Pharmacogenomic Testing to Inform Prescribing in Patients with Behavioral and Psychiatric Symptoms of Dementia (BPSD): Results from Two Small, Randomized, Controlled Trials

Introduction Behavioral and psychiatric symptoms of dementia (BPSD) occur frequently, representing a significant driver of the total costs of dementia in the US. Although antidepressants and atypical antipsychotics are often used in BPSD, they have limited effectiveness and significant toxicity. As a result, recent initiatives have focused on reducing polypharmacy and psychotropic prescribing in this population. Combinatorial pharmacogenomic testing has demonstrated utility in guiding prescribing for psychotropic medications in patients with depression by identifying medications unlikely to be safe or effective due to significant gene-drug interactions. Here we present data from two small, randomized, controlled trials (RCTs) designed to test the hypothesis that combinatorial pharmacogenomic testing could aid in treatment selection for BPSD. Methods The first RCT included residents being treated at the Bayleigh Chance retirement community in Maryland (inpatient RCT). The second RCT included patients from the University of Alabama at Birmingham Memory Disorders Clinic (outpatient RCT). Patients were eligible if they had a diagnosis of dementia with psychotic symptoms and/or behavioral disturbance and their condition was severe enough to trigger a consultation with a physician (inpatient RCT) or their physician was considering starting/changing a psychotropic medication (outpatient RCT). For the outpatient RCT, patients were also required to have a caregiver who spends at least 10 hours a week with them. Informed consent was obtained at the screening visit from patients’ legally authorized representatives. The inpatient RCT intended to enroll 50 patients and the outpatient RCT intended to enroll 100 patients. Both studies were stopped early due to slow enrollment. Patients were randomized 1:1 to treatment as usual (TAU) or the combinatorial pharmacogenomic-guided care arm. All patients received combinatorial pharmacogenomic testing (GeneSight, Assurex Health). Variants in multiple pharmacokinetic and pharmacodynamic genes were assessed and a weighted combinatorial algorithm categorized medications according to the level of predicted gene-drug interactions (GDI). For patients in the guided-care arm, physicians had access to the test report at the time of the baseline visit. Physicians were blinded to the test report until after the trial for patients in TAU. Assessments were performed at baseline, week 2 (AEs only, inpatient RCT only), week 4 (outpatient RCT only), week 8, and week 12. The primary outcome was the Neuropsychiatric Inventory (NPI), which assesses the presence and severity of BPSD across 12 domains. The nursing home version (NPI-NH) was used for the inpatient RCT and the NPI questionnaire (NPI-Q) was used for the outpatient RCT. A Mixed Model for Repeated Measures (MMRM) was utilized to evaluate NPI and included treatment, week, treatment-by-week interaction, baseline score, baseline score-by-week interaction as fixed effects. The presence of a condition (i.e. depression) was evaluated as a score >0 for the relevant domain on the NPI. Side effects were evaluated using the SA-EPS and MOSES scale (inpatient RCT) or the FIBSER scale (outpatient RCT). Changes in prescribing relative to a pre-test intended medication plan were evaluated in the outpatient RCT. Results A total of 12 patients were enrolled in the inpatient RCT (5 in TAU; 7 in guided-care) and 38 patients were enrolled in the outpatient RCT (19 in TAU; 19 in guided-care). At week 12, there were no significant differences in NPI or side effects between guided-care and TAU in either RCT (Table 1). However, the proportion of outpatients experiencing depression at week 12 was significantly lower in the guided-care arm versus TAU (p=0.0479). In the outpatient RCT, the proportion of patients prescribed at least one medication subject to GDI decreased from 58.8% (pre-test) to 25.0% (week 12) in the guided-care arm (Table 2). In contrast, there was an increase in the proportion of patients in TAU taking medications subject to GDI throughout the trial. A significantly higher proportion of patients in the guided-care arm had a reduction in the number of prescribed psychotropic medications by week 12 compared to TAU (Table 2). Conclusions Overall, there were no observed differences in overall neuropsychiatric symptoms or side effects among inpatients or outpatients who received pharmacogenomic-guided care compared to TAU. It should be noted that these studies were stopped early due to slow enrollment and were likely underpowered to detect any differences. However, there was a significant reduction in the proportion of outpatients with depression in the guided care arm, which is consistent with the validated use of combinatorial pharmacogenomic testing among patients with depression. There was also evidence that pharmacogenomic-guided care did inform prescribing, with reduced prescribing of medications subject to GDI and reduced psychotropic medication polypharmacy. Funding Myriad Neuroscience (Formerly Assurex Health)