Twenty-eight RCTs were included (n=2,731 participants, range 16 to 654). All were rated as having satisfactory allocation concealment. Twelve trials reported blinding. Reported losses to follow-up ranged from nil to 16 participants.
No statistically significant difference was found between intervention and control arms in rates of mortality (28 RCTs, I2=0%), myocardial infarction (15 RCTs, I2=0%) and stroke (13 RCTs, I2=0%). No studies reported events of epidural haematoma or abscess.
The intervention was associated with a significantly lower rate of supraventricular tachyarrhythmias (RR 0.68, 95% CI 0.50 to 0.93, I2=62%; 14 RCTs). Statistical heterogeneity in this analysis was attributed partially to selective use of clonidine in one RCT, which may have biased results in favour of the intervention. The intervention was associated with a significantly lower rate of respiratory complications (RR 0.53, 95% CI 0.40 to 0.69, I2=0%; 13 RCTs).
No evidence was found of significant publication bias. None of the variables investigated in meta-regression had a significant effect on outcomes.