CYP2D6-Guided Opioid Prescribing Falls Short After Surgery

Key Takeaways

• The 10-day Silverman integrated analgesic assessment was not different between guided care and usual care.
• Phenotype-concordant prescribing was more common with genotype guidance, with less hydrocodone prescribing and more hydromorphone prescribing.
• Overall opioid use and other pain-related outcomes were similar, no study-related adverse events were reported, and the authors concluded the data did not support a role for CYP2D6-guided opioid therapy in this multimodal setting.

The comparison was detailed in a JAMA Network Open trial report. The actionable subgroup was defined by genotype together with CYP2D6 inhibitor use and informed recommendations to avoid tramadol, hydrocodone, and codeine after surgery. The prespecified 10-day primary outcome did not improve.

Recruitment ran from March 2021 to September 2023, follow-up concluded in March 2024, and assessments occurred at 10 days, 1 month, 3 months, and 6 months. The study enrolled and randomized 1602 participants overall; 351 who had an actionable CYP2D6 poor- or intermediate-metabolizer phenotype and completed surgery formed the primary analytical cohort. That cohort included 176 assigned to guided care and 175 assigned to usual pain management, with total knee arthroplasty most common, followed by total hip arthroplasty.

Guided care used preoperative CYP2D6 genotyping and recommendations to avoid tramadol, hydrocodone, and codeine in poor and intermediate metabolizers, while prescriber autonomy was preserved. Genotype results were available before surgery for 165 of 176 guided participants, making the workflow largely feasible before surgery.

The primary outcome was the 10-day Silverman integrated analgesic assessment score, which combined pain intensity with opioid exposure. At 10 days, mean scores were 1.4 with guided care and -1.4 with usual care, for a difference of 2.8 with 95% CI -18.3 to 23.8 and P=.80. Overall opioid use was also similar at 13.7 versus 13.2 morphine milligram equivalents per day, with P=.91. Numeric pain intensity at 10 days was 5.2 versus 5.1, and the individual pain and opioid-use measures were similar between groups at 10 days.

Phenotype-concordant prescribing was higher with guidance at 64% versus 27%, a 37-percentage-point difference with 95% CI 27 to 46 and P<.001. Hydrocodone was prescribed less often in the guided arm, while hydromorphone was prescribed more often. Use of nerve blocks and nonopioid analgesics was common in both arms, and no clear benefit emerged for pain intensity, composite pain intensity, mobility, overall well-being, or pain trends. Per-protocol and subgroup analyses, including nonoxycodone, nonactionable, arthroplasty-only, genotype-awareness, and oxycodone interaction analyses, also did not show a clear benefit, and no study-related adverse events were reported.

In this contemporary multimodal postoperative setting, the authors concluded the data did not support a role for CYP2D6-guided opioid therapy.