Severe COVID-19 is associated with a hypercoagulable state, with a high risk of thrombotic phenomena such as pulmonary thromboembolism (PE). Its diagnostic suspicion is complicated, due to the overlap of symptoms of PE with those of COVID-19 itself. Therefore, it is essential to improve PE prediction to optimise the performance of confirmatory imaging tests such as thoracic CT angiography. Early diagnosis has relevant therapeutic implications, as it justifies starting anticoagulant treatment early, with a possible positive impact on the clinical evolution of these patients.

The CHOD risk scale has recently been described: the acronym for C-reactive protein concentration, heart rate, oxygen saturation, and D-Dimer levels. Its initial description was carried out in a study in a single hospital centre. proving to be an easy-to-apply tool, useful for predicting the appearance of PE in patients hospitalized for COVID-19.

The objective of this study is to carry out an external validation of this scale in patients hospitalized for COVID-19 pneumonia, through an observational, cross-sectional, multicenter, real-life study in patients hospitalized for severe COVID-19 pneumonia, confirmed by objective methods, and showing high D-dimer values.

Imaging tests with CT angiography will be performed in patients with elevated D-Dimer, following international clinical practice regulations. Given that they will be consecutive patients, CT angiography will be performed in all patients regardless of the patient's clinical probability of PE as long as they meet the inclusion criteria and none of the exclusion criteria.

To calculate the PE predictive power of the CHOD scale in the validation cohort, a methodology similar to that used in the construction cohort will be used, that is, the use of a ROC curve.

Taking into account that a similar predictive value (with a maximum error of 5%) between the CHOD scale in the construction cohort and that of this study (validation cohort) will be considered as an adequate external validation, and taking into account a statistical power of 80%, an alpha error of 5% and a maximum loss of patients of 15%, the required sample size is 245 patients. Since 7 centres initially participate, each of which will have to contribute 35 valid consecutive patients for the analysis.

Inclusion Criteria:

1. Diagnostic confirmation of SARS-CoV-2 infection by positive result in at least one test:

   • Rapid nasopharyngeal smear antigen test.
   • PCR smear or nasopharyngeal aspirate.
   • Sputum PCR.
   • PCR of bronchoalveolar lavage.
   • Positive IgM in the serology for SARS-CoV2

2. Hospitalization for COVID-19 pneumonia

3. More than 5 days from the date of onset of symptoms (this criterion can be met during hospitalization).

4. Elevated D-dimer, by any of the following criteria:

   • DD> 500 ng/ml; In the case of patients over 50 years of age, the DD cut-off point adjusted for age (age x 10) will be applied.
   • With the DDU measurement, the cut-off point is 250 mg/dl (instead of 500) and the adjustments for age in those over 50 years are made by multiplying the age x 5 (instead of x 10).

Exclusion Criteria:

1. Pregnancy.

2. Age <18 years.

3. Hemodynamic instability: SBP <90 mmHg or vasopressors to achieve ≥ 90 mmHg.

4. Contraindication for performing CT angiography:

   • Severe kidney failure (GFR <30)
   • Allergy to iodinated contrasts.
   • Intolerance to decubitus.

Consecutive patients hospitalized for pneumonia caused by SARS-CoV-2 confirmed by objective methods, and with elevated D-dimer values.

Since it is a prevalence and validation study, pulmonary artery CT angiography will be performed consecutively in all patients (regardless of the pre-test clinical probability of PE that they present) who meet the criteria defined in the construction study of the CHOD risk scale, and that are managed by the group of participating physicians. The random assignment of admitted patients by the inpatient department manager to participating physicians ensures that the sample is representative of the admissions set, which will then allow inferences to be established without the need to establish a priori randomization that would imply ethical aspects.