Predictive Value of the Neutrophil-to-Lymphocyte Ratio(NLR) for Diagnosis and Worse Clinical Course of the COVID-19: Findings from Ten Provinces in China
Nannan Shi,Yan Ma,Yipin Fan,Jingya Wang,Chen Zhao,Guangxi Li,Xiaofeng Zhang,Weiguo Bai,Furong Xiao,Jingwei Wang,Hui Na,Quntang Li,Mei Shi,Nong Tang,Jin Huang,Shaozhen Huang,Haihao Jin,Baihua Jiang,Hongming Xu,Tianqing Zhu,Shoufang Xu,Yanhua Xiao,Qiao Feng,Shusen Zhao,Hong Chen,Yingchun Zhou,Zhang Liu,Huaben Zhang,Dongting Wang,Yinzhen Wang,Xiaoyan Wang,Fangli Song,Kai Zheng,Lanping Wu,Jinping Liu,Haijun Xie,Ruixia Xue,Guifen Hu,Ya Mao,Gongqi Zhang,Wei Wu,Puye Yang,Mingbo Yang,Liying Chen,Jinbo Zhang,Tuanmao Guo,Yuanjun Li,Jiangfeng Bai,Xiao Lei,Minqing Li,Linsong Zhang,Junteng Zhu,Sheng Sun,Xiaomei Hu,Yunhong Hu,Shengli Yuan,Lin Tong,Sihong Liu,Renbo Chen,Zhan Shi,Li Li,Jia Liu,Hao Gu,Yingjie Zhi,Zhifei Wang,Yuanyuan Li,Ning Liang,Huizhen Li,Liwen Jiao,Ruili Huo,Yongyan Wang,Yanping Wang,Huamin Zhang,Wei Wang
DOI: https://doi.org/10.2139/ssrn.3569838
2020-01-01
SSRN Electronic Journal
Abstract:Background: Novel coronavirus pneumonia (NCP) is often changing rapidly and fatal. Early detection and early triage of coronavirus disease 2019 (Covid-19) is the key to success management of the disease. An easily obtainable yet accurate variable for both diagnosis and prognosis is urgently needed. We aim to report predictive Value of the Neutrophil-to-Lymphocyte Ratio(NLR) for diagnosis and worse clinical course of the COVID-19, which have not been well demonstrated. Methods: Our study consisted of two stages, at the first stage, a retrospective, single-center, cohort study including was conducted in Heilongjiang, on admission, demographic, clinical, and laboratory data were collected and compared between patients with COVID-19 and patients with non COVID-19; we used multivariable logistic regression methods to explore the risk factors associated with COVID-19;A receiver operating characteristic(ROC) analysis was conducted to calculate the area under the curve(AUC) to assess predictive value of NLR for diagnosis of COVID 19. At the second stage, we conducted retrospective, multi-center and large sample study in 43 hospitals from ten provinces of China, COVID-19 patients with laboratory-confirmed divided into three groups including mild cases, ordinary cases and severe cases. Multivariate logistic regression methods were used to identify the risk factors for the deterioration of COVID-19, along with, a receiver operating characteristic (ROC) curve was also drawn to assess impact on the clinical course of the COVID-19. Findings: We recruited a total of 635 patients with COVID-19 and 27 cases with non COVID-19(Viral pneumonia) from 28 January to 25 February. A total of 88 cases were enrolled with a retrospective, single-center, cohort study from Heilongjiang province, of these, COVID-19 cases were 61(69%) and non COVID-19 cases were 27(31%). On admission, fever (69%) was the most common symptoms, cough (56%) and fatigue(53%). An average(SD) of NLR of COVID-19 patients and non- COVID-19 patients were3.48±2.04 and 2.21±1.14, respectively. multivariable regression showed increasing odds of COVID-19 patients associated with NLR(odds ratio 1.752, 95% CI 1.111-2.763, per 1 unit increase; p=0.016). In addition, the area under the curve (AUC) of NLR was 0.707 and cutoff value was 2.22. At the second stage, 635 patients with COVID were enrolled with a retrospective, multi-center, large sample study in the 43 settings from 10 provinces, of these, mild case were 86(14%), ordinary cases [486(76%)],severe cases[63(10%)], common symptoms was at onset of disease were cough[356(56%)], an average of NLR of 635 patients was 4.04±4.68, and elevated NLR with the deterioration of clinical course[mild case(2.73±2.28), ordinary cases(3.58±3.07), severe cases(9.38±10.52), P<0.0001], in multivariable logistic regression model, compared to mild group, fever(OR 5.739, 95% CI 2.849-11.564) and cough(OR 3.265, 95% CI 1.675-6.331) were associated with ordinary cases, increasing odds of NLR was associated with ordinary cases (OR 1.199, 95% CI 1.010-1.422, per 1 unit increase; p=0.038). Fever(OR 7.587, 95% CI 2.601- 22.132),cough(OR 6.493, 95% CI 2.257-18.682) and shortness of breath(OR 4.133, 95% CI 1.125-15.179) were associated with severe cases, increasing odds of NLR (OR 1.342, 95% CI 1.122-1.605, per 1 unit increase; p=0.001), and increasing odds of age (OR 1.036, 95% CI 1.001- 1.073, per 1 unit increase; p=0.041) were associated with severe cases. The area under the curve (AUC) of NLR was 0.727 and cutoff value was 4.06, additionally, AUC of lymphocytes was 0.719 and cutoff value was 0.765. Interpretation: NLR as inflammatory markers with rapid, convenient characteristics, NLR≥2.22 could be utilized as a predicting indicator for the early recognition COVID-19 and facilitate detection timely; meanwhile, NLR≥4.06 and lymphocytes≤0.765 were as predicting indicator for severe COVID-19 and can facilitate to further prevent worse progress of clinical course. In addition, cough, shortness of breath, abnormal chest radiological findings were also associated with worse outcome.Funding Statement: The study was supported by“ National Science and Technology Major Project” (2018ZX10101001-005-003, 2018ZX10101001-005-004).Declaration of Interests: All authors declare that they have no competing interests.Ethics Approval Statement: The study was approved by National Administration of Traditional Chinese Medicine, Administration of Traditional Chinese Medicine of 10 provinces and the institutional board of 43 participating setting.