316 Use of Opioids and Outcomes of Pneumonia: Results from the United States Nationwide Inpatient Sample

Abstract

The prevalence of opioid abuse has increased significantly in the past few decades. According to the data of 2017 National Survey on Drug Use and Health, 11.4 million Americans reported misuse of opioid prescriptions, a dramatic increase from 4.9 million in 1992. One recent Centers for Disease Control and Prevention (CDC) report also showed that emergency department visits for opioid overdoses rose 30% in all parts of the US from July 2016 through September 2017. Opioids have some adverse effects on the respiratory system, like respiratory depression, sedation with increasing risk of aspiration, reduced mucus clearance due to cough suppression, and even immunosuppressive effects. The aim of this study was to compare outcomes of pneumonia among patients with an opioid disorder, other drug dependence, and patients without drug dependence problems. Given the challenges in treating these patients, we hypothesized that the patients with drug use disorder were at an increased risk in developing comorbidity, health care expenses, and length of stay. One outcome that may be of particular interest is the length of stay (LOS), considering its significant impact on health care costs and its ability to be modulated. We conducted a US population-based cohort study using the Nationwide Inpatient Sample (NIS) database between 2005 and 2014. We defined opioid disorder with ICD-9 CM codes for opioid dependence, other drug dependence with ICD-9 CM codes for cannabis, hallucinogens, sedatives, hypnotics, or anxiolytics, cocaine, amphetamines, antidepressants, or other mixed or unspecified drugs. Pneumonia-related outcomes including pleural effusion, empyema, use of ventilator, in-hospital mortality, length of hospital stay, the total cost of hospital stay, cost per day for a hospital stay. To minimize baseline imbalance between patients with different characteristics, we performed a propensity score (PS)-matched analysis using 1:1:1 matching technique. The PS contains several variables including demographics, socioeconomic status, and comorbidities. We assessed the association between drug dependence and pneumonia mortality using Cox proportional hazard regression analysis. A total of 11,357,032 pneumonia hospitalization met the inclusion criteria, of which 11,085,783 were without drug dependence, 170,468 were with other drug dependence, and 100,781 were with opioid disorder. After propensity score matching, compared with patients without drug dependence, patients with opioid disorder and other drug dependence were associated with increasing use of ventilators, length of hospital stay, total cost and cost per day of hospital stay. By using this robust analytical approach, we found that patients with opioid disorder and other drug dependence were indeed significantly associated with increasing use of ventilators, length of hospital stay, total cost, and cost per day of hospital stay in pneumonia patients.