Background

COVID-19 is an infectious respiratory disease caused by severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) [1], which was first reported in Wuhan, China, in December 2019, and declared a global pandemic by the World Health Organization on March 2020 [2]. Patients infected with SARS-CoV-2 may complain of fever, cough, sore throat, fatigue, shortness of breath, and/or loss of appetite due to changes in taste and smell, and most patients recover over time with mild symptoms. However, some patients with underlying medical conditions and elderly patients are susceptible to COVID-19 pneumonia, which can progress to sepsis or acute respiratory distress syndrome, multiple organ failure, and death [3-6]. Many countries have reported sudden deaths from COVID-19.

Approximately 5% of COVID-19 patients reportedly require intensive care [7]. While the risk factors affecting COVID-19 mortality remain uncertain, malnutrition resulting from long-term intensive care appears to reduce the mass and function of skeletal muscle. Subsequent malnutrition is associated with increase in morbidity and mortality [8].

Malnutrition is an imbalance in nutrition due to excessive or insufficient intake of energy, protein, and other nutrients. Among surgeons, it is well known that patients with malnutrition experience poor wound healing, higher rates of surgical complications, higher mortality, increased medical costs, and longer hospital stays than do those with good nutritional status.

Even in medical diseases, the influenza pandemic of 1918 showed that malnutrition was a risk factor for the severity and mortality of viral pneumonia [9]. The severity and transmission of pandemic influenza viruses depend on complex interactions among the virus, host, and external factors. Malnutrition is reportedly an important contributing factor in increasing the risk of mortality and the severity of viral pneumonia [9]. Even in patients with viral pneumonia, nutritional management during hospitalization, such as screening patients at risk of malnutrition at an early stage of hospitalization and providing appropriate nutritional support, is strongly recommended [10]. Research on factors related to various clinical features and severity associated with COVID-19 is underway worldwide, and malnutrition is expected to be related to mortality and complications in patients with pneumonia caused by COVID-19 [11].

Objectives

We hypothesized that malnutrition can affect treatment outcomes in COVID-19 pneumonia in intensive care units (ICUs). This study is an attempt to identify how nutritional status and nutritional support affect the survival of critically ill COVID-19 patients.

Ethics statement

This is a retrospective study conducted by collecting electronic medical records with deliberation and approval from the Institutional Review Board (IRB) of Dong-A University Hospital (approval number: DAUHIRB-23-016). Obtainment of informed consent was waived from the IRB.

Study design

It is a case-control study. It was described according to the STROBE statement available at: https://www.strobe-statement.org/.

Setting

We collected data on patients admitted to the ICU of Dong-A University Hospital for COVID-19 pneumonia from January to December 2021. COVID-19 was diagnosed through real-time polymerase chain reaction amplification tests. During the COVID-19 pandemic, Dong-A University Hospital was tasked with treating only critically ill patients. The decision to admit a patient to the ICU was made by specialists in infectious diseases and pulmonology. Most patients were transferred from a primary hospital to receive ventilator care, extracorporeal membrane oxygenation (ECMO), or continuous renal replacement therapy (CRRT), if needed.

Participants

We selected patients who received enteral nutrition (EN) and parenteral nutrition (PN) support after monitoring by a Nutrition Support Team. Our team evaluates the nutritional status of patients who have malnutrition or require EN and PN support, calculates their nutritional requirements, and establishes nutrition plans such as formulating and dosing nutritional support. If no physical measurements, such as height and weight, or no record from the Nutrition Support Team is available, the patient was excluded from the analysis.

Variables

Primary outcome variable is a nutritional status and the other variables are medications, ICU treatment, and clinical prognostic indicators.

Data sources/measurement

We collected clinical parameters of age, sex, weight, height, underlying disease, Acute Physiology and Chronic Health Evaluation II (APACHE II) score, PaO₂/FiO₂ ratio, serum C-reactive protein (CRP) level, and treatment provided. To compare nutritional factors between survivors and deceased patients, we analyzed nutritional status, body mass index (BMI), serum albumin level, and requirements for and supplied amounts of calories and protein through EN and PN support. Data on height and last known weight (i.e., before COVID-19 within the past 6 months) were collected by questioning patients during or shortly after their stay in the hospital.

Initial nutritional status assessments were classified based on the serum albumin level and percentage of ideal body weight (PIBW) according to the nutrition assessment guidelines of our hospital. A serum albumin of 3.5 g/dL or more and a PIBW of 90% or more were classified as adequate; a serum albumin level of 3.0–3.5 g/dL and a PIBW of 75%–90% were classified as mild malnutrition; a serum albumin level of 3.0–3.5 g/dL and a PIBW of 60%–75% were classified as modulate malnutrition; a serum albumin level less than 3.0 g/dL and a PIBW greater than 90% were classified as kwashiorkor; and a serum albumin level greater than 3.0 g/dL and a PIBW of no more than 80% were classified as marasmus.

Bias

There was no potential source of bias in selecting participants.

Study size

Since it is a retrospective study, sample size estimation was not done. All participants during the study period were selected according to the inclusion and exclusion criteria.

Statistical methods

We summarized the data as mean±standard deviation for continuous variables and numbers and percentages for categorical variables. We analyzed categorical variables using chi-square and Fisher’s exact tests. Continuous variables were compared using a two-tailed t-test for normally distributed data using IBM SPSS version 27 for Windows (IBM Corp.). A P-value less than 0.05 was considered statistically significant.

Participants

A total of 70 patients was included in the study, of whom 57 (81.4%) were discharged from the hospital after COVID-19 treatment and 13 (18.6%) died. All patients able to ingest food orally were fed a high-protein diet. EN or PN support was started an average of 3.2 days after admission to the ICU. During an average of 20.4 days of hospitalization, EN or PN support was supplied for an average of 12.5 days.

Basic information and clinical characteristics of the study subjects are presented in Table 1. The average age was 59.8±17.6 years in the survival group and 74.2±10.6 years in the deceased group; the difference was statistically significant (P=0.006). There was no statistically significant difference between the two groups in terms of sex (P=0.753) and height (P=0.511). The weights at hospitalization (P=0.766) and at discharge (P=0.493) were lower in the death group but showed no significant difference. BMI was also lower in the death group, but there was no significant difference between the two groups (P=0.676).

Table 1

Baseline characteristics of the study population

Variable	Total (n=70)	Survival (n=57)	Deceased (n=13)	P-value
Age (yr)	62.49±17.4	59.8±17.6	74.2±10.6	0.006*
Sex				0.753
Male	46 (65.7)	38 (66.7)	8 (61.5)
Female	24 (34.3)	19 (33.3)	5 (38.5)
Height (cm)	166.0±10.1	165.6±10.6	167.8±7.4	0.511
Weight at admission (kg)	66.9±15.2	67.1±15.5	65.7±14.4	0.766
Weight at discharge (kg)	66.2±14.5	67.2±15.4	62.5±10.5	0.493
Body mass index (kg/m2)	24.9±3.9	24.1±3.9	23.6±3.9	0.676
Underlying disease
Hypertension	9 (12.9)	5 (8.8)	4 (30.8)	0.055
Diabetes mellitus	13 (18.6)	8 (14.0)	5 (38.5)	0.056
Chronic kidney disease	1 (1.4)	0	1 (7.7)	0.186
Cancer	2 (2.9)	2 (3.5)	0	>0.999
Asthma	1 (1.4)	1 (1.8)	0	>0.999
Cerebral infarction	3 (4.3)	3 (5.3)	0	>0.999
Myocardial infarction	2 (2.9)	1 (1.8)	1 (7.7)	0.339
Liver cirrhosis	1 (1.4)	1 (1.8)	0	>0.999
Atrial fibrillation	2 (2.9)	2 (3.5)	0	>0.999
Etc.	13 (18.6)	13 (22.8)	0	>0.999

Values are presented as mean±standard deviation or number (%).

*P<0.05.

With respect to underlying diseases, 13 patients (18.6%) had diabetes mellitus, 9 (12.9%) had hypertension, 3 (4.3%) had a cerebral infarction, 2 (2.9%) had cancer, and there were 2 cases of myocardial infarction (2.9%) and 2 cases of atrial fibrillation (2.9%). Patients with 3 additional comorbidities showed a higher mortality rate (Table 1).

Main results

Clinical prognosis indicators

The two groups were compared by examining APACHE II scores, PaO₂/FiO₂ ratios at hospitalization, and CRP levels at hospitalization and discharge (Table 2). APACHE II score was higher in the group of patients who died, but the difference was not statistically significant (P=0.944). A total of 37.1% of patients received high-flow oxygen therapy or mechanical ventilation at the time of admission, and the average PaO₂/FiO₂ was 237.3. The PaO₂/FiO₂ ratio at admission (P=0.670) and CRP level at admission (P=0.102) were also higher in the deceased patients, but not to a significant degree. At discharge, CRP level was significantly higher in the deceased group than in the survival group (P=0.002).

Table 2

Clinical prognostic indicators

Variable	Total (n=70)	Survival (n=57)	Deceased (n=13)	P-value
APACHE II score	14.8±7.7	14.7±7.1	15.0±10.3	0.944
PaO2/FiO2 ratio at admission	237.3±145.5	233.7±139.1	252.9±176.1	0.670
Median (IQR)		199.3 (128.0–306.4)	212 (141.8–335.0)
CRP at admission (mg/dL)		10.2±6.8	13.6±5.4	0.102
CRP at discharge (mg/dL)		1.5±2.5	7.2±5.1	0.002**

Values are presented as mean±standard deviation.

APACHE II = Acute Physiology and Chronic Health Evaluation II; CRP = C-reactive protein.

**P<0.01.

Decisions about whether to administer drugs (dexamethasone, remdesivir, regdanvimab, or tocilizumab) to treat COVID-19 or top use a ventilator, ECMO, or CRRT were described and compared between the two groups (Table 3). The proportions of patients prescribed dexamethasone (P=0.336), remdesivir (P=0.763), regdanvimab (P>0.999), and tocilizumab (P=0.123) to treat COVID-19 were not significantly different between the two groups. The proportion of patients on ventilators was significantly higher in the death group than in the survival group (P=0.012). The proportion of patients with ECMO and CRRT was significantly higher in the death group than in the survival group (P<0.001).

Table 3

Comparison of COVID-19 medications and intensive care unit treatments

Variable	Total (n=70)	Survival (n=57)	Deceased (n =13)	P-value
Medicationsa
Dexamethasone	62 (88.6)	49 (86.0)	13 (100)	0.336
Remdesivir	36 (51.4)	30 (52.6)	6 (46.2)	0.763
Regdanvimab	2 (2.9)	2 (3.5)	0	>0.999
Tocilizumab	33 (47.1)	24 (42.1)	9 (69.2)	0.123
ICU treatment
Ventilator	26 (37.1)	17 (29.8)	9 (69.2)	0.012*
ECMO	11 (15.7)	3 (5.3)	8 (61.5)	<0.001***
CRRT	8 (11.4)	1 (1.8)	7 (53.8)	<0.001***

Values are presented as number (%).

ICU = intensive care unit; ECMO = extracorporeal membrane oxygenation; CRRT = continuous renal replacement therapy.

^aMedications available at Dong-A University Hospital for COVID-19 treatment as of 2021.

*P<0.05 and ***P<0.001.

Nutritional status and support

The nutritional status of patients was based on serum albumin level and PIBW at hospitalization (Table 4). There was no significant difference in nutritional status at the onset of hospitalization between the two groups (P=0.280). The amounts of calories and protein supplied through oral, EN, and PN support were calculated, and the calorie and protein requirements per kilogram of reference weight were calculated and compared (Table 5). There was no significant difference between the two groups in calorie and protein requirements (P=0.163 and P=0.408, respectively). The total calorie supply was significantly lower in the deceased group (P<0.001), as was the total protein supply (P=0.019). Daily calories and protein supplied per weight (P=0.002 and P=0.015, respectively) were significantly lower in the deceased group than in the survival group. Serum albumin level at the time of hospitalization was not significantly different between the two groups (P=0.205), but serum albumin level at discharge was significantly lower among non-survivors (P<0.001; Table 6).

Table 4

Comparison of nutritional status at admission

Variable	Total (n=70)	Survival (n=57)	Deceased (n=13)	P-value
Normal	15 (21.4)	14 (24.6)	1 (7.7)	0.280
Abnormal
Mild malnutrition	32 (45.7)	26 (45.6)	6 (46.2)
Moderate malnutrition	7 (10.0)	5 (8.8)	2 (15.4)
Severe malnutrition	1 (1.4)	1 (1.7)	0
Kwashiorkor	10 (14.3)	6 (10.5)	4 (30.7)
Marasmus	5 (7.2)	5 (8.8)	0

Values are presented as number (%).

Table 5

Comparison of nutritional requirements and nutrition supply status

Variable	Total (n=70)	Survival (n=57)	Deceased (n=13)	P-value
Nutritional requirement
Calorie (kcal)	1,685.9±347.4	1,721.3±342.2	1,556.2±349.6	0.163
Protein (g)	79.3±15.4	80.2±13.9	76.1±18.7	0.408
Nutrition supply
Total calorie (kcal)	1,420.2±382.4	1,495.0±387.1	1,145.8±201.7	<0.001***
Total protein (g)	62.4±22.4	64.8±24.0	53.2±11.0	0.019*
Calorie/BW (kcal/kg)	21.8±6.8	22.9±7.0	17.8±3.9	0.002**
Protein/BW (g/kg)	1.0±0.3	1.0±0.3	0.8±0.2	0.015*

Values are presented as mean±standard deviation.

BW = body weight.

*P<0.05, **P<0.01, and ***P<0.001.

Table 6

Comparison of serum albumin level

Variable	Total (n=70)	Survival (n=57)	Deceased (n=13)	P-value
Serum albumin level
At admission (g/dL)	3.4±0.5	3.4±0.5	3.2±0.4	0.205
At discharge (g/dL)	3.1±0.5	3.2±0.4	2.6±0.2	<0.001***

Values are presented as mean±standard deviation.

***P<0.001.

Interpretation

In this study, we investigated only patients with critical illness requiring ICU hospitalization. These patients were assumed to be a relatively homogeneous group among those infected with SARS-CoV-2, and nutritional status and requirements were assessed according to a common protocol. The average duration of admission to the ICU for the patients in this study was 20.4 days, and malnutrition was observed in all but 15 patients, who had good nutritional status at admission. Although the nutritional status at the time of hospitalization was similar between survivors and non-survivors, the amount of nutrients actually supplied was lower in the non-survivor group. The American Society for Parenteral and Enteral Nutrition (ASPEN), which set guiding principles for COVID-19 management, recommends an energy goal of 15–20 kcal/kg actual body weight (ABW)/day, which should provide 70% to 80% of a patient’s energy requirements, and a protein goal of 1.2–2 g/kg ABW/day [7]. ESPEN expert statements and practical guidance for nutritional management of individuals infected with SARS-CoV-2 recommend an energy intake of 27–30 kcal/kg of body weight/day and a protein intake of at least 1 g/kg of body weight/day [8]. In this study, the amounts of calories and protein supplied to patients who died were less than those recommended by the guidelines. This can be interpreted in several ways.

Critically ill patients are prone to deteriorating nutritional status and can quickly enter a catabolic state due to complex metabolic changes in cytokines and stress hormones. Hemodynamic instability due to sepsis, ventilator care due to respiratory complications, and application of ECMO or CRRT are risk factors that hinder adequate nutritional support. In addition, for patients in a state of shock, the use of inotropic agents may be limited to ensure sufficient nutrition. The deceased group was older on average than the survivors, their CRP level was higher at discharge, and the proportion of patients who used a ventilator, ECMO, or CRRT during treatment was also high. Although the patients were in a recovery period and required sufficient nutritional support, oral intake was often difficult due to difficulty swallowing after removing the ventilator. This was accompanied by a loss of taste and smell and a loss of appetite. To minimize deterioration of nutritional status, sufficient calories and protein must be supplied to severely ill COVID-19 patients during hospitalization [8,12].

Because an immunocompromised state can be a risk factor for respiratory infections, adequate nutritional support and maintaining adequate nutritional status are believed to be crucial for ensuring optimal immunity against infection [13]. Although mortality rates are decreasing due to the development of vaccines and new pharmaceutical treatments, clinical research on COVID-19 may continue to reveal the nature of COVID-19 and protect populations against the emergence of mutant viruses. This study represents an attempt to identify risk factors that affect the survival of critically ill patients with COVID-19 and demonstrates the benefits of adequate nutritional support.

Comparison with previous studies

Kananen et al. [14] reported that a low BMI was associated with an increase in the mortality rate among elderly patients with COVID-19. In a large cohort study of COVID-19 patients, Bouziotis et al. [15] reported that overweight patients had a significantly lower risk of in-hospital death compared with those who were underweight. This indicates that a nutritional status outside the standard range, rather than weight itself, is a risk factor. Body measurements, such as body weight, are used as an index to indicate imbalances in calorie and protein intake and consumption over a certain period of time. They are used to diagnose chronic caloric and protein intake deficiencies, sluggish growth, and overweight and can also be used to evaluate the response to nutritional treatment. It is important to accurately measure weight when evaluating nutritional status during hospitalization. Although no significant differences were found in weight and BMI at admission and discharge, both tended to be lower in the deceased group than in the surviving group. However, it was difficult to interpret the results in critically ill patients due to difficulty accurately measuring patient weights due to their clinical conditions as well as water retention caused by stress.

Among the nutrients that affect immune function, vitamin D, magnesium, vitamin B12, and vitamin C have been reported to significantly reduce the proportion of severely ill patients and prolong survival [13,16-19]. Although immune-boosting nutrients, such as glutamine, selenium, and zinc, reportedly have antiviral effects [13,20,21], appropriate supply levels, methods, types, and administration periods have yet to be established. European Society for Clinical Nutrition and Metabolism guidelines suggest providing the recommended daily amounts of vitamins and trace elements [8]. No blood measurements of trace elements and vitamin D were conducted for this study. However, the total calories and protein mass supplied during hospitalization appeared to be lower in the deceased group compared with the survival group, indicating insufficient nutritional support.

The serum albumin level, which can be used as an index of nutritional status, may be affected by an inflammatory response. In previous studies, serum albumin level was not considered an appropriate indicator of malnutrition [22]. However, Eckart et al. [23] reported that serum albumin level can be used to evaluate nutritional status in patients with acute disease conditions, regardless of the presence of inflammation. Chronic protein deficiency, even when there is adequate intake of non-protein calories, can lead to significant hypoalbuminemia. Body weight and serum albumin concentrations are the minimum objective data needed to substantiate clinical observations and establish reference values for monitoring. The cost-effectiveness of adding biochemical markers other than albumin has not been established [24]. We used serum albumin level to assess the nutritional status of patients, finding a lower level lower in the mortality group.

Considering that multiple factors affect the severity of COVID-19, inadequate nutritional support alone is not responsible for patient death. However, there was no significant difference between the two groups in the type and number of underlying diseases, nutritional status, and albumin and CRP levels at the time of admission. These results suggest that sufficient nutritional support during hospitalization in the ICU can contribute to the survival of critically ill patients with COVID-19. Calculations of actual nutritional indicators and supplying nutritional requirements in a relatively uniform group would be worthy of research resources.

Limitations

However, this study had several limitations. Measuring ABW as a reference to assess nutritional status was difficult because many patients were bedridden. In addition, when calculating caloric requirements, the calculation method we followed (according to our guidelines) did not use an indirect calorimeter.

Generalizability

The result of this study may be able to generalized to case of COVID-19 patients in another hospital in Korea.

Suggestion for further studies

Further studies are required to identify additional factors that affect patient survival.

Conclusion

Due to various factors affecting the severity of COVID-19, it was not possible to determine whether the nutritional status of critically ill patients with COVID-19 at hospitalization had an effect on survival. Adequate nutritional support during ICU stays may contribute to the survival of critically ill patients with COVID-19. The severity of COVID-19, the unstable hemodynamic status of the deceased group, and difficulties providing active nutritional support are potential influences.

Conceptualization: NGL, SHN. Data curation: NGL, HJK, JGK, DHJ, MSK. Formal analysis: NGL, SHN. Investigation: NGL, HJK, JGK, DHJ, MSK. Methodology: SHN. Project administration: SHN. Resources: NGL, SHN. Software: NGL, SHN. Supervision: SHN. Validation: SHN. Visualization: NGL, SHN. Writing – original draft: NGL. Writing – review & editing: SHN.

None.