A Prognostic Role of Mean 24-h Pulse Pressure Level for Cardiovascular Events in Type 2 Diabetic Subjects Under 60 Years of Age

The blood pressure (BP) monitoring program recruited 392 consecutive subjects with type 2 diabetes (249 men and 143 women, aged between 17 and 88 years) who were initially admitted to our hospital from December 1988 to June 1998 and then followed in our outpatient clinic (14,15). In this study, 237 type 2 diabetic subjects under 60 years of age, who were defined as those without any history of vascular complications based on atherosclerosis, were subanalyzed. The study protocol and analytical methods have been described previously (14,15). After excluding nine dropout subjects (four cancer, one pneumonia, one traffic accident, one suicide, and two unknown), 228 subjects (46 ± 11 years of age [mean ± SD]; 69% men) entered the study. The mean follow-up period was 100 ± 44 months (range 3–168).

Data are presented as means ± SD. Because all data for the comparison of group means were almost normally distributed, ANOVA was applied, except for serum triglyceride, for which variables were log-transformed to approximate normality. When testing the number of subjects in groups, the χ² test was used. Histograms divided by 5 mmHg each were depicted to evaluate the distribution of mean 24-h, daytime, and nighttime PPs, and skewness and kurtosis were examined for normality. On each histogram, a theoretical normal distribution curve was superimposed. An association between ambulatory PP and age was determined using Pearson’s correlation analysis. The rates of occurrences of cardiovascular, cerebrovascular, and all events were expressed in cases per 1, 000 patient-years in diabetic subgroups divided by quartile of the distribution of mean 24-h PPs. Cumulative incidence of cardiovascular, cerebrovascular, and all events were estimated using the Kaplan-Meier product-limit method and compared using the log-rank test, and relative risks between two diabetic subgroups were calculated using the Cox-Mantel test. The effects of prognostic factors on occurrences of various vascular events were evaluated by the Cox proportional-hazards model. In this analysis, dummy variables were used for hypertension (absent or present), and the other measured data were treated as continuous variables. Relative risks and their 95% CIs in the Cox’s model were calculated as previously described (14). The statistical significance was defined as a two-tailed P value <0.05. Multivariate models were used to adjust for potential confounding factors at P < 0.05 in the comparison of baseline characteristics.

In Fig. 1, histograms of mean 24-h, daytime, and nighttime PP were depicted with a theoretical normal distribution curve. Distribution of PP in any time periods for all subjects (Fig. 1, upper panels) was positively skewed, meaning that there might be a diabetic subgroup that had a wide PP. When diabetic subjects were divided into two subgroups (young subjects [≤40 years of age] and middle-aged subjects [between 41 and 59 years of age]), all histograms of 24 h, daytime, and nighttime tended to be similar to those for all subjects. Further analysis was therefore performed by stratifying the diabetic subjects by quartile of the distribution of mean 24-h PP. The cut points for quartiles of mean 24-h PP were 41.0, 46.3, and 53.3 mmHg. Then the widest quartile of diabetic subjects (n = 58) was compared with the other narrower quartiles (n = 170).

In the diabetic subjects with a wide PP, age was older, the prevalence of hypertensive subjects was higher, duration of diabetes was longer, and serum levels of creatinine, total cholesterol, and lipid ratio were higher than in individuals with a narrow PP. There was however no difference in sex, BMI, HbA_1c, serum triglyceride, or prevalence of insulin use and current smokers between the two groups (Table 1).

Moreover, Table 1 shows the comparison of mean 24-h, daytime, and nighttime levels of systolic BP (sBP), diastolic BP (dBP), PP, and pulse rate in diabetic subjects with a narrow PP or a wide PP. Mean levels of not only PP and sBP, but also sBP, in any time periods were consistently higher in diabetic subjects with a wide PP than in those with a narrow PP. Furthermore, mean 24-h PP levels for all subjects were significantly correlated with age (r = 0.257, P < 0.001). Pulse rate levels were similar, even in the nighttime, between the two groups. Although the prevalence of nondippers in the two diabetic subgroups was similar, nocturnal fall in sBP was significantly smaller in diabetic subjects with a wide PP than in those with a narrow PP.

Various vascular events that occurred during the follow-up period were summarized in Table 2. Prevalence of cardiovascular and all events were significantly higher in diabetic subjects with a wide PP than those with a narrow PP (P < 0.001 for each), whereas there was no difference in the prevalence of cerebrovascular events between the two groups. Moreover, cardiovascular events, but not cerebrovascular events, seemed to be more fatal in diabetic subjects with a wide PP than in those with a narrow PP.

Rates of occurrences of various vascular events in diabetic subjects divided by quartiles of mean 24-h PP and age stratification were illustrated in Fig. 2. In spite of age stratification, cardiovascular and all events were more frequent in the highest quartile than the other quartiles.

The Kaplan-Meier analysis of cumulative incidence of various vascular events in diabetic subjects with narrow and wide PPs revealed that there was a significant difference in cardiovascular and all events between the two groups (P < 0.001 for each, log-rank test). Relative risks for these events were 25.3 and 5.0, respectively (P < 0.001 for each, Cox-Mantel test). However, there was no statistical difference in cumulative incidences of cerebrovascular events between the two subgroups.

In the Cox proportional-hazards model, the adjusted relative risks for all factors listed are shown in Table 3. Although mean 24-h sBP (model 1) and dBP (model 2) were independent predictors for both cardiovascular and all events, mean 24-h PP was a better predictor than sBP and dBP. In model 3, mean 24-h PP as well as presence of hypertension and lipid ratio exhibited a statistically significant adjusted relative risk for cardiovascular events (P < 0.05 for each), whereas only duration of diabetes exhibited risk for cerebrovascular events (P < 0.01). Furthermore, duration of diabetes, lipid ratio, nocturnal fall in sBP, and mean 24-h PP exhibited a significant adjusted relative risk for all events (P < 0.05 for each).

In this prospective study of diabetic subjects under 60 years of age, we clearly demonstrated that a wide 24-h PP at baseline was a better predictor than sBP and dBP and independently predicted subsequent cardiovascular and all events but not cerebrovascular events. By contrast, only the duration of diabetes was the risk factor for cerebrovascular events.

Atherosclerotic changes are generally more pronounced in type 2 diabetic subjects compared with those with a normal glucose tolerance (6,7). Mean 24-h PP of relatively young diabetic subjects in this study was significantly correlated with age. Therefore, when a wide 24-h PP appears, even in young diabetic subjects, it may be indicative of large artery stiffness and thereby may be an early marker of vascular disease, especially cardiovascular events. However, histograms of distribution of mean 24-h PP showed that there was a diabetic subgroup in which 24-h PPs were wider than those thought to already have wide PPs. Thus, it is clinically important to detect and manage such a diabetic subgroup to prevent vascular events.

Many investigators have reported that nephropathy (microalbuminuria and gross proteinuria) in type 2 diabetic subjects is related to the occurrence of cardiovascular events (16,17), and even in nondiabetic subjects (18), microalbuminuria is a cardiovascular disease risk factor. Riu et al. (19) recently confirmed this relationship based on the observation that mean urinary albumin excretion during follow-up was closely associated with occurrences of cardiovascular disease in type 2 diabetic subjects. Mattock et al. (20), however, demonstrated that although microalbuminuria predicted incident clinical coronary heart disease in men with type 2 diabetes, preexisting coronary heart disease was also a risk factor for microalbuminuria in such subjects, suggesting that both factors were not causally related but rather resulted from common determinants. Thus, the causal relationship between nephropathy and cardiovascular events in type 2 diabetic subjects is not clear, but it may be clinically undoubting that type 2 diabetic subjects with nephropathy are at high risk for cardiovascular disease, suggesting that management for PP as well as sBP and dBP focused on arterial stiffness is required, even in relatively young diabetic subjects.

In the previous reports dealing with middle-aged subjects with essential hypertension (3–5), sBP increased progressively from the lower to the higher PP group, whereas dBP was significantly lower in the subgroups with the higher PP. In contrast to those findings, our data showed that sBP and dBP as well as PP in the diabetic subjects with a wide PP were consistently higher than those with a narrow PP (Table 1). Therefore, it seems that this difference could not be explained merely by age-related progression of stiffening in large arteries. Knudsen et al. (21) demonstrated that ambulatory daytime and nighttime sBP but also dBP values, when grouped according to severity of each microvascular and macrovascular complication, were consistently higher in type 2 diabetic subjects with advanced microvascular and macrovascular complications than those without them, accompanying a widening of the PPs as well as blunted diurnal BP variations. Taken together, underlying mechanisms for widening PP in relatively young diabetic subjects may differ from those in subjects with essential hypertension. Although diabetes-specific atherosclerotic changes are not known, it is feasible that widening of PP with increases in both sBP and dBP in the diabetic subjects may be involved in hemodynamic alterations associated with the development of microvascular complications. In fact, in this study, duration of diabetes was longer, and serum creatinine level was higher at baseline in diabetic subjects with a wide PP compared with individuals with a narrow PP. In addition, duration of diabetes was an independent predictor for cerebrovascular and all events. Another possibility is the blunted nocturnal fall in sBP or nondipper status, which was pronounced at baseline in the diabetic subjects with a wide PP compared with those with a narrow PP. Accordingly, altered diurnal pattern of BP may contribute to not only the progression of microvascular complications but also to the stiffening of large arteries. Finally, autonomic dysfunction causing altered vascular reactivity may be associated with the worse outcomes, although at least cardiac sympathovagal imbalance was not largely different at baseline between the two subgroups.

In this study, the cut point for mean 24-h PP between the diabetic subjects with a wide PP and a narrow PP was 53.3 mmHg, which was closely equal to the observations of Verdecchia et al. (12). They showed that both total and fatal cardiovascular event risks in untreated subjects with uncomplicated essential hypertension (mean age 51.7 years) were highest in the top tertile, who had a mean 24-h PP value wider than 53 mmHg, while the cut point of 65 mmHg for office PP was shown. Thus, this finding suggests that the ambulatory 24-h PP value is a clinically useful predictor for cardiovascular risk, even in a hemodynamically different population.

The main limitation of our study is the lack of information on the degree of BP and glycemic control during the follow-up periods in all participants and on the effect of antihypertensive drugs on BP level, pattern, and protection from organ damage in the long term. However, we made every effort to obtain excellent BP and glycemic control in each individual during the whole follow-up period by using a standard treatment.

In conclusion, this study showed that a wide 24-h PP is predictive of cardiovascular events in the relatively young and low-disease risk group, suggesting that this measure can be used as an additional tool in assisting risk stratification of young diabetic subjects, where other conventional risk factors are not helpful. Further studies are needed to establish the strategies of pharmacological and nonpharmacological management for type 2 diabetic subjects with a high cardiovascular disease risk.