Corona Virus Disease 2019 (COVID-19) is the third fatal coronavirus infection of the 21st century, which is significantly more lethal than the previous Severe Acute Respiratory Syndromes (SARS) and Middle East respiratory syndrome¹. The World Health Organization reported 29.37 million new cases confirmed in the third week of December, 2022². As of May 3, 2023, the global count of confirmed cases has exceeded 760 million, with over 6.92 million deaths.² Due to the high infectivity and high case fatality rate of COVID-19, people throughout the world are affected by the pandemic. COVID-19 has now become a major public health problem.

Studies have demonstrated an association between nutritional status and mortality from COVID-19 infection³. Malnutrition is an risk factor that negatively affects the clinical outcomes of patients, being associated with an increased risk of adverse events, increased hospital stay and higher mortality, especially in elderly people⁴. However, there is no clinically clear screening tool for the nutritional status of COVID-19.

The use of a blood-based objective malnutrition index can overcome the limitations of subjective malnutrition screening tools. The prognostic nutritional index (PNI) and controlling nutritional status score (CONUT) can be computed based on three conventional blood parameters: peripheral blood lymphocytes, serum albumin, and total cholesterol. Originally employed to evaluate surgical risk and predict the nutritional and immune status of surgical patients⁵, other studies have found a correlation between PNI and CONUT and the prognosis of patients with malignant tumors and heart failure^6,7.

Therefore, this study retrospectively explores the the influence of nutritional status on the outcome of elderly COVID-19 patients based on PNI and CONUT, while investigating the value of PNI and CONUT in predicting elderly COVID-19 hospital new outcomes. Meanwhile, this study will provide reference opinions for clinicians to assess the nutritional status of COVID-19 patients and help reduce the poor prognosis of elderly COVID-19 patients.

Data source

In this single-center retrospective study, we obtained data from the Big Data Center Information Management System of a third-class hospital. We collected information and hospital new outcomes on patient sex, age, admission time, underlying disease, surgical status, hospital duration, hospitalization details, discharge outcome, medical records and so on. In addition, laboratory test data, including serum albumin levels and total lymphocyte counts from the first test after admission, were included.

Filters for patient selection included the following criteria: ① Patients diagnosed with novel coronavirus first infection during hospitalization from November 1, 2022, to January 31, 2023; ② According to the COVID-19 Diagnosis and Treatment Protocol (Trial Ninth Edition), patients were confirmed as COVID-19;③ Age ≥ 60 years old. Exclusion criteria: ① Long-term inpatients who were admitted before November 2022; ② Patients with severe loss of clinical data or laboratory diagnostic data; ③ Patients with extremely abnormal clinical data or laboratory diagnostic data; ④ Patients who were still not discharged as of the statistical date (January 31, 2023); ⑤ Patients who are not first-time infected. ⑥Patients who potentially have clinical conditions that alter serum albumin, total lymphocyte count, and total cholesterol, such as the use of statins, patients with advanced liver disease, or patients with lymphoma/leukemia, excluded.

This study included COVID-19 patients admitted to the Second Affiliated Hospital of Nanchang University from November 1, 2022, to January 31, 2023. After applying the Not Applicable (Na) exclusion criteria to exclude duplicate and incomplete information cases, 4241 patients were finally included for retrospective analysis.

This study has been approved by the Ethics Committee of the Second Affiliated Hospital of Nanchang University.

Patients were screened for malnutrition employing PNI and CONUT. The CONUT score was computed based on serum albumin level, total cholesterol, and lymphocyte count, with scores of 0 to 1 classified as no malnutrition, 2–4, 5–8, and 9–12 indicating mild, moderate, and severe malnutrition, respectively. The PNI score was computed using the formula PNI = serum albumin (g/L) + 5 total lymphocyte count (10⁹/L), with scores > 38 considered free of malnutrition, 35–38 and < 35 representing moderate and severe malnutrition, see Table A1 in Appendix A.

According to the fifth edition of the National Health Council COVID-19 management guidelines, the severity of inpatient outcomes in hospitalized COVID-19 patients was defined. Patients that exhibited any of the following characteristics were considered critical: respiratory failure requiring mechanical ventilation, shock, organ dysfunction, or the need for admission to the intensive care unit.

According to the above standards, patients were classified as “hazardous” and “noncritical” in this study.

We employed Python version 3.7 and R version 3.6.3 statistical software for statistical processing. In addition, we expressed the data with a normal distribution as mean ± standard deviation (x ± s) and performed two independent sample t-tests to compare groups. Continuous variables that did not follow a normal distribution were expressed as median and interquartile spacing [M (Q1, Q3)], and the groups were compared using the Mann–Whitney U test. Categorical variables were expressed as frequency (%) and comparison between groups was performed using the χ² test.

According to the in-hospital outcome, we conducted a univariate analysis of the basic characteristics of COVID-19 patients with different outcomes, multi-model multivariate adjustment analysis, and restricted cubic spline comparative analysis to clarify the impact of nutritional status on the outcome of COVID-19 patients. In the multivariate regression analysis model, the variables were selectively included. In Model 1, only nutrition was included without adjusting for other factors, while age and sex were adjusted in Model 2. Model 3 included all other univariate analysis of meaningful variables. Furthermore, we assessed the highest predictive value according to the subject characteristic (ROC) curve. Restricted cubic spline regression (RCS) was used to flexibly model and visualize the nonlinear relationship between PNI/CONUT and COVID-19 critical hospitalization outcome. All hypothesis tests considered two-sided p<0.05 as a statistically significant difference.

Baseline characteristics

This study included 4241 hospitalized COVID-19 patients, 2507 (59.1%) were male, the mean age of the patients was 67 years and the median hospital stay was 9 days, with 437 (10.3%) in the critical group.The findings demonstrated that compared with the noncritically ill group, the critically ill group exhibited older age, and the differences in the proportion of men, the proportion of combined underlying diseases, and the proportion of surgeries were all statistically significant (p<0.05). Additionally, the PNI scores were significantly lower while the CONUT scores were higher in the critical group than in the noncritical group(Table 1).

The findings of the first laboratory test demonstrated lower levels of red blood cell count, lymphocyte count, platelet count, total protein, albumin, hemoglobin, and total cholesterol, while higher levels of sodium, potassium, chloride, white blood cell count, neutrophil count, total bilirubin, creatinine, and lactate dehydrogenase were observed. These differences between the two groups were statistically significant (p<0.05)(Table 2).

In Model 1, only nutrition was included without adjusting for other factors, while age and sex were adjusted in Model 2. Model 3 included surgical treatment, underlying disease, sodium, potassium, chlorine, hemoglobin, red cell count, white cell count, platelet count, neutrophil count, total protein, total bilirubin, creatinine, and lactate dehydrogenase. The findings demonstrated that severe malnutrition, as indicated by PNI (adjusted odds ratio (OR) 4.07,95%CI = 2.79–5.92) and CONUT (adjusted OR 10.00,95%CI = 4.54–22.01), was a significant factor in the multivariate analysis(Table 3).

PNI, CONUT, and their independent risk factors

We investigated the ability of PNI, CONUT, and each of its components (serum albumin, total cholesterol, and total lymphocyte count) to predict COVID-19 outcomes using the area under the ROC curve (AUC). The AUCs for PNI, CONUT, serum albumin, total cholesterol, and total lymphocyte counts were 0.817, 0.804, 0.773, 0.702, and 0.774, respectively (Fig. 1). Notably, the AUC of PNI was significantly higher than that of the other four components, and a PNI < 38.04 was the best threshold for predicting critical outcomes in COVID-19 inpatients, with a sensitivity of 70.7% and a specificity of 79.6%.

ROC curves of PNI and CONUT and their components for predicting critical illness.

Full size image

Restrictive cubic splines of PNI and CONUT

A restrictive cubic spline was conducted to further investigate the association between the two nutritional scores (considered as continuous variables) and inpatient outcomes of COVID-19.The findings indicated a linear relationship between the OR of PNI and CONUT (with P-values of 0.207 and 0.171, respectively). There was a negative dose-response relationship between PNI and COVID-19 critical hospitalization outcome, showing a trend of rapid decline first and then slow decline, that is, the higher the PNI level, the lower the risk of critical hospitalization outcome. A positive dose-response relationship was observed between CONUT and COVID-19 critical hospitalization outcome, with a trend of slow rise first and then rapid rise, that is, the higher the CONUT level, the higher the risk of critical hospitalization outcome. This trend is continuous throughout the range, with no apparent turns.At the reference point (OR = 1), PNI was 40.72 and CONUT was 6, respectively(Fig. 2).

Restrictive cubic spline between PNI, CONUT, and hospital outcomes. Notes: The odds ratio (OR) is represented by the red solid line, and the 95% confidence interval (CI) by the red shaded area.

Full size image

In this study, the critically ill patients were older compared with the noncritically ill group, aligning with similar observations in the US, Saudi Arabia, and Japan^8,9,10. This correlation may be attributed to the age-related decline in the function of immune system caused by immune senescence. Moreover, males have higher rates of critical illness because of weaker immunity in these patients¹¹. The reduced female susceptibility to viral infection can be attributed to protection from the X chromosome and sex hormones, which play an important role in both innate and adaptive immunity¹². Nutrition is a factor that determines the immune response and can strongly influence the infection trajectory by improving or inhibiting the immune system¹³. Maintaining good nutrition is crucial for supporting the immune response¹⁴. In this study, approximately 34.5% of hospitalized patients in this study showed malnutrition, which is similar to the conclusion reached in the study by Lucie Allard et al. (38.9% of COVID-19 inpatients reported malnutrition).¹⁵ The critically ill group exhibited a more unfavorable nutritional status compared with the noncritical group, and it was observed that nutritional risk was associated with severe COVID-19, probably as both a cause and outcome¹⁵.

The etiology of malnutrition is complex and multifactorial. Malnutrition associated with disease or injury typically comprises reduced food intake or assimilation and varying degrees of acute or chronic inflammation, leading to changes in body composition and decreased biological function^16,17,18. In COVID-19 patients, the origin of malnutrition may be due to reduced food intake, increased digestive and skin loss (diarrhea, vomiting, sweating), olfactory and taste dysfunction, and high catabolism, resulting from high levels of inflammation and muscle atrophy¹⁹. When SARS-CoV-2 infects the respiratory tract, it can lead to mild respiratory infection or severe ARS, which triggers an inflammatory response²⁰. Excessive inflammation, cytokine storms, acute respiratory distress syndrome, and damage to the lungs, heart, and kidneys characterize severe COVID-19³. Huang et al. discovered that SARS-CoV-2 attacked the mucosal epithelium, resulting in gastrointestinal symptoms and worsening the nutritional status of elderly patients²¹. What’s more, the development of malnutrition in the elderly is also most likely facilitated by the aging process²².

Insufficient nutrition due to malnutrition may lead to inflammation, oxidative stress disorders and inadequate availability of optimal functioning immune system, resulting in increased susceptibility to infection and inability to control infection, ultimately leading to adverse consequences such as frailty and infection²³.Malnutrition in the elderly presents with involuntary weight loss or low body mass index, and a higher risk of malnutrition is associated with decreased muscle mass at admission in elderly inpatients²⁴. The reduction in lean body mass usually represents no metabolic reserve in the elderly. In addition to undernutrition, overnutrition, represented by obesity and overweight, constitutes adverse factors for novel coronavirus infection in older patients, and a higher Body Mass Index (BMI) in older patients is associated with negative COVID-19 outcomes²³.

According to the laboratory test results, the critically ill COVID-19 inpatients had lower serum albumin, total cholesterol, and lymphocyte values. Serum albumin, particularly in Chinese elderly individuals, serves as an indicator of nutritional status²⁵. Hypoalbuminaemia is closely associated with adverse clinical outcomes²⁶. In severe COVID-19 patients, SARS-CoV-2 may not only trigger an antiviral immune response, but may also lead to excessive production of proinflammatory cytokines, leading to an uncontrolled inflammatory response. This condition can result in lymphopenia and lymphocyte dysfunction, rendering patients more susceptible to infection, septic shock, and severe multiorgan dysfunction²⁷. Lymphopenia is a crucial factor in disease severity and is a characteristic feature of severe COVID-19²⁸.

In this study, the AUC value of PNI was higher than that of CONUT, serum albumin, total cholesterol, and lymphocytes, which had a good predictive value for the prognosis of patients hospitalized with COVID-19. Low PNI was an independent risk factor for critical inpatient outcome in COVID-19 patients and had a positive dose-response relationship with the risk of critical outcome. This is in line with the findings of shang et al., who showed a monotonic decline trend of 90-day all-cause mortality risk in patients with spontaneous ICH with increasing PNI levels²⁹. PNI, reflecting the intersection of nutritional and inflammatory statuses, may provide more accurate predictive values for adverse outcomes compared to other nutritional scores³⁰. Therefore, PNI is a reliable predictor to adverse the prognosis of COVID-19 elderly patients.

The Global Leadership Initiative on Malnutrition (GLIM) have noted that once malnutrition is diagnosed, skeletal muscle function should be investigated as a relevant component of sarcopenia and for complete nutrition assessment of persons with malnutrition³¹. The role of micronutrients in supporting the immune system has been widely investigated^32,33. Existing evidence indicates that supplementing multiple micronutrients with immune-supporting effects can modulate immune function and mitigate the risk of infection³². Early nutritional management can lead to recovery from malnutrition after COVID-19 hospitalization¹⁹. Considering that the mean delay between the first onset of symptoms and admission is about one week, this observation suggests that nutritional support should be provided to each COVID-19 patient before admission. In addition, doctors should focus on nutritional status in clinical practice.early nutritional support for patients with high nutritional risk can effectively reduce the incidence of critical illness¹⁵.

This study is characterized by a large sample size, a large amount of data, and high confidence. However, this study also presents some limitations, such as: (1) The samples are solely from one hospital, resulting in a potential single-center bias; (2) This study lacks external validation due to the availability of the data; (3) Since clinical treatment did not explicitly perform standardized nutritional intervention, this retrospective study lacks nutritional intervention and pre-post comparison for malnourished patients; (4) The types of laboratory test data are limited, and the research is not comprehensive; (5) In this study, the linear relationship may limit the power of the restrictive cubic spline in revealing potential non-linear features between PNI and COVID-19 prognosis. Therefore, the results of this study may not fully reflect the complex relationships between the variables.Future studies could attempt to combine restrictive cubic splines with other nonlinear methods to form a more powerful analytical framework to improve model prediction and interpretation capabilities.

In conclusion, the critical illness of elderly COVID-19 patients shows a linear relationship with malnutrition at admission. PNI is a highly valuable biomarker independently associated with elderly COVID-19 severity, and the inpatient outcome of elderly COVID-19 hospitalized patients with PNI < 38.04 may be critical. Future studies are warranted to explore nutritional intervention treatment and pre-post comparison for malnourished elderly COVID-19 hospitalized patients, as well as external and internal validities.