Abstract
The introduction of sensitive laboratory assays for the analysis of cyclic antidepressant blood levels more than 15 years ago promised a rational basis for drug treatment of serious depressive illness. Early studies demonstrated a linear or sigmoid relation between the blood level of imipramine and the therapeutic response. For nortriptyline, a curvilinear inverted U, the so-called therapeutic window, was established. In other studies of patients who had taken overdoses of antidepressants, blood levels also correlated well with cardiac toxicity: all patients with blood levels above 1000 ng per milliliter had a QRS of more than 100 msec. Unfortunately, in recent . . .