| Abstract | Uvod: Ultrazvuk pluća (LUS) korisna je metoda snimanja za identifikaciju upale pluća COVID-19. Cilj naše prve studije bio je istražiti ulogu LUS-a u predviđanju težine bolesti i smrtnosti u bolesnika s COVID-19. Cilj druge studije je bio predložiti model predviđanja koji uključuje ultrazvuk pluća i usporediti ga s RTG-om i CT-om. Metode: studija provedena od 1. studenoga 2020. do svibnja 2021. LUS skor temeljio se na procjeni 14 plućnih zona s ukupnom ocjenom do 42, što je uspoređeno s težinom bolesti i smrtnim ishodom, CT skor se temeljio na 14 plućnih zona koje odgovaraju istim zonama kao i LUS skor, a RTG skor je bio podijeljen na 6 zona. Modeli predviđanja za nužnost mehaničke ventilacije (MV) ili smrtnog ishoda razvijeni su kombiniranjem slikovnih, biometrijskih i biokemijskih parametara. Rezultati: U prvu studiju uključena su ukupno 133 bolesnika s COVID-19 pneumonijom potvrđena RT-PCR-om, a medijan je od prijema u bolnicu na ultrazvuk pluća od jednog dana. LUS skor bio je u korelaciji s kliničkom težinom na prijemu u bolnicu (Spearmanov rho 0,40, 95% CI 0,24 do 0,53, str. < 0,001). Bolesnici s višim LUS skorovima imali su veću težinu bolesti; nosna kanila visokog protoka imala je omjer koeficijenta 1,43 (5% CI 1,17–1,74) u bolesnika s LUS skorom > 29; istim skorom predviđena je i potreba za mehaničkom ventilacijom (1,25, [1,07–1,48]). LUS skor > 30 (1,41 [1,18–1,68]) i stariji od 68 godina (1,26 [1,11–1,43]) bili su značajni prediktori smrtnih slučajeva. U drugo istraživanje uključeno je ukupno 255 bolesnika s COVID-19 pneumonijom. U regresijskom modelu identificirana su četiri neovisna prediktora za potrebu MV: LUS skor, dan bolesti, broj leukocita i kardiovaskularne bolesti (χ2 = 29,16, str. < 0,001). Model je točno klasificirao 89,9% slučajeva. Za smrtni ishod, samo dva neovisna prediktora pridonijela su regresijskom modelu: LUS skor i dob bolesnika (χ2 = 48,56, p < 0,001, 93,2% ispravno klasificirano). Zaključci: Pokazalo se da LUS pri prijemu u bolnicu ima visoku prediktivnu snagu težine i smrtnosti COVID-19 pneumonije. Prediktivni model identificirao je četiri ključna parametra pri prijemu bolesnika koji bi mogli predvidjeti štetan ishod. |
| Abstract (english) | Introduction: Lung ultrasound (LUS) is a useful imaging method to identify COVID-19 pneumonia. The aim of our first study was to investigate the role of LUS in predicting the severity of the disease and mortality in COVID-19 patients. The aim of the second study was to propose a prediction model involving lung ultrasound and compare it with CXR and CT scans. Methods: The study was conducted from 1 November 2020 to May 2021. The LUS score was based on an assessment of 14 lung zones with an overall score of up to 42, which was compared with the severity of the disease and mortality. CT score was based on 14 lung zones corresponding to the same zones as the LUS score, and the CXR score was divided into 6 zones. Prediction models for the necessity of mechanical ventilation (MV) or fatal outcome were developed by combining imaging, biometric and biochemical parameters. Results: The first study included a total of 133 patients with COVID-19 pneumonia confirmed by RT-PCR, and the median is from hospital admission to a one-day lung ultrasound. The LUS score was correlated with severity on hospital admission (Spearman's rho 0.40; 95% CI 0.24 to 0.53; p. 12) < 0.001). Patients with higher LUS scores had a higher severity of the disease; highflow nasal cannula had a coefficient ratio of 1.43 (5% CI 1.17–1.74) in patients with a LUS score > 29; the same result predicts the need for mechanical ventilation (1.25; [1.07–1.48]). LUS score > 30 (1.41 [1.18–1.68]) and age over 68 (1.26 [1.11–1.43]) were significant predictors of death. A total of 255 patients with COVID-19 pneumonia were included in the second study. In the regression model, four independent predictors for the need of MV were identified: LUS result, day of the disease, number of leukocytes and cardiovascular diseases (χ2 = 29.16; p. < 0.001). The model accurately classified 89.9% of cases. For mortality, only two independent predictors contributed to the regression model: LUS score and patient age (χ2 = 48.56; p < 0.001; 93.2% correctly classified). Conclusions: LUS has been shown to have a high predictive strength for severity and mortality of COVID-19 pneumonia. The predictive model identified four key parameters when admitting a patient that could predict a fatal outcome. |
