Dynamic changes in serum IL-6, IL-8, and IL-10 are associated with the outcome of patients with severe COVID-19 in ICU

Background: Biomarkers that would help prognosticate outcomes and guide treatment of patients with severe coronavirus disease 2019 (COVID-19) are currently required. We aimed to investigate whether the dynamic variation of cytokines was associated with the survival of patients admitted to an intensive care unit (ICU). Methods: A retrospective study was performed on 40 patients with COVID-19 admitted to an ICU in Wuhan, China. Demographic, clinical, and laboratory variables were collected, and serum cytokines were kinetically assessed. A multivariable- adjusted generalized linear regression model was used to evaluate the differences in serum cytokine levels between survivor and non-survivors. Results: Among the 40 patients included, a signicant positive correlation was found between multiple cytokines. Serum levels of IL-6, IL-10, and tumor necrosis factor alpha in non-survivors were consistently elevated compared to that of the survivors. Kinetic variations of IL-6, IL-8, and IL-10 were associated with a fatal outcome in severe patients with COVID-19, independent of sex, age, absolute lymphocyte count, direct bilirubin, hypertension, chronic obstructive pulmonary disease, and cancer. Conclusion: Dynamic changes in serum IL-6, IL-8, and IL-10 levels were associated with survival in ICU and could serve as a predictive biomarker in patients with severe COVID-19 to determine therapeutic options. early of disease progression in severe COVID-19.


Introduction
The pandemic caused due to coronavirus disease 2019 (COVID- 19), has led to more than 30 million infections and 946,000 deaths worldwide. The pathogenesis and effective treatments for the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection are currently less known.
The survival rate of patients with severe COVID-19 in an intensive care unit (ICU) is extremely lower than that of patients with mild COVID-19 patients. Hyperin ammation and pro-in ammatory cytokines have been considered to be key factors in the clinical course of acute respiratory distress syndrome and multiple organ dysfunction syndrome in severe patients with COVID-19 [1] Studies have demonstrated a correlation between clinical and laboratory characteristics and the outcomes in patients with COVID-19 [2,3]. Clinical trials performed on patients with COVID- 19, involving targeting cytokine signals, including interleukin (IL)-6, IL-1, granulocyte-macrophage colony-stimulating factor (GM-CSF) antagonists, or small molecules, have reported bene ts or failure [4][5][6][7][8][9][10]. Studies showed that serum cytokine pro le predicted COVID-19 severity [11,12].
However, little is known about the correlation between the dynamics of serum cytokine alterations and the different prognoses of patients with severe COVID-19 during their hospitalization. More evidence is needed to determine whether cytokine pro ling serves as an optimal biomarker to study in ammation processes and predict the outcome in severely ill patients with heterogeneity [13]. In this retrospective study, we explored the relationship between the kinetic variations of cytokines and the clinical outcomes in patients with severe COVID-19 admitted to an ICU.

Data collection
Demographic, clinical, laboratory parameters including cytokine test, and outcome data of the ICU patients were obtained from electronic medical records, and the data were checked by two physicians. The normal concentration ranges for IL-1β, IL-8, IL-10, and tumor necrosis factor alpha (TNFα) were below 5 pg/mL, 62 pg/mL, 9.1 pg/mL, and 8.1 pg/mL, respectively. The normal concentration range of IL-6 is between 0-7 pg/mL and that of normal IL-2 receptor (IL-2R) ranges from 223 to 710 U/mL. The sera of patients were repeatedly assessed and further classi ed into four groups (≤3d, 4-7d, 8-13d, and ≥14d) according to the length of stay in the ICU, for further analyzing the dynamic variation of cytokines.

Statistical analysis
We performed all statistical analyses using SAS version 9.4 (SAS Institute, Inc, Cary, USA). The formal hypothesis testing performed was two sided with a signi cance level of P <0.05. Classi cation variables were expressed as percentage, and the differences between two groups were tested using chi square test. The quanti ed variables were expressed as mean ± standard deviation, and the differences were analyzed using t-test for normal distribution, while the data were expressed in median and interquartile range (IQR) for abnormal distribution, and the differences were analyzed using the Kruskal-Wallis H test. We used a multivariable-adjusted generalized linear regression model to evaluate the differences in serum cytokine levels between survivor and non-survivor patients, after adjusting for the covariates that were selected using univariate analysis. Figures were drawn using GraphPad Prism version 8.0 (GraphPad Software, San Diego, California, USA).

The correlation between time-dependent cytokines and outcomes
We analyzed the relationship between cytokines and explored the effect of longitudinal change of cytokines on the outcomes of severely ill patients in ICU. The concentrations of serum cytokines including IL-1β, IL-2R, IL-6, IL-8, IL-10, and TNFα were measured at different points of time since the patients were admitted to the ICU (Table 2). Among them, IL-1β was observed to be within the normal range in almost all of the patients; this was in agreement with the previous studies. We analyzed the correlation between six types of cytokines and found a signi cant positive correlation (Table 3). IL-2R had the highest correlation with TNFα (r = 0.69652, P < 0.0001) and IL-6 showed a strong correlation with IL-10 (r = 0.63088, P < 0.0001). The dynamic changes in cytokine pro les of survivors and non-survivors are illustrated in Figure 1. The serum levels of IL-6, IL-10, IL-8, and TNFα in non-survivors showed a gradual increase compared with that of the survivors during the ICU stay. Upon admission to the ICU, the serum levels of IL-6 was signi cantly high in non-survivors. The serum levels of IL-10 showed statistically different between survivors and non-survivors after three days committed to ICU. The IL-8 serum levels were statistically different in two weeks of hospitalization (P < 0.05).
3.3 IL-6, IL-8, and IL-10 are associated with the survival of patients with severe COVID-19 admitted in ICU To further explore the role of cytokines in the outcome of severe patients with COVID-19, we used a generalized linear regression model to analyze the differences in serum cytokine levels between two outcome groups (survivor vs. non-survivor). The following factors were considered for analysis: age, gender, absolute neutrophil count, absolute lymphocyte count, direct bilirubin, lactate dehydrogenase, blood urea nitrogen, C-reactive protein, and D-dimer, as well as the history of hypertension, diabetes mellitus, chronic obstructive pulmonary disease, and cancer [13,14]. By using univariate analysis, we found that age (P < 0.001), absolute lymphocyte count (P = 0.005), direct bilirubin (P < 0.001), history of hypertension (P < 0.001), chronic obstructive pulmonary disease (P = 0.002), and cancer (P < 0.001) were associated with the outcome and were included in the next analysis (Table 4 and Online Resource ). We further found that IL-6, IL-8, and IL-10 were associated with the outcome in patients with severe COVID-19 after factoring sex, age, lymphocyte count, direct bilirubin, history of chronic obstructive pulmonary disease, cancer, and hypertension into the next analysis (Table 5).

Page 4/9
The prevalence of COVID-19 and the resulting high fatality in the ICU are major problems. Finding an ideal therapeutic strategy for patients in the ICU has been a challenge under the current circumstances. Identifying biomarkers that are produced during the course of the disease would be helpful for guiding the treatment and supervising the response. Cytokine storm syndrome (CSS) or hyperin ammation has been recognized in the pathogenesis of COVID-19. Several studies have shown that different in ammatory cytokines are correlated to the severity of the disease and survival of patients [15,16]. This study provided evidence that the kinetic variation of IL-6, IL-8, and IL-10 was closely associated with the prognosis in patients with severe COVID-19, independent of other risk factors including demographics, comorbidities, and common laboratory markers.
It has been reported that IL-6 is a potent predictive biomarker in patients with COVID-19 [11,[17][18]. IL-6, produced by multiple cell types including monocytes and macrophages, is shown to be involved in in ammation and immune response and plays a role in CSS or cytokine release syndrome (CRS). Blocking IL-6 has been a main focus in mitigating cytokine storm-related symptoms in patients with severe COVID-19. Nevertheless, regarding the clinical symptoms of COVID-19, including lung injury and the circulation of various cytokines, the cytokine pro les in patients with COVID-19 were distinct from those of typical cytokine storm including CRS, hemophagocytic lymphohistiocytosis, or macrophage activation syndrome [19,20]. IL-6 antagonists, which have been approved for treating chimeric antigen receptor T cell-induced CRS, have demonstrated various effects on patients with COVID-19 in multiple studies [21]. It is di cult to draw conclusions because of the variable participants and intervention criteria. As the most common adverse event, serious infections were additional concerns. Soners et al. found that tocilizumab is associated with an increased proportion of superinfections in mechanically ventilated patients with COVID-19 [22]. It is of importance that who might bene t from these therapy regimes considering heterogeneity of patients with COVID-19. Rossotti et al. performed a retrospective, single-center analysis on 74 patients treated with tocilizumab and 148 matched controls [23]. They found that a steep increase in IL-6 levels in patients indicated a poor clinical outcome. Similarly, Quartuccio et al. showed that out of 24 patients who received tocilizumab treatment, 6 patients died, and 24-48 h post-tocilizumab treatment, IL-6 serum levels were signi cantly higher in non-survivors than in survivors [24]. Two studies found that low levels of IL-6 at baseline were associated with a response to sarilumab or tocilizumab treatment [25,26]. Together with our results, supervising the dynamic changes in serum IL-6 levels might offer an opportunity to de ne inclusion population, predict the prognosis, and adjust the treatment regime in clinical trials or off-label treatment with IL-6 antagonists.
We found that continuous elevated level of serum IL-8 in non-survivors compared to the survivors despite statistic difference was only observed in patients after two weeks' hospitalization. IL-8 is normally secreted by multiple cells exposed to stimuli and is considered to be the primary molecules of acute in ammation. Increased neutrophils have been thought to correlated to the severity of patients with COVID-19 in the previous clinical observation [27]. IL-8 has the effect on potent neutrophil chemotactic and promotion of angiogenesis, might result in deterioration of COVID-19. IL-6 signaling promotes IL-8 recruitment and enhances the in ammation cascade.
We found that the increase in serum levels of IL-10, which is normally considered an anti-in ammatory molecule, was correlated with the increase in levels of pro-in ammatory factors including IL-6 and was also a predictor of survival in the ICU patients regardless of other severity factors. SARS-CoV-2 invasion is a trigger that induces a broad range of immunological events. High levels of circulating pro-in ammatory cytokines can lead to tissue damage and multi-organ failure. As a result, a negative feedback loop might be activated to alleviate the response and promote the transition from immune hyperactivity into the resolution phase. Studies show that anti-in ammatory and pro-in ammatory reactions arise concomitantly in patients with sepsis [28,29]. A shortage of IL-10 is indicated by the sustained elevation of TNF, IL-12, and interferon gamma levels in infected mice [30]. Therefore, the increased levels of circulating IL-10 in patients with severe COVID-19 suggested that an immune regulatory network was activated to limit the magnitude of the immune response and repair the damage. The combination of pro-in ammatory and anti-  [31]. In another study that enrolled patients with moderate and severe COVID-19, IL-6 and IL-10 were used as predictors for recognizing patients at high risk of deterioration [12].
Our study is the rst to show that the dynamic variation of speci c cytokine pro les is correlated with the outcome of patients with COVID-19 admitted in the ICU. Limitations of this single-center study involve the small sample size and missing laboratory parameters that may induce potential bias. We observed the survival of patients during their ICU stay; this might slightly differ from the nal clinical outcome. Notwithstanding, our results signi ed that circulating cytokine biomarkers might prove useful for evaluating therapeutic response and optimizing the therapeutic strategies in patients with severe COVID-19.