Five years of anticonvulsant monitoring on site at the epilepsy clinic.

The influence of antiepileptic drug monitoring on site at the epilepsy clinic was assessed in a prospective way by recording physicians' decisions before and after the drug concentration became available over the five-year period including 1986 to 1990. A total of 2,857 assays (1,121 carbamazepine, 798 phenytoin, 622 sodium valproate, and 72 phenobarbital) were performed during 2,696 hospital attendances on 618 epileptic patients. A gradual reduction in the proportion of visits monitored annually was observed (1986: 42%; 1987: 51%; 1988: 44%; 1989: 39%; 1990: 28%; p < 0.01). In 481 (17.8%) visits a change in management resulted from knowledge of the drug concentration. This was not an automatic response, as a large minority of unexpected results did not produce such an alteration. The proportion of decisions to adjust the drug dose after an unexpected result, however, tended to increase over the years of the study (1986: 43%; 1987: 70%; 1988: 60%; 1989: 78%; 1990: 54%; p < 0.01). Changes in anticonvulsant dose following an unexpected result were observed in "low" (59%), "therapeutic" (61%), and "high" (52%) concentration categories. Possible benefit for patients can be inferred from the observation that more assay results fell within the target range in the last (61%) compared with the first (33%) year of monitoring (p < 0.01).