Hello, my name is Michael Martini. I'm presenting our work from the Icahn School of Medicine at Mount Sinai. The title of the presentation today is Increased Risk of Vasospasm Following Subarachnoid Hemorrhage in Patients with Premorbid Opioid Use Disorders, a Nationwide Analysis of Outcomes. Subarachnoid hemorrhage is the most morbid sequelae of intracranial aneurysms and often results in debilitating complications or death. Despite advances in treatment, transient cerebral ischemia, or TCI, secondary to subarachnoid hemorrhage is one of the major causes of morbidity and mortality following aneurysm rupture. While overall subarachnoid hemorrhage mortality has been declining, opioid use in the United States population has surged, and this has had a profound impact on patient health and the healthcare system, leading to increased complications, hospitalization, and death. Physicians are increasingly tasked with operating on patients with opioid use disorders, or OUDs, who present with opioid-related complications and histories of opioid addiction. However, scant information exists for how a history of opioid abuse affects subarachnoid hemorrhage outcomes. In addition, despite many studies describing vasoactive properties of opioids, no study to date has examined how chronic opioid use might impact vasculated outcomes following subarachnoid hemorrhage, particularly TCI. We hypothesize that withdrawal-like effects in hospitalized subarachnoid hemorrhage patients with pre-morbid opioid use disorders may increase the rates of adverse vascular outcomes, including TCI. This study aims to investigate the influence of clinically diagnosed OUDs on the outcomes of microsurgical clipping and endovascular coiling following acute subarachnoid hemorrhage, with a specific focus on the rate of symptomatic TCI. This information may help inform medical professionals of ways to improve the care of this patient population during the critical times surrounding subarachnoid hemorrhage treatment, and may facilitate more effective dialogue on this topic in the realms of healthcare delivery, health policy, and research. Data was acquired from the 2012-2014 National Inpatient Sample, or NIS, which includes data for over 7 million hospitalizations annually, making it the largest all-payer national database. As such, the NIS has been used by several studies in the literature to examine questions surrounding subarachnoid hemorrhage. Patients were included in this study if they had ICD-9 codes for subarachnoid hemorrhage and underwent either a microsurgical clipping or endovascular coiling procedure. Patients were included in the opioid use disorder cohort if they also carried a diagnosis of opioid dependence or non-dependent opioid abuse, along with a diagnosis of subarachnoid hemorrhage. Patients were excluded if they had a diagnosis of a cerebrovascular malformation or cerebral arteritis. Patients who underwent treatment of an AVM through surgery or radiosurgery were also excluded. A symptomatic TCI event was recorded if the patient had a diagnosis of a TCI event by ICD-9 coding that produced a clinical deficit. This is code 435.x, which reflects methodologies by previous studies in the literature. This definition approximates the preferred definition provided by Vergawan et al., which is clinical deterioration caused by transient cerebral ischemia. The coding for these inclusion and exclusion criteria are in line with previously validated studies with the NIS, which have examined various questions pertaining to subarachnoid hemorrhage or opioid use disorders. In addition, a recent study by Cheng et al. found that the ICD-9 coding for cerebral transient ischemic attacks displayed a greater than 90% concordance with actual clinical diagnoses documented by physicians for patients in the Paul Coverdale National Acute Stroke Program, a large registry of stroke-related cases established by the CDC, suggesting that these codes are accurately capturing the correct patient population. Data was also collected on patients' age, sex, race, insurance status, discharge status, hospital census division, admission type, and other measures. In addition, the zip code was used as—median income by zip code was used as a proxy for socioeconomic status. The Elixhauser Comorbidity Index, or ECI, which is a measure of global comorbidity burden, was calculated using the administrative coding methodology described by Kwan et al. and used as a covariate in multivariate analyses. In addition, the NIS SAW Severity Scale, or NISSS, is an externally validated metric described by Washington et al. with a strong correlation with Hunt-Hesgrade and has been used regularly in NIS studies exploring subarachnoid hemorrhage. The NISSSS was used to adjust for severity across cases in the present study. The primary outcome was the incidence of symptomatic TCI following treatment for subarachnoid hemorrhage. Secondary outcomes included the incidence of overall complication, in-hospital mortality, and discharge disposition. The NIS Subarachnoid Hemorrhage Outcome Measure, or NISSOM, an externally validated binary metric of subarachnoid hemorrhage outcome, was also considered as a measure of patient outcome. A poor outcome designated by the NISSOM is highly correlated with MRS scores greater than or equal to 3. The results from this investigation are as follows. A total of 25,330 patients undergoing either microsurgical clipping or endovascular coiling procedures for subarachnoid hemorrhage were included in this study. Of these, 310 patients, or 1.2% of the population, also carried a clinical diagnosis of opioid use disorders. Members of the OUD cohort were more likely to be younger than 65 years old, female, and insured by Medicaid, compared to patients without OUDs. There was a greater proportion of OUD patients treated at hospitals in certain census divisions, including the Middle Atlantic and Pacific regions, which are both regions known to be heavily impacted by the opioid epidemic. Of note, however, patients cohorts did not differ significantly in terms of median income by zip code, which has been used as an indirect proxy of socioeconomic status, nor did they differ by admission type, treatment technique, risk mortality, and disease severity scores, or subarachnoid hemorrhage severity, as measured by the NISSSS. In addition to using the Elixhauser Comorbidity Index as a measure of glomal comorbidity burden, patient cohorts were profiled for a range of common surgical comorbidities to better understand the clinical baselines of the cohorts. This is shown in this figure. The OUD patient cohort had significantly higher rates of AIDS, alcohol abuse, liver disease, chronic lung disease, psychosis, and weight loss. Importantly, the patient cohorts did not differ significantly in any neurological or cardiovascular surgical comorbidities at baseline, which may predispose to stroke or complicate aneurysm or subarachnoid hemorrhage treatment. This table displays the frequencies and univariate odds ratios of the primary and secondary study outcomes by opioid use disorder status for the total study population. In the total study population, 30.6% of OUD patients experienced a TCI event following subarachnoid hemorrhage treatment, compared to only 18.7% of non-OUD patients, representing a p-value of 0.012. Similarly, OUD patients faced significantly higher mean total charges for their hospitalization compared to non-OUD patients. Of note, the patient cohorts did not differ significantly in any of the other primary or secondary outcomes, including total complication rate, in-hospital mortality rate, poor functional outcome by the NIS-SOM metric, rate of non-home discharge, or mean hospitalization length. Next, a more global perspective of complication risks was acquired by examining a cold set of common postoperative complications classified by organ system, as shown in this figure. Interestingly, this univariate analysis revealed that OUD patients experienced significantly increased odds of infectious complications, but not neurological, cardiac, pulmonary, renal, gastrointestinal, or hematological or circulatory complications. In addition, we performed a more focal analysis of the neurological complications, shown on the left, and found that the cranial nerve palsy presented with an increased frequency in the OUD cohort. A similar focal analysis of non-neurological complications, shown on the right, revealed few associations other than that OUD patients faced higher rates of postoperative anemia. Finally, multivariate regression analysis was performed to assess the effect of OUD status on primary and secondary outcomes, as shown in this table. Models used for this analysis controlled for variables that were significant at the univariate level, as well as those that have recognized clinical significance. In the model, OUD patients faced significantly increased odds of postprocedural TCI, but not other adverse primary-secondary outcomes. This included the odds of any complication, in-hospital mortality, poor outcome by NIS-SOM, non-home discharge, extended hospitalization, or higher total charges. These findings suggest utility in screening patients for OUDs to identify individuals who may benefit from a higher level of perioperative management in the setting of subarachnoid hemorrhage. To conclude, several limitations must be acknowledged in this study. First, the retrospective study design prevents any causational analysis and may introduce potential bias into the results. Second, the NIS database is also intrinsically limited, as measures not encoded by the ICD-9 coding system could not be examined, which included potentially important measures of pain, extent of opioid use, and the amount of hemorrhage on presentation. Nevertheless, the evidence from this study suggests that OUD patients faced significantly higher odds of vasospasm events during hospitalization for acute subarachnoid hemorrhage. Future work investigating potential pharmacological mechanisms may provide a better understanding and enable future improved treatments and management guidelines for this patient population. Finally, these studies suggest utility in screening patients for OUD to identify individuals who may benefit from a higher level of clinical scrutiny for post-subarachnoid hemorrhage vasospasm. Finally, I'd like to thank everyone at Mount Sinai who made this project possible, including our study team and my co-authors, as well as the neurosurgery department leadership, including Dr. Joshua Bederson and Dr. Jay Mako. Thank you for your attention.

Icahn School of Medicine at Mount Sinai

opioid use disorders

subarachnoid hemorrhage

intracranial aneurysms

transient cerebral ischemia