Back

Hypophosphataemia in suspected seizures evaluated in the first seizure clinic and neurology consults

Binks, S. N.; Zorkin, D.; Liem, B.; Sen, A.

2025-07-01 neurology
10.1101/2025.06.26.25329859 medRxiv
Show abstract

Transient loss of consciousness (TLoC) is a leading cause of referrals to acute neurological services. A witness account is often lacking and ancillary investigations are a critical diagnostic adjunct. Hypophosphataemia was recently identified as a potential maker of epileptic seizures in ward and emergency department presentations. We evaluated the real-world utility of checking phosphate in an unselected cohort of people presenting with TLoC. We retrospectively reviewed 182 episodes (91 referrals to first seizure clinic and 91 consults) from 173 patients. We assessed nine pre-specified serological markers frequently measured in people presenting with TLoC. Raw P-values comparing mean levels showed significant difference between epileptic seizures and non-epileptic episodes only for phosphate (0.98 vs. 1.19 mmol/L, P=0.006) and lactate (2.82 vs. 1.82 mmol/L, P=0.007). No blood biomarkers were significant after multiple comparison correction, although a phosphate below 0.8mmol/L was significantly more likely to associate with epileptic seizures than non-epileptic episodes (17/64, 26.5% vs. 7.4%, 2/27, P=0.049). Logistic regression showed that a model including lactate and phosphate was most accurate to predict epileptic seizures with an area-under-the ROC curve of 0.728 (96% CI 0.607-0.848). Checking serum phosphate may be valuable in helping to determine the aetiology of an episode of TLoC. Plain language summaryWe studied blood test results of 173 people suspected of having had an epileptic seizure who presented to a UK hospital neurology service. Although blood phosphate tests were infrequently requested, our results suggest a low phosphate level could be useful to help distinguish between epileptic and non-epileptic attacks.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.