Faced with an emerging COVID-19 pandemic, considerable efforts have been implemented in order to identify clinical and laboratory prognostic factors aiding in triage and resource allocation 40,41.
One such factor, Neutrophils to lymphocytes ratio (NLR), has been extensively studied and yields a good predictive capacity using varying cutoffs. The immune dysregulation that characterizes COVID-19 has been implicated as a possible explanation for its strong predictive value 31. Anticipating the winter and influenza season, we set out to compere the predictive value of NLR across three respiratory viruses; SARS-CoV-2 (The causative agent of the infection named COVID-19), Influenza (A and B) and Respiratory syncytial virus (RSV).
We compared the demographic, laboratory and clinical characteristics between these three respiratory viruses. RSV patients were significantly older (mean 79 years) with more comorbidities (expressed in a mean Charlson score of 5.0) (table 1). Differences in mean body temperature were found to be statistically significant, with slightly higher values for COVID-19 patients, but the difference was not clinically meaningful. Interestingly, oxygen saturation at presentation was similar between the three groups. COVID-19 patients had a unique inflammatory profile, with significantly lower NLR values at presentation but with higher ferritin and triglycerides. CRP levels at presentation showed no significant difference across viruses. COVID-19 patients had higher 30-day mortality rates and a larger proportion of patients required mechanical ventilation. RSV patients had longer hospitalizations and more readmissions, perhaps pertaining to their general frailty (older age and with more comorbidities).
Reviewing the literature we found few small - scale studies comparing laboratory parameters of adult patients with COVID-19 and Influenza 32,42,43 and none compared with RSV. Only two studies investigated the NLR levels in COVID-19 compared to Influenza. One found no difference, and the other found that NLR was significantly lower in the COVID-19 patients (median of 2.6), as was described in our report 32,44. In all reports, as in ours, CRP levels at admission were similar across the viruses.
Focusing on the prognostic value of NLR, we directly compared the association between NLR and a composite adverse outcome score between the three viruses. Only in the COVID-19 group poor clinical outcome was associated with higher NLR. Furthermore, when performing ROC analysis, only in the COVID-19 group NLR had a discrimination ability between patients with poor and favorable outcomes. Taken together, these results strongly suggest that NLR is a more valuable prognostic marker in COVID-19 compared to Influenza and RSV. This is opposed to former reports, which identified the role of NLR as a predictor of poor prognosis in Influenza patients 15-18.
The application of NLR as a prognostic factor in COVID-19 has been extensively studied. A high NLR implies an aberrant immune response, with increased neutrophils and decreased lymphocytes 31. Neutrophil production can be triggered by virus-related inflammatory factors, such as interleukin-6 and interleukin-8, tumor necrosis factor-α, granulocyte colony stimulating factor, and interferon-γ. Lymphopenia is common in COVID-19, as a result of direct cytokine - induced inhibition 45. The massive over - production of these cytokines, or "cytokine storm", can result in acute respiratory distress syndrome (ARDS), the hallmark of severe cases of COVID-19 46. In one cohort, all COVID-19 patients with ARDS had an aberrant immune response [either macrophage activation syndrome (MAS) or "immune paralysis" (depletion of HLA-DR expression)] 47. These patients had a sustained production of IL-6 and TNF-α, which was unique when compared to a retrospective cohort of influenza H1N1 2009 patients. Moreover, in vitro, higher levels of IL-6, TNF- α and IL1-β were released from monocytes derived from COVID-19 patients compared to H1N1 2009 patients.
In contrast to previous reports, our results did not demonstrate a prognostic value for NLR in influenza patients. This discrepancy may possibly be attributed to the small number of patients reported in previous studies.
In RSV infection, NLR has not been hitherto studied. Current understanding of the immune response to RSV infection holds that severe cases are initially associated with neutrophilia and lymphopenia and thus would be expected to be associated with a high NLR.
Following the neutrophilic response, RSV infection is associated with a pulmonary CD8 T-cell response and lymphocytes levels rise, coinciding with viral clearance 48. Thus, theoretically, a low NLR would be expected to imply a favorable prognosis. Nonetheless, in our real-life, large cohort of RSV patients, NLR at admission did not have prognostic value.
In order to determine the applicable threshold for high NLR we used a ROC curve. The optimal NLR threshold at 4.7 had the highest sensitivity and specificity for differentiating favorable and poor outcome. This result is in concordance with previous studies, in which the proposed optimum cutoff values ranged from 3 to 6. Two studies stratified the prognostic capacity of NLR by age 21,49. In our study, a logistic regression model was used in order to test the discrimination ability of NLR (above or below 4.7) in predicting clinical outcomes adjusted to age, sex and Charlson comorbidity score. Our results show an OR of 1.6 for NLR and a total AUC of 0.74 (figure 3). This reinforces the conclusion that NLR is a valuable prognostic marker in COVID-19 patients, even after stratification according to sex, age and comorbidities.
Our study has several limitations. First, since it is a retrospective observational study, subject selection bias was unavoidable. Second, the COVID-19 pandemic resulted in the referral of patients with minimal upper respiratory symptoms and general good health to the ER. Thus, the COVID-19 cohort included a large subset of relatively healthy subjects with a mild disease skewing NLR levels towards a lower baseline. The respective RSV or influenza patients in past years were probably treated in the community and their data is absent or alternatively, presented to the ER later in the course of their disease, at a worse clinical stage. This poses an additional limitation, the timing of the first complete blood count, which was performed at an earlier stage among COVID-19 patients.
In conclusion, we present the first large scale study to compare laboratory markers and outcome between COVID-19, influenza and RSV. We show that COVID-19 patients had a lower NLR at admission. We calculated a NLR cutoff of 4.7 which was able to differentiate between poor and favorable outcome. Finally, when comparing the three viruses, only in COVID-19 NLR had a prognostic value. Thus, during the upcoming winter months, physicians faced with the daunting challenge of treating and triaging patients presenting with acute respiratory infections, may benefit from the implementation of this readily accessible tool in their clinical decision – making.