Stroke Risk of Blood Pressure Indices Determined by Home Blood Pressure Measurement

Methods

The present study is part of a longitudinal observational cohort study of subjects who have been participating in our HBP measurement project in Ohasama, Japan. The socioeconomic and demographic characteristics of Ohasama, the selection procedure of the study populations, the HBP and CBP measurement procedures, and the data collection procedures have been described previously.^5–7 The present study population consisted of 2369 individuals (40% men), ≥35 years of age (mean, 59.2 years), without a history of stroke and with ≥3 days of morning HBP and a CBP.⁶

The subjects were followed from the dates of HBP measurement (approximately 1992) until December 31, 2004. The procedures for diagnosing stroke were described previously.^7,8 Cerebral infarction (International Classification of Diseases, 10th Revision code I63) was defined as ischemic stroke, and intracerebral hemorrhage (I61) and subarachnoid hemorrhage (I60) were defined as hemorrhagic stroke.

The relative hazard and 95% CI were estimated for a 1-SD difference for each index using Cox proportional hazards regression, adjusted for age, sex, smoking status, use of antihypertensive medication, and history of heart disease, diabetes mellitus, or hypercholesterolemia. The likelihood ratio (LR) χ² value was used as a measure of the improvement of goodness of fit⁹ between the model containing a single BP index (and confounding factors) and the model containing 2 indices. A significant LR χ² indicates that the index represents a significantly stronger association with stroke.⁹

Results

During follow-up (mean, 11.7 years), 238 strokes occurred (169 ischemic, 69 hemorrhagic).

All baseline BP indices were significantly higher in subjects who developed stroke than in those who did not (Table 1).

Figure A shows relative hazards for total stroke. Of the HBP indices, SBP and MBP were strongly associated with total stroke followed by DBP. The association between total stroke and PP was relatively weak. Associations between stroke and the CBP indices were generally weak. Table 2 shows the LR χ² values when 2 HBP indices were analyzed simultaneously. For total stroke, the LR χ² values of SBP and MBP (LR χ² ≥11.8, P<0.01) were significantly higher than those of DBP and PP (LR χ² ≤3.8, P≥0.05). The LR χ² value of SBP tended to be larger than that of MBP, although the difference was not significant. The LR χ² values of DBP and PP were not significantly different when DBP and PP were compared.

• Download figure
• Open in new tab
• Download powerpoint

Figure. Relative hazards (RHs) and 95% CIs of BP indices for stroke risk. Boxes represent RHs. Horizontal lines represent 95% CIs. (A), Total stroke; (B) ischemic stroke; (C) hemorrhagic stroke.

Figure B shows the relative hazards for ischemic stroke. The association between ischemic stroke and SBP was the strongest followed by MBP. Associations between ischemic stroke and DBP and PP were relatively weak. As shown in Table 2, the LR χ² of SBP tended to be larger than that of MBP when SBP was compared with MBP, but there were no significant differences. The LR χ² values of SBP and MBP were significantly stronger than those of DBP and PP when SBP and MBP were compared with DBP and PP. The LR χ² values of DBP and PP were not significantly different when DBP and PP were compared, although that of PP tended to be larger than that of DBP.

Figure C shows the relative hazards for hemorrhagic stroke. SBP, DBP, and MBP were similarly associated with hemorrhagic stroke, and PP was not significantly associated. No significant differences were observed when the LR χ² values of SBP, DBP, and MBP were compared; these values were also significantly larger than that of PP (Table 2). These results did not change for the analysis of intracerebral hemorrhage alone.

Discussion

Recent studies using CBP^3,4 and ambulatory BP,⁸ which investigated associations between stroke and BP indices, showed that PP was not very strongly associated with stroke. In the present study using HBP, like in these previous studies, PP was not strongly associated with stroke. LR tests showed that SBP and MBP were significantly more strongly associated with total or ischemic stroke than DBP and PP. SBP, MBP, and DBP were all significantly more strongly associated with hemorrhagic stroke than PP. Thus, PP may not be important for predicting stroke. However, PP tended to be slightly more strongly associated with ischemic stroke than DBP, suggesting that different BP indices are specifically associated with different stroke types.

SBP tended to be more strongly associated with total and ischemic stroke than MBP, although the differences were not significant. Considering that MBP has to be calculated from the formula as an approximation, SBP may be the most useful index to follow in daily medical practice.

In conclusion, PP obtained from HBP measurements was not strongly associated with stroke. SBP showed a strong association with stroke and appears to be a useful index. DBP and PP may be associated with different stroke types.