Notched P-Wave on Digital Electrocardiogram Predicts Cardiovascular Events in Patients with Cardiovascular Risks: The Japan Morning Surge Home Blood Pressure Study

The P-wave on 12-lead electrocardiography represents atrial activation. Atrial remodeling causes intra/interatrial conduction delay, and results in a change in P-wave morphology [1]. The notched P-wave is one of the changes in P-wave morphology, and reflects the atrial conduction delay [2]. An abnormality of the P-wave is a predictor of the onset of atrial fibrillation [2‒5].

Left ventricular hypertrophy (LVH) is a feature of myocardial remodeling and is an important pathological change in hypertensive patients and/or patients with atherosclerosis. Atherosclerosis causes an increase in afterload and left atrial load. Thus, left atrial load is not only a reflection of an increase in afterload, but also a preliminary stage of LVH [5]. The change in the P-wave is also a reflection of left ventricular remodeling and is associated with LVH [6].

We previously reported that the notched P-wave was associated with cardiovascular events in a community-dwelling population [7]. However, the association between the notched P-wave and prognosis remains unclear among patients with cardiovascular risk factors, and the data regarding the association between the notched P-wave and morphology of the left atrium and left ventricle are also insufficient. The aim of this study was to investigate the association among notched P-waves on digital electrocardiogram (ECG), left atrial/ventricular morphology, and cardiovascular events.

We enrolled subjects from the Japan Morning Surge Home Blood Pressure (J-HOP) study who had one or more of the four cardiovascular risk factors: hypertension, dyslipidemia, diabetes, and smoking [8]. The recruitment of patients for the J-HOP study was consecutively conducted from January 2005 to May 2012, by 75 doctors at 71 institutions throughout Japan (45 primary practices, 22 hospital-based outpatient clinics, and 4 specialized university hospitals). The Ethics Committee of the Internal Review Board of the Jichi Medical University School of Medicine approved the protocol. The study protocol was registered on the University Hospital Medical Information Network Clinical Trials Registry website under registration no. UMIN000000894. Written informed consent was obtained from all patients who were enrolled in the study [8]. The J-HOP study is a prospective observational study conducted to evaluate the predictive values of home measurements of blood pressure for cardiovascular events in Japanese patients with any of the following cardiovascular risk factors: hypertension, diabetes mellitus, hyperlipidemia, smoking (including patients with chronic obstructive pulmonary disease), chronic renal disease, atrial fibrillation, metabolic syndrome, and sleep apnea syndrome.

We recorded digital 12-lead electrocardiography at a sampling rate of 500 Hz and 4.88-μV resolution over 10 s (equipment supplied by Fukuda Denshi, Tokyo). The peak-to-peak distance in the M shape in lead II was calculated and stored automatically using a 12-lead ECG analysis system (Fukuda Denshi, Tokyo) (Fig. 1). We obtained ECG data for 834 subjects. We excluded 24 patients with atrial fibrillation because we could not analyze the notch of the P-wave, leaving 810 patients who were enrolled in the present analysis. (1) We found a valley in the section from the start point of the P-wave to the endpoint of the P-wave with respect to the averaging waveform of lead II, (2) we searched for the maximum value A between the start point of the P-wave and the valley, (3) we searched for the maximum value B between the valley and the end of the P-wave, and (4) the notch interval was defined by the distance between A and B. We compared two definitions: one in which P-waves were defined as “notched” if the peak-to-peak distance in the M shape was ≥40 ms (which is the smallest unit on paper ECG recordings) in lead II [1, 7], and the another in which “notched” P-waves were those with a peak-to-peak distance in the M shape of ≥20 ms (i.e., half of the smallest unit on paper ECG recordings). We also manually checked for notched P-waves in lead II by printed ECG (at a paper speed of 25 mm/s). The agreement between the notched P-waves by the 40 ms criterion identified by manual checking and those identified by the ECG analysis system, was found to be relatively weak (κ statistic = 0.40). The 40-ms criterion identified by the ECG analysis system had a sensitivity of 0.30 and a specificity of 0.99 for the notched P-waves identified by the manual checking. We calculated the Cornell product as the product of Cornell voltage and QRS duration. The LVH diagnosed by ECG (ECG-LVH) was defined as ≥244 mV × ms according to a previous report of the Losartan Intervention for Endpoint Reduction in Hypertension study [9].

Echocardiography was performed at Jichi Medical University. The two-dimensional M-mode or B-mode image was recorded using an ultrasound machine according to the guidelines of the American Society of Echocardiology and the European Association of Echocardiography [10]. The left ventricular (LV) mass was obtained using the formula validated by the American Society of Echocardiology: LVM = 0.8 × {1.04 [(LVIDd + PWTd + SWTd) 3 − (LVIDd) 3]}+0.6 g, where LVIDd is LV internal diameter in diastole, PWTd is posterior wall thickness in diastole, and SWTd is septal wall thickness in diastole. The left ventricular mass index (LVMI) was calculated as LV mass/body surface area. LVH diagnosed by echocardiography (echo-LVH) was defined as an LVMI >115 g/m² and >95 g/m² in both men and women, respectively [10].

Blood samples were collected at each patient’s hospital arrival in the morning in a fasting state. The blood samples were centrifuged at 3,000 g for 15 min at room temperature. Plasma/serum samples after separation were stored at 4°C in refrigerated containers and sent to a commercial laboratory (SRL, Tokyo) within 24 h. All assays were performed within the next 24 h at this single laboratory center. The plasma level of brain natriuretic peptide (BNP) was measured using a chemiluminescent enzyme (MI02 Shionogi BNP; Shionogi, Osaka, Japan).

The primary endpoint was defined as a composite endpoint that combines fatal events (stroke, coronary artery disease, heart failure, and sudden death) and nonfatal events (acute myocardial infarction, angina, congestive heart failure, stroke, and aortic dissection). The outcomes were categorized as follows. (1) Fatal and nonfatal stroke, defined as the sudden onset of a neurologic deficit persisting for ≥24 h in the absence of another disease that could account for the symptoms, based on the findings of brain computed tomography or magnetic resonance imaging; transient ischemic attack was not included. (2) Fatal and nonfatal coronary artery disease, defined as acute myocardial infarction, angina pectoris requiring percutaneous coronary intervention, or sudden death within 24 h of the abrupt onset of symptoms. (3) Fatal and nonfatal heart failure and artery disease that required admission. (4) Other sudden death, and (5) aortic dissection. The endpoint committee adjudicated all events by reviewing the patient files and source documents or requesting more detailed written information from investigators. The committee was blinded to the individual clinical characteristics, including home blood pressure data. A final follow-up survey to reconfirm the clinical outcomes was performed from September 2014 to March 2015.

Data are presented as the mean (±standard deviation), the median, with the interquartile range as 25th and 75th percentiles, or a percentage. The χ² test was used for the categorical data. Unpaired t-tests were used for the normally distributed data and comparisons between two groups. The Kaplan-Meier curves of the cumulative incidence of cardiovascular events in each group were calculated, and the differences in the rate of events between groups were assessed by the log-rank test. We used Cox proportional hazard models to examine the associations between the presence or absence of a notched P-wave and the primary endpoints, including the CP-LVH. The covariates in model 1 included traditional risk factors such as age, sex, current smoking, regular drinking, and history of hypertension, dyslipidemia, and diabetes; the factors in model 2 were those in model 1 plus a history of atrial fibrillation; the factors in model 3 were those in model 2 plus ECG-LVH and echo-LVH; and the factors in model 4 were those in model 3 plus left atrial diameter. To find the best cut-off value for predicting cardiovascular events, receiver operating characteristic analysis was performed. All statistical analyses were performed using the computer software package SPSS (ver. 25; IBM, Chicago, IL, USA). p values <0.05 were considered significant.

The average age was 62.6 ± 10.9 years, and the percentage of male subjects was 51.5% in this study. The number of patients categorized into the notched P-wave group by the ≥40 ms definition was 23 (2.8%) and that by the ≥20 ms definition was 92 (11.4%).

The baseline characteristics by the two definitions of a notched P-wave are shown in Table 1. The average ages in the notched P-wave group by the ≥20 ms definition and the notched P-wave group by the ≥40 ms definition were higher than those in the other groups (notched P-wave (≥20 ms in M shape): 65.7 ± 10.6 versus 62.4 ± 10.9 years, p = 0.006; notched P-wave (≥40 ms in M shape): 68.4 ± 11.2 versus 62.6 ± 10.8 years, p = 0.008 (Table 1). The percentage of patients with a history of AF in the notched P-wave group by the ≥20 ms definition and the notched P-wave group by the ≥40 ms definition was higher than those in the other groups (notched P-wave [≥20 ms in M shape]: 10.9% vs. 3.5%, p = 0.001; notched P-wave [≥40 ms in M shape]: 17.4% vs. 3.9%, p = 0.002) (Table 1). The percentage of male patients was similar between the presence and absence of a notched P-wave using the ≥20 ms definition and between the presence and absence of a notched P-wave using the ≥40 ms definition.

The associations between cardiac damage and notched P-wave are shown in Table 2. The left atrial diameter and LVMI in the patients with the notched P-wave by the ≥20 ms definition were significantly higher than those in patients without the notched P-wave by this definition (left atrial diameter 38.8 ± 5.9 vs. 36.8 ± 5.0 mm, p = 0.001; LVMI 103.9 ± 27.7 vs. 96.3 ± 25.7 g/m², p = 0.010). The left atrial diameter in the patients with the notched P-wave by the ≥40 ms definition was significantly higher than that in the patients without the notched P-wave by this definition (left atrial diameter 41.8 ± 7.1 vs. 36.9 ± 5.0 mm, p < 0.001). BNP values in patients with the notched P-wave by the ≥20 ms definition and in patients with the notched P-wave by the ≥40 ms definition were higher than those in the other groups (notched P-wave by the ≥20 ms definition: 33.8 vs. 16.9 pg/dL, p = 0.003; notched P-wave by the ≥40 ms definition: 45.7 vs. 17.9 pg/dL, p < 0.001, Table 2).

The mean follow-up period was 101 ± 34 months, and 85 cardiovascular events occurred (1 case of fatal myocardial infarction, one of fatal stroke, 4 of fatal heart failure, 3 of sudden death, 25 of angina pectoris, 6 of nonfatal myocardial infarction, 28 of nonfatal stroke, 12 of nonfatal heart failure, and 5 of aortic dissection). The Kaplan-Meier curves of the incidence of cardiovascular events are given in Figure 2. The patients with a notched P-wave by the ≥20 ms definition had significantly poorer prognoses compared to those without a notched P-wave by the ≥20 ms definition (log rank 8.89, p = 0.003). The patients with a notched P-wave by the ≥40 ms definition also had significantly poorer prognoses compared to those without a notched P-wave by the ≥40 ms definition (log rank 6.81, p = 0.009).

The results of Cox proportional hazard models to examine the associations between the presence or absence of a notched P-wave and the primary endpoints are shown in Tables 3and4. When we defined a notched P-wave as a peak-to-peak distance of ≥20 ms in the M shape (n = 92), a notched P-wave was a significant predictor of cardiovascular events after adjustment for age, gender, comorbidity, ECG-LVH, echo-LVH, and left atrial diameter (hazard ratio: 1.83; 95% confidence interval: 1.01–3.31; p = 0.045; Table 3). When we defined a notched P-wave as a peak-to-peak distance of ≥40 ms in the M shape (n = 25), the hazard ratio of cardiovascular events in the notched P-wave group was not significant after adjustment for age, gender, comorbidity, ECG-LVH, echo-LVH, and left atrial diameter (hazard ratio: 1.52; 95% confidence interval: 0.51–4.53; p = 0.455; Table 4).

In the results of the forward conditional analysis, a notched P-wave defined as a peak-to-peak distance of ≥20 ms in the M shape was an explored predictor of cardiovascular events (p = 0.026), but a notched P-wave defined as a peak-to-peak distance of ≥40 ms in the M shape was not. The results of the receiver operating characteristic analysis showed that a peak-to-peak distance of 18 ms in the M shape was the best cut-off value for predicting the incidence of cardiovascular events.

Finally, we examined the relation between notched P-wave (40 ms by manual measurements) and cardiovascular events. However, the association between notched P-wave defined as ≥40 ms by manual measurements and cardiovascular events was weak (log rank: 3.48; p = 0.062).

The main findings of this study were as follows: (1) the notched P-wave on digital ECG was associated with cardiovascular events, and (2) the notched P-wave was associated with left atrial enlargement. We previously reported that the risk of cardiovascular events among subjects with a notched P-wave was 1.59-fold greater than that in subjects without a notched P-wave in a community-dwelling population [7]. In the present study, the risk of cardiovascular events in subjects with a notched P-wave was more than twice that of subjects without a notched P-wave. This difference in the risk of cardiovascular events might be due to the difference in the prevalence of hypertension and diabetes. Holmqvist et al. [11] found that abnormal P-wave on signal-averaged ECG was associated with poor cardiac outcome and atrial fibrillation. Platonov et al. [12] showed that a notched P-wave on signal-averaged ECG was associated with interatrial conduction in paroxysmal atrial fibrillation, which might contribute to cardiovascular events. De Bacquer et al. [1] showed that a deflected P-wave was associated with a 6.89-fold increase in the onset of atrial fibrillation. Deflected P-waves were classified by a clear notch-free pattern of 40 ms or longer, and notched P-waves were defined as those with a clear notch pattern of 40 ms or longer. In our present analysis, we employed two definitions of “notched”, i.e., a peak-to-peak distance in the M shape of ≥20 ms or ≥40 ms. Digital analysis might be useful in discriminating such small notches.

In this study, the notched P-wave defined at 40 ms was not a significant predictor of cardiovascular events after adjusting for LVH, but the notched P-wave defined at 20 ms was a significant predictor. Afterload due to atherosclerosis causes left atrial load, which will eventually develop into LVH [6]. In the present study, patients in the notched P-wave group defined at 40 ms had a higher BNP, a larger left atrial diameter, and a heavier LVMI than those in the notched P-wave group defined at 20 ms. The increase in notch width might reflect remodeling of the left atrium and left ventricle. Since LVH progresses when the notch width is extended to 40 ms, the notched P-wave defined at 40 ms was not associated with cardiovascular events after adjusting for LVH.

Our results showed that a notched P-wave was associated with a left atrial load and an increase in BNP. The morphology of the P-wave is determined by the propagation of excitation in the right atrium and left atrium [13]. The latter half of the P-wave is a reflection of the excitation of the left atrium, and the notched P-wave is a reflection of the delayed excitation of the left atrium [1]. Bachmann bundle conduction is important for right atrium and left atrium conduction [14], and its delay has been shown to cause intra-atrial block and to be associated with notched P [15]. Intra-atrial block has been shown to be associated with LV diastolic dysfunction and an increase in atrial natriuretic peptide, as well as with atrial fibrillation [16‒18]. Left atrial load is a reflection of myocardial remodeling, and the notched P-wave reflects left atrial excitation delay and might be associated with the risk of developing heart failure.

Hypertension is a common and important cardiovascular risk factor. Approximately 90% of patients in the present study had hypertension. ECG is an important modality for risk stratification in hypertension [19, 20]; as a noninvasive, low-cost test, ECG merits further research. The advantages of interpretation of digitally obtained ECG are as follows: (1) noncardiologists can assess the ECG findings, and (2) the interobserver differences might be reduced. In addition, a small disturbance of the P-wave, such as a notched P-wave with a peak-to-peak distance of ≥20 ms in the M shape, can hardly be measured manually on a paper ECG, but when detected by automatic ECG analysis such a disturbance is associated with worse patient outcomes in the long term.

The main limitation of this study was that we enrolled only high-risk Japanese patients. Because the risk of developing atrial fibrillation and cardiovascular events varies by race, large-scale international research is needed in future. The small number of subjects was also a limitation of this study. More significant P-wave changes might be associated with worse outcomes; however, the small number of patients with a notched P-wave defined as a peak-to-peak distance in the M shape of ≥40 ms would have resulted in the larger range of the 95% confidence interval. As a result, cardiovascular events were significantly associated with a notched P-wave ≥20 ms, but not with more significant P-wave abnormality (≥40 ms). We could not check the agreement between the notched P-waves by the 20 ms criterion identified by manual checking, because 40 ms is the smallest unit in paper ECG recordings. The agreement between the notched P-waves by manual checking and those identified by the ECG analysis system was relatively weak. It may be necessary to improve the algorithm used for the manual identification of notched P-waves.

Notched P-wave on digital ECG was associated with cardiovascular events and left atrial enlargement.

All enrolled patients provided written informed consent. The Ethics Committee of the Internal Review Board of the Jichi Medical University School of Medicine approved the protocol (No. 04-17). The study protocol was registered on the University Hospital Medical Information Network Clinical Trials Registry website under registration no. UMIN000000894.

K. Kario received research grants from Omron Healthcare and A&D Co. The other authors have no conflict of interest to declare.

This study was financially supported in part by a Grant from the 21st Century Center of Excellence Project run by Japan’s Ministry of Education, Culture, Sports, Science, and Technology (to K. Kario); a grant from the Foundation for Development of the Community (Tochigi, Japan); a Grant from Omron Healthcare, Co., Ltd.; a Grant-in-Aid for Scientific Research (B) (21390247) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT) of Japan, 2009–2013; and funds from the MEXT-Supported Program for the Strategic Research Foundation at Private Universities, 2011–2015 Cooperative Basic and Clinical Research on Circadian Medicine (S1101022).

Kabutoya T. wrote the manuscript, performed the statistical analysis, and takes primary responsibility for this paper. Kabutoya T., Hoshide S., and Kario K. collected the patients’ data and reviewed/edited the manuscript. Kario K. acquired research grants for the J-HOP study.

All supporting data within the article are available upon reasonable request from any qualified investigator.