Polypharmacy and pharmacogenomics in high-acuity behavioral health care for autism spectrum disorder: a retrospective study

Abstract

Background: This study evaluated pharmacogenomic (PGx) testing in children and adolescents with autism spectrum disorder (ASD). ASD frequently presents with co-occurring depression and anxiety. This complex phenotype often results in psychotropic medication polypharmacy. Incorporating PGx testing into the medical work-up may reduce polypharmacy and improve quality of life with symptom reduction.

Methods: A retrospective electronic health record (EHR) review between January 2017 and May 2023. Individuals either received PGx testing or treatment as usual (TAU). The co-primary outcomes were instance of polypharmacy and the Pediatric Quality of Life Enjoyment and Satisfaction Questionnaire (PQ-LES-Q). Secondary outcomes included length of stay, average number of psychotropic medications, readmissions and assessments measuring severity of symptoms or behavioral impact. When at least one daily psychotropic medication was prescribed and reported to have an increased probability of gene-drug interactions, the individual was classified as "incongruent" (PGx-I). Individuals were categorized as "congruent" (PGx-C) if all prescribed psychotropic medications were without potential gene-drug interactions. Polypharmacy was evaluated and compared within the PGx-C and PGx-I subgroups.

Results: A total of 99 individuals with ASD were analyzed. At the time of admission, 93% of individuals were prescribed at least one psychotropic medication and over half of these individuals were prescribed medications with potential gene-drug interactions. Following PGx testing, there was an overall reduction in prescribed medications with potential gene-drug interactions. No differences were observed between the PGx and TAU groups in polypharmacy, quality of life, or symptom assessments of depression, anxiety, obsessive-compulsive disorder and body-focused repetitive behaviors. Subanalysis comparing congruent ("use as directed") or incongruent ("use with caution"), as well as exploratory analysis of only CYP2D6 and CYP2C19 gene-drug interactions, were observed to have a similar profile between treatment groups for all primary and secondary outcomes, except for the average number of psychotropic medications prescribed.

Conclusions: Incorporating PGx testing into the medical workup did not improve outcomes, with all treatment groups achieving similar levels of polypharmacy and quality of life. Analysis of secondary outcomes revealed some differences in medication prescribing when stratifying by congruency; however, no differences were observed between treatment groups for all other secondary outcomes.

Keywords: ASD; Antidepressants; Antipsychotics; Anxiety; Autism; CYP2D6; Depression; GeneSight; PGx; Pharmacogenomics; Polypharmacy.

Fig. 1

Polypharmacy rate between the PGx-C and PGx-I cohorts. a When evaluating all psychotropic medications at admission, the polypharmacy rate was similar between PGx-C and PGx-I. The PGx-C polypharmacy rate was 55.0% (11/20) and 53.8% for PGx-I (14/26) (P = 0.938). At discharge, the polypharmacy rate was also similar between PGx-C and PGx-I. The PGx-C polypharmacy rate was 65.0% (13/20) and 80.8% for PGx-I (21/26) (P = 0.227). The polypharmacy rate did not differ between admission and discharge for PGx-C (P = 0.519), but it did increase for PGx-I (P = 0.039). b When evaluating antidepressants at admission, the polypharmacy rate was similar between PGx-C and PGx-I. The PGx-C polypharmacy rate was 0.0% (0/20) and 11.5% for PGx-I (3/26) (P = 0.245). At discharge, the polypharmacy rate was also similar between PGx-C and PGx-I. The PGx-C polypharmacy rate was 25.0% (5/20) and 30.8% for PGx-I (8/26) (P = 0.667). The polypharmacy rate increased between admission and discharge for PGx-C (P = 0.047), but did not differ for PGx-I, with only a trend (P = 0.090). *P < 0.05; n.s. = not significant. Data reported as mean ± SEM

Fig. 2

The average number of medications was similar between treatment groups. a Within the PGx group, the average number of medications increased between admission (2.7 ± 0.2) and discharge (3.2 ± 0.2) (P = 0.010). b At admission, the average number of medications did not differ between PGx-C (2.5 ± 0.3) and PGx-I (2.8 ± 0.2) (P = 0.450). At discharge the average number of medications did not differ between PGx-C (3.0 ± 0.2) and PGx-I (3.6 ± 0.3), although there was a trend (P = 0.115). The average number of medications did not differ between admission and discharge for PGx-C (P = 0.180), but they did increase for PGx-I (P = 0.027). n.s. = not significant. Data reported as mean ± SEM

Fig. 3

Quality of life, anxiety and depression improved between admission and discharge for both the TAU and PGx groups. a PQ-LES-Q scores at admission were similar between the TAU and PGx groups (P = 0.664). Quality of life for both groups improved to reach similar levels at the time of discharge (P = 0.805), with both groups reporting average life satisfaction. b LSAS-CA scores at admission were also similar between the TAU and PGx groups (P > 0.999). Social anxiety for both groups improved to similar levels at the time of discharge (P = 0.339), with both groups reporting mild levels of social anxiety. c PROMIS-D scores at admission did not differ between the TAU and PGx groups (P = 0.671). Depressive symptom severity for both groups improved to reach similar levels at the time of discharge (P = 0.223), with both groups reporting none to slight levels of depressive symptoms. d PQ-LES-Q scores at admission were similar between PGx-C and PGx-I (P = 0.059), with quality of life improving for both subgroups at discharge (P = 0.188) and reporting average life satisfaction. e LSAS-CA scores at admission were similar between PGx-C and PGx-I (P = 0.309), with both subgroups reporting improvements in social anxiety to mild levels by discharge (P = 0.483). f PROMIS-D scores at admission did not differ significantly between PGx-C and PGx-I (P = 0.079), with depressive symptom severity improving for both subgroups to none to slight levels by discharge (P = 0.295). *P < 0.05; **P < 0.01; ***P < 0.001; n.s. = not significant. Data reported as mean ± SEM