Neutrophil-to-lymphocyte ratio on admission is an independent risk factor for the severity and mortality in patients with coronavirus disease 2019.

The study by Liu et al. had been published in your journal, and reported that the neutrophil-to-lymphocyte ratio (NLR) was an independent risk factor for the mortality of the COVID-19 patients. Based on it, we reported the association between levels of NLR at admission and the disease severity in COVID-19 and further explored the predictive role of NLR for mortality of the COVID-19 patients in more subgroups.

Key epidemiological, clinical, laboratory, radiological and outcomes data were obtained through a detailed medical chart review from January 1st to February 10th, 2020 at the Renmin Hospital of Wuhan University. All the peripheral venous blood samples were collected on admission and were examined at the laboratory following standard procedures. Multivariable logistic regression analyses with the stepwise procedure were used to estimate odds ratios (OR) and 95% confidence intervals (CI). Then, the subgroup and interaction analyses for NLR were conducted according to statistically significant variables in former logistic regression analyses.

A cohort of 140 patients with the confirmed disease was identified. 52 patients had severe diseases and 32 patients eventually died. Compared to the lower NLR group, patients with higher NLR in this study were 29 years older, more likely to have current smoking, had more comorbidities such as diabetes, hypertension, cardiovascular and chronic obstructive pulmonary disease (COPD), and had various symptoms, especially sputum production, headache, upper airway symptoms, and dyspnea (Table 1 ). It was consistent with the baseline characteristic of the study by Liu et al. Table 2 shows the correlation of NLR with severe disease and death in the final analysis. Upon multivariate adjustment, most of the estimated correlations were attenuated. Increased NLR (severity: OR, 8.56, 95% CI, 1.39 - 52.61, p = 0.021; mortality: OR, 1.30, 95% CI, 1.14 - 1.49, p < 0.001) was an independent risk factor for both severity and mortality. Adjusting interacted variables for NLR did not eliminate its significant correlation with the severity risk and mortality risk. Liu et al. reported that NLR exhibited an increase in the risk of mortality for the third tertile (NLR ≧ 4.85). And we further found that the risk of severity and mortality for patients with NLR ≧ 2.14 were significantly higher than those with NLR ≦ 0.48. It suggested that the cut-off point of the NLR could be lower down. For disease severity, age group (≤60, >60; p = 0.053), cancer (p = 0.037) had interactions with NLR. For mortality, age group (p = 0.059), hypertension (p = 0.005), dyspnea (p = 0.006) had interactions with NLR. The correlations of NLR with severity and mortality were statistically significant in subgroups of patients with the age of ≤60 years old, without diabetes, cancer, hypertension, and symptom of dyspnea.

Table 1

The baseline characteristics of patients with COVID-19 by the level of NLR.

Characteristics	All cases	NLR
	(n = 140)	Group 1 (< 2.14, n = 93)	Group 2 (≧2.14, n = 47)
Demographics
Age, median (IQR)	48.00 (29.75)	37.00 (21.00)	66.00 (19.00)
Age range, years
≤30	25 (17.9%)	22 (23.7%)	3 (6.4%)
30–60	79 (56.4%)	62 (66.7%)	17 (36.2%)
>60	36 (25.7%)	9 (9.7%)	27 (57.4%)
Sex
male	51 (36.4%)	35 (37.6%)	16 (34.0%)
female	89 (63.6%)	58 (62.4%)	31 (66.0%)
BMI, median (IQR)	23.45 (2.88)	23.20 (3.30)	23.80 (2.30)
Overweight	46 (38.3%)	33 (35.5%)	13 (48.1%)
Exposure history (yes)	4 (3%)	2 (2.2%)	2 (4.3%)
Current smoking (yes)	8 (6%)	2 (2.2%)	6 (12.8%)
Comorbidity3
Diabetes (yes)	15 (11%)	3 (3.2%)	12 (25.5%)
Hypertension (yes)	32 (23%)	6 (18.8%)	26 (55.3%)
Cardiovascular disease (yes)	16 (11%)	1 (1.1%)	15 (31.9%)
COPD (yes)	6 (4%)	1 (1.1%)	5 (10.6%)
Cancer (yes)	10 (7%)	6 (6.5%)	4 (8.5%)
Chronic liver disease (yes)	2 (1%)	1 (1.1%)	1 (2.1%)
Other disease (yes)	14 (10%)	4 (4.3%)	10 (21.3%)
Clinical symptoms
Fever (yes)	96 (68.6%)	59 (63.4%)	37 (78.7%)
Cough (yes)	84 (60.0%)	54 (58.1%)	30 (63.8%)
Sputum (yes)	13 (9.3%)	1 (1.1%)	12 (25.5%)
Myalgia (yes)	59 (42.1%)	38 (40.9%)	21 (44.7%)
Headache (yes)	28 (20.0%)	10 (10.8%)	18 (38.3%)
Dyspnoea (yes)	45 (32.1%)	21 (22.6%)	24 (51.1%)
UAS (yes)	34 (24.3%)	16 (17.2%)	18 (38.3%)
GIS (yes)	11 (7.9%)	6 (6.5%)	5 (10.6%)
Laboratory findings2
WBC, median (IQR), 109/L	4.66 (2.37)	4.34 (2.00)	5.61 (4.00)
N, median (IQR), 109/L	1.63 (2.30)	1.14 (0.80)	4.39 (3.37)
L, median (IQR)), 109/L	1.87 (2.05)	2.60 (1.71)	0.66 (0.42)
NLR
LDH, median (IQR), U/L	214.5 (113.50)	197.00 (62.00)	293.00 (173.00)
LDH > 250 U/L	43 (30.7%)	16 (17.2%)	27 (57.4%)
CRPR, median (IQR)	0.99 (2.78)	0.48 (1.41)	2.78 (5.56)
Total severity score, median (IQR)	2.00 (3.00)	2.00 (2.00)	7.00 (11.0)
Disease severity
Common type	86 (61.4%)	79 (84.9%)	7 (14.9%)
Severe type	54 (38.6%)	14 (15.1%)	40 (85.1%)
Survival status
Alive	108 (77.1%)	87 (93.5%)	21 (44.7%)
Dead	32 (22.9%)	6 (6.5%)	26 (55.3%)

Abbreviation: COVID-19, coronavirus disease 2019; n, number of cases; BMI, body mass index; IQR, interquartile range; COPD, chronic obstructive pulmonary disease; UAS, upper airways symptoms (sore throat, sneeze, and rhinorrhea); GIS, gastrointestinal symptoms (diarrhea, gastrointestinal discomfort, and loss of appetite); WBC, white blood cell; N, neutrophil; L, lymphocyte; NLR, neutrophil-to-lymphocyte ratio; LDH, lactate dehydrogenase; CRPR, ratio of C-reactive protein (the ratio of CRP value/ upper limit of the CRP value). Exposure history meant Huanan seafood market exposure history.

.1For categorical variables, P values were derived from χ2-test or Fisher's exact test. For continuous variables, P values were derived from Student-t-test or Mann-Whitney U test.

On admission, liver and renal function tests were all found to be within normal range, and thus excluded from the data collection.

Table 2

Final multivariable analyses of the correlation of neutrophil, lymphocyte, and NLR with the severe disease and death of COVID-19.

Characteristics	Severity		Mortality
	OR (95% CI)	P value1	OR (95% CI)	P value1
Final Model 12
N (109/L)	2.27 (1.04–4.93)	0.039	1.06 (0.82–1.36)	0.674
L (109/L)	0.64 (0.25–1.18)	0.155	0.78 (0.56–1.09)	0.146
Final Model 23
NLR	8.56 (1.39 - 52.61)	0.021	1.30 (1.14 - 1.49)	< 0.001
Final Model 34
NLR	1.87 (1.42–2.46)	<0.001	1.24 (1.08–1.41)	0.002
Tertile 1 (≦0.48, as ref.)	1.00		1.00
Tertile 2 (0.49- 2.13)	0.74 (0.18–3.05)	0.382	1.15 (0.12–11.05)	0.902
Tertile 3 (≧2.14)	20.48 (5.87–71.46)	<0.001	11.79 (2.05–67.94)	0.006
Subgroup analyses for NLR5
Age
≤ 60	2.73 (1.28–5.84)	0.010	1.36 (1.10–1.68)	0.005
>60	1.92 (1.00–3.69)	0.049	1.16 (0.99–1.35)	0.067
Fever
No	1.69 (0.92–3.10)	0.088	5.06 (0.16–155.78)	0.354
yes	2.98 (1.52–5.87)	0.002	1.11 (0.96–1.30)	0.171
dyspnea
No	2.77 (1.42–5.41)	0.030	1.80 (1.10–2.94)	0.020
Yes	2.09 (0.86–5.08)	0.105	1.11 (0.96–1.28)	0.160
Hypertension
No	2.34 (1.06–5.16)	0.035	1.53 (1.22–1.92)	<0.001
Yes	2.17 (1.00–4.73)	0.050	1.05 (0.92–1.19)	0.487
Diabetes
No	1.88 (1.18–7.00)	0.020	1.16 (1.00–1.34)	0.050
Yes	/	/	1.34 (0.70–2.58)	0.383
Cancer
No	2.54 (1.43–4.51)	0.001	1.18 (1.02–1.37)	0.030
Yes	/	/	1.51 (0.58–3.92)	0.396

Abbreviation: COVID-19, coronavirus disease 2019; OR, odds ratio; CI, confidence interval; N, neutrophil; L, lymphocyte; NLR, neutrophil-to-lymphocyte ratio.

“/” means that the number of cases in the subgroup is not sufficient for analysis.

Derived from multivariate stepwise analysis of logistic regression model.

Final model 1: for disease severity, retained in this model after the stepwise selection was age (as a continuous variable), fever (yes, no), dyspnea (yes, no), neutrophil (109/L), lymphocyte (109/L), C-reactive protein ration (CRPR); for mortality, retained in this model after the stepwise selection was age (as a continuous variable), neutrophil (109/L), lymphocyte (109/L), CRPR.

Final model 2: for disease severity, retained in the final model was fever (yes, no), dyspnea (yes, no), NLR (as a continuous variable), and CRPR (as a continuous variable); for mortality, retained in the final model was hypertension (yes, no), cancer (yes, no), NLR (as a continuous variable), and CRPR (as a continuous variable).

Final model 3: for disease severity, retained in the final model was age group (<60, ≧60), COPD (yes, no), cancer (yes, no), fever (yes, no), dyspnea (yes, no), and NLR (as a categorical variable); for mortality, retained in the final model was age group (<60, ≧60), hypertension (yes, no), dyspnea (yes, no), CRPR, NLR (as a categorical variable). 5NLR was used as a continuous variable in the subgroup analysis.

For COVID-19, the increased neutrophils indicate the degree of the inflammatory response, and the decreased lymphocytes indicate the degree of immune imbalance. These associations are amplified by the concept of NLR. In severe diseases, the rapid replication of coronavirus induces the delayed IFN response, which sensitizes the T cells to apoptosis. Thus, the virus cannot be cleared in time. A large number of neutrophils and monocyte/macrophages are recruited to the infectious sites, and infiltrate into the lungs. Previous studies of coronavirus suggested that the lymphocyte loss might be associated with the immune-escape mechanism of the virus. Nowadays, it has also been suggested that lymphocyte loss might be associated with the direct infection of lymphocytes by virus or the myelosuppression by antiviral responses. The proliferation of the virus results in the toxic effect on lymphocytes, and the decrease of lymphocytes further weakened the immune response to the virus, forming a vicious circle. Uncontrolled and overreacted immune responses lead to the cytokine storm, causing diffuse alveolar damage or multi-organ failure, finally resulting in death from COVID-19.

The depletion of hematopoietic stem cell bank or immune cell function, which were caused by the aging of the body or a long time of chronic inflammation, were more likely to lead to the cytokine storm when responding to the severe infections. Thus, patients with these risk factors had a higher level of NLR. Liu et al. conducted subgroup analyses by gender, female, body mass index, and the presence of hypertension. They only found that the male had a more significant association with the risk of mortality than the female. Interestingly, we further found that NLR was of greater value in predicting severity and mortality for patients with no other clinical risk factors (i.e., those with a theoretically better prognosis), such as patients with younger age, or without comorbidities.

In conclusion, we found that the level of NLR on admission could be an independent risk factor for the prognosis of COVID-19, not only for the mortality but also for the disease severity. Also, the predictive value was more significant in patients without other potential risk factors. NLR could help physicians rapidly identify high-risk patients and adopt timely intervention, to reduce the rates of severe disease and mortality.

Disclosure

The authors state that they have no conflicts of interest to disclose.

Funding

There was no funding source for this study.