Does Transient Opioid Use Increase Risk of Short-Term Respiratory Exacerbation among Older Adults with Chronic Obstructive Pulmonary Disease?

Abstract

The objective of this study was to examine the association between transient opioid use and acute respiratory exacerbations among older Medicare beneficiaries with COPD. This study was conducted using national Medicare 5% sample administrative claims data between 2012 and 2016 and employed a case-crossover design. The date of eligible COPD exacerbation events was defined as the index date and the presence of opioid prescriptions during a 7-day exposure window prior to index date was compared to a set of 10 control periods, each 7-days long. The association between opioid exposure and COPD exacerbation was estimated using a conditional logistic regression with robust sandwich estimators, after accounting for known time-varying confounders. Among 16,290 eligible COPD exacerbations included in the study sample, the average patient age was 77.08 years, and 64.2% of events occurred in women. Transient exposure to opioids was associated with a 76% increase in the odds of an acute COPD exacerbation (OR: 1.76, 95%CI: 1.67–1.84), and each 25 mg increase in morphine milligram equivalent dose was associated with a 18% increase in the odds of exacerbation (OR: 1.18, 95% CI: 1.15–1.21). Effect estimates were consistent across subgroup analyses conducted among events identified in the emergency department versus hospital, and among individuals with a single exacerbation event versus those with multiple exacerbations. Transient exposure to opioids was associated with an increased short-term risk of respiratory exacerbation among older adults with COPD. Treatment decisions for breathlessness among individuals with COPD need to account for the benefit-risk profile of opioids.

Supplemental data for this article is available online at https://doi.org/10.1080/15412555.2021.2013460 .

Keywords:

• COPD
• opioid
• older adults
• exacerbation
• case-crossover
• Medicare

Introduction

Among Americans aged over 64 years (older adults), chronic obstructive pulmonary disease (COPD) has been reported to be one of the leading causes of death, and results in considerable economic and clinical burden [1]. The prevalence of COPD increases with age and it is reported to be 2–4 times more prevalent in older adults than in adults aged 45–64 years of age. Studies report older age to be associated with mortality resulting from COPD [2]. Older patients with advanced COPD experience an average of two to three exacerbations each year, with each lasting 12 days [3]. COPD exacerbation is defined as an acute worsening of respiratory symptoms [4] that affect lung functioning leading to increased mortality, healthcare utilization and costs, lost work productivity and exercise capacity, and reduced quality of life [5–9]. Among older adults hospitalized with acute exacerbation of COPD, the estimated 30-day all-cause rehospitalization rate is as high as 20 to 23% [10]. Factors such as infection, smoking, poor adherence to medication, and other environmental factors are known to increase the risk of COPD exacerbations [11–13]. Use of opioid analgesics is recommended by current clinical guidelines for alleviation of symptoms of breathlessness among older adults with COPD [4,14], but there is inadequate evidence about the adverse effects associated with such opioid use. Clinical trials assessing the impact of opioid use in severe COPD were conducted in small populations and failed to examine adverse effects [15]. Evidence from observational studies provide some evidence of an elevated risk of exacerbation as well as mortality from opioid use among patients with COPD [16–18]. Existing evidence about these relationships have been limited to periods of past opioid use ranging from 30 to 180 days. However, the Centers for Disease Control and Prevention recommended pattern of opioid use is less than 3 days, and not greater than 7 days, except in unusual situations [19]. This type of short-term use can be characterized as transient or intermittent use. It is critical to investigate the impact of transient use of opioids because the pharmacological mechanism of action of opioids suggests that adverse effects due to opioid use frequently occur in the short term [20–23]. One study by Rong et al. examined the association of transient opioid use and acute respiratory exacerbations among adults with COPD using Medicaid data [24]. However, the Rong et al. study focused on a younger population (average age of 53) and the results may not be generalizable to older adults. Evidence surrounding the risk of COPD exacerbations among older adults is much needed because of the higher burden associated with management of COPD among individuals who are likely to have multiple comorbidities [25]. Therefore, the objective of this study was to examine the association of transient opioid use and short-term risk of acute respiratory exacerbations among older Medicare beneficiaries with COPD.

Methods

This study was conducted using National Medicare 5% sample administrative claims data for the time period from January 1, 2012 through December 31, 2016. The National Medicare data were licensed for use from the Centers for Medicare and Medicaid Services under a data use agreement (DUA# RSCH-2018-52319) and all study procedures were approved by the institutional review board (21x-203). The Medicare data capture all individual enrollee demographics, diagnoses, and healthcare service utilization including pharmacy, hospital, and outpatient claim records.

Study design

The research question in this study examines the effect of transient use of opioid medication on risk of COPD exacerbations. Existing research shows that among Medicare-enrolled older adults, only 10% to 12% of all incident opioid prescriptions translated to long-term use [26], indicating that the majority of opioid exposure in this population is transient. The Global Initiative for Chronic Obstructive Lung Disease (GOLD) defines COPD exacerbations, the outcome in this study, as periods of acute worsening of respiratory symptoms [4]. Further, our understanding of the mechanism of action of opioid medications shows that the respiratory depressive effect of opioids occurs in the short-term [20–23], allowing us to reasonably conclude that the induction period for COPD exacerbations is relatively short. Existing evidence from Rong et al. also shows that the impact of opioids on onset of exacerbation among adults with COPD happens in the short-term, likely within 3 to 5 days of opioid use [24].

Case-crossover designs are best suited to examine associations when the exposure of interest is transient and when the outcome of interest is an acute event with a short induction period [27]. In this design, all eligible individuals are those who have experienced the outcome of interest in the study population (also termed cases). Once these cases are identified, the patterns of exposure are examined to identify whether the hypothesized risk factor is present in the time period immediately preceding the outcome event more often than a random set of time points in the historical record for that same individual. This comparison of exposure prevalence conducted within the same individual at different points of time is called a self-controlled design [28]. Therefore, a case-crossover design compares an individual to themselves at various time points, which allows for all time-invariant characteristics such as sex, race, and chronic clinical comorbidities, to be automatically controlled without the need for statistical adjustment [27].

Eligibility criteria

The population for this study includes Medicare-enrolled older adults diagnosed with COPD who had a respiratory exacerbation between July 1, 2012 and December 31, 2016. The identification of COPD exacerbation was conducted using an algorithm from previous research [24]. An individual was identified as having experienced a COPD exacerbation if a diagnosis code for COPD exacerbation (International Classification of Diseases Ninth Revision [ICD-9] diagnosis code of 491, 492, 493.2, 496, or International Classification of Diseases Tenth Revision [ICD-10] diagnosis code of J41, J43.9, or J44) or respiratory failure (ICD-9 diagnosis code of 518.81, 518.82, 518.84, or ICD-10 diagnosis code of J96.0, J96.20, J96.9, R06.03) was identified in the emergency department (ED) or hospital claims. An individual’s date of occurrence of COPD exacerbation was identified as the index date. All individuals with an exacerbation were included in the study only if they had a confirmed diagnosis of COPD, operationalized as the presence of either one claim with an ICD code for COPD in the inpatient setting or two or more claims with an ICD code for COPD in the outpatient setting, during the 6-month period prior to the index date. Individuals also needed to be at least 65 years or older as of the index date, continuously eligible in Medicare Parts A and B for at least 6 months prior to the index date, and enrolled in Medicare Part D for at least 3 months prior to the index date in order to be included in the analysis. Individuals who had a medical claim with a diagnosis of cystic fibrosis, tuberculosis, or cancer in the 6-months before index date were excluded because the patterns of opioid use and respiratory depression among these individuals may be distinct from the COPD population. Finally, because COPD exacerbations are often recurring, we aimed to ensure that exacerbation events could be considered independent events. We accomplished this by only including exacerbation events that did not have a claim for a previous exacerbation within 90 days. Effectively, this allowed us to include individuals who experience repeated exacerbations multiple times in the study, as long as the exacerbation events were more than 90 days apart from each other.

Identification of exposure within case and control periods

Individual exposure to opioids was evaluated during a hazard period (or case window), defined as a seven-day period immediately preceding an eligible exacerbation event, i.e. a duration spanning from the 7th day before index date up to the 1st day before index date. Ten control periods were also identified, each of them a seven-day period placed consecutively before the case window. Therefore, the first control period spanned from the 14th to the 8th day before the index date, the second control period spanned from the 21st to 15th day before the index date, and so on and so forth until the 10th control period extended from the 77th to 71th day before the index date. Each case and control period was classified as having exposure to opioid medications if a prescription for opioid medications was found in the pharmacy claims file with a sufficient days of supply to allow opioid use for at least one day during a given 7-day period. Medications used for the treatment of opioid use disorder (buprenorphine, naloxone, and naltrexone) were excluded from identification of opioids. The use of opioids in each period was identified as a dichotomous variable indicating presence or absence of use. As part of follow up analyses aimed at identifying a dose-response relationship, we also calculated the dose of each opioid prescription as the morphine equivalent daily dose (MEDD) [29]. In this analysis, the average MEDD for each case and control period was identified and converted into units of 25 mg MEDD, in order to aid the ease of interpretation.

Each case and control period was also examined for the presence of other commonly used medications. These medications included long-acting bronchodilators, short-acting bronchodilators, inhaled and oral corticosteroids, methylxanthines, antibiotics, phosphodiesterase type 4 (PDE-4) inhibitors, benzodiazepines, and beta-blockers. These medication classes were chosen because they are known to affect lung functioning or be indicative of the presence of risk factors that may affect lung functioning. Finally, because the risk of a COPD exacerbation may be elevated by presence of other time-varying acute comorbidities, we also captured claims for presence of an upper or lower respiratory tract infection and asthma exacerbation in each of the case and control periods.

Statistical analysis

All data management and analyses were conducted using SAS version 9.4. Descriptive characteristics were calculated for all study-eligible individuals and exacerbation events, as each individual was allowed to have more than one eligible event included in the analysis. Conditional logistic regression was used to evaluate the association between exposure to opioid medications and the occurrence of COPD exacerbations, after accounting for all hypothesized confounders. In order to account for the fact that each individual may have been included more than once in the analysis, we used the robust sandwich estimate for the covariance matrix [30].

We also conducted a variety of sensitivity analyses to assess the robustness of the potential association between opioid use and risk of COPD exacerbations. First, we classified opioids as being long-acting opioids, schedule II opioids, and schedule III opioids, to evaluate if the risk of exacerbations differed between these types of medications. The classification of opioids into these groups was based on prior research [31] and the Drug Enforcement Administration schedule. Second, in order to account for the severity of COPD exacerbations, we repeated our main analyses among exacerbations that only resulted in ED use and again among those that resulted in a hospital admission. Third, we evaluated the impact of including individuals that have repeated exacerbations by conducting our main analyses in individuals that only had a single exacerbation event versus those that had multiple eligible exacerbation events during the study period. Finally, because it is possible that opioids were being used for symptomatic relief among individuals that were already experiencing exacerbations, we conducted a subgroup analysis that examined the effect of opioids only in events that were not preceded by oral corticosteroids or antibiotics in the 7-days before the COPD exacerbation.

Results

A total of 12 446 Medicare beneficiaries and 16 290 COPD respiratory exacerbation events were included in the study. Baseline characteristics for the events included in the study are presented in Table 1. The mean age for beneficiaries as of the event date was 77.08 years. Evaluation of medication use across all case/control periods prior to each event found that 31.1% of events had a claim for long-acting bronchodilators, 29.5% for short-acting bronchodilators, 34.2% for oral corticosteroids, 41.2% for antibiotics, 29% for benzodiazepines, and 44.1% for β-blockers. Baseline characteristics of the beneficiaries included in the study are shown in Table 1. The majority of the beneficiaries were female (63.3%), white (93.3%), and had an average of 1.31 exacerbations during the study period.

Table 1. Baseline characteristics of COPD respiratory exacerbations.

Characteristics	Event-level characteristics (N = 16 290)	Patient-level characteristics (N = 12 446)
Age
Mean (SD)	77.08 (7.34)	77.29 (7.53)
Sex
Male	5838 (35.8%)	4572 (36.7%)
Female	10 452 (64.2%)	7874 (63.3%)
Race
Caucasian	15 235 (93.5%)	11 617 (93.3%)
African American	794 (4.9%)	627 (5.0%)
Other	233 (1.4%)	183 (1.5%)
Unknown	28 (0.2%)	19 (0.2%)
Use of long-acting Bronchodilators	5059 (31.1%)	4024 (32.3%)
Use of short-acting Bronchodilators	4801 (29.5%)	3935 (31.6%)
Use of inhaled corticosteroids	385 (2.4%)	340 (2.7%)
Use of oral corticosteroids	5563 (34.2%)	4,529 (36.4%)
Use of methylxanthines	393 (2.4%)	297 (2.4%)
Use of PDE4 inhibitors	301 (1.9%)	218 (1.8%)
Use of antibiotics	6707 (41.2%)	5623 (45.2%)
Use of benzodiazepines	4729 (29.0%)	3667 (29.5%)
Use of β-blockers	7190 (44.1%)	5874 (47.2%)
Lower respiratory tract infection	2998 (18.4%)	2729 (21.9%)
Upper respiratory tract infection	1184 (7.3%)	1120 (9%)
Asthma exacerbation	38 (0.2%)	37 (0.3%)

Adjusting for all hypothesized confounders, transient opioid exposure was found to be associated with a 76% increase in the odds of an acute COPD exacerbation (odds ratio [OR] = 1.76, 95% confidence interval [CI]: 1.67–1.84) (Table 2). Measuring opioid exposure as a continuous variable, the adjusted analysis demonstrated that each 25-mg increase in MEDD was associated with an 18% increase in the odds of an acute COPD exacerbation (OR = 1.18, 95% CI: 1.15–1.21) (Table 3).

Table 2. Conditional logistic regression of opioid use (dichotomous) and respiratory exacerbation of COPD among medicare beneficiaries, 2012–2016.

	Within seven-day respiratory exacerbation
	OR	95% CI		P-value
Use of opioid (Y/N)	1.76	1.67	1.84	<0.001
Use of long-acting bronchodilators	2.80	2.51	3.13	<0.001
Use of short-acting bronchodilators	2.04	1.86	2.23	<0.001
Use of inhaled corticosteroids	2.50	1.75	3.57	<0.001
Use of oral corticosteroids	2.68	2.47	2.90	<0.001
Use of methylxanthines	4.78	2.93	7.80	<0.001
Use of PDE4 inhibitors	2.46	1.51	4.01	<0.001
Use of antibiotics	2.32	2.18	2.48	<0.001
Use of benzodiazepines	2.28	2.05	2.54	<0.001
Use of β-blockers	8.45	7.49	9.54	<0.001
Lower respiratory tract infection	2.12	1.93	2.34	<0.001
Upper respiratory tract infection	1.49	1.28	1.74	<0.001
Asthma exacerbation	2.88	1.16	7.14	0.022

Table 3. Conditional logistic regression of opioid use (continuous) and respiratory exacerbation of COPD among medicare beneficiaries, 2012–2016.

	Within seven-day respiratory exacerbation
	OR	95% CI		P-value
Use of opioid (25 mg MEDD)	1.18	1.15	1.21	<0.001
Use of long-acting bronchodilators	2.84	2.55	3.17	<0.001
Use of short-acting bronchodilators	2.08	1.90	2.28	<0.001
Use of inhaled corticosteroids	2.56	1.79	3.66	<0.001
Use of oral corticosteroids	2.75	2.53	2.98	<0.001
Use of methylxanthines	4.84	2.97	7.89	<0.001
Use of PDE4 inhibitors	2.49	1.52	4.08	<0.001
Use of antibiotics	2.38	2.23	2.54	<0.001
Use of benzodiazepines	2.39	2.15	2.66	<0.001
Use of β-blockers	8.89	7.88	10.03	<0.001
Lower respiratory tract infection	2.13	1.93	2.35	<0.001
Upper respiratory tract infection	1.48	1.26	1.72	<0.001
Asthma exacerbation	2.86	1.17	7.00	0.021

The results of the sensitivity analyses are presented in supplemental tables. Supplemental Tables 1 and 2 report the association of opioid use and risk of COPD exacerbation for exacerbations occurring in ED and hospital inpatient settings respectively. Opioid use was associated with an 86% increase in the odds of an acute COPD exacerbation occurring in ED (OR = 1.86, 95% CI: 1.79–2.02), and a 70% increase in the odds of an acute COPD exacerbation occurring in the inpatient setting (OR = 1.70, 95% CI: 1.61–1.81). Additionally, impact of opioid use on COPD exacerbations among beneficiaries with multiple and single exacerbation events are reported in Supplemental Tables 3 and 4 respectively. Opioid use was associated with a 59% increase (OR= 1.59, 95% CI: 1.47–1.72) in the odds of an acute COPD exacerbation among beneficiaries with multiple exacerbations, and an 85% increase (OR = 1.85, 95% CI: 1.79–1.91) in the odds of an acute exacerbation among beneficiaries with a single exacerbation event. Furthermore, adjusted analysis for association of different types of opioid exposure with an acute COPD exacerbation are presented in Supplemental Table 5. Long-acting opioid exposure was found to be associated with a 125% increase (OR = 2.25, 95% CI: 1.76–2.88), schedule II opioid exposure with a 78% increase (OR = 1.78, 95% CI: 1.69–1.88), and schedule III opioid exposure with a 50% increase (OR = 1.50, 95% CI: 1.39–1.62) in the odds of an acute COPD exacerbation. Finally, Supplemental Table 6 shows that the study results are consistent (OR for opioid use = 1.69; 95% CI: 1.60–1.79) in a subgroup of events that were not preceded by use of oral corticosteroids or antibiotics in the 7-days before the index date.

Discussion

This study used a case-crossover design with data from a national sample of Medicare enrollees to investigate the short-term risks associated with transient opioid use in older adults with COPD. Given the considerable economic and humanistic burden of COPD among older adults, identifying opportunities to improve the standard of care and reduce serious adverse reactions in this population will be of critical benefit. Our findings indicate that acute administration of any opioid medications can result in elevated short-term risk of COPD exacerbation among older adults. This finding is in line with previous research conducted among individuals with COPD – four different studies have found similar risk of exacerbation associated with historical use of opioid medications, either alone or in combination with other medications such as benzodiazepines [16–18,32]. However, these studies only captured opioid use during the 3 to 12 months prior to COPD exacerbation.

Our investigation of this association was based upon both the prevailing patterns of opioid use, and our understanding of the mechanism of action of the hypothesized effect of opioid medications on respiratory depression. Opioids act to decrease the rate of breathing and oxygen saturation by binding with the µ-opioid receptor in the lungs [33–35]. Evidence from randomized, double-blind controlled trials show that the respiratory depressive effect of opioids can be observed in as little as 24 h after administration of the medications [33]. The duration of this effect may be prolonged after the use of extended release or long-acting formulations, immediate onset of respiratory depression is to be expected after any opioid use. Given that older adults with COPD often have lower baselines of respiratory function, this depressive effect may lead to increased risk of COPD exacerbation in the short-term. The use of a case-crossover design offers two critical advantages – first, it allows for an investigation of the short-term impact of transient exposure on acute outcomes better than other pharmacoepidemiological designs [25,27], and second, it offers a self-controlled approach where each individual serve as their own control therefore automatically controlling for all intransient confounders that may or may not available in our dataset. Previous research investigating the relationship between opioid use and short-term COPD exacerbation among adults employed similar methodology using administrative claims data and found that the risk of short-term exacerbations increased when the period of examination decreased from 8 days to 5 [24]. Our findings are consistent with these results and demonstrate this effect in a population of older adults.

Our findings were also consistent across several sensitivity analyses. First, we found a dose-response relationship that is consistent with the effect of opioids on several other adverse reactions and their mechanisms of action. Second, our analyses demonstrate that severity of COPD does not confound the effect of opioids on risk of exacerbation. This finding was demonstrated by the consistency of our results both across severity of COPD exacerbation and across individuals with and without multiple exacerbation events during the study period. Finally, because our results were consistent among events that were not already receiving treatment using corticosteroids or antibiotics, there is minimal evidence of confounding by indication.

Findings from this study are critical in light of existing evidence for the treatment of dyspnea among older adults with COPD. Recommendations from GOLD as well as clinical guidelines from the American Thoracic Society suggest that dyspnea and breathlessness among individuals with COPD may be treated by symptom management and palliative care, using opioid medications [4,14]. In light of our findings and other existing evidence, there is a need for a revisiting of the benefit-risk profile associated with treatment of breathlessness with opioid medications among individuals with COPD. Improvements in clinical treatment of breathlessness can help reduce hospitalizations, readmissions, and even risk of mortality leading to improvements in both quality of life and cost of care. Further research is needed to identify levels of MEDD with lower risk of exacerbation and investigate potential interactions of opioids with other medications known to cause a respiratory depressive effect such as benzodiazepines.

While evidence from this study is valuable for future clinical practice, it is important to note a few key limitations: First, the use of administrative claims has certain drawbacks. For instance, these data only capture whether a prescription was filled, but not whether medication was actually consumed by an individual. They also do not capture all potential information for each study participant – for example, exacerbations that did not result in an ED or hospital visit may have been missed, and identification of other critical confounders such as exposure to air pollution or other use of over-the-counter medication were not available in these data. However, the use of a case-crossover design automatically controls for all intransient risk factors, whether or not they were captured in the data. Second, the key limitation of a case-crossover design is its inability to account for time trends in exposure that may confound the effect of interest in the study. However, previous studies show that prescribing rates of opioid medications during the study period are decreasing [26]. Therefore, any potential effect of exposure trends is likely to have biased our findings toward the null, indicating that our findings are likely robust. Nevertheless, future research should consider other self-controlled designs that include a comparator group such as case-time-control design to identify the impact of potential exposure trends [28]. Finally, the case-crossover design does not adequately account for the presence of time-varying confounders or other time-invariant variables that could moderate the relationship of interest in this study. Our selection of confounders was based on clinical knowledge of COPD and existing literature, but future studies are needed using different data sources that can confirm our findings.

Conclusions

In conclusion, this study found that the presence of any opioid exposure is associated with an increased short-term risk of exacerbation among older adults with COPD. Treatment decisions for breathlessness among individuals with COPD need to account for the benefit-risk profile of opioids and require careful evaluation of existing treatment options.

Declaration of interest

The authors report no conflict of interest

Funding

No external funding was obtained for this study.

Data availability statement

All study data are available from the Centers for Medicare and Medicaid Services upon request.