Analysis of Causalities and Outcomes in Trauma Patients Who Self-Extubate

Transcription

1 Analysis of Causalities and Outcomes in Trauma Patients Who Self-Extubate Ana Negrete, PharmD, BCPS 1 Stephanie N Thompson, PhD 2 Julton Tomanguillo Chumbe, MD 3 Damayanti Samanta, MS 4 Audis Bethea, PharmD, BCPS 5 1 Methodist University Hospital anegrete1@hotmail.com 2 CAMC Health Education and Research Institute Center for Health Services and Outcome Research stethompson@camc.org 3 CAMC Health Education and Research Institute Center for Health Services and Outcome Research julton.tomanguillochumbe@camc.org 5 CAMC Health Education and Research Institute Center for Health Services and Outcome Research damayanti.samanta@camc.org 5 CAMC Health Education and Research Institute Center for Health Services and Outcome Research traumaresearchcamc@gmail.com Corresponding Author: Audis Bethea, PharmD, BCPS Clinical Research Scientist/Pharmacy Specialist, Trauma/Surgery 3200 MacCorkle Ave. SE Charleston, WV Phone , Fax audis.bethea@camc.org Conflict of Interest: The authors have no conflict of interest to report. Study Presentations: Presented as a poster at Society of Critical Care Medicine s 42 nd Annual Clinical Congress, 2013.

2 Running Head: Outcomes in Trauma Patients Who Self-Extubate Financial Disclosure Ana Negrete has no financial disclosures. Stephanie N Thompson is a member of the West Virginia Clinical Translational Science Institute. Research reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health under Award Number U54GM The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Julton Tomanguillo Chumbe is a member of the West Virginia Clinical Translational Science Institute. Research reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health under Award Number U54GM The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Damayanti Samanta is a member of the West Virginia Clinical Translational Science Institute. Research reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health under Award Number U54GM The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health. Audis Bethea is a research scientist for the West Virginia Clinical Translational Science Institute. Research reported in this publication was supported by the National Institute of General Medical Sciences of the National Institutes of Health under Award Number U54GM The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.

3 Background: Literature identifying risk factors for self-extubation in trauma patients is limited. This study endeavors to identify characteristics, risk factors, and evaluate outcomes of trauma patients who self-extubate. Methods: Self-extubated patients admitted to a Level 1 trauma center were matched to individuals who did not self-extubate in a 1:1 ratio based on injury severity score and admission date. Results: Self-extubation occurred in 81 of the 162 included patients. Self-extubated patients had a lower mortality rate (2.5% vs. 13.6%, p=0.02). Male gender and polysubstance exposure were risk factors for self-extubation. Polysubstance patients were 3-times more likely to self-extubate (OR=3.20, 95%CI= , p=0.003). Self-extubated patients requiring reintubation had longer ICU stays (13 vs. 6.5 days, p<0.001) and mechanical ventilation duration (11 vs. 4 days, p<0.001) compared to non- reintubated patients. Conclusion: Presence of polysubstances is associated with a 3-fold increase in self-extubation. Reintubation after self-extubation is associated with longer mechanical ventilation and ICU stays. Key Words: Self-Extubation, Trauma, Substance Abuse

4 Introduction Endotracheal intubation and mechanical ventilation (MV) are lifesaving measures for many critically ill patients including trauma patients. Current practice standards support weaning of MV as quickly and safely as possible in order to prevent complications such as ventilator associated pneumonia, atrophy of respiratory muscles, and ventilator dependence. Despite even the best attempts to wean MV up to 11% of patients will prematurely remove their endotracheal tubes themselves. 1-4 Serious complications can arise following unplanned extubation including tracheal edema, vocal cord damage, aspiration pneumonia, and most commonly, respiratory compromise resulting in the need for reintubation. Self and unplanned -extubation in medical ICU patients has been associated with a higher need for reintubation, longer duration of MV, ICU, and hospital stays. 5-8 In previous studies patients requiring reintubation following unplanned extubation have demonstrated higher mortality rates compared to patients not requiring reintubation. 7-8 While several studies have included trauma patients as part of their total patient population, there are no published studies evaluating self-extubation exclusively in the trauma patient population. 2,3,9,10 Studies conducted in critically ill adults have attempted to determine factors associated with self-extubation. 2, 6, Disease states, co-morbidities, vitals, laboratory values, process of care variables, patients condition immediately prior to the event, and substance use/abuse are some of the risk factors identified in the existing critical care literature Scientific literature regarding self-extubation in trauma patients is limited, thus the purpose of the current study was to identify commonalties in histories, comorbidities, and medical management of trauma patients who self-extubated. This study also examined risk factors and differences in patient outcomes in those who self-extubated compared to a matched no self-extubation cohort. 2

5 Methods This is a retrospective case-control study of MV trauma patients admitted to a Surgical Trauma Intensive Care Unit (STICU) at a Level I trauma center between January 1, 2007 and September 30, All patients were under the care of a trauma surgery intensivist. Patients were included if they were 18 years of age or older and required MV for 48 hours. Patients who selfextubated were identified through the trauma registry maintained by the hospital s Department of Trauma. Those admitted to the STICU during the same time period who were mechanically ventilated, but did not self-extubate were identified through the trauma registry and then matched to the self-extubation patients on a one to one ratio based on Injury Severity Score (ISS, ±5), and date of admission (± 3months). ISS is an anatomical score that evaluates trauma severity based on various body regions injuries. The score has been shown to correlate linearly with morbidity and mortality based on trauma victim s injuries. 14 Patients were excluded if they self-extubated following discharge from the STICU, had a tracheostomy prior to admission to the STICU, extubation was accidental (staff manipulation or moving the patient), or if documentation of the self-extubation could not be found in nursing or physician notes. This study was approved by the facility s Institutional Review Board. Electronic patient records were evaluated to collect patient demographics, admitting diagnoses and injuries, past medical history and comorbidities, and Glasgow Coma Scale (GCS) score upon arrival (Figure 1). Serum alcohol and urine drug screens were used to identify patients who were potential substance users and/or abusers. A patient was considered positive for substance use/abuse, if an agent was present in the admission urine drug screen and had not been administered by emergency service personal (EMS) prior to presentation in emergency room. 3

6 Patients were also considered to be substance users/abusers if there was documentation of prescribed or recreational use of controlled substances in their medical records. Additional information was collected for the 12 hours preceding a self-extubation. Patients who self-extubated more than once had peri-extubation data collected for the 1 st selfextubation event only. This data included: Richmond Agitation and Sedation Scale (RASS) scores, sedative, and analgesic pharmacotherapy in the 12 hours prior to the self-extubation event (Figure 2). Pharmacotherapy data consisted of changes to/from a continuous infusion, frequency of scheduled medications, or dose of scheduled medications. The need for re-intubation was recorded. Patient outcomes of ICU length of stay (LOS), MV duration and in-hospital mortality were evaluated in both patient cohorts. Statistics The statistical program IBM-SPSS 19.0 was used to analyze data. Basic descriptive statistics for continuous variables (means ± SD or median [inter-quartile range]), proportions and frequencies for categorical variables, were used to analyze patient characteristics. Comparisons between patients who self-extubated and those who did not was performed using contingency table analysis with a Chi-square test or Fisher s exact test, as appropriate, to determine statistically significant differences for categorical variables and t-tests or Mann-Whitney U test for continuous variables. Significance was reported with a p value less than Variables with p values of <0.05 in univariate analysis were entered into a logistic regression model to determine independent predictors for self-extubation. Odds ratios (OR) and 95% confidence intervals (CI) were calculated. 4

7 Results Ninety patients were identified by the Trauma Registry as having self-extubated during the study period. After review of medical records, five patients were found to have been extubated per the treating practitioner s orders and were excluded. Additionally, four patients had inadequate documentation in the nursing and physician notes to determine that extubation was a true self-extubation event. The remaining 81 patients were included in the self-extubation arm and compared to 81 patients matched in the control arm resulting in a total study population of 162. The median ISS score for all study patients was 24 [4-45]. No difference in baseline characteristics and comorbidities was found between the two groups, except for a greater percentage of males in the self-extubation cohort (80% vs. 63%, p=0.04). Traumatic injuries did not differ between the two groups (Table 1). There was also no gender specific differences in the incidence of substance use/abuse (males 73.5% vs. 66.7% females, p = 0.503). Significantly higher rates of alcohol (53.1% vs. 32.5%, p = 0.01), nicotine (64.2% vs. 45.7%, p = 0.03), and tetrahydrocannabinol (THC) (30.9% vs. 12.3%, p=0.01) use/abuse were seen, however, in patients who self-extubated (Table 2). There was also a 10% higher rate of opiate use/abuse in the self-extubation group, however, this did not reach statistical significance (p = 0.15) There were no additional significant differences between the study cohorts when comparing the incidence of use/abuse with other substances, including benzodiazepines. The incidence of polysubstance ( 2 substance) use/abuse was significantly higher in the self-extubation cohort when compared to the cohort that did not self-extubate (66.7% vs. 45.7%, p=0.008). In patients who had a positive history for nicotine use, in-hospital nicotine replacement therapy (NRT) was not significantly different with 37.8% of patients who self-extubated receiving a nicotine patch compared to 28.8% in non-self-extubated patients (p=0.51). 5

8 Logistic regression was conducted with factors significantly associated with selfextubation in univariate analyses. The following independent variables were entered into a logistic regression model: male gender, THC, alcohol, and nicotine use. Only male gender and THC use were independent predictors of self-extubation (THC: OR=2.550, 95% CI= , p=0.03; male gender: OR=2.240, 95% CI , p=0.03, Table 4). Of note, all THC positive patients in the analysis were also polysubstance positive. Thus, polysubstance use was entered into a separate regression model with the confounder male gender. Patients with polysubstance use/abuse had a 3-fold (OR=3.201, 95% CI= , p=0.003) increase in the likelihood of self-extubation. When evaluating outcomes, patients who self-extubated had a lower incidence of hospital mortality as compared to patients who did not (2.5 % vs. 13.6%, p =0.02). There was no statistically significant difference in duration of MV, ICU LOS, or other evaluated outcomes (Table 5). Based on a review of the medication administration records of all patients who selfextubated, there were 16 medication changes made in 11 (13.6%) patients in the 12 hours preceding the self-extubation event. Twelve of these changes involved adjustments to the frequency of a scheduled analgesic or the conversion of a continuously infused analgesic to an intermittently administered regimen. Out of the 81 self-extubating patients, 29 (35.8%) had RASS scores (+1 to +4) indicative of restlessness or agitation immediately prior to the extubation event. There were 47 (58%) patients that self-extubated and required reintubation. Over onehalf (55.3%) of those who required re-intubation were mechanically ventilated within an hour of self-extubation. This percentage increased to 87.2% within 24 hours. There was no difference seen in GCS, ISS, or RASS scores between the patients who self-extubated and required reintubation and those who did not. Comparison of self-extubated patients who required re- 6

9 intubation versus self-extubated patients that did not require reintubation, revealed a longer duration of ICU LOS (13 [9-17] vs. 6.5 [4-11] days, p <0.001) and MV (11 [6-16] vs. 4 [3-6] days, p<0.001) for the re-intubation group. Examination of the baseline characteristics between these two groups revealed no significant difference in age, GCS, ISS, or RASS scores observed in the peri-extubation period (Table 6). Mortality between the two groups did not differ (4.3% vs. 0%, p=0.51). Discussion In this study, characteristics and outcomes of trauma patients who self-extubated during ICU hospitalization were compared to matched controls. Mortality rates for patients who selfextubated were significantly lower compared to patients who did not self-extubate. There were no differences in baseline characteristics including ISS, type of injury, or comorbidities, therefore, the lower mortality rate seen in the self-extubation group may be attributed to an unstudied variable. In addition the difference in mortality may also be explained by any, or a combination of the following factors. The first explanation could be the fact that regimented, protocol-based assessments to evaluate readiness for extubation were not routinely performed during the study period. Accordingly, a portion of the self-extubation patients may represent those who no longer required MV. This conclusion appears to be supported by the higher mortality rate seen in the group of patients who required re-intubation after an initial selfextubation. While this difference did not reach statistical significance, it suggests a dichotomy among the self-extubation group given that the overall trend in mortality for this group was lower than that of the no self-extubation group. A portion of the self-extubation patients may also have been more alert, in better clinical condition, and therefore physically stronger, affording them to the ability to extubate themselves. Additionally, patients who self-extubated 7

10 and did not require reintubation had a shorter duration of MV and ICU stay placing them at a lower risk for nosocomial complications and their associated morbidity and mortality. Similar comparisons that included, but were not limited to trauma patients, have had mortality rates that were similar between patients who self-extubate and those with planned extubations. 2,5 These findings are in contrast, however, to the study conducted by Atkins and colleagues where a non-significant decrease in the mortality rate was observed in patients who did not self-extubate. Accordingly, given the variability in outcomes among the available body of literature, it is difficult to conclusively determine the impact of self-extubation on hospital mortality. While conclusions cannot be drawn regarding the causality between self-extubation and mortality from the comparisons in this study, interesting outcomes were observed in some of the evaluated morbidity comparisons. The rate of reintubation (58%) is similar to that reported in a recent review of 50 articles examining unplanned extubation. This review found the median rate for reintubation following self-extubation to be 45.8% with an interquartile range of %. 15 The current study found that 58.1% of patients who self-extubated required reintubation with over ½ of those patients requiring MV within 1 hour of self-extubation. These findings further support the paradigm that methods used during the study period for determining when patients were candidates for extubation likely need to be revisited or more consistently practiced. Furthermore, nearly half of the patients who self-extubated did not require reintubation suggesting that the frequency of evaluation for possible extubation should be performed more frequently to ensure that liberation from MV occurs in a timely fashion. Additional findings of the current study assessed the association between pre-existing substance use/abuse and self-extubation. There are a limited number of additional studies reviewing this type of association, however, one study of medical and surgical ICU patients by 8

11 Atkins and colleages found that patients who self-extubated were more likely to have a history of nicotine use. 6 The current study also found this to be prevalent in the trauma patient population. In addition, our study found that the use of NRT was not different between the patients who selfextubated and those that did not, suggesting that NRT may not be an effective preventative measure for self-extubation. There are potential limitations associated with this finding however, as the study was not specifically designed to test the utility of NRT as an intervention to prevent self-extubation. This study also did not perform an analysis comparing the strength of nicotine patch utilized versus the reported amount of nicotine consumption. In addition, there were no assessments evaluating NRT in relation to the timing of patients self-extubation. Despite limitations, these findings are supportive of previous data evaluating the efficacy of nicotinebased NRT in patients with chronic nicotine exposure. In a study by Teneggi and colleagues the investigators found that replacement therapy with nicotine patches was effective in curbing craving symptoms, but ineffective in preventing other common withdrawal symptoms such as irritability, anxiety, restlessness, and hostility. 16 Accordingly in the current study NRT with patches appeared to have no effect in the prevention of self-extubation in patients with a history of nicotine use. Unlike the current study, previous studies have failed to find an association with alcohol use in patients who self-extubate. 2,6, 12 It is possible that the patients in the self-extubation group of the current study did not receive adequate prophylaxis for alcohol withdrawal and were therefore predisposed to experience complications such as self-extubation. There is limited data assessing the incidence and implications of alcohol withdrawal in critically ill trauma patients who are admitted with positive serum alcohol concentrations. A recent study by Sharp and colleagues evaluated the incidence of withdrawal symptoms in non-icu trauma patients who were admitted with positive serum alcohol levels. This study found that only 3.6% of these 9

12 patients experienced alcohol withdrawal symptoms. 17 These findings suggest that a low proportion of trauma patients require treatment for alcohol withdrawal. However the study was conducted in non-critically ill patients and caution should be made in applying the results of Sharp and colleagues study to trauma patients in the intensive care setting. The current study did not evaluate the implementation of preventative measures for alcohol withdrawal. However, the findings of the current study highlight the importance of insuring that measures to prevent misadventures in care are consistently implemented in critically ill trauma patients with a history of alcohol use. Several studies have identified lack of sedation therapy prior to self-extubation as a risk factor for self-extubation. 2,11 The current study found that 13.6% of patients had a change in their sedation and analgesia therapy prior to self-extubating. Due to the design of this study a side by side comparison of changes in sedation and analgesia therapy and its effect on consciousness levels between the two study cohorts could not be performed. However, it appears that changes in pharmacotherapy did not frequently occur prior to incidents of selfextubation. Male gender and THC were found to be predictors of self-extubation in the current study population. The finding of male gender as a predictor of self-extubation in the critically ill is consistent with previously published findings. 8,15,18 THC however, the strongest predictor of selfextubation in the current study, appears to be a novel finding in the critical care literature as previous studies have not reported a similar association. The finding in the current study is likely influenced by the significant association seen with concomitant substance use in this patient population as 100% of patients who were positive for THC were also positive for other substances. The predictive qualities of THC on self-extubation may be largely reflective of the cumulative influence of polysubstance exposure versus the influence of THC alone. 10

13 Accordingly, patients positive for multiple substances in our second regression model were more than 3-times more likely to self-extubate. While trauma patients have been included in several previous studies, this is the first study to be conducted exclusively in the trauma patient population 2,3,10 In addition the current study is one of the largest to date, providing a large sample size to detect differences in several important patient characteristics. Limitations consistent with that of retrospective research are pertinent to the current study. Inherent variation and availability of documentation resulted in the exclusion of four potential study patients. Serum alcohol and urine drug screens were used as a screening tool for substance use and abuse. Although the investigators attempted to identify urine drug screens that were positive due to iatrogenic exposure to opioids and benzodiazepines by EMS, it is possible that patients may have been erroneously documented as substance users. Conclusion Critically ill trauma patients who self-extubated were found to have significantly lower mortality rates and duration of hospitalization compared to those who did not self-extubate. These findings in conjuction with a reintubation rate of 58% suggest that many patients were likely intubated despite being clinically ready for liberation from MV. Accordingly, these findings suggest that methods for determining patients readiness for extubation need to be consistently practiced in a regimented, prospective fashion. Patients positive for 2 or more substances were 3.2 times more likely to self-extubate. Self-extubated patients who required reintubation had a significantly longer duration of MV and ICU LOS. These findings highlight a subset of trauma patients that would likely benefit from specialized patient care strategies geared towards the management of the various clinical issues 11

14 commonly associated with pre-existing substance use such as higher sedative and analgesic requirements, and the propensity to develop withdrawal syndromes. Accordingly, future efforts should focus on early identification of these patients and the implementation of individualized clinical management strategies geared towards prevention of self-extubation and timely extubation from MV when clinically indicated. Acknowledgement The authors would like to thank Mrs. Kelly-Jo Evans, RN, BSN and Lisa Shamblin for their assistance in generating the patient list of trauma victims who self-extubated. 12

15 References 1. Coppolo D, May J. Self-extubations. A 12-month experience. Chest. 1990;98: Boulain T. Unplanned extubation in the adult intensive care unit. Am J Respir Crit Care. 1998;157: Betbese A, Perez M, Bak E, et al. A prospective study of unplanned endotracheal extubations in intensive care unit patients. Crit Care Med. 1998;26: Tindol G, DiBenedetto R, Kosciuk L. Unplanned extubations. Chest. 1994;105: Epstein S, Nevins M, Chung J. Effect of unplanned extubations on outcomes on mechanical ventilation. Am J Respir Crit Care Med 2000;161: Atkins P, Mion L, Mendelson W. Characteristics and outcomes of patients who self-extubate from ventilatory support: a case-control study. Chest. 1997;112: Krinsley JS, Barone JE. The drive to survive: unplanned extubation in the ICU. Chest Aug;128(2): de Groot RI, Dekkers OM, Herold IH, de Jonge E, Arbous MS. Risk factors and outcomes after unplanned extubations on the ICU: a case-control study. Crit Care. 2011;15(1):R Listello D, Sessler C. Unplanned extubation. Clinical predictors for reintubation. Chest. 1994;105: Balon J. Common factors of spontaneous self-extubation in a critical care setting. Int J Trauma Nurse 2001;7: Moons P, Sels K, De Becker et al. Development of a risk assessment tool for deliberate self-extubation in intensive care patients. Intensive Care Med 2004;30: Chang L, Liu P, Huang Y et al. Risk factors associated with unplanned endotracheal self-extubation of hospitalized intubated patients: a 3-year retrospective case-control study. Appl Nurs Res 2011;24: Whelan J, Simpson S, Levy H. Unplanned extubation. Predictors of successful termination of mechanical ventilator support. Chest. 1994;105: Baker SP, O Neill B, Haddon W Jr, Long WB. The Injury Severity Score: a method for describing patients with multiple injuries and evaluating emergency care. J Trauma 1974 Mar; 14(3): da Silva PS, Fonseca MC. Unplanned endotracheal extubations in the intensive care unit: systematic review, critical appraisal, and evidence-based recommendations. Anesth Analg May;114(5): Teneggi V, Tiffany ST, Squassante L, Milleri s, Ziviani L, Bye A. Smokers deprived of cigarettes for 72h: effect of nicotine patches on craving and withdrawal. Psychopharmacology (Berl) 2002; 164(2): Sharp B, Schermer C, Esposito T et al. Alcohol Withdrawal Syndrome in Trauma Patients: A Prospective Cohort Study. J Trauma Treat 2012; Kiekkas P, Aretha D, Panteli E, Baltopoulos G, Filos KS. Unplanned extubation in critically ill adults: clinical review. Nursing and Crit Care 2013; 18(3): Teasdale G, Knill-Jones R, van der Sande J. Observer variability in assessing impaired consciousness and coma. J Neurol Neurosurg Psychiatry 1978;41: Sessler CN 1, Gosnell MS, Grap MJ, Brophy GM, O'Neal PV, Keane KA, Tesoro EP, Elswick RK. The Richmond Agitation-Sedation Scale: validity and reliability in adult intensive care unit patients. Am J Respir Crit Care Med Nov 15;166(10):

16 Table 1: Baseline Characteristics No Self-Extubation Self-Extubation P value N=81 N=81 Age, Mean ± SD 48.5± ± ISS, median [IQR] 24 [17-29] 24 [18-32] 0.71 Gender: Male 52 (64.2%) 65 (80.2%) 0.04 Race: Caucasian 79 (97.5%) 75 (92.6%) 0.28 Non-Caucasian 2 (2.4%) 6 (7.4%) Comorbidities CRI 1 (1.2%) 1 (1.2%) 1.00 HD/CRRT 2 (2.5%) 4 (4.9%) 0.68 COPD 9 (11.1%) 9 (11.1%) 1.00 ARDS 2 (2.5%) 4 (4.9%) 0.68 Pneumonia 24 (29.6%) 27 (33.3%) 0.74 CHF 3 (3.7%) 1 (1.2%) 0.62 DM 14 (17.3%) 6 (7.4%) 0.95 Anxiety 8 (9.9%) 2 (2.5%) 0.10 Mood disorder 14 (17.3%) 16 (19.8%) 0.84 Dementia 1 (1.2%) 4 (4.9%) 0.37 Initial GCS, median [IQR] 3[3-15] 5 [3-15] 0.46 Type of Traumatic Injury No Self-Extubation Self-Extubation P value N=81 N=81 Lower extremity fracture 20 (24.7%) 20 (24.7%) 1.00 Pelvic fracture 14 (17.3%) 19 (23.5%) 0.44 Upper extremity fracture 10 (12.3%) 6 (7.4%) 0.43 Spinal fracture/cord injury 29 (35.8%) 30 (37.0%) 1.00 Rib fracture 45 (55.6%) 42 (51.9%) 0.75 Traumatic brain injury 41 (50.6%) 46 (56.8%) 0.53 Facial fracture 34 (42.0%) 32 (39.5%) 0.87 Liver injury 4 (4.9%) 11 (13.6%) 0.10 Renal injury 4 (4.9%) 6 (7.4%) 0.74 Other Abdominal injury 24 (29.6%) 37 (45.8%) 0.52 Vascular injury 8 (9.9%) 6 (7.4%) 0.78 Abbreviations, ISS = Injury Severity Score, CRI = Catheter-related infection, HD/CRRT = Hemodialysis/continuous renal replacement therapy, COPD = Chronic obstructive pulmonary disease, ARDS = acute respiratory distress syndrome, CHF = congestive heart failure, DM = diabetes mellitus, GCS = Glasgow Coma Scale (Figure 3), IQR = inter-quartile range 14

17 Table 2: Substance use/abuse No Self-Extubation Self-Extubation P value N=81 N=81 Alcohol 26 (32.5%) 43 (53.1%) 0.01 Alprazolam 9 (11.1%) 8 (9.9%) 1.00 Amphetamines 1 (1.2%) 6 (7.4%) 0.12 Benzodiazepine 15 (18.5%) 15 (18.5%) 1.00 Clonazepam 2 (8.6%) 4 (4.9%) 0.68 Cocaine 7 (8.6%) 6 (7.4%) 1.00 THC 10 (12.3%) 25 (30.9%) 0.01 Nicotine 37 (45.7%) 52 (64.2%) 0.03 Opiates 10 (12.3%) 18 (22.2%) 0.15 Nicotine patch 14 (17.3%) 15 (18.5%) 1.00 Abbreviations, THC = tetrahydrocannabinol 15

18 Table 3. Independent risk factors for self-extubation using multivariate analyses Model 1 Odds Ratio 95% CI P value Male gender Alcohol use THC use Nicotine use Constant Model 2 Odds Ratio 95% CI P value Male Gender Polysubstance use Use of 1 substance vs no substance use -Use of 2 substances vs. no substance use Constant CI = confidence interval. Model 1 examined substances each individually while Model 2 examined substances as the composite, polysubstance use. 16

19 Table 4: Outcomes Outcomes No Self-Extubation Self-Extubation P value N=81 N=81 In hospital mortality 11 (13.6%) 2 (2.5%) 0.02 ICU LOS (days), median [IQR] 8 [4-14] 10 [6-15] 0.06 MV (days), median [IQR] 5 [3-12] 8 [4-14] 0.18 Need for tracheostomy 25 (30.9%) 19 (23.5%) 0.38 Need for MV at discharge 5 (6.2%) 4 (4.9%) 1.00 Discharge to extended care facility 39 (48.1%) 39 (48.1%) 1.00 Abbreviations, ICU = intensive care unit, LOS = length of stay, MV = mechanical ventilation 17

20 Table 5. Baseline Characteristics of patients who self-extubated No Re-intubation Re-intubation P value N = 34 N = 47 Age, Mean ± SD 44.0± ± Gender: Male 26 (76.5%) 39 (83.0%) 0.66 ISS, median [IQR] 22 [18-32] 24 [18-30] 0.91 Initial GCS, median 7.5[3-15] 3 [3-15] 0.32 [IQR] RASS (+1 to +4) 12 (35.3%) 17 (36.2%)

21 Figure 1 Glasgow Coma Scale 19 Best Eye Response. (4) 1. No eye opening. 2. Eye opening to pain. 3. Eye opening to verbal command. 4. Eyes open spontaneously. Best Verbal Response. (5) 1. No verbal response 2. Incomprehensible sounds. 3. Inappropriate words. 4. Confused 5. Orientated Best Motor Response. (6) 1. No motor response. 2. Extension to pain. 3. Flexion to pain. 4. Withdrawal from pain. 5. Localizing pain. 6. Obeys Commands. The Glasgow Coma Scale is a tool used to assess patients neurologic function. Three parameters are evaluated to calculate GCS: visual, verbal, and motor responsiveness. Score ranges from 3 to 15, with 3 being the worst and 15 being the best. 19

22 Figure 2 RICHMOND AGITATION SEDATION SCALE (RASS) 20 Score Term Description +4 Combative Overtly combative or violent; immediate danger to staff +3 Very agitation Pulls on or removes tube(s) or catheter(s) or has aggressive behavior toward staff +2 Agitated Frequent nonpurposeful movement or patient ventilator dyssynchrony +1 Restless Anxious or apprehensive but movements not aggressive or vigorous 0 Alert and calm -1 Drowsy Not fully alert, but has sustained (more than 10 seconds) awakening, with eye contact, to voice -2 Light sedation Briefly (less than 10 seconds) awakens with eye contact to voice -3 Moderate sedation Any movement (but no eye contact) to voice -4 Deep sedation No response to voice, but any movement to physical stimulation -5 Unarousable No response to voice or physical stimulation The RASS scale has shown high reliability and validity as a reliable tool for evaluating patients level of sedation in ventilated and non-ventilated ICU patients. Sessler, Gosnell, Grap, et al.: ICU Sedation Scale Validation. Respiratory and Critical Care Medicine