Upsala J Med Sci 97: 195-200 Pulse Pressure, Mean Blood Pressure and Impaired Glucose Tolerance-A Study in Middle-aged Subjects Jan Cederholm1,2 and Lars Wibell' I Departments of Internal Medicine and 2Farnily Medicine, University HoApital, Uppsala, Sweden A B S T R A C l In a study of 695 middle-aged subjects, without antihypertensive agents, and without more pronounced obesity, both pulse pressure (PP) and mean blood pressure (MBP) were strongly related to 2-h blood glucose in 75 g OSTTs (p < 0 . 0 0 1 ) . All hypertensives (DBP 290 mm Hg) were separated into 39 with higher PP (260 m Hg) and 137 with lower PP (<60 mm Hg). The high PP hypertensives, compared with the low PP hypertensives and all 519 normtensives, had higher frequency of impaired glucose tolerance (IGT; WHO-criteria), 3 3 % , 606, and 4%, respectively (p <0.001), and also higher mean 2-h blood glucose, 5.9, 4.5, and 4,2 mmo1.1-1, respectively (p <0.001). These differences were independent of MBP levels. Similarly, all 54 hypertensives with higher MBP (2110 mm Hg) had more IGT and higher 2-h glucose than the 122 hypertensives with lower MBP (<110 m Hg) or the normtensives, 30%, 5% and 4%, respectively (p < 0 . 0 0 1 ) , and 5.8, 4.4, 4.2 m~l.l-~, respectively (p <0.001), independently of PP. Thus, both high PP and high MBP were related to IGT, independently of each other. I N T R O D U C T I O N Blood pressure can be divided into two components: a steady component, represented by mean blood pressure, and a pulsatile component, represented by pulse pressure (3,4,9). It was the aim of this study to analyse the independent relationships between blood glucose and these two components, in middle-aged untreated subjects. 195 M A T E R I A L A N D M E T H O D S A sample of 695 subjects, 47-54-years-old, was obtained from a health survey in Uppsala, previously described, with a participation rate of 71% (1). All subjects in the health survey with mre pronounced obesity or antihypertensive agents were excluded from the present study. The hypertensives, 176 subjects (45.5% males, 54.5% females), were the subjects with diastolic blood pressure (DBP) 290 mm Hg and no antihypertensive agents. The normotensives were 519 subjects (50.3% males, 49.7% females). Body mass index (BMI) was computed as weight - height-2 (kg. m - 2 ) and expressed as relative BMI (RBMI, % ) , based on ideal BMI values. Blood pressure (BPI was measured sitting after >15 min of rest and no previous smoking, using Korotkoff fifth phase sounds, with a mercury sphygmomanometer (cuff size 12.5 x 35 cm), by the same observer. Pulse pressure (PP) was the difference of systolic BP and diastolic BP. Mean blood pressure (MBP) was diastolic BP + one-third of PP. All subjects with RBMI >130% at the health survey were excluded from the present study, to avoid the problem of greater arm circumference in obese subjects. Oral glucose tolerance tests (CGTT) were performed in the morning after 10 hours of fasting (11). Venous whole blood glucose was measured at 0-h and 2-h levels by a glucose oxidase method (YSI Model 23 A M ) . Impaired glucose tolerance (IGT) was diagnosed according to strict WHO-criteria, based on two subsequent OGTTs (11). Subjects with manifest diabetes mellitus were excluded from the present study. A questionnaire was used to obtain information on smoking and physical activity during leisure time and at work, the latter evaluated with a 4-point scale as used in the Gothenburg studies. Statistical analysis. Analyses were performed with the SAS program. A p value of p <0.05 was considered statistically significant. Student's t-test, chi-square statistics and Pearson's correlations were used. Multiple regression analysis (PRCC GLM) was used (Table 1) with t-values of the predictors and the coefficient of determination (R ) given. Analysis of covariance (PRO= GLM) yielded mean values (Table 3 ) , after adjustment for confounding covariates. 2 N-way analysis of frequency distribution (PRCC FREQ) yielded the odds ratio (Table 3) , after adjustment for n confounding covariates, with Cochran - Mantel - Haenszel correlation and general association statistics. 196 Table 1. Multiple regressions, PP and MBP as dependent variables (n=695). ........................................................................... Pulse pressure Mean blood pressure t-va lue t-va lue Predictors: 2-h glucose Body mass index Age 4.8 *** 3.1 ** 5.2 *** 6.9 *** 4.2 *** 1.9 R E S U L T S In all 695 participants, the correlation coefficients (r) were: PP-SBP 0.84, PP-DBP 0.11, and MBP-SBP 0.89, MBP-DBP 0.91. Other correlations were: PP-MBP 0.52, SBP-DBP 0.62. All correlations were mainly similar in females and mles. Table 1 shows that 2-h glucose was independenly related to both PP and MBP (p <0.001). Hypertensives with low or high PP levels. All hypertensives were divided in two groups, with the mean + 1 SD value of PP (60 mm Hg) as the dividing level (Table 2). Group 2 hypertensives (high PP level) had higher 2-h glucose than group 1 hypertensives (low PP level) or normotensives, also after adjustment for MBP levels (Table 3, left part). The frequency of IGT (Table 4, left part) was clearly increased in group 2 hypertensives, 33%, compared with group 1 hypertensives, 7 % , and normtensives, 4%, also differing as odds ratios for IGT, adjusted for MBP levels (p <0.001). Hypertensives with low or hiqh MBP levels. All hypertensives were also divided in two groups, wih roughly the mean + 1 SD value of MBP (110 mm Hg) as the dividing level (Table 2). Group 4 hypertensives (high MBP) had higher mean 2-h glucose than group 3 hypertensives (low MBP) and normotensives, even when adjusted for PP levels (Table 3, right part), and the frequency of IGT was higher in group 4 hypertensives, 30 %, than in group 3 hypertensives, 5%, or normtensives, 4%, independently of PP according to adjusted odds ratios (p