Association between Posttransplant Opioid Use and Immunosuppressant Therapy Adherence Among Renal Transplant Recipients

Abstract

Introduction:

Little is known about the effect of posttransplant opioid use on adherence to immunosuppressant therapy (IST) among adult renal transplant recipients (RTRs).

Objective:

To examine the relationship between opioid use and IST adherence among adult renal transplant recipients during the first year posttransplant.

Methods:

Longitudinal data were analyzed from a retrospective cohort study examining US veterans undergoing renal transplant from October 1, 2007 through March 31, 2015. Data were collected from the US Renal Data System, Centers for Medicare and Medicaid Services Data (Medicare Part D) and Veterans Affairs pharmacy records. Dose of opioid prescriptions were collected and divided based on Annual Morphine Milligram Equivalent (AMME) within a year of transplant. Proportion of days covered of greater than or equal to 80% indicated adherence to tacrolimus. Unadjusted and multivariable adjusted logistic regression analyses were performed.

Results:

Study population of 1,229 RTRs included 258 with no opioid use, while 971 opioid users were identified within the first year after transplantation. Compared to RTRs without opioid usage, RTRs with opioid usage had a lower probability of being adherent to tacrolimus in unadjusted logistic regression [OR (95% CI): 0.22 (0.07–0.72)] and adjusted logistic regression [OR (95% CI): 0.11 (0.03–0.44)]. These patterns generally remained consistent in unadjusted and adjusted main and sensitivity analyses.

Conclusions:

Findings indicate RTRs who use prescription opioids during the first year posttransplant, regardless of dosage/amount, are less likely to be adherent to tacrolimus. Future studies are needed to better understand underlying causes of the association between opioid use and tacrolimus nonadherence.

Introduction

Renal transplantation is the preferred treatment for end-stage renal disease (ESRD), with graft survival rates greater than 90% and acute rejection rates less than 10% within one-year posttransplant.[1] To maintain transplant function and minimize graft failure rate, renal transplant recipients (RTRs) are prescribed immunosuppressant therapy (IST) following transplant. However, as summarized by Morissey et al., between 15% and 30% of RTRs become nonadherent to their IST, while a meta-analysis by Dew et al. estimated an annual nonadherence rate of approximately 36% among RTRs.[2–4] Moreover, IST nonadherence increases as more time elapses posttransplant.[5,6] IST nonadherence increases the risk of graft failure sevenfold, and 21% of nonadherent patients experience late acute rejection compared to only 8% of adherent patients by 5 years posttransplant.[3,7] Gaston et al. found that 35% of grafts lost within the first 6 months following transplant were due to nonadherence.[8] RTRs may not adhere to their IST for multiple reasons including complex or burdensome treatment, side effects, financial constraints, mental health issues (such as depression or severe anxiety), forgetfulness, and irregular lifestyles.[2,9,10]

Prior to transplant, RTRs frequently face issues with chronic pain which may have ramifications for posttransplant adherence and outcomes. Kimmel et al. noted that among Medicare recipients, opioid prescriptions are more common in dialysis patients than the general population.[11] Between 2006 and 2010, greater than 60% of ESRD patients on dialysis were prescribed an opioid at least once annually, and over 20% had a chronic (defined as greater than or equal to 90 days) opioid prescription annually.[11] While opioids can be effective for pain management, there may be serious consequences as outcomes such as increased mortality and hospitalization are associated with opioid use among ESRD patients.[11,12]

ESRD patients who filled opioid prescriptions prior to transplantation were more likely to fill opioid prescriptions in the year following transplant, and level of opioid use pre-transplant corresponded to level of use posttransplant.[13] Pre-transplant opioid use is associated with an array of complications (e.g., increased hypotension, cardiac arrest, mental status changes, among others) and increased risk of mortality and graft loss among RTRs.[13–15] Increased risk of mortality and graft loss were also noted with posttransplant opioid use.[13]

Given the critical role of IST adherence in facilitating graft survival, it is important to develop a better understanding of the effects of posttransplant opioid use on IST adherence. Although Lentine et al. noted that deceased donor RTRs who used opioids pre-transplant at a level of ≥23.8 mg/kg were found to have increased nonadherence with their medical treatment regimen compared to non-users, they did not focus exclusively on IST nonadherence.[14] Instead, they examined a composite of IST and other medication, follow-up appointments and lab testing, and thus there is a lack of clarity on the relationship between opioid use and adherence to IST.[14] A PubMed search (years unlimited) further revealed that little is known about the impact of posttransplant opioid use on adherence to IST following renal transplant. Notably, Surbhi et al. found that opioid use was associated (in adjusted analysis which considered sociodemographic factors and comorbidities such as depression/anxiety, among other covariates) with increased risk of medication nonadherence in treatment of chronic disease states in a sample of Medicare patients classified as high-cost healthcare users.[16] Jeevanjee et al. likewise noted that misuse of opioids is associated (in adjusted analysis which considered illicit substance use and depression, among other covariates) with decreased adherence to medication therapy among individuals with Human Immunodeficiency Virus (HIV) infection.[17]

Therefore, the objective of the current study was to examine the relationship between opioid use and IST adherence among adult RTRs during the first year posttransplant. The study hypothesis is that posttransplant opioid usage, regardless of dosage/amount, is associated with risk of IST nonadherence in the first year after transplantation.

Materials and Methods

Data Source and Cohort Definition

The study analyzed longitudinal data of RTRs from the Transition of Care in CKD (TCCKD) study, a retrospective cohort study examining US veterans with late-stage non-dialysis dependent chronic kidney disease (NDD-CKD) transitioning to renal replacement therapy from October 1, 2007 through March 31, 2015.[18,19] This dataset was utilized as it included Veterans Affairs (VA) records, facilitating access to clinical data and thus enabling use of propensity score methods and logistic regression modeling (described later). A total of 102,477 US veterans were identified from the US Renal Data System (USRDS) as a source population. Only individuals who received preemptive renal transplantation or transitioned to receive dialysis therapy then subsequently received renal transplantation were included in the study. The algorithm for the cohort definition is shown in Figure 1. Patients who were never transplanted (n=97,220) were excluded as were those without at least two prescriptions of tacrolimus within a year of transplant (n=4,028), which resulted in a study population of 1,229 RTRs. Among the excluded transplant patients, 3802 (94.4%) received their renal transplant at a transplant center other than the VA, while 28 (0.7%) patients had their transplant performed at the VA. Transplant center information was missing for 198 (4.9%) excluded patients. Out of 1,229 patients in the study population, 323 (26.3%) had their transplant performed at a VA center, 848 (69%) had a transplant at a center other than the VA, and 58 patients (4.7%) had missing information for the center. From these 1,229 patients, 258 RTRs did not use an opioid any time before or after transplantation, while 971 opioid users were identified within the first year after transplantation.

Figure 1:

Flow chart of the renal transplant recipients’ selection for study inclusion.

Exposure Variable

Medication data was collected from both Centers for Medicare and Medicaid Services (CMS) Data (Medicare Part D) and VA pharmacy dispensation records. Medication data (VA & CMS) included refill data from mail order pharmacy prescriptions. Dose of opioid prescription after transplantation was collected and divided based on Opioid Annual Morphine Milligram Equivalent (AMME) within a year of transplant.[13,20,21] The following exposure categories were defined: (a) any opioid use versus no opioid use; (b) 1–30 analysis: 1–30 AMME dose vs. no opioid use; (c) 31–60 analysis: 31–60 AMME dose vs. no opioid use; and (d) >60 analysis: >60 AMME dose vs. no opioid use.

Covariates

Data from the USRDS Patient and Medical Evidence Form were used to determine RTRs’ baseline demographic characteristics at the time of renal transplantation and type of renal transplant center (VA or external to the VA). Information on comorbidities such as depression, diabetes, and hypertension, among others, within the 12 months pre-transplantation period were extracted from VA Inpatient and Outpatient Medical SAS Datasets, using ICD-9-CM Diagnostic and Current Procedural Terminology codes, as well as from VA/CMS data (MedPAR, Outpatient, Carrier and Inpatient files) and were used as baseline covariates. Pre-transplant medication was used as a baseline covariate, and data were collected from both CMS (Medicare Part D) and VA pharmacy dispensation records. RTRs who received at least one dispensation of medication within the 12 months pre-transplantation period were recorded as having been treated with these medications. Laboratory data were obtained from VA research databases, as previously described, and baseline values were defined as the average of each covariate during the 12 months pre-transplantation period.[22–24]

Outcome Assessment

Detailed information about each tacrolimus prescription was collected during the first year after renal transplantation using both CMS data (Medicare Part D) and VA pharmacy dispensation records. Although not a direct measure of medication adherence, prescription dispensation (refill) records have demonstrated relationships with health outcomes in previous adherence studies among transplant recipients.[25–27] Prescriptions outside CMS and VA pharmacy dispensation records were not available. Proportion of days covered (PDC) was calculated for use of outpatient tacrolimus prescriptions. A detailed description of PDC has been published previously.[19,24,28–30] Briefly, PDC was defined as the proportion of days when the drug was available in the measurement period, capped at 100%.[28,29] The index tacrolimus date was the date of the first available prescription after transplantation. The last prescription had to be dispensed before the first-year transplantation anniversary, and the full prescription period was included in the denominator, regardless of whether the supply lasted until after the date of the first-year transplantation anniversary. Only outpatient prescriptions were considered. Similar to previous studies, PDC greater than or equal to 80% indicated adherence to tacrolimus and less than 80% indicated nonadherence.[16,31,32]

Statistical Analyses

Baseline patient characteristics were summarized according to opioid exposure (yes vs. no) and AMME dose and presented as percent for categorical variables and mean ± standard deviation (SD) or median and interquartile range (IQR) for continuous variables. Differences between opioid groups (yes vs. no) were assessed using standardized differences.[33]

To examine the association of opioid use with tacrolimus adherence, we used the following approaches: (a) logistic regression model to examine unadjusted and multivariable adjusted association between various opioid exposures and tacrolimus adherence; (b) inverse probability of treatment weighting (IPTW); (c) propensity score matching (PSM); and (d) propensity score adjustment (PSA).

The association of opioid exposure (yes vs. no) and AMME dose with tacrolimus adherence within the first year of transplant was examined using logistic regression in unadjusted and adjusted analyses. The following potential confounders were included in the multivariable adjusted models based on theoretical considerations in accordance with previous results from the literature [34]: Model 1: unadjusted; Model 2: demographic characteristics (age, gender, race, marital status, transplant center type and service connected disability [disability that is incurred or aggravated in the line of duty during active military service]); Model 3: model 2 variables plus comorbidities (chronic pulmonary disease, peptic ulcer disease, paraplegia & hemiplegia, AIDS/HIV, diabetes, cancer, liver disease, atrial fibrillation, depression and hypertension), Charlson Comorbidity Index (CCI), donor type, and smoking status; Model 4: model 3 variables plus medications (Vitamin D Active, sevelamer, lanthanum, calcium acetate, anticoagulants, aspirin, β-blockers, alpha blockers, calcium channel blockers, statins, vasodilators, loop diuretics, renin-angiotensin-aldosterone system inhibitors and insulin) and renal replacement therapy modality; and Model 5: model 4 variables and dialysis vintage, systolic and diastolic blood pressure and body mass index.

For IPTW analyses, we used logistic regression to calculate propensity scores based on all demographic and comorbidity variables, which were then used directly as stabilized inverse weights to estimate average treatment effects. In order to handle extreme weights, we used stabilized weights, which have a mean weight close to 1 and maximum weight less than 10.[35] To address stabilized weights greater than 9, we recoded the weights to 9. Stabilized inverse probability treatment weight estimates the effect in the entire cohort as opposed to the subset of a cohort in the case of PSM. In PSM analysis, we used 1:1 greedy matching using propensity scores calculated from the same demographic and comorbidity characteristics used for IPTW (Table 1 and Supplemental Tables 1 and 3).[35] To capitalize on data from the entire cohort, in addition to PSM we also applied a PSA analysis by adjusting the association between opioid exposure for individual propensity scores in a logistic regression model.

Table 1:

Baseline characteristics of the study population

	All	No opioid use	Opioid use	Missing	Standardized differences
	(N=1,229)	(N=258)	(N=971)	N (%)	Before IPTW	After IPTW
Demographics
Mean age at transplant (SD), years	60 (10)	61 (12)	60 (10)		−0.057	−0.191
Males, n (%)	1,187 (97)	252 (98)	935 (96)		−0.081	−0.041
Race, n (%)					0.161	0.069
Whites	821 (67)	187 (72)	634 (65)
African Americans	346 (28)	59 (23)	287 (30)
Other	62 (5)	12 (5)	50 (5)
Transplant Center, n (%)					1.019	0.839
VA	323 (26)	0 (0)	323 (33)
Not a VA center	848 (69)	250 (97)	598 (62)
Missing	58 (5)	8 (3)	50 (5)
Donor type, n (%)				18 (1.5)	0.257	0.068
Deceased	877 (71)	155 (60)	722 (74)
Living	334 (27)	91 (35)	243 (25)
Comorbidities, n (%)				18 (1.5 )
Chronic Pulmonary Disease	278 (23)	54 (21)	224 (23)		0.030	−0.130
Liver Disease	374 (30)	78 (30)	296 (30)		−0.022	−0.143
Diabetes	747 (61)	127 (49)	620 (64)		0.258	−0.019
Cancer	237 (19)	49 (19)	188 (19)		−0.011	−0.098
Paraplegia and Hemiplegia	58 (5)	9 (3)	49 (50)		0.070	−0.030
Atrial Fibrillation	156 (13)	33 (13)	123 (13)		−0.020	−0.159
Depression	277 (23)	51 (20)	226 (23)		0.065	−0.118
Hypertension	1,051 (86)	202 (78)	849 (87)		0.165	−0.055
Median CCI (IQR)	3 (1 −5)	2 (1 −5)	3 (1–6)		0.101	−0.208
Smoker, n (%)				204 (17)	0.603	0.171
Never	411 (33)	62 (24)	349 (36)
Current	307 (25)	48 (19)	259 (27)
Past smoker	307 (25)	57 (22)	250 (26)
RRT modality, n (%)				18 (1.5)	0.349	0.087
Hemodialysis	813 (66)	136 (53)	677 (70)
Peritoneal Dialysis	149 (12)	32 (12)	117 (12)
Transplant	219 (18)	70 (27)	149 (15)
Uncertain Dialysis	30 (2)	8 (3)	22 (2)
Median dialysis duration (IQR), days	1,954 (1,390–2,438)	1,840 (1,203–2,386)	1,967 (1,456–2,450)		0.202	0.091

CCI, Charlson Comorbidity Index; IPTW, inverse probability of treatment weighting; IQR, interquartile range; RRT, renal replacement therapy; SD, standard deviation; VA, Veterans Affairs

Sensitivity analyses were conducted to evaluate the robustness of the main findings. More than one thousand (n=1,068 [87%]) patients had complete data for analysis in the final model (Model 5). Therefore, this model was regarded as the main multivariable adjusted model without replacing missing data; however, due to the moderately high proportion of missingness (13%), missing covariates were imputed using multiple imputations as sensitivity analysis for multivariable adjusted models and IPTW model.[36]

The P values were two-sided and reported as significant at <0.05 for all analyses. All analyses were conducted using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA) and STATA/MP Version 15 (STATA Corporation, College Station, TX). The study was approved by the Institutional Review Board of the Memphis VA Medical Center (protocol number 555872), with exemption from informed consent.

Results

Baseline Characteristics

The mean age of the cohort at baseline was 60±10 years, 97% were male, 67% were white, 28% were African American, 66% were married, 61% were diabetic, and 18% of transplants were preemptive. In the entire cohort, 258 and 971 RTRs without and with opioid usage were identified, respectively. Of 1,229 RTRs included in the study, 52 (4.2%) were classified as nonadherent based on PDC <80%, while 1,177 (95.8%) were adherent. Based on this categorization, with a power of 0.8 in combination with α of 0.05, it would be feasible to detect >25% difference in outcome.[37] Baseline characteristics of RTRs categorized by opioid exposure (yes vs. no) and by AMME opioid dose are shown in Table 1 and Supplementary Tables 1 and 2. In the original cohort (n=1,229), RTRs with increasing dose of AMME were more likely to be African American, a current smoker, and had higher prevalence of several comorbidities (Table 1 and Supplementary Tables 1 and 2). These differences narrowed after IPTW, indicating reduced differences between groups (Table 1 and Supplementary Table 1). A similar narrowing of differences was also noted after propensity score matching (Supplementary Table 3).

Tacrolimus Nonadherence

Compared to RTRs without opioid usage, RTRs with any opioid usage had a lower probability of being adherent to tacrolimus in unadjusted (N=1,229, i.e., 52 had PDC <80% [nonadherent] and 1,177 had PDC ≥80% [adherent]) logistic regression model [OR (95% CI): 0.22 (0.07–0.72)] and adjusted (N=1,068, i.e., 47 had PDC <80% and 1,021 had PDC ≥80%) logistic regression model [OR (95% CI): 0.11 (0.03–0.44)], as shown in Model 1 and Model 5 in Figure 2. Similar results were found in the comparison of RTRs without opioid usage versus RTRs who had AMME opioid dose of 1–30 [OR (95% CI): 0.17 (0.05–0.57)], 31–60 [OR (95% CI): 0.21 (0.06–0.77)], and >60 [OR (95% CI): 0.18 (0.05–0.6)] in adjusted logistic regression models (Model 5 in Figure 2). Similar results were also found after multiple imputation (data not shown).

Figure 2:

Association of opioid exposure overall and by Annual Morphine Milligram Equivalent (AMME) dose within a year of transplant with adherence to tacrolimus observed in the same period (measured as proportion of days covered [PDC] ≥80%) in unadjusted (Model 1) models in 1,229 US veteran renal transplant recipients and adjusted (Models 2–5) logistic regression models in 1,068 US veteran renal transplant recipients.Overall: Any opioid use vs. No opioid use; 1–30: 1–30 AMME dose vs. No opioid use; 31–60: 31–60 AMME dose vs. No opioid use; >60: >60 AMME dose vs. No opioid use; Reference: No opioid use.Model 1: unadjusted; Model 2: demographic characteristics (age, gender, race, marital status, transplant center type and service connected disability); Model 3: model 2 variables plus comorbidities (chronic pulmonary disease, peptic ulcer disease, paraplegia & hemiplegia, AIDS/HIV, diabetes, cancer, liver disease, atrial fibrillation, depression and hypertension), Charlson Comorbidity Index (CCI), donor type, and smoking status; Model 4: model 3 variables plus medications (Vitamin D Active, sevelamer, lanthanum, calcium acetate, anticoagulants, aspirin, β-blockers, alpha blockers, calcium channel blockers, statins, vasodilators, loop diuretics, renin-angiotensin-aldosterone system inhibitors and insulin) and renal replacement therapy modality; and Model 5: model 4 variables and dialysis vintage, systolic and diastolic blood pressure and body mass index.

In the PSA analysis, opioid use (vs. no opioid use) was associated with significantly lower tacrolimus adherence [OR (95% CI): 0.32 (0.12–0.88)], with similar findings in patients exposed to AMME doses of 1–30 [OR (95% CI): 0.28 (0.09–0.87)] and 31–60 [OR (95% CI): 0.16 (0.05–0.55)] (Figure 3). The results of the IPTW analysis were similar (Figure 3). In the PSM analysis, RTRs with any opioid usage (vs. no opioid usage) showed a trend toward lower probability of being adherent to tacrolimus [OR (95% CI): 0.37 (0.13–1.06), p=0.06], with similar point estimates for the OR in the three categories of AMME dose, although only AMME dose of 31–60 showed a statistically significant association (Figure 3). Similar results were also obtained with multiple imputation analysis (data not shown).

Figure 3:

Association of opioid exposure overall and by Annual Morphine Milligram Equivalent (AMME) dose within a year of transplant with adherence to tacrolimus observed in the same period (measured as proportion of days covered [PDC] ≥80%) in propensity score (PS)-adjusted, inverse probability treatment weighted (IPTW) and propensity score-matched analyses, . Overall: Any opioid use vs. No opioid use; 1–30: 1–30 AMME dose vs. No opioid use; 31–60: 31–60 AMME dose vs. No opioid use; >60: >60 AMME dose vs. No opioid use; Reference: No opioid use.

Discussion

The purpose of the current study was to examine the association between opioid use and IST (i.e., tacrolimus) adherence among RTRs during the first year posttransplant. The findings support our hypothesis and indicate that RTRs who have any opioid usage in the first year posttransplant are less likely to be adherent to tacrolimus. Furthermore, RTRs who use opioids at any given dosage level (AMME of 1–30, 30–60, or >60) are less likely to be adherent to tacrolimus. With a few exceptions, these patterns generally remain consistent in both unadjusted and adjusted main analyses, as well as sensitivity analyses. The findings of this study are consistent with Surbhi et al. who noted that risk of medication nonadherence to treat chronic diseases, such as hypertension and diabetes, increased with opioid use among a sample of Medicare patients classified as high-cost healthcare users, or “superutilizers.”[16]

The association between any posttransplant opioid use and IST nonadherence is worrisome given that IST nonadherence increases the risk of graft failure and loss among RTRs.[3,8] Several barriers or challenges to IST adherence have been identified including negative health/medication beliefs, socioeconomic factors, psychosocial issues (e.g., depression), and tobacco and alcohol use.[4,9] The most significant finding of the current study reveals that post-transplant opioid use, regardless of dosage, may be an additional challenge to IST adherence, regardless of opioid dosage/amount. We postulate that there are several factors contributing to the relationship between opioid use and IST nonadherence. Cost of medications may be one such factor. If RTRs have limited funds available to obtain medications, they may prioritize one prescription over another (note, veterans with nonservice-connected conditions must pay copayments for outpatient medications up to the annual cap, unless they meet criteria for exemption).[38] As suggested by Surbhi et al., in the choice between medications, RTRs may favor the medication that provides more obvious or immediate symptom/pain relief (opioid) over the medication whose effects are not directly or immediately noticeable (tacrolimus).[16] Likewise, Massey and colleagues found that the perceived necessity of IST decreased significantly over the first 18 months post-renal transplant.[39,40] Because IST is not associated with direct physical symptoms, in the same way opioids are associated with alleviation of pain, RTRs may experience greater lapses in IST adherence. Opioid use to treat pain (whether resulting from posttransplant surgery complications or comorbid health conditions) may also be a proxy for poorer perceived health, which is associated with increased IST nonadherence.[4] Another possible factor is that side effects attributable to tacrolimus may be more challenging for RTRs to cope with compared to opioid-related side effects, resulting in nonadherence to IST. Additionally, Dew et al. found that poorer social support was associated with increased IST nonadherence; in our study, a greater percentage of RTRs in the opioid use group were not married compared to RTRs in the nonopioid use group, which may be indicative of decreased social support.[4] However, additional studies are needed to better understand the underlying causes of the association between opioid use and IST nonadherence.

To address the issue of IST nonadherence in the context of opioid use, several interventional strategies and tools are available to healthcare providers. Routine, proactive monitoring of IST adherence may be limited, with nonadherence often identified in hindsight following a negative health outcome such as graft rejection. Therefore, we suggest that providers closely monitor IST adherence among opioid users through such mechanisms as patient self-report, pill counts, and electronic monitoring systems, as monitoring may prompt early interventions when nonadherent behaviors become apparent.[41] Myaskovsky et al. supported implementing ongoing education for RTRs regarding the importance of IST and its role in maintaining graft function and survival, in conjunction with other interventions to promote adherence.[41] When patients are also receiving opioids, such education should be expanded to explain the role of opioids in treatment, differentiate the roles of opioids and IST, and emphasize the function and necessity of IST for graft survival. Behavioral interventions such as behavioral contracts and motivational interviewing have been used successfully to promote IST adherence in adult transplant recipients, as they are patient-centered and engage RTRs as collaborators in their own treatment.[42,43] Future studies should explore and compare the efficacy of these interventions among RTRs who do and do not use opioids. Other suggested adherence strategies such as tailoring the medication regimen or setting up medication reminder systems are described in previous reports.[41,44,45]

Since an association was noted between any posttransplant prescription opioid use and tacrolimus nonadherence, an additional intervention to bolster IST adherence is consideration of appropriate pain management strategies among RTRs including nonopioid alternatives. This is critical as evidenced not only by the increased risk of IST nonadherence found in this study, but also the increased risk of mortality and graft loss among RTRs who use opioids posttransplant.[13] In the general patient population, opioids are prescribed at a notably high rate of 58.7 per 100 persons.[46] Prior studies indicate prescription opioid rates among dialysis patients exceed the general population, and RTRs with pre-transplant prescription opioid use are likely to continue opioid use posttransplant.[11,13] Given the prevalence of opioid use in the ESRD and renal transplant populations, the guideline for opioid prescribing published by the Centers for Disease Control and Prevention (CDC) may assist providers in managing chronic pain in RTRs.[47]

The CDC guideline recommends utilizing nonpharmacologic or nonopioid pharmacologic treatment options for non-cancer and/or palliative care pain management.[47] If opioids are used, it is preferable for utilization to occur in combination with these alternative nonpharmacologic or nonopioid pharmacologic therapies.[47] In the setting of prescription opioid use, the guideline further suggests providers and patients should set realistic treatment goals for “pain and function” and have ongoing discussions of the risks and benefits of opioid treatment.[47] Among other recommendations, the CDC guideline also suggests prescribing the lowest effective dose of opioids, conducting periodic evaluation of the benefits and harms of opioid therapy in a given patient, and periodically reviewing state prescription drug monitoring systems to review patient history of controlled substance prescriptions to assist in decision-making regarding opioid prescribing/management.[47,48] Additional recommended strategies include educating and counseling patients concerning appropriate pain management, screening patients for risk of opioid abuse using instruments such as the Opioid Risk Tool, and referring patients to addiction treatment specialists as needed, among other strategies.[49] RTRs, particularly those who use high levels of opioids chronically, would likely also benefit from education about the use of opioid overdose reversal agents such as naloxone.[49]

This study highlights the association between prescription opioid use at any level and tacrolimus nonadherence among adult RTRs. We focused on 1-year post-transplant outcome as these data serve as the basis for assessment of US centers by UNOS and CMS. However, there are some limitations. Prescription dispensation records were used to determine PDC as the adherence measure. Although an objective measure, prescription dispensation records do not directly measure actual consumption of medication. However, IST adherence as calculated based on these records has been significantly associated with health outcomes among transplant recipients in prior studies, which provides evidence supporting the use of prescription dispensation records in determining adherence.[25–27] Because the sample was selected from the VA patient population, demographics differ somewhat from the national RTR population. For example, our sample was greater than 95% male, whereas the national adult RTR population is approximately 61% male. The lack of prescription data outside CMS and VA prescription records is a limitation and makes the results less generalizable to other populations.

Nevertheless, this is the first published study to demonstrate the relationship between posttransplant opioid use and IST nonadherence. Thus, future studies should explore this relationship in a more generalizable population. An additional limitation is that, as this was a retrospective study, causality cannot be established and residual confounding may exist. However, we used two different propensity score methods to minimize residual confounding and the findings provide substantive evidence of a relationship between posttransplant opioid use and nonadherence to IST (tacrolimus). Also, although the study was also underpowered to detect difference of <25% in the study outcome (e.g., adherence), we found that any opioid use in the first year posttransplant is associated with nonadherence to tacrolimus.

In conclusion, RTRs who use prescription opioids during the first year posttransplant, regardless of dosage/amount, are less likely to be adherent to tacrolimus. This is noteworthy, as nonadherence to IST is associated with negative outcomes such as graft failure.[3,7,8] Future studies are needed to better understand opioid use as a risk factor for IST nonadherence.

Additionally, several interventional strategies are suggested to facilitate IST adherence among RTRs. Future studies are also needed to assess the efficacy of the suggested strategies in improving adherence and health outcomes of RTRs.