Background

Monitoring the nutritional status of hospitalized patients and the implementation of appropriate dietary support are crucial for disease improvement. Malnutrition occurs when oral intake is persistently inadequate, leading to changes in physical function, weight, and body composition, and a failure to meet nutritional requirements [1]. Malnutrition is associated with complication rates in hospitalized patients, length of hospital stay, mortality, and healthcare costs [2,3]. In Western countries, the prevalence of malnutrition in hospitals has been estimated to be 30%–50% [4]. Additionally, it has been reported in 30%–55% of hospitalized patients in Korea [5], and there is a concern that this condition may worsen during hospitalization [6].

Recently, the focus of nutritional support has shifted toward objectives such as preventing oxidative damage to cells, reducing metabolic overreactions, and enhancing immune responses, as part of nutritional therapy [7]. Multiple studies have reported the necessity of early nutritional support (ENS), which involves providing enteral or intravenous nutrition to patients unable to achieve adequate oral intake within 48 hours of hospitalization [3,7-11]. However, most studies and guidelines related to ENS focus on critically ill patients, overlooking the significance of ENS for patients in general wards. Inpatients, not only those in the intensive care unit (ICU), face challenges in achieving adequate nutritional intake due to several factors, including disease-related appetite loss, medication-related side effects, nil per os (NPO) prescriptions for diagnostic tests, and gastrointestinal disorders arising from disease [12]. Böhne et al. (2022) [2] reported that patients admitted to medical facilities, including hospitals, ICUs, and nursing homes, exhibited low oral intake even after admission. Particularly noteworthy was the observation that while oral intake was not compromised before admission, a notable 27% reduction occurred following admission. In connection with this, the omission or delay in nutritional support for general ward patients may result in supplying inadequate nutrition, thereby exacerbating the progression of malnutrition during hospitalization. Additionally, malnutrition could accelerate infection rates and immune function impairment, potentially increasing the length of hospital stay, mortality rates, and healthcare costs [7].

A recent trend in the increasing severity among patients is related to the aging of the population, the rising incidence of chronic disease, and the complexity of the disease. However, due to limitations in the availability of ICU beds, most patients are admitted to general wards for treatment [13,14]. Hence, by preventing unnecessary transfers from general wards to the ICU, those critically ill patients who absolutely require ICU care have the opportunity to receive it. Moreover, ICU admission influences unnecessary healthcare expenditures, increases the total length of hospital stay, and jeopardizes patient safety [15]. In a study by Lee and Choi (2014) [13], the Sequential Organ Failure Assessment (SOFA) score, representing nutritional status, emerged as a risk factor for ICU admission among general ward patients [16-18].

Objectives

We aim to investigate whether ENS affects the ICU admission rate of general ward patients.

Ethics statement

This was a single-center retrospective study approved by the Institutional Review Board (IRB) of Jeonbuk National University Hospital (CUH 2023-12-044). Informed consent from the subjects was exempted due to the retrospective nature of the study. All procedures adhered to institutional and national ethical standards, as established by the committee overseeing human experimentation, and complied with the principles of the Helsinki Declaration of 2013.

Study design

It is a single institute case-control study. It was described according to the STOBE statement, which is available at: https://www.strobe-statement.org/

Participants

The study included 1,077 patients who were receiving prescriptions related to enteral nutrition or parenteral nutrition and were admitted to Jeonbuk National University Hospital in Jeonju, Jeollabuk-do, Korea, from January 2020 to December 2020, with hospital stays lasting more than 7 days. Exclusion criteria were as follows: patients with hospital stays of less than 7 days or exceeding one year, patients not receiving enteral or parenteral nutrition during hospitalization, those admitted to the ICU, duplicate patients, and those without accessible electronic medical records (EMRs). The final analysis involved 146 patients. In cases of repeated admissions by the same patient, only the first occurrence was considered. The ENS group included 38 patients who received enteral or parenteral nutrition during the first 48 hours after admission as followed in international guidelines [10]. The control group comprised 108 patients who received enteral or parenteral nutrition 48 hours after admission.

Variables

The primary outcome of this study is admission to the ICU, and the secondary outcomes are the amount of calories and protein supplied, length of hospital stay, and in-hospital mortality.

Data sources/measurement

The characteristics of the subjects were collected via EMRs. The data, including age, sex, height, weight, admission type, smoking and alcohol consumption, and nutritional status, were collected at admission. The body mass index (BMI), calculated as weight (kg)/height (m²), ideal body weight, calculated as height (m)×height (m)×22 (male) or 21 (female), and the percentage of ideal body weight (PIBW), calculated as current body weight/ideal body weight×100, were determined using height (cm) and weight (kg). The obesity classification followed the criteria of the Korean Society for the Study of Obesity, categorizing BMI below 18.5 kg/m² as underweight, 18.5 to 22.9 kg/m² as normal, 23.0 to 24.9 kg/m² as overweight, and 25.0 kg/m² or above as obese. Nutritional status at admission was based on the Global Leadership Initiative on Malnutrition criteria [19]. Consultations for nutritional support, as well as the types of nutritional support, were confirmed through prescription during the hospitalization period. The length of NPO, the length of hospital stay, and mortality during the hospitalization period were collected via EMRs.

We calculated daily calorie requirements, applying a minimum value of 25 kcal/kg/d. Protein requirements were calculated by referencing the protein needs per weight based on the stress status of the patient. Protein requirements were calculated in the range of 1.0 to 1.5 g/kg/d, with restrictions to 0.8 g/kg/d based on renal and hepatic failure. In stress status, protein requirements were calculated as 1.2 to 1.5 g/kg for postoperative patients and 1.5 to 2.5 g/kg for patients undergoing continuous renal replacement therapy. Caloric delivery rate (%) was calculated as the supplied calories divided by the calorie requirements, multiplied by 100, whereas the protein delivery rate (%) was calculated as the supplied protein divided by the protein requirements, multiplied by 100.

Biochemical markers analyzed in this study included albumin, total lymphocyte count (TLC), hemoglobin, hematocrit, glucose, blood urea nitrogen (BUN), creatinine, alanine transaminase (ALT), aspartate transaminase (AST), sodium, potassium, chloride, calcium, phosphorus, and C-reactive protein (CRP) and these data were collected via EMRs.

Bias

There was no selection bias reportable.

Study size

Sample size estimation was not conducted because it is a retrospective study. All participants were selected based on the inclusion and exclusion criteria.

Statistical methods

All the data collected were analyzed using IBM SPSS Statistics version 23.0 (IBM Corp.), R software (version 4.3.2., GUI 1.80 for Statistical Computing), and the R Studio interface (version 2023.09.1). Categorical variables such as sex, admission type, nutrition consultation, and ICU admission were presented as frequencies and percentages. Continuous variables such as height, weight, and BMI were expressed as means and standard deviations. For the analysis of continuous variables, either the Student’s t-test or the Mann–Whitney test was applied. The paired t-test or Wilcoxon signed-rank test was used for comparing variables at admission and discharge within each group. Categorical variables were analyzed using Pearson’s chi-square test.

To eliminate selection and confounding bias, propensity score matching (PSM) was conducted [20,21]. Covariates with P<0.1 were chosen to adjust for significant differences [22], so confounding factors included in the propensity score were age, height, admission type, nutrition support access, and nutritional status. Additionally, a univariable logistic model was used to confirm the effect of the following factors: age, height, admission type, nutrition support access, and nutritional status on ICU admission. PSM was executed using a 1:1 nearest neighbor approach with a caliper width of 0.2. Matching was conducted within the caliper values. To validate the covariate balance between the two groups, standardized differences were computed pre- and post-PSM. A balanced covariate was defined as having a standard deviation of less than 10% [20,23]. Crude odds ratios (ORs) and adjusted ORs with a 95% confidence interval (CI) were calculated. A P-value less than 0.05 was considered statistically significant.

After performing PSM, subgroup analysis was conducted for internal medicine as the surgical type was omitted in the admission type. For subgroup analysis, variables such as length of hospital stay, NPO period, ICU admission, laboratory data, calorie supply, and protein supply were analyzed, and the analysis method remains the same as above.

General characteristics of participants

A total of 146 subjects were included in this study (Fig. 1). The demographic characteristics of the subjects are shown in Table 1. There were no significant differences in sex between the ENS group and the control group, but age was significantly lower in the ENS group (77.56±9.79 years vs. 81.66±9.34 years, P=0.028). The admission type in the ENS group consisted of 83.3% medical type and 16.7% surgical type, whereas it was 100% medical type in the control group, indicating a significant difference between the two groups (P=0.007). Moreover, the distribution of medical departments was significantly different between the two groups (P<0.001; Supplement Table 1). Anthropometric data including height, weight, BMI, BMI range and percentage ideal body weight, smoking status, and drinking status were not significantly different between the ENS and control groups. There was a significant difference in access to nutritional support between the two groups (P=0.045). The distribution of enteral and parenteral nutrition combined was the highest in the ENS group at 58.3%, while the distribution of enteral nutrition alone was the highest in the control group at 47.4%. Nutritional status did not show a significant difference between the two groups. However, severe malnutrition was most prevalent in the ENS group at 41.7%, whereas moderate malnutrition was most prevalent in the control group at 50.0%.

Table 1

Baseline characteristics of the two groups

Variable	ENS (n=108)	CON (n=38)	P-value
Sex
Male	64 (59.3)	21 (55.3)	0.705b
Female	44 (40.7)	17 (44.7)
Age (yr)	77.56±9.79	81.66±9.34	0.028c
Admission type
Medical	90 (83.3)	38 (100)	0.007b
Surgical	18 (16.7)	0 (0)
Height (cm)	160.50±8.90	163.18±10.27	0.061c
Weight (kg)	50.15±10.47	53.00±12.67	0.233c
BMI (kg/m2)	19.41±3.56	19.86±4.54	0.703c
BMI range (kg/m2)
<18.5 (underweight)	51 (47.2)	17 (44.7)	0.677b
18.5–22.9 (normal)	39 (36.1)	14 (36.8)
23.0–24.9 (overweight)	10 (9.3)	2 (5.3)
≥25.0 (obese)	8 (7.4)	5 (13.2)
PIBW (%)	89.57±16.87	91.71±21.51	0.703c
Smoking status
Former smoker	25 (23.1)	9 (23.7)	>0.999b
Nonsmoker	79 (73.1)	28 (73.7)
Current smoker	4 (3.7)	1 (2.6)
Drinking status
Former	22 (20.4)	12 (31.6)	0.252b
None	78 (72.2)	22 (57.9)
Current	8 (7.4)	4 (10.5)
Nutrition consultationa
Yes	26 (24.1)	6 (15.8)	0.365b
No	82 (75.9)	32 (84.2)
Nutrition support access
EN only	28 (25.9)	18 (47.4)	0.045b
PN only	17 (15.7)	3 (7.9)
EN & PN	63 (58.3)	17 (44.7)
Nutritional status
No malnutrition	30 (27.8)	5 (13.2)	0.057b
Moderate malnutrition	33 (30.6)	19 (50.0)
Severe malnutrition	45 (41.7)	14 (36.8)

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

^aConsultation for nutrition support.

^bStatistical analysis by χ²-test, ^cstatistical analysis by Mann–Whitney U-test.

ENS = patients who received enteral or parenteral nutrition within 48 hours after admission; CON = patients who received enteral or parenteral nutrition support 48 hours after admission; BMI = body mass index; PIBW = percentage ideal body weight; EN = enteral nutrition; PN = parenteral nutrition.

Fig. 1. low chart of this study. LOS = length of stay; ENS = patients who received enteral or parenteral nutrition 48 hours after admission; CON = patients who received enteral or parenteral nutrition support 48 hours after admission; PSM = propensity score matching.

Outcomes of patients in the two groups

The length of NPO and length of hospital stay for the subjects were significantly shorter in the ENS group compared to the control group (P<0.001 and P=0.016, respectively). Additionally, there was a significant difference in the proportion of ICU admissions during hospitalization, with 10.2% in the ENS group and 26.3% in the control group (P=0.019). Furthermore, in-hospital mortality showed a significant difference, with 9.3% in the ENS group and 23.7% in the control group (P=0.029; Table 2).

Table 2

Outcomes of patients in the two groups

Variable	ENS (n=108)	CON (n=38)	P-value
Length of NPO (day)	4.23±6.45	6.32±5.69	<0.001a
Length of stay (day)	22.83±14.89	27.42±13.83	0.016a,c
ICU admission
Yes	11 (10.2)	10 (26.3)	0.019b,c
No	97 (89.8)	28 (73.7)
In-hospital mortality
Yes	10 (9.3)	9 (23.7)	0.029b
No	98 (90.7)	29 (76.3)

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

ENS = patients who received enteral or parenteral nutrition within 48 hours after admission; CON = patients who received enteral or parenteral nutrition support 48 hours after admission; NPO = nil per os; ICU = intensive care unit.

^aStatistical analysis by Mann–Whitney U-test; ^bstatistical analysis by χ²-test; ^cP<0.05 by χ²-test in internal medicine patients of sub-group.

Daily amounts of required and delivered nutrition

The results of the target and amount of calorie and protein supply for the subjects are shown in Table 3. The proportion of average daily calorie supply relative to the target calories was significantly higher in the ENS group compared to the control group (50.12%±23.30% vs. 38.56%±18.02%, P=0.006). Similarly, the proportion of average daily protein supply showed a significant difference between the ENS and control groups (44.61%±25.07% vs. 32.07%±22.76%, P=0.002). The proportion of calories received through enteral nutrition was significantly higher in the ENS group compared to the control group (31.46%±22.96% vs. 18.76%±15.44%, P=0.004; Supplement Table 2). However, there was no significant difference in the proportion of protein supply through enteral nutrition (Supplement Table 2). Additionally, there were no significant differences between the two groups in the proportion of calories and protein received through parenteral nutrition (Supplement Table 2).

Table 3

Daily required and delivered nutrition amount in the two groups

Variable	ENS (n=108)	CON (n=38)	P-valuea
Energy (kcal)
Requirement	1,388.23±181.52	1,434.87±203.64	0.113
Order (delivered)	678.02±294.79	546.18±260.58	0.016b
Total ordered/required calorie ratio (%)	50.12±23.30	38.56±18.02	0.006b,c
Protein (g)
Requirement	59.99±13.43	52.00±12.83	0.002
Order (delivered)	25.24±12.39	16.74±12.54	<0.001d
Total ordered/required protein ratio (%)	44.61±25.07	32.07±22.76	0.002c

Values are presented as mean±standard deviation.

ENS = patients who received enteral or parenteral nutrition within 48 hours of admission; CON = patients who received enteral or parenteral nutrition support 48 hours after admission.

^aStatistical analysis by Mann–Whitney U-test; ^bstatistical analysis by independent t-test; ^cP<0.05 by Mann–Whitney U-test in internal medicine patients of sub-group; ^dP<0.05 by independent t-test in internal medicine patients of sub-group.

Laboratory data

The serum albumin levels showed no significant differences from admission to discharge in both the ENS and control groups. However, in both groups, the levels of TLC significantly increased at discharge compared to admission (both, P<0.001), with no significant differences between the two groups at admission and discharge. At discharge, serum hemoglobin and hematocrit levels were significantly higher in the ENS group compared to the control group (P=0.015 and P=0.008, respectively). However, hemoglobin levels significantly decreased at discarge compared with levels as admission in both groups (ENS, P<0.001; control, P=0.002, respectively). The levels of serum calcium showed no significant differences between the two groups at admission but were significantly higher in the ENS group compared to the control group at discharge (P=0.027). In both groups, serum CRP levels significantly decreased at discharge compared to levels at admission (both, P<0.001). Additionally, the level of CRP was significantly lower in the ENS group compared to the control group at discharge (P=0.013; Table 4).

Table 4

Laboratory data in the two groups

Variable	ENS (n=108)				CON (n=38)			P-valueb	P-valuec
	Admission	Discharge	P-valuea		Admission	Discharge	P-valuea
Albumin (g/dL)	3.21±0.66	3.10±0.47	0.061		2.97±0.51	2.95±0.42	0.879	0.098	0.093
TLC (cell/mm3)	856.18±650.48	1,191.21±586.11	<0.001g		920.16±670.09	1,309.48±828.75	<0.001g	0.641	0.882
Hb (g/dL)	10.71±2.50	9.85±1.58	<0.001g		10.89±2.83	9.15±1.27	0.002d,g	0.717e	0.015f,h
Hct (%)	32.19±8.23	32.15±21.27	0.985d,g		33.34±8.36	28.51±3.91	0.003d,g	0.418e	0.008h
Glucose (mg/L)	156.92±77.00	152.82±71.55	0.258		191.82±100.62	136.25±42.89	0.007g	0.114	0.117f
BUN (mg/dL)	33.23±32.42	29.14±29.55	0.172		64.61±48.74	28.47±22.11	<0.001g	<0.001h	0.503
Creatinine (mg/dL)	1.63±2.50	1.19±1.47	<0.001g		2.71±3.42	1.18±0.07	<0.001g	<0.001h	0.478
AST (IU/L)	43.06±40.42	57.94±125.30	0.537		165.82±507.39	36.97±25.59	0.004g	0.012h	0.635
ALT (IU/L)	26.80±29.87	28.36±40.84	0.490		64.84±145.14	27.75±24.22	0.295	0.077	0.720
Na (mmol/L)	136.47±7.68	137.52±7.07	0.268		140.34±11.53	137.71±5.11	0.359	0.118	0.878f
K (mmol/L)	4.07±1.02	4.05±0.79	0.867		4.22±1.03	3.87±0.65	0.064	0.172	0.175
Cl (mmol/L)	107.73±28.77	103.60±17.88	0.003g		107.61±10.75	102.74±7.53	0.022d,g	0.260	0.837
Ca (mg/dL)	8.46±1.06	8.73±0.77	0.105		8.42±0.99	8.35±0.77	0.713d	0.749	0.027
P (mg/dL)	3.44±1.91	3.28±1.38	0.855		4.46±2.41	3.06±1.11	0.001g	0.041h	0.460
CRP (mg/L)	87.90±85.11	35.81±41.24	<0.001g		142.62±70.01	54.81±50.05	<0.001g	<0.001h	0.013

Values are presented as mean±standard deviation.

ENS = patients who received enteral or parenteral nutritional support within 48 hours of admission; CON = patients who received enteral or parenteral nutritional support 48 hours after admission; TLC = total lymphocyte count; Hb = hemoglobin; Hct = hematocrit; BUN = blood urea nitrogen; AST = aspartate aminotransferase; ALT=alanine aminotransferase; CRP = C-reactive protein.

^aStatistical analysis by Wilcoxon signed-rank test within the group; ^bstatistical analysis by Mann–Whitney U-test at admission between group; ^cstatistical analysis by Mann–Whitney U-test at discharge between groups; ^dstatistical analysis by paired t-test; ^estatistical analysis by independent t-test at admission; ^fstatistical analysis by independent t-test at discharge; ^gP<0.05 by Mann–Whitney U-test in internal medicine patients of sub-group between groups; ^hP<0.05 by Mann–Whitney U-test in internal medicine patients of sub-group between groups.

Factors influencing ICU admission

Details on the factors influencing ICU admission are shown in Table 5, Supplement Tables 3, 4. Before PSM, ENS and admission type were significantly associated with ICU admission (OR 0.32, 95% CI 0.21–0.82, P=0.018; OR 5.18, 95% CI 1.73–15.54, P=0.003, respectively). Moreover, ENS significantly influenced ICU admission after adjusting for age, admission type, and nutrition support access (aOR 0.14, 95% CI 0.04–0.53, P=0.004). Furthermore, ENS was significantly associated with ICU admission after PSM (OR 0.08, 95% CI 0.01–0.69, P=0.022).

Table 5

Effect of ENS on ICU admission for general ward patients

Variable	Model 1 (n=146)			Model 2 (n=146)			Model 3 (n=62)
	OR (95% CI)	P-value		aOR (95% CI)	P-value		OR (95% CI)	P-value
No	1			1			1
Yes	0.32 (0.12–0.82)	0.018		0.14 (0.04–0.53)	0.004		0.08 (0.01–0.69)	0.022

Model 1 = univariable logistic regression; Model 2 = multivariable analysis with adjustment of age, admission type, and nutrition support access; Model 3 = logistic regression after propensity score matching.

ENS = patients who received enteral or parenteral nutrition within 48 hours of admission; ICU = intensive care unit; OR = odds ratio; aOR = adjusted odds ratio.

Key results

This study was conducted to determine whether ENS impacted the ICU admission of general ward patients. The rates of energy and protein supply in the ENS group were significantly higher than those in the control group. Furthermore, providing ENS to patients in the general wards can prevent ICU admissions.

Interpretation/comparison with previous studies

Patients with poor oral intake or undergoing NPO often find it challenging to meet nutritional requirements through diet alone. Therefore, it is advisable to provide supplemental enteral or intravenous nutrition concurrently. However, patients face barriers to receiving sufficient nutritional support due to various factors [12]. Inadequate nutritional support may lead to malnutrition and is reported to affect almost half of hospitalized patients by previous studies [4,5]. Nutritional support aims to prevent negative outcomes such as increased infection rates, muscle loss, and delayed wound healing due to nutritional deficiencies [2]. The guidelines for nutritional support in critically ill patients strongly recommend initiating ENS within 48 hours of admission to prevent complications [24]. Nutritional support is essential not only for critically ill patients but also for all hospitalized patients when considering factors such as restricted activity, sensory loss (taste, smell), and surgery, which can contribute to nutritional deficiencies [25]. Therefore, we determined the impact of ENS within 48 hours of admission for general ward patients undergoing fasting or insufficient oral intake.

In this study, the average amounts of calories and protein supplied compared to nutritional goals were analyzed to determine whether ENS achieves the nutritional requirements for general ward patients. The rates of the calories and protein supplied were significantly higher in the ENS group compared to the delayed nutritional support group. Similarly, in a study by Koga et al. (2018) [8], a significant difference was observed in the amounts of calories and protein supplied between the group receiving early enteral nutrition within 48 hours of admission and the group receiving delayed enteral nutrition. The early enteral nutrition group had significantly higher amounts than the delayed enteral nutrition group. Furthermore, Pardo et al. (2023) [10] reported that the ENS group reached the target calories faster than the delayed nutritional support group, and the amount of calories supplied was also higher. Based on both this study and prior studies, ENS has a positive effect on achieving the nutritional goals for general ward patients.

Proper nutritional supply is related to biomarkers reflecting nutritional status. In a systematic review and meta-analysis by Zhang et al. (2017) [26], serum albumin, prealbumin, hemoglobin, total cholesterol, and total protein were found to be useful biomarkers to access malnutrition in adults. Unfortunately, albumin level was not significantly different between the ENS group and the control group. Albumin is influenced not only by the nutritional status of the patient but also by hepatic diseases, acute inflammatory conditions, and dehydration [27,28]. Particularly, serum albumin decreases sharply in inflammatory stress status, and guidelines suggest that recovery through active nutritional support may be unattainable [27]. For these reasons, we concluded that ENS had an insufficient effect on albumin in this study.

Hemoglobin is also a biomarker that reflects malnutrition [26]. In this study, hemoglobin levels, which are known to decrease in cases of malnutrition, improved significantly at discharge in the ENS group [27]. The components of hemoglobin require amino acids, vitamins, including vitamin B6, and iron [29,30]. This means that an inadequate nutrient supply can decrease the synthesis of hemoglobin. Therefore, we suggest that ENS improves hemoglobin and nutritional status by providing necessary nutrients.

In this study, the levels of AST, BUN, creatinine, and CRP in the control group were significantly higher compared to the ENS group at admission. These biomarkers are indicators that reflect impaired liver function, renal dysfunction, and inflammation [27]. Therefore, abnormalities associated with these biomarkers might be one reason to delay nutritional support in this study. However, an absolute contraindication of nutritional support is hemodynamic instability, requiring fluid resuscitation, septic shock, mean arterial pressure below 60 mmHg, or an escalating demand for vasoactive support [31], and not the abnormality of biomarkers. Moreover, as the hemodynamic stability of general ward patients tends to be constant, ENS may be administered. This demonstrates the importance of clinicians acquiring specialized knowledge regarding the indications and contraindications of nutritional support. Additionally, according to recent studies, inadequate medical staff education, along with factors such as excessive workload and personnel shortages, was identified as causes for inadequate nutritional support for patients [3,32,33]. Therefore, it is believed that the causes of delayed nutritional support in this study may also be related to these factors. Hence, it is deemed necessary to establish systems at the hospital and national levels to address this issue, including the education of medical staff.

In this study, the prevalence of patients with medical conditions was significantly higher in the control group than in the ENS group. Additionally, the biomarkers indicating renal and liver functions, as previously mentioned, were higher in the control group. This suggests that patients with chronic illnesses or complex underlying diseases might be considered for transfer to the ICU [34]. However, both of the groups exhibited elevated levels of BUN, creatinine, ALT, AST, and CRP at admission, surpassing the normal range. Given that BUN with a threshold of 20 mg/dL is a predictive factor for ICU admission, both the control and ENS groups indicate a high risk of ICU admission [35]. Therefore, the significantly lower ICU admission rate in the ENS group compared to the control group indicates the importance of ENS for general ward patients.

The transfer of patients from general wards to the ICU demonstrates higher severity and mortality rates compared to transfers that occur through alternative routes such as emergency or operating rooms [13]. Preventing transfers from general wards to the ICU is considered to reduce patient mortality and have a positive impact on prognosis. One risk factor influencing the decision to transfer patients from the general ward to the ICU is the SOFA score [13,16-18]. According to prior studies, the SOFA score was reported as indicative of a patient’s nutritional status [36,37], and ENS can improve a patient’s nutritional status [3,7-11].

Limitations

There are several limitations in this study. Firstly, we did not include the data of patients with underlying disease and clinical status, such as operation, trauma, and others. Secondly, the study did not conduct various blood biomarkers related to malnutrition (prealbumin, total protein). Thirdly, the significant differences between the two groups in indicators related to renal and liver function at admission suggest that they may act as factors attenuating the positive effects of ENS. Lastly, we could not compare the nutritional status of patients at admission and discharge due to insufficient data for nutritional status assessment at the time of discharge.

Suggestion for further studies

We suggest conducting a multi-center prospective study on ENS in the future. This would address the limitations of this single-center retrospective study and further confirm the im- pact of ENS.

Conclusion

Improving nutritional status through ENS is reflected in the improvement of the SOFA score, which can decrease the risk factor for ICU admission. The significantly lower ICU admission in the ENS group compared to the control group of this study also supports the evidence shown in previous studies.

This study contributes to the limited studies on ENS for general ward patients and provides scientific evidence to substantiate the necessity of ENS for general ward patients. Furthermore, it is strengthened by its focus on the domestic population.

Conceptualization: CYK, DSK. Data curation: HSK, MRL. Formal analysis: SL, SWH. Investigation: MRL, DSK, HSK. Methodology: CYK, SWH, SHK. Project administration: SHK, HSK, CYK. Resources: CYK, JDY, SWH, MRL, DSK, SL, SHK, HSK. Supervision: CYK. Validation: JDY, SWH, SHK. Visualization: MRL, DSK, HSK. Writing – original draft: HSK, SL. Writing – review and editing: JDY, SWH, MRL.

The authors of this manuscript have no conflicts of interest to disclose.

None.

Contact the corresponding author for data availability.

None.

Supplementary materials can be found via https://doi.org/10.15747/ACNM.2024.16.2.57

Supplement Table 1. Distribution of medical departments.

Supplement Table 2. Daily required and delivered nutritional amount in the two groups.

Supplement Table 3. Logistic regression analysis of general characteristics and ENS on ICU admission.

Supplement Table 4. Standardized mean difference before and after PSM.