Introduction

Covid-19 is a serious pandemic characterized by severe acute respiratory disease needing mechanical ventilation and intensive care unit (ICU) treatment.

Among the factors predisposing to poor survival, thrombotic complications have been suggested to have an important role. Accordingly, clinical studies showed a high prevalence of venous thromboembolism in the Covid-19 patient population with a variable incidence, ranging from 7 to 31% [1]; a small series of autoptic studies demonstrated a relationship between thrombosis in the circulation lung and poor survival [2]. Conversely, the incidence of arterial thrombosis has been studied in small series and appears to be much lower than venous thromboembolism; in a recent review arterial thrombosis was reported to range from 2.8 to 3.8% [3]. While there are consistent data on the association between hypercoagulability and mortality [4], the impact of venous and arterial thrombosis on survival is less clear. Thus, our aim was to evaluate the occurrence of thrombotic events in the artery and venous circulation, and its relationship with mortality in patients recruited by 5 Italian hospital centers dedicated to the management of severe acute respiratory syndrome coronavirus 2019 (SARS-Cov-2).

Methods

Study design and population

This is a multicentre observational cohort study performed at University hospitals located in Rome (2 centres), Latina, Perugia and Chieti (all in Italy).

We included consecutive adult patients (age ≥ 18 years) with confirmed Covid-19-related pneumonia, with or without mechanical ventilation, hospitalized from 1 to 31st March 2020 and admitted to infectious disease wards or ICU. Covid-19 was diagnosed on the basis of the WHO interim guidance [5]. A confirmed case was defined as a person with laboratory confirmation of Covid-19 infection, irrespective of clinical signs and symptoms. Oropharyngeal and nasopharyngeal swabs for laboratory diagnosis of Covid-19 were performed in duplicate: SARS- Cov2 E and S gene were detected by a reverse transcriptase-polymerase chain reaction (RT-PCR).

High-resolution computed tomography (CT) scan was used to identify lung involvement according to the Official diagnosis and treatment protocol (6th edition) declared by the National Health Commission of China. Typical CT findings of SARS-CoV2 related pneumonia were considered: consolidation, ground-glass opacities, crazy paving and/or reticular pattern [6]. Radiologic abnormalities were reviewed by attending physicians in respiratory medicine who extracted the data. Major disagreement between two reviewers was resolved by consultation with a third reviewer. Ethical approval for this study was obtained from the Ethics Committee of “Azienda Ospedaliero Universitaria Policlinico Umberto I” (approval number/ID Prot. 109/2020).

Baseline assessment

Demographic, clinical, laboratory and radiological results were extracted from electronic medical records of patients enrolled. Cancer history was reported on the basis of the history-taking or from medical records. Laboratory assessments consisted of routine blood and chemical analysis included coagulation tests (D-dimer, prothrombin time (PT) and activated partial thromboplastin time (aPTT) and fibrinogen), high-sensitivity C-reactive protein (hs-CRP), procalcitonin and High-Sensitivity Cardiac Troponin T (hs-cTnT). Prevalence of diabetes mellitus, hypertension, coronary artery disease (CAD), chronic kidney disease, obesity were recorded, defined as previously described [7]. Assessment of in-hospital thrombotic events.

Patients were followed-up until discharge or in-hospital death. We registered the occurrence of thrombotic events including ischemic/embolic events that were categorized as follows: pulmonary embolism (PE), detected by lung CT scan [8]; deep vein thrombosis (DVT), assessed by ultrasonography or by CT scan; superficial vein thromboses (SVT), assessed by ultrasonography in symptomatic patients; acute myocardial infarction, diagnosed on the basis of the diagnostic electrocardiographic findings associated with elevation of serum markers of myocardial necrosis [9]; acute ischemic strokes, identified by observing the onset of new focal neurological signs and symptoms and confirmed with MRI or CT imaging [10]; acute limb ischemia diagnosed according to guidelines [11]. Inferior leg ultrasonography was performed in patients with signs and/or symptoms for inferior leg DVT and routinely performed in patients with PE. Uncommon vein thrombosis in symptomatic patients (i.e. gonadic vein thrombosis) was detected by CT-scan.

Statistical analysis

Categorical variables are reported as counts and percentages and continuous variables as mean ± standard deviation (SD), or medians and interquartile ranges (IQRs). Differences between percentages were assessed by chi-square or Fisher exact tests. All continuous variables were tested for normality with the Shapiro–Wilk test. Student unpaired t-tests were used for normally distributed continuous variables. Appropriate nonparametric tests (Mann–Whitney and Spearman rank correlation tests) were used for the other variables. The bivariate and multivariate effects of prognostic factors on thrombotic events were assessed by means of logistic regression models. Wald confidence intervals and tests for odds ratios (OR) and adjusted OR were computed on the basis of the estimated standard errors. Survival curves were estimated using the Kaplan–Meier product-limit estimator and compared using the log-rank test. Cox proportional hazards analysis was used to calculate the adjusted hazard ratios (HR) and 95% confidence interval (CI) for each clinical variable. Only p values < 0.05 were considered statistically significant. All tests were 2-tailed, and analyses were performed using computer software packages (IBM SPSS Statistics 25).

Results

Thrombotic events in Covid-19

Three-hundred seventy-three patients were hospitalized for a median length of stay of 18 days (IQR: 12–27), 306 (82%) have been admitted to infectious disease wards and 67 (18%) in ICU. Baseline clinical and laboratory characteristics, as well as in-hospital treatments at admission, are reported in Table 1. During the in-hospital stay, 75 (20%) patients experienced thrombotic events: 41 (54%) were in the venous circulation (7 superficial venous thrombophlebitis, 1 gonadic venous thrombosis, 15 deep venous thrombosis, 18 pulmonary embolism) and 34 (46%) were in the artery circulation (9 acute limb ischemia, 15 acute myocardial ischemia and 10 TIA/strokes).

Table 1 Clinical and laboratory characteristics of study patients, according to thrombotic events
Full size table

Compared to patients who did not experience an in-hospital thrombotic event, patients with thrombotic events were older (71.5 ± 15.5 vs. 66.3 ± 17.0; p = 0.012), were more likely to have a history of CAD (25% vs. 13%; p = 0.015), and to have been admitted to ICU (41% vs. 12% p < 0.001). Moreover, PaO2/FiO2 ratio at hospital admission were lower in patients who developed thrombotic events. Among the laboratory variables, baseline values of white blood cells, serum D-dimer, high sensitivity cardiac troponin T (hs-cTnT), were higher, and albumin lower, in patients who developed thrombotic events (Table 1).

In a multivariable regression logistic model, ICU admission (OR. 6.0; 95% CI 2.8–12.6; p < 0.001), CAD (OR: 2.4; 95% CI 1.4–5.0; p = 0.022), and low albumin levels (OR: 0.49; 95% CI 0.28–0.87; p = 0.014) were independently associated with thrombotic events.

When separately analyzed, no significant differences in comorbidities were found between patients who experienced in-hospital arterial or venous thrombotic events. However, patients with arterial events showed to be older and higher baseline hs-cTnT, while patients with venous event showed lower PaO2/FiO2 (Table 2).

Table 2 Clinical and laboratory characteristics of study patients, according to arterial of venous thrombotic events
Full size table

Mortality in Covid-19

During the in-hospital follow-up, 75 (20%) patients died. Clinical and laboratory characteristics of patients according to mortality are represented in Table 3. Patients who died were older, more likely male and with a greater prevalence of COPD, diabetes, heart failure, coronary artery disease, ICU admission and intra-hospital thrombotic events than patients who survived. Among the laboratory variables, baseline D-dimer, hs-CRP, hs-cTnT, and albumin levels were associated with mortality. Kaplan-Meier analysis showed that patients with thrombotic events had a higher probability of in-hospital mortality compared to thrombotic events-free patients (44% vs. 14%; p < 0.001) (Fig. 1). A multivariable COX regression analysis showed that in-hospital thrombotic events (HR: 2.481; 95% CI 1.336–4.609; p = 0.004), age (HR: 1.048; 95% CI 1.022–1.075; p < 0.001), baseline low albumin levels (HR: 0.447; 95% CI 0.277–0.723; p = 0.001) and low PaO2/FiO2 ratio (HR: 0.996; 95% CI0.993–0.999; p = 0.022) predicted mortality, after adjusting for sex, ICU admission, diabetes, CAD, heart failure and COPD.

Table 3 Clinical and laboratory characteristics of study patients, according to mortality
Full size table
Fig. 1
figure 1

Kaplan–Meier estimates of time to mortality in patients with or without thrombotic events during the in-hospital stay

Full size image

Discussion

The study shows that patients hospitalized for Covid-19 experience both venous and arterial ischemic events, that are a warning sign of poor survival.

Compared to previous reports showing that Covid-19 is prevalently associated to venous thromboembolism [3], the present study shows, conversely, that SARS-Cov-2 is complicated by an equipollent rate of venous and arterial thrombosis; thus among 75 patients with ischemic events 34 (46%) experienced arterial thrombosis, which occurred in the coronary, cerebral and peripheral circulation. The incidence of thrombosis was even higher than that reported in community-acquired pneumonia, which is essentially associated with arterial thrombosis [12,13,14], and was detectable infrequent sites of thrombosis such as, for example, superficial or gonadic venous thrombosis or acute leg ischemia.

The mechanisms accounting for thrombosis -related vascular complications may be dependent on the severe inflammatory response which accompanies the disease; this may activate platelets and clotting system via oxidative stress or over-production of cytokines with pro-thrombotic properties [15,16,17,18]; in accordance with this, we found an independent association between elevated D-dimer and thrombotic complications. Hypoalbuminemia may be another precipitating factor as it is associated with an increased risk of arterial and venous thrombosis [19, 20]. The present study supports and extends the results of a previous study from our group including a small group of Covid-19 population and reporting a close association between hypoalbuminemia and thrombosis [21]. Hypoalbuminemia is a feature of acute and chronic inflammation, as depicted by the inverse association with hs-CRP, and might favor thrombosis at the site of vascular lesion as albumin encompasses anticoagulant [19] and antiplatelet properties [22, 23]. However, we cannot exclude that other mechanisms such as increased albuminuria may also be implicated [19]; conversely, hypoalbuminemia as a reflection of concomitant acute liver failure can be reasonably excluded by the present and previous studies [4, 24].

The present study reinforces and extends the results of an autoptic study indicating that thrombosis predisposes to poor survival; thus, on multivariable analysis, intra-hospital venous and arterial thrombotic events along with age and albumin were factors associated with death. The novelty of the study is, as depicted by Kaplan-Meyer survival analysis, in the divergence of the curves within the first 10 days from hospitalization, suggesting thrombosis as an early warning sign of poor outcome.

The study has implications and limitations. The fact that a large number of Covid-19 patients experience arterial thrombosis casts doubts over the use of anticoagulants alone to prevent thrombosis sequelae. The mechanism accounting for thrombosis in Covid-19 has not been clarified yet. Hypoalbuminemia may be a factor favoring thrombosis but an interventional trial with albumin supplementation is needed to establish a cause-effect relationship. Furthermore, we have no information regarding the quo ante nutritional status, that may be an important element predisposing to low albumin levels. Finally, venous and arterial thrombosis may not share the same pathogenetic mechanisms, they could reflect different evolving conditions during hospitalization and could be associated with a different risk of poor prognosis. For example, acute limb or cerebral ischemic events could be related to a underlying new-onset atrial fibrillation, that is a common during pneumonia [25, 26]. Further studies will be necessary to understand the potential different implication of the different kinds of vascular complication and of their different timing in the disease course.

In conclusion, the present study provides further insights into the clinical picture of Covid-19, which is characterized by an equipollent incidence of venous and arterial thrombosis. Patients with thrombosis were affected by more severe disease and showed laboratory features of systemic inflammation and hypercoagulability. Such ischemic episodes are an ominous sign of poor survival, thereby early identification and appropriate therapy of at risk Covid-19 patients could manage the thrombotic risk and ameliorate mortality.