Abstracts

Phenotypes of Comorbidity in Veterans with Epilepsy: A Foundation for Understanding Risk for Adverse Outcomes

Abstract number : 3.379
Submission category : 16. Epidemiology
Year : 2021
Submission ID : 1826125
Source : www.aesnet.org
Presentation date : 12/6/2021 12:00:00 PM
Published date : Nov 22, 2021, 06:52 AM

Authors :
Mary Jo Pugh, PhD, RN - University of Utah/VA Salt Lake City; Hamada Altalib - Yale University; Jacob Kean - University of Utah School of Medicine; Christine Baca, MD, MPH - Virginia Commonwealth University; Eamonn Kennedy - University of Utah; Megan Amuan - VA Salt Lake City Health Care System; Anne Van Cott - Neurology - VA Pittsburgh Health Care System; Alan Towne - Virginia Commonwealth University; Sidney Hinds - Wounded Warrior Foundation; W. Curt LaFrance - Brown University; Amy Henion - University of Utah School of Medicine; Samin Panahi - University of Utah; Chen-Pin Wang - UT Health San Antonio

Rationale: The common comorbidities of epilepsy can be more burdensome to patients with epilepsy than primary symptoms and seizures. However, little is known about how the phenotypic patterns of comorbidity in epilepsy can emerge and diverge over time.

Methods: We identified epilepsy and its index date using our previously validated algorithm in a cohort of Post-9/11 Era Veterans in Veterans Health Administration (VHA)/ Department of Defense care (FY00-FY19). We restricted analyses to those with data available for a 5-year ‘trajectory’ period centered on the index date (-2, index [0], +2). We identified diagnoses for 26 specific conditions previously associated with epilepsy each year: traumatic brain injury (TBI), post-concussion symptoms (e.g., headache, pain, sleep disorders), chronic disease (e.g., cerebrovascular disease (CVD), cardiac disease, hypertension), and mental health (e.g., depression, bipolar disorder, substance use disorder (SUD), and post-traumatic stress disorder [PTSD]). We used latent class analysis (LCA) to group Veterans based on the similarity of their comorbidity patterns over the 5-year period. We modeled demographic characteristics in the phenotypes and examined mortality and cause of death after the trajectory period.

Results: 15,716 Veterans met inclusion criteria (mean age at index 33.5; SD 9.9). LCA identified 6 phenotypes defined by distinct patterns of comorbidity. Figure 1 shows the change in probabilities (risk experienced by each group) for each condition two years before and two years after epilepsy index.

Relatively Healthy (RH): low probabilities of all comorbidities except for pain and headache, which were like the overall post-9/11 population; brain tumor was highest in this group.

Deterioration: moderate probabilities of mental health conditions that increased over time.

Notable high probabilities are highlighted below for the other groups:
Chronic Disease (CD): hypertension, CVD, cardiac disease, diabetes
PTSD: PTSD, TBI, headache
Bipolar/SUD: bipolar, SUD, anxiety, personality disorder
Polytrauma: TBI, PTSD, pain, headache, obstructive sleep apnea, lung disease

The Deterioration group was significantly older than all other groups and was more likely to be female (Table 1). Suicidality and overdose were highest in the index year and were significantly more likely in Bipolar/SUD and Polytrauma groups. PTSD and Polytrauma had the highest service-connected disability. All-cause mortality was significantly higher for chronic disease and bipolar/SUD groups (p< .01). Cause of death also varied by phenotype. Accidents and suicide were significantly higher in PTSD, Bipolar/SUD and Deterioration; cancer was significantly higher in the Healthy group; cardiovascular disease was significantly higher in Chronic Disease (p< .01).
Epidemiology