BACKGROUND We conducted a systematic review to estimate the effect of acupuncture on blood pressure (BP) in hypertensive patients. METHODS Electronic literature searches for randomized controlled trials (RCTs) of acupuncture were performed in six electronic databases to June 2007 without language restrictions. RESULTS Eleven RCTs testing acupuncture either as an adjunct or an alternative met our inclusion criteria and they showed a wide variety of methodological quality, mainly due to poor reporting. Three sham-controlled trials out of 11 studies were statistically pooled: systolic BP (SBP) change was not statistically significant (mean difference -5 mm Hg, 95% CI (-12, 1), P = 0.12) and acupuncture only marginally reduced diastolic BP (DBP) by 3 mm Hg (95% CI (-6, 0), P = 0.05), but substantial heterogeneity was observed (I(2) = 92% for SBP, I(2) = 79% for DBP). When given with antihypertensive medication, acupuncture significantly reduced SBP (-8 mm Hg, 95% CI (-10, -5), P < 0.00001) and DBP (-4 mm Hg, 95% CI (-6, -2), P < 0.0001) and no heterogeneity between studies was detected. Four studies that investigated acupuncture against antihypertensive medication indicated noninferiority of acupuncture in lowering BP, albeit the quality of them was poor, and their sample sizes were not satisfactory as an equivalence study. Other studies comparing acupuncture with various control procedures had inconsistent findings and most of them were of low methodological quality. CONCLUSIONS Considering the limitation of the four positive noninferiority studies and the results of the meta-analysis of the three sham-controlled studies, the notion that acupuncture may lower high BP is inconclusive. More rigorous trials are warranted.