Abstracts

Rationale:
Little is known about the disease burden, healthcare resource use (HCRU) and costs in refractory epilepsy (RE) patients treated with rescue medications (RMs). This study characterized and compared these outcomes among RE patients with RM claims and those without.
Method:
A retrospective cohort study was performed using IBM MarketScan administrative claims data from 01JAN2008 to 31DEC2018. RE was defined by two antiepileptic drugs (AEDs) taken consecutively for ≥3 months and the initiation of a third AED. RMs included rectal diazepam, intranasal midazolam (MDZ) and oral benzodiazepines (BZPs). Cases (RM+) were defined as individuals with a pharmacy claim for RM after RE was confirmed and were matched to controls without RM claims (RM-) by age, gender, type of epilepsy, and date of RE status. Index date was the date of first RM claim for RM+; the time between RE status and index date of RM+ was used to define a pseudo-index date for RM-. Outcomes included seizure-related injuries (SRIs: falls, burns, fractures, drowning, vehicle accidents and suicidality), status epilepticus (SE) episodes, HCRU (tests and procedures, hospitalizations, emergency room visits and outpatient visits) and healthcare costs, in the 6 and 12 months after index/pseudo-index date. Conditional multivariate analyses estimated incidence rate ratios (IRRs) and odds ratios (ORs), adjusted for year of index date, number of SE episodes and SRIs during the one-year baseline period and number of concomitant AEDs at index.
Results:
In total, 2306 RM+ were matched to 5155 RM-. Overall, 42% were male and median age was 45 years (IQR: 34). RM+ were more likely than RM- to receive 3+ concomitant AEDs at index (42% vs 33%, p< 0.0001) and to have experienced SE during the 12-month baseline period (11.0% vs 4.8%, p< 0.0001). 75.3%, 25.7% and 2.2% of RM+ received oral BZP, rectal diazepam and intranasal MDZ, respectively. During the 6-month post-index period, RM+ experienced a greater likelihood of SRI (adjusted OR=1.29 [1.05−1.59, 95%CI]), mainly driven by a greater proportion of burns (0.8% vs 0.4%, p=0.042) and fractures (6.6% vs 5.1%, p< 0.001). RM+ also had a greater likelihood and frequency of SE episodes (adjusted OR=1.59 [1.23−2.06]; adjusted IRR=2.39 [1.57−3.62]). HCRU was more likely (adjusted ORs ranging from 1.41 to 2.40, all p< 0.0001) and more frequent (adjusted IRRs ranging from 1.57 to 2.23, all p< 0.0001) in RM+ than RM-; general healthcare costs were 119% higher [100%−138%] in RM+ than in RM-. Most of the results remained statistically significant during 12-month follow-up.
Conclusion:
This study showed that RE patients with RM claims have a greater disease burden and higher HCRU and costs than those without RM. Further research is required to assess whether RE patients who need RM have a worse underlying disease or more comorbid conditions, and how RM use (e.g. low usage versus high usage) can improve patient outcomes in RE patients.
Funding:
:UCB Pharma-sponsored