Abstract Aims ST-depression at 24hECG has not been studied in relation to atrial fibrillation (AF) risk. We aimed to determine whether ST-depression at 24hECG was associated with incident AF in two Swedish population-based cohorts - a sub-cohort of the Malmö Diet and Cancer study (MDCS), and the cohort ‘Men born in 1914’, and to determine whether 24hECG could be used to predict AF development. Methods and results There were 378 acceptable 24hECG recordings in the MDCS (mean age 64.5 years, 43% men) and 394 acceptable recordings in ‘Men born in 1914’ (mean age 68.8 years). Incidence of AF was monitored using national registers of hospitalizations and outpatient visits in Sweden. Mean follow-up ± SD (cumulative incidence) was 10.4 ± 2 years (11.3%) in MDCS, and 10.9 ± 4 years (7.3%) in ‘Men born in 1914’. ST-depressions were independently associated with incident AF; hazard ratio (HR) (95% CI) 2.41 (1.29–4.50, P = 0.006) and 2.28 (1.05–4.95, P = 0.038) after adjustment [age, sex, height, weight, systolic blood pressure, smoking, anti-hypertensive drugs, LDL/total cholesterol, and HOMA-IR (in MDCS)]. AF incidence was substantially lower in individuals who had neither ST-depressions or high supraventricular activity (SVA, negative predictive value 0.97 and 0.94, in MDCS and ‘Men born in 1914’, respectively), and similar in men and women. Conclusion ST-depression at 24h-ECG is independently associated with incident AF, and incidence is substantially lower in individuals with neither ST-depression or high SVA. 24hECG can be used not only to diagnose AF but also to identify individuals at high and low AF risk. Atrial fibrillation, 24-hour electrocardiography, Holter ECG, ST-segment depression, Population , Prediction What’s new? ST-segment depressions at 24hECG predict incident atrial fibrillation. The negative predictive value of 24hECG is high. 24hECG may be a useful for atrial fibrillation screening as well as identifying individuals at low risk for atrial fibrillation. Introduction Even though atrial fibrillation (AF) is a serious public health problem there are no validated strategies for primary prevention of the disease.1 The lack of such measures can in part be attributed to difficulties in prediction. It has recently been shown that supraventricular activity (SVA) at 24hECG recordings is associated with increased incidence of AF,2,3 but other 24hECG parameters have not been studied as regards AF risk, nor has the predictive value of 24hECG for future development of AF been studied. Episodes of ST-depressions during 24hECG are generally considered to be a sign of myocardial ischemia and population-based studies have shown associations between ST-segment depression and increased cardiovascular mortality.4 Coronary artery disease (CAD) and markers of subclinical atherosclerosis are known AF risk factors.5 Therefore, ST-depressions at 24hECG could hypothetically be a risk factor/risk marker for AF, although this has not previously been studied. The aim of this study was to determine whether ST-depressions at 24hECG was associated with incident AF in two prospective population-based cohorts, the Malmö Diet and Cancer Study (MDCS) cohort and the cohort ‘Men born in 1914’, during long-term follow-up. We also aimed to determine whether combining information on ST-depressions and SVA from a 24hECG could be used to improve prediction of future AF. Methods Study populations The study is based on data from two prospective cohorts, derived from the MDCS and the ‘Men born in 1914’ cohorts. The MDCS consists of 30 447 individuals (age range 44–73 years; 40% men).6 A cardiovascular sub-study was conducted in a random sample (n = 6103, 40% men) of those with complete data on dietary registration (n = 28 098).7 From this population 909 (42% men) individuals were randomly invited based on homeostasis model assessment of insulin resistance (HOMA-IR). High HOMA-IR (defined as the sex-specific 75th percentile) was slightly oversampled (15% from each of quartiles 1 and 2, 30% from quartile 3 and 40% from quartile 4). From this sub-population a random subsample of 389 individuals underwent 24hECG screening in 1998–2000. Finally, 378 remained after exclusion of cases of prevalent AF and inadequate recordings. The derivation of the study cohort has been described elsewhere,3 and is outlined in Figure 1. Figure 1 View largeDownload slide Derivation of study population in the MDCS. Figure 1 View largeDownload slide Derivation of study population in the MDCS. The ‘Men born in 1914’ cohort has been described in more detail elsewhere.4 Briefly, it is based on invitation of a random 50% sample of the men born in 1914 who were residing in the city of Malmö, Sweden in 1982 (621 men aged 68 years). Of the invitees 500 (80.5%) agreed to participate, of whom 456 were able to record a 24hECG. Subjects with recordings where ST-segment analysis was not possible (n = 62) due to bundle branch block (n = 10), technical reasons (n = 2), AF (n = 19) and digoxin use (n = 31), were excluded. The final study population thus consisted of 394 men. Data collection Both cohorts included physical examinations and blood sampling after an overnight fast, at baseline and follow-up investigations. Blood samples were analysed using standard laboratory procedures at the Malmö University Hospital. In the MDCS a self-administered questionnaire was used at baseline and from this smoking status, physical activity score and self-reported alcohol use based on 7-day dietary registration was derived. Leisure-time physical activity and alcohol use were derived from the baseline questionnaire. Current smoking was defined as self-reported smoking at the time of 24hECG. Use of anti-hypertensive medication was defined as a positive answer to questions regarding current use of ß-blockers, calcium-channel blockers, ACE-inhibitors or diuretics. The physical activity score in the MDCS was calculated through a previously described modification of the Minnesota Leisure Time Physical Activity Questionnaire.8 Low physical activity was defined as the lowest sex-specific quartile in the MDCS. In ‘Men born in 1914’ low physical activity was defined as those who reported that the mostly were engaged in sedentary activities in leisure time. Alcohol use was self-reported in the context of a self-documented questionnaire, and measured in g/week. Height (in cm) and weight (in kg) were measured standing in light indoor clothes but no shoes. Blood pressure (mmHg) was measured after 10 min supine rest in the MDCS and after 15 min seated rest in ‘Men born in 1914’. In the MDCS 24hECG was recorded with a 3-lead (X, Y, Z coupling), 256 Hz sampling rate for arrhythmia detection, using a Lifecard CF digital Holter recorder with 12-bit resolution. Analyses were performed using the Pathfinder SL analysis tool (both from Spacelabs Healthcare, Issaquah, WA). The 24hECG in ‘Men born in 1914’ were recorded using an Oxford Medilog II frequency-modulated ECG tape recorder (0.05–100 Hz) with two pairs of bipolar electrodes in the V2–V6 location, and a ground electrode under the right clavicle. Analysis was performed using the Oxford Medilog MA 20 ECG analyser. ST-segment analysis in ‘Men born in 1914’ has been described in detail previously.4 Mean analysis time was 23.2 h in the MDCS and 23.4 h in ‘Men born in 1914’. On the day of recording no anti-hypertensive medications were used. ST-depressions were defined as the presence of any episode of planar or downsloping shift of 0.1 mV or greater from the isoelectric baseline, occurring at least 0.08 s after the J-point and with a duration of at least 30 s in the ‘Men born in 1914’ cohort and at least 60 s in the MDCS cohort. A high level of SVA was defined as either the top quartile of number of supraventricular extrasystoles (SVESs) or the top quartile of number of supraventricular tachycardia (SVT) episodes in the MDCS and as the top quartile of SVESs in ‘Men born in 1914’.9 Endpoint definition The endpoint was clinical AF diagnosed in a hospital setting, and was retrieved from the Swedish Registers for in- and outpatients, administered by The Swedish National Board of Health and Welfare. The inpatient register has been in use in the south of Sweden during the entire follow-up period and the register became nation-wide in 1987. The outpatient register has been maintained since the year 2000. The AF diagnosis in MDCS has recently been validated and found to have high diagnostic validity.6 Participants were followed until first episode of AF (diagnosis codes 427D for the 9th revision of International Classification of Diseases, ICD-9, and I48 for the 10th revision, ICD-10), or until death or censoring by emigration from Sweden. Follow-up ended at 31st of December 2010 in the MDCS and at 31st of December 1996 in ‘Men born in 1914’. This date was chosen in order to provide similar follow-up times in both cohorts. A sub-analysis with follow-up until the death of the last participant (1 May 2013) was also performed. AF and atrial flutter have not been distinguished due to the similarities of these diagnoses. The study conforms to the Declaration of Helsinki and the regional ethics review board in Lund has approved the study. Statistical methods All analyses were performed using Stata for windows version 12.1. Continuous variables were assessed for normality, and skewed variables (alcohol use and HOMA-IR) were log-transformed with the natural logarithm before inclusion in analyses. The association between ST-depressions and incident AF was analysed using multivariable adjusted Cox regression. The proportional hazards assumption was assessed using cumulative hazard plots. Established risk factors for AF and other plausible confounders were selected for inclusion in multi-variable models. Two pre-specified models were used; Model 1 adjusted for age and, in MDCS, gender. Model 2 included further adjustment for systolic blood pressure, height, weight, LDL-cholesterol, current smoking, anti-hypertensive medication, and HOMA-IR in the MDCS; and height, weight, systolic blood pressure, current smoking, anti-hypertensive medication, and total-cholesterol in ‘Men born in 1914’. We assessed the influence of alcohol use and leisure-time physical activity in the MDCS cohort by adding these variables to Model 2. Alcohol use was included as g/week after transformation by the natural logarithm after the addition of a small constant, 1. These variables did not influence the association between ST-depressions and AF, and were therefore left out of the final model. In order to assess the possibility of an interaction between ST-segment depressions and gender as regards AF risk an interaction term was added to Model 2 in the MDCS cohort. A P-value of <0.05 was considered statistically significant. The predictive power of SVES, SVT was analysed using the likelihood ratio test as well as Harrell’s C-statistic with 95% confidence interval (CI). Furthermore, the negative predictive value (NPV) and positive predictive value (PPV) (with 95% CI) for incident AF during follow-up time was calculated for ST-depressions, high levels of SVA as well as a combination of the two. Results Mean follow-up time ± standard deviation (SD) was 10.4 ± 2 years in the MDCS and 10.9 ± 4 years in ‘Men born in 1914’. Mean age at baseline was 64.5 years in MDCS and 68.8 years in ‘Men born in 1914’. There were 43 incident cases of AF (53.5% men) in MDCS and 29 cases in ‘Men born in 1914’. Cumulative incidence was 11.3% in MDCS and 7.3% in ‘Men born in 1914’. Baseline characteristics are otherwise reported in Table 1. Kaplan–Meier survival curves for AF are reported in Figures 2 and 3. Table 1 Baseline characteristics MDCS Men born in 1914 All Men Women Number, n 378 171 207 394 Male gender, % 43 100 Age, years 64.5 (5.9) 64.6 (6.0) 64.5 (5.7 68.8 (0.7) Height, cm 168.9 (9.3) 176.3 (6.8) 162.8 (6.3) 174.2 (6.5) Weight, kg 77.4 (13.4) 83.9 (12.2) 72.0 (11.8) 75.6 (10.9) BMI, kg/m2 27.1 (3.9) 27.0 (3.5) 27.1 (4.3) 24.9 (3.1) HOMA-IR, median (IQR) 2.1 (1.8) 2.2 (1.8) 1.7 (1.7) SBP, mmHg 143.7 (18.5) 145.1 (18.6) 142.1 (18.3) 153.6 (21.5) DBP, mmHg 88.2 (9.6) 90.3 (9.7) 86.5 (9.3) 92.8 (10.7) S-LDL, mmol/L 3.7 (0.8) 3.7 (0.7) 3.8 (0.9) S-HDL, mmol/L 1.5 (0.4) 1.4 (0.3 1.6 (0.4) S-cholesterol, mmol/L 6.2 (1.1) 6.0 (1.0) 6.4 (1.1) 6.1 (1.0) S-triglycerides, mmol/L 1.4 (0.8) 1.5 (0.7) 1.4 (0.8) 1.5 (0.7) Alcohol, g/daya 10.5 (12.4) 15.1 (15.2) 6.7 (7.6) 14.2 (21.8) Current smokers, % 23 26 19 46 Anti-hypertensive treatment, % 36.5 43.7 31.8 2.5 ST-segment depressions, % 27 22 31 25 High levels of supraventricular activityb, % 34 35 34 25 MDCS Men born in 1914 All Men Women Number, n 378 171 207 394 Male gender, % 43 100 Age, years 64.5 (5.9) 64.6 (6.0) 64.5 (5.7 68.8 (0.7) Height, cm 168.9 (9.3) 176.3 (6.8) 162.8 (6.3) 174.2 (6.5) Weight, kg 77.4 (13.4) 83.9 (12.2) 72.0 (11.8) 75.6 (10.9) BMI, kg/m2 27.1 (3.9) 27.0 (3.5) 27.1 (4.3) 24.9 (3.1) HOMA-IR, median (IQR) 2.1 (1.8) 2.2 (1.8) 1.7 (1.7) SBP, mmHg 143.7 (18.5) 145.1 (18.6) 142.1 (18.3) 153.6 (21.5) DBP, mmHg 88.2 (9.6) 90.3 (9.7) 86.5 (9.3) 92.8 (10.7) S-LDL, mmol/L 3.7 (0.8) 3.7 (0.7) 3.8 (0.9) S-HDL, mmol/L 1.5 (0.4) 1.4 (0.3 1.6 (0.4) S-cholesterol, mmol/L 6.2 (1.1) 6.0 (1.0) 6.4 (1.1) 6.1 (1.0) S-triglycerides, mmol/L 1.4 (0.8) 1.5 (0.7) 1.4 (0.8) 1.5 (0.7) Alcohol, g/daya 10.5 (12.4) 15.1 (15.2) 6.7 (7.6) 14.2 (21.8) Current smokers, % 23 26 19 46 Anti-hypertensive treatment, % 36.5 43.7 31.8 2.5 ST-segment depressions, % 27 22 31 25 High levels of supraventricular activityb, % 34 35 34 25 All values are mean (standard deviation) unless otherwise stated. a Derived from MDCS baseline questionnaire. b ‘Defined as top quartile of either SVES or SVT in the MDCS and top quartile of SVES in Men born in 1914’. Figure 2 View largeDownload slide Kaplan–Meier curves showing proportion of atrial fibrillation free survival by ST-depression status, in the MDCS. AF, atrial fibrillation; MDCS, Malmö Diet and Cancer Study; SVA, high levels of supraventricular activity; ST, ST-segment depressions. Figure 2 View largeDownload slide Kaplan–Meier curves showing proportion of atrial fibrillation free survival by ST-depression status, in the MDCS. AF, atrial fibrillation; MDCS, Malmö Diet and Cancer Study; SVA, high levels of supraventricular activity; ST, ST-segment depressions. Figure 3 View largeDownload slide Kaplan–Meier curves showing proportion of atrial fibrillation free survival by ST-depression status, in ‘Men born in 1914’. AF, atrial fibrillation; MDCS, Malmö Diet and Cancer Study; SVA, high levels of supraventricular activity; ST, ST-segment depressions. Figure 3 View largeDownload slide Kaplan–Meier curves showing proportion of atrial fibrillation free survival by ST-depression status, in ‘Men born in 1914’. AF, atrial fibrillation; MDCS, Malmö Diet and Cancer Study; SVA, high levels of supraventricular activity; ST, ST-segment depressions. ST-depressions were independently associated with incident AF in both cohorts, Table 2. There was no evidence of an interaction between gender and ST-segment depressions in the multivariable adjusted Cox regression in the MDCS cohort, (P for interaction = 0.48). After adjustment for Model 2 covariates ST-depressions were associated with significantly increased risk of AF among men in the MDCS cohort (HR 3.34, 95% CI 1.33–8.37, P = 0.01). Among women in the MDCS cohort the association was close to statistical significance (HR 2.50, 95% CI 0.93–6.69, P = 0.07). Table 2 ST depressions at 24hECG and incidence of AF, analysed using Cox regression Model 1 Model 2 HR 95% CI P HR 95% CI P MDCS cohorta 2.37 1.29–4.35 0.009 2.55 1.35–4.82 0.004 Men born in 1914b 2.31 1.09–4.89 0.03 2.28 1.05–4.95 0.038 Model 1 Model 2 HR 95% CI P HR 95% CI P MDCS cohorta 2.37 1.29–4.35 0.009 2.55 1.35–4.82 0.004 Men born in 1914b 2.31 1.09–4.89 0.03 2.28 1.05–4.95 0.038 a Model 1 adjusted for age and sex. 378 cases, 43 events. Model 2 Adjusted for Model 1 + height, weight, systolic blood pressure, smoking, LDL-cholesterol, anti-hypertensive medication and HOMA-IR. 369 subjects, 43 events. b Model 1 is adjusted for age 394 subjects and 29 events. Model 2 is adjusted for Model 1 + height, weight, systolic blood pressure, current smoking, anti-hypertensive medication and total-cholesterol. 392 subjects and 29 events. Table 3 describes the incidence of AF by 24hECG results, as well as PPV and NPV for 24hECG changes. Among the individuals with high levels of SVA the median (inter-quartile range) SVES count per hour was 9 (18), compared with 0.7 (1.1) among those without high levels of SVA, in the MDCS and 24 (75) compared with 1.2 (1.8). The negative predictive value for 24hECG changes was high in both cohorts. The highest NPV was 0.97, for a combination of high SVA (top quartile of either SVESs or SVTs) and ST-depressions in the MDCS cohort. The results were similar for men and women. Table 3 Predictive value of 24hECG changes for incident AF No STD, low SVA STD High SVAa STD and high SVA MDCS All Incidence of AFb 2.6 (1.1–6.2) 17.5 (11.0–27.8) 20.1 (13.7–29.5) 19.3 (14.0–26.5) PPV (95% CI) 0.18 (0.11–0.26) 0.20 (0.14–0.28) 0.19 (0.14–0.25) NPV (95% CI) 0.91 (0.87–0.94) 0.93 (0.89–0.96) 0.97 (0.94–0.99) Men Incidence of AFb 3.2 (1.0–10.0) 27.9 (15.0–51.9) 25.5 (15.1–43.1) 25.6 (16.5–39.6) PPV (95% CI) 0.26 (0.13–0.43) 0.24 (0.14–0.37) 0.24 (0.15–0.35) NPV (95% CI) 0.90 (0.84–0.95) 0.92 (0.85–0.96) 0.97 (0.90–0.99) Women Incidence of AFb 2.0 (0.5–7.9) 12.0 (6.0–23.9) 16.1 (9.2–28.4) 15.2 (9.6–24.1) PPV (95% CI) 0.13 (0.06–0.23) 0.17 (0.09–0.28) 0.16 (0.10–0.24) NPV (95% CI) 0.92 (0.86–0.96) 0.94 (0.29–0.97) 0.98 (0.93–1.00) Men born in 1914 Incidence of AFb 5.2 (3.1–8.9) 11.7 (6.5–21.0) 13.5 (8.0–22.9) 9.2 (5.6–15.3) PPV (95% CI) 0.11 (0.06–0.19) 0.14 (0.08–0.23) 0.09 (0.05–0.15) NPV (95% CI) 0.94 (0.91–0.97) 0.95 (0.92–0.97) 0.94 (0.90–0.97). No STD, low SVA STD High SVAa STD and high SVA MDCS All Incidence of AFb 2.6 (1.1–6.2) 17.5 (11.0–27.8) 20.1 (13.7–29.5) 19.3 (14.0–26.5) PPV (95% CI) 0.18 (0.11–0.26) 0.20 (0.14–0.28) 0.19 (0.14–0.25) NPV (95% CI) 0.91 (0.87–0.94) 0.93 (0.89–0.96) 0.97 (0.94–0.99) Men Incidence of AFb 3.2 (1.0–10.0) 27.9 (15.0–51.9) 25.5 (15.1–43.1) 25.6 (16.5–39.6) PPV (95% CI) 0.26 (0.13–0.43) 0.24 (0.14–0.37) 0.24 (0.15–0.35) NPV (95% CI) 0.90 (0.84–0.95) 0.92 (0.85–0.96) 0.97 (0.90–0.99) Women Incidence of AFb 2.0 (0.5–7.9) 12.0 (6.0–23.9) 16.1 (9.2–28.4) 15.2 (9.6–24.1) PPV (95% CI) 0.13 (0.06–0.23) 0.17 (0.09–0.28) 0.16 (0.10–0.24) NPV (95% CI) 0.92 (0.86–0.96) 0.94 (0.29–0.97) 0.98 (0.93–1.00) Men born in 1914 Incidence of AFb 5.2 (3.1–8.9) 11.7 (6.5–21.0) 13.5 (8.0–22.9) 9.2 (5.6–15.3) PPV (95% CI) 0.11 (0.06–0.19) 0.14 (0.08–0.23) 0.09 (0.05–0.15) NPV (95% CI) 0.94 (0.91–0.97) 0.95 (0.92–0.97) 0.94 (0.90–0.97). PPV, positive predictive value; NPV, negative predictive value; STD, ST-depression; SVA, supraventricular activity. a Defined as top quartile of either SVES or SVT in the MDCS and top quartile of SVES in M1914. b Per 1000 person-years. Added value of ST-depressions In the MDCS the inclusion of ST-depression to a model composed of Model 2 covariates resulted in a significantly likelier model, P = 0.005, as measured by the likelihood ratio test. The model without ST-depressions had a Harrell’s C-statistic (95% CI) of 0.7813 (0.7160–0.8466) which increased to 0.7990 (0.7337–0.8644) after inclusion of ST-depressions. In the ‘Men born in 1914’ cohort the likelihood ratio test comparing the model with Model 2 covariates with and without ST-depression yielded a P-value of 0.046. Harrell’s C-statistic (95%CI) for Model 2 covariates was 0.6170 (0.5139–0.7201) and improved to 0.6640 (0.5555–0.7725) with the addition of ST-depression. Sensitivity analyses Since presence of ST-depressions have been associated with incidence of coronary artery disease (CAD)18 we explored whether the association between ST-depressions and incident AF was independent of CAD. A sub-analysis was performed, wherein prevalent cases of CAD were excluded and individuals with CAD during the follow-up period were followed up to the date of CAD and censored after that. The association between ST-depressions and incident AF remained significant, HR 2.72 (95% CI 1.42–5.22, P = 0.003), in MDCS and 2.45 (95% CI 1.09–5.53, P = 0.03) in ‘Men born in 1914’, after adjustment for Model 2 covariates. Similarly, we assessed the effect of incident heart failure (HF) on the association between ST-depressions and AF by excluding prevalent cases of HF and censoring at the time of any incident HF event. The results were essentially unchanged (HR 2.58, (95% CI 1.36–4.91, P = 0.004), in MDCS and 2.28, (95% CI 1.05–4.95, P = 0.038) in ‘Men born in 1914’, in Model 2). We also performed a sub-analysis in the ‘Men born in 1914’ cohort with follow-up until the death of the last subject. This analysis, which thus includes more cases of AF in old ages, had a mean (SD) follow-up of 13.8 ± 7 years. There were 73 incident cases of AF (cumulative incidence 18.5%). The incidence of AF per 1000 person-years (with 95% CI) among those with a normal 24hECG was 11.1 (8.0–15.2), compared with 18.3 (15.6–32.3) among those with ST-depressions, 22.4 (15.6–32.2) for those with high levels of SVA and 17.4 (12.5–24.2) in the group with both ST-depressions and high levels of SVA. The HR (95% CI) for AF was 1.91 (1.13–3.23), P = 0.016 in men with ST-depressions, adjusted in Model 2. The NPV (95% CI) for a 24ECG without ST-depressions or high levels of SVA was 0.84 (0.79–0.88). Finally, we tested a model with reduced number of covariates (age, sex height, weight, smoking status and systolic blood pressure), with no change in results: HR (95% CI) was 2.59 (1.40–4.82) P = 0.002 in MDCS and 2.28 (1.06–4.93) P = 0.036 in ‘Men born in 1914’. Discussion Presence of episodes of ST-depression at 24hECG was significantly associated with incident AF in both cohorts. The association was not attenuated by multi-variable adjustment for age, gender, SBP, BMI, LDL-cholesterol (MDCS)/total cholesterol (‘Men born 1914’), anti-hypertensive medication, and current smoking. Furthermore, adjustment for sedentary life style and alcohol use did not alter results. When ST-depressions were combined with high levels of SVES, 24hECG alone yielded a high negative predictive value, over 10 years of follow-up. This finding may have both immediate clinical relevance, as well as usefulness in planning of future studies of AF prevention and studies of screening for AF. There was no evidence that the association between ST-depressions and incident AF was explained by intermediate CAD; in a sub-analysis in the MDCS consisting only of individuals free of symptomatic CAD the association was somewhat stronger. Symptomatic CAD is a well-known AF risk factor,6,10 and several recent studies have reported associations between AF and markers of subclinical CAD such as intima media thickness and calcifications at coronary computer tomography angiography (CCTA).11–13 The findings of the present study are thus congruent with previous research, even though there are no previous reports on ST-depressions at long-term ECG and incident AF. Possible mechanisms by which cardiac ischemia could induce AF include fibrosis and scarring of the atrial wall. Fibrosis can lead to AF not only through electrical disturbances related to the interposition of fibrous tissue between myocardial tissue, which then leads to conduction abnormalities, but also through direct effects of fibroblasts on myocardial cells.14,15 It is also possible that some of the episodes of ST-depression were due to left-ventricular hypertrophy. We found that the negative predictive value of 24hECG for incident AF was high, in the context of the long follow-up times in both cohorts. The incidence of AF over 10 years of follow-up was substantially lower in the absence of ST-depressions and high levels of SVA. This has clinical relevance, since individuals who undergo 24hECG for detection of AF currently represent a diagnostic problem, in that an individual without signs of AF on 24hECG may still have AF episodes on days that have not been recorded, and it is well known that many paroxysmal episodes of AF are asymptomatic.16 Thus, the present study suggests that 24hECG can be used not only to diagnose AF, but also to detect low risk individuals who do not need to undergo repeat examinations. Extensive monitoring with extended periods of ambulatory ECG recordings, including with implantable devices, has been discussed.17 The implication of this study is that this may not be necessary in individuals with a normal 24ECG recording, although further studies are needed to evaluate this hypothesis. Conversely the results in the present study also indicate that individuals with ST-depressions or high SVA could possibly benefit from risk factor interventions or repeated examinations. Further studies that assess the benefit of such interventions in high risk individuals are needed to confirm these observations. Screening for AF with intermittent hand-held devices is currently being tested, and has been shown to be both cost effective and to identify previously undiagnosed cases of AF among 75–76 year olds.18,19 The prevalence of AF continues to increase after this age, however,20 which raises the question of when to re-screen. Individuals whose 24hECG suggest low risk for developing AF could participate in fewer rescreening activities, thereby reducing the cost and increasing effectiveness of screening programs. The results are also relevant when considering the use of novel long-term ECG recording devices, and intermittent ECG recording devices. Any device that does not measure ST-depressions may have limitations in its ability to predict AF compared with conventional Holter monitors. There are currently no proven strategies for primary prevention of AF, and calls for research that could improve the predictive power of risk models have been made.1 The inclusion of ST-depression at 24hECG resulted in a significantly likelier model, as measured by the likelihood ratio test, in both the MDCS and the ‘Men born in 1914’ cohorts. Furthermore, there was an improvement in Harrell’s C-statistic, although not statistically significant. Perhaps 24h ECG can be used to devise better prediction models, and thereby make effective prevention programs for AF possible. Strengths and limitations A major strength of this study is that it was performed in two separate population-based cohorts, which reduces the possibility of chance findings significantly. Both cohorts have been screened at a similar age, but at different points in calendar time, which indicates that results are not likely to be influenced by any temporal trend in management of AF and AF risk factors. Furthermore, the age range in the study population corresponds to ages at which the prevalence of AF is high,20 and diagnostic work up to rule it out is highly relevant. Follow-up has been through national registries, which have been validated and found to be of good quality.6 Subclinical cases could be missed, however, unless the diagnosis has been recorded in the context of a hospital visit for another cause. Since follow-up time is long we assume that the proportion of cases that have been entirely missed is rather low, and has not influenced results to a substantial degree. Several potential confounders have been adjusted for, without any reduction in effect size. We cannot rule out residual confounding, such as by lung function or glucometabolic factors, that are known AF risk factors. However, we don’t believe the effect of any such potential residual confounding to be large. Results are consistent in several different multi-variable models, and two separate populations. The study is observational in its nature, and causality cannot be established. A causal link is plausible, however, since CAD is a known AF risk factor. Furthermore, for diagnostic purposes the results are useful regardless of the underlying causal relationships. ST-depressions at 24hECG, alone and in combination with high levels of supraventricular extrasystolic activity, is a new risk factor for AF with a high negative predictive value. The results could have important implications for screening of AF and primary prevention of AF and stroke. Conclusions ST-depression at 24hECG is independently associated with incident AF. Incidence of AF is substantially lower in individuals without ST-depression and without high levels of SVA, and 24hECG has a high negative predictive value for AF over 10 years of follow-up. Thus, 24hECG may be useful to identify individuals at high and low AF risk. Conflict of interest: SJM is the Medical director of Cardiome. MR has been consultant, given lectures, and received research grants from Bayer, Boehringer-Ingelheim, Bristol-Myers-Squibb, Medtronic, Pfizer, Sanofi, St Jude Medical, and Zenicor. Funding This work was supported by EU Interreg IV A, the Swedish Heart-Lung Foundation [grant numbers 2006-0169, 2013-0249; 2016-0315], Swedish research council [2014-2265, 521-2007-3533, 521-2010-2917], Governmental funding of clinical research within the NHS (national Health Services) and the strategic research area Epidemiology for Health (EpiHealth) at the Lund University. JSH has a personnel award from the Heart and Stroke Foundation, Ontario Provincial office (MC7450). References 1 Gorenek B, Pelliccia A, Benjamin EJ, Boriani G, Crijns HJ, Fogel RI et al. European Heart Rhythm Association (EHRA)/European Association of Cardiovascular Prevention and Rehabilitation (EACPR) position paper on how to prevent atrial fibrillation endorsed by the Heart Rhythm Society (HRS) and Asia Pacific Heart Rhythm Society (APHRS). Europace 2017; 19: 190– 225. Google Scholar PubMed 2 Johnson LS, Juhlin T, Juul-Moller S, Hedblad B, Nilsson PM, Engstrom G. A prospective study of supraventricular activity and incidence of atrial fibrillation. Heart Rhythm 2015; 12: 1898– 904. Google Scholar CrossRef Search ADS PubMed 3 Binici Z, Intzilakis T, Nielsen OW, Kober L, Sajadieh A. Excessive supraventricular ectopic activity and increased risk of atrial fibrillation and stroke. Circulation 2010; 121: 1904– 11. Google Scholar CrossRef Search ADS PubMed 4 Hedblad B, Juul-Moller S, Svensson K, Hanson BS, Isacsson SO, Janzon L et al. Increased mortality in men with ST segment depression during 24 h ambulatory long-term ECG recording. Results from prospective population study ‘Men born in 1914’, from Malmo, Sweden. Eur Heart J 1989; 10: 149– 58. Google Scholar CrossRef Search ADS PubMed 5 Heeringa J, van der Kuip DA, Hofman A, Kors JA, van Rooij FJ, Lip GY et al. Subclinical atherosclerosis and risk of atrial fibrillation: the Rotterdam study. Arch Intern Med 2007; 167: 382– 7. Google Scholar CrossRef Search ADS PubMed 6 Smith JG, Platonov PG, Hedblad B, Engstrom G, Melander O. Atrial fibrillation in the Malmo Diet and Cancer study: a study of occurrence, risk factors and diagnostic validity. Eur J Epidemiol 2010; 25: 95– 102. Google Scholar CrossRef Search ADS PubMed 7 Manjer J, Carlsson S, Elmstahl S, Gullberg B, Janzon L, Lindstrom M et al. The Malmo Diet and Cancer Study: representativity, cancer incidence and mortality in participants and non-participants. Eur J Cancer Prev 2001; 10: 489– 99. Google Scholar CrossRef Search ADS PubMed 8 Li C, Aronsson CA, Hedblad B, Gullberg B, Wirfalt E, Berglund G. Ability of physical activity measurements to assess health-related risks. Eur J Clin Nutr 2009; 63: 1448– 51. Google Scholar CrossRef Search ADS PubMed 9 Engstrom G, Hedblad B, Juul-Moller S, Tyden P, Janzon L. Cardiac arrhythmias and stroke - increased risk in men with high frequency of atrial ectopic beats. Stroke 2000; 31: 2925– 9. Google Scholar CrossRef Search ADS PubMed 10 Alonso A, Krijthe BP, Aspelund T, Stepas KA, Pencina MJ, Moser CB et al. Simple risk model predicts incidence of atrial fibrillation in a racially and geographically diverse population: the CHARGE-AF consortium. J Am Heart Assoc 2013; 2: e000102. Google Scholar CrossRef Search ADS PubMed 11 Adamsson Eryd S, Ostling G, Rosvall M, Persson M, Smith JG, Melander O et al. Carotid intima-media thickness is associated with incidence of hospitalized atrial fibrillation. Atherosclerosis 2014; 233: 673– 8. Google Scholar CrossRef Search ADS PubMed 12 Chaikriangkrai K, Valderrabano M, Bala SK, Alchalabi S, Graviss EA, Nabi F et al. . Prevalence and implications of subclinical coronary artery disease in patients with atrial fibrillation. Am J Cardiol 2015; 116: 1219– 23. Google Scholar CrossRef Search ADS PubMed 13 O'Neal WT, Efird JT, Qureshi WT, Yeboah J, Alonso A, Heckbert SR et al. Coronary artery calcium progression and atrial fibrillation: the multi-ethnic study of atherosclerosis. Circ Cardiovasc Imaging 2015; 8. 14 Anderson KR, Sutton MG, Lie JT. Histopathological types of cardiac fibrosis in myocardial disease. J Pathol 1979; 128: 79– 85. Google Scholar CrossRef Search ADS PubMed 15 Yue L, Xie J, Nattel S. Molecular determinants of cardiac fibroblast electrical function and therapeutic implications for atrial fibrillation. Cardiovasc Res 2011; 89: 744– 53. Google Scholar CrossRef Search ADS PubMed 16 Camm AJ, Corbucci G, Padeletti L. Usefulness of continuous electrocardiographic monitoring for atrial fibrillation. Am J Cardiol 2012; 110: 270– 6. Google Scholar CrossRef Search ADS PubMed 17 Camm AJ. The role of continuous monitoring in atrial fibrillation management. Arrhythm Electrophysiol Rev 2014; 3: 48– 50. Google Scholar CrossRef Search ADS PubMed 18 Svennberg E, Engdahl J, Al-Khalili F, Friberg L, Frykman V, Rosenqvist M. Mass screening for untreated atrial fibrillation: the STROKESTOP study. Circulation 2015; 131: 2176– 84. Google Scholar CrossRef Search ADS PubMed 19 Aronsson M, Svennberg E, Rosenqvist M, Engdahl J, Al-Khalili F, Friberg L et al. Cost-effectiveness of mass screening for untreated atrial fibrillation using intermittent ECG recording. Europace 2015; 17: 1023– 9. Google Scholar CrossRef Search ADS PubMed 20 Friberg L, Bergfeldt L. Atrial fibrillation prevalence revisited. J Intern Med 2013; 274: 461– 8. Google Scholar CrossRef Search ADS PubMed Published on behalf of the European Society of Cardiology. All rights reserved. © The Author 2017. For permissions, please email: firstname.lastname@example.org.
Europace – Oxford University Press
Published: Mar 1, 2018
It’s your single place to instantly
discover and read the research
that matters to you.
Enjoy affordable access to
over 12 million articles from more than
10,000 peer-reviewed journals.
All for just $49/month
Read as many articles as you need. Full articles with original layout, charts and figures. Read online, from anywhere.
Keep up with your field with Personalized Recommendations and Follow Journals to get automatic updates.
It’s easy to organize your research with our built-in tools.
Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.
All the latest content is available, no embargo periods.
“Hi guys, I cannot tell you how much I love this resource. Incredible. I really believe you've hit the nail on the head with this site in regards to solving the research-purchase issue.”Daniel C.
“Whoa! It’s like Spotify but for academic articles.”@Phil_Robichaud
“I must say, @deepdyve is a fabulous solution to the independent researcher's problem of #access to #information.”@deepthiw
“My last article couldn't be possible without the platform @deepdyve that makes journal papers cheaper.”@JoseServera