OBJECTIVE To test the hypothesis that sudden unexplained death (SUD) in epilepsy is related to geomagnetic activity. BACKGROUND Prior studies presume that geomagnetic activity (with average amplitudes above 50 nanotesla [nT]) is associated with SUD in epileptic human patients and in epileptic laboratory rats. METHODS In a retrospective study, 46 epileptic patients with definite SUD were compared with 108 epileptic patients with known cause of death (KCD) who died between 1981 and 1992. A complete postmortem examination was performed in all cases. The time of the day and date of death, as well as two international geomagnetic indices concerning Bartels' planetary 3-hour signs (Kp) and the mean planetary daily amplitudes (Ap) at time of death, were assessed. RESULTS Among 45 SUD individuals, the local time (37.8%) and the universal time of death (35.6%) peaked within the critical period between 3 to 9 AM. However, the SUD and KCD group did not substantially differ in regard to the distribution of local or universal time of death (p > 0.2, Fisher test). Neither the Kp signs at death and 2 hours before death nor the Ap values showed considerable differences between the SUD and KCD series (p > 0.2, Mann-Whitney test). Merely 4.3% of SUD patients and 3.7% of KCD patients were associated with Ap indices above 50 nT (p > 0.2, Fisher test). CONCLUSIONS The results do not support the hypothesis that geomagnetic activity is related to occurrence of sudden unexplained death in epileptic patients.