Objectives To examine the health risks associated with smoking and blood pressure in Chinese hypertensive patients and to compare the bene®t of antihypertensive drug treatment with the risk attributable to smoking.Methods We used multiple Cox regression to correlate outcome with blood pressure and smoking status in 2284 older (aged > 60 years) patients enrolled in the Systolic Hypertension in China (Syst-China) Trial (systolic/diastolic blood pressure > 160/< 95 mmHg).Results Median follow-up was 3.0 years. After adjustment for sex, age, active antihypertensive treatment and various entry characteristics, the relative hazard rates associated with smoking more than 20 cigarettes per day were 2.04 (P = 0.04), 4.66 (P < 0.001) and 4.74 (P = 0.002) for allcause, noncardiovascular and cancer mortality, respectively. With similar adjustments applied, the relative hazard rates for total (fatal and non-fatal) stroke associated with smoking 10±20 and more than 20 cigarettes per day were 1.78 (P = 0.04) and 2.23 (P = 0.03), respectively. Furthermore, both smoking and systolic blood pressure were associated with higher risk of stroke. Compared with the overall risk in the whole group, treating 1000 patients for 5 years prevented 40 [95% con®dence interval (CI), 5±75] strokes in smokers as well as never smokers. Prognosis in never smokers and past smokers was similar. Quitting smoking had the potential to prevent 51 (95% CI, 221 to 122) strokes in untreated hypertensive patients and to prevent 45 (95% CI, ±14 to 104) additional strokes in treated patients, over and above the effects of antihypertensive treatment. If, in addition to stroke, noncardiovascular mortality was also accounted for, the estimated absolute bene®t of quitting smoking increased to 69 (95% CI, 218 to 155) and 97 (95% CI, 23±171) events in the untreated and the treated group, respectively.Conclusions In elderly Chinese patients, smoking was a risk factor for all-cause, non-cardiovascular and cancer mortality, as well as fatal and non-fatal stroke. The potential bene®ts of antihypertensive treatment and quitting smoking were approximately similar. In our view, these ®ndings are important in terms of public health policies and health economics.