Lamotrigine serum concentrations as longitudinal biomarkers for seizure risk prediction in pregnant women with epilepsy: a secondary analysis of the EMPiRE study
John Allotey, Fatima Junaid, Teresa Peréz, Valeria Rolle, María del Carmen Pardo, Javier Zamora, Shakila ThangaratinamBackground
Serum lamotrigine concentrations fall during pregnancy, and therapeutic drug monitoring is often used to guide dose adjustment. We do not know whether serial antenatal measurements improve prediction of seizure risk in pregnant women with epilepsy.
Objective
To evaluate whether changes in maternal serum lamotrigine concentration levels are associated with seizure occurrence in pregnancy and whether serial measurements improve prediction of seizure risk beyond baseline clinical factors.
Design
Secondary analysis of data from the AntiEpileptic drug Monitoring in PREgnancy (EMPiRE) study, a multicentre, double-blind randomised trial nested within a prospective cohort, using joint longitudinal-survival modelling.
Setting
Fifty maternity units in the UK between November 2011 and May 2015.
Methods
We included pregnant women with epilepsy taking lamotrigine (monotherapy or polytherapy) at baseline and who had a baseline and at least one follow-up serum lamotrigine concentration measured before a seizure or up to 6 weeks postpartum. The primary outcome was time to first seizure during pregnancy or within 6 weeks postpartum. We used joint longitudinal-survival models to assess the association between serial lamotrigine concentrations and seizure occurrence, adjusting for baseline clinical predictors from the validated EMPiRE seizure risk model. Associations were expressed as hazard ratios (HRs) with 95% confidence intervals (CIs).
Results
Of 560 women recruited to the EMPiRE trial, 183 women contributed data; 46 women had seizures in pregnancy. Longitudinal serum lamotrigine concentrations were not associated with seizure occurrence (HR 0.421, 95% CI 0.157 to 1.133, p=0.087). Baseline lamotrigine daily dose was the only independent predictor of seizure (HR 1.005, 95% CI 1.003 to 1.007, p<0.001). The model based on serum lamotrigine levels had poor discrimination with an AUC ranging from 0.46 to 0.55 throughout pregnancy.
Conclusion
Longitudinal serum lamotrigine concentrations do not appear to predict seizure occurrence in pregnancy. Seizure risk was better explained by baseline lamotrigine dose.