Evaluation of comorbidities in epilepsy and their effects on outcome


  • Comorbidities of differing underlying mechanisms are present in approximately one-quarter of people with epilepsy and may influence outcomes through indirect mechanisms.

Why this matters

  • Approximately half of adults with epilepsy have a comorbid condition that may affect their quality of life, treatment response, or mortality rates. This study, including adults and children, provides evidence that comorbid conditions should be assessed at time of epilepsy diagnosis to potentially improve outcomes.

Study design

  • Objective: in a retrospective cohort study, to evaluate prevalence and types of comorbidities and to determine the association of comorbidities with long-term epilepsy outcomes.

  • Population: adults and children diagnosed with epilepsy before 2005 Dec 31 and followed for ≥10 years at 13 Italian epilepsy centers (n=486 females; n=520 males).

  • During follow-up, the following classifications were applied:

    • Seizure freedom: <2 years remission or 2+ years remission

    • Prognostic patterns: early remission; late remission; relapsing-remitting course; worsening course; no remission.

  • Comorbidities identified in medical records were defined as any clinical condition that occurred prior to epilepsy diagnosis and was recorded at diagnosis.

    • Each comorbidity was classified by 2 raters, by type and by organ/system categories, and also classified according to underlying mechanisms as:

      • Causal mechanism: preceding seizure onset and having a causal role in seizure development

      • Shared risk factors: common cause for epilepsy and comorbid condition

      • Chance association: occurs in patients with epilepsy without any causal association.

  • General characteristic comparisons were made between patients with ≥1 comorbidity and those without, for patients achieving 2-year remission and those never achieving remission, and according to 3 different prognostic categories (no remissions; relapsing-remitting and worsening course; and sustained remission).

  • Categorical and continuous variables were compared, and multivariable analysis was performed.

Key results

  • Of the total population, 26.4% had ≥1 comorbid condition, with total comorbidities per patient ranging from 1 to 7.

    • Organ/systems most commonly involved were: developmental/perinatal (21.3%), psychiatric (16.7%), cardiovascular (15.7%), endocrine/metabolic (10.3%), cerebrovascular (7.6%), and neurological (4.9%).

    • Mechanisms, in order of decreasing prevalence, were: chance association (42.2%); shared risk factors (31.1%); and causal association (26.7%).

  • Developmental/perinatal (7.5%), psychiatric (6.2%), cardiovascular (5.3%), and endocrine/metabolic (3.8%) comorbidities predominated in patients with ≥1 comorbidity.

  • In patients with <2-year remissions, endocrine/metabolic, psychiatric, and respiratory diseases were more common, and shared risk factors were more frequent.

    • Multivariable analysis confirmed a lower probability of achieving 2-year remission for psychiatric, endocrine/metabolic, and respiratory diseases.

    • Shared risk factors were not associated with remission.

  • In patients with no remission, when compared with relapsing-remitting and sustained remission, endocrine/metabolic, psychiatric, and respiratory comorbidities were more frequent.

    • Multivariable analysis confirmed a lower probability of relapsing-remitting or sustained remission for psychiatric and respiratory diseases, and a significant association for endocrine/metabolic for relapsing-remitting course only.

    • Shared risk factors were not associated with prognostic categories.

  • The authors concluded that at least 1 comorbidity was present in approximately one-quarter of patients with epilepsy, and developmental/perinatal, psychiatric, cardiovascular, and endocrine/metabolic comorbidities predominated; chance association was the most common mechanism.


  • The retrospective design may have impacted data quality and completeness.

  • The population base was lacking.

  • Interpretation of mechanisms underlying the comorbidities was subjective.