Concurrent Use of Prescription Opioids and Gabapentinoids in Older Adults

Abstract

Introduction:

Concurrent use of prescription opioids with gabapentinoids may pose risks of serious drug interactions. Yet, little is known about the trends in and patient characteristics associated with concurrent opioid–gabapentinoid use among older Medicare opioid users with chronic noncancer pain.

Methods:

A cross-sectional study was conducted among Medicare older beneficiaries (aged ≥65 years) with chronic noncancer pain who filled ≥1 opioid prescription within 3 months after a randomly selected chronic noncancer pain diagnosis (index date) in a calendar year between 2011 and 2018. Patient characteristics were measured in the 6-month baseline before the index date, and concurrent opioid–gabapentinoid use for ≥1 day was measured in the 3-month follow-up after the index date. Multivariable modified Poisson regression hwas used to assess the trends and characteristics of concurrent opioid–gabapentinoid use. Analyses were conducted from January to June 2021.

Results:

Among 464,721 eligible older beneficiaries with chronic noncancer pain and prescription opioids, the prevalence of concurrent opioid–gabapentinoid use increased from 17.0% in 2011 to 23.5% in 2018 (adjusted prevalence ratio=1.48, 95% CI=1.45, 1.53). Concurrent users versus opioid-only users tended to be non-Black, low-income subsidy recipients, and Southern residents. The clinical factors associated with concurrent opioid–gabapentinoid use included having a diagnosis of neuropathic pain, polypharmacy, and risk factors for opioid-related adverse events.

Conclusions:

Concurrent opioid–gabapentinoid use among older Medicare beneficiaries with chronic noncancer pain and prescription opioids has increased significantly between 2011 and 2018. Future studies are warranted to investigate the impact of concurrent use on outcomes in older patients. Interventions that reduce inappropriate concurrent use may target older patients with identified characteristics.

INTRODUCTION

The use of prescription gabapentinoids (i.e., gabapentin and pregabalin) among older adults aged ≥65 years has more than tripled from 2.6% in 2002 to 8.4% in 2015.¹ This growth was likely attributable to the 2009 American Geriatric Society guideline’s recommendation of gabapentinoids as adjuvant analgesics, with or without opioids, to manage chronic noncancer pain (CNCP).² Concurrent opioid–gabapentinoid use is particularly appealing to achieve the desired analgesia while reducing the consumption of opioids and their associated adverse events.²

However, recent growing concerns have arisen over the safety of concurrent opioid–gabapentinoid use. Such use potentiated the respiratory depressant effects of opioids in an animal study.³ Evidence from case reports and observational studies also linked the concurrent use with the risk for respiratory complications,⁴ opioid overdose,⁴ and opioid-related deaths.^5,6 These risks were postulated to arise from an additive depressant effect on the central nervous system (CNS) from both opioids and gabapentinoids.⁷ The accumulating evidence prompted the U.S. Food and Drug Administration to issue a warning concerning potential serious respiratory adverse events when prescribing gabapentinoids to patients who are taking opioids.⁸ The 2019 Beers criteria, an important guideline for prescribing safe medications to older adults, also cautioned against the co-use of opioids and gabapentinoids in this population, although most evidence was derived from nonelderly samples.⁷

Limited data exist to understand the extent to which gabapentinoids and opioids were concurrently used by older patients with CNCP. A recent study reported that 20.8% of opioid users received gabapentinoids concurrently among Medicare older adults who had Medicare Supplement plans, constituting only 28% of older adults enrolled in the Medicare fee-for-service (FFS) program.⁹ It remains unknown about the prevalence of concurrent opioid–gabapentinoid use among older Medicare FFS beneficiaries with CNCP, whether the prevalence has changed over time, and which patient subgroups are more likely to have the concurrent use. To fill these knowledge gaps, this study examines the trends in and the characteristics associated with concurrent opioid–gabapentinoid use among older adults enrolled in the Medicare FFS program who had active CNCP diagnoses and filled opioid prescriptions between 2011 and 2018.

METHODS

Study Population

A multiyear, cross-sectional study was conducted using a 5% national sample of Medicare FFS beneficiaries between 2011 and 2018. Medicare claims data contain billing records for services covered under Parts A (inpatient), B (office-based visits), and D (pharmacy dispensing) and information on sociodemographics. The study was approved by the University of Florida IRB and followed the STROBE reporting guidelines.

The study sample included older adults aged ≥65 years who had active CNCP (defined as having ≥2 inpatient or outpatient claims ≥30 days apart with a primary or secondary diagnosis of a CNCP condition; Appendix Table 1 (available online) provides the diagnostic codes¹⁰ in a calendar year between 2011 and 2018. Patients with CNCP were further required to have continuous enrollment in Medicare Parts A, B, and D without enrollment in Medicare Advantage plans during the 6 months before (baseline) and 3 months after (follow-up) the date of a CNCP diagnosis (i.e., index date) in a given year. The baseline period was used to measure patient characteristics, and the follow-up period was used to ascertain concurrent use (Appendix Figure 1, available online). Given its interest in patients with CNCP that warranted prescription opioid use, the study was restricted to patients filling opioid prescriptions within 3 months after a CNCP diagnosis. This follow-up period was chosen on the basis of clinical guidelines’ recommendation that patients’ pain treatment, especially opioid use, should be re-evaluated at least every 3 months.¹¹ The 3-month follow-up after a CNCP diagnosis also allowed for an increased likelihood that gabapentinoids were prescribed for the management of CNCP.

Patients were excluded if they (1) had a diagnosis of epilepsy or restless legs syndrome and nonpain indications approved by the Food and Drug Administration for gabapentinoids¹² or (2) received a cancer diagnosis; palliative care; or hospice service, which requires different opioid regimens,¹¹ during the baseline or follow-up periods. Also excluded were patients having Part A–covered inpatient or skilled nursing facility stay during the follow-up period because information on the prescriptions filled during such stays is incomplete. For patients with multiple eligible CNCP diagnoses in a year, 1 diagnosis was randomly selected to generate population-based (as opposed to diagnosis-based) prevalence of concurrent use. This random selection increases the representativeness of the selected diagnoses and has been used by claims-based studies to estimate the population-based prevalence of drug exposure.^13,14 Thus, a patient contributed only 1 CNCP diagnosis each year but may contribute to multiple years of observation if meeting the eligibility criteria in multiple years.

Measures

Use of opioids and gabapentinoids was measured on the basis of the filling date and days’ supply of respective prescriptions recorded in Medicare Part D event files. The medications of interest included gabapentinoids and opioid analgesics (Appendix Table 2, available online, provides the full list). Concurrent use was defined as having gabapentinoid and opioid days’ supply overlapping for ≥1 day during the follow-up period. Literature indicated that as short as a 1-day overlap between opioids and gabapentinoids could increase the risk of opioid-related adverse events,^4–6 and a 1-day overlap has been used previously to define concurrent use of opioids and other CNS depressants such as benzodiazepines.^15,16 Because there are no definitive cut off days for determining concurrent opioid–gabapentinoid use, sensitivity analyses were conducted using alternative concurrent use definitions (i.e., ≥7 or ≥30 cumulative days). To fully capture concurrent use at the follow-up, prescriptions of opioids and gabapentinoids filled at the baseline with days’ supply overlapping the follow-up period were also used.

Given that high-dose gabapentinoids and high-dose opioids could magnify the risk of opioid overdose,¹⁷ this study also examined the proportion of concurrent users who received high-dose gabapentinoids and those who received high-dose opioids within a concurrent episode. High-dose opioid use was defined as having ≥1 day with dosage exceeding 90 morphine milligram equivalents,¹¹ and high-dose gabapentinoid use was defined as having ≥1 day with dosage exceeding 1,800 mg/day for gabapentin and 600 mg/day for pregabalin.^5,6

Patient characteristics were measured during the 6-month baseline period. Sociodemographic characteristics included age (categorized as 65–74, 75–84, and ≥85 years), sex, race (White, Black, and others), receipt of low-income subsidy (LIS) (yes versus no), and census region (Northeast, Midwest, South, and West). Clinical characteristics included chronic pain types (musculoskeletal, neuropathic, or idiopathic pain),^2,11 comorbidities that may affect gabapentinoid treatment (chronic kidney disease,^18’19 chronic pulmonary disease,² fall, or fracture²⁰), polypharmacy (defined as the use of ≥5 generic drugs, excluding opioids and gabapentinoids), and several risk factors for opioid-related adverse events.²¹ These factors included a diagnosis of alcohol or tobacco use disorder, diagnosis of opioid use disorder or overdose²²; diagnoses of injection-related infections; diagnosis of depression²³; receipt of a urine drug test^24–26; use of long-acting opioids²⁷; use of chronic opioid therapy (defined as ≥90 consecutive days of opioid use)^28,29; use of high-dose opioids³⁰; concurrent use of opioids with other CNS-active drugs (including benzodiazepines, nonbenzodiazepine sedatives, tricyclic antidepressants, selective serotonin reuptake inhibitors, muscle relaxants, and antipsychotics)³¹; and the total number of opioid prescribers as a proxy for continuity of care, with a larger number indicating a higher degree of receiving fragmented continuity of care.³² To account for the number of comorbidities that are related to polypharmacy, this study also considered the number of Elixhauser comorbidities (excluding the conditions listed individually; categorized as 0–1, 2–3, 4–5, and ≥6). A detailed definition of each variable is provided in Appendix Table 1 (available online). Gabapentinoid use at baseline was not included as a covariate because the days’ supply of prescription gabapentinoids filled at baseline could overlap with days of follow-up and thus were used for defining the concurrent use outcome.

Statistical Analysis

To illustrate crude trends, the proportion of the eligible sample with concurrent use was plotted against the year of the CNCP diagnosis. The trends were also stratified by age, sex, race, region, and LIS status. Across the study years, to identify the characteristics associated with concurrent use versus those with opioid use only, a multivariable modified Poisson regression model was fitted with concurrent use (yes versus no) as the dependent variable.³³ To test the adjusted trend of concurrent use, each calendar year was added as a dummy variable in the model. The coefficients of these year indicators represented changes in the prevalence of concurrent use for a certain year compared with that of the reference year 2011, adjusting for patient characteristics. Generalized estimating equations were incorporated in all models to account for within-individual correlation because of the multiple years of observations that a patient may contribute.³³ Adjusted prevalence ratios (APRs) and 95% CIs were reported for trends and characteristics.

To further understand the patterns of concurrent use, several subgroup analyses were conducted among the concurrent users, in terms of (1) the type of gabapentinoid agents; (2) the top 3 opioid agents; (3) short-acting or long-acting opioids; (4) the use of high-dose gabapentinoids; (5) the use of high-dose opioids; and (6) whether gabapentinoids and opioids were prescribed by a single prescriber, within a concurrent episode. Across the years, a fold change calculated by the equation measure of 2018 – measure of 2011/measure of 2011 was reported. All analyses were performed between January and June 2021 using SAS, version 9.4. A 2-sided p<0.05 was considered statistically significant.

RESULTS

A total of 464,721 eligible Medicare FFS beneficiaries were identified (the flowchart is shown in Figure 1 Appendix Figure 1). The mean age was 75.3 (SD=8.1) years, 68.4% were female, and 81.4% were White. Musculoskeletal pain was the most common type of CNCP (94.4%), followed by neuropathic pain (38.6%). During the baseline period, 75.8% had polypharmacy, 19.8% were on chronic opioid therapy, and 24.7% used opioids with other CNS-active drugs concurrently (Table 1).

Figure 1.

Flow chart of sample selection.

CNCP, chronic noncancer pain; SNF, skilled nursing home.

Table 1.

Baseline Characteristics of the Sample and Their Associations With Concurrent Opioid–Gabapentinoid Use

Baseline characteristics	Overall samplea n (%)	Concurrent use versus opioid onlyb
		Crude prevalence ratio (95% CI)	Adjusted prevalence ratio (95% CI)
Year
2011	76,070 (16.4)	ref	ref
2012	67,146 (14.4)	0.98 (0.97, 1.00) *	1.03 (1.01, 1.05) **
2013	70,885 (15.3)	1.06 (1.04, 1.07) **	1.11 (1.09, 1.13) **
2014	60,245 (13.0)	1.09 (1.08, 1.11) **	1.16 (1.14, 1.18) **
2015	52,040 (11.2)	1.17 (1.15, 1.19) **	1.24 (1.22, 1.27) **
2016	58,267 (12.5)	1.23 (1.21, 1.25) **	1.34 (1.32, 1.37) **
2017	47,205 (10.2)	1.30 (1.28, 1.32) **	1.42 (1.39, 1.44) **
2018	32,863 (7.1)	1.38 (1.35, 1.40) **	1.48 (1.45, 1.51) **
Age, years, mean ± SD	75.3 ± 8.1
≥85	73,165 (15.7)	ref	ref
65–74	251,259 (54.1)	1.16 (1.14, 1.18) **	1.09 (1.07, 1.10) **
75–84	140,297 (30.2)	1.12 (1.10, 1.14) **	1.06 (1.05, 1.08) **
Female	317,974 (68.4)	1.11 (1.09, 1.13) **	1.07 (1.05, 1.08) **
Race
Black	39,824 (8.6)	ref	ref
White	378,486 (81.4)	0.95 (0.93, 0.97) **	1.11 (1.09, 1.13) **
Othersc	46,411 (10.0)	1.00 (0.97, 1.02)	1.07 (1.04, 1.10) **
Receiving Part D low-income subsidy	139,588 (30.0)	1.48 (1.46, 1.49) **	1.38 (1.36, 1.39) **
U.S. region
Northeast	75,440 (16.2)	ref	ref
Midwest	110,309 (23.7)	1.14 (1.11, 1.16) **	1.12 (1.10, 1.14) **
South	194,681 (41.9)	1.28 (1.25, 1.30) **	1.20 (1.18, 1.22) **
West	84,291 (18.1)	1.13 (1.11, 1.16) **	1.09 (1.07, 1.11) **
Chronic pain typed
Musculoskeletal	438,927 (94.4)	1.08 (1.05, 1.10) **	1.01 (0.99, 1.03) **
Neuropathic	179,361 (38.6)	2.56 (2.54, 2.59) **	1.56 (1.57, 1.60) **
Idiopathic	156,433 (33.7)	1.43 (1.42, 1.45) **	1.02 (1.01, 1.03) **
Comorbidity
Chronic kidney disease	83,355 (17.9)	1.40 (1.38, 1.41) **	1.04 (1.03, 1.05) **
Chronic pulmonary disease	83,221 (17.9)	1.39 (1.37, 1.40) **	1.04 (1.03, 1.05) **
Fracture/fall	61,231 (13.2)	1.19 (1.17, 1.20) **	0.98 (0.97, 0.99) **
Number of Elixhauser comorbiditiese
0–1	158,439 (34.1)	ref	ref
2–3	168,051 (36.2)	1.27 (1.25, 1.29) **	1.06 (1.04, 1.07) **
4–5	84,543 (18.2)	1.58 (1.56, 1.60) **	1.11 (1.09, 1.12) **
≥6	53,688 (11.6)	1.97 (1.94, 2.00) **	1.15 (1.13, 1.17) **
Polypharmacyf	352,209 (75.8)	2.09 (2.05, 2.13) **	1.26 (1.24, 1.28) **
Risk factors for opioid-related adverse events
Alcohol/tobacco use disorder	32,612 (7.0)	1.33 (1.31, 1.35) **	1.00 (0.99, 1.02)
Opioid use disorder or overdose	5,136 (1.1)	1.90 (1.86, 1.94) **	1.03 (1.01, 1.05) *
Injection-related infections	39,478 (8.5)	1.33 (1.32, 1.35) **	1.03 (1.02, 1.05) **
Depression	77,126 (16.6)	1.48 (1.46, 1.49) **	1.04 (1.03, 1.05) **
Urine drug test	16,270 (3.5)	1.83 (1.80, 1.86) **	1.05 (1.03, 1.06) **
Use of long-acting opioids	21,623 (4.7)	1.87 (1.84, 1.90) **	1.10 (1.08, 1.12) **
Use of chronic opioid therapy	91,856 (19.8)	1.98 (1.96, 2.00) **	1.20 (1.19, 1.21) **
High dose opioid use	33,975 (7.3)	1.80 (1.77, 1.82) **	1.07 (1.05, 1.08) **
Concurrent use of opioids with other CNS-active drugs	114,830 (24.7)	1.86 (1.84, 1.88) **	1.08 (1.07, 1.10) **
Total number of opioid prescribers
1	152,124 (32.7)	ref	ref
0g	234,644 (50.5)	0.87 (0.70, 1.09)	0.81 (0.80, 0.82) **
2	53,934 (11.6)	1.00 (0.91, 1.10)	1.03 (1.02, 1.04) **
≥3	24,019 (5.2)	1.04 (0.90, 1.21)	1.03 (1.02, 1.05) **

Note: Boldface indicates statistical significance (*p<0.05, **p<0.001).

^aOverall sample included all eligible Medicare beneficiaries with CNCP filling an opioid prescription (N=464,721 patients). A patient may contribute multiple eligible CNCP diagnoses across the study years after random selection; for these patients, we used their first CNCP diagnosis to describe their baseline characteristics.

^bAll randomly selected CNCP diagnoses were used for the multivariable regression model (Nn=1,025,328 CNCP diagnoses). Fewer than 2% of them had both gabapentinoids and opioids but did not meet the concurrent use criterion; they were grouped as opioid-only users.

^cOthers include Hispanic, Asian, the Natives of North America, and individuals with other or unknown races and ethnicities.

^dPatients could have multiple types of chronic pain.

^eThe following conditions were not counted because they were considered separately: arthritis, chronic pulmonary disease, renal failure, alcohol abuse, drug abuse, and depression. Cancer-related conditions (lymphoma, metastatic cancer, and solid tumor without metastasis) were also uncounted because patients with cancer were excluded.

^fPolypharmacy was defined as the use of ≥5 different generic medications, excluding gabapentinoids and opioids.

^gZero means having no opioid use at baseline.

CNCP, chronic noncancer pain; CNS, central nervous system.

Overall, 19.8% of the study sample had concurrent opioid–gabapentinoid use during their follow-up period between 2011 and 2018. The prevalence of concurrent use increased from 17.0% in 2011 to 23.5% in 2018 (Figure 2). Similar trends were seen in sensitivity analyses using ≥7 and ≥30 days to define concurrent use (Figure 2) and when stratified by sociodemographics (Appendix Figure 2, available online). Results from multivariable modified Poisson regression showed that adjusted prevalence of concurrent use has been increasing steadily since 2013 (APR=1.11, 95% CI=1.09, 1.13) and was 48% higher in 2018 (APR=1.48, 95% CI=1.45, 1.51) than in 2011 (Table 1).

Figure 2.

Crude prevalence of concurrent opioid–gabapentinoid use by year, 2011–2018. The overall prevalence of concurrent opioid–gabapentinoid use using different overlapping day definitions: 1-day threshold: 19.8%, 7-day threshold: 17.3%, and 30-day threshold: 11.2%.

CNCP, chronic noncancer pain.

Table 1 also presents the characteristics associated with concurrent use versus with opioid use only. Concurrent users were more likely to be non-Black (APR=1.11, 95% CI=1.09, 1.13 for White versus Black; APR=1.07, 95% CI=1.04, 1.10 for other race versus Black), recipients of LIS (APR=1.38, 95% CI=1.36, 1.39 versus non-LIS), and Southern residents (South APR=1.20, 95% CI=1.18, 1.22; Midwest APR=1.12, 95% CI=1.10, 1.14; West APR=1.09, 95% CI=1.07,1.11; versus Northeast). Of the examined clinical factors, having neuropathic pain (APR=1.59, 95% CI=1.57, 1.60 versus no such diagnosis) and polypharmacy (APR=1.26, 95% CI=1.24, 1.28 versus no polypharmacy) were the strongest characteristics of concurrent use. When analyzing the risk factors of opioid-related adverse events, significant associations were observed for most of them (except for alcohol/tobacco use disorder), with APR values ranging from 1.03 (95% CI=1.01, 1.05) for diagnosis of opioid use disorder or overdose to 1.20 (95% CI=1.19, 1.21) for use of chronic opioid therapy. Similar patterns were seen in sensitivity analyses (Appendix 3, available online).

Table 2 shows the patterns of concurrent opioid–gabapentinoid use. Most (83.0%–86.9%) of the concurrent users received gabapentin, whereas less than one fifth (15.1%–19.3%) received pregabalin. Although hydrocodone use decreased significantly from 54.3% in 2011 to 41.2% in 2018, potentially owing to the rescheduling of hydrocodone combination medications (e.g., hydrocodone and acetaminophen) from Schedule III to Schedule II in October 2014,³⁴ the drug remained the most commonly used opioid with gabapentinoids. Among concurrent users, the use of high-dose opioids decreased from 18.2% in 2011 to 13.6% in 2018, whereas the use of high-dose gabapentinoids increased from 17.0% in 2011 to 18.2% in 2018. The majority (63.2%–69.8%) of concurrent users received both medications from a single prescriber.

Table 2.

Patterns of Concurrent Opioid-Gabapentinoid Use Among Concurrent Users by Year, 2011–2018

Measures of interest	Overall	2011	2012	2013	2014	2015	2016	2017	2018	Fold change, %
Concurrent users, n	203,186	12,939	18,890	24,033	26,281	26,605	32,189	32,699	29,550
Type of gabapentinoid agents concurrently used with prescription opioids (%)a
Any gabapentin	85.5	83.0	84.2	84.4	84.6	85.5	86.2	86.8	86.9	5
Any pregabalin	16.8	19.3	18.0	18.2	17.9	17.0	16.1	15.1	15.1	−22
Type of opioid agents concurrently used with prescription gabapentinoids (%)a,b
Any hydrocodone	47.0	54.3	53.3	51.8	50.4	46.1	44.5	42.9	41.2	−24
Any tramadol	31.6	27.8	29.7	31.0	31.5	31.4	32.5	32.8	33.1	19
Any oxycodone	22.3	19.5	19.7	19.6	20.5	23.3	23.2	23.8	25.3	30
Duration of action of opioids concurrently used with prescription gabapentinoids (%)a
Any short acting	96.0	95.1	95.4	95.4	95.8	96.1	96.2	96.4	96.5	1
Any long acting	12.2	14.2	13.4	13.1	13.0	12.4	12.1	11.2	10.4	−26
High-dose opioids concurrently used with prescription gabapentinoids (%)c	15.8	18.2	17.3	16.8	16.6	16.4	15.2	14.6	13.6	−25
High-dose gabapentinoids concurrently used with prescription opioids (%)c	17.9	17.0	16.6	17.0	17.3	18.3	18.4	18.8	18.2	7
Receiving prescription opioids and gabapentinoids for ≥1 day from the same prescriber (%)	67.4	69.8	63.2	70.2	69.2	67.2	66.4	66.1	66.1	−2

^aA patient can contribute to multiple concurrent use diagnoses; thus, the sum of proportions exceeds 100%.

^bOnly the top 3 opioids were reported.

^cMeasured during concurrent opioid–gabapentinoid diagnoses.

DISCUSSION

This study is the first to examine the trends and characteristics associated with concurrent opioid–gabapentinoid use among older Medicare beneficiaries with CNCP between 2011 and 2018. Overall, 1 in 5 older opioid users had concurrent use within 3 months after a CNCP diagnosis. This estimate aligns with previous reports of 16.2% of opioid-related ambulatory visits involving coprescribing of gabapentinoids³⁵ and 20.8% of opioid users receiving gabapentinoids concurrently among a sample of Medicare beneficiaries with Supplement plans.⁹ A 48% increase was noted in the prevalence of concurrent use in 2018 compared with that in 2011, after adjusting for patient characteristics. Consistent results were seen in sensitivity analyses using cut offs of 7 and 30 overlapping days to define concurrent use. These findings indicate that concurrent opioid–gabapentinoid use has become increasingly common among older patients with CNCP who received opioid treatment.

The upward trend of concurrent use among older opioid users may reflect changes in clinicians’ prescribing practice of pain medications in recent years. The 2009 American Geriatric Society and 2016 Centers for Disease Control and Prevention guidelines for CNCP management have recommended combining opioids with non-opioid alternatives, including gabapentinoids, to minimize opioid requirements and the adverse effects from opioid use.^2,11 These recommendations were supported by multiple clinical trials that showed gabapentinoids’ efficacy in reducing pain and opioid dosing.^36–38 However, because these trials were conducted among patients with postherpetic neuralgia or diabetic neuropathy, for which gabapentinoid monotherapy was effective, whether such benefits can be extended to patients with other CNCP conditions remains debatable. Indeed, in a cohort of patients with chronic neck or low back pain, concurrent opioid–gabapentinoid use did not reduce pain or opioid consumption over a 2-year follow-up, compared with using opioids alone.³⁹ Furthermore, emerging evidence from nonelderly patients have linked concurrent use with opioid-related respiratory depression,⁴ overdose,¹⁷ and deaths.^5,6 Given the increasing concurrent opioid–gabapentinoid use observed in this older population, future studies that focus on older adults are warranted to understand the effect of this concurrent use on opioid utilization, pain control, and safety outcomes.

Concurrent use was more common among patients with (than among those without) the risk factors for opioid-related adverse events, including those with a diagnosis of opioid use disorder or overdose, and injection-related infections. The reasons for these associations were not clear, but this finding may be present if clinicians prescribed gabapentinoids as opioid substitutes to patients at risk of opioid misuse to reduce opioid dose and subsequently opioid-related adverse events. The wide use of gabapentinoids as opioid substitutes could be largely attributable to the belief that gabapentinoids are without abuse potential.⁴⁰ However, recent animal studies found that gabapentinoids could induce drug-seeking–like behaviors when used alone at high doses^41,42 and reinforce drug-rewarding effects when coadministered with morphine.⁴³ Heroin users also reported coingesting gabapentinoids to potentiate euphoria.³ Given the current evidence indicating gabapentinoids’ abuse potential, the observed associations between the risk factors for opioid-related adverse events and concurrent opioid–gabapentinoid use raised concerns about gabapentinoid misuse. Future studies are needed to understand whether the increased concurrent use among older Medicare patients with CNCP who are at higher risk for opioid-related adverse events was an artifact of gabapentinoids being prescribed as opioid substitutes or a sign for potential gabapentinoid misuse. If gabapentinoid misuse is common among these high-risk patients, strategies should be implemented to ensure their access to non-pharmacological therapy for CNCP management, naloxone for overdose prevention, and services for addiction treatment.⁴⁴

Several characteristics provide additional research and policy implications. The findings that older patients with LIS (versus those without) or residing in the South (versus those residing in other regions) had a higher prevalence of concurrent use are consistent with previous reports among mixed samples of young and older adults.^9,17,35 These results seem to support targeted interventions in the South or areas with lower SES to address high-risk opioid–gabapentinoid coprescribing.⁴⁵ In fact, several Southern states, including Kentucky, Tennessee, and Alabama, have recently rescheduled gabapentin as a controlled substance in light of accumulating evidence implicating gabapentinoids in opioid-related deaths. Future studies should evaluate the impact of this reschedule on concurrent opioid–gabapentinoid use. The association between having multiple opioid prescribers and increased likelihood of concurrent use echoes previous studies on opioid–benzodiazepine coprescribing.^16,45 Taken together, these observed findings highlight the importance of coordinated care, such as Prescription Drug Monitoring Programs, in reducing inappropriate opioid prescribing.^16,45 Finally, the observation that most concurrent users received both medications from a single prescriber suggested that educational interventions, such as letters to prescribers or guidelines discouraging coprescribing,^46,47 might be needed to address inappropriate concurrent use.

Notably, 1 in 6 (17.9%) concurrent users had gabapentinoid doses over the recommended threshold. Research has shown that high-dose gabapentinoids did not show appreciable pain relief^18,48 but an elevated risk of opioid-related death by up to 65%.^5,6 These results support practice measures to prevent gabapentinoid overuse among older opioid users. The practice measures could mimic those used to prevent high-dose opioid use, including restrictions on daily doses, previous authorizations, claim denials, and alert systems.⁴⁹

Limitations

This study has several limitations. First, the 1-day definition of concurrent use may misclassify patients who gradually switched from opioids to gabapentinoids as concurrent users of these 2 drugs. Yet, sensitivity analyses using 7- and 30-day definitions yielded similar results, suggesting the extent of concurrent use to be substantial. Second, Medicare Part D data captured only prescriptions filled and reimbursed by Medicare; thus, the prevalence of medication use may be underestimated. Third, the indication of prescription use is not available in the Part D data. To increase the likelihood that opioids and gabapentinoids were used for CNCP, this study excluded patients with nonpain indications of gabapentinoids and measured prescription use within 3 months after an eligible CNCP diagnosis. Finally, the study results can only be generalized to older Medicare opioid users enrolled in the FFS program who had no diagnosis of cancer, hospice, or palliative care.

CONCLUSIONS

In this sample of older Medicare opioid users with CNCP, concurrent opioid–gabapentinoid use was prevalent and increased significantly from 2011 to 2018. Concurrent use was more common among patients of non-Black race, receiving LIS, residing in the South, diagnosed with neuropathic pain, having polypharmacy, and at higher risk for opioid-related adverse events than their counterparts. Further investigations are warranted to understand the impact of concurrent opioid–gabapentinoid use on the health outcomes of older patients. Interventions to reduce inappropriate concurrent use may target those with identified characteristics.