To investigate the lung ultrasound characteristics in different clinical classifications of coronavirus disease 2019 (COVID-19), and to provide references for clinical diagnosis and efficacy evaluation.

Thirty-nine patients with clinically diagnosed COVID-19 from March to July 2020 in Beijing Ditan Hospital, Capital Medical University were analyzed, retrospectively, including 13 patients with general type, 15 patients with severe type and 11 patients with critical type. All patients underwent lung ultrasound examination, and the features of lung ultrasound images were observed.

Ultrasound findings of pleural line: continuous but irregular and unsmooth pleural lines, or with thickened pleural were observed in 23.08% (9/39) patients; discontinuous and damaged pleural lines were detected in 46.15% (18/39) patients; blurred and serious damaged pleural lines occurred in 30.77% (12/39) patients. B lines: 2.56% (1/39) patients detected B lines (≤ 3); 2.56% (1/39) patients with B increased lines (≥ 4); significantly increased B lines which fused into "waterfall sign" were observed in 94.87% (37/39) patients. Consolidation: 23.08% (9/39) patients were without consolidation; small subpleural consolidation occurred in 53.85% (21/39) patients; large-scale consolidation with air bronchogram sign occurred in 20.51% (9/39) patients. Lung ultrasound features in COVID-19 patients with different clinical classifications: among all the patients with continuous but irregular and unsmooth pleural lines or combined with pleural thickening and no consolidation, the general type accounted for the largest proportion, accounting for 66.7% (6/9) and 77.8% (7/9) of all clinical classifications. Discontinuous pleural line, rupture, and small subpleural consolidation mostly occurred in severe type patients, accounting for 50% (9/18) and 50.0% (11/22), respectively. Blurred and serious damaged pleural lines and large-scale consolidation with air bronchogram sign mostly presented in critical patients, accounting for 50% (6/12) and 62.5% (5/8), respectively. The absence of consolidation was significantly different among different clinical subtypes (P = 0.003).

Different clinical classifications of COVID-19 patients have different features in lung ultrasound, which could provide reference for clinical diagnosis and efficacy evaluation.