Regional Intensity of Outpatient Cardiovascular Testing and Cardiac-Origin Out-of-Hospital Cardiac Arrest Authors: Hiroki Sato, Naohiko Takahashi, Naohiro Yonemoto, Tetsuya Matoba, Yoshio Tahara, for the Japanese Circulation Society Resuscitation Science Study (JCS-ReSS) Group Published: 2026 ABSTRACT BackgroundCardiac-origin out-of-hospital cardiac arrest (CO-OHCA) remains a public health challenge, and variations in its incidence may reflect differences in cardiovascular care. We investigated whether the intensity of regional outpatient cardiovascular diagnostic testing was associated with the CO-OHCA incidence among adults aged 40–74 years in Japan.Methods and ResultsThis prefecture-level ecological study (2014–2023) used the All-Japan Utstein Registry and National Database Open Data. The outcome was the age- and sex-standardized incidence ratio (SIR) of CO-OHCA. The intensity of outpatient 12-lead electrocardiography (ECG) testing was quantified as an observed/expected ratio. Among 248,633 patients, higher 12-lead ECG testing intensity was significantly associated with lower SIR after adjusting for calendar year and cardiometabolic risk factors (rate ratio [RR] 0.95 per 0.1-unit increase; 95% confidence interval [CI] 0.93–0.98; P<0.001). Conversely, higher obesity prevalence was associated with higher SIR (RR 1.35 per 10-percentage-point increase; 95% CI 1.16–1.57; P<0.001). No significant associations were seen for transthoracic echocardiography and B-type natriuretic peptide/N-terminal pro B-type natriuretic peptide testing.ConclusionsIn this nationwide ecological study, higher prefecture-level 12-lead ECG testing intensity was associated with lower CO-OHCA SIR among adults aged 40–74 years. Although causal interpretation is not possible, 12-lead ECG testing intensity may represent a prefecture-level indicator of regional outpatient cardiovascular care and offer a useful perspective for understanding geographic variations in CO-OHCA burden. Central Figure Out-of-hospital cardiac arrest (OHCA) remains a critical public health challenge worldwide. Although outcomes after OHCA have gradually improved in recent decades, the overall survival to hospital discharge remains very low.1–4 In particular, OHCA of cardiac origin among middle-aged and young-old populations warrants specific attention. This group constitutes a vital segment of the productive workforce, making such deaths a profound socioeconomic loss.5 Importantly, a proportion of cardiac-origin OHCA may be preventable through proactive cardiovascular risk assessment and evidence-based preventive interventions delivered in routine clinical care,6–8 underscoring the urgent need for targeted strategies in this age group. International studies suggest substantial geographic variation in the incidence of cardiac-origin OHCA.9 This variation has been linked to differences in population risk profiles and community socioeconomic conditions.10,11 However, these factors may not fully account for the observed regional variation in the incidence of cardiac-origin OHCA.12 This suggests that factors related to healthcare delivery may also contribute to geographic variations in its incidence.12 Therefore, this study investigated whether a higher regional intensity of outpatient cardiovascular diagnostic testing activity is associated with a lower incidence of cardiac-origin OHCA among adults aged 40–74 years. Methods Study Design and Population We conducted a nationwide prefecture-level ecological study in Japan using individual-level data from the All-Japan Utstein Registry and prefecture-level indicators from the National Database of Health Insurance Claims and Specific Health Checkups (NDB) Open Data. The study period was from 2014 to 2023. We included patients with OHCA of presumed cardiac origin aged 40–74 years in the All-Japan Utstein Registry to match the target age range of Japan’s National Specific Health Checkups program, because several prefecture-level indicators in the NDB Open Data are based on data from this program. All-Japan Utstein Registry OHCA data were obtained from the nationwide All-Japan Utstein Registry maintained by the Fire and Disaster Management Agency. This registry is a prospective, population-based registry of OHCA patients treated by emergency medical services that covers almost the entire Japanese population.13 Data quality is reviewed by the Resuscitation Science Study (JCS-ReSS) Group of the Japanese Circulation Society. The All-Japan Utstein Registry captures Utstein-style variables for all emergency medical services-treated OHCA patients transported to hospital, including patient characteristics, presumed etiology, key prehospital process measures, and short-term outcomes. The presumed etiology of OHCA is coded as cardiac or non-cardiac using standard Utstein registry definitions.13,14 The present study selected patients for whom OHCA was recorded as being of presumed cardiac origin. Patients with a presumed OHCA etiology recorded as non-cardiac were excluded, in accordance with international Utstein recommendations.14 Causes for these non-cardiac OHCA cases included trauma, drowning, asphyxia, drug overdose, exsanguination, and other external or non-cardiac medical causes. Hereafter, for brevity, we refer to OHCA of presumed cardiac origin as “cardiac-origin OHCA.” NDB Open Data Prefecture-level indicators of cardiovascular diagnostic testing intensity and cardiometabolic risk profiles were obtained from the NDB Open Data.15 The NDB is a large nationwide database that compiles health insurance claims data from almost all medical insurers in Japan and covers approximately 98% of healthcare services provided by medical institutions nationwide.16 The NDB Open Data provide annual prefecture-level counts of outpatient cardiovascular diagnostic tests, which served as the source data for deriving standardized measures of prefectural diagnostic testing intensity. The NDB Open Data also include annual prefecture-level aggregated results from Japan’s National Specific Health Checkups program, an annual screening program targeting adults aged 40–74 years that focuses on cardiometabolic risk factors. In fiscal year 2023 (April 1, 2023 to March 31, 2024), 31,230,198 of 52,103,341 eligible individuals participated, corresponding to a national participation rate of 59.9%.17 These prefecture-level annual aggregated results were used to characterize prefectural cardiometabolic risk profiles in the present analysis. Study Period and Data Sources The study period was defined as 2014 through 2023, based on the availability of NDB Open Data on outpatient cardiovascular diagnostic test counts for fiscal years 2014–2023 (April 1, 2014 to March 31, 2024). Although data in the All-Japan Utstein Registry are available from 2005, we restricted the analysis to cases of cardiac-origin OHCA occurring between January 1, 2014 and December 31, 2023 to maximize overlap with the NDB Open Data period. We linked the All-Japan Utstein Registry calendar-year prefecture-level data to NDB fiscal-year prefecture-level data using the same year label, yielding 10 annual prefecture–year intervals. The calendar–fiscal year mismatch resulted in a fixed 3-month offset at the start and end of each interval, applied uniformly across prefectures and years. Outcome The primary outcome was the annual prefecture-level incidence of cardiac-origin OHCA among adults aged 40–74 years, expressed as an age- and sex-standardized incidence ratio (SIR). For each year from 2014 to 2023, we determined the observed number of cases of cardiac-origin OHCA in this age group for each prefecture. The expected number of cases of cardiac-origin OHCA was calculated by applying year-specific national incidence rates, stratified by sex and 5-year age groups (from 40–44 to 70–74 years), to the corresponding prefecture populations and summing these values across all strata.18 Cardiovascular Diagnostic Testing Intensity Counts of outpatient cardiovascular diagnostic tests, including 12-lead electrocardiogram (ECG), Holter ECG monitoring, transthoracic echocardiography (TTE), B-type natriuretic peptide (BNP), and N-terminal pro-BNP (NT-proBNP), were obtained from the NDB Open Data and summarized for each prefecture in each fiscal year. Prefecture-level test counts are provided without stratification by age or sex, and test counts restricted to adults aged 40–74 years are not available. Accordingly, the fiscal-year total number of outpatient tests in each prefecture was used as the observed number of tests (O). The expected number of tests (E) was calculated by applying national testing rates by sex and 5-year age group to prefecture populations in the corresponding strata and summing across all age groups. An observed-to-expected (O/E) ratio was then calculated. This approach follows the general principle of indirect standardization, in which observed counts are compared to expected counts calculated by applying reference age- and sex-specific rates to the target population structure. The O/E ratio was used as a standardized indicator of regional outpatient cardiovascular testing activity.19 An O/E ratio >1 indicates that the observed number of tests was higher than expected based on national age- and sex-specific testing rates. In this study, the 12-lead ECG was prespecified as the primary indicator of prefectural cardiovascular diagnostic testing intensity because it is the most widely performed non-invasive cardiovascular test in outpatient practice and is routinely used to evaluate suspected cardiac conditions.20 Abnormal findings on a 12-lead ECG have been consistently associated with subsequent cardiovascular events, cardiovascular mortality, and sudden cardiac death.21,22 Cardiometabolic Risk Indicators From the results of Specific Health Checkups in the NDB Open Data, the total annual number of participants and the number meeting the criteria for obesity, hypertension, dyslipidemia, and diabetes were obtained for each prefecture from 2013 through 2022, stratified by sex and 5-year age groups (from 40–44 to 70–74 years). Specific Health Checkups data were available in the NDB Open Data for fiscal years 2013 through 2022, which defined the period used for these variables. The cardiometabolic risk factors were defined based on the following criteria: diabetes (HbA1c ≥6.5%); dyslipidemia (low-density lipoprotein cholesterol ≥140 mg/dL); hypertension (systolic blood pressure ≥140 mmHg); and obesity (body mass index ≥25 kg/m2). Subsequently, sex- and age-standardized prevalences for each year were calculated using an internal standard population, derived by pooling the total Japanese population data aggregated over the entire 10-year study period (fiscal years 2013–2022). These standardized prevalences were then computed for each prefecture to represent the baseline risk burden. Principal components analysis was performed on the standardized prevalences of diabetes, dyslipidemia, hypertension, and obesity to derive a prefecture-level composite cardiometabolic risk score. The first principal component was extracted and standardized as a Z-score, with higher values indicating a greater overall burden of the measured cardiometabolic risk factors. The individual cardiometabolic risk factors were also evaluated. For regression analyses, cardiometabolic risk measures were treated as time-invariant prefecture-level covariates aggregated over the study period. Statistical Analysis Prefecture-level variables across 47 prefectures were summarized as median values with an interquartile range (IQR). To examine the association between 12-lead ECG testing intensity and the incidence of cardiac-origin OHCA, a prefecture–year panel dataset for 2014–2023 (470 prefecture–year observations) was constructed. Annual counts of cases of cardiac-origin OHCA were modeled using negative binomial regression with a log link, with the log of the expected counts as an offset. Generalized estimating equations with an exchangeable working correlation structure were used to account for clustering of repeated observations within prefectures across years. In the primary analyses, the year-specific 12-lead ECG O/E ratio was modeled as a continuous variable and rescaled so that regression coefficients represented rate ratios (RRs) per 0.1-unit increase in the O/E ratio. In secondary analyses, prefectures were categorized into tertiles based on the 10-year 12-lead ECG O/E ratios, and this classification was held constant across all years. RRs were estimated for the middle and high tertiles compared with the low tertile. Three prespecified models were fitted. Model 1 included the 12-lead ECG O/E ratio and calendar year only. Model 2 additionally included a composite cardiometabolic risk defined as a standardized principal component score that summarized the prefecture-level prevalence of obesity, hypertension, dyslipidemia, and diabetes. Model 3 included the individual cardiometabolic risk factors in place of the composite risk. In all models, calendar year was included as a categorical variable to control for secular trends. The same modeling framework was applied to other cardiovascular diagnostic tests, including Holter ECG, TTE, and natriuretic peptide testing (combined BNP and NT-proBNP). As a sensitivity analysis to examine temporal ordering, we also fitted models using 1-year lagged testing intensity, linking prefecture-level cardiovascular testing intensity O/E ratio in the preceding year to cardiac-origin OHCA SIR in the subsequent year. Statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA). Maps were created using R version 4.4.1 (R Foundation for Statistical Computing, Vienna, Austria), and figures (including forest plots) were generated using Python version 3.9. Two-sided P<0.05 was considered statistically significant. Ethical Considerations This study complied with the Declaration of Helsinki and the Ethical Guidelines for Medical and Health Research Involving Human Subjects in Japan. This study was approved by the Medical Ethics Committee of Oita University (Approval no. 3375). The ethics committee waived the requirement for individual written informed consent and approved the use of an opt-out approach. Study details were publicly disclosed, and individuals were afforded an opportunity to decline participation. Results Between 2014 and 2023, 248,633 presumed cases of cardiac-origin OHCA were identified among adults aged 40–74 years (Figure 1). Table 1 summarizes the prefecture-level distributions of cardiometabolic risk prevalence, cardiovascular testing intensity, and 10-year cardiac-origin OHCA SIR in Japan. The prefecture-level distributions of cardiovascular testing intensity and cardiac-origin OHCA SIR were broadly stable over fiscal years 2014–23 (Supplementary Table 1). The annual numbers of 12-lead ECGs, Holter ECGs, and TTEs per 100,000 population-years changed little between fiscal years 2014 and 2023, and the proportions attributable to adults aged 40–74 years also remained broadly stable (Supplementary Figure 1). In contrast, BNP/NT-proBNP testing showed an increasing trend over time, observed primarily in age groups outside the 40–74 years range. Substantial prefectural variation was observed in cardiac-origin OHCA SIRs and 12-lead ECG testing O/E ratios in Japan (Figure 2). Supplementary Figure 2 shows prefecture-level testing intensity for each cardiovascular testing modality. Figure 1.Flow diagram. NDB, National Database; O/E, observed-to-expected ratio; OHCA, out-of-hospital cardiac arrest; SIR, standardized incidence ratio. Table 1.Prefecture-Level Distributions of Cardiometabolic Risk, Cardiovascular Testing Intensity (O/E), and 10-Year Cardiac-Origin OHCA SIR in Japan Median [IQR]MinimumMaximumCardiometabolic risk prevalence (%) Diabetes7.0 [6.6, 7.4]6.18.1 Dyslipidemia28.9 [28.0, 30.0]25.531.8 Hypertension19.9 [19.2, 21.1]17.222.8 Obesity26.9 [25.6, 29.2]23.938.9Composite cardiometabolic risk−0.05 [−0.76, 0.74]−1.9312.676Cardiovascular testing intensity (O/E) 12-lead ECG1.01 [0.94, 1.08]0.821.23 Holter ECG0.87 [0.81, 1.04]0.571.92 TTE0.97 [0.84, 1.08]0.671.54 BNP/NT-proBNP0.99 [0.81, 1.13]0.351.6810-year cardiac-origin OHCA SIR0.94 [0.82–1.10]0.631.25Values are presented as the median with interquartile (IQR) across the 47 prefectures. The cardiometabolic risk factors are sex- and age-standardized prevalences among adults aged 40–74 years from Specific Health Checkups data (2013–2022), directly standardized using an internal standard population pooled across all prefectures and years. The composite cardiometabolic risk score is the first principal component of the 4 standardized risk prevalences, expressed as a Z-score. The cardiometabolic risk factors were defined based on the following criteria: diabetes (HbA1c ≥6.5%); dyslipidemia (low-density lipoprotein cholesterol ≥140 mg/dL); hypertension (systolic blood pressure ≥140 mmHg); and obesity (body mass index ≥25 kg/m2). The 10-year cardiovascular testing intensity was summarized as an observed-to-expected ratio (O/E), defined as the ratio of the total observed number of outpatient tests to the total expected number derived from national age- and sex-specific testing rates over the 10-year period for each prefecture. The 10-year cardiac-origin out-of-hospital cardiac arrest (OHCA) standardized incidence ratio (SIR) is defined as O/E aggregated over the study period. BNP, B-type natriuretic peptide; ECG, electrocardiogram; NT-proBNP, N-terminal pro-B-type natriuretic peptide; TTE, transthoracic echocardiography. Figure 2.Geographic distribution of (A) the 10-year standardized incidence ratio (SIR) of cardiac-origin out-of-hospital cardiac arrest (OHCA) among adults aged 40–74 years and (B) the 10-year 12-lead ECG observed-to-expected ratio (O/E) across 47 prefectures. Expected numbers of cases of cardiac-origin OHCA and 12-lead ECG counts in the outpatient care setting were calculated using national age- (5-year bands) and sex-specific rates applied to prefecture population counts. An SIR >1 indicates a higher incidence of cardiac-origin OHCA than expected based on national age- and sex-specific incidence rates. An O/E >1 indicates higher 12-lead ECG testing intensity than expected based on national age- and sex-specific testing rates. Flow diagram. NDB, National Database; O/E, observed-to-expected ratio; OHCA, out-of-hospital cardiac arrest; SIR, standardized incidence ratio. Table 1.Prefecture-Level Distributions of Cardiometabolic Risk, Cardiovascular Testing Intensity (O/E), and 10-Year Cardiac-Origin OHCA SIR in Japan Median [IQR]MinimumMaximumCardiometabolic risk prevalence (%) Diabetes7.0 [6.6, 7.4]6.18.1 Dyslipidemia28.9 [28.0, 30.0]25.531.8 Hypertension19.9 [19.2, 21.1]17.222.8 Obesity26.9 [25.6, 29.2]23.938.9Composite cardiometabolic risk−0.05 [−0.76, 0.74]−1.9312.676Cardiovascular testing intensity (O/E) 12-lead ECG1.01 [0.94, 1.08]0.821.23 Holter ECG0.87 [0.81, 1.04]0.571.92 TTE0.97 [0.84, 1.08]0.671.54 BNP/NT-proBNP0.99 [0.81, 1.13]0.351.6810-year cardiac-origin OHCA SIR0.94 [0.82–1.10]0.631.25Values are presented as the median with interquartile (IQR) across the 47 prefectures. The cardiometabolic risk factors are sex- and age-standardized prevalences among adults aged 40–74 years from Specific Health Checkups data (2013–2022), directly standardized using an internal standard population pooled across all prefectures and years. The composite cardiometabolic risk score is the first principal component of the 4 standardized risk prevalences, expressed as a Z-score. The cardiometabolic risk factors were defined based on the following criteria: diabetes (HbA1c ≥6.5%); dyslipidemia (low-density lipoprotein cholesterol ≥140 mg/dL); hypertension (systolic blood pressure ≥140 mmHg); and obesity (body mass index ≥25 kg/m2). The 10-year cardiovascular testing intensity was summarized as an observed-to-expected ratio (O/E), defined as the ratio of the total observed number of outpatient tests to the total expected number derived from national age- and sex-specific testing rates over the 10-year period for each prefecture. The 10-year cardiac-origin out-of-hospital cardiac arrest (OHCA) standardized incidence ratio (SIR) is defined as O/E aggregated over the study period. BNP, B-type natriuretic peptide; ECG, electrocardiogram; NT-proBNP, N-terminal pro-B-type natriuretic peptide; TTE, transthoracic echocardiography. Prefecture-Level Distributions of Cardiometabolic Risk, Cardiovascular Testing Intensity (O/E), and 10-Year Cardiac-Origin OHCA SIR in Japan Values are presented as the median with interquartile (IQR) across the 47 prefectures. The cardiometabolic risk factors are sex- and age-standardized prevalences among adults aged 40–74 years from Specific Health Checkups data (2013–2022), directly standardized using an internal standard population pooled across all prefectures and years. The composite cardiometabolic risk score is the first principal component of the 4 standardized risk prevalences, expressed as a Z-score. The cardiometabolic risk factors were defined based on the following criteria: diabetes (HbA1c ≥6.5%); dyslipidemia (low-density lipoprotein cholesterol ≥140 mg/dL); hypertension (systolic blood pressure ≥140 mmHg); and obesity (body mass index ≥25 kg/m2). The 10-year cardiovascular testing intensity was summarized as an observed-to-expected ratio (O/E), defined as the ratio of the total observed number of outpatient tests to the total expected number derived from national age- and sex-specific testing rates over the 10-year period for each prefecture. The 10-year cardiac-origin out-of-hospital cardiac arrest (OHCA) standardized incidence ratio (SIR) is defined as O/E aggregated over the study period. BNP, B-type natriuretic peptide; ECG, electrocardiogram; NT-proBNP, N-terminal pro-B-type natriuretic peptide; TTE, transthoracic echocardiography. Geographic distribution of (A) the 10-year standardized incidence ratio (SIR) of cardiac-origin out-of-hospital cardiac arrest (OHCA) among adults aged 40–74 years and (B) the 10-year 12-lead ECG observed-to-expected ratio (O/E) across 47 prefectures. Expected numbers of cases of cardiac-origin OHCA and 12-lead ECG counts in the outpatient care setting were calculated using national age- (5-year bands) and sex-specific rates applied to prefecture population counts. An SIR >1 indicates a higher incidence of cardiac-origin OHCA than expected based on national age- and sex-specific incidence rates. An O/E >1 indicates higher 12-lead ECG testing intensity than expected based on national age- and sex-specific testing rates. Figure 3 shows that, after adjustment for prefecture-level cardiometabolic risk and calendar year, higher 12-lead ECG testing intensity was significantly associated with lower cardiac-origin OHCA SIR (RR 0.95 per 0.1-unit increase in the O/E ratio; 95% confidence interval [CI] 0.93–0.98; P<0.001), whereas obesity prevalence was significantly associated with a higher cardiac-origin OHCA SIR (RR 1.35 per 10-percentage-point increase; 95% CI 1.16–1.57; P<0.001). As indicated in Table 2, the inverse association between 12-lead ECG testing intensity and cardiac-origin OHCA SIR remained robust after adjustment for calendar year (Model 1) and for composite cardiometabolic risk and calendar year (Model 2). In contrast, Holter ECG, TTE, and BNP/NT-proBNP testing intensity were not significantly associated with cardiac-origin OHCA SIR in any model. Obesity prevalence was significantly associated with higher cardiac-origin OHCA SIR in all models, with similar RRs across test modalities. In the sensitivity analysis using 1-year lagged testing intensity, the inverse association between the 12-lead ECG O/E ratio and cardiac-origin OHCA SIR remained broadly consistent with the primary analysis (Model 3; RR 0.94 per 0.1-unit increase; 95% CI 0.91–0.97; P<0.001; Supplementary Table 2). Figure 3.Prefecture-level 12-lead electrocardiogram (ECG) testing intensity (observed-to-expected ratio [O/E]) and standardized incidence ratio (SIR) of cardiac-origin out-of-hospital cardiac arrest (OHCA). Forest plot showing rate ratios (RRs) and 95% confidence intervals (CIs) for cardiac-origin OHCA SIR estimated using the model adjusted for calendar year and prefecture-level cardiometabolic risk factors (Model 3). Testing intensity was expressed as the year-specific prefecture-level 12-lead ECG O/E per 0.1-unit increase. Cardiometabolic risk factor prevalences were expressed per 10-percentage-point increase. For each variable, RRs >1 indicate an association with higher cardiac-origin OHCA SIR, whereas RRs <1 indicate an association with lower cardiac-origin OHCA SIR. Table 2.Prefecture-Level Cardiovascular Testing Intensity (O/E) and Cardiac-Origin OHCA SIR 12-lead ECGHolter ECGTTEBNP/NT-proBNPRR (95% CI)P valueRR (95% CI)P valueRR (95% CI)P valueRR (95% CI)P valueModel 1 (adjusted for year only) Testing intensity (O/E)0.96 (0.93–0.99)0.0041.01 (1.00–1.02)0.1660.99 (0.96–1.01)0.2591.01 (1.00–1.02)0.251Model 2 (adjusted for year and composite risk) Testing intensity (O/E)0.96 (0.93–0.98)0.0021.01 (0.99–1.02)0.2880.99 (0.96–1.01)0.2551.01 (1.00–1.02)0.149 Composite cardiometabolic risk1.04 (0.99–1.09)0.1161.02 (0.98–1.07)0.3131.03 (0.98–1.08)0.2391.04 (0.99–1.09)0.121Model 3 (adjusted for year and separated risks) Testing intensity (O/E)0.95 (0.93–0.98)<0.0011.01 (0.99–1.02)0.2590.98 (0.96–1.00)0.0901.01 (1.00–1.02)0.103 Diabetes0.68 (0.29–1.62)0.3840.73 (0.29–1.84)0.5000.67 (0.26–1.69)0.3950.78 (0.30–2.04)0.609 Dyslipidemia1.14 (0.85–1.53)0.3971.21 (0.88–1.65)0.2421.27 (0.93–1.72)0.1361.17 (0.85–1.63)0.338 Hypertension0.93 (0.69–1.26)0.6370.95 (0.71–1.28)0.7600.93 (0.68–1.27)0.6460.95 (0.70–1.30)0.751 Obesity1.35 (1.16–1.57)<0.0011.28 (1.08–1.50)0.0041.35 (1.15–1.57)<0.0011.34 (1.12–1.61)0.001The cardiometabolic risk factors are sex- and age-standardized prevalences among adults aged 40–74 years from Specific Health Checkups data (2013–2022), directly standardized using an internal standard population pooled across all prefectures and years. The cardiometabolic risk factors were defined based on the following criteria: diabetes (HbA1c ≥6.5%); dyslipidemia (low-density lipoprotein cholesterol ≥140 mg/dL); hypertension (systolic blood pressure ≥140 mmHg); and obesity (body mass index ≥25 kg/m2). Risk factor prevalences were expressed per 10-percentage-point increase. The composite cardiometabolic risk score is the first principal component of the 4 standardized risk prevalences, expressed as a Z-score. Cardiovascular testing intensity was summarized as O/E, defined as the ratio of the observed number of outpatient tests in each prefecture in each year to the expected number derived from national age- and sex-specific testing rates. Testing intensity O/E was expressed per 0.1-unit increase. The cardiac-origin OHCA SIR was defined as the ratio of observed to expected cases, calculated separately for each prefecture in each year. The global test for calendar year was not significant in any model. Full model results, including estimates for the individual calendar year indicator variables, are presented in Supplementary Table 4. CI, confidence interval; RR, rate ratio. Other abbreviations as in Table 1. Prefecture-level 12-lead electrocardiogram (ECG) testing intensity (observed-to-expected ratio [O/E]) and standardized incidence ratio (SIR) of cardiac-origin out-of-hospital cardiac arrest (OHCA). Forest plot showing rate ratios (RRs) and 95% confidence intervals (CIs) for cardiac-origin OHCA SIR estimated using the model adjusted for calendar year and prefecture-level cardiometabolic risk factors (Model 3). Testing intensity was expressed as the year-specific prefecture-level 12-lead ECG O/E per 0.1-unit increase. Cardiometabolic risk factor prevalences were expressed per 10-percentage-point increase. For each variable, RRs >1 indicate an association with higher cardiac-origin OHCA SIR, whereas RRs <1 indicate an association with lower cardiac-origin OHCA SIR. Table 2.Prefecture-Level Cardiovascular Testing Intensity (O/E) and Cardiac-Origin OHCA SIR 12-lead ECGHolter ECGTTEBNP/NT-proBNPRR (95% CI)P valueRR (95% CI)P valueRR (95% CI)P valueRR (95% CI)P valueModel 1 (adjusted for year only) Testing intensity (O/E)0.96 (0.93–0.99)0.0041.01 (1.00–1.02)0.1660.99 (0.96–1.01)0.2591.01 (1.00–1.02)0.251Model 2 (adjusted for year and composite risk) Testing intensity (O/E)0.96 (0.93–0.98)0.0021.01 (0.99–1.02)0.2880.99 (0.96–1.01)0.2551.01 (1.00–1.02)0.149 Composite cardiometabolic risk1.04 (0.99–1.09)0.1161.02 (0.98–1.07)0.3131.03 (0.98–1.08)0.2391.04 (0.99–1.09)0.121Model 3 (adjusted for year and separated risks) Testing intensity (O/E)0.95 (0.93–0.98)<0.0011.01 (0.99–1.02)0.2590.98 (0.96–1.00)0.0901.01 (1.00–1.02)0.103 Diabetes0.68 (0.29–1.62)0.3840.73 (0.29–1.84)0.5000.67 (0.26–1.69)0.3950.78 (0.30–2.04)0.609 Dyslipidemia1.14 (0.85–1.53)0.3971.21 (0.88–1.65)0.2421.27 (0.93–1.72)0.1361.17 (0.85–1.63)0.338 Hypertension0.93 (0.69–1.26)0.6370.95 (0.71–1.28)0.7600.93 (0.68–1.27)0.6460.95 (0.70–1.30)0.751 Obesity1.35 (1.16–1.57)<0.0011.28 (1.08–1.50)0.0041.35 (1.15–1.57)<0.0011.34 (1.12–1.61)0.001The cardiometabolic risk factors are sex- and age-standardized prevalences among adults aged 40–74 years from Specific Health Checkups data (2013–2022), directly standardized using an internal standard population pooled across all prefectures and years. The cardiometabolic risk factors were defined based on the following criteria: diabetes (HbA1c ≥6.5%); dyslipidemia (low-density lipoprotein cholesterol ≥140 mg/dL); hypertension (systolic blood pressure ≥140 mmHg); and obesity (body mass index ≥25 kg/m2). Risk factor prevalences were expressed per 10-percentage-point increase. The composite cardiometabolic risk score is the first principal component of the 4 standardized risk prevalences, expressed as a Z-score. Cardiovascular testing intensity was summarized as O/E, defined as the ratio of the observed number of outpatient tests in each prefecture in each year to the expected number derived from national age- and sex-specific testing rates. Testing intensity O/E was expressed per 0.1-unit increase. The cardiac-origin OHCA SIR was defined as the ratio of observed to expected cases, calculated separately for each prefecture in each year. The global test for calendar year was not significant in any model. Full model results, including estimates for the individual calendar year indicator variables, are presented in Supplementary Table 4. CI, confidence interval; RR, rate ratio. Other abbreviations as in Table 1. Prefecture-Level Cardiovascular Testing Intensity (O/E) and Cardiac-Origin OHCA SIR The cardiometabolic risk factors are sex- and age-standardized prevalences among adults aged 40–74 years from Specific Health Checkups data (2013–2022), directly standardized using an internal standard population pooled across all prefectures and years. The cardiometabolic risk factors were defined based on the following criteria: diabetes (HbA1c ≥6.5%); dyslipidemia (low-density lipoprotein cholesterol ≥140 mg/dL); hypertension (systolic blood pressure ≥140 mmHg); and obesity (body mass index ≥25 kg/m2). Risk factor prevalences were expressed per 10-percentage-point increase. The composite cardiometabolic risk score is the first principal component of the 4 standardized risk prevalences, expressed as a Z-score. Cardiovascular testing intensity was summarized as O/E, defined as the ratio of the observed number of outpatient tests in each prefecture in each year to the expected number derived from national age- and sex-specific testing rates. Testing intensity O/E was expressed per 0.1-unit increase. The cardiac-origin OHCA SIR was defined as the ratio of observed to expected cases, calculated separately for each prefecture in each year. The global test for calendar year was not significant in any model. Full model results, including estimates for the individual calendar year indicator variables, are presented in Supplementary Table 4. CI, confidence interval; RR, rate ratio. Other abbreviations as in Table 1. In secondary analyses using tertiles of the 10-year 12-lead ECG O/E ratio, the middle tertile was not significantly associated with cardiac-origin OHCA SIR, whereas the high tertile was associated with a lower SIR in Model 3 (RR 0.85; 95% CI 0.77–0.93; P<0.001); similar results were observed in Models 1 and 2 (Supplementary Table 3). For the 10-year Holter ECG O/E ratio, the middle tertile was not significantly associated with cardiac-origin OHCA SIR, whereas the high tertile was associated with a higher SIR in Model 3 (RR 1.17; 95% CI 1.05–1.31; P=0.004); similar findings were observed in Models 1 and 2. In these tertile-based analyses, obesity prevalence was also significantly associated with a higher cardiac-origin OHCA SIR across models for all testing modalities. The global test for calendar year was not significant in any model. Results for individual calendar year indicators are presented in Supplementary Tables 4 and 5 for the continuous and tertile-based analyses, respectively. Discussion In this 10-year nationwide ecological study across 47 prefectures in Japan, higher regional intensity of outpatient 12-lead ECG testing was consistently associated with a lower standardized incidence of cardiac-origin OHCA among adults aged 40–74 years. This inverse association was robust across models with cardiometabolic risk adjustment and was also supported by tertile-based analyses. In contrast, Holter ECG testing showed an association in the opposite direction in tertile-based analyses, whereas TTE and BNP/NT-proBNP testing intensity were not significantly associated with the incidence of cardiac-origin OHCA. Obesity prevalence was consistently associated with a higher incidence of cardiac-origin OHCA across models. Calendar year effects were not statistically significant, suggesting no clear evidence of residual year-to-year variation in prefecture-level cardiac-origin OHCA SIR. The inverse association observed for outpatient 12-lead ECG testing may reflect broader outpatient cardiovascular care, healthcare accessibility, and local practice patterns rather than a direct protective effect of the ECG test itself. Because the 12-lead ECG is a basic and widely used non-invasive cardiovascular test in routine practice, higher regional use may indicate a setting in which clinically important cardiac abnormalities are more likely to be detected.20–22 Such an environment may be associated with broader opportunities for cardiovascular evaluation and risk management in outpatient care. This interpretation is consistent with recent Japanese work suggesting that routine ECG screening may contribute to the early detection of cardiovascular disease, treatment optimization, and improved prognosis.23 The different patterns observed for Holter ECG, TTE, and BNP/NT-proBNP testing may reflect differences in their clinical roles. Unlike the 12-lead ECG, which is widely used as an initial test in routine outpatient practice, these modalities are generally used more selectively after symptoms develop, initial findings are abnormal, or established cardiovascular disease is suspected.20,24–26 Higher regional Holter ECG testing intensity may indicate more frequent evaluation of patients with symptoms or initial findings suggestive of clinically relevant arrhythmias, which may explain, in part, the observed positive association with cardiac-origin OHCA SIR in the tertile-based analyses. TTE and BNP/NT-proBNP testing, in turn, are used for broader and more heterogeneous indications, which may explain, in part, why significant associations were not observed in the present ecological analysis. Of the 4 cardiometabolic risk factors examined, only obesity prevalence showed a consistent positive association with the incidence of cardiac-origin OHCA across models. These results are consistent with previous findings in Western populations linking obesity to an elevated risk of sudden cardiac death.27–29 Although sudden cardiac death and cardiac-origin OHCA are not identical outcomes, these previous findings are directionally consistent with our results. In Japanese populations, obesity has been associated with cardiovascular disease and cardiovascular mortality among middle-aged and young-old adults.30,31 However, few studies have examined its association with the incidence of OHCA in these age groups. Our findings suggest that obesity may be an important prefecture-level factor associated with the incidence of cardiac-origin OHCA among adults aged 40–74 years in Japan. Study Limitations Several study limitations should be considered. First, this was a prefecture-level ecological study, and the observed associations do not necessarily reflect individual-level relationships or establish causality. Although the 1-year lagged sensitivity analysis showed broadly similar findings, the temporal relationship between testing intensity and the incidence of cardiac-origin OHCA remains uncertain. The use of prefectures as the unit of analysis may also have masked within-prefecture heterogeneity. Second, although we adjusted for prefecture-level cardiometabolic risk factors, the possibility of residual confounding by unmeasured regional factors cannot be ruled out. Regional healthcare accessibility, socioeconomic conditions, and other healthcare system characteristics may have differed across prefectures,32–34 and may have been associated with both outpatient cardiovascular testing intensity and the incidence of cardiac-origin OHCA. Moreover, cardiometabolic risk prevalences were calculated from the Specific Health Checkup data and may not have fully reflected the underlying cardiovascular risk burden of each prefecture, given incomplete participation and the inability to account for treatment status. In addition, kidney-related indicators were not incorporated into the primary models because estimated glomerular filtration rate (eGFR)-based data were available from fiscal year 2018 onward and serum creatinine testing with eGFR assessment is classified as a detailed health checkup item performed selectively rather than measured in all participants.35 The absence of kidney-related adjustment may have contributed to residual confounding in the assessment of prefecture-level cardiometabolic risk burden. Third, because prefecture-level counts of outpatient cardiovascular testing stratified by age were unavailable in the NDB Open Data, testing intensity was defined using all-age prefecture-level O/E ratios. Although this O/E ratio was indirectly age and sex standardized to reduce the influence of regional differences in population structure, it represents all-age testing intensity rather than testing intensity specifically among adults aged 40–74 years. The measure may have been affected if testing intensity in age groups outside the 40- to 74-year age range, particularly older adults, differed substantially across prefectures. In addition, the available outpatient diagnostic test counts did not include 12-lead ECGs performed as part of health checkups or screening programs, which may have led to underestimation of overall regional ECG testing activity. Fourth, this study analyzed OHCA of presumed cardiac origin based on the Utstein-style registry classification, an approach widely used in large-scale registry-based OHCA studies.13,14,36–39 However, this operational classification does not confirm the underlying cause in each case, and some cases of OHCA may have involved non-cardiac medical causes that could not be distinguished in the registry data.40 Conclusions In this nationwide ecological study in Japan, higher prefecture-level outpatient 12-lead ECG testing intensity was associated with lower cardiac-origin OHCA SIR among adults aged 40–74 years. This association was not consistently observed for other cardiovascular diagnostic modalities. Although causal inference is not possible in this ecological design, regional differences in 12-lead ECG testing intensity may reflect differences in outpatient cardiovascular care environments at the prefecture level and provide insights into geographic variations in the cardiac-origin OHCA burden. Disclosures N.T., T.M. are members of Circulation Reports’ Editorial Team. IRB Information This study was approved by the Medical Ethics Committee of Oita University (Approval no. 3375) and was conducted under an opt-out consent procedure. Supplementary Files Supplementary Table 1. Supplementary Table 2. Supplementary Table 3. Supplementary Table 4. Supplementary Table 5. Supplementary Figure 1. Supplementary Figure 2.