2025 年 32 巻 12 号 p. 1601-1612
Aim: To investigate the association between a prolonged heart rate-corrected QT (QTc) interval on a 12-lead electrocardiogram and the risk of developing dementia and its subtypes using long-term prospective longitudinal data from a Japanese community.
Methods: A total of 1,082 residents ≥ 60 years old without dementia were followed up for 24 years. The QT interval was corrected for the heart rate using Bazett’s equation. QTc prolongation was defined as QTc ≥ 440 ms, and participants with QTc ＜440 ms were divided into tertiles. Therefore, QTc interval levels at baseline were divided into 4 ranges: ≤ 401, 402–417, 418–439, and ≥ 440 ms. The Cox proportional hazards model was used to estimate the hazard ratios (HRs) of QTc interval levels on the risk of dementia.
Results: During the follow-up period, 475 participants developed all-cause dementia, 146 had vascular dementia (VaD), and 295 had Alzheimer’s disease (AD). Compared with the lowest QTc level (≤ 401 ms), the multivariable-adjusted HRs for VaD increased significantly with longer QTc intervals (HR [95% confidence interval] 1.80 [1.05 to 3.08] for 402–417 ms, 1.93 [1.12 to 3.34] for 418–439 ms, and 2.64 [1.49 to 4.68] for ≥ 440 ms; p for trend = 0.01). No significant association was found between QTc interval and the risk of both all-cause dementia and AD.
Conclusion: The present findings suggest that QTc prolongation serves as a potential indicator for identifying individuals at a high risk of developing VaD. QTc measurement may assist in the primary prevention of VaD.
The number of people with cognitive impairment and dementia has steadily increased with increasing life expectancy. Dementia is recognized as a public health priority because of its physical, psychological, social, and economic impact1, 2), making early risk assessment and preventive risk reduction for dementia increasingly important. Recently, modifiable risk factors, such as hypertension, smoking, obesity, and diabetes, have drawn particular attention for their potential impact on future dementia development3). An impaired cardiac function resulting from long-term exposure to these risk factors, which are also risk factors for cardiovascular disease, may contribute to accelerated cognitive decline through increased cerebral hypoperfusion and atherosclerosis4, 5).
Several prospective cohort studies have reported an association between cardiac function markers, as measured by electrocardiography (ECG), and cognitive decline and incident dementia6, 7). The relationship between ECG markers of subclinical cardiac dysfunction and dementia is important for possible preventive strategies. Substantial evidence indicates that a prolonged heart rate-corrected QT (QTc) interval within normal limits is associated with an increased risk of subclinical arterial disease, cardiovascular disease, and mortality8-11). Conversely, only a limited number of prospective cohort studies have investigated the association of ventricular repolarization with cognitive decline12, 13) or dementia14, 15), and their results have been inconsistent16). Indeed, while one of the hospital-based studies of a Western population reported a significant association between a prolonged QTc interval and increased risk of non-Alzheimer’s disease14), another found Alzheimer’s disease (AD) in a Western population15). Furthermore, no community-based prospective studies have examined the association between QTc prolongation and the development of dementia in non-Western populations.
AimThe present study investigated the association between QTc prolongation and the risk of all-cause dementia, vascular dementia (VaD), and AD, using data from a long-term prospective cohort study in a general Japanese population.
Since 1961, the Hisayama Study has been conducting a long-term prospective cohort study of cardiovascular disease in the town of Hisayama, a suburb of Fukuoka City in Southern Japan17). Health examinations are conducted every one to two years. Beginning in 1985, surveillance for dementia prevalence including neuropsychological tests has also been performed among older residents at five- to seven-year intervals in conjunction with this study18).
In 1988, independent of the aforementioned health examinations and dementia prevalence survey, we conducted a screening survey of Hisayama residents for the purposes of the present analysis. A total of 1,228 residents ≥ 60 years old (90.3% of town residents in this age group) consented to participate in the screening survey examination. Among those, 31 participants with a history of dementia at baseline using data from the 1985 dementia prevalence survey and a follow-up survey from 1985 to 1988 and 4 participants who died before starting the follow-up were excluded from the present analysis. After further excluding 6 participants whose ECG recordings were not available, 17 participants who had atrial fibrillation or ventricular conduction defects (QRS interval of ≥ 120 ms), 80 participants with high heart rates (＞100 beats per minute), and 8 participants whose QTc data were missing, 1,082 participants (429 men and 653 women) were included in the present study.
The study protocol was approved by the Kyushu University Institutional Review Board for Clinical Research (Approval No. 23061-04), and all participants provided their written informed consent.
Measurements of QTcTwelve-lead ECG was performed with an FCP-270 device (Fukuda Denshi, Tokyo, Japan) in the supine position. Heart rate (beats per minute) and QT interval duration (ms) were determined using an ECG analysis software program (PI-01; Fukuda Denshi), as described previously19). The QT interval was measured from the start of the QRS to the end of the T-wave on 12-lead ECG. QTc was calculated based on Bazett’s formula QTc = QT/√RR, where RR is the RR interval in seconds20). A prolonged QTc interval was defined as ≥ 440 ms based on major clinical guidelines21), and participants with QTc ＜440 ms were divided into tertiles. Therefore, the participants were categorized into 4 QTc ranges (≤ 401, 402–417, 418–439, and ≥ 440 ms) for the present analysis.
The Diagnosis of DementiaDementia, VaD, and AD were diagnosed using the guidelines of the Diagnostic and Statistical Manual of Mental Disorders, Third Edition, Revised22), the criteria of the National Institute of Neurological Disorders and Stroke-Association Internationale pour la Recherche et l’Enseignement en Neurosciences23), and the criteria of the National Institute of Neurological and Communicative Disorders and Stroke and the Alzheimer’s Disease and Related Disorders Association24), respectively. Clinical information and neuroimaging morphology were used to define probable or possible dementia and its subtypes. For autopsy dementia cases, neuropathological and clinical information was used to clearly diagnose the dementia subtype. The detailed procedures used for the autopsy diagnosis of dementia have been reported previously25). For all participants with suspected dementia and those who died during the follow-up period, expert psychiatrists and stroke physicians discussed all available medical information to ascertain the presence or absence of dementia, the date of the onset of dementia, and subtype.
Follow-Up SurveyParticipants were followed for 24 years (December 1988 to November 2012). As previously reported18), new-onset cases of dementia and stroke were identified based on an organized, continuous surveillance network system that included our study team and local physicians or members of the town’s Health Office. In this surveillance system, physicians of the study team regularly visited hospitals and clinics to collect information on incident dementia, including suspected cases. We also obtained information from annual health examinations. For participants who did not participate in the annual health examination or had moved out of the town, we collected information on their health status through postal and telephone surveys. Once dementia or other neurological conditions were suspected, a psychiatrist and a stroke physician from the study team carefully examined the individual for the absence or presence of dementia. In addition, we conducted surveys on dementia prevalence in 1985, 1992, 1998, 2005, and 2012 to identify incident dementia cases to the greatest extent possible. When a subject died, we interviewed the family and the attending physician and reviewed all available clinical information. We also attempted to obtain consent from the family to perform an autopsy. During the follow-up period, 776 subjects died, of whom 549 (70.8%) underwent an autopsy. Except for the deceased cases, none of the subjects were lost to follow-up.
Relevant FactorsA self-administered questionnaire covering information on educational status, medical history, alcohol consumption, smoking habits, regular exercise, and daily physical activity levels was completed in advance by each participant and checked by trained interviewers. Alcohol consumption and smoking habits were classified as habitual or not. Participants engaging in sports or other forms of exertion, including recreational walking ≥ 3 times a week during their leisure time, were assigned to the regular exercise group. Daily physical activity was reported in one of four categories (mostly sitting or lying, mixed activity, walking, or heavy labor) and classified as sedentary (mostly sitting or lying) or not. Blood pressure was measured using a mercury sphygmomanometer in a seated position after a rest period of ＞5 min. Three measurements were taken, and the mean blood pressure was used in the present analysis. Hypertension was defined as a blood pressure ≥ 140/90 mmHg and/or the use of antihypertensive agents. Body height and weight were measured in light clothing without shoes, and the body mass index (BMI) was calculated as the weight (kg) divided by the height (m) squared. Obesity and leanness were defined as a BMI ≥ 25.0 kg/m2 and ＜18.5 kg/m², respectively. ECG abnormalities were defined as left ventricular hypertrophy (Minnesota code, 3-1) and/or ST depression (4-1, 4-2 or 4-3). A history of cardiovascular disease was defined as any prior event of stroke or coronary heart disease, including myocardial infarction or coronary intervention. All cardiovascular events were adjudicated based on physical examination and a comprehensive review of available clinical information, including medical records and imaging. Total serum cholesterol levels were measured using an enzymatic method. Hypercholesterolemia was defined as a serum total cholesterol level ≥ 220 mg/dL. To evaluate diabetic status, a 75 g glucose tolerance test was performed. Plasma glucose levels were measured using the glucose-oxidase method. Diabetes was defined according to the 1998 World Health Organization criteria26): fasting plasma glucose ≥ 126 mg/dL, 2-h plasma glucose ≥ 200 mg/dL, casual blood glucose ≥ 200 mg/dL, or the use of hypoglycemic agents.
Statistical AnalysesTrends across the QTc interval levels were tested using the linear regression model for mean values and logistic regression for frequencies. The incidence rates of dementia were calculated using the person-year method, and the differences among QTc interval levels were tested using a Cox proportional hazards model. The hazard ratios (HRs) and 95% confidence intervals (CIs) of dementia according to QTc interval levels were also estimated using a Cox proportional hazards model. The proportional hazard assumption was checked graphically using a log cumulative hazard plot. No violations of this assumption were observed.
In the multivariable analysis, we selected previously reported traditional or non-traditional dementia risk factors, namely age, sex, education, hypertension, heart rate, electrocardiogram abnormalities, diabetes, serum total cholesterol, body mass index, alcohol intake, smoking habits, and regular exercise. To assess the shape of the associations between QTc interval levels and the development of dementia, we used restricted cubic splines with 3 knots placed at the 10th, 50th, and 90th percentiles of QTc interval levels (386, 417, and 455 ms, respectively); 386 ms was set as the reference value. In addition, a sensitivity analysis was performed, excluding participants with a history of cardiovascular disease. Finally, we investigated the association between QTc interval levels and the risk of all-cause dementia and its subtypes, occurring either without prior stroke events or after the onset of stroke during the follow-up period. To investigate the influence of the QTc interval on the prediction ability of incident VaD, Uno’s C statistics27), continuous net reclassification improvement (NRI), and integrated discrimination improvement (IDI) were calculated before and after adding the QTc interval to the statistical model, including known risk factors for VaD28, 29).
Statistical significance was defined as a 2-tailed p-value ＜0.05. The SAS software program (version 9.4; SAS Institute, Cary, NC, USA) was used for all statistical analyses.
The age- and sex-adjusted baseline characteristics according to QTc interval levels are shown in Table 1. Individuals with a higher QTc interval were older and more likely to have hypertension, ECG abnormalities, a history of cardiovascular disease, and a current smoking habit than those with lower QTc interval levels. The mean values of systolic and diastolic blood pressure, heart rate, and fasting plasma glucose levels increased significantly across the QTc interval. The proportion of male participants and mean values of serum total cholesterol decreased significantly with increased QTc interval levels.
| Variables | Heart rate-corrected QT interval, ms | p for trend | |||
|---|---|---|---|---|---|
|
≤ 401 (n = 276) |
402–417 (n = 276) |
418–439 (n = 297) |
≥ 440 (n = 233) |
||
| Age, y | 67.8 (0.4) | 69.3 (0.4) | 70.5 (0.4) | 73.5 (0.5) | ＜0.001 |
| Men, % | 54.8 | 42.3 | 32.4 | 27.8 | ＜0.001 |
| Education ≤ 9 years, % a | 68.6 | 68.2 | 73.5 | 71.3 | 0.29 |
| Systolic BP, mmHg | 135.5 (1.4) | 140.6 (1.3) | 141.9 (1.3) | 145.1 (1.5) | ＜0.001 |
| Diastolic BP, mmHg | 74.0 (0.6) | 75.9 (0.6) | 77.1 (0.6) | 78.1 (0.7) | ＜0.001 |
| Hypertension, % | 46.8 | 54.6 | 59.5 | 65.7 | ＜0.001 |
| Heart rate, b.p.m. | 62.0 (0.6) | 65.6 (0.6) | 68.3 (0.6) | 69.8 (0.7) | ＜0.001 |
| ECG abnormalities, % | 15.1 | 21.2 | 23.1 | 26.8 | 0.002 |
| Serum total cholesterol, mg/dLb | 212.5 (2.6) | 208.8 (2.5) | 211.7 (2.4) | 199.8 (2.8) | 0.008 |
| Hypercholesterolemia, % b | 37.7 | 33.6 | 39.7 | 28.0 | 0.15 |
| Fasting plasma glucose, mg/dLc | 102.7 (1.5) | 104.8 (1.5) | 107.8 (1.5) | 108.7 (1.9) | 0.005 |
| Diabetes, % | 10.1 | 13.4 | 15.6 | 14.7 | 0.09 |
| Body mass index, kg/m2 b | 22.5 (0.19) | 22.3 (0.2) | 22.0 (0.2) | 22.2 (0.2) | 0.14 |
| History of cardiovascular disease, % | 3.0 | 6.6 | 5.3 | 9.9 | 0.006 |
| Alcohol intake, % d | 14.3 | 17.5 | 19.7 | 16.6 | 0.28 |
| Smoking habit, % | 13.6 | 15.0 | 15.1 | 24.5 | 0.02 |
| Regular exercise, % | 12.8 | 16.9 | 13.3 | 11.9 | 0.55 |
Abbreviations: BP, blood pressure; b.p.m., beats per minute; ECG, electrocardiogram.
Data are given as the mean (standard error) or percentage.
Electrocardiogram abnormalities were defined as left ventricular hypertrophy (Minnesota Code, 3-1) or ST depression (4-1, 2, 3).
Number of participants with missing value(s): a 11 for education, b 1 for serum total cholesterol, c 92 for fasting plasma glucose, d 1 for alcohol intake.
During the 24-year follow-up period, 475 patients developed all-cause dementia. Overall, 146 participants developed VaD and 295 developed AD. Of these, 35 participants in each group were diagnosed with mixed VaD and AD, and these cases were counted as events in the analyses for each subtype. Table 2 shows the incidence rates and HRs for dementia and its subtypes according to the QTc interval levels. The age- and sex-adjusted incidence rates of all-cause dementia tended to increase across QTc interval levels, with rates of 24.6, 35.3, 38.0, and 37.33 per 1000 person-years, respectively (p for trend = 0.052). When examined by dementia subtype, the age- and sex-adjusted incidence rate of VaD increased significantly with QTc interval levels (p for trend ＜0.001), whereas there was no significant association between QTc interval levels and the incidence rate of AD (p for trend = 0.78). The age- and sex-adjusted HRs for developing VaD increased significantly with QTc prolongation (p for trend ＜0.001). This association remained unchanged after additional adjustment for potential confounding factors, such as the education level, hypertension, heart rate, ECG abnormalities, diabetes, serum total cholesterol, BMI, history of cardiovascular disease, current smoking, current drinking, and regular exercise.
|
Heart rate-corrected QT interval levels (ms) |
Number of events/ person-years at risk |
Age- and sex- adjusted incidence rates (per 103 person-years) |
Hazard ratios (95% confidence intervals) |
|||
|---|---|---|---|---|---|---|
| Age- and sex-adjusted | p value | Multivariable-adjusted＊ | p value | |||
| All-cause dementia | ||||||
| ≤ 401 | 105/4552 | 24.6 | 1.00 (reference) | 1.00 (reference) | ||
| 402–417 | 121/4101 | 35.3 | 1.26 (0.97 to 1.64) | 0.09 | 1.22 (0.93 to 1.60) | 0.14 |
| 418–439 | 142/4006 | 38.0 | 1.38 (1.06 to 1.79) | 0.02 | 1.30 (0.99 to 1.71) | 0.06 |
| ≥ 440 | 107/2605 | 37.3 | 1.45 (1.08 to 1.93) | 0.01 | 1.39 (1.03 to 1.88) | 0.04 |
| p for trend | 0.05 | 0.16 | ||||
| Vascular dementia | ||||||
| ≤ 401 | 22/4552 | 4.0 | 1.00 (reference) | 1.00 (reference) | ||
| 402–417 | 39/4101 | 10.7 | 2.10 (1.24 to 3.56) | 0.006 | 1.80 (1.05 to 3.08) | 0.03 |
| 418–439 | 42/4006 | 11.8 | 2.32 (1.37 to 3.94) | 0.002 | 1.93 (1.12 to 3.34) | 0.02 |
| ≥ 440 | 43/2605 | 15.6 | 3.49 (2.02 to 6.03) | ＜0.001 | 2.64 (1.49 to 4.68) | ＜0.001 |
| p for trend | ＜0.001 | 0.01 | ||||
| Alzheimer’s disease | ||||||
| ≤ 401 | 75/4552 | 17.5 | 1.00 (reference) | 1.00 (reference) | ||
| 402–417 | 78/4101 | 23.6 | 1.05 (0.76 to 1.45) | 0.77 | 1.07 (0.77 to 1.48) | 0.65 |
| 418–439 | 85/4006 | 22.2 | 0.99 (0.72 to 1.37) | 0.97 | 1.00 (0.71 to 1.39) | 0.97 |
| ≥ 440 | 57/2605 | 19.1 | 0.87 (0.61 to 1.26) | 0.47 | 0.93 (0.63 to 1.36) | 0.73 |
| p for trend | 0.78 | 0.89 | ||||
＊The model was adjusted for age, sex, education, hypertension, heart rate, electrocardiogram abnormalities, diabetes, serum total cholesterol, body mass index, history of cardiovascular disease, alcohol intake, smoking habits, and regular exercise.
The multivariable-adjusted HR for developing VaD significantly increased with QTc prolongation (p for trend = 0.01). In contrast, although the multivariable-adjusted HR for all-cause dementia was significantly higher in the prolonged QTc group (≥ 440 ms) than in the lowest QTc interval group (308–401 ms), the trend across the 4 groups was not statistically significant (p for trend = 0.16). Meanwhile, there were no significant associations between QTc interval and the risk of developing AD. We examined the association between the QTc interval and the risk of dementia by sex (Supplementary Table 1). Compared to the lowest QTc level, the multivariable-adjusted HRs for VaD increased significantly with longer QTc intervals in both sexes (both p for trend ＜0.05). Fig.1 shows the shape of the association between the QTc interval and the risk of dementia and its subtypes using restricted cubic splines. The risk of VaD increased with increasing QTc interval levels and then plateaued above 440 ms. We also performed a sensitivity analysis by excluding individuals with a history of cardiovascular disease, because the QTc interval is influenced by pre-existing cardiovascular disease. The results of the sensitivity analysis were similar to those of the primary analysis (Supplementary Table 2).
|
Heart rate-corrected QT interval levels (ms) |
No. of events/ person- years at risk |
Age-adjusted incidence rates (per 103 person- years) |
Hazard ratios (95% confidence intervals) |
|||
|---|---|---|---|---|---|---|
| Age-adjusted | p value | Multivariable-adjusted＊ | p value | |||
| Men | ||||||
| All-cause dementia | ||||||
| ≤ 395 | 36/1939 | 24.9 | 1.00 (reference) | 1.00 (reference) | ||
| 396–412 | 48/1783 | 29.7 | 1.43 (0.93 to 2.21) | 0.11 | 1.24 (0.78 to 1.95) | 0.36 |
| 413–439 | 41/1472 | 28.7 | 1.62 (1.03 to 2.56) | 0.04 | 1.32 (0.80 to 2.19) | 0.28 |
| ≥ 440 | 18/552 | 30.6 | 2.01 (1.11 to 3.67) | 0.02 | 1.53 (0.81 to 2.90) | 0.19 |
| p for trend | 0.08 | 0.57 | ||||
| Vascular dementia | ||||||
| ≤ 395 | 6/1939 | 2.2 | 1.00 (reference) | 1.00 (reference) | ||
| 396–412 | 25/1783 | 16.1 | 4.43 (1.81 to 10.83) | 0.001 | 3.83 (1.54 to 9.55) | 0.004 |
| 413–439 | 20/1472 | 14.0 | 4.49 (1.79 to 11.26) | 0.001 | 3.27 (1.23 to 8.67) | 0.02 |
| ≥ 440 | 8/552 | 14.5 | 4.72 (1.58 to 14.16) | 0.006 | 3.32 (1.05 to 10.49) | 0.04 |
| p for trend | 0.008 | 0.04 | ||||
| Alzheimer’s disease | ||||||
| ≤ 395 | 24/1939 | 10.7 | 1.00 (reference) | 1.00 (reference) | ||
| 396–412 | 16/1783 | 10.1 | 0.71 (0.38 to 1.34) | 0.29 | 0.69 (0.35 to 1.35) | 0.28 |
| 413–439 | 17/1472 | 11.1 | 1.07 (0.57 to 2.03) | 0.83 | 1.03 (0.51 to 2.12) | 0.93 |
| ≥ 440 | 9/552 | 13.3 | 1.88 (0.83 to 4.25) | 0.13 | 1.52 (0.63 to 3.64) | 0.35 |
| p for trend | 0.17 | 0.35 | ||||
| Women | ||||||
| All-cause dementia | ||||||
| ≤ 404 | 68/2526 | 28.3 | 1.00 (reference) | 1.00 (reference) | ||
| 405–421 | 87/2625 | 41.3 | 1.18 (0.86 to 1.62) | 0.31 | 1.26 (0.91 to 1.73) | 0.16 |
| 422–439 | 88/2314 | 42.6 | 1.23 (0.90 to 1.70) | 0.20 | 1.22 (0.87 to 1.70) | 0.25 |
| ≥ 440 | 89/2053 | 42.4 | 1.27 (0.92 to 1.76) | 0.15 | 1.33 (0.94 to 1.87) | 0.11 |
| p for trend | 0.49 | 0.40 | ||||
| Vascular dementia | ||||||
| ≤ 404 | 10/2526 | 3.1 | 1.00 (reference) | 1.00 (reference) | ||
| 405–421 | 21/2625 | 9.3 | 1.94 (0.91 to 4.12) | 0.09 | 1.97 (0.91 to 4.24) | 0.08 |
| 422–439 | 21/2314 | 9.2 | 2.00 (0.94 to 4.26) | 0.07 | 2.01 (0.92 to 4.42) | 0.08 |
| ≥ 440 | 35/2053 | 16.6 | 3.46 (1.69 to 7.09) | ＜0.001 | 3.01 (1.42 to 6.37) | 0.004 |
| p for trend | 0.005 | 0.03 | ||||
| Alzheimer’s disease | ||||||
| ≤ 404 | 58/2526 | 24.2 | 1.00 (reference) | 1.00 (reference) | ||
| 405–421 | 70/2625 | 32.9 | 1.10 (0.78 to 1.56) | 0.59 | 1.21 (0.85 to 1.73) | 0.29 |
| 422–439 | 53/2314 | 27.9 | 0.84 (0.58 to 1.23) | 0.39 | 0.85 (0.57 to 1.27) | 0.43 |
| ≥ 440 | 48/2053 | 23.3 | 0.76 (0.51 to 1.12) | 0.16 | 0.84 (0.55 to 1.27) | 0.40 |
| p for trend | 0.20 | 0.19 | ||||
＊The model was adjusted for age, education, hypertension, heart rate, electrocardiogram abnormalities, diabetes, serum total cholesterol, body mass index, history of cardiovascular disease, alcohol intake, smoking habits, and regular exercise.
Solid lines represent HRs, and dashed lines represent 95% CIs. Knots were placed at the 10th, 50th, and 90th percentiles (386, 417, and 455 ms, respectively) of the QTc interval. The reference point was set at 386 ms. The QTc values under the 1st percentile (363 ms) and over the 99th percentile (500 ms) were not plotted. Risk estimates were adjusted for age, sex, education, hypertension, heart rate, electrocardiographic abnormalities, diabetes, serum total cholesterol, body mass index, history of cardiovascular disease, current smoking, current drinking, and regular exercise.
|
Heart rate-corrected QT interval levels (ms) |
No. of events/ person- years at risk |
Age- and sex- adjusted incidence rates (per 103 person-years) |
Hazard ratios (95% confidence intervals) |
|||
|---|---|---|---|---|---|---|
| Age- and sex-adjusted | p value | Multivariable-adjusted＊ | p value | |||
| All-cause dementia | ||||||
| ≤ 401 | 101/4447 | 23.8 | 1.00 (reference) | 1.00 (reference) | ||
| 402–417 | 114/3886 | 35.3 | 1.25 (0.96 to 1.64) | 0.10 | 1.24 (0.94 to 1.63) | 0.13 |
| 418–439 | 132/3800 | 37.6 | 1.38 (1.06 to 1.81) | 0.02 | 1.34 (1.01 to 1.77) | 0.04 |
| ≥ 440 | 94/2430 | 35.8 | 1.36 (1.01 to 1.83) | 0.046 | 1.36 (0.99 to 1.86) | 0.06 |
| p for trend | 0.10 | 0.18 | ||||
| Vascular dementia | ||||||
| ≤ 401 | 21/4447 | 3.9 | 1.00 (reference) | 1.00 (reference) | ||
| 402–417 | 32/3886 | 9.6 | 1.86 (1.06 to 3.25) | 0.03 | 1.70 (0.97 to 2.98) | 0.06 |
| 418–439 | 35/3800 | 10.7 | 2.12 (1.21 to 3.70) | 0.01 | 1.95 (1.11 to 3.44) | 0.02 |
| ≥ 440 | 35/2430 | 13.8 | 3.08 (1.73 to 5.50) | ＜0.001 | 2.61 (1.43 to 4.76) | 0.002 |
| p for trend | 0.002 | 0.02 | ||||
| Alzheimer’s disease | ||||||
| ≤ 401 | 72/4447 | 16.8 | 1.00 (reference) | 1.00 (reference) | ||
| 402–417 | 77/3886 | 24.3 | 1.09 (0.79 to 1.51) | 0.6 | 1.11 (0.80 to 1.54) | 0.54 |
| 418–439 | 83/3800 | 22.7 | 1.04 (0.75 to 1.44) | 0.81 | 1.04 (0.74 to 1.46) | 0.84 |
| ≥ 440 | 53/2430 | 19.3 | 0.87 (0.60 to 1.26) | 0.46 | 0.93 (0.63 to 1.38) | 0.71 |
| p for trend | 0.65 | 0.80 | ||||
＊The model was adjusted for age, sex, education, hypertension, heart rate, electrocardiogram abnormalities, diabetes, serum total cholesterol, body mass index, history of cardiovascular disease, alcohol intake, smoking habits, and regular exercise.
We also investigated the association between prolonged QTc and the risk of all-cause dementia and its subtypes either without prior stroke events or after the onset of stroke during the follow-up period, because QTc prolongation has been reported to be linked to an increased risk of stroke onset8, 9) (Supplementary Table 3). The results showed that QTc prolongation was associated with an increased risk of VaD, regardless of the presence of stroke onset during the follow-up period (Table 3).
|
Heart rate-corrected QT interval levels (ms) |
No. of events/ No. of participants |
Hazard ratios (95% confidence intervals) |
|||
|---|---|---|---|---|---|
| Age- and sex-adjusted | p value | Multivariable-adjusted＊ | p value | ||
| ≤ 401 | 47/270 | 1.00 (reference) | 1.00 (reference) | ||
| 402–417 | 46/258 | 1.02 (0.68 to 1.54) | 0.93 | 1.04 (0.69 to 1.57) | 0.86 |
| 418–439 | 67/283 | 1.38 (0.94 to 2.02) | 0.10 | 1.39 (0.94 to 2.06) | 0.10 |
| ≥ 440 | 52/211 | 1.39 (0.91 to 2.10) | 0.12 | 1.37 (0.88 to 2.12) | 0.16 |
| p for trend | 0.19 | 0.25 | |||
＊The model was adjusted for age, sex, education, hypertension, heart rate, electrocardiogram abnormalities, diabetes, serum total cholesterol, body mass index, history of cardiovascular disease, alcohol intake, smoking habits, and regular exercise.
|
Heart rate-corrected QT interval levels (ms) |
Number of events | Person-years at risk |
Hazard ratios (95% confidence intervals) |
|||
|---|---|---|---|---|---|---|
| Age- and sex-adjusted | p value | Multivariable-adjusted＊ | p value | |||
| Dementia occurring without prior stroke events | ||||||
| All-cause dementia | ||||||
| ≤ 401 | 89 | 4482 | 1.00 (reference) | 1.00 (reference) | ||
| 402–417 | 88 | 3900 | 1.06 (0.79 to 1.43) | 0.71 | 1.06 (0.79 to 1.44) | 0.70 |
| 418–439 | 111 | 3828 | 1.22 (0.91 to 1.63) | 0.18 | 1.20 (0.89 to 1.62) | 0.24 |
| ≥ 440 | 75 | 2439 | 1.12 (0.81 to 1.55) | 0.51 | 1.17 (0.83 to 1.65) | 0.37 |
| p for trend | 0.58 | 0.65 | ||||
| Vascular dementia | ||||||
| ≤ 401 | 11 | 4482 | 1.00 (reference) | 1.00 (reference) | ||
| 402–417 | 9 | 3900 | 0.95 (0.39 to 2.32) | 0.92 | 0.91 (0.37 to 2.22) | 0.83 |
| 418–439 | 20 | 3828 | 1.96 (0.91 to 4.22) | 0.08 | 1.96 (0.90 to 4.26) | 0.09 |
| ≥ 440 | 18 | 2439 | 2.52 (1.12 to 5.66) | 0.03 | 2.44 (1.05 to 5.64) | 0.04 |
| p for trend | 0.046 | 0.05 | ||||
| Alzheimer’s disease | ||||||
| ≤ 401 | 68 | 4482 | 1.00 (reference) | 1.00 (reference) | ||
| 402–417 | 72 | 3900 | 1.07 (0.77 to 1.50) | 0.69 | 1.09 (0.78 to 1.54) | 0.61 |
| 418–439 | 79 | 3828 | 1.02 (0.73 to 1.43) | 0.90 | 1.02 (0.72 to 1.44) | 0.93 |
| ≥ 440 | 45 | 2439 | 0.75 (0.51 to 1.12) | 0.16 | 0.81 (0.53 to 1.23) | 0.32 |
| p for trend | 0.31 | 0.50 | ||||
| Dementia occurring after the onset of stroke | ||||||
| All-cause dementia | ||||||
| ≤ 401 | 13 | 4482 | 1.00 (reference) | 1.00 (reference) | ||
| 402–417 | 26 | 3900 | 2.51 (1.28 to 4.91) | 0.007 | 2.35 (1.19 to 4.65) | 0.01 |
| 418–439 | 21 | 3828 | 2.22 (1.09 to 4.51) | 0.03 | 2.02 (0.98 to 4.19) | 0.06 |
| ≥ 440 | 19 | 2439 | 3.02 (1.43 to 6.39) | 0.004 | 2.52 (1.15 to 5.51) | 0.02 |
| p for trend | 0.02 | 0.07 | ||||
| Vascular dementia | ||||||
| ≤ 401 | 10 | 4482 | 1.00 (reference) | 1.00 (reference) | ||
| 402–417 | 23 | 3900 | 2.91 (1.38 to 6.14) | 0.005 | 2.64 (1.24 to 5.65) | 0.01 |
| 418–439 | 15 | 3828 | 2.09 (0.92 to 4.73) | 0.08 | 1.84 (0.80 to 4.23) | 0.15 |
| ≥ 440 | 17 | 2439 | 3.55 (1.55 to 8.12) | 0.003 | 2.79 (1.17 to 6.61) | 0.02 |
| p for trend | 0.01 | 0.06 | ||||
| Alzheimer’s disease | ||||||
| ≤ 401 | 5 | 4482 | 1.00 (reference) | 1.00 (reference) | ||
| 402–417 | 5 | 3900 | 1.10 (0.31 to 3.84) | 0.88 | 1.03 (0.29 to 3.70) | 0.96 |
| 418–439 | 4 | 3828 | 0.89 (0.23 to 3.44) | 0.87 | 0.87 (0.22 to 3.45) | 0.84 |
| ≥ 440 | 8 | 2439 | 2.47 (0.74 to 8.31) | 0.14 | 2.31 (0.66 to 8.17) | 0.19 |
| p for trend | 0.28 | 0.34 | ||||
＊The model was adjusted for age, sex, education, hypertension, heart rate, electrocardiogram abnormalities, diabetes, serum total cholesterol, body mass index, alcohol intake, smoking habits, and regular exercise.
Finally, to investigate the influence of the QTc interval on the accuracy of the risk assessment for incident VaD, we evaluated the differences in Uno’s C-statistics, continuous NRI, and IDI by adding the QTc interval to the basic model consisting of the above-mentioned potential risk factors for VaD (Supplementary Table 4). The predictive performance of the model was significantly improved by adding the QTc interval (from 0.731 to 0.742; p for the difference in Uno’s C-statistics = 0.166; continuous NRI = 0.193, p = 0.031; IDI = 0.016; p = 0.005). Significant improvements in continuous NRI and IDI were also observed when the QTc interval was added to the known-risk-factor-based model, namely the age, sex, education, hypertension, diabetes, leanness, history of stroke, smoking habits, and sedentariness30).
| Uno’s c- statistics |
p value for Uno’s c- statistics difference |
Continuous NRI (95% CI) |
p value for NRI |
IDI (95% CI) |
p value for IDI |
|
|---|---|---|---|---|---|---|
| Basic model＊ | 0.731 | 0.166 | 0.193 | 0.031 | 0.016 | 0.005 |
| Basic model + QTc | 0.742 | (0.018 to 0.368) | (0.005 to 0.027) | |||
| Honda’s risk model＊＊ | 0.744 | 0.370 | 0.192 | 0.032 | 0.013 | 0.012 |
| Honda’s risk model＊＊ + QTc | 0.751 | (0.017 to 0.366) | (0.003 to 0.023) |
Abbreviations: QTc, heart rate-corrected QT; NRI, net reclassification improvement; IDI, integrated discrimination improvement; CI, confidence interval.
＊The basic model included age, sex, education, hypertension, heart rate, electrocardiogram abnormalities, diabetes, serum total cholesterol, body mass index, history of cardiovascular disease, alcohol intake, smoking habits, and regular exercise.
＊＊Honda’s risk model included age, sex, education, hypertension, diabetes, leanness, history of stroke, smoking habits, and sedentariness.
In the present study, QTc prolongation was significantly associated with an increased risk of VaD, but not AD, in the general older Japanese population. The association remained unchanged even after adjusting for confounding factors and sensitivity analysis, excluding participants with a history of cardiovascular disease. Our findings suggest that QTc interval prolongation, even within the normal range, is an important indicator of the risk of developing VaD in the general population. However, we found that the risk of developing VaD plateaued beginning with the highest QTc interval (≥ 440 ms), possibly because individuals with QTc ≥ 440 ms might be at an increased risk of death from cardiovascular disease and coronary vascular disease10). Nevertheless, considering that the QTc interval can be measured inexpensively and noninvasively, the present study underscores the value of QTc assessment for effectively identifying individuals at a high risk for VaD.
To our knowledge, only two longitudinal hospital-based studies using data from the Copenhagen Primary Care Laboratory (CopLab) database have addressed the association between QTc interval prolongation and dementia14, 15). The first was a longitudinal study of 170,605 primary care patients (mean follow-up of 7.6 years) using data from the CopLab database14). The results showed that QTc interval prolongation was significantly associated with an elevated risk of VaD, which is consistent with our findings. With regard to the association between the QTc interval and the risk of developing AD, our findings did not agree with those of the second study, which was performed by the same team using the same database and found that QTc interval prolongation was significantly associated with a decreased rate of AD except in the highest quintile of QTc interval levels (＞439 ms)15). This inconsistency may have been due to differences in the study populations (i.e. population-based vs. primary care patients) and the methods used for diagnosing dementia.
The pathophysiology underlying the association between QTc interval prolongation and VaD has not been fully clarified; however, QTc interval prolongation, as a marker of ventricular hypertrophy10, 11) or subclinical arterial disease, has been reported to be associated with the risk of clinical cerebrovascular disease9, 19, 31, 32). Other studies have reported that QT interval prolongation is associated with white matter hyperintensity33, 34) on brain MRI, a marker of cerebral small-vessel disease, which is known to be a risk factor for dementia35). Altered or impaired cardiac function detected by QTc interval prolongation may lead to cerebrovascular insults and cerebral amyloid angiopathy, including small-vessel disease, increased risk of thrombosis, and decreased cerebral perfusion36), potentially promoting amyloid β accumulation and synergistically interacting with neurodegenerative processes to contribute to dementia37). However, in the present study, there was no significant association between the QTc interval and AD. Taken together, these findings suggest that the QTc interval on ECG may be associated with vascular damage rather than with neurodegenerative changes in the brain. In addition, in the present study, QTc interval prolongation was associated with the risk of developing VaD even in the sensitivity analysis, excluding participants with existing cardiovascular disease at baseline or the prior onset of stroke, suggesting that QTc prolongation may be an early marker for the development of VaD in the preclinical phase of cardiovascular events.
The strengths of the present study were the longitudinal community-based design, high participation rate in the baseline survey, accurate determination of dementia by expert psychiatrists, and absence of loss to follow-up among the study participants. However, our study also had several limitations. First, the influence of medications on the findings could not be sufficiently assessed. Several medications, including antiarrhythmic medications, antibiotics, antipsychotic agents, and antihistamines, can alter the QTc interval duration. Second, we had no information on the participants with congenital long-QT syndrome. However, because the prevalence of congenital long-QT syndrome has been reported to be ＜0.1%38), the influence of congenital long-QT syndrome on our results was likely negligible. Third, QTc intervals were only assessed at baseline, so the effects of changes in QTc interval duration and other risk factors were not taken into account. This limitation could lead to random misclassification of the QTc interval, which would tend to underestimate the results of the present study. Finally, the generalizability of these findings is limited, because the present study was conducted in a single Japanese community. Therefore, the findings of this study should be validated in other populations with different backgrounds.
The present study showed that QTc prolongation was significantly associated with an increased risk of developing VaD in the general older Japanese population. Our findings suggest that QTc measurement on ECG, which is an inexpensive, noninvasive, and reproducible test, might permit the identification of individuals at high risk for VaD. Further epidemiological studies are needed to validate the association between QTc prolongation and the risk of VaD and the potential use of QTc interval to predict the development of VaD.
The authors thank the residents of the town of Hisayama for their participation in the survey, and the staff of the Division of Health of Hisayama for their cooperation in this study. Statistical analyses were performed using computer resources offered under the category of General Projects by the Research Institute for Information Technology, Kyushu University.
This study was supported in part by the Ministry of Education, Culture, Sports, Science and Technology of Japan (JSPS KAKENHI Grant Number JP22K07421, JP23K09692, JP23K09717, JP23K16330, JP23K06787, JP23K09060, and JP25K13589) ; by the Health and Labour Sciences Research Grants of the Ministry of Health, Labour and Welfare of Japan (JPMH23FA1006, and JPMH24GB1002) ; and by the Japan Agency for Medical Research and Development (JP25dk0207053, JP25km0405209, JP25tm0524003, and JP25he2202021) ; by JST Grant Number (JPMJPF2210).
The authors declare that they have no conflicts of interest.
