QT and Outcomes in Cirrhosis: A prolonged debate on causality in need of correction

As the field of transplant has advanced, cardiovascular events have become the leading cause of early death after liver transplantation (LT) ahead of graft failure and infection.¹ This trend has been bolstered by the transplantation of older and sicker patients who have an increasing amount of cardiometabolic risk factors, accentuating the need to determine which patients should undergo more extensive cardiac evaluation. The electrocardiogram (ECG) has been used for decades for perioperative risk stratification, and in particular the QT interval corrected for heart rate (QTc) has been of interest. The QTc is a measure of ventricular repolarization, a time during which ectopic beats may cause potentially lethal arrhythmias, and it is calculated by the ratio of the QT interval to R-R interval.² In patients with end stage liver disease (ESLD) many risk factors are present that may prolong the QTc interval—electrolyte derangements, pH dependent calcium binding to albumin, chronic alkalosis, elevated sympathetic tone, and medications such as diuretics, antibiotics (e.g., quinolones) and beta blockers.³ Furthermore, QTc appears to be correlated with severity of liver disease and normalizes within a few months after LT.⁴

In this issue of AJT, Koshy et al. develop a ‘Cardiac Arrest Risk Index’ (CARI) using a retrospective cohort of 408 patients undergoing LT at a state-wide transplant center in Melbourne, Australia. The primary outcome, defined as a composite of either ventricular arrhythmia (e.g., ≥ 3 beats of ventricular tachycardia (VT) or ventricular fibrillation) or need for cardiopulmonary resuscitation occurred in 26 (6.4%) patients within the first 30 days of LT. Of note, 14 of the 26 patients’ outcomes were due to ≥ 3 beats of VT, nearly all of which occurred within 72 hours of surgery. The authors use a logistic regression model to determine the most significant parameters and propose the CARI which was rated on a scale of 0–5 (two points for QTc ≥480 milliseconds, one point for age ≥65 years, one point for MELD score ≥30 and one point for male sex). They note that a score >2 conferred a positive likelihood ratio of 5 for an arrhythmia in the first 30 days after LT. These findings support prior literature that QTc may be an important marker for prognosis in patients with ESLD independent of MELD score.^4,5 Although no longer a major criterion for cirrhotic cardiomyopathy,⁶ the current study suggests that prolonged QTc in ESLD may be a marker for subclinical cardiac dysfunction that may be unmasked by LT.

The study by Koshy et al. attempts to fill an important gap in predicting perioperative cardiac arrest, which accounts for half of all perioperative events after LT.^1,7,8 A risk assessment tool for cardiac arrest is critical to target intensive prevention strategies. For example, beta blockade has been shown to reduce prolonged QTc among patients with cirrhosis, but its relationship to cardiac outcomes after LT is unknown.^9,10 Importantly, there are several issues with the proposed CARI risk tool that should be addressed in future studies before clinical implementation can be considered. First, although the authors sought to predict cardiac arrest, approximately half of the patients met the primary composite outcome due to ‘≥3 beats of VT’ which is of unclear clinical significance and does not equate to cardiac arrest. VT is typically defined as sustained (lasting longer than 30 seconds) which can lead to hemodynamic instability as compared to non-sustained (≥3 beats of VT but less than 30 seconds) which may be completely asymptomatic. In fact, practice-based guidelines in cardiology advise against intervention for non-sustained VT and overzealous treatment of ventricular ectopy with anti-arrhythmic therapy has been linked with increased mortality.^11,12 Although these guidelines are not specific to perioperative VT, it is imperative that future studies with higher sustained VT event rates separate non-sustained VT from sustained VT in order to determine the true hemodynamic and clinical significance of VT in LT. Furthermore, 9 of the 26 events in the current study were due to asystole during reperfusion. The relevance of prolonged QTc, which classically degenerates into ventricular arrhythmias, remains to be examined in these cases. Second, the group of patients who met the primary composite outcome not only had a higher MELD score, but more encephalopathy, more hepatorenal syndrome, more ascites, and less hepatocellular carcinoma suggesting a later stage of ESLD at LT. The overall low composite event rate may have hampered efforts to adjust for these potential confounders which are known to correlate with both QT and perioperative LT complications. Although the authors adjusted for MELD score as a surrogate for ESLD severity, it is well-established that MELD alone may not be the best predictor of perioperative complication.¹³ Similarly, most (24 of 26) events occurred in the early post-operative period (e.g., within 1 week of LT) and could potentially be mediated by the much higher vasopressor dose administered to the sicker patients in this group.

In summary, as the field of LT trends towards higher cardiac risk patients larger studies are needed to fully address whether the QTc interval is a modifiable risk factor in perioperative cardiac arrest or if it is simply a marker of sicker patients.