Diagnostic Performance of Clinical Probability Scores for Pulmonary Embolism in COVID

Abstract

<b>Introduction:</b> The use of clinical probability scores allows to standardizing diagnostic approaches for pulmonary embolism (PE) by rationalizing the use of diagnostic means. So, what would be the place of these scores in the diagnosis of PE in COVID 19. <b>Aim:</b> To study the diagnostic performance of the Wells and Geneva scores in PE in COVID 19. <b>Methods:</b> We included patients hospitalized in the pulmonary department of Charles Nicolle hospital in Tunisia (September 2020 to September 2021), for COVID pneumonia. Patients were divided in two groups: G1: with PE (n=20) and G2: without PE (n=450) ROC curves were generated to measure the diagnostic performance of the Wells and Geneva scores in PE in COVID-19 patients. <b>Results:</b> Among the 470 patients included, 20 had PE. The two groups had similar mean age (G1: 62 years old, G2: 60 years old, p= 0.5) with a male predominance (60% vs 54%, p= 0.6).&nbsp; A significant difference in Wells score between the two groups was noted (G1: 0.75 vs G2: 0.4, p= 0.032), without any difference for Geneva score (4.3 vs 3.8, p= 0.76). All patients with PE had a low probability according to Wells and 25% according to Geneva. The relevance of the Wells and Geneva scores was analyzed on the ROC curve. The area under the curve (AUC) was 0.617 ([0.485-0.750] 95% CI) and 0.608 ([0.475-0.741] 95% CI) respectively. The couple sensitivity and specificity for a Wells score &lt;2 was 50% and 75%). It was 75% and 44% for a Geneva score &lt;4. <b>Conclusions:</b> Our results are consistent with those of the literature in terms of diagnostic performance of clinical probability scores in PE in COVID 19. This performance was poor for both scores. Hence, the interest in associating certain biological markers like ddimer.