Self-inflicted gunshot wounds to the face are one of the most challenging clinical scenarios encountered by oral and maxillofacial surgeons. Knowledge is lacking regarding which factors might influence survival after these devastating injuries, especially pertaining to psychiatric history and substance use. The purpose of the present study was to evaluate the risk factors that might influence the survival of subjects with self-inflicted gunshot wounds to the face.
Materials and Methods
A retrospective cohort study was designed to analyze the data from subjects presenting to the University of Louisville Trauma Center with self-inflicted gunshot wounds to the face from February 2010 to September 2019. The predictor variables included demographic (eg, age, gender, race), medical and psychiatric history, and toxicology test results. The primary outcome variable was death before hospital discharge. Descriptive, bivariate, and logistic regression models were computed.
The sample included 120 subjects, with an age range of 16 to 85 years old (average age, 43.5 years); 90.8% were male, and 56.7% had survived their suicide attempt. Of the 120 patients, 35% had a history of depression, 23.3% tested positive for benzodiazepines, and 33% had a social history positive for smoking, alcohol use, and/or drug use. Depression was the single largest predictor of mortality. Patients with depression were significantly more likely to survive their injuries than were patients without depression (odds ratio, 0.230; P = .003). The presence of benzodiazepines in toxicology tests was also a significant predictor of mortality (odds ratio, 0.297; P = .018); patients testing positive were more likely to survive than were patients with negative test results or positive test results for other drugs.
Subjects who attempt suicide via self-inflicted gunshot wounds to the face were more likely to survive their injury if they had a reported history of depression or test results positive for a benzodiazepine.
Gun violence is a public health crisis. Of the 47,173 suicide attempts in 2017, 23,854 were with a firearm. However, the incidence of deaths from firearm-related suicide in the United States has been increasing. In addition, it has been estimated that up to 95% of individuals who attempt suicide will have a diagnosed mental health disorder, with most being mood disorders (eg, major depression, bipolar disorder). The results from the 2017 National Survey on Drug Use and Health revealed that in 2017 alone, it was estimated that 17.3 million adults had experienced at least 1 major depressive episode, representing 7.1% of all US adults. Considering that self-inflicted gunshot wounds to the face represent one of the most challenging clinical scenarios encountered by oral and maxillofacial surgeons, several studies have investigated the injury patterns, complications, survival, and gun orientation. However, few studies have focused on the psychiatric history and substance use patterns within the context of self-inflicted gunshot wounds.
The purpose of the present study was to determine the risk factors associated with mortality in subjects with a self-inflicted gunshot wound to the face, with specific interest in the psychiatric history, social history, and substance use results. The null hypothesis was that no statistically significant difference would be found between the predictor variables and the primary outcome variable of death. The specific aims of the present study were to 1) identify those subjects who had sustained a self-inflicted gunshot wound to the face at the University of Louisville Trauma Center; 2) record pertinent variables; and 3) identify those variables that might be associated with an increased or a decreased mortality rate.
Materials and Methods
Study Design and Sample
A retrospective cohort study was designed and implemented using patients from a level 1 tertiary care center. Once the institutional review board granted approval (approval no. 19.1107), a search of the medical center's trauma database for facial gunshot wounds was completed. Data were collected from the patients who presented to the trauma center from February 2010 to September 2019. For a subject to be included in the present study, the injury had to have been a gunshot wound to the face only and the mechanism of injury documented as self-inflicted. Patients were excluded if the mechanism of injury had not been self-inflicted, if the injury had not been to the face, if the records were not complete, or the patients were dead on arrival.
The predictor variables included demographic data (ie, age, gender, race), medical and psychiatric history, and toxicology test results. The use of controlled substances was determined from the results of blood laboratory tests that had been completed on immediate patient arrival to the trauma center; if a substance had been given by emergency medical services en route to the hospital, the test results were not included. The social history and psychiatric history were derived from the reported admission history and physical examination findings. The primary outcome variable was death before hospital discharge.
Data Collection, Management, and Analysis
De-identified data were exported to a standardized database software (Excel, version 14.4.2; Microsoft Corp, Redmond, WA). Descriptive statistics were calculated for each variable. Categorical variables are reported using counts and percentages. Continuous variables are reported as the mean ± standard deviation. Nonparametric tests (eg, χ 2 tests or Fisher's exact tests) were used to identify dependency between the rates of mortality and patient demographic, psychiatric, and substance abuse information among the patients with self-inflicted injuries. All statistical tests were assessed at the P = .05 significance level. Logistic regression models were also used to identify significant relationships between mortality and the predictor variables to quantify the effects of these patient characteristics on their survival probability. Stepwise regression analysis was used to identify significant covariates and determine optimal models to predict mortality. All statistical tests and modeling were performed using the statistical software R (R Foundation for Statistical Computing, Vienna, Austria).