Summary of findings for the main comparison. Perineural dexamethasone versus placebo.
|
Patient or population: participants undergoing surgery with peripheral nerve block
Setting: participants undergoing upper and lower limb surgery in hospitals in Australia, Bangledesh, Belgium, Brazil, India, Iran, Japan, Korea, Nepal, Turkey and USA Intervention: perineural dexamethasone Comparison: placebo | ||||
| Outcomes | Anticipated absolute effects* (95% CI) | № of participants (studies) | Quality of the evidence (GRADE) | |
| Risk with placebo | Risk with perineural dexamethasone | |||
| Duration of sensory block (we included all studies describing duration of sensory block, regardless of how it was described) |
The mean duration of sensory block was 10.2 hours | The mean duration of sensory block in the perineural dexamethasone group was 6.70 hours longer (5.54 longer to 7.85 longer) | 1625 (27 RCTs) | ⨁⨁◯◯ LOW a |
| Incidence of serious adverse events (we used the NIH definition of adverse events. A serious event includes death, a life‐threatening event that requires hospitalization or prolonged hospitalization, disability or congenital anomaly) |
In seven studies, authors reported that they assessed for serious adverse events. Five serious adverse events were reported in three studies: one block‐related adverse event (pneumothorax) occurred in one participant in a trial comparing perineural dexamethasone and placebo; however, group allocation was not reported. The remaining non‐block‐related events occurred in two trials comparing perineural dexamethasone, intravenous dexamethasone and placebo. Two participants in the placebo group required hospitalization within one week of surgery; one for a fall, and one for a bowel infection. One participant in the placebo group developed Complex Regional Pain Syndrome Type I and one in the intravenous dexamethasone group developed pneumonia. | 620 (7 RCTs) |
⨁◯◯◯
VERY LOWb |
|
| Postoperative pain intensity at 12 hours (assessed by pain scores on an 11‐point VAS) |
The mean postoperative pain intensity at 12 hours was 3.0 | The mean postoperative pain score at 12 hours in the perineural dexamethasone group was 2.08 points lower (1.52 lower to 2.63 lower) | 257 (5 RCTs) | ⨁◯◯◯
VERY LOWc |
| Postoperative pain intensity at 24 hours. (assessed by pain scores on an 11‐point VAS) |
The mean postoperative pain intensity at 24 hours was 3.9 | The mean postoperative pain score at 24 hours in the perineural dexamethasone group was 1.63 points lower (0.93 lower to 2.34 lower) | 469 (9 RCTs) | ⨁⨁◯◯ LOW d |
| Postoperative pain intensity at 48 hours (assessed by pain scores on an 11‐point VAS) |
The mean postoperative pain intensity at 48 hours was 3.3 | The mean postoperative pain score at 48 hours in the perineural dexamethasone group was 0.61 points lower (1.24 lower to 0.03 higher) | 296 (3 RCTs) | ⨁⨁◯◯ LOW e |
| *The risk in the intervention group (and its 95% confidence interval) is based on the assumed risk in the comparison group and the relative effect of the intervention (and its 95% CI). CI: Confidence interval; MD: Mean difference; NIH: National Institute of Health; RCT: randomized controlled trial; VAS: Visual Analogue Scale. | ||||
| GRADE Working Group grades of evidence High‐quality: we are very confident that the true effect lies close to that of the estimate of the effect. Moderate‐quality: we are moderately confident in the effect estimate; the true effect is likely to be close to the estimate of the effect, but there is a possibility that it is substantially different. Low‐quality: our confidence in the effect estimate is limited; the true effect may be substantially different from the estimate of the effect. Very low‐quality: we have very little confidence in the effect estimate; the true effect is likely to be substantially different from the estimate of effect. | ||||
aDowngraded by one level for risk of bias as 19 out of 27 studies are at unclear risk of bias. Reasons include lack of reporting on random sequence generation, concealment allocation, blinding, and attrition bias. Downgraded by one level for inconsistency (I2 = 99%, P value for heterogeneity is < 0.00001) and heterogeneity is not explained by subgroup analyses; point estimates vary widely among studies, confidence intervals show minimal overlap).
bDowngraded by one level for risk of bias as four out of the seven studies are at unclear risk of bias. Reasons include lack of reporting on random sequence generation, concealment allocation, blinding, and evidence of selective reporting bias. Downgraded by two levels for imprecision due to very low number of events.
c Downgraded by one level for risk of bias. Three out of five studies are at unclear risk of bias. Reasons include lack of
reporting on random sequence generation and allocation concealment and evidence of attrition bias, selective reporting bias, and stopping early for benefit. Downgraded by one level for inconsistency (I2 = 61%, P value for heterogeneity is 0.03) and heterogeneity is not explained by subgroup analyses; point estimates vary widely among studies, confidence intervals show minimal overlap
dDowngraded by one level for inconsistency (I2 = 80%, P value for heterogeneity is < 0.00001) and heterogeneity is not explained by subgroup analyses; point estimates vary widely across studies. Downgraded by one level for imprecision (95% confidence interval includes no clinical effect and a clinical effect). By no clinical effect we mean the lower bound of the CI did not surpass our chosen MID threshold of 1.2 on VAS.
e Downgraded by two levels for imprecision because of a sparse number of participants (n=296) and a very wide confidence interval demonstrating that the treatment effect is not statistically significant and of questionable clinical significance.