Table 4. Association of hypertension categories with CVD mortality in women and men combined.

		Hypertensive group
	Total	Normotensive	Untreated hypertensive	Poorly treated hypertensive	Well treated hypertensive
Using subset with no missing values for any covariable/hypertensive group
N	4796	1264	2427	941	164
CVD deaths	107	15	59	29	4
Hazard ratios (95% CI)
Model 1*	-	1	1.60 (0.90, 2.85)	1.98 (1.05, 3.72)	1.85 (0.61, 5.58)
Model 2*	-	1	1.66 (0.93, 2.96)	2.09 (1.09, 4.00)	1.70 (0.56, 5.22)
Model 3*	-	1	1.09 (0.54, 2.20)	1.29 (0.58, 2.89)	1.67 (0.55, 5.12)
Using imputed data for missing values (20 imputations)
N**	6476	1688	3328	1245	215
CVD deaths**	158	21	92	39	6
Hazard ratios (95% CI)
Model 1*	-	1	1.80 (1.10, 2.94)	1.97 (1.14, 3.39)	2.04 (0.82, 5.08)
Model 2*	-	1	1.91 (1.17, 3.12)	2.13 (1.23, 3.71)	1.97 (0.78, 4.94)
Model 3*	-	1	1.19 (0.65, 2.18)	1.26 (0.64, 2.47)	1.92 (0.77, 4.83)

*Model 1: adjusted for sex, age (as time variable in Cox model).

Model 2: additionally adjusted for BMI, smoking, physical activity, LDL cholesterol, HDL cholesterol, triglycerides (log scale), glucose (log scale), insulin (log scale), fathers occupational social class (childhood social class), own adult occupational social class.

Model 3: additionally adjusted for systolic and diastolic blood pressure (BP).

**Imputed numbers per group and associated numbers of events are means over 20 imputed datasets.

Numbers per group differ slightly between Tables 2, 3, 4 as time to event and the censoring variable are included as a covariable in the imputation prediction model.