(202) Association Between Rapid Opioid Dose Tapering and Polydrug Overdose Death Among High-dose, Long-term Opioid Therapy Recipients in North Carolina Between 2006-2018

Background
Chronic pain patients who receive high-dose, long-term opioid therapy (HDLTOT) and subsequently are exposed to rapid dose tapering may be at high risk of opioid withdrawal and illicit drug use, which can lead to opioid and polydrug overdose.

Objectives
To quantify the association between rapid tapering/discontinuation of HDLTOT and polydrug overdose death (POD) over 4 years, accounting for time-varying biases.

Methods
We constructed a retrospective cohort of 21,478 commercially insured North Carolina residents aged 18-64 who received HDLTOT (≥90 morphine milligram equivalents for 90% of 90 consecutive days) before entering the cohort during 01/2006-09/2018. The time-varying binary exposure was assessed monthly as rapid opioid tapering (dose reduction >10% per week) or discontinuation, versus dose maintenance, increase, or gradual tapering. POD outcomes were defined as due to the combined use of opioids and benzodiazepines, methamphetamine, and/or cocaine. We estimated the intention-to-treat effect of rapid tapering by risk differences (RD) and risk ratios (RR) during follow-up, using inverse probability of treatment and censoring (IPTC)-weighted Aalen-Johansen (AJ) estimator and Fine-Gray (FG) estimator, to account for time-varying biases and competing risk due to deaths of other cause(s). Bootstrapping with 1000 iterations was used for statistical inference of estimates.

Results
About half of the total person-time (255,901 person-months, 49%) was exposed to rapid opioid dose tapering during 4-year follow-up. Sixty-seven polydrug overdose deaths (POD) occurred due to the combined use of opioid(s) with benzodiazepines, methamphetamine, and/or cocaine, with a median time-to-event of 1.9 years. Compared to no rapid tapering, rapid tapering did not affect POD risk until the end of the follow-up. At year four, both AJ and FG estimates suggested increased risk of POD associated with rapid opioid tapering, with better precision in FG estimates (AJ-RD[95%CI]: 0.0022[-0.0006 - 0.0051], FG-RD[95%CI]: 0.0029[0 - 0.0059]; AJ-RR[95%CI]: 1.59[0.835 - 3.02], FG-RR[95%CI]: 1.68 [0.962 - 2.93]). When restricted to HDLTOT recipients of stable opioid dosing at baseline (N=12,609), rapid opioid tapering was not found to be associated with POD risk during follow-up.

Conclusion
Increased risk of polydrug overdose death 4 years after rapid opioid tapering compared to no rapid tapering suggests a long-term safety concern in tapering among chronic pain patients who received HDLTOT. These findings warrant future examination of this association in the general population under the subsequent polydrug overdose epidemic (“fourth wave”).