Demographic characteristics of the patients
As of March 11, 2020, a total of 146 confirmed patients in Taizhou were transferred to isolation hospitals and were discharged after the treatment. Three cases were not included in this study owing to the partial data. We obtained basic information, clinical data, and CT images of 143 patients (96.6%), among whom 66 were women (46.2%), and 77 were men (53.8%). The demographic and clinical characteristics of the patients are shown in Table 1. On admission, the degree of severity of COVID-19 was categorized as non-severe in 110 patients and severe in 33 patients. Of all the patients, 60 patients (42.0%) had contacts with a confirmed patient, 70 (49.0%) had contacts with a person in the Wuhan area, 4 (2.8%) had contacts with both a confirmed patient and a person in the Wuhan area, and 9 (6.3%) had an unclear contact history. The median age of all the patients was 47 years (range, 4–86 years), 56.8% were aged 30–60 years. The median age of the non-severe patients was 44.5 years (range, 4–80 years), while the median age of the severe patients was 55.0 years (range, 27–86 years). 129 patients (90.2%) had a history of smoking, including 100 patients (90.9%) in the non-severe group and 29 patients (87.9%) in the severe group.
The mean incubation period was 6.9 days (standard deviation [SD], 3.472 days), and the mean treatment delay period was 3.0 days (SD, 2.631 days). Fever (73.4%) was the most common symptom. A total of 91 patients (63.6%) developed cough: sputum (28.7%), yellow sputum (5.6%), and dry cough (29.4%). Sore throat (10.5%), nasal congestion (5.6%), muscle soreness (9.8%), and chest tightness (17.5%) were relatively few. Of all the patients, 24.5% had at least one coexisting illness (hypertension and/or chronic obstructive pulmonary disease). The average body temperature of the patients was 38.0℃ (range, 37℃–39℃; SD, 3.741℃), and only 1 case (0.7%) had a high fever (>40℃).
Laboratory examination results of the two groups
As some of the patients were first admitted to the county hospital, some laboratory tests were not carried out owing to the conditions. A total of 123 cases was included for the analysis of laboratory examination results, and those without such examination were recorded as UNKOWN. There were 28 cases in the severe group and 95 cases in the non-severe group. Table 2 shows the details of the laboratory results of all the cases. In the non-severe group, except for the increase in blood glucose level (14 ± 60 mmol/L), the decrease of serum albumin (39 ± 5.2 g/L), erythrocyte sedimentation rate (35 ± 24 µmol/L), and serum sodium (130 ± 25 mmol/L), and the increase of CRP (15 ± 19 mg/L) and amyloid A (190 ± 330 mg/L) were observed. All other test results were within the normal range (Table 2).
In the severe group, following parameters showed a decreased value: lymphocyte count decreased ((0.91 ± 0.43)×109), erythrocyte sedimentation rate (38 ± 24 µmol/L), serum albumin (130 ± 25 g/L), glomerular filtration rate (82 ± 26 mL/min), arterial blood oxygen partial pressure (83 (± 24) mmHg), calcitonin level (0.086 ± 0.090 µg/L), total CD3 value (56 ± 13%), absolute CD4 value (290 ± 200/µL), and absolute CD8 value (190 ± 130/µL). Following parameters showed an increased value in the severe group: blood glucose level (13 ± 53 mmol/L), lactate dehydrogenase level (350 ± 180 U/L), CPR level (35 ± 31 mg/L), amyloid protein A level (430 ± 420), fibrinogen detection value (4.5 ± 1.3 s), D dimmer level (0.54 ± 0.63 g/L), pH level (7.8 ± 2.0), myoglobin (84 ± 80 ng/mL), IL-10 (9.4 ± 9.2 pg/mL), C1q (240 ± 38 mg/L), PT (16 ± 19 s), absolute CD3 value (490 ± 330/µL). Other test results were within the normal range.
CT manifestationsof the two groups
Five (3%) CT images could not be evaluated owing to poor respiratory artefact quality. CT images of 138 (97%) patients at admission were obtained. Among them, 34 patients were severe (31%) and 104 patients were non-severe (75%). The lesions in 135 patients (98%) were mainly distributed in the external or subpleural of the lung; 37 patients (27%), middle or inner zone of lung; and 34 patients (25%), both. Among them, the lesions in the middle or inner band of the lungs were more commonly observed in the severe patients (47% vs 20%). Mass (81%) was the most common lesion shape followed by patchy (68%). Irregularity in the shapes of the lung segments was observed only in severe patients, while the nodular shape was observed only in non-severe patients. Most of the lesions presented with mixed density (63%).
In 60% of the CT images, the lesion was distributed along the pulmonary bronchial tree, and in 62% of the CT images, the lesion was adjacent to vasodilation. Fifty‑two percent of the CT images showed interlobular septal thickening, and 30% of the CT images showed fibrous foci. Other concomitant signs such as cavitation or calcification (1%), enlarged lymph nodes (3%), pleural effusion (2%), chronic bronchitis (4%), emphysema, or pulmonary bullous (4%) were rare. Consolidation (68% vs 41%), bronchial inflation signs (59% vs 41%), and bronchiectasis (71% vs 39%) were more common in the severe group.
Results of univariate logistic regression analysis predicting the severity of COVID-19 patients
Univariate logistic regression was performed for all the collected variables, and the results are shown in Table 4. For all the factors such as clinical characteristics, laboratory results, and CT findings, the variables associated with the severity of COVID–19 were age (Odds ratio [OR]: 1.052, 95% confidence interval [CI]: 1.020–1.086, P = 0.001), days from the symptom onset to hospitalisation (OR: 1.213, 95% CI: 1.034–1.939, P = 0.016), days from the symptom onset to diagnosis (OR: 1.213, 95% CI: 1.084–1.357, P <0.001), body temperature (OR: 2.252, 95% CI: 1.139–4.450, P = 0.020), Neutrophil count (OR: 0.087, 95% CI: 0.026–0.274, P <0.001), Lymphocyte count (OR: 1.128, 95% CI: 1.000–1.272, P = 0.049), IgM (OR: 2.226, 95% CI: 1.015–4.883, P = 0.046), ADA (OR: 1.163, 95% CI: 1.023–1.323, P = 0.021), albumin (OR: 0.847, 95% CI: 0.725–0.988, P <0.035),CRP (OR: 1.024, 95% CI: 1.007–1.042, P = 0.006),Glomerular filtration rate (OR: 0.965, 95% CI: 0.942–0.988, P = 0.004), amyloid A (OR: 1.002, 95% CI: 1.001–1.003, P = 0.001), PCT (OR: 1.43E+11, 95% CI: 1334.315–1.53E+19, P = 0.006), PaO2 (OR: 0.972, 95% CI: 0.953–0.992, P = 0.007), oxygen concentration (OR: 1.027, 95% CI: 1.001–1.053, P = 0.044), oxygenation index (OR: 0.992, 95% CI: 0.988–0.996, P <0.001), myoglobin (OR: 1.011, 95% CI: 1.002–1.020, P = 0.014), IL‑10 (OR: 1.184, 95% CI: 1.037–1.351, P = 0.012), consolidation (OR: 3.103, 95% CI: 1.399–6.885, P = 0.005), and bronchiectasis (OR: 3.818, 95% CI: 1.694–8.605, P = 0.001).
To construct a clinical-image nomogram for the prediction of severe COVID-19.
All the variables with P <0.05 were screened for multivariable logistic regression. Finally, variables selected for predicting severe COVID-19 were age (OR: 1.03, 95% CI: 0.99–1.07, P = 0.1565011), body temperature (OR: 2.28, 95% CI: 1.01–5.18, P = 0.0480145), Lymphocyte count (OR: 0.11, 95% CI:0.03–0.38, P = 0.0006512), ADA (OR: 1.27, 95% CI: 1.08–1.49, P = 0.0037539), PaO2 (OR: 0.98, 95% CI: 0.95–0.99, P = 0.0511345), IL- 10 (OR: 1.15, 95% CI: 0.99–1.34, P = 0.063646), bronchiectasis (OR: 5.25, 95% CI: 1.77–5.62, P = 0.0028594). The final equation is: Nomoscore = 34.663 + Age × 0.028 + Temperature × 0.826 + L × −2.250 + ADA × 0.239 + PaO2 × −0.023 + IL‑10 × 0.143 + Bronchiectasis × 1.659.
The constructed equation was used for drawing the ROC, with the area under the curve being 0.877. The Youden index was used for determining the optimal threshold, and the corresponding sensitivity and specificity were 74.29% and 82.73%, respectively.