Incidence, Predictors, and Outcomes of Thrombotic Events in Hospitalized Patients With Viral Pneumonia

Viral pneumonia represents a major cause of morbidity and mortality, especially among older patients.¹ Viral pneumonia may be associated with a systemic inflammatory response, which has been linked with arterial and venous thrombotic cardiovascular events.² During the recent COVID-19 pandemic due to the novel SARS-CoV-2 virus, anecdotal and published experience suggests high rates of thrombotic complications among hospitalized patients.³ Due to a paucity of studies describing the incidence, predictors, and outcomes of thrombotic events among patients with viral pneumonia, the observation of thrombosis in patients with COVID-19 lacks context. Hence, we aimed to address this knowledge gap using a large nationally representative dataset.

The National Inpatient Sample (NIS) database years 2005 to 2015 was used to identify hospitalizations with primary diagnosis of viral pneumonia (influenza or noninfluenza) using International Classification of Diseases, Ninth Edition, Clinical Modification (ICD-9-CM) codes 480.x, 487.0, 487.1, 488.x1, 488.x2, 484.1, 052.1, or 055.1. Thrombotic events were defined as acute myocardial infarction (AMI), acute ischemic stroke (AIS), venous thromboembolism (VTE), or acute limb ischemia (ALI). Outcomes included in-hospital mortality, hospital costs, and length of stay (LOS) and discharge disposition (among survivors). Hospitalizations with missing data for mortality or LOS were excluded. Complex samples multivariable regression models that included patient demographics, comorbidities, and in-hospital complications (bacterial/fungal superinfection, sepsis, septic shock, acute respiratory failure, and mechanical ventilation), were used to identify factors independently associated with thrombotic events among admissions with viral pneumonia. Statistical analyses were performed using IBM SPSS Statistics, Version 20 (Armonk, New York).

The study included 455,629 hospitalizations with a primary diagnosis of viral pneumonia (83.6% with influenza pneumonia). Among these, 13,465 (3.0%) had thrombotic events (AMI 5,843 [1.3%], VTE 4,739 [1.0%], AIS 3,153 [0.7%], and ALI 217 [0.05%]). Those with thrombotic events were older, less likely to be women, and had a higher prevalence of hypertension, hyperlipidemia, diabetes mellitus, prior MI, peripheral vascular disease (PVD), and chronic kidney disease (CKD). On multivariable analysis, a number of key variables were independently associated with thrombotic complications, with mechanical ventilation posing the highest risk (OR 3.96, 95% CI 3.38 to 4.64) (Figure 1).

Figure 1.

Forest plot for factors independently associated with thrombotic events on multivariable regression analysis. CABG = coronary artery bypass surgery; CAD = coronary artery disease.

In-hospital mortality was significantly higher among those with versus without thrombotic events (12.8% vs 1.8%; unadjusted OR 7.90, 95% CI 6.99 to 8.93; adjusted OR 1.68, 95% CI 1.37 to 2.05; p <0.001). Compared with those who did not develop thrombotic events, those who developed thrombotic events had longer LOS (10.1 vs 4.5 days; parameter estimate 5.67, 95% CI 5.19 to 6.16; adjusted parameter estimate 2.94, 95% CI 2.56 to 3.32; p <0.001), higher total hospital costs ($30,782 vs $9,641; unadjusted parameter estimate 21,141, 95% CI 19,217 to 23,065; adjusted parameter estimate 9,322, 95% CI 7,880 to 10,764; p <0.001), and were more likely to be discharged to skilled nursing facility (41.2% vs 16.7%; unadjusted OR 3.51, 95% CI 3.22 to 3.81; adjusted OR 2.11, 95% CI 1.89 to 2.35; p <0.001).

In this nationwide observational study of >455,000 hospitalizations for viral pneumonia, we found that thrombotic events were not uncommon (∼3.0%). Several baseline characteristics and comorbidities were independently associated with thrombotic events. The development of thrombotic events among these hospitalized patients was associated with increased in-hospital mortality, longer LOS, increased utilization of skilled nursing care following discharge, and higher costs. Prior studies have focused on AMI and were limited to influenza infection only.⁴ The current study extends our knowledge by examining the prevalence of any thrombotic event and included any etiology of viral pneumonia. These findings are relevant to the current COVID-19 pandemic caused by SARS-CoV-2, which has been linked to an increased risk of arterial and venous thrombotic events potentially related to direct viral infection or the profound inflammatory response.³ This analysis is limited by the observational nature of the study and potential residual confounding. The unit of analysis is the number of hospitalizations rather than individual patients, and the NIS lacks data on therapies or investigations performed as well as outcomes beyond the index hospitalization. Despite these limitations, our study provides important insights into the incidence, risk factors, and outcomes of thrombotic events in hospitalized patients with viral pneumonia and may serve as a comparator once similar data emerge for SARS-CoV-2 pneumonia.

Disclosures

The authors have no conflicts of interest to disclose.