Comparison of thrombotic and clinical outcomes in SARS-CoV-2-pneumonia versus other viral pneumonia in an urban academic medical center

Abstract

Background

Infection with viral pneumonia (PNA) is known to offset the coagulation cascade. Recent studies assessing novel SARS-CoV-2 infection observed a high frequency of systemic thrombotic events resulting in ambiguity if severity of infection or specific viral strain drive thrombosis and worsen clinical outcomes. Furthermore, limited data exists addressing SARS-CoV-2 in underrepresented patient populations.

Objectives

Assess clinical outcomes events and death in patients diagnosed with SARS-CoV-2 pneumonia compared to patients with other types of viral pneumonia.

Methods

Retrospective cohort study evaluated electronic medical records in adult patients admitted to University of Illinois Hospital and Health Sciences System (UIHHSS) with primary diagnosis of SARS-CoV-2 PNA or other viral (H1N1 or H3N2) PNA between 10/01/2017 and 09/01/2020. Primary composite outcome was the following event incidence rates: death, ICU admission, infection, thrombotic complications, mechanical ventilation, renal replacement therapy, and major bleeding.

Results

Of 257 patient records, 199 and 58 patients had SARS-CoV-2 PNA and other viral PNA, respectively. There was no difference in primary composite outcome. Thrombotic events (n = 6, 3%) occurred solely in SARS-CoV-2 PNA patients in the ICU. A significantly higher incidence of renal replacement therapy (8.5% vs 0%, p=0.016) and mortality (15.6% vs 3.4%, p=0.048) occurred in the SARS-CoV-2 PNA group. Multivariable logistic regression analysis revealed age, presence of SARS-CoV-2, and ICU admission, aOR 1.07, 11.37, and 41.95 respectively, was significantly associated with mortality risk during hospitalization; race and ethnicity were not.

Conclusion

Low overall incidence of thrombotic events occurred only in the SARS-CoV-2 PNA group. SARS-CoV-2 PNA may lead to higher incidence of clinical events than those observed in H3N2/H1N1 viral pneumonia, and that race/ethnicity does not drive mortality outcomes.

List of Abbreviations

ASH

American Society of Hematology

CHEST

American College of Chest Physicians

DVT

Deep Vein Thrombosis

DIC

Disseminated Intravascular Coagulation

H1N1

Hemagglutinin Type 1 and Neuraminidase Type 1 Influenza

H3N2

Influenza A virus subtype H3N2

ICU

Intensive Care Unit

LMWH

Low Molecular Weight Heparin

PE

Pulmonary Embolism

RRT

Renal Replacement Therapy

SARS-CoV-2

Severe Acute Respiratory Syndrome CoronaVirus-2

UIHHS

University of Illinois Health & Hospital Sciences System

VTE

Venous Thromboembolism

Introduction

As severe acute respiratory syndrome coronavirus-2 (SARS-CoV-2) continues to spread both on a global level and within the United States the clinical complications attributed to this virus and its disproportionate toll on minority groups remains ill-defined.^{1 , 2} Previous literature have noted increasing rates of viral PNA as a serious complication of SARS-CoV-2, with patients at increased risk of multi-organ failure and death.3, 4, 5 Additionally, increased rates of coagulation abnormalities, leading to a greater risk for arterial and venous thrombotic events, have also been observed in patients infected with SARS-CoV-2, and linked to a poor prognosis.^{4 , 5}

A prior study revealed that the prothrombic state in SARS-CoV-2-infected patients manifested as an elevated D-dimer and prolonged prothrombin time (PT).⁴ Importantly, significantly increased D-dimer levels were observed in non-survivors versus survivors (median D-dimer 2.12 ug/mL versus 0.66 ug/mL, respectively) with an overall mortality rate of 11%. It has been well described that viral infections lead to a prothrombotic state, however the extent to which certain strains have this effect compared to SARS-CoV-2 is not well understood.⁶ The viral PNA variant strains of influenza Type A, H1N1 and H3N2, are of special interest because these strains are most frequently associated with causing human illness during the influenza season.⁷ Additionally, the H1N1 strain emerging in 2009 caused the first global pandemic of this century while H3N2, a variant strain of H1N1 that appeared in 2011, has been linked to high hospitalization rates.^{7 , 8} Coagulation complications secondary to seasonal influenza PNA have previously been attributed to respiratory tract inflammation, resulting in activation of tissue factor and the coagulation system.⁹ In 2018, the American Society of Hematology (ASH) 2018 clinical practice guidelines on VTE prophylaxis in hospitalized patients rendered a strong recommendation for pharmacologic VTE prophylaxis over mechanical prophylaxis for hospitalized acutely ill patients since infection is recognized as a risk enhancer for clot formation.¹⁰ As a result of data on venous thromboembolism rates in hospitalized SARS-CoV-2 patients and evidence suggesting pharmacologic thromboprophylaxis led to improved outcomes in early SARS-CoV-2 trials, ASH updated the VTE prophylaxis recommendations to specifically include patients hospitalized with SARS-CoV-2 infection.¹¹ Interestingly, expert consensus panel recommendations remained unclear due to inconsistent clinical findings until recently.12, 13, 14, 15, 16 Thus, thrombotic complications both in SARS-CoV-2 and other viral PNA patients have been reported^{9 , 11}; however, it is crucial to understand the differences in coagulopathy and other clinical outcomes between viral PNA strains to improve patient care.

In addition to the clinical and thrombotic consequences observed in patients with SARS-CoV-2 infection, heightened racial disparity in outcomes has also been reported.² Black-non-Hispanic, Hispanic, and Latinx racial or ethnic groups experienced disproportionately higher likelihood of hospitalization and death from SARS-CoV-2 infection compared to those belonging to white non-Hispanic groups.17, 18, 19 Furthermore, most expert recommendations regarding SARS-CoV-2, which often translate to hospital protocols, cite studies featuring a predominantly white patient demographic.¹¹ Prioritizing research representing patients from all backgrounds is necessary to examine inequalities and prevent further health disparities.

The investigators of the current study examined thrombotic and clinical events in SARS-CoV-2 PNA compared to other viral PNA in an urban medical center featuring a primarily Hispanic and Black non-Hispanic population, with the intention of addressing gaps in care, adding a unique perspective to the current literature. Understanding differences in outcomes could help health-care providers tailor care for these diverse patients.

Methods

This study was a single-center, non-randomized, retrospective study conducted at UIHHSS under University of Illinois Institutional Review Board-approved protocol # 2020-1942. Using the International Classification of Diagnosis-10 (ICD-10) codes, medical records for patients hospitalized from 10/01/2017 to 09/01/2020 were queried to extract clinical events related to molecular pathology results indicating infection with H1N1, H3N2, or SARS-CoV-2 viral strain. The two cohorts: SARS-CoV-2 cohort patients were those in the first wave of the SARS-CoV-2 pandemic beginning in late 2019 through September 2020, and the other viral PNA cohort was compiled from three full influenza seasons from October 2017 to September 2020. Patients with H1N1 or H3N2 were grouped together labeled as “other viral PNA” cohort for statistical analysis. Confirmation of the primary admission diagnosis was verified by examining information from admission notes, imaging studies, and laboratory results. Electronic health records were utilized to collect demographic information, pertinent laboratory values, past medical history, clinical outcomes, use of mechanical and pharmacological thromboprophylaxis, and other relevant information such as length of hospital stay, whether an ICU admission occurred, and in-hospital death. The maximum and minimum pertinent laboratory values were obtained and analyzed. The term mechanical thromboprophylaxis referred to the use of an intermittent compression device, while the term pharmacological thromboprophylaxis referred to the use of medications such as heparin 5000 or 7500 units (U) every 8–12 h or enoxaparin 40 mg every 12–24 h. Compliance with thromboprophylaxis (mechanical or pharmacologic) was charted by nurses per hospital protocol using the flowsheet function and abstracted for retrospective review ^†Mechanical prophylaxis was charted as “On” or “Off” on electronic health record several times a day; charted as “On” for every 24 h period per hospital admission coded as compliant; if charted as “Off” at any point per 24 h period coded as non-compliant.

Inclusion criteria for this study was age ≥ 18 years, positive molecular pathology results indicating infection with SARS-CoV-2, H1N1, or H3N2, and a primary diagnosis of viral PNA based on ICD-10 code. Primary diagnosis of viral PNA refers to patients whose admission to the hospital was attributed to viral PNA. The investigators of this study were trained to verify the primary diagnosis during chart review per study protocol. Patients presenting with positive virology but no findings of PNA by imaging and/or clinical characteristics, or who developed viral PNA secondary to their admission due to an unrelated primary admission, were excluded from this study.

The primary outcome was defined as a composite of the following clinical events: thrombotic events (venous thromboembolism-pulmonary embolism, deep vein thrombosis; arterial thromboembolism-myocardial infarction, ischemic stroke, or other systemic arterial thrombosis), major bleeding (defined as bleeding into a critical organ, fatal bleeding, or bleeding causing a decrease of hemoglobin level of ≥ 2 g/dL, or leading to a transfusion of two or more units of packed red blood cells), newly initiated on renal replacement therapy (RRT), use of mechanical ventilation, admission to the intensive care unit (ICU), co-infection during admission, or in-hospital death. A multivariable logistic regression analyses was performed as secondary outcome to determine the risk of mortality during hospital stay using pre-specified variables (age, SARS-CoV-2 status, gender, ethnicity, ICU admission, obesity, and anticoagulation compliance in a multivariable logistic regression analysis.

Statistical analyses

Continuous variables were compared using the Student's t test if parametric, and the Mann-Whitney U test if non-parametric. Categorical variables were evaluated with the χ2 or Fisher's exact test, as appropriate.

Results

Medical records of 257 patients hospitalized with viral PNA between 10/01/2017 and 09/01/2020 were reviewed. Of these, 199 patients were admitted with the primary diagnosis of SARS-CoV-2 PNA, and 58 patients with other viral PNA (n=25 H1N1, n=33 H3N2 viral PNA strains). The mean age of the patients in the SARS-CoV-2 PNA group and the other viral PNA group was 58 and 63 years, respectively. About half (46%) were men, with a similar proportion of Hispanic (46%) and non-Hispanic Black (44%) patients among the two cohorts. Demographics and medical histories of the two groups are shown in Table 1 .

Table 1.

Demographic information and past medical history.

Demographic	SARS-CoV-2 PNA n=199	Other Viral PNA n=58
Age (year), mean ± SD	58 ± 15	63 ± 17
Male, n (%)	95 (48)	24 (41)
Race/Ethnicity *
Hispanic, n (%)	105 (53)	14 (24)
Non-Hispanic Black, n (%)	78 (39)	35 (60)
White, n (%)	11 (6)	6 (10)
Weight (kg), mean ± SD *	92 ± 27	80 ± 25
BMI (kg/m2), mean ± SD *	33 ± 10	29 ± 9
Obesity1, n (%) *	109 (55)	21 (36)
Morbid Obesity2, n (%)	37 (19)	7 (12)
Past Medical History, n (%)
Hypertension	115 (58)	35 (60)
Diabetes	79 (40)	26 (45)
Lung Disease3	33 (17)	15 (26)
CKD on Dialysis *	19 (10)	15 (26)
Chronic Heart Failure *	13 (7)	10 (17)
Atrial Fibrillation *	11 (6)	8 (14)
Prior Stroke/TIA	13 (6)	3 (5)
Venous Thromboembolism	9 (5)	5 (9)
Myocardial Infarction *	7 (4)	6 (10)
Pertinent Laboratory Value¥
Maximum D-Dimer§ (ug/mL), n=145, 9	1.9 [4.0]	2.7 [5.9]
Minimum Fibrinogen (mg/dL), n=60, 9	655 [307]	634.5 [360.8]
Maximum INR, n=155, 40	1.2 [0.3]	1.3 [0.6]
Minimum Platelets (109/L), n= 195, 50	176 [109]	137 [99]
Maximum Serum Creatinine (mg/dL), n=195, 50	1.14 [2.0]	1.37 [1.9]
ICU Admission, n (%)	63 (32)	17 (29)
ICU LOS, days (IQR)	104,-20	42,-7
Hospital LOS, days (IQR)	64,-14	54,-9

¹Obesity: BMI >30 kg/m²; 2 Morbid obesity >40 kg/m²; 3 Lung disease: asthma, chronic obstructive pulmonary disease (COPD), emphysema, pulmonary hypertension, or home oxygen therapy.

^⁎p<0.05; Maximum/Minimum Measured Value is mean ± SD if parametric, median [IQR] if non-parametric or N (%) for categorical variables; §Normal D-dimer per institution guidelines is ≤0.50 µg/mL. Mean calculated from n of respective laboratory values per patient per admission; depicted in each row as follows: n=SARS-Cov2, Other Viral PNA; BMI, body mass index kg/m2; CKD, chronic kidney disease; TIA, transient ischemic attack, IQR, interquartile range; LOS, length of stay, SD, standard deviation; INR, International Normalized Ratio; ICU, intensive care unit.

In both cohorts, the average maximum D-dimer was ≥ 3x ULN and serum creatinine was mildly elevated. Anticoagulation, either prophylactic or treatment dose, was prescribed at similar rates for SARS-CoV-2 PNA and other viral PNA patients (95% versus 84%) with prophylactic dosing most frequently prescribed for both groups (89% versus 74%). High compliance with mechanical and pharmacologic thromboprophylaxis regimens (Fig. 1 ) was observed in both groups (91% vs. 90% mechanical; 89% versus 83% pharmacologic). Duration of hospital stay was similar among patients with SARS-CoV-2 PNA versus other viral PNA however the former experienced longer ICU stays (median 10 days versus 4 days; p=0.056). Patients with SARS-CoV-2 were also admitted to the ICU slightly more frequently.

Fig. 1.

Compliance Rate (%) with Mechanical and Pharmacological Thromboprophylaxis for SARS-CoV-2 PNA versus Other Viral PNA

Compliance rate (%) in both mechanical (91% versus 90%) and pharmacological (89% versus 83%) were similar in the SARS-CoV-2 PNA and other viral PNA cohort respectively.

There was no difference in the incidence of the composite outcome between the two groups (Table 2 ). Incidence of composite clinical events was relatively similar among both groups (p=0.338). A significantly higher number of patients in the SARS-CoV-2 PNA group were initiated on RRT (9% vs 0%; p = 0.0210) or died (16% versus 3%; P= 0.0130) compared to the other viral PNA group. The SARS-CoV-2 group also required longer durations of mechanical ventilation (median 12 versus 5 days). Seven (3%) thrombotic events were observed in SARS-CoV-2 patients in the ICU, of these 3 were pulmonary embolisms (PE), 1 distal lower extremity deep vein thrombosis (DVT), and 3 non-ST elevated myocardial infarctions (NSTEMI).

Table 2.

Frequency of thrombotic and clinical events during hospitalization.

	SARS-CoV-2 PNA n=199	Other Viral PNA n=58	P-value
Composite Outcome *, n (%)	82 (41)	28 (48)	0.338
Venous Thrombosis	4 (2)	0	0.276
Arterial Thrombosis	3 (2)	0	0.347
Major Bleeding	3 (2)	0	0.347
RRT	17 (9)	0	0.021
Mechanical Ventilator	41 (21)	6 (10)	0.073
Co-Infection◊	67 (34)	26 (45)	0.120
Death	31 (16)	2 (3)	0.013

^⁎Composite clinical outcome refers to total number of thrombotic (arterial, venous) or clinical (major bleeding, mechanical ventilator, co-infection, RRT, or death) events per patient per hospital admission ◊ Co-infection was defined as having a concomitant bacterial pneumonia verified by imaging and diagnostic testing.

Multivariable logistic regression analysis showed that age (aOR 1.07, 95%CI 1.03–1.11), a diagnosis of SARS-CoV-2 pneumonia (aOR 13.01, 95%CI 2.17–78.5), and ICU admission (aOR 44.41, 95%CI 11.9–165.24) were significantly associated with risk of mortality during the hospital stay. Race and ethnicity were not associated with mortality.

Discussion

The data suggests that infection with SARS-CoV-2 pneumonia causes a stronger systemic and inflammatory insult to the body than the seasonal flu.²² A retrospective study comparing acute respiratory distress syndrome (ARDS) in SARS-CoV-2 patients compared to ARDS in influenza patients observed patients infected with SARS-CoV-2 had a significantly longer duration on mechanical ventilation compared to other viral infections (P=0.06) and increased frequency of renal replacement therapy in SARS-CoV-2 patients; however, the latter was not statistically significant (P=0.99).²¹ In our study, a similar observation was noted. SARS-CoV-2 PNA group required longer mechanical ventilator support (median 12 vs 5 days) and required higher rate of renal replacement therapy (9% versus 0, P=0.021) than other viral PNA group. The results observed in the current study also demonstrated significantly higher rate of in-hospital mortality (16% versus 3%, P=0.013) in SARS-CoV-2 patients than in other viral PNA patients. While not statistically significant, use of mechanical ventilation occurred more frequently in the SARS-CoV-2 cohort (21% versus 10%, P=0.073). Implications of these results suggest that inflammation caused by SARS-CoV-2 virus compared to other viruses may negatively impact organ systems, evidenced by the more frequent occurrence of mechanical ventilation, longer days requiring mechanical ventilator support, renal replacement therapy, and death. Lastly, multivariable regression analysis yielded that the presence of SARS-CoV-2 pneumonia diagnosis, age, and intensive care unit admission were considered significant risk factors for mortality. Knowledge of these outcomes can inform clinicians to improve health-care utilization during times of increased transmission.

The investigators in the current study observed more frequent thrombotic events in the SARS-CoV-2 PNA cohort versus the other-viral PNA cohort (7 versus 0) consisting of 3 arterial thrombotic events and 4 venous thrombotic events. Although not statistically significant, our results are consistent with prior studies reporting high occurrence of thrombotic events in SARS-CoV-2 patients.^{4 , 5 , 16} It has previously been reported in a retrospective cohort study that thrombotic complications occurred more frequently in SARS-CoV-2 versus other influenza patients (25% versus 11%), mainly driven by high incidence of VTE in the SARS-COV-2 cohort (23% versus 3.6%).²² Interestingly, the prior study reported more frequent arterial complications in the influenza cohort (7.5% versus 4.4%), while the current study did not observe any thrombotic complications in the other viral PNA cohort. Previous literature on viral mechanisms and thrombotic complications have reported a higher incidence of arterial events secondary to viral mechanisms in patients with cardiovascular comorbidities.²⁰ Our study had an overall equal distribution of cardiovascular comorbidities between both cohorts, and differences in data from that study may be explained by potentially inconsistent diagnostic testing, screening methods, and different hospital practice guidelines compared to the current study.²² Importantly, the previous study did not report anticoagulation use or specific influenza viral strain while our study observed those admitted to UIHHS with SARS-CoV-2 or other viral PNA strains (specifically H1N1 and H3N2) with the majority receiving prophylactic or treatment dose pharmacologic thromboprophylaxis under compliance with strict hospital protocols.²³ Standards for thromboprophylaxis use may also differ which may explain the overall low number of thrombotic complications observed in the current study.

The body's inflammatory response to certain triggers, such as infection, alerts the body to a potential attack. This response is also intertwined with driving platelet activation, increasing fibrin levels, and generation of thrombin, therefore increasing risk of thrombotic events.²⁰ On a cellular level, inflammation perpetuates a positive feedback loop of further leukocyte, cytokine, and inflammatory mediator release. Furthermore, this complex cascade is implicated in causing a broad effect on other organ systems and leads to more clinical events observed in viral pneumonia. Importantly, understanding the patient's clinical picture is dependent on realizing the pathogen-specific impact on patients.^{4 , 5 , 9 , 20}

Racial disparities owing to poorer outcomes in SARS-CoV-2 been previously reported; however, race and ethnicity of patients in the current study was not associated with risk of mortality.^{2 , 17 , 18} It is important to highlight the unique demographic of this study which includes a majority of Hispanic and non-Hispanic Black patients in both patient groups studied. Throughout the SARS-CoV-2 pandemic, underserved communities and minority ethnic groups have been disproportionately affected and reports suggest that Black patients are at greater risk for mortality.^{17 , 22} However, a large, multi-center study, featuring 92 hospitals in the United States, showed no difference in risk of mortality between Black and White patients for those able to access hospital care.¹⁸ The current study may be viewed as an ample area to access hospital care, located in a metropolitan area of Chicago featuring a mid-sized hospital with >20 outpatient clinics, and federally qualified health centers (FQHC).²⁴ The differences observed in our study and that of studies reporting worse outcomes for Hispanic and non-Hispanic black patients may be attributed to our area.^{2 , 17 , 22} Outcomes for SARS-CoV-2 patients may be dependent on geographic location that has more access to resources and may not be dependent on race. Further studies should address non-medical social determinants of health outcomes.²⁵

Limitations of current study include a small sample size, different waves of the COVID-19 pandemic, and challenges diagnoses of thrombotic complications. First, rates of thrombotic complications may have been underreported in our health-system due to the lack of resources that were available at the beginning of the pandemic. Given that the patients in our data set came from the first wave of the pandemic, extra precautions had to be taken to reduce the risk to both providers and patients, and shortages of staff and various medical equipment (e.g., personal protective equipment) were widespread. In addition, early reports suggested that patients were experiencing higher rates of thrombotic complications during their hospitalization which may have lowered our sensitivity to screening as compared to other viral pneumonia patients seen at the health-system.^{4 , 11 , 16} Due to these limitations, the incidence of thrombotic complications may have been underreported in the SARS-CoV-2 PNA group. Another limitation of this study is the study period from March 2020 to September 2020; since then, there has been at least two more waves of SARS-CoV-2 and immense vaccinations efforts.²⁶ Lastly, another limitation was potential for inconsistent interpretation of race and ethnicity between patient and healthcare facility. The results in our current study may have limited external validity due to the constantly evolving virus and categorization of race and ethnicity.

Given the continuous nature of this pandemic, future studies are needed to validate our findings and whether there's any outcome differences in the evolving virus causing SARS-CoV-2 infection.

Conclusion

Overall, the current study found a similar incidence of composite clinical events between patients with SARS-CoV-2 PNA and those with H1N1/H3N2 PNA. The incidence of venous or arterial thrombotic events was low and occurred only in the SARS-CoV-2 PNA group. Low rates of thrombotic complications may be attributed to robust, hospital-wide implementation of anticoagulation protocols beginning shortly after the pandemic commenced in March of 2020, as demonstrated by the high rates of prophylaxis therapy prescribed and the high rates of compliance. However, it may also be attributed to a lack of staff and the resources required to confirm the diagnosis of a thrombotic event.

SARS-CoV-2 PNA was associated with a higher rate of kidney failure leading to RRT and mortality compared to other viral PNA. The study did not observe a higher association of mortality based on race/ethnicity.

Financial disclosures

None.

Declaration of Competing Interest

Scott Benken (CareDx, Eurofins, VericiDx, Daichii Sankyo, BMS, Velovis); none relevant to this publication; all authors have no conflicts of interest to disclose.