Background A higher incidence of thrombotic events, mainly pulmonary embolism (PE), has been reported in hospitalized patients with COVID-19. The main objective was to assess clinical and laboratory differences in hospitalized COVID-19 patients according to occurrence of PE. Methods This retrospective study included all consecutive patients hospitalized with COVID-19 who underwent a computed tomography (CT) angiography for PE clinical suspicion. Clinical data and median blood test results distributed into weekly periods from COVID-19 symptoms onset, were compared between PE and non-PE patients. Results Ninety-two patients were included, 29 (32%) had PE. PE patients were younger (63.9 (SD 13.7) vs 69.9 (SD 12.5) years). Clinical symptoms and COVID-19 CT features were similar in both groups. PE was diagnosed after a mean of 20.0 (SD 8.6) days from the onset of COVID-19 symptoms. Corticosteroid boluses were more frequently used in PE patients (62% vs. 43%). No patients met ISTH DIC criteria. Any parameter was statistically significant or clinically relevant except for D-Dimer when comparing both groups. Median values [IQR] of D-dimer in PE vs non-PE patients were: week 2 (2010.7 [770.1–11208.9] vs 626.0 [374.0–2382.2]; p = 0.004); week 3 (3893.1 [1388.2–6694.0] vs 1184.4 [461.8–2447.8]; p = 0.003); and week 4 (2736.3 [1202.1–8514.1] vs 1129.1 [542.5–2834.6]; p = 0.01). Median fold-increase of D-dimer between week 1 and 2 differed between groups (6.64 [3.02–23.05] vs 1.57 [0.64–2.71], p = 0.003); ROC curve AUC was 0.879 (p = 0.003) with a sensitivity and specificity for PE of 86% and 80%, respectively. Conclusions Among hospitalized COVID-19 patients, D-dimer levels are higher at weeks 2, 3 and 4 after COVID-19 symptom onset in patients who develop PE. This difference is more pronounced when the fold increase between weeks 1 and 2 is compared.