Abstracts

Rationale: There is emerging evidence that cytochrome P-450 inducing anticonvulsant medications have adverse effects on atherogenic serum lipids (Mintzer S, Curr Opin Neurol 2010; 23: 164-169). Data, however, are as yet preliminary. Methods: The Geisinger Health System (GHS) Clinical Decision Intelligence System (CDIS) database of all patients under GHS clinical management was electronically queried for all patients prescribed any anticonvulsant medication for an epilepsy diagnosis. Initial review of database records revealed accurate medication records only for patients with GHS primary care, and were limited to these 52419 records, of which 4102 carried an epilepsy diagnosis. Results were included in the study for subjects with medication records to substantiate a 6 month time period on an enzyme-inducing AED, any separate and distinct 6 month time period on a non-enzyme-inducing AED without enzyme-inducing AED, and with atherogenic lipid levels drawn and reported during both time periods, with the off-inducer lipid levels accepted only if drawn at least 60 days after any inducer had been stopped. Low density lipoprotein (LDL) levels, high density lipoprotein (HDL) levels, and total cholesterol to HDL ratio were available in high numbers and were included; homocyst(e)ine, CRP and lipoprotein A levels were too few for use and were excluded. 40 records fell within the parameters identified. Results: HDL on inducing AED: 52.5mg/dL; off inducing AED 50.7; t=0.38 (p>0.10). LDL on inducing AED: 104.5mg/dL; off inducing AED 96.5; t=0.15 (p>0.10). Chol/HDL ratio on inducing AED: 4.5; off inducing AED 4.7; t=0.22 (p>0.10). Conclusions: These results do not provide prima facie evidence of a significant difference between in LDL and HDL cholesterol levels between time periods on inducing and non-inducing epileptic medications in patients treated for epilepsy. While results are not controlled for confounds such as age, smoking status and antihyperlipidemic drug dose, initial raw results do not suggest such control will prove fruitful. Careful prospective analysis will likely be necessary to demonstrate this possible effect if it exists.