Nine parallel-group RCTs that reported 12 comparison groups were included in the review (n=623, range 17 to 144). Three RCTs stratified participants by blood pressure status. Nine comparison groups were double blinded and three were single blinded. All studies received a Jadad score of 4 or 5. Drop-out rates ranged from 5% to 10%. No RCTs reported use of intention to treat analysis.
When all 12 comparisons were pooled, IPP and VPP significantly reduced blood pressure compared to controls: systolic blood pressure reduced by -4.8mmHg (95% CI -6.0 to -3.7) and diastolic by -2.2mmHg (95% CI -3.1 to -1.3). There was no evidence of significant heterogeneity or publication bias.
Subgroup analysis showed a more marked effect in the hypertensive group than in the pre-hypertensive group; the difference (2.4mmHg) was statistically significant for systolic blood pressure. Other subgroup analyses showed no significant differences between groups.
Most RCTs reported minor adverse events in both groups, none of which were considered to be treatment related.
An additional analysis (apparently post-hoc) suggested a trend for effects to increase over follow up.