CT pulmonary angiography in COVID-19 pneumonia: relationship between pulmonary embolism and disease severity

Background Pulmonary embolism (PE) is one of the known sequels of COVID-19 infection. We aimed to assess the incidence of PE in patients with COVID-19 infection and to evaluate the relationship between the CT severity of the disease and the laboratory indicators. This was a retrospective study conducted on 96 patients with COVID-19 infection proved by positive PCR who underwent CT pulmonary angiography (CTPA) with a calculation of the CT severity of COVID-19 infection. Available patients’ complaint and laboratory data at the time of CTPA were correlated with PE presence and disease severity. Results Forty patients (41.7%) showed positive PE with the median time for the incidence of PE which was 12 days after onset of the disease. No significant correlation was found between the incidence of PE and the patients’ age, sex, laboratory results, and the CT severity of COVID-19. A statistically significant relation was found between the incidence of PE and the patients’ desaturation, hemoptysis, and chest pain. A highly significant correlation was found between the incidence of PE and the rising in the D-dimer level as well as the progressive CT findings when compared to the previous one. Conclusion CT progression and the rising in D-dimer level are considered the most important parameters suggesting underlying PE in patients with positive COVID-19 infection which is commonly seen during the second week of infection and alert the use of CT pulmonary angiography to exclude or confirm PE. This is may help in improving the management of COVID-19 infection.

Background COVID-19 pandemic outbreak started in Egypt in February 2020 and is still going on. The Egyptian protocol includes doing a non-contrast CT chest for patients with suspicious respiratory symptoms. However, many patients with worsened symptoms especially during the second week of disease showed symptoms of thrombotic events of pulmonary arteries, and that is why CT pulmonary angiography (CTPA) became recommended by multiple reports [1].
Association with elevated levels of D-dimer has been found in many patients with pulmonary embolism especially severely ill patients in intensive care units (ICU). Thromboprophylaxis is recommended for those patients with worsened symptoms and elevated laboratory indicators of coagulopathy [2].
Many reports emerged in the literature showing an increased incidence of a pulmonary embolism (PE) at segmental and subsegmental levels. The percentage was 30% in patients with COVID-19 pneumonia as compared to 1% in critically ill patients without COVID-19 disease [3].
Another study showed that 23% of CTPA performed at their institution were positive for PE. Those were diagnosed on average 12 days after onset of symptoms and most of them required ICU admission [4].
Most studies showed that patients with high body mass index (BMI) are more prone to embolic events. Those studies were trying to link thrombotic events to the severity of the disease, elevated D-dimer, and patient weight and associated diseases rather than a complication of hospitalization [5].
CT pulmonary angiography not only can assess the presence of pulmonary embolus but also can assess the severity of the embolus as well as heart function and strain on the right ventricle. Thus, we can predict the need for ICU admission and patient outcome [6].
This study aimed to assess the presence of PE in patients with COVID-19 infection with severe respiratory symptoms and to evaluate the relation to disease severity on CT as well as laboratory indicators as D-dimer, CRP, ferritin, and lymphocyte count. This is may help in more understanding, diagnosis, and management of COVID-19 infection.

Methods
This was a retrospective study that was approved in our institutional ethical review board with the written consent waived. The authors had access to the data with no conflict of interest to declare. The study was conducted from March 1 till July 31.

Patients
The study was conducted on 96 patients at one of the University Hospitals in Cairo, Egypt.

Inclusion criteria
All patients with COVID-19 infection proved by positive PCR who underwent CT pulmonary angiography. No age or sex predilection. Available patients' complaints and laboratory data at the time of CTPA scanning.

Exclusion criteria
We excluded patients with unavailable PCR results, as well as patients with no data about the clinical condition and unavailable laboratory results.

CT technique
Patients' preparation: Patients were required to fast for 4 h, and normal kidney function was needed using serum creatinine as a reference. Adequate hydration was instructed before and after the procedure. An 18-20-gauge cannula was inserted into the antecubital vein. CT machine: 80-slice CT machine (Prime Aquilion, Toshiba, USA) and Dual-source CT (Somatom Definition, Siemens Medical Solutions, Forchheim, Germany). The infection control parameters were applied under the guidance of the hospital infection control unit. The patients were scanned in a supine position with the arm above the head. A breath-hold was requested from the patients trying to avoid respiratory motion artifact. A region over interest was drawn on the main pulmonary artery. Bolus IV injection of nonionic contrast medium (Ultravist 370) 1.2 ml/kg was used at rate 4 ml/s using injector pump (Medrad injector) followed by 40 ml saline at rate 4 ml/s. CT parameters: The scan area extended from diaphragm to lung apex with scan time used = 10 s. The tube voltage was 140/80 KVP and tube current: 51/213 mAs. Rotation time was 0.33 s, 0.75 mm thickness, 0.7 reconstruction increment (mm), and 0.7 Pitch. Image processing and interpretation: The images were transferred to the workstation where the axial cuts and multi-planar reformation were reviewed by two radiologists experienced in chest imaging with at least 5 years of experience. They were blinded to the aim of the study and the clinical condition of the patients or the laboratory results. The following items were fulfilled: -Presence or absence of pulmonary embolism. If positive pulmonary embolism, site if unilateral or bilateral and extent if segmental, lobar, or main arterial. -CT probability of COVD-19 infection according to RSNA recommendations [7]. -CT severity score for COVID-19 infection: The lung was divided into five lobes and was visually scored on a scale of 0 to 5. Score 0 if no CT evidence of pulmonary involvement; score 1 if the pulmonary involvement was less than 5%; score 2 if 5-25%; score 3 if 26-49%; score 4 if 50-75%; and score 5 if more than 75%. The highest total CT score was 25. -Presence or absence of CT progression comparing the CTPA with the last available study regarding the parenchymal involvement.
The patients were divided according to the CTPA results into two groups: a group with positive PE and a group with negative PE.
The clinical and laboratory data included the following: -The timing of the CTPA after the onset of the symptoms. -The main complaint at the time of CTPA scanning.
-The laboratory results including the WBC count, neutrophil count, lymphocyte count, D-dimer level, ferritin, and CRP levels. All laboratory results included must be within 24 h from the time of CTPA.

Sample size calculation
Using PASS 11 program for sample size calculation, confidence level 95%, and a margin of error ± 0.1 and by reviewing previous study results by Grillet et al. [4] showed the rate of pulmonary embolism among COVID-19-positive patients (23%); based on that, the required sample size will be 76 patients with COVID-19 to be sufficient to estimate the rate of pulmonary embolism in COVID-19 patients.

Statistical methods
All the data were collected, compared, and analyzed using IBM SPSS statistics (V. 26.0, IBM Corp., USA, 2019) for data analysis. Data were expressed as mean ± SD for quantitative parametric measures in addition to median and percentiles for quantitative nonparametric measures and both number and percentage for categorized data. The following tests were done: 1. Comparison between two independent groups for non-parametric data using the Wilcoxon rank-sum test. 2. Chi-square test to study the association between every 2 variables or comparison between 2 independent groups as regards the categorized data. The probability of error at 0.05 was considered significant, while at 0.01 and 0.001 are highly significant.

Results
The female patients. No significant difference was found between the incidence of pulmonary embolism and the patients' age and sex. The main complaint at the time of CTPA was progressive dyspnea detected in 64 patients (66.7%). Although progressive dyspnea was the highest in incidence regarding the complaint at the time of CTPA scanning, yet it showed no significant relationship with the incidence of PE. However, oxygen desaturation, chest pain, and hemoptysis showed a statistically significant relationship with the incidence of PE being significantly higher between patients with positive PE (Table 1).

Regarding the laboratory data
Patients with positive PE showed higher D-dimer and CRP levels and lower lymphocytic counts compared to the patients with negative PE. However, no significant relation was found between the level of the D-dimer, ferritin, CRP, WBC count, neutrophil, and lymphocyte count and the incidence of PE (Table 2). However, a highly significant correlation was found between the incidence of PE and the rising in the D-dimer level ( Table 1, Fig. 1).

Regarding the CT
The median time for CTPA was 9 days for patients with negative PE and 12 days for patients with positive PE with no statistically significant relationship found between the timing of CTPA and incidence of PE (Table 2). CT with typical criteria for COVID-19 infection was the highest in the two groups yet with no significant relationship found between the COVID-19 probability as reported in CT guided by the RSNA criteria and the incidence of PE (Table 1). The median CT severity score of the COVID-19 was statistically insignificant lower between patients with positive PE (Table 2). However, Progressive CT findings compared to the previous CT showed a highly significant correlation with the incidence of PE (Table 1, Fig. 1).

Discussion
The current study focused on highlighting the relationship between the incidence of pulmonary embolism and COVID-19 disease severity either laboratory or by the CT severity score trying to help in early anticipation, diagnosis, and treatment of complicated COVID-19 infection cases. The study included 96 patients with COVID-19 infection and suspected to have PE, out of which 40 were positive; meaning 41.7% of patients with suspected PE. The most common symptom was worsening dyspnea with an incidence of 66.7% in this study population. However, it was not statistically significant. The statistically significant symptoms were oxygen desaturation, chest pain, and hemoptysis. A similar study by Kaminetzky et al. [8] showed a PE incidence of 37% which is close to our study. This was higher than Poyiadji et al. [5] who found a 22% incidence of PE between their study group, yet they are also similar to our study which concluded no significant difference in age and gender regarding the incidence of PE.
In this article, we studied the relationship between the D-dimer, ferritin, CRP, WBC count, neutrophil, and lymphocyte count and the incidence of PE which was found insignificant although patients with positive PE Fig. 1 The incidence of rising D-dimer level and the progressive CT findings between the patients with positive and negative pulmonary embolism (PE) according to CTPA interpretation showed higher levels of D-dimer and CRP compared to the patients with negative PE. Also, the rising D-dimer was found to be highly significant. This mandates incorporation of the follow-up of D-dimer in the management of patients. This could explain the result demonstrated by Kamunetzky et al. [8] that concluded that only a significant indicator was the D-dimer closest to the incidence of PE. Another study by Garcia-Olivé et al. [9] demonstrated that patients with high levels of D-dimer have a higher probability of developing PE. Poyiadji et al. [5] concluded an increase in the D-dimer level of 6 μg/ml had an odds ratio of 2.7 for developing a PE. However, in controversy to our study, they found a significant difference not only with D-dimer but also with CRP. The patients with high D-dimer and high CRP were more susceptible to develop PE. All these inflammatory markers are alarming the use of anticoagulants as advised by many authors [10][11][12].
The median time to develop PE was found to be 12 days; however, that was statistically insignificant. The duration of illness in our opinion could raise a flag to continuous follow-up of other alarming symptoms and serial measures of D-dimer. The duration in the study done by Garcia-Olivé et al. [9] was 9.7 which is in keeping with our results. This is also similar to Grillet et al. [4] who diagnosed PE at a mean of 12 days from the onset of the symptoms.
The severity of infection detected by CT severity score calculation was found to be not significant with the incidence of PE which is similar to the results concluded by Bompard et al. [1]. However, the progression of the disease was found to be highly significant with the incidence of PE and this was against Bompard et al. [1] who found the higher incidence of CT progression among patients with negative PE. Again, the follow-up of critically ill patients is considered the key to early diagnosis and life-saving for those critically ill patients.
The incidence of segmental PE was higher among our study population with only 3 cases representing 7.5% of the positive cases developed RV strain. This is almost close to the results found by Poyiadji et al. [5] who found more incidence of segmental PE (51% of cases) and the RV strain occurred in 11% of the patients under study.

Limitation
Our study had few limitations; first, a retrospective study was done in one center, and larger multicentric studies are needed to further understand the nature of this novel complex illness and better manage the second wave. We did not include anticoagulation in our study which will be done in future studies.

Conclusion
We conclude from our study that for PE to occur during COVID-19 infection, several factors are associated which include disease progression with worsening symptoms as hemoptysis, oxygen desaturation, and chest pain. Rising D-dimer during the second week of disease which all are alert to do CT pulmonary angiography to exclude or confirm PE. This also raises the utmost need for anticoagulation prophylactically.