Factors in the middle age that are associated with the risk for development of diastolic dysfunction in long term are not fully established. The Oulu Project Elucidating Risk of Atherosclerosis OPERA study randomly selected middle-aged subjects with hypertension and age- and gender-matched control subjects (n = 1,045, age 51 ± 6 years, men 49.8%). After >20 years of follow-up, majority of the subjects still alive were available for reexaminations (n = 600). After excluding the subjects with mitral regurgitation, left ventricular ejection fraction <50%, and those from whom echocardiographic septal E/E′ could not be reliably measured, the present analysis included 460 subjects. E/E’ was divided into 3 subgroups (subgroup 1: E/E’ ≤8, subgroup 2: 8 < E/E′ < 15, subgroup 3: E/E′ ≥15), subgroup 3 suggesting a significant diastolic dysfunction. Several baseline variables were associated with diastolic dysfunction: greater age (p = 0.001), female gender (p = 0.001), shorter height (p <0.001), larger body mass index (p = 0.008), greater systolic blood pressure (p = 0.001), greater pulse pressure (p <0.001), lower baroreflex sensitivity (p = 0.007), lower estimated glomerular filtration rate (p = 0.02), greater atrial natriuretic peptide (p = 0.001), greater fasting plasma glucose (p = 0.001), more common occurrence of diabetes (p = 0.011), and more common usage of antihypertensive medication (p = 0.001). After adjustments in the multivariate model, only systolic blood pressure (p = 0.001), shorter height (p = 0.002), and estimated glomerular filtration rate (p = 0.006) retained a significant association with the risk of developing diastolic dysfunction. In conclusion, greater systolic blood pressure, short height, and lower estimated glomerular filtration rate of the middle-aged subjects were the main determinants of development of diastolic dysfunction during a 20-year follow-up.
The prevalence and proportion of diastolic dysfunction rises with increasing age. Currently, diastolic heart failure is usually termed as heart failure with a preserved left ventricular ejection fraction (HFpEF). Patients with HFpEF are more often hypertensive, obese, elderly, and women compared with patients with heart failure with reduced left ventricular ejection fraction (HFrEF). They have a higher prevalence of chronic obstructive pulmonary disease, psychiatric disorders, anemia, and cancer. Many patients have also diabetes, coronary artery disease, and atrial fibrillation. The prevalence of HFpEF is rising, and it may soon be the most prevalent type of heart failure. Nevertheless, factors that predispose middle-aged subjects to diastolic dysfunction in long term are not completely understood. Therefore, we aimed to identify baseline factors that are associated with the development of diastolic dysfunction in a prospective study in which middle-aged subjects were followed up over 20 years.
Methods
The Oulu Project Elucidating Risk of Atherosclerosis study was initiated in early 1990s. For the study, hypertensive subjects were randomly selected from the Social Insurance Institute Register for reimbursement of antihypertensive medication. For each hypertensive subject, an age- and gender-matched control subject was randomly selected from the National health register excluding any subjects with the right to reimbursement for hypertension medication. The study recruited middle-aged 1,045 subjects (age 51 ± 6 years [mean ± standard deviation], age range from 40 to 59 years, 520 men of whom 261 were hypertensive and 259 controls, and 525 women of whom 258 were hypertensive and 267 controls). Initially, the subjects who had had myocardial infarction or stroke were excluded. After the exclusions, the study included 1,004 subjects. The study was designed to assess the risk factors and disease end points of cardiovascular diseases. The details of the study population are published earlier. At baseline, the study subjects (n = 1,004) underwent thorough clinical examinations including a standardized blood pressure assessment, laboratory tests, an evaluation of autonomic cardiovascular regulation, and an echocardiographic examination. After over 20 years of follow-up, of the 813 subjects still alive, majority (n = 600) were available for reexaminations, such as echocardiography using modern techniques. After excluding the subjects with mitral regurgitation, left ventricular ejection fraction <50%, and those from whom septal E/E′ could not be reliably measured, 460 subjects were included in the present analysis. E wave in the pulsed Doppler registration describes the early mitral inflow in diastole, and E′ in the tissue Doppler registration measures the mitral annular longitudinal motion in early diastole. The ratio of E to E′ is considered to be one of the best echocardiographic measurements of diastolic dysfunction, and patients with HFpEF and E/E′ >15 after medical therapy have increased risk of death and heart failure hospitalization. The study was approved by the Ethical Committee of the Medical Department of the University of Oulu.
At baseline, blood pressure was measured from the right arm in a sitting position after an overnight fast and after 10 to 15 minutes rest using an automatic oscillometric blood pressure recorder (Dinamap, Critikon Ltd). The blood pressure was measured 3 times at 1-minute intervals, and the means of the last 2 measurements were used in the analyses. Similar protocol was used in the blood pressure measurements during the follow-up visit.
The subjects performed the Valsalva maneuver for the assessment of baroreflex sensitivity at baseline. Beat-to-beat arterial blood pressure was recorded noninvasively using the Finapres finger-cuff method (Ohmeda Inc., USA). Electrocardiographic and continuous blood pressure data were analyzed later using a menu-driven software package (CAFTS, Medikro Oy, Kuopio, Finland). Baroreflex sensitivity was calculated from the overshoot phase after the Valsalva maneuver as the slope of the linear relation between the RR interval and the preceding systolic blood pressure.
At baseline, the same experienced cardiologist performed echocardiographic measurements blinded to the patients’ clinical data. The echocardiographic measurements were done using a Hewlett-Packard 77020A ultrasound color system for M-mode, 2-dimensional, and Doppler examinations by a previously described method. The M-mode measurements were based on the American Society of Echocardiography recommendations. The formula of Troy was used to calculate the left ventricular mass, and the left ventricular mass index was determined by dividing left ventricular mass by body surface area. After the follow-up, the study subjects had an echocardiographic examination in a core laboratory using a GE Healthcare Vivid E 9 version 110. x.x ultrasound system. Standard and modern parameters including tissue Doppler–based measurements were determined according to the American Society of Echocardiography recommendations.
The one-way analysis of variance was used to assess whether continuous variables differed significantly between the subjects divided in 3 subgroups based on the E/E’ values (subgroup 1: E/E’ ≤8, subgroup 2: 8 < E/E’ < 15, subgroup 3: E/E’ ≥15). The statistical significance of categorical variables between the subgroups was evaluated using the chi-square test. The power of different factors to predict diastolic dysfunction after adjusting with other relevant univariate predictors was assessed in the logistic regression analysis multivariate model. An E/E’ value ≥15 was used as a criterion for diastolic dysfunction in the logistic regression analysis. The association of different echocardiographic variables with E/E’ in the follow-up examination was assessed using the Pearson correlation coefficients. The data were analyzed using the IBM Statistics software 21. A p value <0.05 was considered to be statistically significant.
Results
Several baseline variables differed significantly between the E/E’ subgroups after over 20 years of follow-up ( Table 1 ). Baseline higher age, female gender, shorter height, larger body mass index, greater systolic blood pressures, greater pulse pressure, lower baroreflex sensitivity, lower estimated glomerular filtration rate, greater atrial natriuretic peptide, greater fasting plasma glucose, more common occurrence of diabetes, and more common usage of antihypertensive medication all showed an association with the development of diastolic dysfunction. After adjustments in the multivariate logistic regression analysis model, only systolic blood pressure (relative risk [RR] 1.047, 95% confidence intervals [95% CI] 1.018 to 1.076, p = 0.001), height (RR 0.897, 95% CI 0.837 to 0.962, p = 0.002), and estimated glomerular filtration rate (RR 0.935, 95% CI 0.892 to 0.981, p = 0.006) retained a significant association with the risk of developing diastolic dysfunction. If only height and gender were entered in the model together, only height had a significant association with the occurrence of diastolic dysfunction (RR 0.934, 95% CI 0.897 to 0.992, p = 0.026; RR 1.748, 95% CI 0.547 to 5.582, p = 0.35, respectively).
