QT interval prolongation by acute gastrointestinal bleeding in patients with cirrhosis

Background & aims QT interval prolongation is frequent in cirrhosis, and stressful conditions could further prolong QT. We aimed to test this hypothesis and, if it proved correct, to assess its prognostic meaning. Methods We reviewed the clinical records of 70 consecutive cirrhotic and 40 non-cirrhotic patients with acute gastrointestinal bleeding. All patients had been evaluated before bleeding (T0) and were re-evaluated at the time of bleeding (T1) and 6 weeks afterwards (T2). Results QT corrected by heart rate (QTc) lengthened at T1, returning towards baseline values at T2 (mean ± SEM; from 415.9 ± 4.3 to 453.4 ± 4.3 to 422.2 ± 5.7 ms, P < 0.001) in cirrhotics; contrariwise, QTc did not change in non-cirrhotic patients. The 6-week mortality was 29.6% among cirrhotic patients, while no control patient died. At T1, patients who died had longer QTc (P = 0.001) and higher model of end-stage liver disease (MELD) score (P < 0.001) than survivors. MELD and QTc independently predicted survival. Their areas under the ROC curve were 0.88 (CI 95% 0.78–0.95) and 0.75 (CI 95% 0.63–0.85) respectively; the best cut-off values were MELD ≥20 and QTc ≥ 460 ms. Based on these factors, the 6-week mortality was: 0% for patients without risk factors, 32.1% for those with one risk factor and 70.6% for those with both (P < 0.001). Conclusions Acute gastrointestinal bleeding further prolongs QTc in cirrhosis. This abnormality independently predicts bleeding-induced mortality. The combined measurement of QTc interval and MELD can clearly identify three patient strata at increasing risk of bleeding-related mortality, thus improving the decision-making for these patients.