Abstract
Objectives
High-dose opioid use is associated with increased morbidity, mortality, and healthcare utilization. People living with HIV (PLHIV) are frequently prescribed these medications to manage their pain. However, little is known about the relationship between being prescribed high doses of opioids (> 90 MME/d) and mortality risk among this population. The objective of this study was to examine the trends in mortality and the relationship between high-dose opioid analgesic prescribing and mortality among PLHIV.
Methods
Utilizing the STOP HIV/AIDS cohort – a population-level linked database of treatment of PLHIV in British Columbia – we conducted bivariable and multivariable generalized estimating equation (GEE) models with a Poisson distribution to examine the relationship between high-dose opioid prescription and all-cause mortality rates in the study sample.
Results
Between 1996 and 2015, 9272 PLHIV were included in the study. Age- and sex-adjusted mortality rate (using the 2011 Canadian population as the reference) was 30.99 per 1000 person-years (95% confidence interval [CI]: 28.11 – 33.88). In a multivariable GEE model with adjustment for various demographic and clinical confounders, there was a positive and independent association between being prescribed high-dose opioids and all-cause mortality rates (adjusted rate ratio [ARR] = 3.01; 95%CI: 2.47 – 3.66).
Discussion
We found that mortality rates were significantly higher among PLHIV who were prescribed high-dose opioids compared to those who were prescribed lower doses. Our results highlight the risk associated with the prescribing of high-dose opioids to manage HIV-related pain and emphasize the need to explore non-opioid approaches to pain management.
INTRODUCTION
Many people living with HIV (PLHIV) experience HIV-related pain, frequently in the form of peripheral neuropathic pain and non-neuropathic pain (nociceptive pain due to tissue injury and musculoskeletal pain) (1,2). Estimates suggest the prevalence of pain among PLHIV ranges from 30 – 90% and it has been noted that this proportion increases in the later stages of HIV (2). Furthermore, PLHIV also experience comorbidities and exposures to socio-structural environments that may increase their risk of pain. For instance, the literature suggests that PLHIV are more likely to have experienced significant trauma in forms such as intimate partner violence and childhood abuse than the general population (3–5). PLHIV also have a high prevalence of psychiatric comorbidities (e.g., Post-traumatic Stress Disorder (PTSD), depression, anxiety) that may make them more vulnerable to experiencing pain (3–5).
Various pharmacological and non-pharmacological pain management modalities exist for PLHIV, including opioid and non-opioid pain relievers, adjuvant therapies, psychotherapies and physical therapies (1,2). However, recent guidelines caution against the prescribing of opioid analgesics as a first line agent for long term management of chronic neuropathic and non-neuropathic pain due to the risk profile of opioids which includes pronociception, cognitive impairment, addiction, misuse and more (1). These guidelines do, however, state that a time limited trial of opioid analgesics may be considered for PLHIV who are experiencing moderate to severe pain and are not responsive to first line therapies such as gabapentin for neuropathic pain or acetaminophen/NSAIDs for non-neuropathic pain (1,6).
Despite this, opioids are commonly, and at an increasing rate, being prescribed in potent, long-acting, high-dose formulations for pain management (1,6,7). These prescribing patterns are not limited to North American populations only; in fact, studies in both Australia and other European countries (e.g., Britain, Germany and Spain) report significant increases in opioid prescribing over the last 10 – 20 years, albeit, at lower rates than in North America (8–11). Moreover, there is evidence to suggest that high-dose and/or long term opioid prescriptions may be particularly problematic for PLHIV (12,15). A 2015 study revealed that long term prescription opioid use is associated with increased mortality risk in HIV patients compared to the general population (12,15). In PLHIV, opioid abuse may accelerate disease progression by disrupting immune-mediated gut homeostasis and by exacerbating the neuropathogenic mechanisms of HIV itself (15–17). Opioids have also been shown to interact with multiple antiretroviral medications, azoles and macrolides resulting in sub-optimal treatment of HIV infection and opioid toxicity (12,15,18,19).
A growing body of literature has demonstrated a wide array of negative outcomes associated with high-dose opioid prescription though much of the research to date has focused on the general population (2,7,20). Specifically, studies found a dose dependent relationship between opioid dosage and opioid-related mortality among the general population (20,21). These studies noted no distinct risk threshold for increased opioid overdose risk and found that while doses exceeding 200 MME/d posed a particularly high opioid overdose risk, intermediate opioid doses (50 – 199 MME/d) still posed elevated opioid-overdose risk (20,21). While the aforementioned studies have explored the adverse effects of high-dose opioid prescription in the general population, there is a limited understanding of the adverse effects of high-dose opioid prescription among PLHIV. It is evident that the prescription of high-dose opioids for PLHIV can have both serious short- and long-term adverse outcomes. However, a better understanding of this relationship, specifically in the context of high dose opioid prescription, is needed to ensure the safety and health of PLHIV. Using population level data, the objective of this study was to examine the adverse outcomes (i.e., mortality) associated with high-dose opioid prescription among PLHIV.
METHODS
Study Overview and Population
We used data from the Seek and Treat for Optimal Prevention of HIV/AIDS (STOP HIV/AIDS) in British Columbia (BC) cohort. Specific details of this cohort and the validity of this linkage methodology have been described in detail elsewhere (22,23). In short, the STOP HIV/AIDS cohort is a database of all identified PLHIV in the province of BC, Canada (22). The cohort, refreshed annually from 1996, links a number of BC provincial treatment, surveillance and administrative databases: The BC Centre for Excellence in HIV/AIDS Drug Treatment Program and Virology Registry (24,25); BC Centre for Disease Control HIV testing database (26,27); BC Medical Services Plan (MSP) (28); Discharge Abstract Database (DAD) (Canadian Institute of Health Information, 2016); the BC PharmaNet database (28); and Vital Statistics Databases (30). These databases collect extensive demographic, clinical, laboratory, prescription drug usage and health service utilization data for all registered HIV patients, which allowed us to analyze key outcomes such as mortality while controlling for a variety of confounders (22). This study was approved by the University of British Columbia – Providence Health Care’s Research Ethics Board.
Study Sample Inclusion Criteria
We included PLHIV who met the following criteria: 1) at least 18 years of age; 2) initiated antiretroviral therapy (ART); and 3) had a CD4 cell count and plasma viral load measurement within 6 months after their ART initiation date.
Measures
The main outcome measure was all-cause mortality, measured using the Vital Statistics Database. We also included descriptive data on overdose-related mortality, which was defined as death where the underlying and contributing cause was drug poisoning of all intents according to the following ICD-10 codes (and related ICD-9 codes): X40-X44, X60-X64, X85, or Y10-Y14 (31). The main explanatory measure we examined was high-dose opioid prescription, defined as > 90 MME/d. We chose >90MME/d as our cut-off point to be consistent with recent North American opioid prescribing guidelines and other research studies that warrant caution to doses prescribed higher than >90MME/d (6,32,33).
We considered a number of confounders we hypothesized would be associated with the main outcome of interest including: 1) sex (female versus [vs.] male), 2) ART initiation era (1996 – 1999 [reference] vs. 2000 – 2003 vs. 2004 – 2007 vs. 2008 – 2011 vs. 2012 – 2015); 3) diagnosed substance use disorder (yes vs. no); 4) co-prescription of opioids and benzodiazepines with overlap of prescriptions for at least one day (yes vs. no); 5) age at baseline (per 10-year increase); 6) Charlson comorbidity index (per unit increase); 7) CD4 cell counts (per 100 cells/mm3 increase); and 8) viral load (per log10 copies/mL increase. Other than sex, ART initiation era, and age, all confounders were time varying in the analysis.
Statistical Analyses
First, we performed descriptive statistics to characterize our patient sample at baseline, stratified by all-cause mortality. Kruskal-Wallis test was used to compare quantitative variables and Chi-square test (or Fisher’s exact test) was used to compare qualitative variables. Next, we calculated all-cause and overdose-related mortality rates as crude measures, adjusted for age and sex using the 2011 Canadian population as a reference, and stratified by sex, age, and year. In total, there were 216950 observations among 9272 individuals. 30517 (14%) observations among 3287 individuals were excluded because of missing data, but no individuals were excluded from the analyses.
To estimate the relationship between high-dose opioid prescription (>90MME/d) and all-cause mortality, we used bivariable and multivariable generalized estimating equation (GEE) models with Poisson distribution after test for overdispersion. To fit the multivariable model, all demographic and clinical confounders that we hypothesized a priori were associated with the main effect of interest were included regardless of statistical significance in bivariable analyses.
As a secondary analysis, we were interested in examining whether there was an interaction effect between high-dose opioid prescription (>90MME/d) and: 1) the presence of a substance use disorder, 2) co-prescription of benzodiazepines on all-cause mortality rates and 3) calendar year on all-cause mortality rates. Therefore, we constructed two additional multivariable GEE models to examine these potential interactions. All p-values are two-sided. All analyses were performed using SAS software version 9.4 (SAS Institute, Cary, NC).
RESULTS
Study sample characteristics
Over the 19-year study period (1996 – 2015) of the STOP HIV/AIDS cohort, 9272 PLHIV met the study inclusion criteria; 1676 (18.1%) were female and the median age at baseline was 38 years (quartile[Q]1 – Q3: 31 – 45 years). In total, 1862 (20.1%) of the identified patients died during the study period; and among these deaths, 244 (13.1%) were overdose-related. In total, 2206 (23.8%) patients were prescribed opioids at >90MME/d at least once during the study period. The baseline characteristics of the included 9272 patients in the cohort are presented in Table 1.
TABLE 1.
Descriptive Characteristics of Study Sample at Baseline
| Exposures of Main Interest | Total (%) (n = 9272) | All-Cause Mortality |
p - value | |
|---|---|---|---|---|
| Yes (20.1%) (n = 1862) | No (79.9%) (n = 7410) | |||
| High-dose Opioid | ||||
| > 90 MME/d | 187 (2.0) | 50 (2.7) | 137 (1.9) | 0.022 |
| ≤ 90 MME/d | 9085 (98.0) | 1812 (97.3) | 7273 (98.2) | |
| Covariates | ||||
| Sex | ||||
| Male | 7596 (81.9) | 1477 (79.3) | 6119 (82.6) | 0.001 |
| Female | 1676 (18.1) | 385 (20.7) | 1291 (17.4) | |
| ART Era | ||||
| 1996 – 1999 | 2596 (28.0) | 966 (51.9) | 1630 (22.0) | < 0.001 |
| 2000 – 2003 | 1353 (14.6) | 392 (21.1) | 961 (13.0) | |
| 2004 – 2007 | 1644 (17.7) | 295 (15.8) | 1349 (18.2) | |
| 2008 – 2011 | 2175 (23.5) | 166 (8.9) | 2009 (27.1) | |
| 2012 – 2015 | 1504 (16.2) | 43 (2.3) | 1461 (19.7) | |
| Substance Use Disorder | ||||
| Yes | 1365 (14.7) | 410 (22.0) | 955 (12.9) | < 0.001 |
| No | 7907 (85.3) | 1452 (78.0) | 6455 (87.1) | |
| Co-prescription of Opioids and Benzodiazepines | ||||
| Yes | 283 (3.1) | 118 (6.3) | 165 (2.2) | < 0.001 |
| No | 8989 (96.9) | 1744 (93.7) | 7245 (97.8) | |
| Age (Years; median, Q1-Q3) | 38 (31–45) | 40 (34–38) | 37 (31–44) | < 0.001 |
| Charlson Comorbidity Index | 4 (4–6) | 5 (4–6) | 4 (4–6) | < 0.001 |
| CD4 Cell Counts (Cells/mm3; median, Q1-Q3) | 260 (130–420) | 190 (80–330) | 280 (150–440) | < 0.001 |
| Viral Load (Log10 copies/ml; median, Q1-Q3) | 4.8 (4.2–5) | 5 (4.6–5) | 4.8 (4.1–5) | < 0.001 |
MME/d: Morphine Milligram Equivalents per day; ART: Antiretroviral Therapy; Q: Quartile
Mortality rates
The crude all-cause mortality rate was 27.10 per 1000 person-years (95% confidence interval[CI]: 25.89 – 28.36 per 1000 person-years), whereas age- and sex-adjusted mortality rate (using the 2011 Canadian population as the reference) was 30.99 per 1000 person-years (95% CI: 28.11 – 33.88). In stratified analyses, among both men and women, mortality rates increased with increasing age groups: among women, from 19.97 per 1000 person-years (95% CI: 13.67 – 29.19) in the 19–30 age group to 29.38 per 1000 person-years (95% CI: 24.91 – 62.24) in the ≥ 60 age group, whereas among men, the mortality rate increased from 15.20 per 1000 person-years (95% CI: 10.77 – 21.44) in the 19–30 age group to 47.05 per 1000 person-years (95% CI: 41.55 – 53.27) in the ≥ 60 age group. Stratification by year revealed a significant decline in mortality rates from 84.43 per 1000 person-years in 1996 to 18.71 per 1000 person-years in 2015.
We also analyzed crude, sex, age and year adjusted overdose-related mortality. The crude overdose-related mortality rate of the cohort was 3.55 per 1000 person-years (95% CI: 3.13 – 4.03). The overall female overdose-related mortality rate was 5.62 per 1000 person-years (95% CI: 4.42 – 7.13) whereas in males it was 3.11 per 1000 person-years (95% CI: 2.68 – 3.61).
Receipt of high-dose opioid analgesics and mortality rates
Table 2 shows unadjusted and adjusted GEE models examining the relationship between high-dose opioid analgesics and all-cause mortality rates. In an unadjusted GEE model, we found a positive and significant relationship between high-dose opioid prescriptions exceeding 90 MME/d and all-cause mortality (rate ratio [RR] = 5.03; 95% CI: 4.38 – 5.78). In a multivariable GEE model adjusted for various demographic and clinical confounders, patients receiving greater than 90 MME/d of opioids remained positively and significantly associated with all-cause mortality risk compared to those who received an opioid prescription of less than 90 MME/day (adjusted rate ratio [ARR] = 3.01; 95% CI: 2.47 – 3.66).
TABLE 2.
Bivariable and multivariable generalized estimating equation modeling of factors associated with all-cause mortality rates (n = 9272)
| Characteristic | Risk Ratio (RR) |
|
|---|---|---|
| Unadjusted RR (95% CI) | Adjusted RR (95% CI) | |
| High-dose Opioid Regimen (>90MME/d vs. ≤90MME/d) | 5.03 (4.38 – 5.78) | 3.01 (2.47 – 3.66) |
| Sex (female vs. male) | 1.22 (1.09 – 1.37) | 1.21 (1.04 – 1.41) |
| ART Era | ||
| 2000 – 2003 vs. 1996 – 1999 | 0.95 (0.84 – 1.07) | 0.78 (0.67 – 0.91) |
| 2004 – 2007 vs. 1996 – 1999 | 0.77 (0.67 – 0.88) | 0.55 (0.46 – 0.65) |
| 2008 – 2011 vs. 1996 – 1999 | 0.52 (0.44 – 0.61) | 0.43 (0.35 – 0.52) |
| 2012 – 2015 vs. 1996 – 1999 | 0.58 (0.43 – 0.79) | 0.61 (0.44 – 0.84) |
| Substance Use Disorder (yes vs. no) | 1.45 (1.30 – 1.62) | 0.93 (0.79 – 1.09) |
| Co-prescription of Opioid and Benzodiazepine (yes vs. no) | 3.95 (3.41 – 4.56) | 1.80 (1.46 – 2.22) |
| Age at Baseline (per 10 years increase)) | 1.38 (1.33 – 1.45) | 1.36 (1.28 – 1.44) |
| Charlson Comorbidity Index (per unit) | 1.26 (1.24 – 1.28) | 1.26 (1.23 – 1.28) |
| CD4 Cell Counts (per 100 cells/mm3) | 0.74 (0.72 – 0.77) | 0.78 (0.75 – 0.81) |
| Viral Load (per log10 copies/ml) | 1.21 (1.17 – 1.25) | 1.21 (1.16 – 1.27) |
MME/d: Morphine milligram equivalents per day; ART: antiretroviral therapy; CI: confidence interval
Table 3 shows unadjusted and adjusted GEE models examining the relationship between high-dose opioid analgesics and overdose-related mortality rates. In an unadjusted GEE model, we found a positive and significant relationship between high-dose opioid prescriptions exceeding 90 MME/d and overdose-related mortality (rate ratio [RR] = 3.14; 95% CI: 2.12 – 4.65). In a multivariable GEE model adjusted for various demographic and clinical confounders, we failed to find a statistically significant relationship between patients receiving greater than 90 MME/d of opioids and overdose-related mortality risk compared to those who received an opioid prescription of less than 90 MME/day (adjusted rate ratio [ARR] = 1.42; 95% CI: 0.84 – 2.43).
TABLE 3.
Bivariable and multivariable generalized estimating equation modeling of factors associated with overdose-related mortality rates (n = 9272)
| Characteristic | Risk Ratio (RR) |
|
|---|---|---|
| Unadjusted RR (95% CI) | Adjusted RR (95% CI) | |
| High-dose Opioid Regimen (>90MME/d vs. ≤90MME/d) | 3.14 (2.12 – 4.65) | 1.42 (0.84 – 2.43) |
| Sex (female vs. male) | 1.81 (1.36 – 2.39) | 1.34 (0.96 – 1.87) |
| ART Era | ||
| 2000 – 2003 vs. 1996 – 1999 | 0.83 (0.60 – 1.16) | 0.82 (0.56 – 1.19) |
| 2004 – 2007 vs. 1996 – 1999 | 0.58 (0.39 – 0.85) | 0.61 (0.40 – 0.93) |
| 2008 – 2011 vs. 1996 – 1999 | 0.42 (0.26 – 0.69) | 0.53 (0.32 – 0.88) |
| 2012 – 2015 vs. 1996 – 1999 | 1.06 (0.57 – 1.96) | 1.53 (0.82 – 2.89) |
| Substance Use Disorder (yes vs. no) | 3.05 (2.35 – 3.96) | 2.03 (1.47 – 2.81) |
| Co-prescription of Opioid and Benzodiazepine (yes vs. no) | 4.75 (3.37 – 6.70) | 2.97 (1.46 – 2.22) |
| Age at Baseline (per 10 years increase) | 0.89 (0.79 – 1.00) | 0.91 (0.78 – 1.05) |
| Charlson Comorbidity Index (per unit) | 1.17 (1.13 – 1.21) | 1.13 (1.07 – 1.18) |
| CD4 Cell Counts (per 100 cells/mm3) | 0.94 (0.89 – 1.00) | 0.95 (0.89 – 1.01) |
| Viral Load (per log10 copies/ml) | 1.07 (0.98 – 1.17) | 1.10 (0.98 – 1.24) |
MME/d: Morphine milligram equivalents per day; ART: antiretroviral therapy; CI: confidence interval
Secondary analyses: interactions with receipt of high-dose opioid analgesics
In secondary analyses, we constructed multivariable GEE models to assess whether the relationship between being prescribed greater than 90 MME/d of opioids and all-cause mortality was dependent on having a substance use disorder (data not shown). Compared to PLHIV who were prescribed ≤90MME/d and did not have a substance use disorder, those who were prescribed >90MME/d and had a substance use disorder (ARR = 1.86; 95%CI: 1.34 – 2.60) and those who were prescribed >90MME/d and did not have a substance use disorder (ARR = 3.98; 95%CI: 3.18–4.98) had an increased all-cause mortality risk. Upon further analysis, all-cause mortality rates among those who were prescribed >90MME/d without a substance use disorder was statistically significantly higher (ARR=2.14; 95% CI: 1.49–3.07) compared to those who were prescribed >90MME/d and had a substance use disorder.
We also hypothesized that the relationship between being prescribed greater than 90 MME/d of opioids and all-cause mortality was dependent on being co-prescribed benzodiazepines (data not shown). Compared to PLHIV prescribed ≤90 MME/d without benzodiazepine co-prescription, PLHIV prescribed ≤90 MME/d with benzodiazepine co-prescription (ARR=2.21; 95% CI: 1.72–2.83) had increased all-cause mortality rates. PLHIV prescribed >90MME/d without benzodiazepine co-prescription (ARR=3.48; 95% CI: 2.79–4.34); and PLHIV prescribed >90MME/d with benzodiazepine co-prescription (ARR=4.75; 95% CI: 3.68–6.13) also had increased all-cause mortality rates.
We also considered whether the association between high-dose opioid prescription and all-cause mortality changed over time. At baseline year (1996), in an adjusted multivariable model, patients receiving greater than 90 MME/d of opioids were positively and significantly associated with all-cause mortality risk compared to those who received an opioid prescription of less than 90 MME/day (adjusted rate ratio [ARR] = 5.35; 95% CI: 3.53 – 8.12). This rate ratio decreases by 5% as calendar year increases by 1 year (ARR = 0.95; 95% CI: 0.92 – 0.98).
DISCUSSION
In an analysis of a province-wide cohort spanning a 19-year period, we found that approximately one-fifth of PLHIV experienced a mortality event. While both all-cause and overdose-related mortality decreased over the study period, we found a positive association between high-dose opioid prescription and all-cause mortality among our study population. In effect modification analyses, we further demonstrated that the relationship between high-dose opioid prescription and all-cause mortality was negatively associated with a) having a SUD and positively associated with b) being co-prescribed benzodiazepines. Our study is one of the first to estimate the relationship between high-dose opioid prescription and all-cause mortality in a population-level cohort of PLHIV.
We found that both all-cause and overdose-related mortality decreased over the 19-year study period in the study cohort. The decline in mortality of PLHIV may be largely explained by advancements in ART therapy, as well as optimization and increased reach of HIV-related care (34,35). While beyond our study period, it is noteworthy that given the opioid overdose public health emergency declared in 2016 in the province, it is anticipated that this decreasing mortality trend would not have continued beyond 2015 (36). In fact, the Coroners Service of British Columbia estimated a tripling in the rate of opioid overdose-related mortality between 2015 and 2018, largely due to the increased fentanyl contamination of the illicit drug supply (37). Future research with data post 2015 should seek to explore the impact of the opioid epidemic on this population in further detail.
Our findings are consistent with studies performed among the general population and PLHIV that showed that all-cause mortality was higher among those who were prescribed high doses of opioids (12,15,20,21). Specifically, in the general population, a study in the US demonstrated almost a three-fold increase in mortality risk among those who were prescribed between 80 – 99.9 MME/d compared to those who were prescribed doses between 0 – 39.9 MME/d (21). Another study conducted in Canada also found a positive dose-dependent relationship between opioid dose and mortality in the general population (20). Furthermore, compared to the general population, high-dose opioid prescription among PLHIV is more prevalent and is associated with greater all-cause mortality (12,15). Our failure to find an association between high-dose opioid prescription and overdose-related mortality may be because overdose is only one potential cause of mortality secondary to opioid use. Long term opioid receipt may contribute to medical mortality via cardiac, endocrine and gastrointestinal disturbances (38,39). Moreover, falls, fractures and vehicular accidents are associated with opioid prescription and may have contributed to all-cause mortality but not overdose-related mortality (38,40–43). Based on findings in the general population, regional, national and international guidelines warn physicians to reassess patient benefits when prescribing over 50 MME/d of opioids and to avoid prescribing over 90 MME/d of opioids unless there is significant evidence of benefit to the patient (6,32,33).
In secondary analyses, we found that the relationship between being prescribed high-dose opioids and mortality was dependent on the presence of a SUD; however, contrary to our hypothesis, the risk of mortality among PLHIV who were prescribed >90MME/d with no SUD was significantly higher compared to those who were prescribed >90MME with a SUD. There may be a few reasons for these unexpected findings. Substance use disorders (SUD) are known to be underdiagnosed by physicians (44) for multiple reasons such as lack of training, emergency medicine’s focus on acute illnesses, clinicians not screening for SUDs in patients who do not fit the stereotype, and due to overlap between symptoms of substance use and other illnesses (45,46). It is likely that the potential under-diagnosis of SUDs could have resulted in a large number of patients with SUDs being misclassified as patients without SUDs. Alternative explanations for this finding may be that patients with an opioid use disorder in particular may possess higher opioid tolerance, making them less likely to overdose. On the other hand, PLHIV with SUDs may be monitored more closely by physicians, which may in turn reduce their risk of mortality.
We also found that the relationship between being prescribed high-dose opioids and mortality was dependent on having been co-prescribed benzodiazepines. In alignment with the literature, we found that co-prescription of benzodiazepines with opioids resulted in a statistically significant increase in all-cause mortality even with low opioid doses (≤90MME/d) (1,12,20,47). In fact, a US study found that overdose mortality was ten times higher in patients co-prescribed opioids and benzodiazepines compared to those prescribed opioids alone (21). Since PLHIV often have higher rates of comorbid mental illness and pain, there are cases where benzodiazepine and opioid co-prescription may be indicated (12,48). Nevertheless, it is imperative that physicians prescribing opioids to PLHIV strike a safe balance between pain management concerns and common drug interactions that put the patient at risk for overdose or mortality (21,48).
Increased all-cause mortality among PLHIV who were prescribed high doses of opioids, alongside a higher prevalence of high-dose opioid prescription among this population, highlights an important need for optimized, well monitored prescription of opioids by physicians. National and international guidelines recommend non-opioid analgesics as a first line treatment for mild chronic pain, followed by a trial of opioids with adjuvant non-pharmacological and non-opioid pharmacological therapies for management of more persistent, moderate to severe pain (6,33,49). Therefore, strategies to ascertain the safety of PLHIV whose pain is managed using opioids should include: 1) improved education of physicians who provide HIV care in addiction medicine, including safer opioid prescribing and recognition and treatment of SUDs; 2) better adherence by physicians to prescribing guidelines for pain among PLHIV; and 3) development of multidisciplinary teams, consisting of pharmacists, addiction medicine specialists and opioid-prescribing physicians working together to monitor patients who are prescribed opioids (50,51). Ultimately, we emphasise the need for an optimized opioid prescription that comprehensively considers patient risks and benefits in accordance with evidence-based guidelines, as opposed to stricter prescribing practices that may leave PLHIV with suboptimal chronic pain management.
Our study has limitations that need to be considered. Firstly, our data is observational in nature and while controlled for a variety of confounders, there may be unmeasured confounders not accounted for. Secondly, opioid doses were ascertained using drug prescription dispensation data; thus, we cannot determine whether the medications were taken as prescribed. It is noteworthy that pain medications are often consumed on an as needed basis, and therefore it is likely that we overestimated the MME/d for each patient. Relatedly, we were unable to incorporate MME/d of additional opioids that may have been prescribed outside of BC and/or obtained illicitly and thus may have underestimated opioid use in these cases. Nevertheless, there are studies that suggest that this method of using drug prescription data is valid for measuring drug exposure in the population (52). Additionally, misclassification in the reporting of ICD codes for overdose-related mortality may have not adequately captured the true rate in our sample. This may have contributed to our failure to find an association between high-dose opioid prescription and overdose-related mortality. Lastly, our findings may not be generalizable to other regions with varying demographics, opioid prescribing habits, and healthcare systems.
Within our large provincial sample of PLHIV, we found a positive association between high-dose opioid prescription and all-cause mortality. These findings demonstrate evidence for the risks associated with prescribing high-dose opioids to manage HIV-related pain and emphasise the importance of assessing patient risks and benefits periodically when prescribing opioids to PLHIV. Strategies and interventions to achieve a better balance between pain relief and related risks are importantly needed. Appropriate opioid prescribing, in alignment with national and international guidelines, is paramount given the comorbidities of PLHIV and in the context of North America’s opioid overdose epidemic.
Figure 1.
All-cause Mortality Rate Stratified by Year
Acknowledgments
This study was funded by the British Columbia Ministry of Health (BCMoH), which-funded Seek and treat for optimal prevention of HIV & AIDS pilot project, and an Avant-Garde Award (number 1DP1DA026182) and grant 1R01DA036307-01 from the National Institute of Drug Abuse, at the US National Institutes of Health. LT is supported by a grant from the Michael Smith Foundation. JM’s Treatment as Prevention (TasP) research, paid to institution, has received support from the Public Health Agency of Canada, BC-Ministry of Health and US NIH (NIDA R01DA036307 and CTN 248). Institutional grants have been provided by J&J, Merck and a Knowledge Translation Award from CIHR. JM has served as an advisor to the federal and BC governments, UNAIDS, WHO in the last year. NF is supported by a grant from the Michael Smith Foundation for Health Research/St. Paul’s Foundation. SN is supported by a MSFHR/St. Paul’s Hospital Foundation and Providence Health Care Research Institute Early Career Research Initiative Award. The funder had no direct role in the conduct of the analysis or the decision to submit the manuscript for publication. All inferences, opinions, and conclusions drawn in this publication are those of the author(s), and do not necessarily reflect the opinions or policies of the data steward.
CONFLICT OF INTEREST
JM’s Treatment as Prevention (TasP) research, paid to institution, has received support from the Public Health Agency of Canada, BC-Ministry of Health and US NIH (NIDA R01DA036307 and CTN 248). Institutional grants have been provided by J&J, Merck and a Knowledge Translation Award from CIHR. JM has served as an advisor to the federal and BC governments, UNAIDS, WHO in the last year.
Footnotes
The STOP HIV/AIDS in BC Study Group:
Rolando Barrios, MD, FRCPC, Senior Medical Director, VCH; Adjunct Professor, School of Population and Public Health, UBC. Patty Daly, MD, Vancouver Coastal Health Authority. Mark Gilbert, Clinical Prevention Services, BC Centre for Disease Control; School of Population and Public Health, University of British Columbia. Reka Gustafson, MD, Vancouver Coastal Health Authority. Perry R.W. Kendall, OBC, MBBS, MSc, FRCPC, Provincial Health Officer, British Columbia Ministry of Health; Clinical Professor, Faculty of Medicine UBC. Ciro Panessa, British Columbia Ministry of Health. Gina McGowan, British Columbia Ministry of Health. Nancy South, British Columbia Ministry of Health. Kate Heath, Robert S. Hogg, and Julio S.G. Montaner, BC Centre for Excellence in HIV/AIDS.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
REFERENCES
- 1.Bruce RD, Merlin J, Lum PJ, Ahmed E, Alexander C, Corbett AH, et al. 2017 HIVMA of IDSA Clinical Practice Guideline for the Management of Chronic Pain in Patients Living With HIV. Clin Infect Dis [Internet]. 2017/10/12. 2017;65(10):e1–37. Available from: 10.1093/cid/cix636 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Krashin DL, Merrill JO, Trescot AM. Opioids in the management of HIV-related pain. Pain Physician [Internet]. 2012;15(3 Suppl):ES157–68. Available from: http://www.ncbi.nlm.nih.gov/pubmed/22786454 [PubMed] [Google Scholar]
- 3.Nightingale VR, Sher TG, Mattson M, Thilges S, Hansen NB. The effects of traumatic stressors and HIV-related trauma symptoms on health and health related quality of life. AIDS Behav [Internet]. 2011. [cited 2018 May 25];15(8):1870–8. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3629911/pdf/nihms348047.pdf [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Plotzker RE, Metzger DS, Holmes WC. Childhood Sexual and Physical Abuse Histories, PTSD, Depression, and HIV Risk Outcomes in Women Injection Drug Users: A Potential Mediating Pathway. Am J Addict [Internet]. 2007. January [cited 2018 May 25];16(6):431–8. Available from: http://doi.wiley.com/10.1080/10550490701643161 [DOI] [PubMed] [Google Scholar]
- 5.Pence BW, Reif S, Whetten K, Leserman J, Stangl D, Swartz M, et al. Minorities, the Poor, and Survivors of Abuse: HIV-Infected Patients in the US Deep South. South Med J [Internet]. 2007. November [cited 2018 May 25];100(11):1114–22. Available from: http://sma.org/southern-medical-journal/article/minorities-the-poor-and-survivors-of-abuse-hiv-infected-patients-in-the-us-deep-south [DOI] [PubMed] [Google Scholar]
- 6.Busse JW, Buckley N, Buna D, Franklin G, Giorshev C, Harris J, et al. The 2017 Canadian Guideline for Opioids for Chronic Non-Cancer Pain Guideline Panel Members. 2017. [cited 2018 May 5]; Available from: http://nationalpaincentre.mcmaster.ca/documents/OpioidGLforCMAJ_01may2017.pdf
- 7.Ottawa OC. Pan-Canadian Trends in the Prescribing of Opioids, 2012 to 2016. 2017. [cited 2018 May 3]; Available from: https://secure.cihi.ca/free_products/pan-canadian-trends-opioid-prescribing-2017-en-web.pdf
- 8.Karanges EA, Blanch B, Buckley NA, Pearson SA. Twenty-five years of prescription opioid use in Australia: a whole-of-population analysis using pharmaceutical claims. Br J Clin Pharmacol. 2016;255–67. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Blanch B, Pearson SA, Haber PS. An overview of the patterns of prescription opioid use, costs and related harms in Australia. Br J Clin Pharmacol. 2014;78(5):1159–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Garcia del Pozo J, Carvajal A, Viloria JM, Velasco A, Garcia del Pozo V. Trends in the consumption of opioid analgesics in Spain. Higher increases as fentanyl replaces morphine. Eur J Clin Pharmacol [Internet]. 2008. April 20 [cited 2018 May 4];64(4):411–5. Available from: http://link.springer.com/10.1007/s00228-007-0419-9 [DOI] [PubMed] [Google Scholar]
- 11.Schubert I, Ihle P, Sabatowski R. Zunahme der opioidverordnungen in Deutschland zwischen 2000 und 2010: Eine studie auf der basis von krankenkassendaten. Dtsch Arztebl Int. 2013;110(4):45–51.23413387 [Google Scholar]
- 12.Becker WC, Gordon K, Edelman EJ, Kerns RD, Crystal S, Dziura JD, et al. Trends in Any and High-Dose Opioid Analgesic Receipt Among Aging Patients With and Without HIV. AIDS Behav [Internet]. 2016. [cited 2018 May 4];20(3):679–86. Available from: http://www.ncbi.nlm.nih.gov/pubmed/26384973 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Silverberg MJ, Ray GT, Saunders K, Rutter CM, Campbell CI, Merrill JO, et al. Prescription Long-term Opioid Use in HIV-infected Patients. Clin J Pain [Internet]. 2012;28(1):39–46. Available from: http://content.wkhealth.com/linkback/openurl?sid=WKPTLP:landingpage&an=00002508−201201000-00006 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Tsao JCI, Dobalian A, Stein JA. Illness burden mediates the relationship between pain and illicit drug use in persons living with HIV. Pain [Internet]. 2005. December 15 [cited 2019 Jul 7];119(1–3):124–32. Available from: http://www.ncbi.nlm.nih.gov/pubmed/16297562 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Weisberg DF, Gordon KS, Barry DT, Becker WC, Crystal S, Edelman EJ, et al. Long-term prescription of opioids and/or benzodiazepines and mortality among HIV-infected and uninfected patients. J Acquir Immune Defic Syndr [Internet]. 2015. June 1 [cited 2019 Jul 7];69(2):223–33. Available from: http://www.ncbi.nlm.nih.gov/pubmed/26009831 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Liu B, Liu X, Tang S-J. Interactions of Opioids and HIV Infection in the Pathogenesis of Chronic Pain. Front Microbiol [Internet]. 2016. February 10 [cited 2019 Jul 7];7:103 Available from: http://journal.frontiersin.org/Article/10.3389/fmicb.2016.00103/abstract [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Meng J, Sindberg GM, Roy S. Disruption of gut homeostasis by opioids accelerates HIV disease progression. Front Microbiol [Internet]. 2015. June 26 [cited 2019 Jul 7];6:643 Available from: http://journal.frontiersin.org/Article/10.3389/fmicb.2015.00643/abstract [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.McCance-Katz EF, Sullivan LE, Nallani S. Drug interactions of clinical importance among the opioids, methadone and buprenorphine, and other frequently prescribed medications: a review. Am J Addict [Internet]. 2010. [cited 2019 Jul 7];19(1):4–16. Available from: http://www.ncbi.nlm.nih.gov/pubmed/20132117 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Gruber VA, McCance-Katz EF. Methadone, buprenorphine, and street drug interactions with antiretroviral medications. Curr HIV/AIDS Rep [Internet]. 2010. August [cited 2019 Jul 7];7(3):152–60. Available from: http://www.ncbi.nlm.nih.gov/pubmed/20532839 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Gomes T, Mamdani MM, Dhalla IA, Paterson JM, Juurlink DN. Opioid dose and drug-related mortality in patients with nonmalignant pain. Arch Intern Med [Internet]. 2011/04/13. 2011;171(7):686–91. Available from: 10.1001/archinternmed.2011.117 [DOI] [PubMed] [Google Scholar]
- 21.Dasgupta N, Funk MJ, Proescholdbell S, Hirsch A, Ribisl KM, Marshall S. Cohort Study of the Impact of High-Dose Opioid Analgesics on Overdose Mortality. Pain Med [Internet]. 2015/09/04. 2016;17(1):85–98. Available from: 10.1111/pme.12907 [DOI] [PubMed] [Google Scholar]
- 22.Heath K, Samji H, Nosyk B, Colley G, Gilbert M, Hogg RS, et al. Cohort profile: Seek and treat for the optimal prevention of HIV/AIDS in British Columbia (STOP HIV/AIDS BC). Int J Epidemiol [Internet]. 2014. August [cited 2018 May 7];43(4):1073–81. Available from: http://www.ncbi.nlm.nih.gov/pubmed/24695113 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Nosyk B, Colley G, Yip B, Chan K, Heath K, Lima VD, et al. Application and validation of case-finding algorithms for identifying individuals with human immunodeficiency virus from administrative data in British Columbia, Canada. PLoS One. 2013;8(1):e54416. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.BC Centre for Excellence in HIV/AIDS. STOP HIV/AIDS | BC Centre for Excellence in HIV/AIDS [Internet]. 2014. [cited 2019 Aug 4]. Available from: http://cfenet.ubc.ca/stop-hiv-aids/about
- 25.BC Centre for Excellence in HIV/AIDS [creator]. Drug treatment program and virology registry. BC Centre for Excellence in HIV/AIDS [publisher]; 2014. [Google Scholar]
- 26.British Columbia Centre for Disease Control Public Health Laboratory [creator]. HIV laboratory testing datasets (tests: ELISA, Western blot, NAAT, p24, culture). Clinical Prevention Services, British Columbia Centre for Disease Control [publisher]; 2016. [Google Scholar]
- 27.BC Centre for Disease Control. HIV in British Columbia: Annual Surveillance Report 2016 [Internet]. 2015. [cited 2018 May 3]. Available from: mailto:http://www.bccdc.ca/resource-gallery/Documents/StatisticsandResearch/StatisticsandReports/STI/HIV_Annual_Report_2015-FINAL.pdf
- 28.British Columbia Ministry of Health. Medical Services Plan (MSP) Payment Information File; Consolidation File (MSP Registration & Premium Billing); Home & Community Care (Continuing Care); Mental Health; PharmaNet. British Columbia Ministry of Health [publisher]; 2016. [Google Scholar]
- 29.Canadian Institute of Health Information [creator]. Discharge Abstract Database (Hospital Separations). British Columbia Ministry of Health [publisher]; 2016. [Google Scholar]
- 30.British Columbia Vital Statistics Agency. Vital Statistics. British Columbia Ministry of Health [publisher]; 2016. [Google Scholar]
- 31.CDC NCIPC DUIP PDO Team H. PRESCRIPTION DRUG OVERDOSE DATA & STATISTICS GUIDE TO ICD-9-CM AND ICD-10 CODES RELATED TO POISONING AND PAIN Introduction to ICD-9-CM and ICD-10 Codes Related to Poisoning and Pain. 2013. [cited 2018 May 7]; Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4543429/pdf/10729_2014_Article_9312.pdf
- 32.College of Physicians and Surgeons of British Columbia. Professional Standards and Guidelines – Safe Prescribing of Drugs with Potential for Misuse/Diversion. 2016. [cited 2018 May 9];1–7. Available from: https://www.cpsbc.ca/files/pdf/PSG-Safe-Prescribing.pdf
- 33.Dowell D, Haegerich TM, Chou R. CDC Guideline for Prescribing Opioids for Chronic Pain — United States, 2016. MMWR Recomm Reports [Internet]. 2016. March 15 [cited 2018 May 9];65(1):1–49. Available from: http://www.cdc.gov/mmwr/volumes/65/rr/rr6501e1er.htm [DOI] [PubMed] [Google Scholar]
- 34.Kok S, Rutherford AR, Gustafson R, Barrios R, Montaner JSG, Vasarhelyi K, et al. Optimizing an HIV testing program using a system dynamics model of the continuum of care on behalf of the Vancouver HIV Testing Program Modelling Group. Heal Care Manag Sci [Internet]. 2015. [cited 2018 May 22];18:334–62. Available from: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4543429/pdf/10729_2014_Article_9312.pdf [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Nosyk B, Zang X, Min JE, Krebs EM, Lima VD, Milloy M, et al. Relative effects of antiretroviral therapy and harm reduction initiatives on HIV incidence in British: a modelling study. www.thelancet.com Artic Lancet HIV [Internet]. 2017. [cited 2018 May 22];4:303–10. Available from: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 36.BC Gov News. Provincial Health Officer Declares Public Health Emergency [Internet]. 2016. [cited 2018 Aug 7]. Available from: https://news.gov.bc.ca/releases/2016hlth0026-000568
- 37.Coroners Service B Ministry of Public Safety and Solicitor General Fentanyl-Detected Illicit Drug Overdose Deaths. 2012. [cited 2018 May 3]; Available from: https://www2.gov.bc.ca/assets/gov/public-safety-and-emergency-services/death-investigation/statistical/fentanyl-detected-overdose.pdf
- 38.Benyamin R, Trescot AM, Datta S, Buenaventura R, Adlaka R, Sehgal N, et al. Opioid complications and side effects. Pain Physician [Internet]. 2008. March [cited 2019 Jul 7];11(2 Suppl):S105–20. Available from: http://www.ncbi.nlm.nih.gov/pubmed/18443635 [PubMed] [Google Scholar]
- 39.Brandenburg MA. Prescription Opioids Are Associated With Population Mortality in US Deep South Middle-Age Non-hispanic Whites: An Ecological Time Series Study. Front Public Heal. 2019. September 6;7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Woolcott JC, Richardson KJ, Wiens MO, Patel B, Marin J, Khan KM, et al. Meta-analysis of the impact of 9 medication classes on falls in elderly persons. Vol. 169, Archives of Internal Medicine. 2009. p. 1952–60. [DOI] [PubMed] [Google Scholar]
- 41.Baldini A, Von Korff M, Lin EHB. A Review of Potential Adverse Effects of Long-Term Opioid Therapy: A Practitioner’s Guide. Prim care companion CNS Disord [Internet]. 2012. [cited 2018 May 5];14(3). Available from: http://www.ncbi.nlm.nih.gov/pubmed/23106029 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Chihuri S, Li G. Use of Prescription Opioids and Initiation of Fatal 2-Vehicle Crashes. JAMA Netw open. 2019. February 1;2(2):e188081. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Saunders KW, Dunn KM, Merrill JO, Sullivan M, Weisner C, Braden JB, et al. Relationship of opioid use and dosage levels to fractures in older chronic pain patients. J Gen Intern Med. 2010. April;25(4):310–5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Modesto-Lowe V, Brooks D, Freedman K, Hargus E. Addiction and chronic pain: diagnostic and treatment dilemmas. Conn Med [Internet]. 2007. March [cited 2018 May 12];71(3):139–44. Available from: http://www.ncbi.nlm.nih.gov/pubmed/17405395 [PubMed] [Google Scholar]
- 45.Basco MR, Jacquot C, Thomas C, Knack JM. Underdiagnosing and overdiagnosing psychiatric comorbidities: insights into common diagnostic oversights. Psychiatr Times [Internet]. 2008. May 23;25(SE1):8 Available from: http://link.galegroup.com.ezproxy.library.ubc.ca/apps/doc/A187963811/HRCA?u=ubcolumbia&sid=HRCA&xid=f5a22535 [Google Scholar]
- 46.Bernstein SL, D’Onofrio G. A Promising Approach For Emergency Departments To Care For Patients With Substance Use And Behavioral Disorders. Health Aff [Internet]. 2013. December 1 [cited 2018 May 22];32(12):2122–8. Available from: http://content.healthaffairs.org/cgi/doi/10.1377/hlthaff.2013.0664 [DOI] [PubMed] [Google Scholar]
- 47.Chou R, Fanciullo GJ, Fine PG, Adler JA, Ballantyne JC, Davies P, et al. Clinical Guidelines for the Use of Chronic Opioid Therapy in Chronic Noncancer Pain. J Pain [Internet]. 2009;10(2):113–130.e22. Available from: 10.1016/j.jpain.2008.10.008 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 48.Reisfield GM, Webster LR. Benzodiazepines in Long-Term Opioid Therapy. Pain Med [Internet]. 2013. October 1 [cited 2018 May 23];14(10):1441–6. Available from: https://academic.oup.com/painmedicine/article-lookup/doi/10.1111/pme.12236 [DOI] [PubMed] [Google Scholar]
- 49.Miller E The World Health Organization analgesic ladder. J Midwifery Women’s Heal [Internet]. 2004. November 1 [cited 2018 May 23];49(6):542–5. Available from: https://www-sciencedirect-com.ezproxy.library.ubc.ca/science/article/pii/S1526952304004374 [DOI] [PubMed] [Google Scholar]
- 50.Khidir H, Weiner SG. A Call for Better Opioid Prescribing Training and Education. West J Emerg Med [Internet]. 2016. [cited 2018 May 12];686(6). Available from: http://escholarship.org/uc/uciem_westjem [DOI] [PMC free article] [PubMed] [Google Scholar]
- 51.Debar LL, Kindler L, Keefe FJ, Green CA, Smith DH, Deyo RA, et al. A primary care-based interdisciplinary team approach to the treatment of chronic pain utilizing a pragmatic clinical trials framework. Transl Behav Med [Internet]. 2012. December 1 [cited 2018 May 23];2(4):523–30. Available from: http://www.ncbi.nlm.nih.gov/pubmed/23440672 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 52.Lau HS, de Boer A, Beuning KS, Porsius A. Validation of Pharmacy Records in Drug Exposure Assessment. J Clin Epidemiol [Internet]. 1997. [cited 2018 May 23];50(5):619–25. Available from: https://ac-els-cdncom.ezproxy.library.ubc.ca/S0895435697000401/1-s2.0-S0895435697000401-main.pdf?_tid=084dfb81-15ac-4857-83bb-ed4d1f8d0bac&acdnat=1527119293_114ad87c00ff1f45f201b4d1e91ffd52 [DOI] [PubMed] [Google Scholar]