Thirty-seven RCTs (n=9,008) were included in the review. Results of the quality assessment were not reported.
None of the clinical outcomes were significantly improved at any level of change in blood pressure. The lowest odds ratios (or best treatment effects) were reported for early death (OR 0.87, 95% CI 0.54 to 1.23) at an on-treatment difference of 8.1mmHg, death at the end of follow-up (OR 0.96, 95% CI 0.31 to 1.65) occurring at a blood pressure difference of 14.4mmHg and combined death or dependency at the end of follow-up (OR 0.95, 95% CI 0.11 to 1.72) present at a blood pressure difference of 14.6mmHg.
The authors stated that increases or large falls in blood pressure in the intervention groups were associated with poor clinical outcome. Sensitivity analyses did not materially alter the main findings, except for treatment duration less than 14 days being significantly associated with early mortality (OR 0.41, 95% CI 0.00 to 0.85).
There was no evidence of publication bias.