April 17, 2020

Tool predicts severe COVID-19, NT-proBNP independently associated with in-hospital death, and bedside ultrasound for COVID-19 lung lesions

By Denise Baez

NEW YORK -- April 17, 2020 -- In today’s DG Alert, we cover a tool to early identify patients who will exacerbate to severe coronavirus disease 2019 (COVID-19), the prognostic value of N terminal pro B type natriuretic peptide (NT-proBNP) in patients with severe COVID-19, and the clinical value of bedside ultrasound in the diagnosis of lung lesions in patients with COVID-19. 

Researchers have created a tool that could help clinicians to early identify patients who will exacerbate to severe COVID-19, which will enable better centralised management and optimise use of medical resources.

Described in Clinical Infectious Diseases, the nomogram -- composed of 7 features of patients with COVID-19 -- had a significantly high area under curve (AUC, 0.912; 95% confidence interval [CI], 0.846-0.978) in the training cohort of 189 patients to discriminate individuals with severe COVID-19 from non-severe COVID-19, with a sensitivity of 85.71 % and specificity of 87.58%.

In a validation cohort of 165 patients, the AUC was 0.853 with a sensitivity of 77.5 % and specificity of 78.4%. In a second validation cohort of 18 patients, the sensitivity and the specificity of the nomogram were 75% and 100%, respectively.

The calibration curve for probability of severe COVID-19 showed optimal agreement between prediction by nomogram and actual observation. Decision curve and clinical impact curve analysis indicated that nomogram conferred high clinical net benefit.

“Our study has several strengths,” wrote Jiao Gong, Third Affiliated Hospital of Sun Yat-sen University, Guangzhou, China, and colleagues. First, we provide a practical quantitative prediction tool based on only 7 features, which are relatively inexpensive and easily obtained from the routine blood tests. Second, to guarantee the robustness of the conclusion, we included the data from 3 centres with a large sample size and validation in independent cohorts. The performance of our nomogram was efficient for clinical practice.”

Another study, published in Respiratory Research, showed that NT-proBNP might be an independent risk factor for in-hospital death in patients with severe COVID-19.

“The best cut-off value of NT-proBNP for predicting in-hospital death was 88.64 pg/mL, with a sensitivity of 100% and a specificity of 66.67%,” reported Lei Gao, MD, The First Affiliated Hospital of Chongqing Medical University, Chongqing, China, and colleagues. “Patients with high NT-proBNP values had a significantly increased risk of death during the days of follow-up compared with those with low values.”

Dr. Gao and colleagues analysed 54 patients with severe COVID-19 according to NT-proBNP levels. A total of 24 patients had levels ≤88.64 pg/mL and 30 had levels >88.64 pg/mL.

Compared with patients with lower NT-pro BNP levels, patients in the high NT-proBNP group were significantly older (51.6 ± 13.9 vs 67.4 ± 14.4 years) and had a higher prevalence of hypertension (8.3% vs 33.3%) and coronary heart disease (4.2% vs 26.7%). They also had higher levels of diastolic blood pressure (69-78 mm Hg vs 70-86 mm Hg), myohemoglobin (25.35 vs 82.53 ng/ml), creatine kinase-MB (0.63 vs 1.90 ug/L), high sensitivity troponin-I (hs-TnI; 0.006 vs 0.021 ng/mL), blood urea (3.4 vs 7.1 mmol/L), creatinine (54 vs 79 umol/L, white blood cell (WBC; 5.89 vs 5.73 109/L), C-reactive protein (7.6 vs 54.3 mg/L), and procalcitonin (0.038 vs 0.137 ng/ml), and lower level of lymphocyte (1.30 ± 0.43 vs 0.98 ± 0.55 109/L). 

Univariate analyses showed that the hazard ratio (HR) of NT-proBNP associated to in-hospital death was 1.369 (95% confidence interval [CI], 1.217-1.541; P < 0.001) per 100 pg/mL increase.

Other factors significantly associated with risk of death included age (per 10 years; HR = 1.975; 95% CI, 1.309-2.981), hs-TnI (per 1 ng/ml increase, HR = 1.862; 95% CI, 1.273-2.722), creatine kinase-MB (per 1 μg/L increase, HR = 1.259; 1.098-1.443), and procalcitonin (per 0.1ng/ml increase, HR = 1.241; 95% CI, 1.142-1.349). 

Multivariate Cox proportional hazards regression analyses were used to evaluate the independent prognostic effect of NT-proBNP level. After adjusting for sex and age, the HR of NT-proBNP for in-hospital death was 1.323 (95% CI, 1.119-1.563, P = 0.001). After adjusting for history of hypertension and coronary heart disease, the HR was 1.342 (95% CI, 1.185-1.520; P < 0.001). After adjusting for myohemoglobin, creatine kinase-MB, and hs-TnI, the HR was 1.360 (95% CI, 1.177-1.572; P < 0.001). After adjusting for urea and creatinine, the HR was 1.373 (95% CI, 1.188-1.586; P < 0.001). After adjusting for WBC and lymphocyte, the HR was 1.248 (95% CI, 1.097-1.419; P = 0.001). After adjusting for WBC, lymphocyte, and c-reactive protein, the HR was 1.230 (95% CI, 1.003-1.509; P = 0.047). After adjusting for WBC, lymphocyte, and procalcitonin, the HR was 1.200 (95% CI, 1.045-1.380; P = 0.010). 

“After adjusting for potential cofounders in separate modes, NT-proBNP presented as an independent risk factor of in-hospital death in patients with severe COVID-19,” the authors concluded. “By investigating the prognostic effect of NT-proBNP level of patients with severe COVID-19 at admission, it might be helpful to early identifying patients with poor prognoses. However, this study was limited by sample size and a single test of NT-proBNP at admission. Larger studies with continuous monitoring of NT-proBNP are necessary to further confirm the prognostic effect of NT-proBNP in patients with severe COVID-19.”

Lastly, a study published in the European Journal of Ultrasound showed the diagnostic efficacy of bedside ultrasound for patients with severe COVID-19.

Wuzhu Lu, MD, Fifth Affiliated Hospital of Sun Yat-sen University, Zhuhai, China, and colleagues, retrospectively analysed 30 patients admitted to their hospital from January 18, 2020 to February 5, 2020. All cases were examined by lung ultrasound and CT and reviewed by blinded observers, with imaging scores being used to analyse the ultrasound findings of lung lesions. 

There was moderate agreement (Kappa  = 0.529) between the noninvasive assessment of bedside ultrasound for lung lesions and CT. Lung ultrasound signs in patients with COVID-19 were mainly manifested as interstitial pulmonary oedema (90.0 %) and pulmonary consolidations (20.0 %).

The ultrasound scores to evaluate mild lung lesions exhibited sensitivity of 68.8 % (11/16), specificity of 85.7 % (12/14), and diagnostic accuracy of 76.7 % (23/30). The scores to evaluate moderate lung lesions exhibited sensitivity 77.8 % (7/9), specificity of 76.2 % (16/21), and diagnostic accuracy of 76.7 % (23/30). The sensitivity, specificity, and diagnostic accuracy for severe lung lesions was 100.0 % (2/2), 92.9 % (26/28), and 93.3 % (28/30), respectively. 

The follow-up dynamic ultrasound examination showed that the condition of all patients worsened gradually, with the ultrasound scores of lung lesions increasing to varying degrees.

“Though the diagnostic efficacy of bedside ultrasound is relatively low for mild to moderate patients, it is high for severe patients,” the authors wrote. “Bedside ultrasound has important clinical significance for noninvasive assessment and dynamic observation of lung lesions in patients with COVID-19, which is worth further consideration.”

SOURCE: Clinical Infectious Diseases, Respiratory Research, and European Journal of Ultrasound