Summary Background Sudden unexplained death in psychiatric patients may be due to drug-induced arrhythmia, of which lengthening of the rate-corrected QT interval (QTc) on the electrocardiogram is a predictive marker. We estimated the point prevalence of QTc lengthening in psychiatric patients and the effects of various psychotropic drugs. Methods Electrocardiograms were obtained from 101 healthy reference individuals and 495 psychiatric patients in various inpatient and community settings and were analysed with a previously validated digitiser technique. Patients with and without QTc lengthening, QTc dispersion, and T-wave abnormality were compared by logistic regression to calculate odds ratios for predictive variables. Findings Abnormal QTc was defined from the healthy reference group as more than 456 ms and was present in 8% (40 of 495) of patients. Age over 65 years (odds ratio 3·0 [95% CI 1·1–8·3]), use of tricyclic antidepressants (4·4 [1·6–12·1]), thioridazine (5·4 [2·0–13·7]), and droperidol (6·7 [1·8–24·8]) were robust predictors of QTc lengthening, as was antipsychotic dose (high dose 5·3 [1·2–24·4]; very high dose 8·2 [1·5–43·6]). Abnormal QT dispersion or T- wave abnormalities were not significantly associated with antipsychotic treatment, but were associated with lithium therapy. Interpretation Antipsychotic drugs cause QTc lengthening in a dose-related manner. Risks are substantially higher for thioridazine and droperidol. These drugs may therefore confer an increased risk of drug-induced arrhythmia. Lancet 2000; 355: 1048–52 Introduction Cardiovascular mortality in psychiatric patients is high. 1 Reports of sudden unexplained death in those taking antipsychotic drugs 2,3 have raised the concern that part of this excess may be due to drug-induced arrhythmias, since many of these drugs have cardiac electro- physiological effects similar to those of quinidine. 2 The polymorphic ventricular arrhythmia known as torsade de pointes has been recorded in patients with psychotropic drug overdose 4 and provides a plausible mechanism for sudden unexplained death associated with drug therapy. 5 Several psychotropic drugs are associated with lengthening of the rate-corrected QT interval (QTc) on the electrocardiogram, 2 which often precedes torsade. 6 There is no direct evidence linking the extent of drug- induced QTc lengthening with the risk of torsade or sudden death. However, QTc-interval lengthening is a predictor of sudden death in patients with cardiac disease 7 and the extent of drug-induced QTc-interval lengthening is thought to be an important marker of arrhythmia risk by drug regulatory authorities (see website: www.emeasearch.is.eudra.org/humandocs/PDFs/ SWP/098696en.pdf). Risk of arrhythmia with drugs that lengthen ventricular repolarisation may also be indicated by the dispersion of repolarisation, which can be assessed by measuring QT dispersion. 8 Abnormal repolarisation may also cause non-specific abnormalities of the T wave, although there is no direct evidence to link such changes with arrhythmia. Clinical guidelines advise caution in the use of high- dose antipsychotic therapy with special reference to the risk of sudden death, as well as regular monitoring of the QTc interval, 9 but evidence for this change in practice is small. Only one study has systematically examined QTc- interval abnormalities in psychiatric patients, and found an increased prevalence of abnormalities in a schizo- phrenic population compared with controls, and an association with high-dose antipsychotic therapy. 10 This sample was not large enough to detect differences between drugs, and no relation was found between drug dosage and QTc dispersion. Therapeutics in mental illness is moving away from high-dose antipsychotic drugs to lower doses and newer drugs; both these strategies have proven benefits, but neither are free of cardiotoxic effects. 2,11 If lengthening of the QTc interval is to be adopted as a marker of arrhythmogenic risk in psychiatric patients, its relation with dose should be clarified and its association with all antipsychotic drugs, antidepressants, and other classes of psychotropics should also be examined. We measured the frequency of QTc lengthening in a large heterogeneous population of psychiatric patients, and assessed whether QTc lengthening was associated with individual antipsychotic drugs, antipsychotic drug dose, and other risk factors. ARTICLES 1048 THE LANCET • Vol 355 • March 25, 2000 QTc-interval abnormalities and psychotropic drug therapy in psychiatric patients J G Reilly, S A Ayis, I N Ferrier, S J Jones, S H L Thomas Academic Department of Psychiatry (J G Reilly MRCPsych, Prof I N Ferrier MD) and Wolfson Unit of Clinical Pharmacology (S A Ayis PhD, S J Jones RGN, S H L Thomas MD), University of Newcastle Upon Tyne, Newcastle upon Tyne NE2 4HH, UK; and Tees and North East Yorkshire NHS Trust, Parkside Community Mental Health Centre, Park Road North, Middlesbrough, UK (J G Reilly) Correspondence to: Dr S H L Thomas