Influence of Risk Factor Combinations on Incidence Rates of SUDEP

Information on absolute risks of sudden unexpected death in epilepsy (SUDEP) in individual patients with epilepsy is scarce. Our main objective was therefore to explore the range in incidence rates of SUDEP to provide a more solid basis for individualized counseling and to characterize patients with high and very high SUDEP incidence for future intervention studies aiming at prevention of SUDEP. We used data on everyone in Sweden diagnosed with epilepsy from 1998 to 2005 (n = 60,952), followed until 2011 and identified SUDEP cases through adjudication of deceased patients who had epilepsy. We conducted a nested case-control study and retrieved detailed information on clinical characteristics for the SUDEP cases and matched living epilepsy controls (5:1). Estimates of the strengths of associations from the case-control study were used to estimate the incidence rate of SUDEP (per 100,000 person-years) in the full cohort by SUDEP risk factors. Two hundred fifty-five SUDEP cases (median age 48 years; 154 males) were identified. The lowest incidence, 8 (95% CI 3-17), was observed among patients without a history of tonic-clonic seizures (TCS), whereas patients with 1 or more TCS the preceding year had an incidence of 287 (95% CI 192-428). Incidence rates above 200 were also found among patients with a history of nocturnal TCS, substance abuse or alcohol dependence, and nonadherence with antiseizure medication (ASM) treatment. Considering combination of risk factors, the incidence rate was very low, 5 (95% CI 2-12), for patients who share bedroom and are free from TCS the preceding year as well as adherent with the prescribed ASM treatment. By contrast, patients living alone who are nonadherent have a history of nocturnal TCS and at least 1 TCS the preceding year have an incidence at 1808 (95% CI 594-5,504), more than 350 times higher than the low-risk patient. In this analysis of Swedish population-based SUDEP data, we have identified a 350-fold difference in the SUDEP incidence depending on individual circumstances and epilepsy characteristics. Although somewhat old, our data should be useful for patient counseling about individual SUDEP risks amenable to modification and for case stratifications for intervention studies aiming at prevention of SUDEP.