Seizure length and clinical outcome in electroconvulsive therapy using methohexital or thiopental

Seizure duration is an extensively studied and controversial indicator of treatment quality in electroconvulsive therapy. Previous research comparing the effect of the barbiturate anesthetics methohexital and thiopental on seizure duration has yielded conflicting results. A recent period of unavailability of methohexital in the United States allowed for retrospective comparison of seizure length as well as clinical improvement in treatment using each agent. Retrospective review was made of 837 treatments administered to 97 patients between January 2, 2002, and May 31, 2003, examining anesthetic, seizure duration, and Global Assessment of Functioning (GAF) scores of inpatients at hospital admission and discharge. Analysis of variance of treatments 2-5 showed no significant effect for anesthetic on seizure duration. Analysis on a treatment-by-treatment basis revealed a marginally significant trend toward shorter EEG seizures in the thiopental group at the second treatment (50.5 +/- 23.6 s vs. 61.1 +/- 27.9 s; P = 0.07) and fifth treatment (41.7 +/- 16.9 s vs. 51.8 +/- 24.0 s; P = 0.07). A difference approaching statistical significance revealed shorter convulsion length in the thiopental group at treatment 5 (29.0 +/- 12.3 s vs. 34.8 +/- 12.3 s; P = 0.07). Comparison of GAF score improvement at hospital discharge revealed no significant difference (GAF increase 26.4 +/- 9.4 for methohexital-treated patients vs. 24.8 +/- 12.0 for thiopental-treated patients; t = 1.00, df = 82, P > 0.1). Trends approaching significance in treatments 2 and 5 revealed shorter seizures in the thiopental group. However, data on clinical recovery reveals no greater improvement in the methohexital group. Thus, this study calls further into question the premise that choice of barbiturate anesthetic may affect clinical efficacy.