Associations between prescribed benzodiazepines, overdose death and buprenorphine discontinuation among people receiving buprenorphine

BACKGROUND AND AIMS

Benzodiazepines are commonly prescribed to patients with opioid use disorder receiving buprenorphine treatment, yet may increase overdose risk. However, prescribed benzodiazepines may improve retention in care by reducing buprenorphine discontinuation and thus may prevent relapse to illicit opioid use. We aimed to test the association between benzodiazepine prescription and fatal opioid overdose, non-fatal opioid overdose, all-cause mortality and buprenorphine discontinuation.

DESIGN AND SETTING

This was a retrospective cohort study using five individually linked data sets from Massachusetts, United States government agencies.

PARTICIPANTS

We studied 63 389 Massachusetts residents aged 18 years or older who received buprenorphine treatment between January 2012 and December 2015.

MEASUREMENTS

Filled benzodiazepine prescription during buprenorphine treatment was the main independent variable. The primary outcome was time to fatal opioid overdose. Secondary outcomes were time to non-fatal opioid overdose, all-cause mortality and buprenorphine discontinuation. We defined buprenorphine discontinuation as having a 30-day gap without another prescription following the end date of the previous prescription. We used Cox proportional hazards models to calculate hazards ratios that tested the association between receipt of benzodiazepines and all outcomes, restricted to periods during buprenorphine treatment.

FINDINGS

Of the 63 345 individuals who received buprenorphine, 24% filled at least one benzodiazepine prescription during buprenorphine treatment. Thirty-one per cent of the 183 deaths from opioid overdose occurred when individuals received benzodiazepines during buprenorphine treatment. Benzodiazepine receipt during buprenorphine treatment was associated with an increased risk of fatal opioid overdose adjusted hazard ratio (HR) = 2.92, 95% confidence interval (CI) = 2.10-4.06, non-fatal opioid overdose, adjusted HR = 2.05, 95% CI, 1.68-2.50, all-cause mortality, adjusted HR = 1.90, 95% CI, 1.48-2.44 and a decreased risk of buprenorphine discontinuation, adjusted HR = 0.87, 95% CI, 0.85-0.89.

CONCLUSIONS

Benzodiazepine receipt appears to be associated with both increased risk of opioid overdose and all-cause mortality and decreased risk of buprenorphine discontinuation among people receiving buprenorphine.

The associated press stylebook changes and the use of addiction-related stigmatizing terms in news media

Background: The May 2017 publication of the updated Associated Press (AP) Stylebook offered guidance that advised against stigmatizing. The objective of this study was to assess the frequency of stigmatizing terms describing substance use and disorder in news articles before and after the update of the AP Stylebook. Methods: We reviewed articles containing terms "opioid" or "addiction" from three major news outlets. We counted the number of AP Stylebook proscribed terms found in each article and compared the proportions of articles from each outlet with proscribed terms before and after AP Stylebook publication. Results: In 2016, 56-94% of articles from each of the three news outlets contained a proscribed term. The use of proscribed terms in articles identified by searching "opioid" published in the New York Times decreased (72% vs. 94%, p = 0.01) after the AP Stylebook change. For other news outlets, there were no significant differences, though all proportions were lower after publication. Conclusions: Evidence for a decrease in the use of stigmatizing terminology for substance use and disorders in news articles after publication of guidance was limited. Additional efforts should address use of such terminology to maximize implementation of effective addiction health policies and practices.

Alcohol and Bone Turnover Markers among People Living with HIV and Substance Use Disorder

BACKGROUND

Although unhealthy alcohol use and low bone density are prevalent among people living with HIV (PLWH), it is not clear whether alcohol use is associated with bone turnover markers (BTMs), and if so, at what quantity and frequency. The study objective was to examine the association between alcohol and BTMs in PLWH with substance use disorder.

METHODS

We studied a prospective cohort recruited from 2 HIV clinics who met criteria for DSM-IV substance dependence or reported ever injection drug use. Outcomes were BTM of (i) bone formation (serum procollagen type 1 N-terminal propeptide [P1NP]) and (ii) bone resorption (serum C-telopeptide type 1 collagen [CTx]). Alcohol consumption measures included (i) mean number of drinks/d (Timeline Follow-Back [TLFB]) (primary predictor), (ii) any alcohol use on ≥20 of the past 30 days, and phosphatidylethanol (PEth), a biomarker of recent alcohol consumption. Linear regression analysis examined associations between (i) each alcohol measure and each BTM and (ii) change in alcohol and change in BTM over 12 months.

RESULTS

Among 198 participants, baseline characteristics were as follows: The median age was 50 years; 38% were female; 93% were prescribed antiretroviral medications; 13% had ≥20 drinking days/month; mean drinks/day was 1.93 (SD 3.89); change in mean drinks/day was -0.42 (SD 4.18); mean P1NP was 73.1 ng/ml (SD 34.5); and mean CTx was 0.36 ng/ml (SD 0.34). Higher drinks/day was significantly associated with lower P1NP (slope -1.09 ng/ml; 95% confidence interval [CI] -1.94, -0.23, per each additional drink). On average, those who drank on ≥ 20 days/month had lower P1NP (-15.45 ng/ml; 95% CI: -26.23, -4.67) than those who did not. Similarly, PEth level ≥ 8ng/ml was associated with lower P1NP. An increase in drinks/d was associated with a decrease in P1NP nonsignificantly (-1.14; 95% CI: -2.40, +0.12; p = 0.08, per each additional drink). No significant associations were detected between either alcohol measure and CTx.

CONCLUSIONS

In this sample of PLWH with substance use disorder, greater alcohol consumption was associated with lower serum levels of bone formation markers.

Development of a tailored, telehealth intervention to address chronic pain and heavy drinking among people with HIV infection: integrating perspectives of patients in HIV care

BACKGROUND

Chronic pain and heavy drinking commonly co-occur and can influence the course of HIV. There have been no interventions designed to address both of these conditions among people living with HIV (PLWH), and none that have used telehealth methods. The purpose of this study was to better understand pain symptoms, patterns of alcohol use, treatment experiences, and technology use among PLWH in order to tailor a telehealth intervention that addresses these conditions.

SUBJECTS

Ten participants with moderate or greater chronic pain and heavy drinking were recruited from a cohort of patients engaged in HIV-care (Boston Alcohol Research Collaborative on HIV/AIDS Cohort) and from an integrated HIV/primary care clinic at a large urban hospital.

METHODS

One-on-one interviews were conducted with participants to understand experiences and treatment of HIV, chronic pain, and alcohol use. Participants' perceptions of the influence of alcohol on HIV and chronic pain were explored as was motivation to change drinking. Technology use and treatment preferences were examined in the final section of the interview. Interviews were recorded, transcribed and uploaded into NVivo^® v12 software for analysis. A codebook was developed based on interviews followed by thematic analysis in which specific meanings were assigned to codes. Interviews were supplemented with Likert-response items to evaluate components of the proposed intervention.

RESULTS

A number of themes were identified that had implications for intervention tailoring including: resilience in coping with HIV; autonomy in health care decision-making; coping with pain, stress, and emotion; understanding treatment rationale; depression and social withdrawal; motives to drink and refrain from drinking; technology use and capacity; and preference for intervention structure and style. Ratings of intervention components indicated that participants viewed each of the proposed intervention content areas as "helpful" to "very helpful". Videoconferencing was viewed as an acceptable modality for intervention delivery.

CONCLUSIONS

Results helped specify treatment targets and provided information about how to enhance intervention delivery. The interviews supported the view that videoconferencing is an acceptable telehealth method of addressing chronic pain and heavy drinking among PLWH.

Strategies used by people who inject drugs to avoid stigma in healthcare settings

BACKGROUND

People who inject drugs (PWID) have limited engagement in healthcare services and report frequent experiences of stigma and mistreatment when accessing services. This paper explores the impact of stigma against injection drug use on healthcare utilization among PWID in the U.S. Northeast.

METHODS

We recruited PWID through community-based organizations (CBOs; e.g., syringe service programs). Participants completed brief surveys and semi-structured interviews lasting approximately 45 min exploring HIV risk behaviors and prevention needs. Thematic analysis examined the emergent topic of stigma experiences in relation to healthcare utilization.

RESULTS

Among 33 PWID (55% male; age range 24-62 years; 67% White; 24% Latino), most used heroin (94%) and injected at least daily (60%). Experiences of dehumanization in healthcare settings were common, with many participants perceiving that they had been treated unfairly or discriminated against due to their injection drug use. As participants anticipated this type of stigma from healthcare providers, they developed strategies to avoid it, including delaying presenting for healthcare, not disclosing drug use, downplaying pain, and seeking care elsewhere. In contrast to large institutional healthcare settings, participants described non-stigmatizing environments within CBOs, where they experienced greater acceptance, mutual respect, and stronger connections with staff.

CONCLUSIONS

Stigma against injection drug use carries important implications for PWID health. Increased provider training on addiction as a medical disorder could improve PWID healthcare experiences, and integrating health services into organizations frequented by PWID could increase utilization of health services by this population.

SUN-540 No Effect Of Alcohol Use On Bone Microarchitecture Of The Distal Radius Among HIV Infected Adults

The effect of alcohol consumption on cortical and trabecular microarchitecture is not well-understood particularly among patients living with HIV who are at higher risk for metabolic bone disease, osteopenia, and osteoporosis. The aim of this study was to assess the effect of heavy alcohol consumption on bone microarchitecture as measured by high-resolution peripheral quantitative computed tomography compared to low or no alcohol consumption among HIV-infected adults. We hypothesized that heavy alcohol consumption would have a detrimental impact on bone microarchitecture in both the cortical and trabecular compartments at the distal radius. We conducted a cross-sectional analysis among 238 participants (148 men, 42 pre-menopausal women, 48 post-menopausal women). Alcohol consumption status was based on the last 30 days (TimeLine Followback) and categorized into 4 categories based on the distribution in the study population: abstinence (0 grams(g)/day(d)) (n=82), lowest tertile (0.47 -7.0 g/d) (n=53), middle tertile (>7.0 - <35.0 g/d) (n=51), and highest tertile (35.0 - 454.1 g/d) (n=52). Bone microarchitecture measures included total volumetric bone mineral density (D100); Trabecular bone mineral density (Tb.BMD); Cortical bone mineral density (Ct.BMD); Trabecular number (Tb.N); Trabecular thickness (Tb.Th); Trabecular separation (Tb.Sp); Cortical thickness (Ct.Th); Cortical area (Ct.Ar). Associations between alcohol intake and bone microarchitecture measures were assessed using unadjusted and adjusted linear regression models. Covariates included sex, age, race, daily calcium intake, weight bearing physical activity, HIV viral load, number of years since first started taking antiretroviral therapy, intact reactive parathyroid hormone (iPTH), body mass index and number of years of heroin use. Among the density parameters, D100 was similar across all alcohol intake levels (adjusted p=0.75). The mean Tb.BMD was lowest among those who consumed low levels of alcohol (150.6 ± 41.5 mg HA/cm³) and highest among those who drank the heaviest (166.6 ± 48.6 mg HA/cm³). Mean Ct.BMD was highest with abstinence (854.8 ± 73.5 mg HA/cm³) followed by the lowest alcohol tertile (853.5 ± 60.5 mg HA/cm³) and then those reporting the heaviest drinking (829.4 ± 88.8 mg HA/cm³). However, the adjusted regression models showed no significant differences (p=0.94 and p=0.21, respectively) and remained non-significant after adjusting for 25-hydroxyvitamin D. There were no significant differences in the mean trabecular or cortical structural parameters including Tb.N (p=0.85), Tb.Th (p=0.90), Tb.Sp (p=0.73), Ct.Th (p=0.71), Ct.Ar (p=0.93) by alcohol intake after adjusting for covariates. We conclude that in this study population, bone microarchitecture measures did not differ by alcohol consumption status at the distal radius.

Alcohol, Age, and Mortality: Estimating Selection Bias Due to Premature Death

OBJECTIVE

Alcohol use causes approximately 10% of deaths among adults ages 20-65 in the United States. Although previous research has demonstrated differential age-related risk relationships, it is difficult to estimate the magnitude of selection bias attributable to premature mortality based on existing cohort studies, the average age of which is greater than 50 years. The objective of our study was to assess the distribution of mortality-related harms and benefits from alcohol among adults ages 20 and older in comparison with the distribution among those older than age 50.

METHOD

Data from the Centers for Disease Control and Prevention's Alcohol-Related Disease Impact software application from 2006-10 were used to determine the distribution of alcohol-attributable deaths (AADs) and the years of potential life lost (YPLLs) that was caused or prevented by alcohol for 54 conditions by 15-year age groupings (20-34, 35-49, 50-64, 65+) in the United States. We also determined the proportion of net deaths and YPLLs occurring in each age group, overall and by cause of death.

RESULTS

Adults ages 20-49 years experienced 35.8% of the deaths and 58.4% of the YPLLs caused by alcohol, whereas the same group accrued only 4.5% of AADs and 14.2% of YPLLs gained. Overall, 46.3% of the total net deaths and 64.7% of the net YPLLs occurred among those ages 20-49; adding net deaths occurring among those ages 20-49 to those occurring after age 50 would result in an 86.3% relative increase in net deaths.

CONCLUSIONS

Because of premature mortality, alcohol-mortality associations based on cohort studies may underestimate negative health consequences compared with those observed among the general population.

Transitions To and From At-Risk Alcohol Use in Adults in the United States

Introduction

The objective of this research was to study transitions to and from at-risk alcohol use.

Methods

Logistic regression analyses (done 2015-2016) assessed transitions to and from past-year at-risk drinking in a representative sample of U.S. adults surveyed twice (in 2001-2 and 2004-5).

Results

Among 34,653 adults, 28% reported at-risk use at time 1. Of those, 73% had at-risk use at time 2. Of those without at-risk use at time 1, 15% reported at-risk use at time 2. Positive high-risk drinking transition predictors were, at time 1, being young, male, white, childless, in good to excellent health, ever smoking, using drugs, military membership (time 1 but not 2), and becoming divorced or separated by time 2. Positive low-risk drinking transition predictors were being elderly (age ≥65), female, non-white, never smoking or using drugs, no alcohol use disorder, alcohol treatment, and, after time 1, having children.

Conclusions

Many adults transition to and from at-risk alcohol use; youth is the strongest positive predictor of transition to at-risk and not transitioning to low-risk drinking. Persons transitioning to legal drinking age are most likely to transition to high-risk and least likely to low-risk drinking.

Alcohol Consumption and Bone Mineral Density in People with HIV and Substance Use Disorder: A Prospective Cohort Study

BACKGROUND

People living with HIV (PLWH) commonly have low bone mineral density (BMD) (low bone mass and osteoporosis) and are at high risk for fractures. Fractures and low BMD are significant causes of morbidity and mortality, increasingly relevant as PLWH age. Alcohol use is common among PLWH and known to affect bone health. The association between alcohol use and changes in BMD among PLWH is not well understood.

METHODS

We conducted a 3.5-year prospective cohort study of 250 PLWH with substance use disorder or ever injection drug use. Annual alcohol consumption was measured as a mean of grams per day of alcohol, mean number of heavy drinking days per month, mean number of days abstinent per month, and any heavy drinking, using the 30-day Timeline Followback method twice each year. The primary outcome was annual change in BMD measured each year by dual energy X-ray absorptiometry in grams per square centimeter (g/cm2 ) at the femoral neck. Additional dependent variables included annual change in total hip and lumbar spine BMD, >6% annual decrease in BMD at any site, and incident fractures in the past year. Regression models adjusted for relevant covariates.

RESULTS

The median age of participants was 50 years. The median duration of HIV infection was 16.5 years and the mean time since antiretroviral therapy initiation was 12.3 years. At study entry, 67% of participants met criteria for low BMD (46% low bone mass, 21% osteoporosis). Median follow-up was 24 months. We found no significant associations between any measure of alcohol consumption and changes in BMD (g/cm2 ) at the femoral neck (adjusted β for g/d of alcohol = -0.0032, p = 0.7487), total hip, or lumbar spine. There was no significant association between any measure of alcohol consumption and >6% annual decrease in BMD at any site, or incident fractures.

CONCLUSIONS

In this sample of PLWH and substance use disorders or ever injection drug use, we detected no association between any of the alcohol measures used in the study and changes in BMD or incident fractures.

Massachusetts Health Reform's Effect on Hospitalizations with Substance Use Disorder-Related Diagnoses

OBJECTIVE

To examine whether Massachusetts (MA) health reform affected substance (alcohol or drug) use disorder (SUD)-related hospitalizations in acute care hospitals.

DATA/STUDY SETTING

2004-2010 MA inpatient discharge data.

DESIGN

Difference-in-differences analysis to identify pre- to postreform changes in age- and sex-standardized population-based rates of SUD-related medical and surgical hospitalizations, adjusting for secular trends.

DATA EXTRACTION METHODS

We identified 373,751 discharges where a SUD-related diagnosis was a primary or secondary discharge diagnosis.

FINDINGS

Adjusted for age and sex, the rates of drug use-related and alcohol use-related hospitalizations prereform were 7.21 and 8.87 (per 1,000 population), respectively, in high-uninsurance counties, and 8.58 and 9.63, respectively, in low-uninsurance counties. Both SUD-related rates increased after health reform in high- and low-uninsurance counties. Adjusting for secular trends in the high- and low-uninsurance counties, health reform was associated with no change in drug- or alcohol-related hospitalizations.

CONCLUSIONS

Massachusetts health reform was not associated with any changes in substance use disorder-related hospitalizations. Further research is needed to determine how to reduce substance use disorder-related hospitalizations, beyond expanding insurance coverage.

Implementing single-item screening for drug use in a Veterans Health Administration outpatient setting

BACKGROUND

Unhealthy drug use is a concern in many settings, including military and veteran populations. In 2013, the Veterans Health Administration (VHA) medical center in Bedford, Massachusetts, started requiring routine screening for unhealthy drug use in outpatient primary care and mental health settings, using a validated single question.

METHODS

This study used descriptive and multivariable analyses of VHA electronic records for patients eligible for the screening program (N = 16,118). The study assessed first-year rates and predictors of screening and of positive screens, both for drug use and for unhealthy alcohol use, for which screening was already required.

RESULTS

During the first year, 70% of patients were screened for unhealthy drug use and 84% were screened for unhealthy alcohol use. In multivariable analyses, screening for drug use was more likely for patients who had 8 or more days with VHA visits or were aged 40 or over. Patients with a prior drug use disorder diagnosis were much less likely to be screened. Three percent of patients screened for unhealthy drug use had a positive screen, and 14% of those screened for unhealthy alcohol use had a positive screen. Strong predictors of a positive drug use screen included a prior-year diagnosis of drug use disorder, any mental health clinic visits, younger age, or being unmarried.

CONCLUSIONS

The drug screening initiative was relatively successful in its first-year implementation, having screened 70% of eligible subjects. However, it failed to screen many of those most likely to screen positive, thereby missing many opportunities to address unhealthy drug use. Future refinements should include better training clinicians in how to ask sensitive questions and how to address positive screens.

Development and differentiability of three brief interventions for risky alcohol use that include varying doses of motivational interviewing

BACKGROUND

While brief intervention (BI) for risky alcohol use generally yields positive effects among those identified by screening, effect sizes are small and there is unexplained heterogeneity in outcome. The heterogeneity may be related to differences in intervention style and content, including elements of motivational interviewing (MI). To date, it has been difficult to interpret the role of MI in BI and these gaps in knowledge interfere with efforts to train, disseminate and implement BI that retains and maximizes efficacy. This study sought to develop BI protocols with varying doses of MI and test their differentiability. Differentiable BI protocols could allow for future studies that prospectively evaluate the role MI plays in affecting BI outcome.

METHODS

We developed three intervention protocols: brief advice, standard BI (NIAAA Clinician's Guide), and MI-enhanced BI and administered them to 45 primary care patients who reported exceeding recommended drinking limits. We then rated the BI sessions for fidelity to the assigned protocol as well as MI consistency based on Motivational Interviewing Treatment Integrity (MITI) scale scores. The differentiability of BI protocols was determined by calculating fidelity to assigned protocols and comparing MITI scores using pairwise, Tukey-adjusted comparisons of least squares mean scores.

RESULTS

High rates of fidelity to each protocol were achieved. The three BI protocols were also highly differentiable based on MITI scores.

CONCLUSIONS

The three interventions can be used in future trials to prospectively examine the role MI has in determining BI outcome. Trial registration clinicaltrials.gov NCT02978027, retrospectively registered 11/28/16.

Perceived discrimination and drug involvement among black primary care patients who use drugs

Perceived discrimination has been associated with disparities for Black patients on a variety of health outcomes. Studies have suggested that perceived discrimination is associated with drug use in Blacks, but they have been limited by use of samples with little drug use and single measures of drug involvement. The current study examined the association between perceived discrimination and multiple measures of drug involvement among a sample of 203 Black adult primary care patients who were participants in a randomized trial of screening and brief intervention for drug use. The main independent variable was everyday perceived discrimination. The three outcomes were frequency of drug use in the past ninety days, drug-related consequences, and total drug involvement risk severity score from the Alcohol, Smoking, and Substance Involvement Test [ASSIST]. Analyses were conducted using negative binomial regression models for frequency and consequence outcomes and median regression models for drug involvement risk. Greater perceived discrimination was not significantly associated with frequency of use, but was associated with more drug-related consequences and a higher drug use risk level. These findings suggest that perceived discrimination may be an important variable to consider when selecting drug intervention approaches for Black primary care patients.

Effect of electronic brief intervention on uptake of specialty treatment in hospital outpatients with likely alcohol dependence: Pilot randomized trial and qualitative interviews

BACKGROUND

A large proportion of hospital outpatients are alcohol dependent (AD) but few are engaged in treatment for their drinking. Brief intervention, designed to raise patients' awareness of their drinking, might encourage uptake of referral to specialty treatment. We assessed the feasibility of conducting a randomized trial evaluating the effectiveness of electronic brief intervention on the uptake of specialty treatment in hospital outpatients with likely AD.

METHODS

This study was conducted in the outpatient department of a large public hospital in Newcastle, Australia. We randomly assigned adults who scored ≥10 on the AUDIT-C and were not currently receiving treatment for their drinking to electronic brief intervention (comprising an assessment of their drinking and personalized feedback) and referral (n = 59), or to referral alone (n = 64). We pre-specified two co-primary outcomes as the proportions of patients who (1) accepted and (2) attended a Drug and Alcohol outpatient clinic appointment. We interviewed 15 study participants to investigate why they had declined the appointment and what sort of help they might prefer to receive.

RESULTS

Ten patients (five in each group) accepted an appointment, and one patient (control) attended. Most interviewees' did not see their drinking as a problem or were confident they could manage it by themselves. Those who identified a preferred source of help expressed a preference for treatment by a GP.

CONCLUSION

Uptake of specialty treatment in hospital outpatients with likely AD was low regardless of whether they received brief intervention. Accordingly, a large randomized trial does not appear to be feasible.

Lifetime marijuana and alcohol use, and cognitive dysfunction in people with human immunodeficiency virus infection

BACKGROUND

Substance use is common among people with human immunodeficiency virus (HIV) infection. Alcohol, marijuana, and HIV can have negative effects on cognition. Associations between current and lifetime marijuana and alcohol use and cognitive dysfunction in people with HIV infection were examined.

METHODS

Some 215 HIV-infected adults with Diagnostic and Statistical Manual of Mental Disorders Fourth Edition (DSM-IV) substance dependence or ever injection drug use were studied. In adjusted cross-sectional regression analyses associations were assessed between current marijuana use, current heavy alcohol use, lifetime marijuana use, lifetime alcohol use, duration of heavy alcohol use (the independent variables), and 3 measures of cognitive dysfunction (dependent variables): both the (i) memory and (ii) attention domains from the Montreal Cognitive Assessment (MoCA) and the (iii) 4-item cognitive function scale (CF4) from the Medical Outcomes Study HIV Health Survey (MOS-HIV). Analyses were adjusted for demographics, primary language, depressive symptoms, anxiety, comorbidities, antiretroviral therapy, hepatitis C virus (ever), duration of HIV infection (years), HIV-viral load (log copies/mL), CD4 cell count, lifetime and recent cocaine use, and recent illicit and prescribed opioid use.

RESULTS

Current marijuana use was significantly and negatively associated with the MOS-HIV CF4 score (adjusted mean difference = -0.40, P = .01). Current marijuana use was not significantly associated with either MoCA score. Lifetime marijuana use and current heavy and lifetime alcohol use and duration of heavy alcohol use were not associated with any measure of cognitive dysfunction.

CONCLUSION

Current marijuana use was associated with one measure of cognitive dysfunction, but there was not a consistent pattern of association with lifetime marijuana use or alcohol use and measures of cognitive dysfunction. Understanding the mechanism by which marijuana, with and without alcohol, are associated with worse cognition warrants larger, longer studies with more precise and diverse measurements of cognitive function.

Polypharmacy and risk of falls and fractures for patients with HIV infection and substance dependence

Although people with HIV infection (PLWH) are at higher risk of polypharmacy and substance use, there is limited knowledge about potential harms associated with polypharmacy such as falls and fractures in this population. The study objective was to determine whether polypharmacy, as measured by the number and type of medication, is associated with falls and fractures among PLWH and DSM-IV substance dependence in the past year or ever injection drug use (IDU). We identified the number of medications by electronic medical record review in the following categories: (i) systemically active, (ii) non-antiretroviral (non-ARV), (iii) sedating, (iv) non-sedating as well as any opioid medication and any non-opioid sedating medication. Outcomes were self-reported (1) fall/accident requiring medical attention and (2) fracture in the previous year. Separate logistic regression models were fitted for medications in each category and each outcome. Among 250 participants, the odds of a fall requiring medical attention were higher with each additional medication overall (odds ratio [OR] 1.12, 95% Confidence Interval [CI] = 1.05, 1.18), each additional non-ARV medication (OR 1.13, 95%CI = 1.06, 1.20), each additional sedating medication (OR 1.36, 95%CI = 1.14, 1.62), and a non-opioid sedating medication (OR 2.89, 95%CI = 1.06, 7.85) but not with an additional non-sedating medication or opioid medication. In receiver operating characteristic (ROC) curve analyses, optimal cutoffs for predicting falls were: ≥8 overall and ≥2 sedating medications. Odds ratios for fracture in the previous year were OR 1.05, 95%CI = 0.97, 1.13 for each additional medication overall and OR 1.11, 95%CI = 0.89, 1.38 for each additional sedating medication. In PLWH and substance dependence or ever IDU, a higher number of medications was associated with greater odds of having a fall requiring medical attention. The association appeared to be driven largely by sedating medications. Future studies should determine if reducing such polypharmacy, particularly sedating medications, lowers the risk of falls.