Objective: Cardiovascular risks are globally elevated in South Asians, but this masks important ethnic subgroup differences in risk factors, such as hypertension, which have not been fully explored. We conducted this study to explore the variations in hypertension within ethnic subgroups among South Asians. Design: Cross-sectional survey [National Health Survey of Pakistan (NHSP) (1990-1994)]. Setting: Population based. Participants: A total of 9442 individuals aged 15 years or over. Methods: Data on sociodemographic and clinical variables were collected. Distinct ethnic subgroups - Muhajir, Punjabi, Sindhi, Pashtun and Baluchi - were defined by mother tongue. Main outcome measure: Hypertension defined as systolic blood pressure ≥ 140 mmHg or diastolic blood pressure 90 mmHg, or currently receiving antihypertensive therapy. Results: The age-standardized prevalence of hypertension was highest among Baluchis (25.3% in men and 41.4% in women), then Pashtuns (23.7% in men and 28.4% in women), Muhajirs (24.1% in men and 24.6% in women), and lowest among Punjabis (17.3% in men and 16.4% in women) and Sindhis (19.0% in men and 9.9% in women) (P = 0.001). While hypertension was more prevalent in urban (22.7%) versus rural dwellers (18.1%) [odds ratio (OR) 1.34; 95% confidence interval (CI), 1.20, 1.49], this difference was no longer significant after adjusting for body mass and waist circumference (OR 1.03; 95% Cl, 0.91, 1.16). However, ethnic differences persisted after adjusting for major sociodemographic, dietary and clinical risk factors (unadjusted OR for Baluchi versus Sindhi, 2.92; 95% CI, 2.20-3.89; adjusted OR, 2.71; 95% Cl, 1.97-3.75). Conclusions: A threefold difference in prevalence of hypertension exists between people of South Asian descent, which, unlike the urban/rural difference, cannot be accounted for by measured risk factors. Further study would provide valuable etiological and therapeutic clues. © 2003 Lippincott Williams & Wilkins.