Endothelial Injury
Autopsy studies early during the pandemic revealed diffuse endothelial inflammation in many organs, including the lung, heart, liver, and kidney, with evidence of direct viral infection of endothelial cells by the SARS-CoV-2 virus. Because in vivo biosynthesis of von Willebrand factor is restricted to endothelial cells and megakaryocytes, high plasma von Willebrand factor concentrations in patients with COVID-19 suggest significant endothelial cell derangement. Immune cell arteritis was found in the lungs of nearly half of those who had died of COVID-19 in one autopsy series. Endothelial injury, particularly in the context of a hypercoagulable milieu, likely is responsible for the high rates of microthrombosis noted in the pulmonary vasculature. Although pulmonary microthrombosis has been noted previously in classical ARDS, the extent evident in COVID-19 is significantly greater. Ackerman et al noted that autopsies from patients with COVID-19 showed nine times more alveolar capillary microthrombi compared with autopsies from patients with ARDS secondary to H1N1 influenza. The high rates of microthrombosis are not limited to the lungs; they also have been reported in the heart53 and skin. Imaging studies demonstrate that thrombotic lesions in COVID-19 are smaller and more peripherally located compared with those in non-COVID acute PE, suggesting that some filling defects on CTPA, particularly isolated subsegmental PE, may reflect in situ pulmonary thrombosis instead of the typical embolization of thrombi originating from peripheral DVT. Mirsadraee et al reported that of the critically ill patients with COVID-19 found to have pulmonary thrombosis via screening CTPA, 77% did not have radiologic evidence of peripheral DVT.
Gas Exchange vs Lung Compliance
Early in the pandemic, Gattinoni et al noted that although many patients with COVID-19 technically fulfilled the Berlin criteria for ARDS, many showed marked hypoxemia and elevated shunt fraction with only minimally affected lung compliance, particularly early in the course of disease. Chiumello et al noted that venous admixture was unrelated to the fraction of nonaerated lung tissue in COVID-19 ARDS, yet was correlated to the fraction of nonaerated lung tissue in a historical cohort of patients with non-COVID ARDS who were matched for both Pao 2 to Fio 2 ratio and compliance. Additionally, many nonintubated patients with COVID-19 demonstrate dramatic hypoxemia, yet lack proportional signs of respiratory distress, a condition coined happy hypoxemia. Some have hypothesized that the presence of pulmonary thrombi, both microthrombi and macrothrombi, may help to explain the disconnect between gas exchange and lung compliance in severe COVID-19.
Hemodynamic Perturbations
Pulmonary emboli increase pulmonary vascular resistance (PVR) and pulmonary artery pressure, with higher thrombotic burden correlating with higher PVR and pulmonary artery pressure. With sufficiently elevated right ventricular (RV) afterload, pulmonary emboli can induce RV dilation and dysfunction. Somewhat strikingly, patients with COVID-19 requiring mechanical ventilation show invasive hemodynamic profiles that are characterized by low, not high, PVR. This finding is surprising given the high prevalence of PE in patients with COVID-19 in the ICU, as well as the high prevalence of elevated PVR in non-COVID ARDS.
It is possible that the hemodynamic effect of pulmonary thrombosis is mitigated by a primary pulmonary vasodilatory process in some patients with COVID-19. Dual-energy CT imaging has demonstrated pulmonary vessel dilation in COVID-19 pneumonia. Ackerman et al, in addition to demonstrating high rates of pulmonary microthrombosis in COVID-19, also noted high rates of intussusceptive and sprouting angiogenesis. Reynolds et al reported that 83% of mechanically ventilated patients with COVID-19 showed positive bubble study findings as assessed by contrast-enhanced transcranial Doppler imaging, likely indicative of abnormal pulmonary capillary dilation, pulmonary arteriovenous malformations, or both. Additionally, the degree of transpulmonary bubble transit correlated with Pao 2 to Fio 2 ratio, suggesting that these pulmonary vascular dilations may be a significant cause of hypoxemia in COVID-19 ARDS. Whereas pulmonary macrothrombi and microthrombi increase PVR, pulmonary vasodilation decreases PVR; when both processes occur simultaneously, each can "cancel out" the hemodynamic effect of the other (Fig 2 ). The coexistence of both obliterative and vasodilatory processes in the pulmonary vasculature is reminiscent of what can occur in chronic liver disease, specifically portopulmonary hypertension (obliterative) and hepatopulmonary syndrome (vasodilatory). Ultimately, in COVID-19 ARDS, the obliterative processes may dominate, leading to severe RV failure and cardiogenic shock.
Although the vasodilatory and obliterative processes may offset each other hemodynamically, their coexistence may amplify hypoxemia in COVID-19. Vasodilated regions experience increased blood flow, creating low V˙/Q˙ ratios. Microthrombi and vasoconstriction in other areas of the lung reroute additional blood flow to the vasodilated regions and drive down the V˙/Q˙ ratio further. Mathematical modeling demonstrates that the large amount of pulmonary venous admixture in the setting of relatively minor parenchymal involvement observed in early COVID-19 can be explained reasonably by a combination of pulmonary thrombosis and vasodilation.
Diagnosis
In the absence of systematic screening, the diagnosis of PE begins with clinical suspicion. Unexplained dyspnea and hypoxemia, particularly in the setting of normal chest radiography findings, raises the clinical suspicion for PE. Clinical suspicion of PE in a patient with COVID-19 pneumonia often is diminished because the signs and symptoms of COVID-19 pneumonia mimic those of PE; a patient's dyspnea and hypoxemia may be attributed solely to COVID-19 pneumonia, and further diagnostic testing for potential PE may be deferred. Although clinical probability scores, such as the Wells score, are helpful in raising clinical suspicion of PE in patients, they have not been validated in patients with COVID-19 and likely underestimate the probability of PE in COVID-19.
D-dimer, in conjunction with clinical probability assessment, has great usefulness in ruling out PE in patients with low or intermediate probability of PE, although its usefulness in COVID-19 is unclear. Although D-dimer levels in COVID-19 correlate with rates of thrombosis, it is not clear whether a particular D-dimer value "rules in" or "rules out" PE. In the study by Mirsadraee et al in which screening CTPA was performed for patients with COVID-19 on admission to the ICU, D-dimer levels did not discriminate between patients with and without PE. Li et al developed a three-factor score consisting of admission fibrinogen, admission D-dimer, and D-dimer increment > 1.5 fold, the score performing with a sensitivity of 0.93 and specificity of 0.71 for symptomatic VTE.
CTPA is the first-choice method for the diagnosis of PE because of its high accuracy, wide availability, and ability to assess for other pulmonary pathologic features. Its use may be limited in critically ill patients with COVID-19 who are not stable enough for transfer and in patients with renal failure, a common complication in severe COVID-19. V˙/Q˙ scanning can be used for patients in whom CTPA is contraindicated or inconclusive. Scans performed on patients with abnormal chest radiography findings, as is often the case in patients with COVID-19 pneumonia, are more likely to result in false-positive results because the images rarely appear as normal or showing a low probability of PE in such patients. Compression ultrasonography can be performed to assess for DVT when chest imaging is contraindicated or indeterminate; however, the absence of DVT does not imply the absence of pulmonary thrombosis, especially because in situ pulmonary thrombosis is a potential mechanism in COVID-19. A diagnosis of DVT may eliminate the need to evaluate for PE because the indication for therapeutic anticoagulation will have been established. Echocardiography can raise suspicion for the diagnosis of PE with the presence of clot in the right side of the heart or new right heart strain. Although it has limited diagnostic value, echocardiography is most useful for risk stratification of confirmed PE.
Prophylaxis
Considering that PE is one of the most common preventable causes of hospital death, thromboprophylaxis is a crucial component in the care of hospitalized patients, and COVID-19 is no exception. A retrospective study by Rentsch et al demonstrated that the administration of prophylactic anticoagulation within 24 h of admission in patients with COVID-19 was associated with decreased mortality when compared with no prophylactic anticoagulation. Multiple society guidelines recommend prophylactic anticoagulation for hospitalized patients with COVID-19 who do not have a contraindication to treatment. However, standard doses of prophylactic anticoagulation likely are insufficient for the prevention of VTE in patients with COVID-19; many VTE events are diagnosed within the first 24 h after admission, and the reported rates of VTE are notably high despite the use of anticoagulant thromboprophylaxis.
As a result of the high rates of VTE despite standard-dose thromboprophylaxis, many institutions have implemented protocols using higher doses, including an intermediate dose and even a therapeutic dose. Tacquard et al reported in a study of 538 patients with COVID-19 in eight French ICUs that high-dose prophylactic anticoagulation (intermediate or therapeutic dose) was associated with a significant reduction in thrombotic complications (hazard ratio, 0.81) without an increase in bleeding risk. In more than 4,000 patients with COVID-19 at a New York hospital with an aggressive anticoagulation protocol, Nadkarni et al noted a trend toward a mortality reduction with therapeutic anticoagulation compared with prophylactic anticoagulation, a finding that did not meet statistical significance (P = .08).
However, recent randomized controlled trials do not support the use of higher than standard doses for prophylactic anticoagulation in critically ill patients with COVID-19. In 600 critically ill patients, intermediate-dose anticoagulation with enoxaparin 1 mg/kg daily was not superior to standard prophylactic anticoagulation with enoxaparin 40 mg daily in reducing the composite outcome of venous or arterial thrombosis, treatment with extracorporeal membrane oxygenation, or mortality within 30 days. Although bleeding events were rare, major and clinically relevant nonmajor bleeding events were nonsignificantly more frequent with intermediate-dose anticoagulation (6.2% for intermediate dose, 3.1% for standard dose; P = .08). A large National Institutes of Health multiplatform randomized controlled trial incorporating three global networks (A Randomised, Embedded, Multi-factorial, Adaptive Platform Trial for Community-Acquired Pneumonia [REMAP-CAP], Anti-Thrombotic Therapy to Ameliorate Complications of COVID-19 [ATTACC], and Accelerating COVID-19 Therapeutic Interventions and Vaccines-4A [ACTIV-4A]) examining the benefit of therapeutic dose vs standard dose prophylactic anticoagulation in more than 1,000 critically ill patients with COVID-19 discontinued enrollment because statistical criteria for futility were met. Importantly, despite therapeutic anticoagulation decreasing major thrombotic events, an 89% probability was found that therapeutic anticoagulation was inferior to standard dose prophylactic anticoagulation in achieving the primary outcome of survival or days free of organ support. The mechanism for this likely harm is unclear, given that major bleeding was increased only mildly with therapeutic anticoagulation (3.1% vs 2.4%). These findings suggest that initiating therapeutic anticoagulation after severe COVID-19 has developed may be too late to alter the clinical course beneficially. In contrast, for moderately ill hospitalized patients with COVID-19, the National Institutes of Health recently announced via press release that therapeutic-dose anticoagulation was superior to standard-dose prophylactic anticoagulation in reducing the need for organ support and mortality. The full results and official publication of these studies are awaited anxiously because they undoubtedly will help to establish the optimal anticoagulation dosing strategies for patients with COVID-19.
Treatment
Anticoagulation is the mainstay of the treatment for acute PE, both for patients with and without COVID-19, to prevent further thrombosis and thromboembolism. Initial treatment options for anticoagulation include unfractionated heparin, low-molecular-weight heparin, fondaparinux, and, in low-risk patients, direct oral anticoagulants. As is the case with non-COVID-19 PE, risk stratification is the central tool used to identify patients at increased risk of early death who may benefit from reperfusion therapy (ie, thrombolysis or embolectomy), mechanical circulatory support, or both. Per the European Society of Cardiology guidelines, high-risk PE is characterized by cardiac arrest, systolic BP < 90 mm Hg, or requiring vasopressors, inotropes, or both. Intermediate-risk PE is characterized by normotension with signs of RV dysfunction on echocardiography or CTPA, elevated troponin levels, or an elevated PE severity index score. Of note, the role of troponin levels as a prognostic biomarker for PE in COVID-19 is confounded by the fact that troponin frequently is elevated in patients with COVID-1982; in that context, it likely reflects myocardial inflammation, cardiac microthrombosis, or both, rather than RV pressure overload. Low-risk PE is characterized by normotension, lack of RV dysfunction, and a low PE severity index score. Figure 3 summarizes the European Society of Cardiology risk stratification algorithm and treatment strategy.
If possible, patients with high-risk PE should undergo reperfusion therapy, mechanical circulatory support, or both. Patients with intermediate-risk PE should be monitored closely for signs of clinical deterioration, with select patients proceeding to reperfusion therapy.