Relation between coffee consumption and risk of seizure-related respiratory dysfunction in patients with drug-resistant focal epilepsy.

Caffeine is an antagonist of the adenosine pathway, which is involved in regulation of breathing. Extracellular concentrations of adenosine are increased in the immediate aftermath of a seizure. Seizure-related overstimulation of adenosine receptors might promote peri-ictal apnea. However, the relation between caffeine consumption and risk of seizure-related respiratory dysfunction in patients with drug-resistant focal epilepsy remains unknown. We performed a cross-sectional analysis of data collected in patients included in the SAVE study in Lyon's epilepsy monitoring unit at the Adult Epilepsy Department of the Lyon University Hospital between February 2016 and October 2018. The video-electroencephalographic recordings of 156 patients with drug-resistant focal epilepsy included in the study were reviewed to identify those with ≥1 focal seizure (FS), valid pulse oximetry (SpO <sub>2</sub> ) measurement, and information about usual coffee consumption. This latter was collected at inclusion using a standardized self-questionnaire and further classified into four groups: none, rare (≤3 cups/week), moderate (4 cups/week to 3 cups/day), and high (≥4 cups/day). Peri-ictal hypoxemia (PIH) was defined as SpO <sub>2</sub> < 90% for at least 5 s occurring during the ictal period, the post-ictal period, or both. Ninety patients fulfilled inclusion criteria, and 323 seizures were analyzed. Both the level of usual coffee consumption (p = .033) and the level of antiepileptic drug withdrawal (p = .004) were independent risk factors for occurrence of PIH. In comparison with FS in patients with no coffee consumption, risk of PIH was four times lower in FS in patients with moderate consumption (odds ratio [OR] = .25, 95% confidence interval [CI] = .07-.91, p = .036) and six times lower in FS in patients with high coffee consumption (OR = .16, 95% CI = .04-.66, p = .011). However, when PIH occurred, its duration was longer in patients with moderate or high consumption than in those with no coffee consumption (p = .042). Coffee consumption may be a protective factor for seizure-related respiratory dysfunction, with a dose-dependent effect