Diabetes mellitus (DM) and hypertension (HT) frequently coexist. Increased central aortic pressures indexes are associated with HT; however, possible associations of these indexes with future development of DM have never been studied in HT.
We recruited 178 patients with uncomplicated nondiabetic HT in this study. Baseline glucose, insulin, lipid profiles, and central aortic pressure indexes obtained using applanation tonometry were measured at the beginning of the study. Patients were followed for new-onset DM.
After a mean follow-up period of 31 ± 12 months, 22 patients (12.4%) developed new-onset DM. In multivariate regression analyses adjusted for age, sex, and mean blood pressure (BP) in model 1, we found that central systolic BP (CSBP; hazard ratio 1.24, 95% CI 1.10–1.41, P < 0.001), and augmentation index (AIx) corrected at heart rate 75/min (AIx75; hazard ratio 1.58, 95% CI 1.11–1.58, P < 0.05) were independent predictors for new-onset DM. After adjustment for age, sex, mean BP, glucose concentration, and β-blocker use in model 2, we found that CSBP (hazard ratio 1.36, 95% CI 1.19–1.55, P < 0.001) and AIx75 (hazard ratio 1.71, 95% CI 1.16–2.52, P < 0.01) were independent predictors for new-onset DM.
CSBP and AIx75 were independent factors for future DM in essential hypertensive patients. Increased central pressure indexes were associated with risk of DM in essential hypertension.
The prevalence of hypertension (HT) and diabetes mellitus (DM) is increasing worldwide.1,2 HT increases cardiovascular disease risk. HT affects 20–60% of all diabetics, contributing to up to 75% of deaths due to cardiovascular disease.3 DM is a state of accelerated atherosclerosis. Patients with coexisting HT and DM may have increased cardiovascular disease risk. Observational studies have identified risk factors for type 2DM in the general population, such as age, obesity, family history of DM, physical inactivity, diet pattern, and increased insulin resistance.4,5 Even HT is a risk factor for development of DM.6 However, factors for development of new-onset DM in hypertensive patients have not been well elucidated.
Most of the information regarding new-onset DM in HT came from large trials of antihypertensive agents.7,8,9 Thiazide,7 β-blocker,8 and angiotensin receptor blocker or angiotensin-converting enzyme inhibitor9 may affect glucose homeostasis and have been found to be associated with increased or decreased risk of future DM.10
Recently, the Hoorn study showed that increased arterial stiffness was associated with deteriorating glucose tolerance status,11 and increased central aortic stiffness was associated with impaired glucose metabolism and type 2DM.12 The roles of central aortic pressures and wave reflection indexes in patients with HT for future development of DM have never been addressed. This study tried to identify the possible role of baseline central aortic pressure indexes in the development of new-onset DM among hypertensive subjects.
Study population. The study subjects came from a previous study regarding target organ damage in uncomplicated essential HT.13 A total of 205 patients from an HT clinic were initially enrolled and remaining 178 patients formed the basis for this study. In total, 27 subjects lost follow-up after the initial visit. The demographic characteristics were similar between patients with or without follow-up. All patients received the appropriate evaluation to exclude secondary HT. Blood pressure (BP) was measured using a standard sphygmomanometer with the patients in sitting position after the patients had rested for at least 5min. HT was diagnosed if systolic BP (SBP) was >140mmHg or diastolic BP (DBP) was >90mmHg on two separate occasions or patients had taken antihypertensive medications. Patients with diabetes, duration of HT >5 years, coronary artery disease, severe valvular heart disease, vascular disease, renal dysfunction (plasma creatinine concentration >1.5mg/dl), or liver disease were excluded. Patients with fasting glucose >126mg/dl, or postprandial glucose >140mg/dl, or HbA1c >6.5, were also excluded at baseline. After initial evaluation, all of the subjects were followed every 3 months in our clinics. The study end point was the development of new-onset DM. DM was diagnosed if subjects had fasting glucose of ≥126mg/dl, or if subjects had been prescribed with oral hypoglycemic agents or insulin during follow-up. Patients who had been followed for at least 3 months formed the basis of this study. Informed consent was obtained at the beginning of this study, and the study protocol was approved by ethics committee in our medical center.
Measurement of insulin resistance. After each patient had fasted overnight, their blood glucose and plasma insulin levels were measured. Insulin was measured using a radioimmunoassay. The insulin resistance index was assessed using a homeostasis model assessment index and calculated as follows: homeostasis model assessment index = fasting plasma insulin (µU/ml) × fasting blood sugar (mmol/l)/22.5.14
Measurement of carotid-to-radial pulse wave velocity and central aortic pressure indexes. Before any testing, all measurements were made with the patient supine for 20min in a quiet, temperature-controlled laboratory at 26 ± 1°C. The right radial and carotid pulse waves were detected directly using a piezoresistive pressure transducer (Millar SPT 301; Millar Instruments, Houston, TX) coupled to an electronic sphygmometer (SphygmoCor Px Aortic BPWaveform Analysis System; AtCor Medical, West Ryde, Australia). The timing of these waveforms was compared with that of the R wave on a simultaneously recorded electrocardiogram. The carotid-to-radial pulse wave velocity was calculated by dividing the transit time by the distance between these two pulses. Central aortic pressure indexes were obtained at the same time by transforming the radial arterial waveform. Central SBP (CSBP), central DBP, and pulse pressure could be obtained. The principle of this noninvasive method consists of registering a pulse waveform at the radial artery and deriving it at the ascending aorta using a mathematical transformation, expressed as the central augmentation pressure, augmentation index (AIx), and AIx corrected at heart rate 75beats/min (AIx75). The central aortic pressure wave is composed of a first (forward) pressure wave and a second (reflected) pressure wave. The augmentation pressure represents the difference between the second and first systolic peaks of the central pressure waveform. The AIx is defined as the percentage of the central pulse pressure attributed to the reflected pulse wave and, therefore, reflects the degree to which central aortic pressure is augmented by wave reflection.15
Statistical analysis. We used the LIFEREG procedure in the SAS software package for our interval-censored and right-censored data.16 The parameters were estimated by the maximum likelihood, using the Newton–Raphson algorithm in the Weibull proportional hazard model. Except for a priori confounders, such as age, gender, and mean BP, we selected other potential confounders using 10% change-in-estimate criteria.17 All analyses were conducted by SAS software, version 9.1 (SAS Institute, Cary, NC).
After a mean follow-up period of 31 ± 12 months, 22 patients (12.4%) developed new-onset DM. First new-onset DM was noted half year after initial evaluation. The study populations were chiefly middle-aged, overweight, and male gender (Table 1). The peripheral BPs, central BPs, and pressure wave reflection indexes are listed in Table 1. The most common antihypertensive drugs were calcium-channel blockers (55.6%) and β-blockers (44.4%). Relatively fewer patients (9.0%) used diuretics as the first-line therapy.
Predictors of new-onset DM
We graphed the log–log survival function by survival time for our hypothesis and the result showed parallel lines to meet the proportional hazard assumption. Table 2 shows the results of univariate analysis for the association between covariates and new-onset DM. Using multivariate regression analysis adjusted for a priori confounders (age, gender, and mean BP) in model 1, we found that CSBP (hazard ratio 1.24, 95% CI 1.10–1.41, P < 0.001) and AIx75 (hazard ratio 1.58, 95% CI 1.11–1.58, P < 0.05) were independent predictors for new-onset DM (Table 3). In model 2, glucose concentration and use of β-blocker were further added as covariates. CSBP (hazard ratio 1.36, 95% CI 1.19–1.55, P < 0.001) and AIx75 (hazard ratio 1.71, 95% CI 1.16–2.52, P < 0.01) were still independent predictors for new-onset DM (Table 3).
Our present observational study demonstrated that CSBP and AIx75 were independent predictors for future DM in essential HT. Central aortic pressure indexes may play an important role in development of new-onset DM in essential HT.
Predictors for new-onset DM in HT
Age, a history of DM in parents or siblings, baseline serum glucose concentration, body mass index, waist circumference, waist-to-height ratio, and waist-to-hip ratio are all risk factors for DM in general populations.18,19 The history of HT is one of nonanthropometric parameters associated with new-onset DM.18,19,20 There have been few studies to elucidate the predictors of new-onset DM in hypertensive patients. Further risk-stratification for new-onset DM in hypertensive patients is clinically important. We have shown that central BP indexes, in addition to other traditional risk factors, such as age and baseline serum glucose concentration, are independent predictors of new-onset DM in hypertensive subjects. These central BP indexes, obtained using a noninvasive, simple, time-saving, nonexpensive modality, seem to be surrogate markers for new-onset DM in essential HT.
Central BPs, pressure wave reflection indexes, and DM
Central BPs might be of higher clinical importance than peripheral BPs for development of target organ damage in HT.21,22 The assessment of central BPs, together with central pressure wave reflection indexes, may give new perspectives and additional data for the risk stratification of new-onset DM in the management of essential HT.23
The proposed mechanisms for the association between central BPs and pressure wave reflection indexes and new-onset DM were possibly due to differences in: (i) aortic pulse wave velocity as a consequence of changes in arterial stiffness; (ii) the proximity of pressure wave reflection sites, and (iii) the timing of systolic ejection resulting from differences in heart rate. However, our study cannot provide evidences for these mechanisms. Increased central pressure indexes and new-onset DM may simply coexist.
Impaired glucose metabolism usually precedes the development of overt type 2DM. Prolonged exposure to hyperglycemic conditions can lead to increased peripheral arterial resistance via collagen crosslinking due to nonenzymatic glycation24,25 and endothelial dysfunction.26 Abnormalities in peripheral vascular resistance may have deleterious consequences for arterial stiffness, and microvascular dysfunction may in turn be further aggravated by increased transmission of the forward wave into microcirculation. In this study, baseline glucose concentration and homeostasis model assessment index were not significantly associated with CSBP and AIx75 (data not shown). However, the possibility of impaired glucose metabolism in new-onset DM subjects at baseline still could not be eliminated, as we did not have oral glucose tolerance test in the study.
The relation between essential HT and insulin resistance is complex.27,28 HT may deteriorate insulin resistance8 but contrarily, insulin resistance may predispose to HT.27 Increased central pressure indexes was noted not only associated with HT but also in patients with impaired glucose metabolism and type 2DM.11,12,28 Increased central BPs and pressure wave reflection indexes in HT patients with new-onset DM were not completely mediated by insulin resistance based on our study results. More studies to identify other underlying processes involved in increased central pressure indexes associated with prediabetes are warranted in hypertensive patients.
Antihypertensive medications could have effects on the development of DM.4,7,8,9,10 In our study, only the use of β-blocker was associated with increased risk of new-onset DM. The small numbers of new-onset DM and numbers of subjects treated with each drug class limit conclusions regarding drug mechanisms and effects in our study.
In a subanalysis of the Candesartan Antihypertensive Survival Evaluation in Japan (CASE-J) Trial,29 they found that peripheral pulse pressure was an independent predictor for new-onset DM in high-risk Japanese hypertensive patients. In comparison with that study (mean age 64 years), our study population was younger (mean age 41 years) and did not show association between pulse pressure and new-onset DM. Pulse pressure and AIx75 are only indirect markers of arterial stiffness. However, both studies have shown the association between arterial properties and development of new-onset DM in HT.
The limitations of this study are as follows. The first is the relatively low number of end points and thus results may not be reliable. The power of the multivariate analysis model may be overestimated. The second is the deficiency of oral glucose tolerance test in our study population initially and follow-up periods. Third, we measure carotid-to-radial pulse wave velocity, not carotid–femoral pulse wave velocity, which is a direct measure of arterial stiffness. Fourth, the interval of 3 months between visits was long, which may contribute to inaccuracies of hazard ratios and evaluation of model assumptions. Finally, our study does not control for antihypertensive agents, we cannot completely eliminate the possible effects of drug modifications. However, our study still demonstrates important information by careful evaluation of most factors and by using multivariate analysis.
Our study showed that 12.4% of subjects with essential HT developed new-onset DM after a mean follow-up period of 31 months. Increased central systolic BP and pressure wave reflection index were associated with future development of new-onset DM in essential HT.
This study was supported by grant NSC 97-2314-B-006-054 from the National Science Council, Executive Yuan, Taipei, Taiwan.
The authors declared no conflict of interest.