Important statistical considerations in the evaluation of post-market studies to assess whether opioids with abuse-deterrent properties result in reduced abuse in the community.

PURPOSE: Abuse, misuse, addiction, overdose, and death associated with non-medical use of prescription opioids have become a serious public health concern. Reformulation of these products with abuse-deterrent properties is one approach for addressing this problem. FDA has approved several extended-release opioid analgesics with abuse-deterrent labeling, the bases of which come from pre-market studies. As all opioid analgesics must be capable of delivering the opioid in order to reduce pain, abuse-deterrent properties do not prevent abuse, nor do pre-market evaluations ensure that there will be reduced abuse in the community. Utilizing data from various surveillance systems, some recent post-market studies suggest a decline in abuse of extended-release oxycodone after reformulation with abuse-deterrent properties. We discuss challenges stemming from the use of such data. METHODS: We quantify the relationship between the sample, the population, and the underlying sampling mechanism and identify the necessary conditions if valid statements about the population are to be made. The presence of other interventions in the community necessitates the use of comparators. We discuss the principles under which the use of comparators can be meaningful. CONCLUSIONS: Results based on surveillance data need to be interpreted with caution as the underlying sampling mechanisms can bias the results in unpredictable ways. The use of comparators has the potential to disentangle the effect due to the abuse-deterrence properties from those due to other interventions. However, identifying a comparator that is meaningful can be very difficult.

Patterns of Immediate-Release and Extended-Release Opioid Analgesic Use in the Management of Chronic Pain, 2003-2014.

Importance: Many stakeholders are working to improve the safe use of immediate-release (IR) and extended-release/long-acting (ER/LA) opioid analgesics. However, little information exists regarding the relative use of these 2 formulations in chronic pain management. Objectives: To describe the distribution of IR and ER/LA opioid analgesic therapy duration and examine adding and switching patterns among patients receiving long-term IR opioid analgesic therapy, defined as at least 90 consecutive days of IR formulation use. Design, Setting, and Participants: A retrospective cohort study of 169 million individuals receiving opioid analgesics from across 90% of outpatient retail pharmacies in the United States from January 1, 2003, to December 31, 2014, using the IQVIA Health Vector One: Data Extract Tool. Analyses were conducted from March 2015 to June 2017. Exposures: Receipt of dispensed IR or ER/LA opioid analgesic prescription. Main Outcomes and Measures: Distribution of therapy frequency and duration of IR and ER/LA opioid analgesic use, and annual proportions of patients receiving long-term IR opioid analgesic therapy who added an ER/LA formulation while continuing to use an IR formulation, switched to an ER/LA formulation, or continued receiving IR opioid analgesic therapy only. Results: Among the 169 280 456 patients included in this analysis, 168 315 458 patients filled IR formulations and 10 216 570 patients filled ER/LA formulations. A similar percentage of women received ER/LA (55%) and IR (56%) formulations, although those receiving ER/LA formulations (72%) were more likely to be aged 45 years or older compared with those receiving IR formulations (46%). The longest opioid analgesic episode duration was 90 days or longer for 11 563 089 patients (7%) filling IR formulations and 3 103 777 patients (30%) filling ER/LA formulations. The median episode duration was 5 days (interquartile range, 3-10 days) for patients using IR formulations and 30 days (interquartile range, 21-74 days) for patients using ER/LA formulations. From January 1, 2003, to December 31, 2014, a small and decreasing proportion of patients with long-term IR opioid analgesic therapy added (3.8% in 2003 to 1.8% in 2014) or switched to (1.0% in 2003 to 0.5% in 2014) an ER/LA formulation. Conclusions and Relevance: Most patients receiving opioid analgesics, whether for short or extended periods, use IR formulations. Once receiving long-term IR opioid analgesic therapy, patients are unlikely to add or switch to an ER/LA formulation.

Trends in the Concomitant Prescribing of Opioids and Benzodiazepines, 2002-2014.

INTRODUCTION: Although many clinical guidelines caution against the combined use of opioids and benzodiazepines, overdose deaths and emergency department visits involving the co-ingestion of these drugs are increasing. METHODS: In this ecologic time series study, the IMS Health Total Patient Tracker was used to describe nationally projected trends of patients receiving opioids and benzodiazepines in the U.S. outpatient retail setting between January 2002 and December 2014. The IMS Health Data Extract Tool was used to examine trends in the concomitant prescribing of these two medication classes among 177 million individuals receiving opioids during this period. The annual proportion of opioid recipients who were prescribed benzodiazepines concomitantly was calculated and stratified by gender, age, duration of opioid use, immediate-release versus extended-release/long-acting opioids, and benzodiazepine molecule. The proportion of patients with concomitancy receiving opioids and benzodiazepines from the same prescriber was also analyzed. Analyses were conducted from April to June 2015. RESULTS: The nationally projected number of patients receiving opioids and benzodiazepines increased by 8% and 31%, respectively, from 2002 to 2014. During this period, the annual proportion of opioid recipients dispensed a benzodiazepine concomitantly increased from 6.8% to 9.6%, which corresponded to a relative increase of 41%. Approximately half of these patients received both prescriptions from the same prescriber on the same day. Concomitancy was more common in patients receiving opioids for ≥90 days, women, and the elderly. CONCLUSIONS: Concomitant prescribing of opioids and benzodiazepines is increasing and may play a growing role in adverse patient outcomes related to these medications.

The incident user design in comparative effectiveness research.

Comparative effectiveness research includes cohort studies and registries of interventions. When investigators design such studies, how important is it to follow patients from the day they initiated treatment with the study interventions? Our article considers this question and related issues to start a dialogue on the value of the incident user design in comparative effectiveness research. By incident user design, we mean a study that sets the cohort's inception date according to patients' new use of an intervention. In contrast, most epidemiologic studies enroll patients who were currently or recently using an intervention when follow-up began. We take the incident user design as a reasonable default strategy because it reduces biases that can impact non-randomized studies, especially when investigators use healthcare databases. We review case studies where investigators have explored the consequences of designing a cohort study by restricting to incident users, but most of the discussion has been informed by expert opinion, not by systematic evidence.

Onset of acute myocardial infarction after use of non-steroidal anti-inflammatory drugs.

PURPOSE: To examine the association between cyclooxygenase-2 (COX-2) selective and traditional nonsteroidal anti-inflammatory drugs (NSAIDs) and incident acute myocardial infarction (AMI), and to address unanswered questions regarding the contour of risk over time. METHODS: A cohort of new NSAID users aged 40-84 years was followed for the occurrence of first AMI. Data were collected within the General Practice Research Database (GPRD) from 1 January 1997 to 31 December 2004. RESULTS: The study population included 1185 AMI events (889 probable and 296 possible) from a cohort of 283 136 patients. After adjustment for demographic and cardiovascular risk factors, the hazard ratio (HR) for AMI was significantly increased for both coxib (2.11, 95% confidence interval (CI) 1.04-4.26) and non-coxib (2.24, 95%CI 1.13-4.42) COX-2 selective NSAIDs when compared to remote exposure to NSAIDs, but was not increased for traditional NSAIDs. Stratifying exposure into the first month of use versus use beyond 1 month, the risk of AMI was increased during the first month of COX-2 selective NSAIDs use, but not later (3.43, 95%CI 1.66-7.07 and 1.88, 95%CI 0.82-4.31, respectively p-value for interaction = 0.6). CONCLUSIONS: The results suggest that the use of coxib and non-coxib COX-2 selective NSAIDs was associated with an elevated risk of AMI within the first month of exposure. Recent past exposure to NSAID was not associated with a similar increase in risk.

Incident diabetes associated with antipsychotic use in the United Kingdom general practice research database.

BACKGROUND: Recent reports suggest an association between antipsychotic use and development or exacerbation of diabetes. This study evaluated the risk of incident diabetes associated with the use of atypical and conventional antipsychotics. METHOD: This nested case-control study included all patients in the U.K. General Practice Research Database treated with antipsychotic drugs between January 1994 and December 1998. The main outcome measures were the odds ratios of current (within prior 6 months) or recent (7 to 12 months) antipsychotic exposure among those with (N = 424) compared with those without incident diabetes (N = 1522). RESULTS: The adjusted odds ratio for current use of any antipsychotic drug compared with no use in the past year among those with diabetes was 1.7 (95% confidence interval [CI] = 1.3 to 2.3). The adjusted odds ratio for current use of atypical and conventional antipsychotic drugs compared with no use in the past year among those with diabetes was 4.7 (95% CI = 1.5 to 14.9) and 1.7 (95% CI = 1.2 to 2.3), respectively. The adjusted odds ratio for recent use of conventional antipsychotic drugs compared with no use in the past year among those with diabetes was 1.0 (95% CI = 0.6 to 1.6). The odds ratio for recent atypical antipsychotic drug use could not be calculated because no study subjects had this exposure. CONCLUSION: This study showed an increased risk of incident diabetes among current users of atypical and conventional antipsychotic medications. These results were independent of other established risk factors. The larger association observed for atypical antipsychotic users should be regarded as preliminary given the small number of incident diabetes cases in this group.