In-Hospital Drug Use Affects a Broad Patient Population and the Health Care System.

INTRODUCTION: Patients who suffer from addiction sometimes take illicit substances while in the hospital (in-hospital drug use [IHDU]), which can lead to unnecessary work-up and preventable treatments when symptoms develop. The purpose of this study was to define the frequency and scope of this problem, who it affects, and its impact on the medical system. METHODS: We reviewed all incident reports from our large, urban level 1 trauma center that involved a patient being found with illicit substances in the hospital between January 2020 and January 2023. Reports were included if patients were witnessed taking the illicit substance or if they admitted to taking one in the hospital. Individual medical charts were then reviewed to determine the details of the incident and the response of the medical team. RESULTS: Thirty-one incidents meeting the inclusion criteria were found. Of the patients involved, 74.2% were male and ranged in age from 21 to 82 years old. The admitting services included internal or family medicine (48.4%), trauma (32.3%), surgical intensive care unit (ICU) (6.5%), medical ICU (6.5%), obstetrics and gynecology (3.2%), and plastic surgery (3.2%). Interventions ranged from no intervention to transfer to the ICU and intubation. 29.0% had visitors immediately prior to the incident, which may be how the substances were obtained. 16.1% left against medical advice after having their drugs confiscated. CONCLUSIONS: Our findings suggest that a broad range of patients are at risk for IHDU, and that IHDU can lead to medical complications, additional medical expense, or patients at risk of incomplete care due to leaving against medical advice. Future work is necessary to understand patient outcomes and financial implications of IHDU, as well as to develop tools to identify those patients most at risk.

A heterogenous population of extracellular vesicles mobilize to the alveoli postinjury.

BACKGROUND: Acute lung injury and subsequent resolution following severe injury are coordinated by a complex lung microenvironment that includes extracellular vesicles (EVs). We hypothesized that there is a heterogenous population of EVs recruited to the alveoli postinjury and that we could identify specific immune-relevant mediators expressed on bronchoalveolar lavage (BAL) EVs as candidate biomarkers of injury and injury resolution. METHODS: Mice underwent 30% TBSA burn injury and BAL fluid was collected 4 hours postinjury and compared with sham. Extracellular vesicles were purified and single vesicle flow cytometry (vFC) was performed using fluorescent antibodies to quantify the expression of specific cell surface markers on individual EVs. Next, we evaluated human BAL specimens from injured patients to establish translational relevance of the mouse vFC analysis. Human BAL was collected from intubated patients following trauma or burn injury, EVs were purified, then subjected to vFC analysis. RESULTS: A diverse population of EVs were mobilized to the alveoli after burn injury in mice. Quantitative BAL vFC identified significant increases in macrophage-derived CD44+ EVs (preinjury, 10.8% vs. postinjury, 13%; p < 0.05) and decreases in IL-6 receptor alpha (CD126) EVs (preinjury, 19.3% vs. postinjury, 9.3%, p < 0.05). Bronchoalveolar lavage from injured patients also contained a heterogeneous population of EVs derived from myeloid cells, endothelium, and epithelium sources, with CD44+ EVs being highly detected. CONCLUSION: Injury causes mobilization of a heterogeneous population of EVs to the alveoli in both animal models and injured patients. Defining EV release after injury will be critical in identifying diagnostic and therapeutic targets to limit postinjury acute lung injury.

A majority of trauma-related deep venous thromboses resolve prior to discharge.

Post-traumatic DVTs present unique challenges in patient populations with specific high-risk injury patterns. Duplex ultrasound (US) can be used to assess evolution of DVTs and may guide treatment for high-risk patients. We hypothesized that many DVTs resolve during the initial admission. Weekly duplex US are ordered on all trauma inpatients regardless of prior DVT at our facility. We reviewed US and outcomes data on all patients with lower extremity DVTs at our Level I trauma center from January 2012-December 2021. 392 patients were diagnosed with lower extremity DVT by US. 261 (67%) patients received follow-up US with a mean time to repeat US of 6 days. Of these, 91 (35%) patients experienced DVT resolution prior to the first follow-up US, and 141 (54%) patients experienced resolution prior to discharge. Mean time to resolution was 10 days. Over 50% of DVTs resolve before discharge and are detected by US. Further studies and post-discharge follow-up are needed to determine if patients with resolved DVTs can be managed without therapeutic anticoagulation.

Funding the war in America: A look in the mirror.

BACKGROUND: Health care political action committees (HPACs) historically contribute more to candidates opposing firearm restrictions (FRs), clashing with their affiliated medical societies. These societies have increasingly emphasized the prevention of firearm violence and it is not known if recent contributions by their HPACs have aligned with their stated goals. We hypothesized that such HPACs still contribute similar amounts toward legislators up for reelection opposing FR. METHODS: We identified HPACs of medical societies endorsing one or both calls-to-action against firearm violence published in the Annals of Internal Medicine (2015, 2019). House of Representatives (HOR) votes on H.R.8, a background checks bill, were characterized from GovTrack. We compiled HPAC contributions between the H.R.8 vote and election to HOR members up for re-election from the National Institute on Money in Politics. Our primary outcome was total campaign contributions by H.R.8 stance. Secondary outcomes included percentage of politicians funded and total contributions. RESULTS: Nineteen societies endorsed one or both call-to-action articles. Three hundred eighty-five of 430 HOR members ran for reelection in 2020. Those endorsing H.R.8 (n = 226, 59%) received $2.8 M for $4,750 (interquartile range [IQR], $1000-$15,500) per candidate. Those opposing (n = 159, 41%) received $1.5 M for $2,500 (IQR, $0-$11,000) per candidate ( p = 0.0057). Health care political action committees donated toward a median of 20% (IQR, 7-28) of candidates endorsing H.R.8 and 9% (IQR, 4-22) of candidates opposing H.R.8 ( p = 0.0014). Those endorsing H.R.8 received 1,585 total contributions for a median of 3 (IQR, 1-10) contributions per candidate, while those opposing received 834 total contributions for a median of 2 (IQR, 0-7) contributions per candidate ( p = 0.0029). CONCLUSION: Politicians voting against background checks received substantial contributions toward reelection from the HPACs of societies advocating for firearm restrictions. However, this is the first study to suggest that HPAC's contributions have become more congruent with their respective societies. Further alignment of medical society goals and their HPAC political contributions could have a profound impact on firearm violence. LEVEL OF EVIDENCE: Prognostic and Epidemiological; Level III.

Post-traumatic Liver Pseudoaneurysms: Rare but Serious Sequela.

INTRODUCTION: The liver is the most commonly injured organ after blunt abdominal trauma. Nonoperative management is the standard of care in stable individuals. Liver injuries, particularly high-grade injuries, can develop pseudoaneurysms (PSAs), which can rupture and cause life-threatening bleeding, even after hospital discharge. There is no consensus on whether patients should receive predischarge contrast computed tomography (CT) screening, or at what time interval after injury, nor which patients are at the highest risk for PSA. The purpose of this study was to identify the rates of PSA in our population and potential risk factors for their formation. METHODS: The trauma registry at our Level 1 urban trauma center was queried for patients admitted with liver injuries between 2015 and 2021. Demographic information was collected from the registry. Individual charts were then reviewed for timing of CT scans, CT findings, interventions, and complications. Liver injury grade was assessed using radiology reports or operative findings. The frequency of PSAs was then analyzed using descriptive statistics using Microsoft Excel and SPSS for odds ratio. RESULTS: A total of 172 patients were admitted with liver injuries during the study period. 130 patients received a CT scan diagnosing liver injury, 42 were diagnosed with liver injury intraoperatively. Of the 130 patients (59.9%) which received follow-up CT scans, six (6.5%) developed PSA, four of which being from penetrating injuries (odds ratio, 6.95). CONCLUSIONS: This study demonstrated a low incidence of PSA consistent with the known literature. We found the majority of the PSA developed following penetrating injury. This may represent a significant indication for follow-up imaging regardless of grade. A larger study will be necessary to identify those most at risk for PSA formation and determine the best PSA screening algorithm.

And Miles to Go Before We Sleep: EAST Diversity and Inclusivity Progress and Remaining Challenges.

OBJECTIVE: The aim of this study was to examine the diversity, equity, and inclusion landscape in academic trauma surgery and the EAST organization. SUMMARY BACKGROUND DATA: In 2019, the Eastern Association for the Surgery of Trauma (EAST) surveyed its members on equity and inclusion in the #EAST4ALL survey and assessed leadership representation. We hypothesized that women and surgeons of color (SOC) are underrepresented as EAST members and leaders. METHODS: Survey responses were analyzed post-hoc for representation of females and SOC in academic appointments and leadership, EAST committees, and the EAST board, and compared to the overall respondent cohort. EAST membership and board demographics were compared to demographic data from the Association of American Medical Colleges. RESULTS: Of 306 respondents, 37.4% identified as female and 23.5% as SOC. There were no significant differences in female and SOC representation in academic appointments and EAST committees compared to their male and white counterparts. In academic leadership, females were underrepresented ( P < 0.0001), whereas SOC were not ( P = 0.08). Both females and SOC were underrepresented in EAST board membership ( P = 0.002 and P = 0.043, respectively). Of EAST's 33 presidents, 3 have been white women (9%), 2 have been Black, non-African American men (6%), and 28 (85%) have been white men. When compared to 2017 AAMC data, women are well-represented in EAST's 2020 membership ( P < 0.0001) and proportionally represented on EAST's 2019-2020 board ( P > 0.05). CONCLUSIONS: The #EAST4ALL survey suggests that women and SOC may be underrepresented as leaders in academic trauma surgery. However, lack of high-quality demographic data makes evaluating representation of structurally marginalized groups challenging. National trauma organizations should elicit data from their members to re-assess and promote the diversity landscape in trauma surgery.

Decreasing use of pancreatic necrosectomy and NSQIP predictors of complications and mortality.

BACKGROUND: Surgical pancreatic necrosectomy (SPN) is an option for the management of infected pancreatic necrosis. The literature indicates that an escalating, combined endoscopic, interventional radiology and minimally invasive surgery "step-up" approach, such as video-assisted retroperitoneal debridement, may reduce the number of required SPNs and ICU complications, such as multiple organ failure. We hypothesized that complications for surgically treated severe necrotizing pancreatitis patients decreased during the period of adoption of the "step-up" approach. METHODS: The American college of surgeons national surgery quality improvement program database (ACS-NSQIP) was used to find SPN cases from 2007 to 2019 in ACS-NSQIP submitting hospitals. Mortality and Clavien-Dindo class 4 (CD4) ICU complications were collected. Predictors of outcomes were identified by univariate and multivariate analyses. RESULTS: There were 2457 SPN cases. SPN cases decreased from 0.09% in 2007 to 0.01% in 2019 of NSQIP operative cases (p < 0.001). Overall mortality was 8.5% and did not decrease with time. CD4 complications decreased from 40 to 27% (p < 0.001). There was a 65% reduction in SPN cases requiring a return to the operating room. Multivariate predictors of complications were emergency general surgery (EGS, p < 0.001), serum albumin (p < 0.0001) and modified frailty index (mFI) (p < 0.0001). Multivariate predictors of mortality were EGS (p < 0.0001), serum albumin (p < 0.0001), and mFI (p < 0.04). The mFI decreased after 2010 (p < 0.001). CONCLUSION: SPNs decreased after 2010, with decreasing CD4 complications, decreasing reoperation rates and stable mortality rates, likely indicating broad adoption of a "step-up" approach. Larger, prospective studies to compare indications and outcomes for "step up" versus open SPN are warranted.

Fluoxetine reduces organ injury and improves motor function after traumatic brain injury in mice.

BACKGROUND: Traumatic brain injury (TBI) is a leading cause of morbidity and mortality in trauma patients worldwide. Brain injury is associated with significant inflammation, both within the brain and in the peripheral organs. This inflammatory response in TBI leads to a secondary injury, worsening the effects of the original brain injury. Serotonin is also linked to inflammation in the intestine and inflammatory bowel disease, but its role in the gut-brain axis is not known. We hypothesized that using fluoxetine to block serotonin reuptake would reduce organ inflammation and improve outcomes after TBI. METHODS: C57/B6 mice were given a severe TBI using a controlled cortical impact. To measure intestinal permeability, a piece of terminal ileum was resected, the lumen was filled with 4-kDa fluorescein isothiocyanate (FITC)-dextran, and the ends were tied. The intestinal segment was submerged in buffer and fluorescence in the buffer measured over time. To measure lung permeability, 70-kDa FITC-dextran is injected retro-orbitally. Thirty minutes later, the left lung was homogenized and the fluorescence was measured. To measure performance on the rota-rod, mice were placed on a spinning rod, and the time to fall off was measured. Those treated with fluoxetine received a single dose of 5 mg/kg via intraperitoneal injection immediately after injury. RESULTS: Traumatic brain injury was associated with an increase in intestinal permeability to FITC-dextran, increased lung vascular permeability, and worse performance on the rota-rod. Fluoxetine significantly reduced lung and intestinal permeability after TBI and improved performance on the rota-rod after TBI. CONCLUSION: Use of fluoxetine has the potential to reduce lung injury and improve motor coordination in severe TBI patients. Further study will be needed to elucidate the mechanism behind this effect.

Youth violence prevention can be enhanced by geospatial analysis of trauma registry data.

BACKGROUND: Geographic information systems (GIS) have been used to understand relationships between trauma mechanisms, locations, and social determinants for injury prevention. We hypothesized that GIS analysis of trauma center registry data for assault patients aged 14 years to 29 years with census tract data would identify geospatial and structural determinants of youth violence. METHODS: Admissions to a Level I trauma center from 2010 to 2019 were retrospectively reviewed to identify assaults in those 14 years to 29 years. Prisoners were excluded. Home and injury scene addresses were geocoded. Cluster analysis was performed with the Moran I test for spatial autocorrelation. Census tract comparisons were done using American Communities Survey (ACS) data by t-test and linear regression. RESULTS: There were 1,608 admissions, 1,517 (92.4%) had complete addresses and were included in the analysis. Mean age was 23 ± 3.8 years, mean ISS was 7.5 ± 6.2, there were 11 (0.7%) in-hospital deaths. Clusters in six areas of the trauma catchment were identified with a Moran I value of 0.24 ( Z score = 17.4, p < 0.001). Linear regression of American Communities Survey demographics showed predictors of assault were unemployment (odds ratio, 4.5; 95% confidence interval, 2.7-6.4; p < 0.001), Spanish spoken at home (odds ratio, 6.6; 95% confidence interval, 3.4-9.8; p < 0.001) and poverty level (odds ratio, 1.9; 95% confidence interval, 1.1-2.7; p < 0.001). Education level of less than high school diploma, single parent households and race were not significant predictors. CONCLUSION: GIS analysis of registry data can identify high-risk areas for youth violence and correlated social and structural determinants. Violence prevention efforts can be better targeted geographically and socioeconomically with better understanding of these risk factors. LEVEL OF EVIDENCE: Prognostic/Epidemiological; Level III.

Precision targeting of the vagal anti-inflammatory pathway attenuates the systemic inflammatory response to burn injury.

BACKGROUND: The systemic inflammatory response (SIRS) drives late morbidity and mortality after injury. The α7 nicotinic acetylcholine receptor (α7nAchR) expressed on immune cells regulates the vagal anti-inflammatory pathway that prevents an overwhelming SIRS response to injury. Nonspecific pharmacologic stimulation of the vagus nerve has been evaluated as a potential therapeutic to limit SIRS. Unfortunately, the results of clinical trials have been underwhelming. We hypothesized that directly targeting the α7nAchR would more precisely stimulate the vagal anti-inflammatory pathway on immune cells and decrease gut and lung injury after severe burn. METHODS: C57BL/6 mice underwent 30% total body surface area steam burn. Mice were treated with an intraperitoneal injection of a selective agonist of the α7nAchR (AR-R17779) at 30 minutes postburn. Intestinal permeability to 4 kDa FITC-dextran was measured at multiple time points postinjury. Lung vascular permeability was measured 6 hours after burn injury. Serial behavioral assessments were performed to quantify activity levels. RESULTS: Intestinal permeability peaked at 6 hours postburn. AR-R17779 decreased burn-induced intestinal permeability in a dose-dependent fashion (p < 0.001). There was no difference in gut permeability to 4 kDa FITC-dextran between sham and burn-injured animals treated with 5 mg/kg of AR-R17779. While burn injury increased lung permeability 10-fold, AR-R17779 prevented burn-induced lung permeability with no difference compared with sham (p < 0.01). Postinjury activity levels were significantly improved in burned animals treated with AR-R17779. CONCLUSION: Directly stimulating the α7nAchR prevents burn-induced gut and lung injury. Directly targeting the α7nAChR that mediates the cholinergic anti-inflammatory response may be an improved strategy compared with nonspecific vagal agonists.

CHRFAM7A expression in mice increases resiliency after injury.

INTRODUCTION: The CHRNA7 gene encodes the α-7 nicotinic acetylcholine receptor (α7nAchR) that regulates anti-inflammatory responses to injury; however, only humans express a variant gene called CHRFAM7A that alters the function of α7nAChR; CHRFAM7A expression predominates in bone marrow and monocytes/macrophages where the CHRFAM7A/CHRNA7 ratio is highly variable between individuals. We have previously shown in transgenic mice that CHRFAM7A increased emergency myelopoiesis from the bone marrow and monocyte/macrophage expression in lungs. MATERIALS AND METHODS: CHRFAM7A transgenic mice are compared to age- and gender-matched wild-type (WT) siblings. We utilized a model of sepsis using LPS injection to measure survival. Lung vascular permeability was measured after severe burn injury in WT vs. CHRFAM7A transgenic mice. Bone marrow CHRFAM7A expression was evaluated using adoptive transfer of CHRFAM7A transgenic bone marrow into WT mice. RESULTS: Here, we demonstrate that CHRFAM7A expression results in an anti-inflammatory phenotype with an improved survival to LPS and decreased acute lung injury in a severe cutaneous burn model compared to WT. CONCLUSIONS: These data suggest that the relative expression of CHRFAM7A may alter resiliency to injury and contribute to individual variability in the human systemic inflammatory response (SIRS) to injury.

Methamphetamine Use is Associated with Increased Surgical Site Infections after Trauma Laparotomy.

BACKGROUND: Methamphetamine (METH) use causes significant vasoconstriction, which can be severe enough to cause bowel ischemia. Methamphetamines have also been shown to alter the immune response. These effects could predispose METH users to poor wound healing, increased infections, and other post-operative complications. We hypothesized that METH users would have longer length of stay and higher rates of complications compared to non-METH users. METHODS: The trauma registry for our urban Level 1 trauma center was searched for patients that received an exploratory laparotomy from 2016 to 2019. A total 204 patients met criteria and 52 (25.5%) were METH positive. Length of stay (LOS), ventilator days, abbreviated injury scale (AIS), and wound class were compared using nonparametric statistics. Age and injury severity score (ISS) were compared using a Student's t-test. A Chi Square or Fisher's Exact test was used to compare sex, mechanism of injury, and rates of infectious complications. RESULTS: Methamphetamine-positive patients had a significantly higher rate of surgical site infections (7.4% versus 0%, P = 0.001). Patients that developed surgical site infection had equivalent rates of smoking and diabetes, as well as equivalent abdominal AIS and wound class compared to those who did not develop surgical site infection. Hospital and ICU LOS, ventilator days, ISS, and mortality were equivalent between METH positive and negative patients. Rates of other infectious complications were the same between groups. CONCLUSIONS: Methamphetamine use is associated with an increased rate of surgical site infection after trauma laparotomy. Other serious complications and mortality were not affected by METH use.

Don't mess with the pancreas! A multicenter analysis of the management of low-grade pancreatic injuries.

INTRODUCTION: Current guidelines recommend nonoperative management (NOM) of low-grade (American Association for the Surgery of Trauma-Organ Injury Scale Grade I-II) pancreatic injuries (LGPIs), and drainage rather than resection for those undergoing operative management, but they are based on low-quality evidence. The purpose of this study was to review the contemporary management and outcomes of LGPIs and identify risk factors for morbidity. METHODS: Multicenter retrospective review of diagnosis, management, and outcomes of adult pancreatic injuries from 2010 to 2018. The primary outcome was pancreas-related complications (PRCs). Predictors of PRCs were analyzed using multivariate logistic regression. RESULTS: Twenty-nine centers submitted data on 728 patients with LGPI (76% men; mean age, 38 years; 37% penetrating; 51% Grade I; median Injury Severity Score, 24). Among 24-hour survivors, definitive management was NOM in 31%, surgical drainage alone in 54%, resection in 10%, and pancreatic debridement or suturing in 5%. The incidence of PRCs was 21% overall and was 42% after resection, 26% after drainage, and 4% after NOM. On multivariate analysis, independent risk factors for PRC were other intra-abdominal injury (odds ratio [OR], 2.30; 95% confidence interval [95% CI], 1.16-15.28), low volume (OR, 2.88; 1.65, 5.06), and penetrating injury (OR, 3.42; 95% CI, 1.80-6.58). Resection was very close to significance (OR, 2.06; 95% CI, 0.97-4.34) (p = 0.0584). CONCLUSION: The incidence of PRCs is significant after LGPIs. Patients who undergo pancreatic resection have PRC rates equivalent to patients resected for high-grade pancreatic injuries. Those who underwent surgical drainage had slightly lower PRC rate, but only 4% of those who underwent NOM had PRCs. In patients with LGPIs, resection should be avoided. The NOM strategy should be used whenever possible and studied prospectively, particularly in penetrating trauma. LEVEL OF EVIDENCE: Therapeutic Study, level IV.

Using Machine Learning to Make Predictions in Patients Who Fall.

BACKGROUND: As the population ages, the incidence of traumatic falls has been increasing. We hypothesize that a machine learning algorithm can more accurately predict mortality after a fall compared with a standard logistic regression (LR) model based on immediately available admission data. Secondary objectives were to predict who would be discharged home and determine which variables had the largest effect on prediction. METHODS: All patients who were admitted for fall between 2012 and 2017 at our level 1 trauma center were reviewed. Fourteen variables describing patient demographics, injury characteristics, and physiology were collected at the time of admission and were used for prediction modeling. Algorithms assessed included LR, decision tree classifier (DTC), and random forest classifier (RFC). Area under the receiver operating characteristic curve (AUC) values were calculated for each algorithm for mortality and discharge to home. RESULTS: About 4725 patients met inclusion criteria. The mean age was 61 ± 20.5 y, Injury Severity Score 8 ± 7, length of stay 5.8 ± 7.6 d, intensive care unit length of stay 1.8± 5.2 d, and ventilator days 0.7 ± 4.2 d. The mortality rate was 3% and three times greater for elderly (aged 65 y and older) patients (5.0% versus 1.6%, P < 0.001). The AUC for predicting mortality for LR, DTC, and RFC was 0.78, 0.64, and 0.86, respectively. The AUC for predicting discharge to home for LR, DTC, and RFC was 0.72, 0.61, and 0.74, respectively. The top five variables that contribute to the prediction of mortality in descending order of importance are the Glasgow Coma Score (GCS) motor, GCS verbal, respiratory rate, GCS eye, and temperature. CONCLUSIONS: RFC can accurately predict mortality and discharge home after a fall. This predictive model can be implemented at the time of patient arrival and may help identify candidates for targeted intervention as well as improve prognostication and resource utilization.

Is there a glass ceiling at national trauma meetings?

BACKGROUND: The purpose of this study was to evaluate the roles of women at national trauma meetings. METHODS: Available scientific programs for the American Association for the Surgery of Trauma (2013-19), Eastern Association for the Surgery of Trauma (2010-19), and Western Trauma Association (2010-19) as well as the Scudder Oration at the American College of Surgeons (1963-2019), were reviewed for names of participants and categorized by gender. RESULTS: Women made up 963 of 2746 (35.1%) of presenters, 252 of 1020 (24.7%) of discussants, 116 of 622 (18.6%) of moderators of scientific sessions, 189 of 707 (26.7%) of panelists, and 69 of 254 (27.2%) of panel moderators. Only 12 of 126 (9.5%) of named lectures or presidential addresses were given by women. CONCLUSIONS: The low rate of female named speakers suggests that there remains a "glass ceiling" when it comes to upper-level participation in national trauma meetings.

Outcomes in Delayed Drainage of Hemothorax.

BACKGROUND: Prompt drainage of traumatic hemothorax is recommended to prevent empyema and trapped lung. Some patients do not present the day of their trauma, leading to their delayed treatment. Delayed drainage could be challenging as clotted blood may not evacuate through a standard chest tube. We hypothesized that such delays would increase the need for surgery or secondary interventions. METHODS: Our trauma registry was reviewed for patients with a hemothorax admitted to our level 1 trauma center from 1/1/00 to 4/30/19. Patients were included in the delayed group if they received a drainage procedure >24 hours after injury. These patients were matched 1:1 by chest abbreviated injury score to patients who received drainage <24 hours from injury. RESULTS: A total of 19 patients with 22 hemothoraces received delayed drainage. All but 3 patients had a chest tube placed as initial treatment. Four patients received surgery, including 3 who initially had chest tubes placed. Longer time to drainage increased the odds of requiring intrathoracic thrombolytics or surgery. In comparison, 2 patients who received prompt drainage received thrombolytics (P = .11) and none required surgery (P = .02). Patients needed surgery when initial drainage was on or after post-injury day 5, but pigtail catheter drainage was effective 26 days after injury. DISCUSSION: Longer times from injury to intervention are associated with increased likelihood of needing surgery for hemothorax evacuation, but outcomes were not uniform. A larger, multicenter study will be necessary to provide better characterization of treatment outcomes for these patients.