Multicenter Study of the Treatment of Appendicitis in America: Acute, Perforated, and Gangrenous (MUSTANG), an EAST Multicenter Study.

OBJECTIVE: We sought to describe contemporary presentation, treatment, and outcomes of patients presenting with acute (A), perforated (P), and gangrenous (G) appendicitis in the United States. SUMMARY BACKGROUND DATA: Recent European trials have reported that medical (antibiotics only) treatment of acute appendicitis is an acceptable alternative to surgical appendectomy. However, the type of operation (open appendectomy) and average duration of stay are not consistent with current American practice and therefore their conclusions do not apply to modern American surgeons. METHODS: This multicenter prospective observational study enrolled adults with appendicitis from January 2017 to June 2018. Descriptive statistics were performed. P and G were combined into a "complicated" outcome variable and risk factors were assessed using multivariable logistic regression. RESULTS: A total 3597 subjects were enrolled across 28 sites: median age was 37 (27-52) years, 1918 (53%) were male, 90% underwent computed tomography (CT) imaging, 91% were initially treated by appendectomy (98% laparoscopic), and median hospital stay was 1 (1-2) day. The 30-day rates of Emergency Department (ED) visit and readmission were 10% and 6%. Of 219 initially treated with antibiotics, 35 (16%) required appendectomy during index hospitalization and 12 (5%) underwent appendectomy within 30 days, for a cumulative failure rate of 21%. Overall, 2403 (77%) patients had A, whereas 487 (16%) and 218 (7%) patients had P and G, respectively. On regression analysis, age, symptoms >48 hours, temperature, WBC, Alvarado score, and appendicolith were predictive of "complicated" appendicitis, whereas co-morbidities, smoking, and ED triage to appendectomy >6 hours or >12 hours were not. CONCLUSION: In the United States, the majority of patients presenting with appendicitis receive CT imaging, undergo laparoscopic appendectomy, and stay in the hospital for 1 day. One in five patients selected for initial non-operative management required appendectomy within 30 days. In-hospital delay to appendectomy is not a risk factor for "complicated" appendicitis.

Anticoagulation therapy in patients with traumatic brain injury: An Eastern Association for the Surgery of Trauma multicenter prospective study.

BACKGROUND: Trauma care providers often face a dilemma regarding anticoagulation therapy initiation in patients with traumatic brain injury owing to the associated risks of traumatic brain injury progression. The aims of this study were the following: (1) to describe the current practice of anticoagulation therapy in traumatic brain injury patients and their outcomes and (2) to identify factors associated with the progression of traumatic brain injury after anticoagulation therapy. METHODS: In this multicenter prospective observational study, we included computed tomography-proven traumatic brain injury patients who received anticoagulation therapy within 30 days of hospital admission. Our primary outcome was the incidence of clinically significant progression of traumatic brain injury after anticoagulation therapy initiation. RESULTS: A total of 168 patients were enrolled more than 22 months. Atrial fibrillation and venous thromboembolism were the most common pre-injury and postinjury anticoagulation therapy indications, respectively. Overall, 16 patients (9.6%) experienced clinically significant traumatic brain injury progression after anticoagulation therapy, out of which 9 (5.4%) patients subsequently required neurosurgical interventions. Between patients with clinical progression of traumatic brain injury and patients who showed no such progression, there were no significant differences in the baseline demographics and severity of traumatic brain injury. However, anticoagulation therapy was initiated significantly earlier in patients of the deterioration group than those of the no-deterioration group (4.5 days vs 11 days, P = .015). In a multiple logistic regression model, patients who received anticoagulation therapy later after injury had significantly lower risk of clinically significant traumatic brain injury progression (odds ratio: 0.915 for each day, 95% confidence interval: 0.841-0.995, P = .037). CONCLUSION: Our results suggest that early anticoagulation therapy is associated with higher risk of traumatic brain injury progression, thus a balance between bleeding and thromboembolic risks should be carefully evaluated in each case before initiating anticoagulation therapy.

Predictors of retained hemothorax in trauma: Results of an Eastern Association for the Surgery of Trauma multi-institutional trial.

BACKGROUND: The natural history of traumatic hemothorax (HTX) remains unclear. We aimed to describe outcomes of HTX following tube thoracostomy drainage and to delineate factors that predict progression to a retained hemothorax (RH). We hypothesized that initial large-volume HTX predicts the development of an RH. METHODS: We conducted a prospective, observational, multi-institutional study of adult trauma patients diagnosed with an HTX identified on computed tomography (CT) scan with volumes calculated at time of diagnosis. All patients were managed with tube thoracostomy drainage within 24 hours of presentation. Retained hemothorax was defined as blood-density fluid identified on follow-up CT scan or need for additional intervention after initial tube thoracostomy placement for HTX. RESULTS: A total of 369 patients who presented with an HTX initially managed with tube thoracostomy drainage were enrolled from 17 trauma centers. Retained hemothorax was identified in 106 patients (28.7%). Patients with RH had a larger median (interquartile range) HTX volume on initial CT compared with no RH (191 [48-431] mL vs. 88 [35-245] mL, p = 0.013) and were more likely to be older with a higher burden of thoracic injury. After controlling for significant differences between groups, RH was independently associated with a larger HTX on presentation, with a 15% increase in risk of RH for each additional 100 mL of HTX on initial CT imaging (odds ratio, 1.15; 95% confidence interval, 1.08-1.21; p < 0.001). Patients with an RH also had higher rates of pneumonia and longer hospital length of stay than those with successful initial management. Retained hemothorax was also associated with worse functional outcomes at discharge and first outpatient follow-up. CONCLUSION: Larger initial HTX volumes are independently associated with RH, and unsuccessful initial management with tube thoracostomy is associated with worse patient outcomes. Future studies should use this experience to assess a range of options for reducing the risk of unsuccessful initial management. LEVEL OF EVIDENCE: Therapeutic/care management study, level III.

MAGIC versus MacDonald treatment regimens for gastric cancer: Trends and predictors of multimodal therapy for gastric cancer using the National Cancer Database.

BACKGROUND: Multimodal therapy is beneficial in gastric cancer, however this practice is not universal. This study examines trends, identifies associative factors, and examines overall survival (OS) benefit from multimodal therapy in gastric cancer. METHODS: Gastric cancer patients staged IB-III from 2005 to 2014, identified using the National Cancer Database, were categorized by treatment: surgery alone, perioperative chemotherapy, and adjuvant chemoradiation. Groups were analyzed to identify associative factors of perioperative therapy. RESULTS: We examined 9243 patients, with the majority receiving multimodal therapy (57%). The proportion of those receiving perioperative chemotherapy rose dramatically from 7.5% in 2006 to 46% in 2013. Academic center treatment was strongly associated with perioperative over adjuvant therapy (p < 0.0001). An OS advantage was clearly seen in those receiving multimodal therapy versus surgery alone (p < 0.0001), with no difference between perioperative and adjuvant therapies. CONCLUSIONS: Treatment of gastric cancer with multimodal therapy has risen significantly since 2005, largely due to increasing use of perioperative chemotherapy. As perioperative therapy becomes more prevalent, more patients will have the opportunity for the improved survival benefit of multimodal therapy.

Outcomes of Endoscopic and Surgical Pancreatic Necrosectomy: A Single Institution Experience.

Pancreatic necrosis can be managed conservatively; however, infection of pancreatic necrosis usually dictates more aggressive management. Our study aimed to assess the outcomes of open pancreatic necrosectomy (OPN) and endoscopic pancreatic necrosectomy (EPN) in a single center. Data from patients undergoing pancreatic necrosectomy at the Geisinger Medical Center from January 1, 2007, to April 25, 2016, were collected and retrospectively analyzed. Cohorts were composed of EPN (n = 22) and OPN (n = 34) groups. The prevalence of preoperative respiratory failure, septic shock, and multiorgan dysfunction syndrome was higher in the OPN group. The OPN group presented with a higher Bedside Index Severity in Acute Pancreatitis score. Postoperative abscess, persistent kidney dysfunction, and death were more frequent in the OPN group. The EPN group had a higher readmission rate. The results of the univariate analysis for complication and mortality demonstrated that higher mortality and persistent kidney dysfunction were associated with the procedure type, specifically OPN and with a higher Bedside Index Severity in Acute Pancreatitis score. Patients who presented with higher severity of disease underwent an OPN, whereas EPN often was performed successfully in a more benign clinical setting. However, patients with infected necrosis are served best in a tertiary medical facility where multiple treatment modalities are available.

Factors Associated with Failure of Nonoperative Management for Complicated Appendicitis.

In recent years, nonoperative management of complicated appendicitis has become more common. Patients managed nonoperatively do well, but there is a paucity of literature on patients who fail nonoperative management. The purpose of this study was to examine the overall failure rate, morbidity associated with failure, and potential predictors of failure in nonop management of appendicitis. This is a descriptive retrospective review of patients from a single hospital system who were diagnosed with advanced appendicitis and underwent nonop management between January 1, 2007, and November of 2017. The data were obtained through review of patient charts from the electronic medical record. Failure was defined as requirement of an operation due to ongoing infection secondary to appendicitis. There were 183 patients initially managed nonoperatively, with 70 patients failing nonoperative management. Patients failing nonoperative management experienced longer hospitalization (6.2 vs 2.9 days, P < 0.0001), and more patients in the failure group required admission to the ICU (10.0% vs 1.8%, P = 0.028). Multivariate analysis revealed that longer duration of symptoms reduced the likelihood of failure (odds ratio: 0.77 [0.64-0.92]). In this retrospective review, 38 per cent of patients failed nonop management of appendicitis. Symptom duration could provide insight for clinicians in assessing the role of nonoperative management because increasing symptom duration reduced the likelihood of failure.

Lymph Node Harvest as a Measure of Quality and Effect on Overall Survival in Esophageal Cancer: A National Cancer Database Assessment.

Surgical therapy for esophageal cancer is the cornerstone of treatment, and the highest quality operation should lead to the highest cure rate. Evaluated lymph node (ELN) count is one quality measure that has been championed. The objective of this study was to explore ELN in esophagectomy, examine predictors of harvesting ≥12 nodes, and determine whether higher ELN improves overall survival (OS). ELN was examined in patients with resected esophageal cancer using the National Cancer Database from 2004 to 2013. In this study, 41,746 patients met the inclusion criteria. Fifty-two per cent of patients had 12 or more nodes harvested. Academic programs were most likely to harvest ≥12 nodes (58% of cases) compared with other programs (43-56% of cases). Seventy per cent of cases with ≥12 nodes harvested were performed at high-volume centers. Preoperative radiation or preoperative chemoradiation led to lower ELN (46% and 48%) versus preoperative chemotherapy alone (66%). Multivariate analysis showed that patients who had ≥12 nodes removed had better OS (Hazard Ratio 0.843 [95 confidence interval 0.820-0.867]). In addition, care at a high-volume facility, care at an academic facility, private insurance, and income ≥$63,000 were all associated with improved OS. Higher ELN count is associated with OS in patients with esophageal cancer. Patients who receive care at high-volume centers and academic centers are more likely to undergo more extensive lymphadenectomy. All centers should strive to examine at least 12 nodes to provide a quality esophagectomy.

Laparoscopic versus robotic adrenalectomy: a review of the national inpatient sample.

BACKGROUND: Laparoscopic adrenalectomy (LA) has become the standard treatment of adrenal lesions. Recently, robotic-assisted adrenalectomy (RA) has become an option, however, short-term outcomes for RA have not been well studied and benefits over LA are debatable. The aim of this study was to explore differences in short-term outcomes between LA and RA using the national inpatient sample (NIS) database. METHODS: Patient data were collected from the NIS. All patients undergoing LA or RA from January 2009 to December 2012 were included. Univariate analysis and propensity matching were performed to look for differences between the groups. RESULTS: A total of 1006 patients (66.4% in LA group and 33.6% in RA group) were identified. Patient age group, gender, race, risk of mortality, severity of illness or indication for adrenalectomy did not differ significantly between the LA or RA cohorts. Insurance type predicted procedure type (45% of medicare patients underwent RA versus 29% of patients with private insurance, p < 0.0001). Patients living in the highest income areas were more likely to receive the laparoscopic approach (31.7 versus 17.4%, p < 0.0001). Hospital volume, bed size and teaching status of the hospital were not significant factors in the decision of RA versus LA. There was no difference in complication and conversion rates between RA versus LA. The mean length of stay was shorter in the RA group (2.2 versus 1.9 days, p = 0.03). Total charges were higher in the RA group ($42,659 versus $33,748, p < 0.0001). There was a significant trend towards more adrenalectomies being performed robotic assisted by year. Only 22% of adrenalectomies were performed robotic-assisted in 2009 compared with 48% in 2012. CONCLUSIONS: The overall benefit for RA remains small and higher total charges for RA may currently outweigh the benefits. These findings may change as more cases are performed robotically assisted and robotic technology improves.

Cancel that PICC line order; cholecystostomy tube and short course of antibiotics.

BACKGROUND: Current guidelines do not specifically address optimal antibiotic duration following cholecystostomy. This study compares outcomes for short-course (<7 days) and long-course (≥7 days) antibiotics post-cholecystostomy. METHODS: This was a retrospective review of cholecystostomy patients (≥18 years) admitted (1/1/2007-12/31/2017) to one healthcare system. RESULTS: Overall, 214 patients were studied. Demographics were similar, except short-course patients had higher Charlson Comorbidity Index (p < 0.0001). There were no intergroup differences in tachycardia (22.5%[short-course] vs 23.3%[long-course]) or leukocytosis (67.1%[short-course] vs 64.4%[long-course]) at drain placement nor time to normalization for pulse, temperature or leukocytosis. There were no differences regarding Clostridium Difficile infection (5.0%[short-course] vs 1.6%[long-course]) or cholecystitis recurrence (8.8%[short-course] vs 10.9%[long-course]). No differences were observed regarding gallbladder-related unplanned readmissions (30-day:18.8%[short-course] vs 17.2%[long-course]; 90-day: 20.0%[short-course] vs 25.8%[long-course]). There were no 30- or 90-day mortality differences (overall mortality: 18.3%). CONCLUSION: Post-cholecystostomy outcomes were comparable between short-course and long-course antibiotics, consistent with emerging literature supporting short-course antibiotics for intra-abdominal infection with source control.

Sarcopenia predicts poor outcomes in urgent exploratory laparotomy.

BACKGROUND: Emergent laparotomies are associated with higher rates of morbidity and mortality. Recent studies suggest sarcopenia predicts worse outcomes in elective operations. The purpose of this study is to examine outcomes following urgent exploratory laparotomy in sarcopenic patients. METHODS: This was a retrospective review of patients in a rural tertiary care facility between 2010 and 2014. Patients underwent a laparotomy within 72 h of admission and had an abdomen/pelvis CT scan were included. Primary outcomes were predictors of morbidity and mortality. Sarcopenia is the lowest quartile cross sectional area of the psoas muscles. RESULTS: Multivariate analysis of 967 patients found that sarcopenic patients had higher mortality, complication rate, were less likely to be discharged home, were more likely to undergo unplanned re-operation, and had a longer length of stay. Increasing abdominal wall fat has favorable outcomes in mortality, discharge destination, and complications. CONCLUSIONS: Sarcopenia is measured from CT scans, making it an accessible outcome predictor. In urgent laparotomies, sarcopenia was associated with higher morbidity, mortality, length of stay, and worse discharge destination.

Neoadjuvant versus Adjuvant Chemotherapy for Resectable Pancreatic Adenocarcinoma: A National Cancer Database Analysis.

There is controversy regarding the role of neoadjuvant versus adjuvant chemotherapy for pancreatic cancer (PAC). Neoadjuvant therapy has been touted as a method to improve survival in PAC patients. This study's objective is to investigate predictors and potential benefits of neoadjuvant therapy in resectable PAC patients. The National Cancer Data Base was used to retrospectively analyze stage I and II surgically resected PAC patients receiving adjuvant or neoadjuvant therapy from 2004 to 2012. A total of 12,983 patients were identified. A significant increase in the rate of neoadjuvant therapy was observed over time with 5 per cent receiving neoadjuvant therapy in 2004 versus 17 per cent in 2012 (P < 0.0001). Multivariate analysis showed that patients were more likely to receive neoadjuvant therapy if they were treated at an academic facility. Private insurance was associated with higher odds of neoadjuvant chemotherapy (P < 0.0001). Pathological outcomes were improved in neoadjuvant patients compared with adjuvant patients on multivariate analysis with neoadjuvant patients having higher rates of negative surgical margins (OR: 1.273, 95% Confidence interval: 1.099-1.474) and negative lymph nodes (OR: 2.852, 95% Confidence interval: 2.547-3.194). Pathological outcomes are improved after neoadjuvant therapy compared with adjuvant therapy, with more patients achieving negative margins and negative lymph nodes. Prospective studies are needed to compare these two treatment modalities in a head to head comparison.

Hospital-Based Study of Compliance with NCCN Guidelines and Predictive Factors of Sentinel Lymph Node Biopsy in the Setting of Thin Melanoma Using the National Cancer Database.

Thin melanoma is the most common form of melanoma in the United States. The National Comprehensive Cancer Network (NCCN) has guidelines for sentinel lymph node biopsy (SLNB) which recommend "discuss and consider" SLNB for invasion >0.75 mm and "discuss and offer" SLNB for invasion >0.75 mm with suspicious features. This study looked at compliance with NCCN guidelines and factors that are predictive of a positive SLNB. This is a retrospective study of patients diagnosed with thin melanoma 2012-2013 using the National Cancer Database. A total of 26,456 patients met study qualifications. Univariate analysis showed that 76 per cent of patients meeting criteria underwent SLNB. Patients recommended to "discuss and consider" received SLNB 53 per cent of the time and those not recommended for SLNB received SLNB 20 per cent of the time. On multivariate analysis, depth was not predictive for positive SLNB whereas mitoses and ulceration were. Other factors predictive of positive SLNB were nodular cell type, lymphovascular invasion, and Clark's level greater than or equal to IV. Patients with thin melanoma that meet NCCN guidelines for SLNB undergo this procedure in good compliance but those who do not meet criteria continue to receive SLNB. Positive predictive factors for positive SLNB include mitoses, ulceration, Clark's level, and primary site.

EAST Multicenter Trial on targeted temperature management for hanging-induced cardiac arrest.

BACKGROUND: We sought to determine the outcome of suicidal hanging and the impact of targeted temperature management (TTM) on hanging-induced cardiac arrest (CA) through an Eastern Association for the Surgery of Trauma (EAST) multicenter retrospective study. METHODS: We analyzed hanging patient data and TTM variables from January 1992 to December 2015. Cerebral performance category score of 1 or 2 was considered good neurologic outcome, while cerebral performance category score of 3 or 4 was considered poor outcome. Classification and Regression Trees recursive partitioning was used to develop multivariate predictive models for survival and neurologic outcome. RESULTS: A total of 692 hanging patients from 17 centers were analyzed for this study. Their overall survival rate was 77%, and the CA survival rate was 28.6%. The CA patients had significantly higher severity of illness and worse outcome than the non-CA patients. Of the 175 CA patients who survived to hospital admission, 81 patients (46.3%) received post-CA TTM. The unadjusted survival of TTM CA patients (24.7% vs 39.4%, p < 0.05) and good neurologic outcome (19.8% vs 37.2%, p < 0.05) were worse than non-TTM CA patients. However, when subgroup analyses were performed between those with an admission Glasgow Coma Scale score of 3 to 8, the differences between TTM and non-TTM CA survival (23.8% vs 30.0%, p = 0.37) and good neurologic outcome (18.8% vs 28.7%, p = 0.14) were not significant. Targeted temperature management implementation and post-CA management varied between the participating centers. Classification and Regression Trees models identified variables predictive of favorable and poor outcome for hanging and TTM patients with excellent accuracy. CONCLUSION: Cardiac arrest hanging patients had worse outcome than non-CA patients. Targeted temperature management CA patients had worse unadjusted survival and neurologic outcome than non-TTM patients. These findings may be explained by their higher severity of illness, variable TTM implementation, and differences in post-CA management. Future prospective studies are necessary to ascertain the effect of TTM on hanging outcome and to validate our Classification and Regression Trees models. LEVEL OF EVIDENCE: Therapeutic study, level IV; prognostic study, level III.

Analyzing the Impact of Compliance with National Guidelines for Pancreatic Cancer Care Using the National Cancer Database.

IMPORTANCE: Management of pancreatic cancer is complex, requiring coordination of multiple providers. National Comprehensive Cancer Network guidelines, developed for standardization and quality improvement, recommend a multimodal approach. OBJECTIVE: This study analyzed national rates of compliance with National Comprehensive Cancer Network recommendations, assessed factors affecting compliance, and evaluated whether compliance with evidence-based guidelines improved overall survival. DESIGN: This is a retrospective review of adults diagnosed with pancreatic cancer entered into the National Cancer Database. Patients included had stage I and II pancreatic cancer, and complete data in the database. Patients were classified as compliant if they underwent both surgery and a second treatment modality (chemotherapy, radiation, or chemoradiation). Clinico-pathologic variables were analyzed using univariate and multivariate models to predict overall survival. SETTING: Hospital-based national study population. PARTICIPANTS: Patients with stage I or II pancreatic cancer. MAIN OUTCOMES AND MEASURES: Compliance with National Comprehensive Cancer Network recommendations, factors affecting compliance, and overall survival based on compliance. RESULTS: A total of 52,450 patients were included; 19,272 patients (37%) were compliant. Patients were found to be most compliant in the 50-59-year-old range (49% complaint), with decreased compliance at the extremes of age. Male patients were more compliant than female patients (39 vs 34%, p < 0.0001). Caucasians were more compliant (39%) than African Americans (32%) or other races (32%, p < 0.0001). Patients treated at academic/research centers were more compliant than patients treated at other facilities (39% compliant, p < 0.0001). Patients with stage II disease were more compliant compared with stage I disease (43 vs 18%, p < 0.0001). Compliance was shown to improve overall survival (p < 0.0001). CONCLUSION: Adherence to National Comprehensive Cancer Network guidelines for pancreatic cancer patients improves survival. Compliance nationwide is low, especially for older patients and minorities and those treated outside academic centers. More studies will need to be performed to identify factors that hinder compliance.

Outcomes of complicated appendicitis: Is conservative management as smooth as it seems?

BACKGROUND: This study characterized the failure rate of non-operative management (NOM) for complicated appendicitis (CA; perforation, abscess, phlegmon), and compared outcomes among patients undergoing acute appendectomy (AA), elective interval appendectomy (EIA), and unplanned appendectomy after failing to improve with NOM. METHODS: Adults treated at one facility between 2007 and 2014 were retrospectively studied. RESULTS: Ninety-five patients presented with CA. Sixty individuals underwent AA. The remaining 35 patients initially underwent NOM: 14 underwent EIA, nine (25.7%) failed NOM, 12 never underwent surgery. All patients failing NOM had an open operation with most (55.6%) requiring bowel resection. AA and EIA were comparable in surgical approach, bowel resection and post-operative readmission. However, AA demonstrated a lower incidence of bowel resection (3.3% vs 17.1%, P = 0.048) when compared to all patients initially undergoing NOM. CONCLUSIONS: Due to the high incidence of failed NOM and the morbidity associated with failure, AA may be appropriate for CA.

Disparities in compliance with the Oncotype DX breast cancer test in the United States: A National Cancer Data Base assessment.

BACKGROUND: Oncotype DX (ODX) is a multi-gene tumor assay for breast cancer patients. Our objective is to assess whether eligible ODX patients received the test and whether recommendations were followed based on respective risk. METHODS: We retrospectively analyzed testing in patients deemed eligible for ODX using the National Cancer Data Base. RESULTS: A total of 158,235 patients met ODX eligibility criteria. Sixty-four percent of eligible patients did not receive the test. Non-testing rose with age. White patients were more likely to be tested (56%) versus black patients (46%, p < 0.0001). Testing was highest at academic facilities (40%). Privately insured patients were more likely to get the test compared to uninsured (45 versus 34%, p < 0.0001). Those in the highest income quartile were more likely to be tested (p < 0.001). CONCLUSIONS: ODX is under-utilized, with racial and socio-economic factors influencing testing. Further studies are necessary to identify ways to remove disparities and increase testing when appropriate.

Management of computed tomography scan detected hemothorax in blunt chest trauma: What computed tomography scan measurements say?

AIM: To investigate the hemothorax size for which tube thoracostomy is necessary. METHODS: Over a 5-year period, we included all patients who were admitted with blunt chest trauma to our level 1 trauma center. Focus was placed on identifying the hemothorax size requiring tube thoracostomy. RESULTS: A total number of 274 hemothoraces were studied. All patients with hemothoraces measuring above 3 cm received a chest tube. The 50% predicted probability of tube thoracostomy was 2 cm. Pneumothorax was associated with odds of receiving tube thoracostomy for hemothoraces below 2 cm (Odds Ratio: 4.967, 95%CI: 2.225-11.097, P < 0.0001). CONCLUSION: All patients with a hemothorax size greater than 3% underwent tube thoracostomy. Prospective studies are warranted to elucidate the clinical outcome of patients with smaller hemothoraces.