Assessment of Prescription Analgesic Use in Older Adults With and Without Chronic Kidney Disease and Outcomes

Key Points Question Is prescription analgesic use higher among Medicare patients with chronic kidney disease (CKD) compared with the general population? Findings In this cohort study of more than 6 million Medicare Part D beneficiaries, use of both opioids and prescription nonsteroidal anti-inflammatory drugs (NSAIDs) increased substantially among patients with CKD aged 65 years or older from 2006 to 2015. Patients with CKD used significantly more opioids but less NSAIDs compared with the non-CKD population. Meaning Results of this study suggest that optimization of pain management in CKD is warranted.


Introduction
Pain is one of most common symptoms experienced by patients with kidney disease. 1,2 More than 70% of patients with chronic kidney disease (CKD) report experiencing pain. 3 Pain has been associated with lower quality of life, high symptom burden, and greater risk of developing kidney disease, its progression, and mortality. 1,4 Thus, management of pain is a vital component of comprehensive care for patients with CKD. 3 The already high medication burden makes pain management even more complex in patients with CKD. There are no well-established guidelines for pain management specific to those individuals with kidney disease. 2,[5][6][7][8] Special precautions are needed to manage treatment of patients with reduced kidney function, who may be more prone to experiencing drug toxic effects, adverse effects, greater dose adjustment requirements, and drug interactions. 2 Opioids and nonsteroidal antiinflammatory drugs (NSAIDs) are the most commonly used analgesics. Opioids may have extended half-life in patients with advanced CKD, with substantial effects on the central nervous system, resulting in respiratory depression, hypotension, and addiction. 2,5,6 Long-term opioid use is a risk factor for opioid abuse and dependence. 9 Use of NSAIDs in patients with CKD may result in nephrotoxicity, fluid and electrolyte imbalances, hypertension, and other complications. 2,6,[10][11][12][13] Therefore, patients with CKD may be especially at risk of receiving suboptimal pain control and inappropriate use of prescription analgesics. 14 The opioid epidemic has recently received considerable attention from clinicians, policy makers, and the public. 15 However, to our knowledge, few studies have assessed use of analgesics and outcomes in patients with reduced kidney function, particularly with regard to the long-term effects. 3 The present study sought first to examine trends from 2006 to 2015 in the use of opioids and prescription NSAIDs among older (age Ն65 years) patients with CKD in the US, and then to explore factors associated with opioid and NSAID use, as well as outcomes associated with their use, including development of end-stage kidney disease (ESKD) and death. We hypothesized that prescription analgesic use would be higher in patients with CKD compared with the general population. Furthermore, we postulated that use of prescription NSAIDs would be associated with more rapid kidney disease progression to ESKD and opioid use with both higher mortality and advanced-stage or progressive kidney disease.

Study Design
Ten annual cohorts of Medicare beneficiaries were identified to evaluate analgesic use among older adults with and without CKD. A visual depiction of the study design and sample selection is provided in eFigure 1 in the Supplement. Briefly, for each calendar year, patient demographic characteristics, CKD status, and comorbidities were assessed in the selection period, the year prior to the calendar year; and patients' use of analgesics was assessed within the calendar year. Beneficiaries aged 65 years or older, continuously enrolled in fee-for-service Medicare and Part D, and with no diagnosis of ESKD in the selection period were selected. Presence of CKD and other comorbidities was defined as having at least 1 inpatient claim or 2 outpatient claims within the selection period with specified International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) codes. 16 Investigated conditions and corresponding ICD-9-CM codes are summarized in eTable 1 in the Supplement. Patient CKD status, referring to whether the patient was diagnosed as having CKD and the stage of CKD, was determined from administrative claims with CKD-related ICD-9-CM codes and CKD stage-specific codes (585.x). Patients without CKD served as a control group in this study. The sample size of the annual CKD cohorts increased from 29 843 in 2006 to 93 755 in 2015.
A longitudinal cohort was developed to examine the association between analgesic prescription and patient outcomes. Study design and sample selection are depicted in eFigure 2 in the Supplement. Briefly, Part D enrollees in 2006-2015 were selected. The first observed date of Part D enrollment was the beginning of the 1-year selection period. Demographic characteristics, patient CKD status, comorbidities, and use of analgesics at baseline were assessed in the selection period using the same algorithm as the annual cohorts. Study participants were then followed after the end of the 1-year selection period. Medicare beneficiaries who were less than age 65 years at the index date, were enrolled in Medicare Advantage, or were without continuous prescription coverage in the 1-year selection period were excluded from this study. A total of 649 339 participants were selected.

Measures
Analgesic use was assessed in the following 3 aspects: (1) overall use, as measured by the proportion of individuals with any prescription opioids/NSAIDs; (2) long-term use, measured by the proportion of use of prescription opioids/NSAIDs for greater than 90 days; and (3) cumulative use, measured by total annual days' supply of opioids/NSAIDs among users in a given calendar year. A list of National Drug Codes for opioids and NSAIDs was developed according to the American Hospital Formulary Society drug categories. Corresponding Medicare Part D claims were then extracted based on these National Drug Codes. Beneficiaries who had any opioid claim in the calendar year were classified as opioid users. Total days' supply of opioids was calculated as the sum of all opioid prescriptions dispensed in the calendar year. A threshold of 91 days was applied to define chronic pain therapy. 10,17 The measures were calculated in the same manner for prescription NSAID use. sample for those without onset of ESKD. Study participants were censored at date of death or end of the study period (December 31, 2015) when progression to ESKD was the event, whereas they were censored at the end date of this study when all-cause mortality was the event.
Demographic information was obtained from Medicare data sets. Investigated comorbidities included CKD-related conditions (hypertension, cardiovascular disease, diabetes) and pain related conditions (cancer, depression, back pain, neck pain, arthritis, headache, and HIV). 18 Cardiovascular disease was defined as having of any of the following conditions: cerebrovascular accident, peripheral vascular disease, atherosclerotic heart disease, congestive heart failure, dysrhythmia, or other cardiac comorbidities. In the outcome analyses, baseline characteristics of the longitudinal cohort were assessed by analgesic use using t test and χ 2 test. Multivariate Cox proportional hazards regression models were applied to assess the association between analgesic use, development of ESKD, and all-cause mortality. The competing risk of death was taken into account when modeling development of ESKD.

Statistical Analysis
Statistical significance was set at 2-sided P < .05. All analyses were conducted using SAS version 9.4 (SAS Institute). Maps were developed using R software (R Project for Statistical Computing).

Results
In this cohort study, a total of 6 260 454 beneficiaries (CKD 9.6%) were selected in the annual cohorts, and 649 339 beneficiaries (CKD 8.3%) were selected in the longitudinal cohort.

Use of Analgesics
The proportion of Medicare Part D beneficiaries aged 65 years or older diagnosed with CKD increased  Figure 1C). The proportion of prescribed opioids for more than 90 days was higher in the CKD users compared with the non-CKD users (35.6% vs 29.1%, 2015) ( Figure 1C). Long-term use of NSAIDs was stable through 2015 ( Figure 1D). Approximately one-third of CKD users had received prescription NSAIDs for more than 90 days, varying from 23.2% in stages 4-5 CKD to 33.2% in stages 1-2 CKD (2015) ( Figure 1D).
In terms of cumulative use, despite a greater proportion of those with advanced CKD using

JAMA Network Open | Nephrology
Prescription Analgesic Use in Older Adults With and Without Chronic Kidney Disease and Outcomes    Table 2, opioid use was higher in patients with CKD, particularly stages 4-5 (odds ratio [OR], 1.35; 95% CI, 1.33-1.37; P < .001) compared with non-CKD with adjustment of demographic characteristics and CKD-and pain-related comorbidities. NSAID use was lower in patients with CKD stages 4-5 (OR, 0.55; 95% CI, 0.54-0.56; P < .001). There were some overlapping factors associated with higher odds of analgesic prescriptions (opioid or NSAID), like younger age (aged 65-75 years), female sex, hypertension, diabetes, back pain, neck pain, arthritis, headache, and HIV. In contrast, individuals who were White and had CKD, cancer, and cardiovascular disease were more likely to use opioids but less likely to use NSAIDs. Having depression was significantly associated with opioid use (OR, 1.25; 95% CI, 1.24-1.26; P < .001) but not prescription NSAID use (OR, 1.00; 95% CI, 0.99-1.01; P = .45) ( Table 2). Factors associated with use of prescribed opioids and NSAIDs for more than 90 days are shown in eTable 5 in the Supplement.    then stabilizing or decreasing slightly thereafter. A higher proportion of opioid vs prescription NSAID use was found in patients with CKD compared with non-CKD Medicare beneficiaries. As expected, cumulative use of analgesics, both opioids and prescription NSAIDs, decreased with CKD progression. This prescription pattern could be associated with patients with advanced CKD being more susceptible to toxic effects and adverse effects in the setting of reduced kidney function.

Discussion
Additionally, NSAIDs were widely used in earlier stages of CKD while opioids were more frequently used in patients with advanced CKD. This finding may reflect both the variation in severity of pain by CKD stage, as opioids are typically used to treat moderate to severe pain, and avoidance of NSAIDs in patients with more advanced CKD, in whom risk of progression is higher with use of this class of medication owing to the inhibitory effect of NSAIDs on prostaglandin synthesis in the kidneys, with resultant potential for intrakidney ischemia. 2 The trends in opioid use observed in this study are consistent with a recent study of the general  The racial/ethnic differences in analgesic use observed in our study may be associated with barriers in accessing health care and different cultural perceptions of pain management. For example, Asian American patients used more prescription NSAIDs than opioids compared with other racial/ ethnic groups. A prior study found that Asian American patients had greater perceived barriers to cancer pain management than patients of other races/ethnicities, which were consistent with their concerns about disease progression, drug tolerance, and addiction and view of cancer pain as inevitable. 24 With respect to comorbidities, our findings were consistent with previous findings that the presence of certain comorbidities such as diabetes, hypertension, depression, cancer, back pain, neck pain, arthritis, and headache increased patients' need for pain relief medications. [25][26][27] High rates of comorbid depression indicate a substantial burden of mental health problems in the CKD population. 17 Patients with comorbid CVD were less likely to use NSAIDs, which may be due to the cardiovascular risks associated with these medications. 28,29 The substantial geographic variations in overall, long-term, and cumulative use of opioids and In our outcomes analyses, opioid use was associated with a higher risk of developing ESKD and death, independent of CKD status, which may be due to opioid users being at high risk of opioid dependence and opioid overdose, with consequent higher hospitalization and mortality. 9,31 It is possible that people in pain requested or received more opioids. In contrast with opioids, no association was observed between NSAID use (except receiving NSAIDs for 91-180 days) and progression to ESKD, but a significant protective association was observed between receiving NSAIDs and death, with adjustment of other patient-level factors. Previous studies have been inconsistent with regard to the association between NSAIDs and progression to ESKD. Some found that regular consumption of NSAIDs was associated with decreased kidney function. [32][33][34][35] Additionally, the risk of developing ESKD increased with greater cumulative exposure to NSAIDs among those with CKD. 36,37 However, others have suggested that use of NSAIDs was not associated with decreased kidney function. [38][39][40][41][42] Use of NSAIDs in CKD requires caution because of their known direct nephrotoxic effects, risk of causing fluid and electrolyte imbalances, hypertension, and other complications. 2,6 However, effective pain management may relieve pain and related stress and could potentially delay disease progression and mortality. 1 The observed protective association of NSAIDs and outcomes may reflect the benefits of pain relief relative to potential nephrotoxic effects, or it may be the consequence of rational clinical decision-making, such that physicians use NSAIDs for patients with CKD at earlier stages and for those who have low risk of progressing to ESKD regardless of stage, while they choose opioids for patients at later stages of disease and high risk of progression.
Our results highlight the need for further investigation of the benefits and risks of choosing NSAIDs vs opioids in pain management.
likely improved over our study period, which could affect the observed trends. 43 Further analysis of use of opioids and prescription NSAIDs in CKD, particularly in patients with stage 1 and stage 2 CKD identified by presence of albuminuria and proteinuria, would have been more informative; however, we do not have laboratory data available in Medicare claims data sets. Second, analgesic use was measured from medication claims and may not reflect actual use of medication. Medication claims cannot capture prescriptions filled outside the Medicare Part D networks. Most NSAID use is over the counter, and NSAID use reported in our study is likely a fraction of actual NSAID use. Third, substantial limits on opioid use in response to the opioid crisis have been implemented since 2015, so our trends may not reflect the most recent changes in opioid use. Fourth, our study may be subject to omitted variable bias, as not all of the risk factors that may be associated with pain and CKD progression were controlled for. Last, we restricted our study patients to fee-for-service Medicare beneficiaries enrolled in Part D owing to lack of data from Medicare Advantage plans. Thus, our results may not be generalizable to other populations. Additionally, Medicare Part D enrollment increased over time, which may also be a factor in the observed increase in analgesic use.

Conclusions
Our study suggests that there has been substantial use of opioids and prescription NSAIDs among older patients with CKD in the US with a clear increasing trend until recently, with early suggestion of decreasing use, perhaps due to the recent heightened awareness of the opioid epidemic in the US.
Clinicians must continue evaluating benefits and risks of using opioids and NSAIDs in this and other patient populations. Specifically, clear clinical guidelines for chronic pain management among patients aged 65 years and older with CKD may be warranted to address potentially suboptimal pain management in this patient population.