Drug-induced torsades de pointes in an underserved urban population. Methadone: is there therapeutic equipoise?

BACKGROUND Although it has been well established that methadone use can result in prolonged QTc/torsades de pointes (TdP) and has been labeled as one of the main drugs that cause TdP, it is still prescribed indiscriminately, and several cases of methadone-associated TdP have been seen in our community. METHODS Our objective was to determine the associated factors for prolonged QTc and the development of torsades de pointes (TdP) in our underserved patient population. We found 12,550 ECGs with prolonged QTc between 2002 and 2013. Medical records were reviewed in order to identify precipitating factors for prolonged QTc and to detect incidence of TdP. RESULTS We identified 2735 patients with prolonged QTc who met the inclusion criteria. Of these, 89 (3%) experienced TdP. There was a greater prevalence of HIV infection in the TdP group (11.2 vs. 3.7%, p < 0.001). Furosemide, hydrochlorothiazide, selective serotonin reuptake inhibitors (SSRIs), amiodarone, ciprofloxacin, methadone, haloperidol, and azithromycin were the drugs most often associated with prolonged QTc (31, 8.2, 7.6, 7.1, 3.9, 3.4 and 3.3%, respectively). However, the agents most commonly associated with TdP were furosemide (39.3%), methadone (27%), SSRIs (19.1%), amiodarone (18%), and dofetilide (9%). The medications with statistical significance in the multivariate analysis for TdP development in descending order were as follows: ranolazine (odds ratios [OR] = 53.61, 95% confidence interval [CI] 5.4-524, p < 0.001), dofetilide (OR = 25, CI 6.47-103.16, p < 0.001), voriconazole (OR = 21.40, CI 3.24-124.25, p < 0.001), verapamil (OR = 10.98, CI 2.62-44.96, p < 0.001), sotalol (OR = 12.72, 1.95-82.81, p = 0.008), methadone (OR = 9.89, CI 4.05-24.15, p < 0.001), and SSRI (OR = 2.26, CI 1.10-5.96, p < 0.001). This multivariate analysis revealed that amiodarone and HIV infection were not implicated in TdP. CONCLUSION Methadone was by far the leading medication implicated in the development of TdP and an independent predictor in both univariate and multivariate analyses despite the fact that it was not the most common QT-prolonging medication in our population