Right Ventricular Dysfunction: An Ominous Sign

See also the article by Jimenez-Juan et al in this issue.

Dr Nazarian earned a doctorate in medicine from Stanford University and PhD in Clinical Epidemiology from the Johns Hopkins Bloomberg School of Public Health. He completed residency at the Brigham and Women’s Hospital, followed by fellowship training in Cardiovascular Disease and Clinical Cardiac Electrophysiology at the Johns Hopkins Hospital. Dr Nazarian is on the faculty of the Perelman School of Medicine at the University of Pennsylvania, where he focuses on translating his experience in electrophysiology, imaging techniques, and ablation into clinically relevant strategies for the care of his patients with cardiac arrhythmias. Dr Nazarian runs an active clinical research program, funded by the U.S. National Institutes of Health.

Dr Zghaib is currently a fellow in Cardiovascular Medicine at the Hospital of the University of Pennsylvania where he will be pursuing further training in clinical cardiac electrophysiology. He did a postdoctoral research fellowship at the Cardiac Electrophysiology Imaging laboratory at the Johns Hopkins Hospital where he also trained in internal medicine. His research interests include the use of advanced cardiac imaging techniques to study the mechanism of atrial and ventricular arrhythmias as well as deriving imaging parameters for prognostication in nonischemic cardiomyopathies. He is interested in using cardiac MRI to guide catheter-based ablation for atrial and ventricular arrhythmias.

Implantable cardioverter-defibrillators (ICDs) are an established therapy for the prevention of arrhythmic sudden cardiac death in patients with reduced left ventricular (LV) function, especially after myocardial infarction (1). However, the majority of sudden cardiac death events occur among individuals that do not meet current patient selection criteria for primary prevention ICD implantation. At the same time, the rate of appropriate shock therapy after ICD implantation using current patient selection criteria is low and complications of device implantation and inappropriate shock therapies are not negligible. As such, there is clear need for further risk stratification tools beyond the currently adopted LV functional criteria to identify patients who would benefit from ICD.

In this issue of Radiology, Jimenez-Juan et al (2) demonstrate an association between reduced right ventricular (RV) ejection fraction (RVEF) with appropriate ICD therapies and all-cause mortality (hazard ratio, 1.21 per 10% drop in RVEF; P = .01) in a retrospective cohort of 411 patients with reduced LV function who underwent cardiac MRI prior to ICD implantation. This association was independent of LV function, LV scar burden on late gadolinium enhancement images, indication for ICD implantation, and cause of cardiomyopathy.

Jimenez-Juan and colleagues (2) should be congratulated for this well-presented study and rigorous statistical analysis. Their findings are biologically plausible, consistent with those of prior reports, and potentially clinically useful. The prognostic impact of poor RV function on mortality in patients with ischemic and nonischemic cardiomyopathies has been demonstrated in multiple studies (3,4). In 314 patients with heart failure, Mikami et al (4) showed that RV dysfunction has incremental prognostic impact on mortality and appropriate ICD therapy beyond that of LV function and percent scar burden. They demonstrated a graded increase in risk of primary outcome with LV dysfunction alone (lowest risk), RV dysfunction alone (intermediate risk), or both (highest risk). In a posthoc analysis of the DANISH trial (Danish Study to Assess the Efficacy of ICDs in Patients with Nonischemic Systolic Heart Failure on Mortality), Elming et al (3) found a significant interaction between RVEF and ICD implantation on prevention of sudden cardiac death. They found that ICD implantation significantly reduced all-cause mortality in patients with RV dysfunction (defined as RVEF <45%) but not in those without RV dysfunction. Interestingly, Mikami et al (4) found that RV function was more predictive of mortality than LV function in patients with mild-moderate LV dysfunction (LV ejection fraction >35%). This finding is of clinical importance because these patients comprise more than 70% of cases of sudden cardiac death and do not meet current criteria for primary prevention ICD implantation. Patients with preimplanted cardiac devices were excluded from prior studies despite having more advanced LV dysfunction and being at high risk of sudden cardiac death. As such, Jimenez-Juan et al confirm the negative impact of RV dysfunction on overall mortality in these patients. Importantly, Jimenez-Juan et al performed a multicenter study and included patients with both ischemic and nonischemic cardiomyopathies as well as those with ICDs implanted for primary or secondary prevention. This renders their findings widely generalizable.

The study by Jimenez-Juan et al also raises important issues that require further investigation. The retrospective design of their study makes it prone to selection bias, misclassification bias, and confounding from unmeasured risk factors (eg, underlying pulmonary embolism, pulmonary hypertension, or severe valvulopathies). Most patients with advanced renal disease are not eligible for cardiac MRI due to the risk of nephrogenic systemic fibrosis; as such, these findings should not be extrapolated to this patient group. The primary outcome included appropriate ICD therapy as a surrogate of sustained life-threatening ventricular arrhythmias. The evolution of patient selection criteria for ICD implantation and programming parameters for ventricular tachycardia detection and therapies that have occurred between 2007 and 2017 (specifically after the publication of the Multicenter Automatic Defibrillator Implantation Trial–Reduce Inappropriate Therapy, or MADIT-RIT, in 2012) may have differentially affected the number of appropriately treated ventricular tachycardia events based on enrollment time (5). Similarly, modifications and enhancements made to imaging sequences and techniques over 10 years may have affected LV scar burden measured with cardiac MRI.

Even though RVEF was a significant predictor of the outcome (mortality or appropriate ICD therapy) on multivariable analysis, it did not improve predictive power of the model beyond clinical parameters and LV function and/or scar. This finding sheds light on potential limitations of using RVEF or RV volumes for the selection of appropriate candidates for ICD therapy. One possible explanation is that RV size and function are more sensitive to underlying loading conditions compared with LV parameters (6). As such, the variability of RV parameters may not allow a reliable prediction of risk of lethal arrhythmias or sudden cardiac death at any time point. Indeed, RV function is inversely proportional to pulmonary artery pressure and even minor increase in PA pressure leads to significant drop in RV stroke volume, in part due to ventriculo-arterial uncoupling (7,8). Pressure overload (due to left-sided heart disease, pulmonary arterial hypertension, pulmonary embolism, etc) leads to hypertrophic remodeling of the RV with preservation of volume and function. When compensatory mechanisms fail, the RV then undergoes eccentric dilation with pump failure and clinical decompensation. In volume overload (valvular heart disease, left-to-right shunts, etc), the response of the RV is primarily dilatation with preservation of RV function. However, concomitant RV pressure overload and contractile dysfunction may also ensue and lead to decompensation and mortality. Furthermore, various causes of nonischemic cardiomyopathy directly involve the RV such as cardiac sarcoidosis, myocarditis, or arrhythmogenic RV cardiomyopathy. The extent of disease, its pathophysiology (infiltrative vs fibrofatty replacement vs edema and inflammation), and concomitant pressure or volume overload, all dictate changes in RV volumes and function. Focal involvement of the RV has been shown to be particularly arrhythmogenic in cardiac sarcoidosis and arrhythmogenic RV cardiomyopathy (7,8). In these cases, early ICD implant may be necessary to mitigate the elevated risk of lethal arrhythmias even before macroscopic changes in RV structure and function become apparent. This renders the use of RVEF as a decision-making tool for ICD less useful. In arrhythmogenic RV cardiomyopathy and cardiac sarcoidosis, regional and global RV strain have been shown to correlate with regional scar and predict arrhythmic outcomes (9,10). Although further studies are needed, advanced imaging techniques such as global and regional longitudinal strain using feature tracking may allow detection of focal RV myopathy, and identification of patients with nonischemic cardiomyopathy who would benefit from ICD implantation for primary prevention of sudden cardiac death.

In conclusion, the study by Jimenez-Juan et al provides the basis for exploring the role of RV structure and/or function in predicting mortality and appropriate ICD therapy in patients with cardiomyopathy. Prospective randomized studies using global or regional RV function in nonischemic cardiomyopathy as a modifier or primary determinant for selecting candidates for ICD implantation are needed.