADE = Grading of Recommendations Assessment, Development and Evaluation; WMD = weighted mean difference; RDQ = Roland-Morris Disability Questionnaire; EQ-5D = European Quality of Life–5 Dimensions; QUALEFFO = Quality of Life Questionnaire of the European Foundation for Osteoporosis; * inadequate blinding, lack of allocation concealed in some trials may increase risk of bias; †inconsistent report of outcomes and significant heterogeneity existed across the trials, but all were well explained by the subgroup analysis; ‡if a study has a wide confidence interval around the estimate of the effect, or included patients less than 400, it may cause imprecision; §“Other” consisted of publication bias and upgraded quality of evidence (large effect, plausible residual confounding and dose-response gradient).