First-Time Seizures Signal Hidden Cancer Risk Far Beyond the Brain

A first-time seizure is often evaluated primarily as a neurological event, but this landmark population-based cohort study from Denmark challenges clinicians to think more broadly. Published in JAMA Neurology, the study examined whether first-time seizures serve as an early clinical signal of undiagnosed cancer — both within the nervous system and elsewhere in the body. The findings are striking: seizures appear to be a sentinel event for occult malignancy at a scale not previously quantified in a nationwide cohort.

The study drew on Danish nationwide medical registries spanning January 1996 through December 2022, identifying 49,894 adults aged 18 or older who received a first-time hospital diagnosis of seizure and had no prior cancer history. The median age at seizure diagnosis was 51.5 years (IQR 35.6–67.8), and 41.4% were women. Outcomes were measured as absolute risks (ARs) and standardized incidence ratios (SIRs) compared to the general Danish population, stratified across three time windows: within 1 year, 1 to less than 5 years, and 5 to 20 years post-seizure.

Within the first year of follow-up, 1,172 neurological and 850 non-neurological cancers were diagnosed. The 1-year AR for any cancer was 4.1%, with an SIR of 5.30 (95% CI, 5.07–5.54). For neurological cancers specifically, the 1-year AR was 2.4% and the SIR was an extraordinary 76.1 (95% CI, 71.8–80.6), reflecting the well-established link between brain tumors and seizures. For non-neurological cancers, the 1-year AR was 1.7% with an SIR of 2.32 (95% CI, 2.17–2.48) — a more than twofold elevation that is clinically meaningful and less widely appreciated.

The elevated risk persisted beyond the first year, though at lower magnitudes. From 1 to less than 5 years, 87 neurological and 1,226 non-neurological cancers were observed, yielding SIRs of 1.85 (95% CI, 1.48–2.28) and 1.15 (95% CI, 1.09–1.22), respectively. From 5 to 20 years, 112 neurological and 2,120 non-neurological cancers were recorded, with SIRs of 1.46 (95% CI, 1.20–1.75) and 1.33 (95% CI, 1.28–1.39). The overall 5-to-20-year AR for any cancer reached 13.4%, with an SIR of 1.34 (95% CI, 1.28–1.40), suggesting a sustained but modest long-term cancer signal.

The clinical implications are significant. The massive short-term SIR for neurological cancers confirms that brain tumors frequently present with seizures, but the doubling of non-neurological cancer risk in the first year is a less-recognized finding that warrants attention. This may reflect paraneoplastic mechanisms, where systemic cancers trigger seizures through immune-mediated or metabolic pathways, or simply that diagnostic workup following a seizure uncovers previously silent malignancies. The authors argue that broader cancer screening — beyond neuroimaging — should be considered in select patients presenting with a first-time seizure, particularly older adults or those with other risk factors.

Key limitations include the registry-based design, which relies on administrative diagnostic codes and may not capture all clinical nuances. The study cannot fully distinguish between seizures that directly caused cancer detection (surveillance bias) and those that are biologically linked to cancer development. Nonetheless, the nationwide scope, long follow-up, and large sample size provide robust epidemiological evidence that first-time seizures are a clinically important signal deserving a more comprehensive diagnostic response than is currently standard practice.