Abstract
Objective:
To define prevalence and changes in opioid use before and after liver transplant.
Background:
Opioid use in liver transplantation is poorly understood and has potential associated morbidity.
Methods:
Using a national dataset of employer-based insurance claims, we identified 1257 adults who underwent liver transplantation between 12/2009-2/2015. We categorized patients based on their duration of opioid fills over the year before and after transplant admission as opioid-naïve/no fills, chronic opioid use (≥120 day supply), and intermittent use (all other use). We calculated risk-adjusted prevalence of peritransplant opioid fills, assessed changes in opioid use after transplant, and identified correlates of persistent or increased opioid use posttransplant.
Results:
Overall, 45% of patients filled ≥1 opioid prescription in the year before transplant (35% intermittent use, 10% chronic). Posttransplant, 61% of patients filled an opioid prescription 0-2 months after discharge, and 21% filled an opioid between 10-12 months after discharge. Among previously opioid-naïve patients, 4% developed chronic use posttransplant. Among patients with pretransplant opioid use, 84% remained intermittent or increased to chronic use, and 73% of chronic users remained chronic users after transplant. Pretransplant opioid use (risk factor) and hepatobiliary malignancy (protective) were the only factors independently associated with risk of persistent or increased posttransplant opioid use.
Conclusion:
Prescription opioid use is common before and after liver transplant, with intermittent and chronic use largely persisting, and a small development of new chronic use posttransplant. To minimize the morbidity of long-term opioid use, it is critical to improve pain management and optimize opioid use before and after liver transplant.
Mini-abstract
We studied prevalence and changes in opioid use before and after liver transplant. Prescription opioid use was common in the year before (45%) and after (68%) transplant. Preoperative intermittent and chronic opioid use largely persisted after transplant, and 4% of patients developed new chronic use.
Introduction
The epidemic of prescription opioid misuse in the United States is a public health priority, and the implications of this problem on surgical care are increasingly recognized.1-4 Opioid use before surgery is a common risk factor associated with poor outcomes and high healthcare utilization.1,2,5 Postoperatively, opioid prescribing is a critical focus for health policy, as overprescribing is common and new persistent postoperative opioid use affects 6-10% of surgical patients.3,4,6-9 In liver transplantation, the impact of this public health crisis in not well characterized.
The transplant community has seen rising opioid overdose deaths contribute to an increase in organs available for donation,10,11 but the implications of opioid use in the liver transplant recipient are unclear. Prevalence of pain and opioid use is high in chronic liver disease yet is poorly understood, and psychiatric and medical comorbidities serve as additional risk factors in these patients.12,13 Postliver transplant, recipients can face a painful and prolonged recovery.14 Thus, this population is particularly vulnerable to opioid misuse, and understanding the extent of opioid use in liver transplant recipients is a priority. Furthermore, the overall change in prescription opioid use following liver transplantation has been understudied to date. In clinical practice, it appears that some patients improve their opioid use posttransplant while others worsen, but these variable clinical trajectories are not yet understood.
In this context, we studied the prevalence of opioid fills, as well as changes in duration of opioid use, over the year before and after liver transplant using a national dataset of commercial insurance claims. We sought to identify changes in chronicity of opioid use after transplant among patients with and without pretransplant opioid use, and to identify factors correlated with such changes.
Methods
Study Population
We utilized the Truven Health MarketScan® Research Databases, including the MarketScan Commercial Claims and Encounters Database and the Medicare Supplemental and Coordination of Benefits Database. This is an insurance claims database that captures patient-level utilization of medical services and prescription drugs across the United States. This dataset represents active employees, early retirees, retirees with employer provided Medicare supplemental plans, and their dependents. Thus, this dataset comprises a commercially insured population which does includes only a portion of liver transplants performed in the United States. Medicare patients other than those with the employer provided Medicare supplemental, and Medicaid patients, are not included. We identified a cohort of patients 18 and older who underwent liver transplantation between December 2009 and February 2015, as identified by Current Procedural Terminology (CPT) codes 47135 and 47136. The cohort was limited to patients who had continuous enrollment in their plan for 1 year before and after surgery (and therefore survived at least 1 year posttransplant), to enable us to capture opioid use during this timeframe. To achieve a more homogenous cohort with respect to postoperative recovery, patients who additionally had a CPT code for kidney transplantation (50360, 50365) during the same admission were excluded. The University of Michigan Institutional Review Board deemed this study exempt from review, and informed consent was waived, as all patient data is de-identified in this dataset.
Prevalence of opioid fills before and after liver transplant
Opioid fills were identified from outpatient pharmacy claims throughout 1 year prior to transplant hospital admission and 1 year following posttransplant hospital discharge. Our primary outcome of interest was prevalence of opioid fills over the year before and after liver transplant. We identified cross-sectional (period) prevalence during the following two-month time periods before transplant admission and after discharge: 0 to 2 months, 2 to 4 months, 4 to 6 months, 6 to 8 months, 8 to 10 months, and 10 to 12 months. We first computed unadjusted prevalence, and for the posttransplant period, we also captured risk-adjusted prevalence by accounting for pretransplant sociodemographic, medical, and liver disease-related factors. To further stratify patients based on their duration of preoperative opioid use, we classified patients as preoperatively opioid-naïve (no opioid fills between 31 and 365 days before transplant admission), chronic users (≥120 days supplied between 31 and 365 days before transplant admission, or 3 or more prescriptions filled in the 3 consecutive months before transplant admission), and intermittent users (duration less than chronic). These definitions are consistent with previous literature.9,15,16 We then computed risk-adjusted prevalence of posttransplant opioid fills stratified by duration of preoperative opioid use.
As these were outpatient claims only, we did not capture opioid use during the initial posttransplant hospital stay. Additionally, for patients who were discharged to a nursing or rehabilitation facility, opioid use during the rehabilitation stay may not be captured. To account for this when computing postdischarge opioid use prevalence, we adjusted for discharge disposition in the multivariate models.
Changes in duration of opioid use after transplant
Our secondary outcome was to evaluate changes in the duration of opioid use posttransplant. To assess this, patients were first grouped based on pretransplant duration of use as defined above. To quantify duration of posttransplant opioid use, patients were classified again into the following groups: no opioid fills, intermittent users, and chronic users in the year after discharge. This classification was done using the same definitions used for classifying preoperative duration of opioid fills. We compared the pre and posttransplant duration of opioid fills to identify patients whose opioid use was persistent or increased after transplantation.
Correlates of persistent or increased opioid use after transplant
To identify factors associated with the changes in duration of opioid use after transplant, we created a binary variable indicating persistent or increased duration of opioid use after transplant. Therefore, this outcome was met if: a) a previously opioid-naïve patient filled opioid prescriptions after discharge; b) a pretransplant intermittent user met criteria for intermittent or chronic use in the year after posttransplant discharge; c) a preoperative chronic user remained a chronic user in the year after discharge.
Patient factors
Patient sociodemographic and clinical factors were obtained from insurance claims over the year prior to transplant. Sociodemographic factors included gender, age, household income (based on metropolitan statistical area, a measure of median household income for patient’s zip code), geographic region, and insurance plan type (comprehensive health insurance, health maintenance organization, preferred provider organization, point of service, and other). Patient medical comorbidities were identified by International Classification of Disease 9 (ICD-9) code and classified using the Elixhauser comorbidity index,17 which shows similar risk-adjustment performance among patients with cirrhosis compared to the Model for End-Stage Liver Disease-Sodium (MELD-Na) and the Child-Turcotte-Pugh score.18 Patient psychiatric and pain diagnoses were also identified by ICD-9 code, using methodology previously published.3 Psychiatric diagnoses included depression, anxiety, substance use disorder, and other diagnosis (e.g. other mood or thought disorders). Pain diagnoses included arthritis, back pain, and neck pain. For analysis, due to association between the psychiatric comorbidities (e.g. patients with multiple diagnoses), we created 1 composite binary variable indicating presence of any psychiatric diagnosis. Similarly, many patients had more than 1 pain diagnosis, and thus a composite binary measure indicating any pain diagnosis was used in the regression analyses. Etiology of liver disease was determined from ICD-9 code19-21 and classified as alcohol-related (571.2, 571.0, 571.1X, 571.3), hepatitis C (070.41, 070.44, 070.51, 070.54, V02.62, 070.70, 070.71), and hepatobiliary malignancy (155.0, 155.1, 155.2, 156.1, 156.9). We were unable to determine primary etiology of liver disease from insurance claims. Thus, for analysis, etiology of liver disease was coded into mutually exclusive categories as hepatitis C virus only, alcohol-related only, both hepatitis C virus and alcohol-related, and other etiology. The “other etiology” category included all remaining patients who underwent liver transplant but did not have any of the aforementioned ICD-9 codes for alcohol or hepatitis C; this would include, for example, nonalcoholic steatohepatitis and biliary etiologies, among others. Hepatobiliary malignancy was accounted for with a distinct variable, as this was not mutually exclusive with other etiologies of liver disease. We also accounted for hospital length of stay and discharge destination (home vs. discharge to nursing or rehabilitation facility).
Statistical Analysis
Descriptive statistics were used to characterize the attributes of the study sample, presented with frequency tables or as mean ± standard deviation (SD). Prevalence of opioid fills over the year before and after transplant was first presented as unadjusted prevalence. Prevalence of posttransplant opioid fills, by duration of pretransplant opioid use, was then reported as risk-adjusted prevalence by using a separate multivariate logistic regression model for each prevalence time-period, adjusting for the aforementioned sociodemographic and clinical factors. We then tabulated pretransplant by posttransplant duration of opioid fills to characterize changes in opioid use before and after transplant. To identify factors independently associated with persistent or increased duration of posttransplant opioid use, we used multivariable logistic regression, including all clinical and sociodemographic variables in the model (without employing any automated variable selection techniques). All significance tests were two-sided with a significance level of P<0.05. All statistical analyses were performed using Stata v13.1 (StataCorp LP, College Station, TX, USA).
Results
Study cohort
A total of 1,340 adult patients with continuous insurance enrollment received a liver transplant during the study period, and 83 (6%) also had a concurrent kidney transplant and were excluded. Our final study cohort therefore included 1,257 patients. Patient characteristics are shown in Table 1 stratified by preoperative opioid use. The majority of the patients were male (68%). Etiology contributing to liver disease included hepatitis C only (24%), alcohol-related only (21%), and both hepatitis C and alcohol-related (18%). A hepatobiliary malignancy was present in 44% of patients. Compared to opioid-naïve patients, patients with chronic opioid use prior to transplant were significantly younger, had more medical comorbidities, were more likely to have 1 or more psychiatric comorbidity, had more pain diagnoses, and had different liver disease etiologies (all P<0.05).
Table 1:
Patient characteristics
| Preoperative Opioid Use | ||||
|---|---|---|---|---|
| Total | Opioid-Naïve | Intermittent Use | Chronic Use | |
| Patient Characteristic | No. (%) or mean (standard deviation) | |||
| Total No. Patients | 1,257 | 687 (54.7%) | 441 (35.1%) | 129 (10.3%) |
| Male | 848 (67.5%) | 460 (67.0%) | 305 (69.2%) | 83 (64.3%) |
| Age | 54.7 (9.9) | 54.7 (9.8) | 55.3 (9.9) | 52.8 (10.0)* |
| Income | ||||
| <$40k | 4 (0.3%) | 3 (0.4%) | 1 (0.2%) | 0 (0.0%) |
| $40k-$50k | 25 (2.0%) | 12 (1.8%) | 11 (2.5%) | 2 (1.6%) |
| $50k-$60k | 204 (16.2%) | 115 (16.7%) | 70 (15.9%) | 19 (14.7%) |
| $60k-$70k | 483 (38.4%) | 256 (37.3%) | 172 (39.0%) | 55 (42.6%) |
| ≥$70k | 376 (29.9%) | 209 (30.4%) | 133 (30.2%) | 34 (26.4%) |
| Missing | 165 (13.1%) | 92 (13.4%) | 54 (12.2%) | 19 (14.7%) |
| Region | * | |||
| Northeast | 268 (21.3%) | 167 (24.3%) | 76 (17.2%) | 25 (19.4%) |
| North Central | 238 (18.9%) | 122 (17.8%) | 93 (21.1%) | 23 (17.8%) |
| South | 529 (42.1%) | 264 (38.4%) | 207 (46.9%) | 58 (45.0%) |
| West | 214 (17.0%) | 127 (18.5%) | 64 (14.5%) | 23 (17.8%) |
| Unknown | 8 (0.6%) | 7 (1.0%) | 1 (0.2%) | 0 (0.0%) |
| Insurance plan | ||||
| Preferred Provider Organization | 752 (59.8) | 423 (61.6%) | 258 (58.5%) | 71 (55.0%) |
| Comprehensive | 76 (6.1%) | 34 (5.0%) | 32 (7.3%) | 10 (7.8%) |
| Health Maintenance Organization | 125 (9.9%) | 64 (9.3%) | 48 (10.9%) | 13 (10.1%) |
| Point of Service | 118 (9.4%) | 60 (8.7%) | 44 (10.0%) | 14 (10.9%) |
| Other | 105 (8.4%) | 56 (8.2%) | 41 (9.3%) | 8 (6.2%) |
| Missing | 81 (6.4%) | 50 (7.3%) | 18 (4.1%) | 13 (10.1%) |
| No. comorbidities (Elixhauser) | 8.0 (3.5) | 7.9 (3.6) | 7.9 (3.2) | 9.1 (3.3)* |
| Psychiatric comorbidity | 429 (34.1%) | 232 (33.8%) | 138 (31.3%) | 59 (45.7%)* |
| Pain diagnosis | * | |||
| Arthritis | 1002 (79.7%) | 527 (76.7%) | 356 (80.7%) | 119 (92.2%) |
| Back | 540 (43.0%) | 270 (39.3%) | 184 (41.7%) | 86 (66.7%) |
| Neck | 221 (17.6%) | 112 (16.3%) | 78 (17.7%) | 31 (24.0%) |
| Liver disease etiology | * | |||
| Hepatitis C | 304 (24.2%) | 161 (23.4%) | 107 (24.3%) | 36 (27.9%) |
| Alcohol-related | 267 (21.2%) | 162 (23.6%) | 83 (18.8%) | 22 (17.1%) |
| Both Hepatitis C and alcohol-related | 229 (18.2%) | 112 (16.3%) | 85 (19.3%) | 32 (24.8%) |
| Other | 457 (36.4%) | 252 (36.7%) | 166 (37.6%) | 39 (30.2%) |
| Hepatobiliary malignancy | 558 (44.4%) | 262 (38.1%) | 241 (54.7%)* | 55 (42.6%) |
| Hospital Length of stay | 18.0 (20.1) | 18.6 (20.1) | 16.2 (18.9)* | 21.0 (23.9) |
| Discharge destination | ||||
| Home | 980 (78.0%) | 527 (76.7%) | 352 (79.8%) | 101 (78.3%) |
| Facility | 161 (12.8%) | 98 (14.3%) | 45 (10.2%) | 18 (14.0%) |
| Missing | 116 (9.2%) | 62 (9.0%) | 44 (10.0%) | 10 (7.8%) |
Indicates statistical significance (P<0.05) when compared to the opioid-naïve group. For categorical variables with multiple levels, this indicates that the distribution is significantly different compared to that of opioid naïve patients (Chi-Squared test).
Types of opioid prescriptions filled
The most common opioid prescribed prior to transplant was oxycodone, accounting for 40% of all opioid prescriptions. Among patients who filled 1 or more opioid prescription in the year before transplant, 55% filled an oxycodone prescription, 54% hydrocodone, 13% hydromorphone, 5% codeine, and 2% of patients filled a methadone prescription. The most common opioid prescribed posttransplant remained oxycodone, accounting for 52% of all opioid prescriptions. Among patients who filled 1 or more opioid prescription in the year after transplant, 71% filled an oxycodone prescription, 46% hydrocodone, 13% hydromorphone, 4% codeine, and 3% of patients filled a methadone prescription.
Prevalence of opioid fills before and after transplant
The unadjusted prevalence of opioid fills in the year before and after transplant admission is shown in Figure 1. Overall, 45% of patients filled 1 or more opioid prescription between 31 and 365 days prior to transplant admission. Prevalence slightly increased over the months leading up to transplant, ranging from 15-21%. After posttransplant hospital discharge, 61% of patients filled an opioid prescription within the first 2 months, and prevalence of opioid fills decreased thereafter and remained between 19% and 25% between months 2 and 12 postdischarge.
Figure 1: Unadjusted prevalence of opioid fills over the year before and after liver transplant hospital stay.
Numbers represent two-month period prevalence. For example, in the first 2 months after hospital discharge, 61% of patients filled 1 or more opioid prescription.
Postdischarge prevalence of opioid fills, by pretransplant opioid use
Preoperatively, 55% of patients were opioid-naïve, 35% were intermittent opioid users, and 10% were chronic users. Figure 2 shows the risk-adjusted prevalence of opioid fills in the year following discharge, stratified by preoperative opioid use status. Among preoperatively opioid-naïve patients, 45% filled 1 or more opioid prescription in the first 2 months postdischarge, and this then leveled out to 9-12% between 2 and 12 months postdischarge. Prevalence of postdischarge opioid fills at all time-points was significantly higher in preoperatively intermittent and chronic opioid users compared to opioid naïve. Among preoperatively chronic opioid users, 91% filled 1 or more opioid prescription within the first 2 months postdischarge, and this decreased to 62% between months 10 and 12.
Figure 2: Adjusted prevalence of opioid fills posttransplant, by pretransplant duration of opioid use.
Numbers are risk-adjusted for patient clinical and sociodemographic factors and represent two-month period prevalence estimates. For example, among patients who were chronic opioid users before surgery, 91% filled 1 or more opioid prescription within the first 2 months after hospital discharge.
Changes in duration of opioid use after transplant
Table 2 displays posttransplant duration of opioid fills relative to pretransplant duration of fills. Among preoperative opioid-naïve patients, 48% filled no opioid prescriptions posttransplant, 48% were classified as intermittent users after transplant, and 4% became chronic users. Among preoperative intermittent users, most (71%) remained intermittent users after transplant. Among preoperative chronic users, most (73%) remained chronic users after transplant. There was a statistically significant shift in prevalence of opioid-naïve vs. intermittent vs. chronic opioid use before and after transplant (Stuart-Maxwell test for marginal homogeneity P<0.001).
Table 2:
Changes in duration of opioid fills after transplant, by pre-transplant opioid fill duration.
| Postoperative opioid fill duration | |||
|---|---|---|---|
|
Preoperative opioid fill duration |
No use | Intermittent use | Chronic use |
| Opioid-naïve | 331 (48.2%) | 327 (47.6%) | 29 (4.2%) |
| Intermittent use | 72 (16.3%) | 314 (71.2%) | 55 (12.5%) |
| Chronic use | 4 (3.1%) | 31 (24.0%) | 94 (72.9%) |
Decrease Use 
stable Use 
Increase Use
This table indicates, for patients of a given preoperative opioid fill duration, the proportion of patients who filled opioids for a given duration over the year post-discharge. Row percentages are presented. The cells filled in with red indicate patients who had maintained or increased post-transplant opioid use. There was a statistically significant shift in prevalence of opioid-naïve vs. intermittent vs. chronic opioid use before and after transplant (P<0.001).
Correlates of persistent or increased opioid use after transplant
Table 3 shows factors independently associated with persistent or increased duration of use posttransplant. All variables in the table were entered into 1 multivariate logistic regression model. The C-statistic of this model was 0.72. Patients with pretransplant intermittent (OR=4.98, 95% C.I. 3.66 to 6.78, P<0.001) or chronic (OR=2.49, 95% C.I. 1.61 to 3.87, P<0.001) opioid use were more likely to have persistent or increased posttransplant opioid use compared to opioid-naïve patients, even after adjusting for the other variables. The only other factor independently associated with this outcome was diagnosis of a hepatobiliary malignancy, which was protective against persistent or increased posttransplant opioid use (OR=0.62, 95% C.I. 0.47 to 0.83, P=0.001).
Table 3:
Factors independently associated with persistent or increased posttransplant opioid use
| Univariate | Multivariate | ||||
|---|---|---|---|---|---|
| Odds Ratio (95% C.I.) |
P- value |
Adjusted Odds Ratio (95% C.I.) |
P- value |
Est. % outcome |
|
| Preoperative opioid fill duration | |||||
| Opioid-naïve (Reference) | 1 | 1 | 52% | ||
| Intermittent use | 4.77 (3.55, 6.39) | <0.001 | 4.98 (3.66, 6.78) | <0.001 | 83% |
| Chronic use | 2.50 (1.65, 3.79) | <0.001 | 2.49 (1.61, 3.87) | <0.001 | 72% |
| Sex | |||||
| Female (Reference) | 1 | 1 | 63% | ||
| Male | 1.11 (0.87, 1.42) | 0.413 | 1.16 (0.88, 1.54) | 0.294 | 66% |
| Age | 0.99 (0.98, 1.01) | 0.330 | 0.99 (0.98, 1.01) | 0.269 | N/A |
| Income | |||||
| <$40k | 1.59 (0.16, 15.39) | 0.691 | 1.74 (0.15, 19.76) | 0.655 | 77% |
| $40k-$50k | 1.36 (0.55, 3.34) | 0.504 | 0.99 (0.37, 2.67) | 0.982 | 67% |
| $50k-$60k | 1.10 (0.77, 1.59) | 0.589 | 0.86 (0.56, 1.31) | 0.472 | 64% |
| $60k-$70k | 0.96 (0.72, 1.27) | 0.752 | 0.87 (0.63, 1.20) | 0.387 | 64% |
| ≥$70k (Reference) | 1 | 1 | 67% | ||
| Missing | 0.88 (0.60, 1.28) | 0.502 | 0.85 (0.55, 1.32) | 0.474 | 64% |
| Region | |||||
| Northeast (Reference) | 1 | 1 | 65% | ||
| North Central | 1.11 (0.77, 1.58) | 0.576 | 0.79 (0.53, 1.19) | 0.255 | 60% |
| South | 1.78 (1.31, 2.43) | <0.001 | 1.36 (0.94, 1.97) | 0.104 | 70% |
| West | 1.02 (0.71, 1.47) | 0.914 | 0.81 (0.54, 1.22) | 0.308 | 60% |
| Unknown | 0.41 (0.10, 1.76) | 0.230 | 0.46 (0.09, 2.31) | 0.347 | 48% |
| Insurance plan | |||||
| Preferred Provider Organization (Reference) | 1 | 1 | 65% | ||
| Comprehensive | 1.24 (0.74, 2.07) | 0.411 | 1.30 (0.74, 2.28) | 0.367 | 70% |
| Health Maintenance Organization | 1.28 (0.85, 1.93) | 0.242 | 1.29 (0.82, 2.02) | 0.269 | 70% |
| Point of Service | 1.18 (0.78, 1.79) | 0.442 | 1.05 (0.66, 1.64) | 0.848 | 66% |
| Other | 1.23 (0.79, 1.91) | 0.364 | 1.22 (0.76, 1.96) | 0.418 | 69% |
| Missing | 0.41 (0.26, 0.65) | <0.001 | 0.40 (0.23, 0.67) | 0.001 | 46% |
| Number of comorbidities (Elixhauser) | 1.00 (0.97, 1.04) | 0.807 | 1.00 (0.96, 1.04) | 0.969 | N/A |
| Psychiatric comorbidity | |||||
| No (Reference) | 1 | 1 | 64% | ||
| Yes | 1.12 (0.88, 1.44) | 0.350 | 1.16 (0.86, 1.57) | 0.338 | 67% |
| Any pain diagnosis | |||||
| No (Reference) | 1 | 1 | 62% | ||
| Yes | 1.29 (0.94, 1.77) | 0.116 | 1.22 (0.86, 1.73) | 0.275 | 66% |
| Liver disease etiology | |||||
| Hepatitis C (Reference) | 1 | 1 | 67% | ||
| Alcohol-related | 1.07 (0.76, 1.51) | 0.700 | 0.93 (0.62, 1.39) | 0.713 | 66% |
| Both Hepatitis C and alcohol-related | 1.12 (0.78, 1.61) | 0.537 | 1.02 (0.67, 1.55) | 0.931 | 68% |
| Other | 0.91 (0.67, 1.23) | 0.553 | 0.76 (0.53, 1.08) | 0.120 | 62% |
| Hepatobiliary malignancy | |||||
| No (Reference) | 1 | 1 | 69% | ||
| Yes | 0.87 (0.69, 1.10) | 0.255 | 0.62 (0.47, 0.83) | 0.001 | 60% |
| Hospital Length of stay >21d | |||||
| No (Reference) | 1 | 1 | 66% | ||
| Yes | 0.76 (0.58, 0.99) | 0.045 | 0.78 (0.57, 1.07) | 0.130 | 61% |
| Discharge destination | |||||
| Home (Reference) | 1 | 1 | 64% | ||
| Facility | 0.80 (0.57, 1.12) | 0.197 | 1.0 (0.67, 1.49) | 0.999 | 64% |
| Missing | 1.62 (1.04, 2.52) | 0.031 | 1.66 (1.02, 2.68) | 0.040 | 74% |
This table shows the results of the univariate and multivariable logistic regression analysis for the outcome of persistent or increased posttransplant opioid use. For the multivariate model, all variables listed in the table were included in the same model. The right-most column lists the estimated incidence of the outcome, persistent or increased posttransplant opioid use, for each level of the categorical variables. This is the output of the logistic regression model, also known as the marginal probability of outcome. This is not possible to compute for the continuous variables. C.I.=confidence interval.
Discussion
Summary
In this study of outpatient national commercial insurance claims, prescription opioid fills were common in the year before and after liver transplant. After posttransplant hospital discharge, roughly one-in-five patients filled opioid prescriptions as far as 1 year after discharge. Prevalence of postdischarge opioid use was highest among patients who were chronic opioid users before transplant, and lowest among preoperatively opioid-naïve patients. Among previously opioid-naïve patients, roughly half had intermittent opioid use posttransplant, and 4% of these patients became chronic users after transplant. Among patients who used opioids prior to transplant, the majority (71-73%) maintained their pattern of opioid use after transplant. Of note, there was a substantial proportion of patients with prior opioid use who decreased their use posttransplant (27% of chronic users and 16% of intermittent users). The only clinical factors associated with persistent or increased posttransplant opioid use included preoperative opioid use (risk factor) and diagnosis of a hepatobiliary malignancy (protective).
Comparison to prior literature
Pain and opioid use have thus far been poorly studied in the chronic liver disease population. In these patients, prevalence of pain is estimated between 17 and 34%.12,22 Opioid use prevalence in this population ranges from 25-29%--a similar range to our pretransplant findings.12,23 One recent large study found that patients with cirrhosis filled opioid prescriptions at higher rates than patients with other chronic diseases.24 Pain and opioid use appear to be common in patients with chronic hepatitis C virus.23 In our study, hepatitis C diagnosis was more common in preoperative chronic opioid users, but was not associated with posttransplant opioid use. In patients with hepatocellular carcinoma, locoregional therapeutic procedures such as ablation, if performed in the year prior to transplant, may contribute to pain and warrant short courses of opioids.25 Our study did not focus specifically on hepatocellular carcinoma or any of these locoregional therapies, but this would be an interesting area for further study. Interestingly, in 1 large singlecenter study of pain in chronic liver disease, patients reported abdominal pain as the most common pain location. Abdominal pain and associated opioid use might be expected to improve after liver transplant, yet we saw persistent posttransplant use in a large proportion of patients, suggesting liver transplantation alone may not lead to significant reduction in opioid use.
In the liver transplant population, the extent of opioid use before and after transplant has not been well characterized. A recent study by Randall et al. assessed daily opioid dose in transplant recipients in relation to posttransplant outcomes but did not focus on duration or prevalence.26 Compared to our results, their study reported a lower opioid fill prevalence of 9% among waitlisted patients. These differences may be attributable to differences in patient populations, as our cohort was limited to commercially insured patients, and our pretransplant analysis was not restricted to the waitlist period. Randall et al. also reported higher doses of pretransplant daily opioids associated with higher prevalence of posttransplant opioid fills, including opioid use beyond 90 days posttransplant.26 By comparison, our study found that pretransplant chronicity of use was also associated with posttransplant opioid fills.
Optimizing pretransplant opioid use
This study highlights the need for transplant centers to develop best practices for pain and opioid management in liver transplant recipients. For patients using opioids before transplant, transplant providers should work with the patients’ other physicians to optimize their pain management regimens. A multi-disciplinary and multi-modal approach to pain management in these patients may help minimize opioid use and improve pain management before and after transplant. Additionally, preoperative opioid use is associated with other poor postoperative outcomes across a range of surgical populations.1,27,28 In liver transplant recipients, preoperative opioid use has been associated with increased readmissions, graft loss, and mortality.26,29 As such, these specific outcomes were not a focus of our paper. We did however address hospital length of stay in our analysis, primarily for the purposes of risk adjustment, but we noted that preoperative chronic opioid users had higher hospital length of stay, consistent with other surgical literature.1 Length of stay was not significantly associated with persistent or increased posttransplant opioid use. In additional to analgesic management, a multi-disciplinary approach to pain and opioid use should include consideration of the pretransplant psychosocial assessment. Psychiatric comorbidities such as mood disorders are commonly associated with pain and opioid use in chronic liver disease patients.12 In our study, psychiatric diagnoses were more common among chronic opioid users, though presence of such a diagnosis was not associated with persistent or increased posttransplant opioid use. Psychiatric evaluation, and ongoing mental health therapy and monitoring, should nevertheless be performed in patients with opioid use and concurrent psychiatric comorbidities. In the authors’ institutional protocol, patients with chronic pain with or without preoperative opioid use are referred to our transplant psychiatrist.
Optimizing posttransplant opioid use
Opioid prescribing after transplant deserves careful attention. Liver transplant recipients receive the majority of their healthcare within multidisciplinary transplant centers for the initial postoperative period. Transplant providers are thus in a unique position to optimize the postoperative pain of transplant recipients. Institutional protocols and guidelines may be useful in directing postoperative opioid prescribing, and a multi-modal approach to pain management will again be useful. More research is needed to develop evidence-based opioid prescribing recommendations in this population. However, in our institution, we prescribe 3 to 5 days’ supply of opioid pain medication on discharge (10 to 12 pills), and reassess opioid requirements at each weekly posttransplant clinic visit, with the expectation that all patients are off opioid pain medications by postoperative day 21. Setting clear expectations for pain management in the perioperative setting is critical to success.
Regarding persistent posttransplant opioid use, even among preoperatively opioid-naïve patients, 4% developed chronic opioid use after transplant. By comparison, new persistent postoperative opioid use is common in other surgical populations, including minor and major general surgery (6%)3 and curative-intent cancer surgery (10%).9 Strategies to mitigate risk of persistent postoperative use should include a focus on transitions of care. When recipients transition from the transplant center back to their primary care physician, this should be done with a clearly communicated plan for any ongoing pain management, and opioid prescriptions should be written by only 1 provider.
Correlates of posttransplant opioid use
In addition to pretransplant opioid use being predictive of postoperative use, we found that patients with hepatobiliary malignancies were less likely to have persistent or increased postoperative opioid use. This may reflect lack of decompensated liver disease in these patients compared to patients with cirrhosis. Understanding the variable opioid use patterns in these different subpopulations will help elucidate reasons underlying posttransplant persistent opioid use.
Limitations
This study has several notable limitations. This was an analysis of commercial insurance claims and thus may not be generalizable to the entire transplant population, namely those publicly insured. Additionally, this was a relatively limited sample size. Benefits of this dataset include the granularity and accuracy with which insurance claims can capture opioid fills. However, claims data are not able to identify actual patient consumption. Additionally, indication for the prescriptions cannot be assessed from claims data, and individual patient experiences that may correlate with opioid use, such as complications, could not be assessed. Furthermore, this was an analysis of outpatient claims only, and thus we did not capture opioid use during the index hospitalization, during subsequent readmissions, or during a stay at a nursing or rehabilitation facility. Our estimates of prevalence should therefore be interpreted as outpatient opioid use prevalence. We took steps to address this limitation. Though we were unable to examine the relationship of hospital course to postoperative opioid utilization, we controlled for length of stay in our analyses. We also accounted for discharge destination when computing adjusted prevalence, and this was not significantly associated with opioid fills at any time point. Lastly, different transplant centers likely have different inclusion criteria regarding listing and transplantation of patients with preoperative opioid use or dependence. For example, in a study of transplant centers, 64% and 38% of centers considered chronic opioid use and opioid maintenance therapy a relative contraindication to transplant, respectively.30 We cannot identify centers in this dataset and thus were not able to account for these differences, which may affect our estimates of opioid use prevalence.
Conclusion
In summary, opioid use is prevalent in the year before and after liver transplant and is prolonged in many patients. Most patients using opioids prior to transplant continue to fill opioids afterward, although there is a proportion of patients who reduce or stop use posttransplant. Additionally, there is a small but notable percentage of opioid-naïve transplant recipients who develop chronic opioid use after transplant. These results highlight a need to direct resources to pain and opioid management in the liver transplant population. Liver transplantation provides a precious opportunity at prolonged survival and improved quality of life, and identifying mechanisms to reduce potential morbidity related to opioid use is critical in order to optimize posttransplant outcomes.
Acknowledgments
Sources of Funding: Supported by the National Research Service Award postdoctoral fellowship (No. 5T32 CA009672-23) (J.S.L.), the National Institute on Drug Abuse (Research Project Grant No. R01 DA042859) (C.M.B., M.J.E., and J.F.W.), and the Michigan Department of Health and Human Services (C.M.B., M.J.E., and J.F.W.). The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health or the Michigan Department of Health and Human Services.
Abbreviations:
- CI
Confidence Interval
- CPT
Current Procedural Terminology
- ICD-9
International Classification of Disease 9
- MELD-Na
Model for End-Stage Liver Disease-Sodium
- OR
Odds Ratio
Footnotes
Conflicts of Interest: The authors of this manuscript have conflicts of interest to disclose. C.M.B. reports a patent for Peripheral Perineural Dexmedetomidine licensed to the University of Michigan, is a consultant for Recro Pharma and Heron Therapeutics, and has received research funding from Neuros Medical.
References
- 1.Cron DC, Englesbe MJ, Bolton CJ, et al. Preoperative Opioid Use is Independently Associated With Increased Costs and Worse Outcomes After Major Abdominal Surgery. Ann Surg. 2017;265(4):695–701. [DOI] [PubMed] [Google Scholar]
- 2.Waljee JF, Cron DC, Steiger RM, et al. Effect of Preoperative Opioid Exposure on Healthcare Utilization and Expenditures Following Elective Abdominal Surgery. Ann Surg. 2017;265(4):715–721. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Brummett CM, Waljee JF, Goesling J, et al. New persistent opioid use after minor and major surgical procedures in us adults. JAMA Surg. 2017;152(6):e170504. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hill MV, McMahon ML, Stucke RS, et al. Wide Variation and Excessive Dosage of Opioid Prescriptions for Common General Surgical Procedures. Ann Surg. 2016;265(4):709–714. [DOI] [PubMed] [Google Scholar]
- 5.Jiang X, Orton M, Feng R, et al. Chronic Opioid Usage in Surgical Patients in a Large Academic Center. Ann Surg. 2017;265(4):722–727. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Hill MV, Stucke RS, McMahon ML, et al. An Educational Intervention Decreases Opioid Prescribing After General Surgical Operations. Ann Surg. 2018;267(3):468–472. [DOI] [PubMed] [Google Scholar]
- 7.Thiels CA, Anderson SS, Ubl DS, et al. Wide Variation and Overprescription of Opioids After Elective Surgery. Ann Surg. 2017;266(4):564–573. [DOI] [PubMed] [Google Scholar]
- 8.Howard R, Waljee J, Brummett C, et al. Reduction in Opioid Prescribing Through Evidence-Based Prescribing Guidelines. JAMA Surg. 2018;153(3):285–287. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Lee JS, Hu HM, Edelman AL, et al. New Persistent Opioid Use Among Patients With Cancer After Curative-Intent Surgery. J Clin Oncol. 2017;35(36):4042–4049. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Gonzalez SA, Trotter JF. The Rise of the Opioid Epidemic and Hepatitis C Positive Organs: A New Era in Liver Transplantation. Hepatology. 2018;67(4):1600–1608. [DOI] [PubMed] [Google Scholar]
- 11.Durand CM, Bowring MG, Thomas AG, et al. The drug overdose epidemic and deceased-donor transplantation in the united states: A national registry study. Ann Intern Med. 2018;168(10):702–711. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Rogal SS, Winger D, Bielefeldt K, et al. Pain and opioid use in chronic liver disease. Dig Dis Sci. 2013;58(10):2976–2985. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Madan A, Barth KS, Balliet WE, et al. Chronic pain among liver transplant candidates. Prog Transplant. 2012;22(4):379–384. [DOI] [PubMed] [Google Scholar]
- 14.Milan Z Analgesia after liver transplantation. World J Hepatol. 2015;7(21):2331–2335. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Von Korff M, Saunders K, Thomas Ray G, et al. De facto long-term opioid therapy for non-cancer pain. Clin J Pain. 2008;24(6):521–527. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Raebel MA, Newcomer SR, Reifler LM, et al. Chronic use of opioid medications before and after bariatric surgery. JAMA. 2013;310(13):1369–1376. [DOI] [PubMed] [Google Scholar]
- 17.Elixhauser A, Steiner C, Harris DR, et al. Comorbidity measures for use with administrative data. Med Care. 1998;36(1):8–27. [DOI] [PubMed] [Google Scholar]
- 18.Xu Y, Li N, Lu M, et al. Comparison of risk adjustment methods in patients with liver disease using electronic medical record data. BMC Gastroenterol. 2017;17(1):5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Liu TL, Trogdon J, Weinberger M, et al. Diabetes Is Associated with Clinical Decompensation Events in Patients with Cirrhosis. Dig Dis Sci. 2016;61(11):3335–3345. [DOI] [PubMed] [Google Scholar]
- 20.Park H, Chen C, Wang W, et al. Chronic hepatitis C virus (HCV) increases the risk of chronic kidney disease (CKD) while effective HCV treatment decreases the incidence of CKD. Hepatology. 2018;67(2):492–504. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Nehra MS, Ma Y, Clark C, et al. Use of administrative claims data for identifying patients with cirrhosis. J Clin Gastroenterol. 2013;47(5):e50–54. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Marchesini G, Bianchi G, Amodio P, et al. Factors associated with poor health-related quality of life of patients with cirrhosis. Gastroenterology. 2001;120(1):170–178. [DOI] [PubMed] [Google Scholar]
- 23.Whitehead AJ, Dobscha SK, Morasco BJ, et al. Pain, substance use disorders and opioid analgesic prescription patterns in veterans with hepatitis C. J Pain Symptom Manage. 2008;36(1):39–45. [DOI] [PubMed] [Google Scholar]
- 24.Konerman MA, Rogers M, Kenney B, et al. Opioid and benzodiazepine prescription among patients with cirrhosis compared to other forms of chronic disease. BMJ Open Gastroenterol. 2019;6(1):e000271. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Christian-Miller N, Frenette C. Hepatocellular cancer pain: impact and management challenges. J Hepatocell Carcinoma. 2018;5:75–80. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Randall HB, Alhamad T, Schnitzler MA, et al. Survival implications of opioid use before and after liver transplantation. Liver Transpl. 2017;23(3):305–314. [DOI] [PubMed] [Google Scholar]
- 27.Lentine KL, Lam NN, Naik AS, et al. Prescription opioid use before and after kidney transplant: implications for posttransplant outcomes. Am J Transplant. 2018;18(12):2987–2999. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Lentine KL, Lam NN, Xiao H, et al. Associations of pre-transplant prescription narcotic use with clinical complications after kidney transplantation. Am J Nephrol. 2015;41(2):165–176. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Rogal S, Mankaney G, Udawatta V, et al. Association between opioid use and readmission following liver transplantation. Clin Transplant. 2016;30(10):1222–1229. [DOI] [PubMed] [Google Scholar]
- 30.Fleming JN, Lai JC, Te HS, et al. Opioid and opioid substitution therapy in liver transplant candidates: A survey of center policies and practices. Clin Transplant. 2017;31(12). doi: 10.1111/ctr.13119. [DOI] [PMC free article] [PubMed] [Google Scholar]