Abstract
Background: Co-prescribing opioids and gabapentinoids (GABA, gabapentin and pregabalin) is associated with increased risk of falls, fractures, opioid overdose and deaths. The Centers for Disease Control and Prevention (CDC) in 2016 and the Food and Drug Administration (FDA) in 2019 recommended caution in such co-prescribing. A key step in updating policy and revising prescribing guidelines aimed at reducing opioid and GABA co-use and its associated consequences is a thorough understanding of the prescriber and the patient factors associated with co-use. We thus examined national trends and patterns in opioid and GABA co-prescribing among Medicare beneficiaries from 2017 to 2022. Methods: We conducted a retrospective study of Medicare beneficiaries with ≥90 consecutive days of opioid use from 2017 to 2022. The study outcome was GABA use during the 90-day opioid use episode. A multivariable logistic regression model was constructed to examine the patient, prescriber and prescription factors associated with receiving a GABA prescription. Results: Our sample comprised 8035 opioid-only and 2818 opioid and GABA users. Non-cancer (e.g., back and neuropathic) pain was a more common diagnosis in the opioid and GABA cohorts than in the opioid-only cohorts. The opioid-GABA co-prescribing rate did not substantially change (2017: 24.5%, 2019: 28.2% and 2022: 25%). Co-prescribing rates were higher in non-white patients, those on Medicaid and Medicare, and those whose initial Medicare entitlement was not based on age. Tramadol and hydrocodone were the most prescribed opioids. Approximately 33% of opioid and GABA users started with an initial daily GABA dose of ≥1200 mg. In the 12-month lookback period, patients on opioids and GABA had nearly 17 clinic visits to approximately 8 different providers. Factors associated with co-prescribing were seeing pain physicians (odds ratio = 1.29, 95% confidence interval-[CI] = 1.11–1.50), having more healthcare encounters (6–11 visits, odds ratio-[OR] = 1.19, 95% CI = 1.02–1.39; 12–19, OR = 1.20, 95% CI = 1.00–1.43; 20+, OR = 1.27, 95% CI = 1.03–1.57) and seeing >10 providers (OR = 1.40, 95% CI = 1.12–1.73). Conclusions: One in four Medicare beneficiaries with long-term opioid use received opioid and GABA prescriptions. Our findings of association in co-prescribing with multiple visits to different clinics/prescribers can inform the development of public health policy and practice guidelines (e.g., prescription-drug monitoring program checks within electronic medical records, EMR alerts with opioid and GABA co-prescribing) to potentially reduce opioid and GABA prescriptions and associated adverse outcomes.
Keywords: co-prescribing, opioids, gabapentinoids, medicare, trends, older adults, prescriptions, geriatrics, long-term opioid use, chronic pain
1. Introduction
Opioid-related morbidity and mortality is an increasing public health problem which has led to myriad government, insurance and health system changes to reduce opioid overprescribing [1,2,3]. Accompanying these changes is a shift by clinicians towards recommending non-opioid analgesic prescriptions as a presumably safer alternative to prescription opioids. This has led to a huge increase in prescribing gabapentinoids (GABA, gabapentin and pregabalin) in clinics, hospitals and post-acute care facilities [4,5,6,7]. Yet, instead of being opioid alternatives, GABA is increasingly being co-prescribed with opioids [5,6]. Opioid and GABA co-prescribing is associated with an increase in the risk of falls, fractures, respiratory depression, opioid overdose and deaths [8,9,10,11,12,13]. Using 2011–2016 fee-for-service Medicare data (n = 71,005 beneficiaries) for chronic pain, Zhou and colleagues (2021) found that the co-use of opioids and GABA was associated with twice the risk of drug overdose compared with opioid-only early discontinuers [10].
Because of the known toxicity of opioid and GABA co-use, both the Centers for Disease Control and Prevention (CDC) and the Food and Drug Administration (FDA) recommend caution in co-prescribing two such central nervous system (CNS) -depressant medications [3,14]. Unclear, however, is the extent to which opioid and GABA co-prescribing has changed since the 2016 CDC Guideline for Prescribing Opioids Policy and the 2019 FDA Drug Safety Communication about the dangers of the co-use of multiple CNS-depressant medications. A key step in updating policy and revising prescribing guidelines aimed at reducing opioid and GABA co-use and its associated consequences is a thorough understanding of the prescriber and patient factors associated with co-use. Gaps exist in the knowledge on the type, specialty and number of prescribers seen by patients; on the type, dose and duration of co-prescribed opioids and GABA; and on the patient behaviors and characteristics associated with opioid and GABA co-use. The current study addresses these gaps by using 2017 to 2022 national Medicare data to examine trends and patterns in prescribing opioid and GABA among long-term opioid users, particularly after the release of the CDC 2016 Opioid Policy Guideline and the 2019 FDA Drug Safety Communication. Because of the recent increase in the prescribing of GABA as an opioid alternative especially among long-term prescription opioid users, we decided to focus on the factors associated with prescribing GABA to long-term opioid users—a practice inconsistent with CDC and the FDA guidelines. Understanding the practice and policy-actionable drivers of the use of two or more CNS-depressing drugs can guide clinical practice guideline revision and inform public health policy aimed at reducing opioid-related morbidity and mortality.
2. Methods
2.1. Data Source
This retrospective cohort study used a 20% national sample of Medicare beneficiaries between 2017 and 2022. Medicare insurance covers hospital admission (Part A), provider visits and outpatient care (Part B) and prescription medications (Part D). Qualification is either by age (65 and older) via disability (<65). It does not pay for dental, vision or hearing services. The Medicare files used included the Master Beneficiary Summary Files (MBSF) and Claims files, including Medicare Provider and Analysis Review Files, Carrier Claims, Outpatient Standard Analytic Files and Part D Event Files (PDE) for all years. The MBSF file contained the demographic and enrollment information and the PDE contained the prescription information that was used to create the cohort. Medicare data were obtained from the United States of America (USA) Centers for Medicare and Medicaid Services and analyses were conducted at the Virtual Research Data Center. This study was approved on 23 July 2019 by the institutional review board (IRB # 16-0247) at the University of Texas Medical Branch.
2.2. Study Cohort
The study included all Medicare beneficiaries who had at least one period of 90 consecutive days of opioid use between 2017 and 2022. The calculation of 90 days allowed for no gaps between the end of one prescription period and the beginning of another and did not extend prescriptions that were filled prior to the end of a previous prescription. This conservative approach to calculating 90 consecutive days ensured that all beneficiaries included in the cohort had at least 90 days of filled prescriptions. Excluded from this group were those who had any opioid or GABA prescription in the one year prior to the 90-day period, those who did not have complete Medicare enrollment (parts A, B and D, with no Medicare Advantage plan) in the year prior to the 90-day period and those who did not live in the 50 states or District of Columbia at the time of their initial opioid prescription. Due to the nature of the lookback (12 months), it was possible for a beneficiary to have multiple periods of 90+ days of consecutive opioid use that met all inclusion/exclusion criteria (e.g., one period in 2017 and another in 2020). In such cases, the most recent 90-day period was used. Figure 1 shows the process of determining the study cohort. The specific opioid types included for analysis in this study were tramadol, buprenorphine, hydrocodone, hydromorphone, morphine, oxycodone and codeine.
Figure 1.
Study cohort selection.
2.3. Study Measures
The measured outcome of this study was GABA use during the 90-day opioid use episode, and only prescriptions with a fill date that occurred during that period were counted. This was measured as any prescription filled during the 90-day period. Because many of the measures used in this study applied to both opioid and GABA prescriptions, beneficiaries in this study were classified as opioid-only or opioid and GABA.
Other variables of interest included those describing the patient, prescription and prescriber characteristics. Patient characteristics included age, sex, race, year of 90-day episode, original Medicare entitlement reason (old age or disabled/end-stage renal disease [ESRD]), Medicaid eligibility, pain-related diagnoses in the year prior to the 90-day period (abdominal/chest, arthritis, back, cancer, chronic, fractures, headaches, musculoskeletal, nerve, other, visceral and wound) and diagnosis of epilepsy or post-herpetic neuralgia. Associated diagnoses were determined using the International Classification of Diseases (ICD-10) codes. Prescription characteristics included opioid dosage (morphine milligrams equivalent [MME] per day), GABA dosage per day, prescription days of supply, total GABA days (for the opioid plus GABA group) and total number of prescriptions (overall and for each drug). Prescriber characteristics included prescriber specialty type (primary care, pain, nurse practitioner [NP], physician’s assistant [PA], other non-physician or other physician), number of prescribers during the 90-day period, number of specialties consulted during the 90-day period and number of prescriptions from each provider specialty type. Also included were characteristics from the 12 months prior to the 90-day period. These characteristics included the number of healthcare visits, number of providers seen, number of specialties seen, number of visits to the original prescriber (first opioid prescriber for the 90-day period), the most visited provider specialty type and whether the patient saw each provider specialty type.
2.4. Statistical Analysis
Descriptive statistics for all variables were presented to compare those who used opioids only and those who used opioids and GABA. These comparisons included all of the patient, prescription and prescriber characteristics described under the study measures. A Cochran–Armitage trend test was used to assess the change in GABA usage between 2017 and 2022. To test the association between the selected variables and the outcome of the receipt of a GABA prescription, a logistic regression model was used to calculate the odds ratios (ORs) for GABA use. Variables included in the model were age, sex, race, original Medicare entitlement reason, dual Medicaid eligibility, specialty of the first opioid prescriber, length of first opioid prescription, specialty of the most-seen provider in the lookback, seeing the original opioid prescriber in the lookback, individual indicators for each specialty seen in the lookback, number of visits in the lookback, number of providers seen in the lookback and number of visits in the 7 days before opioid initiation. All analyses were conducted using SAS Enterprise Guide 7.1 (SAS Inc., Cary, NC, USA).
3. Results
During the study period, the opioid and GABA co-prescribing rate remained similar across the years: 24.5% in 2017, 27.02% in 2018, 28.2% in 2019, 25.9% in 2020, 25.5% in 2021 and 25% in 2022 (Cochran–Armitage trend test p = 0.9225). The demographics of patients receiving opioids only versus opioid and GABA prescriptions were relatively comparable by age, race and sex, with a majority of patients identifying as white and female (Table 1). Key pain diagnoses in the year prior to opioid use more common among those with opioid-only prescription utilization were back pain (55.6%), arthritis (40.6%) and chronic pain (37.2%). These primary diagnoses differed among patients co-prescribed opioid and GABA, with 69.9% of this group having back pain, 51.6% having chronic pain and 44.6% having arthritis. Of note, the co-use of opioid and GABA was more common in those with diagnoses of back pain, chronic pain and nerve pain, while opioid-only use was more common among persons with diagnoses of cancer and visceral pain.
Table 1.
Beneficiary characteristics by opioid/GABA use group.
| Opioid Only | Opioid + GABA | |||||
|---|---|---|---|---|---|---|
| N | % | N | % | p-Value | ||
| Total | 8035 | 100% | 2818 | 100% | ||
| Age | Mean, SD | 70.42 | 14.45 | 67.03 | 14.75 | |
| <65 | 2218 | 27.6% | 1078 | 38.3% | <0.0001 | |
| 65–69 | 1305 | 16.2% | 431 | 15.3% | ||
| 70–74 | 1283 | 16.0% | 426 | 15.1% | ||
| 75–79 | 1025 | 12.8% | 316 | 11.2% | ||
| 80–84 | 845 | 10.5% | 234 | 8.3% | ||
| 85+ | 1359 | 16.9% | 333 | 11.8% | ||
| Sex | Male | 3338 | 41.5% | 1180 | 41.9% | 0.7595 |
| Female | 4697 | 58.5% | 1638 | 58.1% | ||
| Race | White | 6660 | 82.9% | 2293 | 81.4% | 0.0752 |
| Black | 661 | 8.2% | 248 | 8.8% | ||
| Other | 210 | 2.6% | 77 | 2.7% | ||
| Asian | 114 | 1.4% | 54 | 1.9% | ||
| Hispanic | 390 | 4.9% | 146 | 5.2% | ||
| Original Entitlement | Old Age | 4622 | 57.5% | 1348 | 47.8% | <0.0001 |
| Disabled/ESRD | 3413 | 42.5% | 1470 | 52.2% | ||
| Dual Medicaid Eligible | No | 4633 | 57.6% | 1373 | 48.7% | <0.0001 |
| Yes | 3402 | 42.4% | 1445 | 51.3% | ||
| Pain Locations | Abdominal/Chest Pain | 2748 | 34.2% | 1101 | 39.1% | <0.0001 |
| Arthritis | 3263 | 40.6% | 1256 | 44.6% | 0.0002 | |
| Back Pain | 4464 | 55.6% | 1971 | 69.9% | <0.0001 | |
| Cancer | 1509 | 18.8% | 470 | 16.7% | 0.0129 | |
| Chronic Pain | 2989 | 37.2% | 1455 | 51.6% | <0.0001 | |
| Fractures | 1148 | 14.3% | 413 | 14.7% | 0.6317 | |
| Headaches | 1365 | 17.0% | 606 | 21.5% | <0.0001 | |
| Musculoskeletal | 1696 | 21.1% | 746 | 26.5% | <0.0001 | |
| Nerve Pain | 1278 | 15.9% | 918 | 32.6% | <0.0001 | |
| Other Pain | 1595 | 19.9% | 591 | 21.0% | 0.2014 | |
| Visceral Pain | 832 | 10.4% | 279 | 9.9% | 0.4939 | |
| Wound Pain | 643 | 8.0% | 232 | 8.2% | 0.6992 | |
| Other Diagnoses | Epilepsy | 312 | 3.9% | 108 | 3.8% | 0.9048 |
| Post-herpetic neuralgia | 6 | 0.1% | 1 | 0.0% | 0.4808 |
GABA: gabapentinoids; SD: standard deviation; ESRD: end-stage renal disease.
For both the opioid-only and opioid and GABA cohorts, the initial opioid prescription comprised a low daily MME, with over 80% in each group receiving a first opioid prescription of <50 MME/day (Table 2). When comparing this to the MME/day average across all opioid prescriptions for the two cohorts, the percentage of prescriptions of <50 MME/day continued to remain the predominant dose type (71.82% of opioid-only prescriptions and 66.9% of opioid + GABA prescriptions). Strikingly, daily dosing of the initial GABA prescription in the opioid and GABA cohort was more variable, with 36.3% of these first-time prescriptions containing a daily GABA dose of <600 mg, 30.5% containing a daily GABA dose of 600–1199 mg and 33.1% containing a daily GABA dose of ≥1200 mg.
Table 2.
Opioid and GABA dosage information, by opioid/GABA use group.
| Opioid Only | Opioid + GABA | ||||
|---|---|---|---|---|---|
| N | % | N | % | ||
| First Opioid Prescription MME/day | <50 | 6528 | 81.2% | 2269 | 80.5% |
| 50-<90 | 414 | 5.2% | 174 | 6.2% | |
| 90+ | 1093 | 13.6% | 375 | 13.3% | |
| All Opioid Prescriptions MME/Day—average across all prescriptions | <50 | 5770 | 71.8% | 1884 | 66.9% |
| 50-<90 | 829 | 10.3% | 372 | 13.2% | |
| 90+ | 1436 | 17.9% | 562 | 19.9% | |
| First GABA Prescription * Dose/Day | <600 | 1024 | 36.3% | ||
| 600-<1200 | 860 | 30.5% | |||
| 1200+ | 934 | 33.1% | |||
| All GABA Prescriptions * Dose/Day—average across all prescriptions | <600 | 983 | 34.9% | ||
| 600-<1200 | 837 | 29.7% | |||
| 1200+ | 990 | 35.1% | |||
GABA: gabapentinoids; MME: morphine milligrams equivalent. * GABA dose is in milligram (mg).
Across a 90-day opioid period, a 30-day supply period was the most frequent duration given for an initial prescription for both the opioid-only (41.1%) and opioid and GABA (47.4%) medications (Table 3). Of note, nearly 25% of first-time opioid-only prescriptions were allocated for >30 days, referred to as “long opioid prescriptions,” compared to approximately 15% of initial opioid and GABA prescriptions. In this group of long opioid prescriptions, the median prescription supply length was 90 days, and tramadol hydrochloride constituted the majority (65.7%) of these prescriptions (Supplemental Table S1). Among those with opioid and GABA prescriptions, 62.8% of the first GABA prescriptions were for 30 days, 24.5% were >30 days and 12.7% were for <30 days.
Table 3.
Prescription and prescriber characteristics in the 90-day long-term opioid period, by opioid/GABA use group.
| Opioid Only | Opioid + GABA | ||||
|---|---|---|---|---|---|
| N | % | N | % | ||
| Total | 8035 | 100% | 2818 | 100% | |
| First Prescription Prescriber Type | Unknown | 59 | 0.7% | 23 | 0.8% |
| Primary Care | 4098 | 51.0% | 1245 | 44.2% | |
| Pain | 718 | 8.9% | 399 | 14.2% | |
| NP | 1185 | 14.7% | 398 | 14.1% | |
| PA | 470 | 5.8% | 200 | 7.1% | |
| Other Non-Physician | 10 | 0.1% | 3 | 0.1% | |
| Other Physician | 1495 | 18.6% | 550 | 19.5% | |
| First Prescription Supply Days | <7 | 1339 | 16.7% | 556 | 19.7% |
| 8–29 | 1405 | 17.5% | 502 | 17.8% | |
| 30 | 3301 | 41.1% | 1335 | 47.4% | |
| 31+ | 1990 | 24.8% | 425 | 15.1% | |
| All Prescriptions Prescriber Type | Unknown | 313 | 0.9% | 152 | 0.7% |
| Primary Care | 15,728 | 45.4% | 9144 | 43.9% | |
| Pain | 3342 | 9.6% | 2438 | 11.7% | |
| NP | 6279 | 18.1% | 3477 | 16.7% | |
| PA | 2201 | 6.4% | 1547 | 7.4% | |
| Other Non-Physician | 50 | 0.1% | 42 | 0.2% | |
| Other Physician | 6733 | 19.4% | 4013 | 19.3% | |
| All Prescriptions Supply Days | <7 | 4795 | 13.8% | 2262 | 10.9% |
| 8–29 | 8512 | 24.6% | 4430 | 21.3% | |
| 30 | 18,360 | 53.0% | 12,437 | 59.8% | |
| 31+ | 2979 | 8.6% | 1684 | 8.1% | |
| Number of Prescribers in 90-Day Period | 1 | 5665 | 70.5% | 1130 | 40.1% |
| 2 | 1662 | 20.7% | 991 | 35.2% | |
| 3 | 451 | 5.6% | 435 | 15.4% | |
| 4 | 160 | 2.0% | 167 | 5.9% | |
| 5 | 61 | 0.8% | 66 | 2.3% | |
| 6 | 25 | 0.3% | 13 | 0.5% | |
| 7 | 9 | 0.1% | 9 | 0.3% | |
| 8 | 0 | 0.0% | 5 | 0.2% | |
| 9 | 1 | 0.0% | 2 | 0.1% | |
| 10 | 1 | 0.0% | 0 | 0.0% | |
| Number of Specialties in 90-Day Period | Unknown | 40 | 0.5% | 4 | 0.1% |
| 1 | 6288 | 78.3% | 1460 | 51.8% | |
| 2 | 1460 | 18.2% | 1074 | 38.1% | |
| 3 | 223 | 2.8% | 247 | 8.8% | |
| 4 | 21 | 0.3% | 32 | 1.1% | |
| 5 | 3 | 0.0% | 1 | 0.0% | |
| GABA Prescription from the Same Prescriber as Original Opioid Prescription | - | - | 1482 | 52.6% | |
| GABA Prescription from the Same Prescriber Specialty as Original Opioid Prescription | - | - | 1795 | 63.7% | |
| GABA Prescription on the Same Day as Original Opioid Prescription | - | - | 129 | 4.6% | |
| Number of Prescriptions | Mean, SD | 4.29 | 2.85 | 7.34 | 3.56 |
| RX by NP | Mean, SD | 0.78 | 1.94 | 1.22 | 2.61 |
| RX by Primary Care | Mean, SD | 1.95 | 2.52 | 3.22 | 3.82 |
| RX by PA | Mean, SD | 0.27 | 1.11 | 0.54 | 1.78 |
| RX by Pain | Mean, SD | 0.41 | 1.32 | 0.86 | 2.07 |
| RX by Other Non-Physician | Mean, SD | 0.01 | 0.16 | 0.01 | 0.26 |
| RX by Other Physician | Mean, SD | 0.83 | 2.02 | 1.42 | 2.64 |
| <7 Day Prescriptions | Mean, SD | 0.59 | 1.87 | 0.79 | 2.22 |
| 8–29 Day Prescriptions | Mean, SD | 1.05 | 2.02 | 1.56 | 2.71 |
| 30 Day Prescriptions | Mean, SD | 2.28 | 2.04 | 4.39 | 2.83 |
| 31+ Day Prescriptions | Mean, SD | 0.37 | 0.6 | 0.59 | 0.93 |
| Number of Opioid Prescriptions | Mean, SD | 4.29 | 2.85 | 4.83 | 2.81 |
| Number of GABA Prescriptions | Mean, SD | 0 | 0 | 2.51 | 1.63 |
| GABA Days | Mean, SD | 0 | 0 | 59.71 | 24.73 |
GABA: gabapentinoids; NP: nurse practitioner; PA: physician assistant; RX: prescription.
In general, primary care physicians were the most common prescriber of opioid-only (45.4%) and opioid and GABA (43.9%) prescriptions, followed by NPs (18.1% for opioid-only and 16.7% for opioid plus GABA) (Table 3). Compared to other provider groups, primary care physicians also represented a larger number of first prescribers of opioid-only and opioid and GABA prescriptions. While primary care physicians prescribed the most initial and refill prescriptions of opioid-only and opioid and GABA medications, slight increases were seen in the percentages of NPs prescribing refills across both cohorts (Supplemental Table S2).
Additionally, differences were observed in the number of prescribers for opioid-only versus opioid and GABA prescriptions in a 90-day period. Across the cohort of patients receiving opioids only, 70.5% had a single provider and 20.7% had two providers prescribing their opioid medications in a 90-day period (Table 3). Approximately 5.6% of these patients received their opioid prescriptions from three providers over 90 days. Comparatively, approximately 40.1% of patients receiving opioids and GABA had one prescriber, while 35.2% had two prescribers and 15.4% had three providers within a 90-day period. When specifically examining the opioid and GABA cohort, 52.6% received their GABA prescription from the same prescriber of their original opioid prescription. Moreover, 4.6% of patients using opioids and GABA received their GABA prescription on the same day as their original opioid prescription, the majority of which came from a primary care physician (Table 3, Supplemental Table S3). The most common prescriber combinations for the first opioid and GABA prescriptions were primary care physicians and pain physicians (Supplemental Table S4).
Over a 12-month lookback period of general visits prior to their first opioid prescription, patients who were co-prescribed opioids and GABA had, on average, nearly seventeen visits, seeing approximately eight different providers (Table 4). On the other hand, patients receiving only opioids had slightly fewer visits (fourteen) and provider types (nearly seven) during this period. Among these average numbers of visits, approximately three were to the patient’s original prescriber of opioid or opioid and GABA identified within the 90-day period. Moreover, an estimated 35.3% of patients prescribed opioids only and 37.7% of patients co-prescribed opioids and GABA did not see their original opioid prescriber in the 12 months prior to receiving their first opioid prescription. For these patients, their original opioid prescriber was most commonly a primary care physician (Supplemental Table S5).
Table 4.
Provider visit information for the 12-month lookback period, by opioid/GABA use group.
| Opioid Only | Opioid + GABA | ||||
|---|---|---|---|---|---|
| N | % | N | % | ||
| Total | 8035 | 100% | 2818 | 100% | |
| Number of Visits 12 Prior Months | Mean, SD | 14.19 | 14.31 | 16.54 | 16.55 |
| Had 0 visits in 12 Prior Months | 373 | 4.6% | 97 | 3.4% | |
| Number of Providers 12 Prior Months | Mean, SD | 6.85 | 6.92 | 8.07 | 7.82 |
| Number of Visits to Original Prescriber (of 90-day period) | Mean, SD | 3.09 | 4.71 | 2.97 | 4.83 |
| Number of Visits to Most Visited Provider | Mean, SD | 5.64 | 5.92 | 6.08 | 6.01 |
| Number of Specialties 12 Prior Months | 0 | 373 | 4.6% | 97 | 3.4% |
| 1 | 1085 | 13.5% | 339 | 12.0% | |
| 2 | 2605 | 32.4% | 798 | 28.3% | |
| 3 | 2442 | 30.4% | 869 | 30.8% | |
| 4 | 1296 | 16.1% | 577 | 20.5% | |
| 5 | 229 | 2.9% | 136 | 4.8% | |
| 6 | 5 | 0.1% | 2 | 0.1% | |
| Most Visited Provider Type | Unknown | 373 | 4.6% | 97 | 3.4% |
| Primary Care | 3342 | 41.6% | 1127 | 40.0% | |
| Pain | 311 | 3.9% | 129 | 4.6% | |
| NP | 1043 | 13.0% | 410 | 14.5% | |
| PA | 370 | 4.6% | 154 | 5.5% | |
| Other Non-Physician | 20 | 0.2% | 10 | 0.4% | |
| Other Physician | 2576 | 32.1% | 891 | 31.6% | |
| Any Visit to Provider Type in Prior 12 Months | Primary Care | 6197 | 77.1% | 2225 | 79.0% |
| Pain | 1012 | 12.6% | 522 | 18.5% | |
| NP | 3611 | 44.9% | 1404 | 49.8% | |
| PA | 2416 | 30.1% | 986 | 35.0% | |
| Other Non-Physician | 101 | 1.3% | 43 | 1.5% | |
| Other Physician | 7016 | 87.3% | 2459 | 87.3% | |
| Had At Least 1 Visit to Original Prescriber (of the 90-day period) | Yes | 5196 | 64.7% | 1757 | 62.3% |
| No | 2839 | 35.3% | 1061 | 37.7% | |
GABA: gabapentinoids; NP: nurse practitioner; PA: physician assistant; SD: standard deviation.
The initial prescriber has an impact on GABA prescriptions in a cohort of long-term (at least 90 days) opioid users (Table 5). In comparison to primary care physicians as the first prescriber, patients seeing pain specialists (OR 1.37, 95% confidence interval [CI] = 1.16–1.62) had higher odds of receiving GABA prescriptions. In the 12-month lookback period, patients also had higher odds of receiving GABA prescriptions if they saw a pain physician (OR 1.29, 95% CI = 1.11–1.50), had more health encounters (6–11 encounters, OR 1.19, 95% CI = 1.02–1.39; 12–19 encounters, OR 1.20, 95% CI = 1.00–1.43; and 20+ encounters, OR 1.27, 95% CI = 1.03–1.57) or saw more than 10 providers (OR 1.40, 95% CI = 1.12–1.73).
Table 5.
Multivariable analysis of the odds ratios of gabapentinoid use among long-term opioid users.
| Variable | Level | Adjusted OR (95% CI) |
|---|---|---|
| Age | <65 | Ref |
| 65–69 | 0.74 (0.63, 0.87) | |
| 70–74 | 0.77 (0.64, 0.91) | |
| 75–79 | 0.69 (0.57, 0.83) | |
| 80–84 | 0.61 (0.49, 0.75) | |
| 85+ | 0.53 (0.43, 0.64) | |
| Sex | Female | Ref |
| Male | 0.93 (0.85, 1.02) | |
| Race | White | Ref |
| Black | 0.92 (0.78, 1.08) | |
| Asian | 1.49 (1.06, 2.09) | |
| Hispanic | 1.02 (0.83, 1.25) | |
| Other | 0.96 (0.74, 1.26) | |
| Original Medicare Entitlement | Disabled/ESRD | Ref |
| Old Age | 1.06 (0.92, 1.22) | |
| Dual Medicaid Eligibility | No | Ref |
| Yes | 1.21 (1.09, 1.34) | |
| First Prescriber | Primary Care | Ref |
| Unknown | 1.04 (0.63, 1.71) | |
| Pain | 1.37 (1.16, 1.62) | |
| NP | 0.92 (0.79, 1.06) | |
| PA | 1.16 (0.96, 1.41) | |
| Other Non-Physician | 0.73 (0.19, 2.77) | |
| Other Physician | 1.10 (0.97, 1.24) | |
| First Prescription Days | <7 | Ref |
| 8–29 | 0.86 (0.75, 1.00) | |
| 30 | 0.99 (0.87, 1.11) | |
| 31+ | 0.66 (0.57, 0.77) | |
| Most Seen Provider in Prior 12 Months | Primary Care | Ref |
| Unknown | 0.89 (0.67, 1.18) | |
| Pain | 0.72 (0.56, 0.94) | |
| NP | 1.10 (0.94, 1.29) | |
| PA | 1.11 (0.88, 1.39) | |
| Other Non-Physician | 1.36 (0.56, 3.33) | |
| Other Physician | 1.01 (0.90, 1.14) | |
| Saw Original Prescriber in the 12 Months Prior | No | Ref |
| Yes | 0.94 (0.85, 1.04) | |
| Specialties Seen in Prior 12 Months (Yes/No) | Primary Care | 1.06 (0.92, 1.22) |
| Pain | 1.29 (1.11, 1.50) | |
| NP | 0.99 (0.88, 1.10) | |
| PA | 1.03 (0.92, 1.14) | |
| Other Non-Physician | 0.92 (0.61, 1.40) | |
| Other Physician | 0.85 (0.73, 1.00) | |
| Number of Visits in Prior 12 Months | ≤5 | Ref |
| 6–11 | 1.19 (1.02, 1.39) | |
| 12–19 | 1.20 (1.00, 1.43) | |
| 20+ | 1.27 (1.03, 1.57) | |
| Number of Providers Seen in Prior 12 Months | ≤3 | Ref |
| 4–5 | 1.05 (0.89, 1.23) | |
| 6–9 | 1.07 (0.90, 1.28) | |
| 10+ | 1.40 (1.12, 1.73) | |
| Number of Visits in Prior 7 Days | 0 | Ref |
| 1 | 0.98 (0.88, 1.09) | |
| 2+ | 0.98 (0.84, 1.15) |
GABA: gabapentinoids; NP: nurse practitioner; PA: physician assistant; OR: odds ratio; CI: confidence interval.
4. Discussion
Our findings can be summarized as follows: the number of patients with any opioid prescription (alone or in combination with GABA) declined from 2017 to 2022, starting soon after the 2016 CDC Guideline for Prescribing Opioids for Chronic Pain and the 2019 FDA drug safety warnings [3,14]. Tramadol was the most prescribed opioid among long-term opioid users, consistent with other research findings showing increasing trends in tramadol prescribing (in part due to various opioid-restricting policies) and a concomitant rise in tramadol-related toxicities (e.g., death, hypoglycemia and hospitalization) [15,16,17].
However, the opioid and GABA co-prescribing rate remained essentially unchanged from 2017 to 2022 (24.5% in 2017 and 25% in 2022), despite the 2016 CDC opioid prescribing guideline and the 2019 FDA drug safety warnings [3,14]. Co-prescribing rates were higher in non-white patients, those on Medicaid and Medicare (dually eligible) and those whose initial Medicare entitlement was based on disability or ESRD, a population at high risk of opioid overdose death [18]. One in four Medicare beneficiaries on long-term opioids received GABA, despite the FDA/CDC warnings and the increasing evidence the high risk of opioid-related overdose, falls and fractures in patients on both drugs [3,11,14,19]. This finding underscores the need for a public health policy update and clinical guideline reform.
Opioid-only users had higher rates of cancer diagnoses than opioid and GABA users. The opioid and GABA users, however, had higher rates of back pain, chronic pain and nerve pain diagnoses. The high opioid and GABA co-prescribing rate for non-cancer pain underscores the need for multidisciplinary (pain, mental health and primary care providers) and multimodal approaches to pain management, including the use of pharmacological (opioids and non-opioids), non-pharmacological (e.g., physical therapy and nerve stimulation) and local injections [20].
Consistent with the CDC guidelines for a low opioid starting dose, over 80% of both cohorts started with a low daily MME dose (<50 MME/day) as their initial opioid prescription. On the other hand, approximately 33% of the opioid and GABA user cohort started with a high initial daily GABA dose of 1200 mg or higher, a dose associated with falls/fractures, delirium and hospitalization in older adults [21,22]. Although prevalence was low at 4.0%, prescribing GABA and opioids on the same day of a visit carries a particularly high risk of drug overdose and deaths, and such practice should be avoided [11,19].
Primary care physicians were the most common prescribers of opioid and GABA (43.9%), followed by NPs (16.7%), with modest increases over time in the percentage of NPs and PAs prescribing opioid and GABA refills. In the 12-month lookback period, patients on opioids and GABA had nearly seventeen clinic visits to approximately eight different providers, with approximately three of the visits to the patient’s original prescriber of opioid or opioid and GABA, thus increasing the fragmentation and discontinuity of care which serve as major contributors to polypharmacy, medical errors, potentially inappropriate medication, misdiagnosis and death [23,24]. Approximately 20.5% of the opioid-GABA co-users visited four different specialties in twelve months. Such multiple visits to different sites, providers and specialties may have been due to myriad reasons, including but not limited to healthcare system fragmentation with frequent changes in insurance or in providers, challenges in accessing optimal pain treatments leading to undertreated pain, socioeconomic barriers to accessing continuity of care, complexity of pain management needs and a lack of coordination among providers. Given these circumstances and the need to reduce the rates of GABA-opioid co-prescribing (and the associated adverse outcomes), future research should examine avenues to effectively identify controlled substance utilization, such as routine use of prescription drug monitoring program (PDMP) checks by prescribers. These checks may show the number of prescribers from whom a patient has received GABA-opioids as well as other controlled substances, helping obtain a more comprehensive picture of a patient’s prescription history and risk status [3,25]. One approach to facilitate and encourage PDMP checking by clinicians is by integrating the PDMP into the electronic health record (EHR), though additional research is needed to identify more streamlined methods for accessing PDMP data across different states and health systems to promote alignment in care outcomes [26].
5. Limitations
This study had limitations. First, we used data from Fee-for-Service Medicare beneficiaries, and our results may not be generalizable to Medicare Advantage or Health Maintenance Organization populations. The calculation of a 90-day consecutive period of opioid use without gaps between prescription periods as defined in the study criteria may also not have fully accounted for patients who experienced delays in prescription re-fills. Additionally, we could not ascertain whether the patients actually took the prescribed opioids and/or GABA, as we analyzed only PDE files. We also had no data on non-prescribed drugs (e.g., drugs received from friends, the street or mail orders) that patients might have obtained outside of Medicare. A limitation of using claims data is the inability to confirm whether patients with opioid medications are actually using them or not, such that starting a new GABA medication prescription during the last month of an opioid prescription might actually be a switch instead of concurrent use. This limitation could have led to the possible misclassification of “less than 30 days” of GABA/OPIOID overlap as concurrency when it could have been switches. Because Medicare claims data are constrained in addressing this misclassification, one way to address this is a qualitative study approach or a patient survey to confirm actual use, discontinuation or concurrency. Our study did not address dynamic changes in sites, sources and severity of pain over time—an important consideration in explaining why additional GABA is prescribed, and this is an area for future study. The NP prescribing data might have been incomplete, as physicians may have included some NP effort as incidental billing, and this may likely have under-reported NP-specific data in the analysis. While our study focused on factors associated with the prescribing of GABA to long-term opioid users—a practice inconsistent with CDC and the FDA guidelines—future studies are needed to also examine the reverse: new opioid prescribing for long-term GABA users. Another area for future study is an examination of whether the decision by clinicians to add or not add GABA to long-term opioid (LTOP) users is influenced by whether the LTOP user is on other CNS-active medications such as benzodiazepine and Z drugs.
6. Conclusions
The co-prescribing of opioids and GABA in the Medicare population—a risk factor for falls, fractures and opioid overdose—is a common and potentially modifiable public health problem. One in four Medicare beneficiaries on long-term opioids were also on GABA medications. The predictors of being on combined opioid and GABA prescriptions include having multiple visits to different clinics, providers and specialties; having both Medicare and Medicaid (dual eligibility); and having qualified for Medicare via disability or ESRD. These findings suggest the need for further research assessing the efficacy of clinical practice interventions (e.g., integrating the PDMP within electronic health record (EHR) systems and incorporating EHR alerts when opioid and GABA are co-prescribed) in helping to reduce co-prescribing two CNS depressant drugs (opioid and GABA) among Medicare patients and decreasing associated morbidity. Designing and implementing targeted, patient-centered educational materials focused on the risks of opioid-GABA co-use for Medicare patients and caregivers have the potential to engage patients and their care partners in the optimal and safe management of their painful conditions.
Supplementary Materials
The following supporting information can be downloaded at: https://www.mdpi.com/article/10.3390/medsci14030345/s1, Table S1: Long Opioid Prescriptions (31+ Days), Table S2: First Rx vs Refill Rx, Table S3: Prescriber Type of those who had Opioid and GABA on the same day (%), Table S4: First Opioid Prescriber and First GABA Prescriber, Table S5: Characteristics of those who did not see their original prescriber in the 12 months prior to the first opioid prescription.
Author Contributions
Concept and design: all authors; acquisition, analysis or interpretation of data: J.W., M.R. and Y.-F.K.; drafting of the manuscript: M.R., A.S., J.W., S.R. and Y.-F.K.; critical review of the manuscript for important intellectual content: M.R., A.S., J.W., S.R. and Y.-F.K.; statistical analysis: J.W., and Y.-F.K.; obtained funding: M.R. and Y.-F.K.; administrative, technical or material support: M.R. and Y.-F.K.; supervision: M.R. and Y.-F.K.; data-sharing statement: the Medicare data used in this study were obtained from the Centers for Medicare and Medicaid (CMS), and the analyses were conducted at the CMS Virtual Research Data Center. CMS should be contacted for use of its data under a data use agreement (DUA). All authors have read and agreed to the published version of the manuscript.
Institutional Review Board Statement
The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Review Board of the University of Texas Medical Branch, Galveston, Texas, (IRB#:16-0247; date of approval: 23 July 2019).
Informed Consent Statement
Not applicable. Informed consent was waived because the research used deidentified data obtained from the Centers for Medicare and Medicaid Services, with analyses conducted at the Virtual Research Data Center.
Data Availability Statement
The data used in this study are not publicly available and require a Data Use Agreement (DUA) from the Centers for Medicare & Medicaid Services.
Conflicts of Interest
The authors declare no conflicts of interest.
Funding Statement
This study was supported by grant funding from the National Institute on Drug Abuse (grant R01DA039192). The funder had no role in the design and conduct of the study; collection, management, analysis, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication. Previous presentations of the information reported in the manuscript: none; human ethics and consent to participate declarations: not applicable.
Footnotes
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
References
- 1.American Medical Association Overdose Epidemic Report 2024: Physicians’ Actions to Help End the Nation’s Drug-Related Overdose and Death Epidemic—And What Still Needs to be Done. Publication 24-117708. 2024. [(accessed on 28 May 2025)]. Available online: https://www.ama-assn.org/system/files/ama-overdose-epidemic-report.pdf.
- 2.Centers for Disease Control and Prevention 2022 CDC Clinical Practice Guideline at a Glance. 7 May 2024. [(accessed on 28 May 2025)]; Available online: https://www.cdc.gov/overdose-prevention/hcp/clinical-guidance/index.html.
- 3.Dowell D., Haegerich T.M., Chou R. CDC Guideline for Prescribing Opioids for Chronic Pain—United States, 2016. MMWR Morb. Mortal. Wkly. Rep. 2016;65:1–49. doi: 10.15585/mmwr.rr6501e1. Erratum in MMWR Morb. Mortal. Wkly. Rep. 2016, 65, 295. [DOI] [PubMed] [Google Scholar]
- 4.Polychronopoulou E., Kuo Y.F., Wilkes D., Raji M.A. Prescribing of gabapentinoids with or without opioids after burn injury in the US, 2012-2018. Burns. 2022;48:293–302. doi: 10.1016/j.burns.2021.12.006. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Chen C., Lo-Ciganic W.H., Winterstein A.G., Tighe P., Wei Y.J. Concurrent use of prescription opioids and gabapentinoids in older adults. Am. J. Prev. Med. 2022;62:519–528. doi: 10.1016/j.amepre.2021.08.024. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Bongiovanni T., Gan S., Finlayson E., Ross J.S., Harrison J.D., Boscardin W.J., Steinman M.A. Trends in the use of gabapentinoids and opioids in the postoperative period among older adults. JAMA Netw. Open. 2023;6:e2318626. doi: 10.1001/jamanetworkopen.2023.18626. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Goodman C.W., Brett A.S. Gabapentin and pregabalin for pain—Is increased prescribing a cause for concern? N. Engl. J. Med. 2017;377:411–414. doi: 10.1056/NEJMp1704633. [DOI] [PubMed] [Google Scholar]
- 8.Gomes T., Juurlink D.N., Antoniou T., Mamdani M.M., Paterson J.M., van den Brink W. Gabapentin, opioids, and the risk of opioid-related death: A population-based nested case-control study. PLoS Med. 2017;14:e1002396. doi: 10.1371/journal.pmed.1002396. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Peckham A.M., Fairman K.A., Sclar D.A. All-cause and drug-related medical events associated with overuse of gabapentin and/or opioid medications: A retrospective cohort analysis of a commercially insured US population. Drug Saf. 2018;41:213–228. doi: 10.1007/s40264-017-0595-1. [DOI] [PubMed] [Google Scholar]
- 10.Zhou L., Bhattacharjee S., Kwoh C.K., Tighe P.J., Reisfield G.M., Malone D.C., Slack M., Wilson D.L., Chang C.Y., Lo-Ciganic W.H. Dual-trajectories of opioid and gabapentinoid use and risk of subsequent drug overdose among Medicare beneficiaries in the United States: A retrospective cohort study. Addiction. 2021;116:819–830. doi: 10.1111/add.15189. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Olopoenia A., Camelo-Castillo W., Qato D.M., Adekoya A., Palumbo F., Sera L., Simoni-Wastila L. Adverse outcomes associated with concurrent gabapentin, opioid, and benzodiazepine utilization: A nested case-control study. Lancet Reg. Health–Am. 2022;13:100302. doi: 10.1016/j.lana.2022.100302. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Mattson C.L., Chowdhury F., Gilson T.P. Notes from the field: Trends in gabapentin detection and involvement in drug overdose deaths—23 states and the District of Columbia, 2019–2020. MMWR Morb. Mortal. Wkly. Rep. 2022;71:664–666. doi: 10.15585/mmwr.mm7119a3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Chen C., Winterstein A.G., Lo-Ciganic W.H., Tighe P.J., Wei Y.J. Concurrent use of prescription gabapentinoids with opioids and risk for fall-related injury among older US Medicare beneficiaries with chronic noncancer pain: A population-based cohort study. PLoS Med. 2022;19:e1003921. doi: 10.1371/journal.pmed.1003921. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.US Food and Drug Administration (FDA) FDA Warns About Serious Breathing Problems with Seizure and Nerve Pain Medicines Gabapentin (Neurontin, Gralise, Horizant) and Pregabalin (Lyrica, Lyrica CR): When Used with CNS Depressants or in Patients with Lung Problems; FDA Drug Safety Communication. 19 December 2019. [(accessed on 28 May 2025)]; Available online: https://www.fda.gov/drugs/drug-safety-and-availability/fda-warns-about-serious-breathing-problems-seizure-and-nerve-pain-medicines-gabapentin-neurontin.
- 15.Rodin R., Li L., McKendrick K., Harrison K., Hunt L.J., Muench U., Smith C.B., Aldridge M.D., Morrison R.S. The 2016 CDC Opioid Guideline and analgesic prescribing patterns in older adults with cancer. JAMA Netw. Open. 2025;8:e259043. doi: 10.1001/jamanetworkopen.2025.9043. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Zeng C., Dubreuil M., LaRochelle M.R., Lu N., Wei J., Choi H.K., Lei G., Zhang Y. Association of tramadol with all-cause mortality among patients with osteoarthritis. JAMA. 2019;321:969–982. doi: 10.1001/jama.2019.1347. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Fournier J., Azoulay L., Yin H., Montastruc J., Suissa S. Tramadol use and the risk of hospitalization for hypoglycemia in patients with noncancer pain. JAMA Intern. Med. 2015;175:186–193. doi: 10.1001/jamainternmed.2014.6512. [DOI] [PubMed] [Google Scholar]
- 18.Kuo Y.F., Raji M.A., Goodwin J.S. Association of disability with mortality from opioid overdose among US Medicare adults. JAMA Netw. Open. 2019;2:e1915638. doi: 10.1001/jamanetworkopen.2019.15638. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Humpert S.R., Reveles K.R., Bhakta K., Torrez S.B., Evoy K.E. Association of gabapentinoids with opioid-related overdose in the inpatient setting: A single center retrospective case-control study. Hosp. Pharm. 2024;59:188–197. doi: 10.1177/00185787231206522. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Pritchard K.T., Baillargeon J., Lee W., Raji M.A., Kuo Y. Trends in the use of opioids vs nonpharmacologic treatments in adults with pain, 2011-2019. JAMA Netw. Open. 2022;5:e2240612. doi: 10.1001/jamanetworkopen.2022.40612. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Fleet J.L., Dixon S.N., Kuwornu P.J., Dev V.K., Montero-Odasso M., Burneo J., Garg A.X. Gabapentin dose and the 30-day risk of altered mental status in older adults: A retrospective population-based study. PLoS ONE. 2018;13:e0193134. doi: 10.1371/journal.pone.0193134. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Muanda F.T., Weir M.A., Ahmadi F., Sontrop J.M., Cowan A., Fleet J.L., Blake P.G., Garg A.X. Higher-dose gabapentinoids and the risk of adverse events in older adults with CKD: A population-based cohort study. Am. J. Kidney Dis. 2022;80:98–107.e1. doi: 10.1053/j.ajkd.2021.11.007. [DOI] [PubMed] [Google Scholar]
- 23.Prior A., Vestergaard C.H., Vedsted P., Smith S.M., Virgilsen L.F., Rasmussen L.A., Fenger-Grøn M. Healthcare fragmentation, multimorbidity, potentially inappropriate medication, and mortality: A Danish nationwide cohort study. BMC Med. 2023;21:305. doi: 10.1186/s12916-023-03021-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Kern L.M., Safford M.M., Slavin M.J., Makovkina E., Fudl A., Carrillo J.E., Abramson E.L. Patients’ and providers’ views on causes and consequences of healthcare fragmentation in the ambulatory setting: A qualitative study. J. Gen. Intern. Med. 2019;34:899–907. doi: 10.1007/s11606-019-04859-1. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Centers for Disease Control and Prevention (CDC) Prescription Drug Monitoring Programs (PDMPs): What Clinicians Need to Know. 6 May 2024. [(accessed on 28 May 2025)]; Available online: https://www.cdc.gov/overdose-prevention/hcp/clinical-guidance/prescription-drug-monitoring-programs.html#:~:text=Check%20the%20PDMP:,frequently%20when%20continuing%20opioid%20therapy.
- 26.Holmgren A.J., Apathy N.C. Evaluation of Prescription Drug Monitoring Program integration with hospital electronic health records by US county-Level opioid prescribing rates. JAMA Netw. Open. 2020;3:e209085. doi: 10.1001/jamanetworkopen.2020.9085. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The data used in this study are not publicly available and require a Data Use Agreement (DUA) from the Centers for Medicare & Medicaid Services.