The Prognostic Nutritional Index is associated with mortality of COVID‐19 patients in Wuhan, China

Abstract Background Declared as pandemic by WHO, the coronavirus disease 2019 (COVID‐19) pneumonia has brought great damage to human health. The uncontrollable spread and poor progression of COVID‐19 have attracted much attention from all over the world. We designed this study to develop a prognostic nomogram incorporating Prognostic nutritional index (PNI) in COVID‐19 patients. Methods Patients confirmed with COVID‐19 and treated in Renmin Hospital of Wuhan University from January to February 2020 were included in this study. We used logistic regression analysis to find risk factors of mortality in these patients. A prognostic nomogram was constructed and receiver operating characteristics (ROC) curve was drawn to evaluate the predictive value of PNI and this prognostic model. Results Comparison of baseline characteristics showed non‐survivors had higher age (P < .001), male ratio (P = .038), neutrophil‐to‐lymphocyte ratio (NLR) (P < .001), platelet‐to‐lymphocyte ratio (PLR) (P < .001), and PNI (P < .001) than survivors. In the multivariate logistic regression analysis, independent risk factors of mortality in COVID‐19 patients included white blood cell (WBC) (OR 1.285, P = .039), PNI (OR 0.790, P = .029), LDH (OR 1.011, P < .015). These three factors were combined to build the prognostic model. Area under the ROC curve (AUC) of only PNI and the prognostic model was 0.849 (95%Cl 0.811‐0.888) and 0.950 (95%Cl 0.922‐0.978), respectively. And calibration plot showed good stability of the prognostic model. Conclusion This research indicates PNI is independently associated with the mortality of COVID‐19 patients. Prognostic model incorporating PNI is beneficial for clinicians to evaluate progression and strengthen monitoring for COVID‐19 patients.


| INTRODUC TI ON
The Corona Virus Disease 2019 (COVID- 19), initially found in Wuhan, China, spread rapidly around the world and becomes a serious global public health issue. Mainly manifested as fever, cough, and fatigue, nearly half of COVID-19 patients would develop dyspnea with concurrent hypoxia one week after onset. [1][2][3] In addition to impaired respiratory function, function of other organs could also be damaged. Complications including cardiac injury, acute kidney injury, acute gastrointestinal injury, coagulopathy, and liver dysfunction are relatively common in critically ill cases 4,5 and were confirmed associated with poor outcome in COVID-19 patients. [6][7][8][9][10] These organs damage is considered resulting from the cytokine release syndrome (CRS) which plays pivotal role in the progression of COVID-19 patients. 11,12 One of the most core cytokines in CRS is the Interkulin-6 (IL-6), which has been acknowledged playing an important role in acute inflammation of various diseases. [13][14][15] The release of excessive cytokines including IL-6 in COVID-19 patients is attributable to the activation of innate and adaptive immune system caused by SARS-CoV-2. 16 The dysregulation of immune response and excessive inflammation actually is key element of pathogenesis in COVID-19. 17,18 And many immunity and inflammation-related markers including C-reactive protein (CRP), IL-6, neutrophil-to-lymphocyte ratio (NLR), and platelet-to-lymphocyte ratio have been confirmed associated with disease severity and mortality of COVID-19 patients. [19][20][21][22][23] The Prognostic Nutritional Index (PNI), a common marker of immune and inflammatory status, has been proved of prognostic value in various clinical settings including cardiovascular diseases, infectious diseases, and cancer. [24][25][26][27][28][29] Incorporating effects of both lymphocyte and albumin, low PNI could indicate poor prognosis of patients. We designed this study to explore the prognostic value of PNI in COVID-19 patients.

| Subjects
Patients admitted to Renmin Hospital of Wuhan University for COVID-19 from January 30 to February 24, 2020 were eligible in this study. The diagnose of COVID-19 patients was confirmed by the positive result for SARS-Cov-2 RNA in nasopharyngeal swabs by using real-time fluorescence reverse transcription-polymerase chain reaction (RT-PCR). Patients died on admission and transferred from other hospitals were excluded from this study. Finally, a total of 450 patients were included in this single-center study.

| Data collection
Demographical and clinical data of included patients were collected by searching records in electronic medical record system (EMRS).
Complicated underlying diseases in admission including hypertension, cardiovascular disease, chronic respiratory or liver disease, and cancer were included as potential risk factors in this study. Results of laboratory tests were obtained by analyzing the blood sample collected on admission. The PNI was calculated as serum albumin (g/L) + 5 × lymphocyte count (10 9 /L). In addition, neutrophil-tolymphocyte ratio (NLR) and platelet-to-lymphocyte ratio (PLR) were also calculated and included as potential risk factors. The primary outcome of this study was in-hospital mortality acquired by following up from admission to discharge. This study was approved by the ethics committee of West China hospital of Sichuan University and Renmin Hospital of Wuhan University. The whole process of this study was accorded with the Declaration of Helsinki. Patients included in this observational study have signed an informed consent.

| Statistical analysis
We used Kolmogorov-Smirnov test to verify the normality of variables. Normally distributed variables were shown as mean ± standard deviation while non-normally distributed variables were shown as median (interquartile range). And categorical variables were presented as the form of numbers (percentage). We respectively performed independent Student's t test and Mann-Whitney U test to analyze differences between two groups of normally distributed and non-normally distributed variables. Chi-square test was performed to examine the difference of categorical variables. Then, univariate and multivariate logistic regression were sequentially performed to explore risk factors of mortality in COVID-19 patients. By multivariate logistic regression, we developed a prognostic nomogram using the rms package in R project. The receiver operating characteristic (ROC) curves were drawn, and area under the ROC curves (AUC) were calculated to evaluate the discrimination ability of PNI and the prognostic nomogram. Finally, we evaluate the stability of the prognostic nomogram by internal validation with 1000 bootstrap samples. Calibration plots were drawn to analyze the consistency between observed probability and predicted probability of poor outcome in COVID-19 patients (Figure 1).
A P value <.05 was considered to be statistically significant.

| D ISCUSS I ON
The mortality rate of previous studies reported ranged from 1% to 28.3%. 1,5,30,31 In this study, there were 78 patients suffered poor outcome with mortality rate of 17.3%. This difference might be attributable to the heterogeneity of included patients, differences in medical treatment level and medical resources. Our results showed non-survivors had older age, higher male ratio, and higher incidence of comorbidities. And underlying diseases including hypertension, cardiovascular disease, and chronic respiratory disease were found associated with mortality in univariate logistic regression analysis.  Length of hospital stay 9 (5-13) 9 (5-14) 6 (4-10) <.001 Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; APTT, activated partial thromboplastin time; AST, aspartate aminotransferase; BUN, blood urea nitrogen; CRP, C-reactive protein; FIB, fibrinogen; INR, international normalized ratio; LDH, lactate dehydrogenase; MAP, mean arterial pressure; NLR, neutrophil-to-lymphocyte ratio; PCT, procalcitonin; PLR, platelet-to-lymphocyte ratio; PNI, prognostic nutritional index; PT, prothrombin time; TT, thrombin time; WBC, white blood cell.
These findings were consistent with results of previous studies. [32][33][34] Moreover, higher NLR and PLR were found positively associated with mortality in univariate analysis. However, after adjustments, only WBC, PNI, and LDH were independently correlated with outcome of COVID-19 patients in multivariate logistic regression analysis.
The PNI, calculated from albumin and lymphocyte levels, is an objective reflection of inflammatory and nutritional status. And it has been confirmed being of prognostic value in various settings such as cardiovascular disease and cancer. 24,27,35 In our study, the level of albumin was significantly lower in non-survivors compared with survivors.
And previous studies have shown that albumin level was inversely associated with unfavorable progression and outcome in COVID-19 patients. 36,37 Low level of albumin in non-survivors might be attributable to intubation induced inadequate intake, reduced synthesis caused by liver dysfunction and increased consumption due to organ damage.
The correlation between poor outcome and low albumin level could be mediated by several mechanisms. Firstly, synthesized by hepatocytes, albumin level is an indicator of liver function. Inflammatory cytokines such as interkulin-6 (IL-6) and tumor necrosis factor-α (TNF-α) could inhibit synthesis ability of hepatocytes so that the serum level Abbreviations: ALP, alkaline phosphatase; ALT, alanine aminotransferase; APTT, activated partial thromboplastin time; AST, aspartate aminotransferase; BUN, blood urea nitrogen; CRP, C-reactive protein; FIB, fibrinogen; INR, international normalized ratio; LDH, lactate dehydrogenase; NLR, neutrophil to lymphocyte ratio; PLR, platelet to lymphocyte ratio; PNI, prognostic nutritional index; PT, prothrombin time; WBC, white blood cell.

TA B L E 3 Predictive value of PNI and the Prognostic model
of albumin decreases. 38 It is the cytokine storm, which is character- patients. 44 It is speculated that direct attack from virus to lymphocyte, antigen presenting cells (APC) dysfunction and apoptosis due to excessive release of cytokines could result in the decrease of T cells. [45][46][47] Whatever, the lymphopenia has been confirmed as an independent risk factor of mortality in COVID-19 patients. 47 And other inflammatory markers incorporating lymphocyte such as NLR and PLR are also associated with severity and outcome in COVID-19 patients. 19,20,48 The decreased lymphocyte might be considered as a reflection of impaired immune function and sharply increasing cytokines. The PNI, composed of albumin level and lymphocyte count, could reflect nutritional and inflammatory status more comprehensively in COVID-19 patients.
The WBC count and LDH were another two significant factors in multivariate logistic regression analysis. Both of them were significantly higher in non-survivors than survivors. patients. 43 And serum LDH level is significantly correlated with indicators of inflammation, cardiac and liver injury such as AST, CRP, and brain natriuretic peptide (BNP). 43 Therefore, the LDH level could

| CON CLUS IONS
The PNI is inversely associated with outcome in COVID-19 patients.
Prognostic model incorporating PNI shows good performance in predicting outcome of COVID-19 patients. The nomogram of our model provides physicians with visual prognostic assessment for COVID-19 patients.

CO N FLI C T O F I NTE R E S T
The authors have no conflicts of interest to disclose.