Changes in Prescribed Opioid Dosages Among Patients Receiving Medical Cannabis for Chronic Pain, New York State, 2017-2019

Key Points

Abstract

Introduction

The US has been experiencing an opioid crisis with multiple waves, starting with increased prescribing of opioids in the 1990s.¹ Recently, opioid-involved overdose deaths in 2021 were the highest in US history, with more than 80 000 deaths.² Opioid prescriptions decreased 47% by 2020, after a peak in 2012 with 255 million prescriptions.^3,4 Despite significant reductions, many people are continuing to use opioids for chronic pain. In 2019, 1 in 5 adults had chronic pain⁵ and 22.1% of these had recently used prescription opioids for relief.⁶

National guidelines for prescribing opioids for chronic pain⁷ and states’ opioid regulations have resulted in discontinuation and reduced initiation of opioid prescriptions.⁸ Nonopioid medication options are limited for many individuals with pain; for example, nonsteroidal anti-inflammatory drugs (NSAIDs) are contraindicated for many patients experiencing chronic pain, particularly among older patients, who often experience a higher burden of pain.⁹ Without humane tapering of opioids and effective alternatives for patients receiving long-term opioid therapy (LOT), many of them are at high risk of overdose (through turning to the illicit market) and of suicide.^8,10

Cannabis has gained popularity as a potential pain management modality.^11,12,13 However, its role in reducing the need for prescription opioids or the dosage of opioids remains controversial due to limited research.¹⁴ Most studies relied on survey data or small samples of individuals who used cannabis in combination with opioids for pain management.¹⁵ Results showed individuals self-reported reduction in opioid use and improvement in pain management. A meta-analysis on the association of cannabinoids with reduced opioid use for analgesia found some support in observational studies but little evidence from clinical trials.¹⁶ Several trials used cannabis-derived products for a short time, which may not reflect how patients actually use cannabis as an alternative therapy.¹⁶

Medical cannabis (MC) programs present opportunities to investigate the potential therapeutic benefits of MC for pain management and its possible adjunctive use with opioid analgesics. Most MC programs regulate the licensure, cultivation, distribution, and sale of products and require patients to be authorized or certified by a health care practitioner to participate in the program. A 2021 study of patients receiving LOT in Canada by Lee et al¹⁷ found that MC authorization was associated with a significant weekly oral morphine equivalent reduction among individuals receiving high dosage of opioids (ie, oral morphine equivalent >100).¹⁷ However, among individuals receiving lower dosages, a weekly increase was observed. These findings provide direct evidence for the association between MC authorization and changes in opioid dosages. However, Lee et al¹⁷ measured exposure by authorization to receive MC rather than the actual receipt of MC and the length of exposure.

New York state implemented a MC program¹⁸ in January 2016 under the authority of the 2014 Compassionate Care Act.¹⁹ Chronic pain was added in regulation as a qualifying medical condition for receiving MC in March 2017.²⁰ This study aims to assess the association between receiving MC for chronic pain and opioid dosage. Specifically, we examined the changes in opioid dosages among individuals receiving LOT who received MC for longer duration compared with shorter duration.

Methods

Results

Figure 1 shows the selection of the study population. Between 2017 and 2019, there were 8165 patients receiving LOT who also received MC for chronic pain. Of those, 4041 patients (median [IQR] age, 57 [47-65] years; 2376 [58.8%] female) received MC for a longer duration and 4124 patients (median [IQR] age, 54 [44-62] years; 2370 [57.5%] female) received MC for a shorter duration.

Table 1 provides distributions for age, sex, insurance, and MME at baseline for all patients and stratified by baseline MME. Approximately half of all patients were prescribed less than 50 MME, a third were prescribed 90 MME or greater, and the remainder were receiving 50 to less than 90 MME at baseline. Overall, there were statistically significant differences in the distributions of age and source of payments between the exposure and nonexposure groups. (Table 1) The exposure group included more patients aged 65 years and older (1083 patients [26.8%] vs 845 patients [20.5%]), and fewer patients aged 18 to 44 years (827 patients [20.5%] vs 1063 patients [25.8%]) than the nonexposure group. More than half of patients in each group had private insurance (2548 patients [63.1%] in the exposure group and 2542 patients [61.6%] in the nonexposure group). The exposure group had fewer patients with Medicaid insurance (123 patients [3.0%] vs 230 patients [5.6%]) than the nonexposure group. Daily MME at baseline was similar for exposure and nonexposure groups (Table 1). There was no missing information for these variables.

Figure 2 shows the observed and model-estimated 30-day mean daily MME trends for the exposure and nonexposure groups, stratified by baseline MME. The mean MMEs for the exposure group generally started higher than the nonexposure group across strata. In the 2 strata with lower MME (Figure 2A and B), both groups showed small downward trends during the preintervention period. However, statistically significant trends were found only in the lowest MME stratum: a −0.20 (95% CI, −0.31 to −0.09) MME monthly reduction was observed in the nonexposure group, while the exposure group experienced a −0.27 (−0.43 to −0.11) MME reduction (Table 2). Before the intervention started, no significant MME changes were observed in either group for patients with a baseline dosage of 50 MME or greater.

At baseline, the observed mean (SD) MMEs for the exposure vs nonexposure groups were 30.7 (17.3) vs 31.4 (18.5) in the lowest stratum, 66.9 (19.9) vs 66.6 (17.9) for the middle stratum, and 176.2 (106.9) vs 174.0 (106.1) for the highest stratum (Figure 2). After the intervention started, larger reductions in the daily MME were observed in the exposure group compared with the nonexposure group (Figure 2).

Among the lowest MME stratum (Figure 2A), mean (SD) daily MME at the end of the intervention period for the exposure group was 16.0 (22.2), a 48% reduction from 30.7 MME at the baseline, compared with 30.0 (25.4) for the nonexposure group, a 4% reduction from a baseline MME of 31.4. The adjusted analyses found the exposure group had a larger monthly MME reduction compared with the nonexposure group, with a difference of −1.52 (95% CI, −1.67 to −1.37) MME (Table 2). This resulted in a total MME net reduction of −14.53 (95% CI, −17.45 to −11.61) MME over 8 months for the exposure group compared with the MME changes in the nonexposure group.

Among patients with a baseline MME of 50 to less than 90 (Figure 2B), the mean (SD) daily MME at the end of the intervention period was 35.4 (37.6) for the exposure group (a 47% reduction) vs 60.8 (35.7) for the nonexposure group (a 9% reduction). The adjusted analysis showed a significant difference of −3.24 (95% CI, −3.61 to −2.87) MME in monthly MME reduction between the exposure group and the nonexposure group (Table 2). After 8 months, a total MME reduction of −29.49 (95% CI, −35.94 to −23.04) MME was observed for the exposure group compared with the nonexposure group.

For adults with a baseline MME of 90 or greater, by the end of the intervention period, the mean (SD) daily MME was 87.2 (106.6) for the exposure group (a 51% reduction from 176.2 MME at the baseline) vs 149.0 (114.0) for the nonexposure group (a 14% reduction from 174 MME at baseline) (Figure 2C). The adjusted analyses showed a monthly reduction difference of −9.33 (95% CI, −9.89 to −8.77) MME between groups. A large MME net reduction of −69.81 (95% CI, −87.09 to −52.53) was observed for the exposure group over the 8-month intervention.

Additional analyses for daily MME differences between the exposure and nonexposure groups found no statistically significant differences before the intervention. After the intervention started, the differences between groups significantly widened over time (Figure 3).

Discussion

This cohort study of patients in New York state receiving LOT found that receiving MC for chronic pain for a longer duration was associated with significant reductions in opioid dosages. Larger reductions were observed for patients receiving higher opioid dosages at the baseline. Among patients with a baseline dosage of less than 50 MME, a significant monthly reduction in the daily opioid dosages (−1.52 MME) was observed among adults receiving long-term MC compared with those with a shorter MC duration. This reduction was 5.6 times larger than the difference between groups prior to receiving MC (−0.27 MME). The higher the patients’ opioid dosage at baseline, the larger the monthly reduction: −3.24 MME for patients receiving 50 to less than 90 MME and −9.33 MME for patients receiving 90 MME or greater. These reductions were 24.9 times larger in the middle stratum and 37.3 times larger in the highest stratum than the differences prior to receiving MC.

Among patients receiving MC for longer, the mean daily dosage for the group with baseline dosage of 90 MME or greater was reduced to less than 90 MME (87.2), and the mean dosage among those in the 50 to less than 90 MME stratum was reduced to less than 50 MME (35.4 MME). If replicable, such reductions might impact future morbidity and mortality of this population.

In the month immediately after the intervention started, both groups receiving MC experienced a reduction in prescribed opioid dosages; patients in the nonexposure group stopped receiving MC after a maximum of 30 supply days, while patients in the exposure group continued to receive MC longer. This could explain the larger and sharper reductions in MME for adults with longer MC. Even after taking the reductions among adults with shorter MC into the calculation, the estimated total MME net reductions for patients with longer MC over the 8-month intervention were 14.53 MME in the lowest baseline stratum, 29.49 in the middle baseline stratum, and 69.81 in the highest baseline stratum. Among patients receiving MC for longer, daily MME was reduced by 47% to 51% of the baseline dosages after the 8-month intervention. Adults receiving MC for less time reduced their baseline dosages by just 4% to 14%.

We did not have information to determine whether patients chose MC as an adjunct therapy to decrease opioid dosages or in response to clinician pressure to decrease opioid dosages. However, no regulations prohibited the use of opioids and MC simultaneously. The clinical significance of the opioid reductions associated with MC is illustrated by the Centers for Disease Control and Prevention guideline on prescribing opioids for chronic pain,⁷ which highlights the escalation of overdose risk as prescribed MME increases. Clinicians caring for patients receiving LOT should discuss the role MC may play in pain management and information on risks and benefits. When patients wish to use MC and their current clinician is unable or unwilling to certify patients, they should be referred to practitioner who could certify patients for MC.

Patients with chronic pain who are insured by Medicaid receive opioids at a higher rate than privately insured individuals and are more likely to be receiving LOT.³⁰ Few patients in the study population were insured by Medicaid, and of those, many who initiated MC did not continue. This may be because MC is typically not covered by insurance and was thus unaffordable. The lack of affordability of a medication that may reduce the need for opioids for pain management and perhaps risk of opioid use disorder and overdose is a health inequity deserving further examination.

MC is documented to have side effects: among the most common are psychoactive effects, dizziness, and dry mouth. Other concerns include cannabis use disorder and withdrawal symptoms.³¹ New York state regulations allowed MC products with THC to be sold by dispensaries; however, THC was limited at a low dose of 10 mg. Therefore, the risk of adverse events and cannabis use disorder is low if patients use MC as directed. A previous study of MC found that while 37.3% of patients experienced 1 or more adverse events, most were mild, and less than 2% of patients required dose adjustment or cessation, suggesting that MC is far safer than opioids.³² However, caution is still urged in certifying patients with some comorbidities, including coronary artery disease, arrhythmias, or a history of psychosis, or patients who are pregnant.³³

This study had several strengths. To our knowledge, this is the first population-based study with a large sample size that evaluated the association between the length of receiving MC and reduction in daily opioid dosages. The length of receiving MC was used to rigorously measure exposure status and concurrent opioids. Using the CITS model methods on patient-level data allowed us to assess preintervention trends and generate the net differences in the daily MME between exposure and nonexposure groups after the intervention started. Even though not all confounders were controlled for, it is highly unlikely that the greater and sharper reductions right after the intervention started and the continuing reductions in the exposure groups are due to factors other than the receiving MC for longer in patients from the 3 baseline MME levels. Future studies should further examine the specific effects of different types or combinations of MC products and dosages on prescribed opioid dosages.

Conclusions

This cohort study found that receiving MC for longer was associated with opioid dosage reductions. The reductions were larger among individuals who were prescribed higher dosages of opioids at baseline. These findings contribute robust evidence for clinicians regarding the potential benefits of MC in reducing the opioid burden for patients receiving LOT and possibly reduce their risk for overdose. Further research is needed to confirm the causal effect of MC and conduct benefit-risk assessment of MC as opioid alternative or companion treatments to address management of chronic pain and the opioid crisis.

Notes

References