Introduction
Methods
Results
Discussion
Limitations
Conclusion

INTRODUCTION

Since the emergence of severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) in Wuhan, China, there have been over one million deaths and over 89 million cases related to coronavirus disease 2019 (COVID-19) in the United States.
¹
Although COVID-19 was initially characterized as a respiratory illness, critically ill patients have proven to have an associated hypercoagulable state.
²
The hypercoagulable state appears to originate in the pulmonary vasculature and evolves into a generalized hypercoagulable state resulting in macro- and microvascular thrombosis such as acute pulmonary embolism (APE).
<sup>3

,

4</sup>

While growing research has documented the incidence of APE in hospitalized patients with COVID-19, few published studies have evaluated patients with SARS-CoV-2 infection and associated diagnosis of APE upon initial presentation to the emergency department (ED). Previous studies of APE risk in ED COVID-19 patients have been either inconclusive and even contradictory.
<sup>5

–

8</sup>
The limited information suggests that rates of APE in non-hospitalized COVID-19 patients may be as high as 18%, more than seven-fold higher than in the non-COVID-19 ED population.
<sup>9

,

10</sup>
With conflicting evidence on the incidence of APE in the ED setting, there remains a paucity of literature discussing diagnostic algorithms and computed tomography pulmonary angiogram (CTPA) diagnostic yield (the percentage of positive scans) for APE in ED COVID-19 patients. Proposed algorithms using D-dimer levels vary greatly and are not ED-specific.
⁸

The diagnosis of APE in COVID-19 patients presents a diagnostic dilemma in the ED. The post-acute SARS-CoV-2 symptoms of dyspnea, chest pain, and tachycardia are all associated with clinical characteristics for APE.
⁸
Further, traditional methods of ruling out APE, such as using D-dimer in low-risk patients, are not feasible because D-dimer levels are commonly elevated in COVID-19 patients.
⁹
In particular, a known relationship exists between the level of D-dimer elevation and COVID-19 severity.
¹⁰

In this derivation study, our primary objective was to identify which of the commonly known risk factors for APE were associated with APE in a COVID-19 patient population in the ED. Our secondary objective was to identify D-dimer values associated with APE in the ED setting.

METHODS

This retrospective review was approved by our institutional review board. We performed a multicenter, retrospective cohort analysis for adult patients who arrived to any of the five EDs within the Atrium Health Wake Forest Baptist system between March 17, 2020–January 31, 2021. The EDs included one academic medical center and four regional community hospitals.

The inclusion criteria for the study were patients >16 years of age who tested positive for SARS-CoV-2 or had a COVID-19-related diagnosis and had a CTPA study ordered. A COVID-19-related diagnosis was based on International Classification of Diseases, 10^th Rev, (ICD-10) codes. (The list of ICD-10 codes used is included in Appendix 1.) Using these criteria, we extracted a patient list from our electronic health record (EHR) via a health analytics software and services company (Roundtable Analytics, Research Triangle Park, NC). Final inclusion was based on confirmation of SARS-CoV-2 based on reverse transcription polymerase chain reaction (RT-PCR) or rapid antigen testing. This article follows the Strengthening and Reporting of Observational Studies in Epidemiology (STROBE) guidelines.
¹¹

The clinical characteristics we focused on were based on commonly used, ED-specific APE decision rules: pulmonary embolism rule-out criteria, Well’s criteria for APE, and the Geneva Score for APE.
<sup>12

–

14</sup>
D-dimer values consisted of both fibrinogen equivalent unit (FEU) and D-dimer unit (DDU). To bring parity to the different assays, DDU results were doubled. This was performed in a manner that has been described in prior COVID-19 D-dimer studies.
¹⁵
The cut-off for one hospital’s FEU assay was 399 micrograms per liter (μg/L), while the other FEU assay cut-offs were 500 μg/L. The cut-off for the DDU assay was 230 μg/L. We calculated chi-square values, odds ratios (OR) with 95% confidence intervals (CI), and Kruskal-Wallis testing of numeric medians, and we used logistic regression to compare characteristics of patients who had APE to those who did not, using P < 0.05 as significant.

RESULTS

We identified 425 patients who underwent evaluation for pulmonary embolism with CTPA in the setting of a suspected COVID-19 infection during the study period. Of this cohort, 408 patients (96%) tested positive for SARS-CoV-2 infection by RT-PCR analysis or rapid antigen testing and were included in our study. Of the CTPAS performed, 72% were done at the four community hospitals and 28% at the academic medical center hospital. Patient demographic and clinical characteristics are summarized in Table 1. Twenty-nine patients (7.1%) were ultimately found to have an APE on CTPA. The diagnostic yield of APE on CTPA varied from a high of 9.9% at one of the medium-sized community hospitals to a low of 2.3% at the smallest community hospital in our system.

The heart rate was significantly higher in patients who were found to have APE (median 102 beats per minute [bpm], interquartile range [IQR] 23 compared to 95 bpm (IQR 28), P = 0.0133]. Patients with APE were significantly more likely to present with hypoxia or a supplemental oxygen requirement (OR 2.4, 95% CI 1.1–5.3; P = 0.02]. Notably, 37.9% of patients found to have APE were not hypoxic. Patients with positive chest radiographs (CXR) experienced significantly more hypoxia (51.7% (89/172) vs 34.1% (46/135), P = 0.002). There was no significant difference between type of PE (saddle, segmental, and subsegmental) and oxygen requirements (P = 0.43). There was no significant association between the presence of an infiltrate on CXR and APE (P = 0.26).

Median age was observed to be slightly higher in patients with PE (61, IQR 24) compared to 58 (IQR 25), but this difference was not significant (P = 0.12). There was no significant difference in body mass index (BMI) between patients who were and were not found to have APE (33, IQR 14 vs 31.5, IQR 11]; P = 0.91]. The proportion of patients found to have APE was not significantly different at 0–5, 6–9, and greater than nine days of illness (P = 0.83, Table 2). Patients with severe COVID-19 were more likely to have APE (20/145, [13.%] vs 9/263 [3.4%}, OR 4.5, 95% CI 2.0–10.2;
P < 0.0001). Of the 29 patients with APE, 27 (93.1%) were admitted and followed through their hospital stay.

In patients with D-dimer testing, 204/228 (89.4%) were found to have elevated values as defined by local laboratory normal values. In this cohort, detailed in Figure 1, an abnormal D-dimer had a sensitivity of 93% (95% CI 66–100%) and a specificity of 11% (95% CI 7–16%) for the diagnosis of pulmonary embolism. A positive D-dimer was not significantly associated with a diagnosis of APE (P = 0.35). Of patients with D-dimer testing and APE, D-dimer values ranged from 134 μg/L to 38,616 μg/L (Table 3). The median D-dimer value was significantly higher in patients who were found to have PE (4,240 μg/L vs 1,030 μg/L; P = 0.0048). Patients with a D-dimer of greater than 1,500 μg/L were significantly more likely to have APE (P = 0.001).

Using logistic regression analysis in the cohort with D-dimer results, we found that a BMI greater than 32, (OR 4.4, 95% CI 1.0–19.3; P = 0.045), a heart rate >90 bpm, (OR 5.0, 95% CI 1.0–24.9; P = 0.048), and a D-dimer greater than 1,500 μg/L (OR 5.6, 95% CI 1.6–20.2; P = 0.008) were significantly associated with APE. Using a D-dimer cut-off of 1,500 μg/L yielded the best balance of sensitivity and specificity (Table 4) for diagnosis of APE. In patients where D-dimer was obtained, no patients with a D-dimer <1,500 μg/L and a heart rate <90 bpm were found to have APE. Use of these thresholds for CTPA testing would have decreased testing by 27.2%, representing 62 CTPA scans of the 228 patients for whom D-dimers were drawn.

DISCUSSION

The primary finding of this multicenter, retrospective cohort analysis showed that the risk factors of BMI greater than 32, a heart rate >90 bpm and D-dimer >1500 μg/L were significantly associated with APE diagnosis in COVID-19 patients. It was interesting that while median BMI did not significantly differ between those with APE and those without APE (33 vs 31.5), a cut-point of >32 was associated with APE. While patients with hypoxia or supplemental oxygen requirement were more likely to have APE, this cohort still showed that a significant portion of APE patients were not hypoxic. The secondary finding shows that D-dimer levels of 1,500 μg/L were significantly associated with APE, while just having a positive D-dimer level was not. The diagnostic yield of APE in our patient population was 7.1%, contradictory to recent literature that showed dramatic increased incidences of APE in COVID-19 patients.
<sup>16

,

17</sup>

Striking were the clinical characteristics not statistically significant for association with APE. For example, multiple studies have shown an increased risk of APE in the inpatient setting for patients requiring admission for COVID-19.
¹⁸
This is thought to be related to elevated pro-inflammatory cytokines and abnormalities in coagulation parameters.
¹⁹
We would have expected to find a statistical significance in findings of APE in patients who are later in their illness of COVID-19, after day 6 of illness when symptoms become more pronounced. While we discovered most of our study population (71.4%) was found to have APE after day 5 of COVID-19 illness, this was not statistically significant.

The use of CTPA has associated risks, costs, and staff resources that must be considered when ordering testing. Risks of CTPA include ionizing radiation exposure, contrast nephropathy, and contrast allergies.
<sup>20

–

22</sup>
Costs associated with CTPA studies are not just to the patient but to health system capacity, both of which are significant.
²³
In a setting of limited healthcare staffing and bed availability, the increased staff resources required for CTPA studies must be considered.
²⁴
Thus, reduction in unnecessary CTPA studies would yield multiple benefits.

Our subset of patients for whom D-dimers were obtained (228) offered the best opportunity for reduction in CTPA studies. With almost 90% of our patient population who had D-dimers drawn having an elevated result, using traditional cut-offs were not helpful in evaluation of APE in COVID-19 patients. However, using elevated D-dimer cut-offs in specific patient populations in the evaluation for APE is not a new concept in emergency medicine. The pregnancy-adapted YEARS algorithm and age-adjusted D-dimer cut-off values for diagnosis of suspected APE are two examples of algorithms that use elevated D-dimer value cut-offs.
<sup>25

,

26</sup>

We found using a D-dimer cut-off of 1,500 μg/L yielded the best balance of sensitivity and specificity (Table 4) for diagnosis of APE. In patients whose D-dimer was obtained, none of them with a D-dimer <1,500 μg/L and a heart rate <90 bpm were found to have APE. Use of these thresholds for CTPA testing would have decreased CTPA usage by 27.2%, potentially eliminating 62 CTPA studies in the 228 patients for whom D-dimers were drawn.

LIMITATIONS

Limitations to our study included the inability to quantify clinician gestalt when choosing to order a CTPA study. A second limitation was the inability to use the same D-dimer assay across all hospitals due to different lab equipment. In our study we included two types of D-dimer assays: FEU and DDU. To achieve parity, our methods used standardization and reporting suggested in COVID-19 D-dimer literature review.
¹⁵
While our study covered a regional health system with five EDs, it still lacks generalizability and would require external validation. External validation is particularly important as our study yielded only 14 patients with APE in our D-dimer subset of 228 patients. However, this low yield could potentially be revealing with regard to low overall findings of APE in ED COVID-19 patients.

CONCLUSION

In patients with acute COVID-19 infections, D-dimer at standard cut-offs was not usable; limiting CTPA using a combination of heart rate >90 bpm or D-dimer >1,500 μg/L would significantly decrease the use of imaging in this population. Future prospective studies are needed to determine whether using this D-dimer threshold and heart rate cut-off in the ED COVID-19 patient population can safely reduce the number of CTPA studies performed.

Footnotes

Section Editor: Patrick Meloy, MD

Full text available through open access at http://escholarship.org/uc/uciem_westjem

Address for Correspondence: Iltifat Husain, MD, Wake Forest School of Medicine, Department of Emergency Medicine, 1 Medical Center Blvd, Winston Salem, NC 27157. Email: ihusain@wakehealth.edu
11 / 2023; 24:1043 – 1048

Submission history: Revision received August 25, 2022; Submitted January 31, 2023; Accepted March 12, 2023

Conflicts of Interest: By the WestJEM article submission agreement, all authors are required to disclose all affiliations, funding sources and financial or management relationships that could be perceived as potential sources of bias. No author has professional or financial relationships with any companies that are relevant to this study. There are no conflicts of interest or sources of funding to declare.

REFERENCES

1.

2.

3.

4.

5.

6.

7.

8.

9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.