AbstractObjective

COVID-19 has been shown to affect outcomes among surgical patients. We hypothesized that COVID-19 would be linked to higher mortality and longer length of stay of trauma patients regardless of the Injury Severity Score.

Methods

We performed a retrospective analysis of trauma registries from two level I trauma centers (suburban and urban) from 3/1/2019-6/30/2019 and 3/1/2020-6/30/2020, comparing baseline characteristics and cumulative adverse events. Data collected included ISS, demographics, and comorbidities. The primary outcome was time from hospitalization to in-hospital death. Outcomes during the height of the first New York COVID-19 wave were also compared to the same timeframe in the prior year. Kaplan-Meier method with Log-rank test and Cox proportional hazard models were used to compare outcomes.

Results

There were 1,180 trauma patients admitted during the study period from 3/2020-6/2020. Of these, 596 were never tested for COVID-19 and were excluded from analysis. A total of 148 COVID+ patients and 436 COVID- patients composed the 2020 cohort for analysis. Compared to the 2019 cohort, the 2020 cohort was older with more associated comorbidities, more adverse events, but lower ISS. Higher rates of historical hypertension, diabetes, neurologic events, and coagulopathy were found among COVID+ patients compared to COVID- patients. D-dimer and ferritin were unreliable indicators of COVID-19 severity; however, CRP levels were higher in COVID+ relative to COVID- patients. Patients who were COVID+ had a lower median ISS compared to COVID- patients, and COVID+ patients had higher rates of mortality and longer LOS.

Conclusions

COVID+ trauma patients admitted to our two Level 1 trauma centers had increased morbidity and mortality compared to admitted COVID- trauma patients despite age and lower ISS. CRP may play a role in monitoring COVID-19 activity in trauma patients. A better understanding of the physiologic impact of COVID-19 on injured patients warrants further investigation.

IntroductionGlobally, a staggering 248 million cases and 5 million deaths have occurred since SARS-CoV-2 (severe acute respiratory syndrome coronavirus-2) spread worldwide in January of 2020.,2

Dong E, Du H, Gardner L. An interactive web-based dashboard to track COVID-19 in real time. COVID-19 Dashboard by the Center for Systems Science and Engineering (CSSE) at Johns Hopkins University. Lancet Inf Dis. 20(5):533-534. doi: 10.1016/S1473-3099(20)30120-1.

COVID-19 (coronavirus disease-2019) affected all patient populations, including patients presenting to the hospital with traumatic injuries, adding an extra burden to patients’ health and possibly changing their outcomes. Data on COVID-19 and trauma patients are scarce with few published studies comparing outcomes between trauma patients with and without COVID-19.Several studies have shown the deleterious effects of COVID-19 on patients with orthopedic trauma.3Kumar A. Haider Y. Passey J. Khan R. Gaba S. Kumar M. Mortality Predictors in Covid-19 Positive Patients with Fractures: A Systematic Review., 4Al Farii H. Al Rawahi S. Samaila E. Lavini F. Magnan B. Al Maskari S. Thirty-Day Mortality in COVID-19 Positive Patients With Hip Fractures: A Case-Series and Literature Review., 5Patralekh M.K. Jain V.K. Iyengar K.P. Upadhyaya G.K. Vaishya R. Mortality escalates in patients of proximal femoral fractures with COVID-19: A systematic review and meta-analysis of 35 studies on 4255 patients. Other studies on the impact of the pandemic on trauma centers and outcomes of trauma patients revealed an increased risk of mortality, longer length of stay and pulmonary complications in COVID-19 positive (COVID 19+) patients, along with fewer traumatic blunt injury encounters.6Kaufman E.J. Ong A.W. Cipolle M.D. Whitehorn G. Ratnasekera A. Stawicki S.P. Martin N.D. The impact of COVID-19 infection on outcomes after injury in a state trauma system.,7Sheets N.W. Fawibe O.S. Mahmoud A. Chawla-Kondal B. Ayutyanont N. Plurad D.S. Impact of the COVID-19 Pandemic on Trauma Encounters. Increased complications have also been seen in other surgical COVID 19+ patients as compared to COVID-19 negative (COVID 19-) patients; mainly pulmonary in nature, followed by multi-organ complications.8Mortality and pulmonary complications in patients undergoing surgery with perioperative SARS-CoV-2 infection: an international cohort study. To our knowledge, none of these studies identified a biochemical marker associated with the risk of worse outcomes for COVID 19+ trauma patients. In the current study, in addition to analysis of COVID 19+/COVID 19- trauma patient outcomes, we performed a further analysis on the usual inflammatory markers monitored during the management of COVID 19+ patients. We hypothesized that COVID 19+ trauma patients would have a higher mortality rate and length of stay (LOS) and that the use of inflammatory biomarkers could help guide the management of trauma patients with COVID-19.Materials and MethodsApproval from our Institutional Review Board (IRB) was obtained, and a waiver of informed consent granted. We then performed a retrospective analysis of the prospectively accrued trauma registries of two Level I trauma centers (one suburban and the other urban) from 3/1/2020 to 6/30/2020 and 3/1/2019 to 6/30/2019, following the STROBE (STrengthening the Reporting of OBservational studies in Epidemiology) guidelines. We compared outcomes and baseline characteristics of the 2020 cohort to trauma patients from the same period in 2019, demonstrating that the groups were comparable before and after the pandemic. Patients from 2020 were categorized into subgroups of COVID 19+ or COVID 19- based on diagnosis codes and COVID-19 testing results found in the EHR (Electronic Health Record). Inflammatory biomarkers including CRP (C-reactive protein), D-dimer and Ferritin were interrogated to identify any patterns in COVID19+ versus COVID 19- injured patients. Comorbidity status was determined using Elixhauser’s algorithm based on ICD-9-CM and ICD-10 diagnosis codes.9Elixhauser A. Steiner C. Harris D.R. Coffey R.M. Comorbidity measures for use with administrative data.,10Quan H. Sundararajan V. Halfon P. Fong A. Burnand B. Luthi J.C. et al.Coding algorithms for defining comorbidities in ICD-9-CM and ICD-10 administrative data. R-package “comorbidity” (https://cran.r-project.org/) was applied to compute comorbidity indices.

Demographic characteristics were summarized by groups and presented using the median (Interquartile Range, IQR) or frequency (percentage) as appropriate. Continuous variables were assessed for normality using the Kolmogorov-Smirnov test, histogram, and Q-Q plot. The baseline characteristics were compared between groups using the Wilcoxon rank-sum test for continuous variables and chi-square test or Fisher’s exact test for categorical variables.

Survival estimates were compared using the Kaplan-Meier method, and the log-rank test was used to compare the survival curves between the treatment groups. Adjusted models were fit using Cox proportional hazards; model selection was performed using a stepwise approach. Model fit was assessed using the Akaike information criterion (AIC). Proportional hazard assumptions were evaluated using graphical analysis, martingale residuals, and the Kolmogorov-Type Supremum test.11Lin D.Y. Wei L.J. Ying Z. Checking the Cox Model with Cumulative Sums of Martingale-Based Residuals.,12Testing Goodness of Fit for Proportional Hazards Model with Censored Observations. Time to hospital discharge was also analyzed utilizing similar techniques using death or end of follow-up as censored. A logistic regression model was used to determine the predictive ability of CRP for death. Area under the curve (AUC) was computed, and Hosmer and Lemeshow test was used to assess the model fit. SAS 9.4 and R 4.0.4 were used to perform all analyses, and p-values of 0.05 or less were considered to indicate statistical significance.ResultsFigure 1 summarizes patient selection for the study. There were 1,180 trauma patients admitted during the study period from March 1, 2020 to June 30, 2020. Of these, 596 were not tested for COVID-19 and were excluded from the study. The remaining 584 patients were divided into two arms: COVID 19+ (n=148) and COVID 19- (n=436).

Figure 1Data flow showing how many patients were included and excluded from the database. Patients not assessed for COVID-19 were excluded.

To reduce errors, we compared trauma patients during the pandemic to patients before the pandemic, looking to identify any difference in ISS, mortality and morbidities. When patients from 2020 were compared to those from the same period in 2019, patients in 2020 were older with an average age of 76.8 compared to 73.0 years old (p<0.001). Trauma patients in 2020 also had significantly more comorbidities: obesity, hypertension, diabetes, heart failure, renal failure, cardiac arrhythmias, coagulopathy, neurological disorders, and pulmonary circulation disorders.

Table 1 compares demographics and clinical characteristics between the 2020 cohorts. COVID 19+ patients were significantly older (79.6 vs 74.7 years; pFigure 2).

Table 1Demographics and clinical characteristics.

ISS: Injury Severity Score; HTN: Hypertension; DM: Diabetes Mellitus; CHF: Congestive Heart Failure; CRP: C-Reactive Protein; LDH: Lactated Dehydrogenase

Figure 2Predictive ability of CRP for mortality.

The primary endpoint was time from admission to in-hospital death, assuming hospital discharge or end of follow-up as censored. Table 2 shows that of 148 COVID 19+ trauma patients, only 37 of them were discharged within fourty days.

Table 2Death censored time to discharge Kaplan-Meier failure curves.

Censored: (46 death + 20 were not discharged prior to 6/25/20)

Multivariable analysis using Cox proportional hazard regression revealed a significantly higher risk of death in COVID 19+ trauma patients compared to those who were COVID 19- (Hazard Ratio, HR [95% Confidence Interval] = 2.7[1.5-5.2], p=0.002) (Table 3). In addition to COVID-19 status, age, coagulopathy, and other neurological disorders were independently associated with higher risk of mortality. Similarly, multivariable analysis demonstrated that the COVID 19+ patients are less likely to get discharged (HR [95% CI] = 0.63[0.49-0.80], pTable 4). The significant factors influencing longer discharge time in COVID 19+ trauma patients were congestive heart failure, fluid/electrolyte disorders, neurological disorders, and pulmonary circulation disorders (Table 5).

Table 3Unadjusted and Adjusted hazard ratios for “time to death” estimated via Cox proportional hazard regression models.

*Patients were assumed COVID+ based on a positive PCR test or COVID diagnosis.

* Please note, our main purpose was to evaluate if the COVID status significantly affects death in trauma patients. So, only characteristics with p-values Table1) were considered for the multivariable model. Factors with p-value>=0.25 were deemed irrelevant in this context.

Table 4Unadjusted and Adjusted hazard ratios for “time to discharge” estimated via Cox proportional hazard regression models.

* Please note, our main purpose was to evaluate if the COVID status significantly affects death in trauma patients. So, only characteristics with p-values Table1) were considered for the multivariable model. Factors with p-value>=0.25 were deemed irrelevant in this context.

Table 5Demographics and clinical characteristics between 2019 and 2020 patients.

HTN: Hypertension; DM: Diabetes Mellitus; CHF: Congestive Heart Failure.

COVID 19+ patients had a lower overall survival probability than COVID 19- patients (Figure 3). Similarly, COVID 19+ patients had longer discharge probability with a longer time to discharge (Median [IQR] = 5[4-7] vs. 3[3-4] days) compared to COVID 19- patients (Figure 3).

Figure 3Comparing Kaplan Meier survival curves for time to death and time to discharge between groups.

Discussion

In this study, we were able to demonstrate increased mortality in COVID 19+ patients, our primary endpoint, despite a lower ISS as compared to the COVID- cohort. We also reported that trauma patients with a diagnosis of COVID-19 requiring a hospital admission could have longer hospital lengths of stay. One major contributing factor was the need for negative COVID-19 PCR results for discharge of our trauma patients to a rehabilitation center, with many of the recovering trauma patients showing no symptoms of COVID-19 continuing to test positive, prolonging their hospital stay.

After the first reported case of COVID-19 in the United States, new protocols were developed by the health care system to diagnose patients presenting with COVID-19 symptoms. Diagnosis included the PCR (polymerase chain reaction), radiographic imaging such as chest CT (computed tomography) imaging, and a high clinical suspicion of COVID-19 after exclusion of other causes.13Holshue M.L. DeBolt C. Lindquist S. Lofy K.H. Wiesman J. Bruce H. et al.First Case of 2019 Novel Coronavirus in the United States. In New York City, a rapid increase in patients was attributed to factors such as dense population and social practices. The health care system was unprepared for the vast number of patients presenting with COVID-19 respiratory symptoms. Due to limited available personnel, facilities, disposable equipment and finite SARS-CoV-2 testing technology, only selective COVID-19 PCR testing was implemented early in the pandemic. Therefore, only 580 out of 1,180 patients were tested for COVID-19 in our study (Figure 1).14Population density, a factor in the spread of COVID-19 in Algeria: statistic study. On admission, we utilized high clinical suspicion coupled with findings on routine trauma chest x-rays and chest CT scans (if obtained) to selectively test and diagnose for COVID-19. This practice was consistent across campuses. As tests became more available however, we routinely tested every trauma patient admitted to the institution.Older adult patients with COVID-19 are at a higher risk of rapid progressive deterioration in comparison to younger patients.15Perrotta F. Corbi G. Mazzeo G. Boccia M. Aronne L. D'Agnano V. et al.COVID-19 and the elderly: insights into pathogenesis and clinical decision-making. In our study, worse outcome of the trauma patient was independent of age and correlated with the presence of COVID-19 disease. When we controlled for age and ISS, COVID 19+ patients had worse outcomes.20Bulger E.M. Arneson M.A. Mock C.N. Jurkovich G.J. Rib fractures in the elderly.,21Comprehensive approach to the management of the patient with multiple rib fractures: a review and introduction of a bundled rib fracture management protocol. Surprisingly, in our study, patients who were COVID 19- had a higher ISS but still had significantly better outcomes. These associations could be explained by the systemic inflammatory response in COVID 19+ patients, leading to multi-organ dysfunction, potentially responsible for worse outcomes.Currently, the Centers for Disease Control and Prevention (CDC) report a nearly similar positive COVID-19 caseload between different ethnicities in the US. In our study, one third of each ethnicity tested for COVID-19 returned positive, echoing similar results to that published by the CDC. The ratio of COVID 19- patients was higher in the suburban trauma centers, perhaps a reflection of the lower population density, which may allow for decreased contact spread.Systemic coagulopathy is now a recognized manifestation of COVID-19 illness.23Biffl W.L. Harrington D.T. Majercik S.D. Starring J. Cioffi W.G. The evolution of trauma care at a level I trauma center. When combined with the coagulopathy of trauma, systemic coagulopathy may complicate the management of those patients. Patients with contraindications to chemoprophylaxis, such as patients with traumatic brain injuries, patients undergoing major procedures that require holding anti-coagulation, or patients with active bleeding are at increased risk.24Maegele M. Schöchl H. Cohen M.J. An update on the coagulopathy of trauma. At the time of our data collection, there were no standardized guidelines for prophylactic anti-coagulant dosing in COVID 19+ trauma patients. At our institution, we implemented a process that relied on the level of D-dimer in those patients. However, elevation of D-dimer levels is a known sequelae of trauma, and it is unclear what the impact of the presence of COVID-19 had on these values in the setting of trauma.25COVID-19 update: Covid-19-associated coagulopathy., 26Grosse C. Grosse A. Salzer H.J.F. Dünser M.W. Motz R. Langer R. Analysis of cardiopulmonary findings in COVID-19 fatalities: High incidence of pulmonary artery thrombi and acute suppurative bronchopneumonia., 27Long B. Brady W.J. Koyfman A. Gottlieb M. Cardiovascular complications in COVID-19.Inflammatory markers were recognized early in the pandemic as methods by which COVID 19+ patients could be managed. Biomarkers such as LDH, CRP, ferritin, and D-dimer were routinely measured in that population. Our data demonstrate an association with CRP and COVID 19+ trauma patients and outcomes. Similarly, D-dimer levels have been used as indicator for initiation of anticoagulation. D-dimer is a useful marker for inflammation, clot breakdown and blood vessel damage.31van Paassen J. Vos J.S. Hoekstra E.M. Neumann K.M.I. Boot P.C. Arbous S.M. Corticosteroid use in COVID-19 patients: a systematic review and meta-analysis on clinical outcomes. D-dimer levels are routinely checked in COVID+ patients to assess their coagulopathy and disease severity, aiding in their management. However, D-dimer could also be elevated due to traumatic injury, minimizing its usefulness in assessing COVID-19 severity and coagulopathy in trauma patients.32Kirchhof M.G. Lee A.Y. Dutz J.P. D-dimer levels as a marker of cutaneous disease activity: case reports of cutaneous polyarteritis nodosa and atypical recurrent urticaria.,33Johna S. Cemaj S. O'Callaghan T. Catalano R. Effect of tissue injury on D-Dimer levels: a prospective study in trauma patients. Ferritin and LDH are other inflammatory markers that are both elevated due to trauma and COVID-19. In our population, COVID 19+ patients had significantly elevated CRP levels as compared to patients who were COVID 19-. This finding persisted in analysis suggesting that COVID 19+ injured patients could have their outcomes predicted by following the CRP trend. Understanding how this could influence care requires assessing a larger population in a controlled setting. Although CRP is elevated in traumatic injury, our COVID+ patients’ CRP levels were significantly elevated in comparison to COVID- patients, suggesting utility as a prognostic tool.34Zhang L. Yan X. Fan Q. Liu H. Liu X. Liu Z. Zhang Z. D-dimer levels on admission to predict in-hospital mortality in patients with Covid-19.

This is a prospectively collected retrospective review of a patient population with inherent limitations. In this model, the outcome is subject to confounding bias unlike a prospective randomized study. Because the data were also reviewed retrospectively, some data were missed or unmeasured further compounding bias. Another limitation is the restricted COVID-19 testing at the start of the pandemic due to limited testing availability and different protocols that stated only patients with symptoms were to be tested. It is possible that some patients may have been positive with the virus while remaining asymptomatic. It is therefore impossible to know if this led to an underestimation of COVID 19+ patients, thereby magnifying the impact of the virus on outcomes. Based on subsequent experience, we do not believe this to be likely however.

Authors’ contributions:

Design: DKJ, PP, GB

Data acquisition: HH, NG

Statistical analysis: SI, JD

Drafting manuscript: HH, AS, AHS, RJ, AS, LV

Supervision: DKJ, SI, PP

Final approval: HH, SI, PP, AS, GB, AHS, NG, RJ, AS, LV, JD, DKJ

The authors declare that they have not conflict of interest.

Article InfoPublication History

Accepted: June 30, 2022

Received in revised form: June 17, 2022

Received: December 6, 2021

Publication stageIn Press Accepted ManuscriptFootnotes

This work was presented as a Poster at the 80th Annual Meeting of the American Association for the Surgery of Trauma. Atlanta, GA. September 29 – October 2, 2021.

Identification

DOI: https://doi.org/10.1016/j.jss.2022.06.068

Copyright

© 2022 Elsevier Inc. All rights reserved.

ScienceDirectAccess this article on ScienceDirect