Two reviewers extracted changes in resting SBP and DBP with variances derived from standard deviations (SD) from
the included trials; where change score standard deviations were missing they were calculated from pre- and poststandard
deviation values. Disagreements were resolved through consensus.