P-wave axis as a predictor of mortality

The first human electrocardiogram (ECG) recording was performed by the British physiologist Augustus D Waller of St Mary’s Medical School London in 1887. It was recorded with a capillary electrometer.1 The implications for modern clinical medicine have been immense.2 New insights into the clinical usefulness of the ECG are still being revealed, 130 years after its discovery. The paper by Lazzeroni et al. is a potential contribution to our understanding of the clinical significance of the P-wave axis.3 The P-wave provides electrocardiographic information of the atria; the information is embedded within the duration, dispersion, amplitude and axis of the P-wave. A normal P-wave is upright in leads I and II (axis between 0 and 75 degrees). The P-wave axis is influenced by multiple factors. First, the P-wave axis is dependent upon the origin of the right atrial depolarisation vector. The origin of the normal P-wave is the sinus node. The sinus node is normally (but not always) located at the superior caval vein opening in the right atrium.4 The propagation of the electric wave fronts follows the complex structure of the atrial myocardial fibres from the right to the left atrium. The P-wave is a summation of the depolarisation of the right (first third of the P-wave) and the left atria (last third of the P-wave). Enlargement of either chamber affects the P-wave: a right atrial enlargement is recognised by increased amplitude, while a left atrial enlargement is characterised by a longer duration, sometimes with a central notch (P-mitrale).4 Several pathological conditions can influence the P-wave axis; for example, an interatrial conduction block due to fibrosis in the area of Bachmann’s bundle leads to a left axis deviation.4 In the general population, P-wave axis abnormalities have been found to predict mortality,5 risk of atrial fibrillation6,7 and stroke.8 The study by Lazzeroni et al. investigates the prognostic impact of an abnormal P-wave axis in patients following coronary revascularisation and/or valve surgery.3 ECGs were acquired in 810 patients at the time of discharge from a 2-week cardiac rehabilitation programme following surgery/intervention. An abnormal P-wave axis was found in 94 patients (11.6%); 53 patients (6.5%) had an abnormal right axis and 39 (4.8%) subjects had an abnormal left axis. Individuals with an abnormal P-wave axis were older and had slightly longer QRS duration and QTc intervals compared with those with a normal axis. They had increased echocardiographic markers of left ventricular and atrial remodeling (left ventricular mass index, left ventricular end-diastolic volume and left atrium volume index (LAVi)), but no difference in left ventricular ejection fraction (LVEF). During a follow-up of 47 ± 27 months, a total of 103 (12.7%) patients died, with 56 (5.7%) deaths being attributed to cardiovascular causes. In patients with an abnormal P-wave axis there was an increased risk of death: a total of 23 patients (24.4%) died with 11 (11.7%) deaths being due to cardiovascular causes. By the use of Cox proportional hazard models, the authors found a 2.5-fold higher overall risk of mortality in patients with abnormal P-wave axis (hazard ratio (HR) 2.5, 95% confidence interval 1.6–4.0, P < 0.001). Similarly, highly significant HRs were reported for cardiovascular mortality, even after adjusting for established risk factors. When patients with abnormal P-wave axis were split into right or left axis deviations, findings suggested worse outcomes for the patients with left compared with right axis. The strength of the present study is the use of hard endpoints and the postoperative study population; the weakness is the use of an ECG recording from a single timepoint and the choice of variables included in the multivariable models. As acknowledged by the authors, the use of only a single 12-lead recording is an important limitation to this study. It is therefore not possible to assess the impact of changes in the P-wave axis over time. The reproducibility of the P-wave axis assessment has been investigated in several populations of clinically stable subjects. In a 2-week reproducibility study of P-wave indices, a variation coefficient of 5.83% was found for the P-wave axis measurement.9 A long-term study demonstrated that fewer than 40% of subjects with abnormal P-wave axis at baseline had a persistent P-wave axis abnormality at follow-up 11 years later, suggesting considerable lability in the measurement of P-wave axis.10 Therefore, a considerable variation in the P-wave axis must be anticipated. This may be of particular importance in the first weeks following transthoracic surgery due to common postoperative complications such as pulmonic complications and pleural and pericardial effusions. Unfortunately, the precise duration between surgery and ECG acquisition is not available in the current publication. It is therefore not possible to use the study to define the optimal timing of P-wave axis assessment following thoracic surgery. Based on the multivariable Cox regression models the authors concluded that abnormal P-wave axis represents an independent predictor of both overall and cardiovascular mortality in patients undergoing myocardial revascularisation or cardiac valve surgery. However, it is necessary to be careful in the interpretation of the models. The multivariate models included age, gender, PR duration, QRS duration, QTc duration, LVEF and LAVi. The authors did not include important prognostic variables such as renal function (estimated glomerular filtration rate), type of operation, prior surgery, functional class and important comorbidities such as diabetes mellitus and lung disease. Right P-wave axis deviation has been strongly linked to emphysema, a disease that leads to abnormal lung hyperinflation, frequently causing a vertical frontal P-wave axis exceeding 60 degrees.11–13 This association might be due to the right atrium being firmly attached to the diaphragm by way of a pericardial ligament around the inferior vena cava, and with progressive flattening of the diaphragm, the right atrium will be displaced.14 The authors state that 10% of those with normal P-wave axis versus 11% of those with abnormal P-wave axis had a diagnosis of chronic obstructive pulmonary disease (COPD) at baseline (P = 0.076). No test to screen for COPD was undertaken and the severity of the disease is not reported. We are therefore unable to assess the potential role of COPD in the increased risk of death found in subjects with right P-wave axis abnormalities. Fibrosis of Bachmann’s bundle can cause interatrial conduction blocks that result in a left P-wave axis. This atrial fibrosis may be secondary to atrial inflammation and strain, particularly in patients with significant valvular disease. Therefore, when considering the additional prognostic significance of the P-wave axis, it is important to bear in mind the underlying pathology. There is increasing interest in the predictive value of P-wave indices. The findings by Lazzeroni et al. are in line with previous reports, but their study adds interesting data by exploring the potential prognostic role of the P-wave axis using hard endpoints in a new patient population. The present study is a single centre study and the analysis is limited to a 12-lead ECG recording at a single timepoint. The findings must therefore be interpreted with caution. However, the work by Lazzeroni et al. provides encouragement to our interest in exploring this field further, using repeated ECG recordings to address the prognostic impact and reproducibility of P-wave axis in a postoperative patient population. The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article. The author(s) received no financial support for the research, authorship, and/or publication of this article.