Non-Interventional retrospective longitudinal study in the United Kingdom and France to investigate the therapeutic strategies after discontinuation of valproate and related substances in clinical practice (VALSE study - VALNAC09344)

Study type

• Study topic:
  Disease /health condition

  Human medicinal product
• Study type:
  Non-interventional study
• Scope of the study:
  Disease epidemiology

  Drug utilisation

  Effectiveness study (incl. comparative)
• Data collection methods:
  Secondary data collection

Non-interventional study

• Non-interventional study design:
  Cohort

Study drug and medical condition

• Anatomical Therapeutic Chemical (ATC) code:
  (N03AG01) valproic acid

  (N03AG02) valpromide
• Medical condition to be studied:
  Epilepsy

  Bipolar disorder

Population studied

• Short description of the study population:
  The study population comprised of patients aged 13-49 years old diagnosed with epilepsy or bipolar disorder who had discontinued valproate and related substances from 2014 to 2017 identified through UK the electronic medical records database (Clinical Practice Research Datalink) and the French nationwide claims (Système National des Données de Santé).
• Age groups:
  Adolescents (12 to < 18 years)

  Adults (18 to < 46 years)

  Adults (46 to < 65 years)
• Special population of interest:
  Other

  Pregnant women
• Special population of interest, other:
  Patients with epilepsy and bipolar disorders
• Estimated number of subjects:
  384

Study design details

• Main study objective:
  To determine clusters of patients that are the most likely to reflect a success in epilepsy/ BD management after VPA discontinuation For each cluster, success will be defined based on the absence of VPA reintroduction in the follow-up period, and contextualised according to several parameters (clinical relapse, number of hospitalizations, polypharmacy) and discussed with a Scientific Committee.
• Outcomes:
  Description of all situations of therapeutic management after discontinuation observed in clinical practice: clusters defined by an unsupervised clustering method. The clusters and their characteristics will be reviewed by a Scientific Committee to determine which, in their experience, are likely to reflect successful management of epilepsy/BD after VPA discontinuation. Secondary outcomes: first occurrence of VPA reintroduction, occurrence of clinical relapse, occurrence of pregnancy. Other outcomes will be also considered (hospitalisation and discharge diagnoses, emergency room visits, diagnoses…). Modelling analyses to identify factors associated with clusters will be conditioned by the size of the clusters and the number of candidate covariates.
• Data analysis plan:
  The following analyses will be performed separately for each database: - Description of women’s recruitment, women’s characteristics, treatment patterns of valproate, - Identification and characterisation of the different clusters identified in each database (if available) during follow-up, and based on unsupervised machine learning methods (clustering analysis) - Description of secondary outcomes - Identification of covariates associated with the most relevant clusters performed using a multivariable multinomial logistic regression model, with dependent variable having several possible categories (cluster 1, cluster 2, … cluster n). Modelling analyses are exploratory conditioned by the size of the clusters and the number of candidate covariates, with at least 10-15 patients expected per covariate modality).

Documents

Study results

PASS VALSE_Abstract Final report-V1.0-v20230630-clean.pdf