Abstracts

RATIONALE: Many factors influence drug effects, including drug concentration at the active site. Steady state blood levels correlate with tissue levels. The first drug levels were bromide levels.
METHODS: Review books & papers from 1927-2001.
RESULTS: Wuth devised a serum bromide assay in 1927. He studied 50 psychiatric inpatients and said that Br levels above 125 mg/dl were [dsquote]unwise[dsquote], without raw data. His idea was embraced by psychiatrists, but few epileptologists. Experimental studies (to be reviewed) found weak correlation between level and clinical state. Greenblatt noted [dsquote]major individual differences in the EEG and clinical picture at the same blood bromide level[dsquote]. Bromide use had declined by the time that these problems became clear.
Buchtal and coworkers devised barbiturate and phenytoin assays in the 50s. At the same time that they proposed a close relationship between PHT level and drug effects, they noted decreased barbiturate effects with continuing use, although serum levels rose. This didn[ssquote]t fit a simple pharmacokinetic model. Buchtal[ssquote]s papers stimulated enthusiastic observational studies. However, the one double blind study of PHT levels, efficacy and adverse effects (Schumacher) found no relationship. [dsquote]the value of the .. therapeutic range in predicting patient response is quite limited.[dsquote] Experimental studies included volunteers, whose AED levels were correlated with cognitive performance. These studies will be reviewed; they suffer from various limitations.
The concept of therapeutic drug levels implies that trough levels are reproducible over time. This is least true of valproate (Jackson found no correlation over time). Development of tolerance varies between drugs (greatest with barbiturates and benzodiazepines) and individuals. Drugs alter tolerance to other drugs, as does stress (Haile).
Skilled clinicians optimize therapy without levels. However no clinician can tell whether a patient received her medication yesterday. Consider a patient who missed several days of medication, had several seizures, took medicine and came to the ER with a [dsquote]high blood level[dsquote]? Something analogous to glycosylated hemoglobin is needed. Hair levels may meet this need (Williams). The various uses of drug levels will be reviewed.
References
Greenblatt, M, et al, Arch Neurol Psychiat 53:431, 1945.
Haile, CN, et al, Psychopharmacology 154:213, 2001
Jackson, M, et al, Seizure 3:225, 1994
Reddy, DS, Rogawski, MA, J Pharmacol Exp Ther. 295: 1241, 2000
Schumacher, GE, et al, Ther Drug Monit. 13: 318, 1991
Svensmark, O, Buchtal, F, Epilepsia 1: 373, 1959/60
Williams, J, et al, Ther Drug Monit 23:15, 2001
Wuth, O, JAMA 88:2013, 1927
CONCLUSIONS: Observational and experimental studies give different perspectives on anticonvulsant levels. Equations to predict drug effect need both pharmacokinetic and pharmacodynamic variables.
Support: Departmental funds.