In two North American databases (Saskatchewan Health and Medicaid), studies were initiated by different investigators. New users were identified as anyone with at least a two-year history in the insurance plan at the time of the first prescription of one of three anticonvulsant drugs. Hospitalizations for cutaneous conditions during the first 60 days of drug exposure were identified, and anonymized medical records were reviewed by dermatologic experts to substantiate the discharge diagnosis. The specific case definitions were left to the discretion of each investigator, but the basic study designs were similar. The Saskatchewan study showed that the risk associated with phenytoin initiation was 1/1000 (Tennis and Stern, 1997) and the risk associated with initiation of carbamazepine was 0.6 per 1000. In the two Medicaid states (US), the risks associated with two alternative therapies ranged from 0.3 to 1.6 per 1000 (Judith Jones, personal communication). Although based on small numbers of cases, these results suggested that the overall risk of serious cutaneous reactions in patients initiating lamotrigine was not dissimilar from that in patients initiating some alternative therapies. However, because of the very small numbers of events, the real-world risk of serious rash in children has not yet been well defined for these other anticonvulsant therapies.
When the International Case-Control Study on Severe Cutaneous Adverse Reactions published data on the association of anticonvulsant therapies with SJS and TEN (Rzany et al., 1999) it appeared that the risk associated with lamotrigine was similar to that of other anticonvulsants (Table 23.1). However, the number of cases using lamotrigine was small.
This study shows how case-control studies can be useful for detecting the role of medications as risk factors for an outcome. However, case-control studies alone cannot quantify the frequency of an event.
The German population-based registry of severe skin reactions was extremely useful for understanding SJS or TEN associated with lamotrigine.
Was this article helpful?