The long and short of the PR-interval: relation to cardiovascular outcome in patients with coronary heart disease

This editorial refers to ‘Clinical outcome as a function of the PR-interval—there is virtue inmoderation: data from the Duke Databank for cardiovascular disease’ by F. Holmqvist et al., on page 978–985.

In this issue of the EP-Europace, Holmqvist et al. analysed the relationship between the PR interval and the outcome in 9637 patients in sinus rhythm, undergoing coronary angiography at Duke University Medical Center between 1989–2010 who had significant stenosis in at least one native coronary artery.¹ Classifying PR interval as short (PR 90–120 ms), normal (PR 120–200 ms), or long (PR 200–500 ms), they demonstrated that both short and long PR intervals were associated with all-cause mortality, the composite of death or stroke, and the composite of cardiovascular death or cardiovascular re-hospitalization compared with the normal PR interval. No difference was detected in sudden cardiac death across the three groups.

Long PR interval, also known as first-degree atrioventricular block (AVB), is a result of conduction delay occurring in the right atrium, AV node, His bundle, or Purkinje network. Potential underlying mechanisms include atherosclerosis of the coronary arteries, atrial fibrosis, dilated cardiomyopathy, infiltrative cardiomyopathies, left ventricular dysfunction, neuromuscular disease and insufficient blood pressure control causing left ventricular hypertrophy.² Therefore, long PR interval may be a marker of underlying heart disease, rather than being an independent predictor of left ventricular dysfunction or hypertrophy. Less is known about the underlying mechanism of an isolated short PR interval (i.e. no delta wave and a normal QRS complex), although this can reflect less common bypass tracts that insert directly into the His-Purkinje system. In the present study only 0.52% had short PR interval. Because heart rate and PR interval have an inverse relationship during physiological increase of the heart rate, Holmqvist et al. suggest that the poor outcome associated with short PR interval could be due to the higher heart rate in patients with short PR interval in their study. Furthermore, compared with patients with a normal PR interval, patients with a short PR interval had more heart failure, longer QRS, and QTc duration, all known to be associated with worse prognosis.

Because the relation between the PR interval and the risk of all four studied endpoints was found to be non-linear, Holmqvist et al. used cubic polynomial spline plots of the unadjusted risks across the range of values of the PR interval for all endpoints, demonstrating a U-shaped relation between the risk and the PR interval. This analysis further showed that a PR interval value of 162 ms split the relationship between the PR interval and the risk of each outcome into two linear portions: analysing a decreasing PR interval in patients with a baseline PR-interval <162 ms and an increasing PR interval for PR-interval values >162 ms, values below but not >162 ms were significantly associated with higher risk of the composite of death or stroke, and the composite of cardiovascular death or cardiovascular re-hospitalization. However, in a sub-analysis of patients not on an antiarrhythmic drug, increasing PR interval above 162 ms was associated with all-cause mortality and the composite of death or stroke.

The current study by Holmqvist et al. was a retrospective analysis of middle-aged participants with significant stenosis in at least one native coronary artery. Compared with patients with a normal PR interval, the 18% of patients with long PR interval (>200 ms) were older, and were more likely to be males, have hypertension, cerebrovascular disease, prior coronary artery bypass grafting, and valve disease. Several previous studies support the adverse prognostic value associated with a long PR interval in the present study. The Heart and Soul Study of 938 patients with stable coronary artery disease showed that a prolonged PR interval, defined as >220 ms, was present in 9.3% and associated with increased risk of heart failure hospitalization, cardiovascular and all-cause mortality, which remained significant after adjusting for several co-variables.³ However, this association seemed to be partly explained by left ventricular systolic and diastolic dysfunction and history of heart failure. In the Framingham Heart Study, the baseline first-degree AVB, defined as PR interval >200 ms, was associated with increased risks of atrial fibrillation, pacemaker implantation, and all-cause mortality in 7575 participants without use of antiarrhythmic agents, cardiac glycosides, history of pacemaker implantation, or history of atrial fibrillation.⁴ In contrast to the Heart and Soul Study, only 124 (1.6%) had baseline first-degree AVB despite the lower PR interval cut-off, suggesting inclusion of a healthier patient sample in the Framingham Heart Study with lower prevalence of cardiovascular morbidity and cardiac structural abnormalities compared with patients in the Heart and Soul Study. The Health ABC study examined a cohort of healthy, well-functioning older adults and showed that the 12% of participants with PR interval >200 ms had a 46% higher risk of incident heart failure after 10 years of follow-up, and when the PR interval was analysed as a continuous variable, every 1-SD of the mean (29 ms) increase in PR interval was associated with a 13% greater 10-year risk of heart failure and a 13% increased risk of incident atrial fibrillation after multivariable adjustment; however, no association was found with all-cause mortality.⁵ Since ageing is associated with fibrosis of the myocardium, the older age of the Health ABC Study participants compared with earlier studies could in part explain the higher prevalence of prolonged PR interval than found in the Framingham Heart Study⁴ and the divergent finding between the previous perception of first-degree AVB as a benign finding and recently published studies suggesting higher rates of morbidity, as previous studies of first-degree AVB have mainly been conducted in healthy, younger participants.

Very few studies have been examining isolated short PR intervals. Possible mechanisms of the short PR interval include accelerated conduction in the AV node, accessory pathways that bypass the AV node and insert directly into the His-Purkinje system, or a prolonged P-wave possibly reflecting a dilated left atria which is known to be associated with cardiovascular morbidity.⁶ It is possible that short PR interval merely represent one end of a continuous spectrum of normal AV nodal physiology. However, given the larger proportion of heart failure among patients with a short PR interval in the present study, it appears more likely that short PR interval reflects underlying structural disease which could explain the impaired left ventricular function of patients in the present study. The associated left ventricular hypertrophy in patients with heart failure could in part explain the worse prognosis in patients with short PR interval. In a database comprising 288 181 electrocardiograms from general practitioners, Nielsen et al. showed that short PR interval was associated with atrial fibrillation,⁷ which has shown to be associated with fatal and non-fatal cardiovascular outcomes in patients with myocardial infarction⁸ and with incident heart failure in patients with hypertension.⁹ Therefore, development of new atrial fibrillation could facilitate the adverse outcomes in patients with short PR interval. Since Cornell duration product has shown to be closely associated with echocardiographic hypertrophy and poor cardiovascular outcome,¹⁰ including Cornell duration product in the analyses and baseline tables would significantly have improved the present paper and made it possible to further discuss the contribution of left ventricular hypertrophy. The absence of an association between PR interval and sudden death in the current study further supports underlying structural heart disease rather than a primarily arrhythmic substrate in the setting of an abnormal PR interval.

The electrocardiogram is an inexpensive and widely available procedure that is already obtained in most patients with signs of cardiovascular disease and provides convincing predictive value in patients with cardiovascular disease beyond modalities such as echocardiography.¹⁰ Holmqvist et al. adjusted their analyses for factors previously known to be associated with cardiovascular morbidity; and thereby support other recently published papers. Their findings suggesting that both the long and short ends of the PR interval spectrum may be valuable predictors of poor outcome should remind the physician to evaluate and monitor patients for development of progressive or new underlying cardiovascular disease after short or long PR interval is detected. Whether the PR interval will become a widely utilized marker of cardiovascular risk in the contemporary evaluation of patients will require further assessment in additional populations and under differing treatment conditions. Importantly, it will be critical to determine whether short and long PR intervals improve the prognostic value of net reclassification index and c-statistics beyond known risk factors for cardiovascular outcomes.

Conflict of interest: none declared.

References

Author notes