To the Editor.—
Davidson et al1 reported that plasma homovanillic acid (pHVA) concentrations decreased in a group of neuroleptic responders (n=7) but were unchanged in nonresponders (n=13). Within the framework of a large experiment,2 we collected analogous data. To replicate their results, we emulated their analyses with our data set.Our experiment2 used 176 acutely exacerbated schizophrenic or schizoaffective patients. After a placebo period, they were randomly assigned to one of three plasma levels of haloperidol for 6 weeks. Weekly measures included haloperidol plasma levels, the Brief Psychiatric Rating Scale (BPRS), Clinical Global Impression (CGI), and, in a subset of patients, pHVA (assayed with a modification of a published method3 using gas chromatography/mass spectroscopy and deuterated HVA as an internal standard).Emulating the sampling criteria of Davidson et al,1 we selected 28 schizophrenic patients who had a minimum of four points on the CGI at baseline