Incident hypertension and its prediction model in a prospective northern urban Han Chinese cohort study

Trends in incidence and prevalence of hypertension are grave in China and identifying high-risk, non-hypertension individuals for intervention may delay hypertension onset. We aimed to investigate the incidence of hypertension in northern urban Han Chinese population and construct multivariable hypertension prediction models through the prospective cohort, which included 7537 men and 4960 women free of hypertension at baseline between 2005 and 2010. During 38 958 person-years of follow-up, 2785 participants (men, 72.57%; women, 27.43%) developed hypertension. The incidence density of hypertension was 71.48 per 1000 person-year. In multivariable backward cox analyses, age, body mass index, systolic blood pressure (SBP), diastolic blood pressure (DBP), fasting blood glucose and current drinking were retained for both men and women, while gamma-glutamyl transferase only for men, total cholesterol, neutrophil granulocyte and current smoking only for women. The area under receiver operating characteristic curve (AUC) was 0.761 (95% confidence interval (CI), 0.752–0.771) for men and 0.753 (95% CI, 0.741–0.765) for women, even after 10-fold cross-validation, the AUC was 0.760 (95% CI, 0.751–0.770) for men and 0.749 (95% CI, 0.737–0.761) for women. Through risk stratification, the absolute risk of incident hypertension in 5 years at moderate, high and very high risk level was 2.13, 3.84 and 6.14 times that of those who were at low risk in men, and 1.30, 2.56 and 6.01 times that of those who were at low risk in women. Our findings identified predictors of incident hypertension and indicated that the sex-specific multivariable prediction models would be simply used to estimate the risk of incident hypertension.