Association of Opioid Use and COPD Exacerbation: A Cohort Analysis Among Older Adult Beneficiaries Within a Medicare Advantage Plan

The Global Initiative for Chronic Obstructive Lung Disease guidelines characterize chronic obstructive pulmonary disease (COPD) as shortness of breath, excess mucus production, coughing, and wheezing; the terms emphysema and chronic bronchitis have been previously used to describe this disease.¹ In 2017, the National Center of Health Statistics stated that COPD remains in the top 10 leading causes of death—fourth behind heart disease, cancer, and unintentional injuries.² In 2015, a higher percentage of patients who were unable to work had COPD than those who did not (20.4% vs 4.8%; 95% CI; 19.3%-21.4%).³ A 2010 study reviewed surveys and the 2010 Centers for Medicare & Medicaid Service (CMS) data and reported 16.4 million workdays were lost to COPD, costing a total of $3.9 billion in medical expenses.^4,5

Tobacco smoking, age, environmental exposure, infections, and suboptimal COPD therapy influence the progression of COPD, lung function, and the risk of acute exacerbations. ¹ Opioids used for pain treatment and medications from other comorbidities can alter exacerbation risk.^6-8 Evidence suggests that patients with COPD are among the highest users of opioids among patients with comorbidities.⁹ The most recent opioid crisis and the rising concern of opioid overuse in older adults aged 65 years and older have increased concerns of respiratory depression in patients with COPD.¹⁰ Comorbidity burden also has an added effect on needing opioid therapy.¹¹

Only a few studies have explored the association between acute COPD exacerbations and opioid use, as well as its associated health care resource utilization and costs.^6,7 The first objective in this study was to examine the incidence of COPD exacerbations that led to hospitalizations within 30 days of opioid use among patients with COPD and to determine predictors associated with COPD exacerbations leading to hospitalization. The second objective was to compare the health care costs between patients with COPD who experienced COPD hospitalizations within 30 days of opioid use with those who did not experience any hospitalization during the study period.

Study Design and Patient Population

This was a retrospective cohort study that was conducted using claims data from members of Medicare Advantage plans (MAPs) between January 1, 2011, and December 31, 2014. The MAP is a health maintenance organization that had approximately 120,000 beneficiaries across Texas. Plan membership and member summary files, hospitalizations, outpatient office visits, and pharmacy claims were provided by the pharmacy benefits manager, Argus/SXC. The study was compliant with the Health Insurance Portability and Accountability Act and approved by the university’s institutional review board.

The study patients were required to have at least one pharmacy claim for any opioid drug between January 1, 2012 and December 31, 2013, and the first observed prescription fill date was considered the index date (Figure 1). Patients were included in the study if they were at least aged 65 years on the index date. Inclusion also required a diagnosis of COPD any time during the pre-index period (International Classification of Disease, Ninth Revision, Clinical Modification [ICD-9-CM] codes 491.xx chronic bronchitis, 492.xx emphysema, 496.xx chronic airway obstruction not elsewhere classified). All patients were required to be continuously enrolled in the plan for at least 12 months before and after the index date. Patients with a diagnosis of end-stage renal disease (ESRD)or kidney failure (ICD-9-CM codes 582.xx, 583.0–583.7, 585.xx, 586.xx, 588.xx) or cancer (including leukemia and lymphoma, ICD-9-CM codes 140.xx–172.xx, 174.xx–195.xx, 200.xx–208.xx) any time within the data period, in hospice pre-index date, or who had any pharmacy claims for opioid drugs 12 months pre-index date (to ensure incident opioid use) were excluded from the study.

Study Outcome Measures

The patients were followed for 12 months after the index date. During the 12-month postindex period, beneficiaries were classified into 2 groups, and study outcomes and health care costs were evaluated.

During the postindex period, the primary study outcome was whether beneficiaries experienced a COPD hospitalization within 30 days. Two mutually exclusive groups were defined, namely hospitalized and not hospitalized, based on whether a hospitalization was indicated within the 30-day postindex period. Patients who experienced a COPD exacerbation that led to hospitalization or an emergency department (ED) visit within 30 days postindex were categorized into the hospitalized group. Patients who did not experience a COPD exacerbation or an ED visit within 30 days were categorized into the not-hospitalized group.

During the 12-month postindex period, health care costs paid by the plan were calculated as inpatient costs, outpatient costs, prescription costs, and total costs. Total costs are the sum of inpatient, outpatient, and prescription costs of both Medicare Part B and Medicare Part D. Health care resource utilization during the postindex period was calculated for 12 months and reported as acute admits, long-term acute care admits, observation admits, rehab admits, skilled nursing facility admits, ED visits, primary care provider visits, and specialist visits. The sum of all resources during the postindex 12-month period was calculated into total admits/visits.

For the secondary objective, to understand the impact of incident opioid use-associated COPD exacerbation on health care resource utilization and costs, the comparison was made between those who experienced an exacerbation and those who did not have any hospitalization during the follow-up period.

Baseline Characteristics

Eligible patients had longitudinal data for 2 years. During the 12-month pre-index date, baseline characteristics such as age, sex, plan type, type of opioid on index date, comorbidities, Deyo’s Charlson Comorbidity Index (CCI), CMS risk score, and total health care costs were evaluated. The type of opioid used on index date was characterized by name into 3 groups: hydrocodone, codeine, and others (including fentanyl, hydromorphone, morphine, and oxycodone). The patients were followed for 12 months after the index date after being classified into the 2 groups (hospitalized vs not hospitalized). All MAPs provided prescription drug coverage and were categorized into 2 groups: Total Care (dual special need plans) and health advantage plans (various MAP drug plans). Deyo’s version of CCI was used to measure comorbidity burden within the administrative claims databases.¹² An aggregate score was computed from the weight based on the 17 medical conditions (the lowest score is 1 and the highest score is 6).¹² Because the diagnosis for COPD was in the inclusion criteria, the diagnoses for COPD were excluded from the CCI calculation. The CMS risk score reflected overall health, which included health status, Medicaid eligibility, and patient demographics that were not reflected by the CCI.^13,14

Statistical Analysis

Descriptive analyses were used to summarize characteristics of patients who had a 30-day COPD exacerbation postindex vs those who did not have a COPD exacerbation within 30 days. T tests were used to compare normally distributed continuous variables (age and CMS risk score), while Wilcoxon–Mann–Whitney tests were used to compare non-normally distributed continuous variables (health care costs, CCI, and health care resource utilization incurred within 12 months pre- and post-index periods). Chi-square and Fisher’s exact tests were used to compare categorical variables (type of opioid drug, gender, type of health plan, type of comorbidities, and CCI category). Multiple logistic regression was used to determine characteristics associated with 30-day COPD exacerbation. All statistical analyses were conducted using SAS version 9.4 (SAS Institute, Inc, Cary, NC).

A total of 7919 beneficiaries had at least 1 claim code for opioids between January 1, 2012 and December 31, 2013 (Figure 2). The final study sample comprised a total of 1742 beneficiaries who were aged 65 years or older, had at least 1 diagnosis for COPD, had 24 months of continuous enrollment, were without any ESRD or cancer diagnosis, were not in hospice, and had no claim for opioids within 12 months pre-index to ensure incident use of opioids. Among them, exacerbation incidences leading to hospitalization within 30 days of opioid use were approximately 5%. Cost incurred within 30 days of incident opioid use was higher than those who were not hospitalized postindex (total costs were $6540 vs $3349).

The sample comprised slightly more females in each of the 2 groups: hospitalized within 30 days (58%) and not hospitalized within 30 days (51.4%) (Table 1). The mean (±SD) age was 74 (± 6.9) years for beneficiaries in the hospitalized group and 75 (±6.9) years for beneficiaries not hospitalized. Approximately 20% of beneficiaries in the hospitalized group and 22% of beneficiaries in the not-hospitalized group were enrolled in Total Care, a type of insurance plan. The majority type of opioid used on the index date was hydrocodone for both groups: 100% for those hospitalized and 87% for those not hospitalized. Patients who were not hospitalized had a slightly higher mean CCI, 0.02 (±0.18), than those who were hospitalized. The mean (±SD) CMS risk score for beneficiaries in the hospitalized group was 1.81 (±1.19), which was not significantly different from the group who was not hospitalized, at 1.73 (±1.17). Multiple logistic regression did not find any significant predictors for 30-day COPD exacerbation (Table 2).

Health Care Costs and Utilization Pre- and Postindex Opioid Use

During the 12 months postindex, compared with those who did not have any hospitalization, patients with COPD who experienced an exacerbation within 30 days of incident opioid use had significantly higher health care costs and health care resource utilization ($15,038 ± $16,962 vs $4776 ± $5192; P < .05; Table 3). Out of the total health care costs, outpatient costs were also higher in patients who were hospitalized within 30 days, $3920 (±$4070), than in those who were not hospitalized within 12 months, $1841 (±$3839)

Our study demonstrates that of patients enrolled in a MAP who had COPD and incident opioid use, approximately 5% were hospitalized within 30 days. Patients who experienced any COPD exacerbation within 30 days of incident opioid use incurred significantly higher costs and health care resource utilization compared with those who did not have any hospitalization during the postindex period.

The results of our study, indicating a 5% incidence of hospitalized patients with opioid use, were not consistent with previously reported findings. One study reported the odds of hospitalization were higher in patients with COPD and opioid use.⁶ Another study found a higher risk of ED visits for patients with COPD with incident opioid use.⁸ However, previous studies did not evaluate the relationship between characteristics associated with exacerbation leading to hospitalization and opioid use. Our research advances the knowledge further by using a MAP database to examine the relationship between characteristics of the hospitalized and the not-hospitalized populations.

This study found that patients who experienced a COPD exacerbation had significantly higher total costs than those who were not hospitalized during the postindex period. Our results were consistent with studies that have shown that patients who had more exacerbations leading to hospitalizations have generally higher out-of-pocket expenses.^15,16 The previous studies, however, did not identify the financial impact opioid use has on patients with COPD. Our research does provide evidence that there is an impact of opioid use and reciprocal hospitalization on the cost of care and resource utilization. The result indicates that such patients should be closely monitored for proper opioid use in addition to COPD management.

Limitations

Data included in this study consisted only of patients with incident opioid use, which excluded chronic users of opioids in the COPD population. This study was unable to confirm whether the opioid pills were taken or continued when dispensed by the pharmacy. Subgroup analysis to quantify opioid use and its impact on hospitalizations was not performed in this study because the data lacked information on how the patients used opioids (ie, scheduled dosing vs as-needed dosing, total morphine milliequivalents per day, or day’s supply per prescription). Considering that more than half of patients who have COPD from our available database were on opioids, the results may not be fully representative of patients with COPD with opioid use. Prospective controlled studies are needed to identify patients with COPD who were on medications within the 2019 American Geriatric Society Beers Criteria that would increase fall risk, such as benzodiazepines. Comprehensive studies that include COPD medication adherence and tobacco use may be needed to test their impact on these populations.

Patients who have COPD and were prescribed opioids had a 5% probability of being hospitalized within 30 days due to the associated exacerbations. Further, those who were hospitalized within 30 days of incident opioid use had significantly higher costs than those who did not experience an incident. 

Authors: Sarah S Laxa, PharmD¹ • Qingqing Xu, MS² • Omar Serna, PharmD, BCACP³ • Sujit S Sansgiry, PhD⁴

Affiliations:

¹Cigna, Houston, TX, USA
²AbbVie Inc, North Chicago, IL, USA
³Heights-Studewood PharmacyHouston, TX, USA
⁴Department of Pharmaceutical Health Outcomes and Policy, College of Pharmacy, University of Houston, Houston, TX, USA

Disclosures:
Authors report no relevant financial relationships, and research did not receive funding.

Author contributions:
All authors contributed to the study conception and design. Material preparation, data collection, and analysis were performed by Dr Laxa and Ms Xu. The first draft of the manuscript was written by Dr Laxa and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Address correspondence to:
Sujit S Sansgiry, PhD
Department of Pharmaceutical Health Outcomes and Policy
University of Houston, College of Pharmacy
Health and Biomedical Sciences Building 2, Office 4050
4849 Calhoun Road
Houston, TX 77204-5047
Phone: (832) 842-8392
Email: sansgiry@central.uh.edu