Incidence and Impact of Refeeding Syndrome in an Internal Medicine and Gastroenterology Ward of an Italian Tertiary Referral Center: A Prospective Cohort Study

Background: Refeeding syndrome (RS) is a neglected, potentially fatal syndrome that occurs in malnourished patients undergoing rapid nutritional replenishment after a period of fasting. The American Society for Parenteral and Enteral Nutrition (ASPEN) recently released new criteria for RS risk and diagnosis. Real-life data on its incidence are still limited. Methods: We consecutively enrolled patients admitted to the Internal Medicine and Gastroenterology Unit of our center. The RS risk prevalence and incidence of RS were evaluated according to ASPEN. The length of stay (LOS), mortality, and re-admission rate within 30 days were assessed. Results: Among 203 admitted patients, 98 (48.3%) were at risk of RS; RS occurred in 38 patients (18.7% of the entire cohort). Patients diagnosed with RS had a higher mean LOS (12.5 days ± 7.9) than those who were not diagnosed with RS (7.1 ± 4.2) (p < 0.0001). Nine patients (4.4%) died. Body mass index (OR 0.82; 95% CI 0.69–0.97), RS diagnosis (OR 10.1; 95% CI 2.4–42.6), and medical nutritional support within 48 h (OR 0.12; 95% CI 0.02–0.56) were associated with mortality. Conclusions: RS incidence is high among clinical wards, influencing clinical outcomes. Awareness among clinicians is necessary to identify patients at risk and to support those developing this syndrome.


Introduction
Refeeding syndrome is a potentially fatal complication that occurs in malnourished patients when an excessive amount of nutrients is too rapidly delivered after a prolonged state of fasting, as often happens in hospital settings [1]. It is defined as severe electrolyte and fluid shifts, associated with metabolic abnormalities in malnourished patients undergoing refeeding, whether orally, enterally, or parenterally [2]. Up-to-date data about the incidence of RS remain heterogeneous. A recent systematic review of thirty-five observational studies found a wide range of RS incidence, varying from 0 to 62%, among studies [3].
The risk of developing RS is high after prolonged fasting, leading to a reduction in insulin release and an increase in glucagon secretion. Metabolic inversion from the use of

Study Design and Ethical Committee Approval
This was a single-center observational prospective cohort study. The study was conducted based on the Declaration of Helsinki and according to Good Clinical Practice guidelines. The study was approved by the Ethical Committee of Fondazione Policlinico A. Gemelli IRCCS-Catholic University of the Sacred Heart (Protocol code 2638/22). All participants signed a consent form recording their agreement to take part in the study and to have the results published. This study was reported according to the STROBE guidelines for cohort studies [7].

Patients
All adult (>18 years old) patients admitted to the Internal Medicine and Gastroenterology Unit at the Fondazione Policlinico Agostino Gemelli IRCCS, Rome, Italy, from March 2021 to January 2022, were prospectively evaluated.
The exclusion criteria were patients already undergoing artificial nutrition during a possible stay in the emergency room, the presence of hypophosphatemia on the day of admission, or refusal to participate in the study. Mortality, LOS, and hospital re-admission within 30 days among enrolled patients were recorded by the analysis of medical records.

Determination of RS Risk
The evaluation of RS risk was performed according to the recently released ASPEN criteria [6].
Patients were considered at "moderate risk" if they met two of the following risk criteria: • BMI was between 16 and 18.5 kg/m 2 ; • A weight loss of 5% of habitual weight was reported; Patients were considered at "significant risk" if they met one of the following risk criteria: A weight loss of 7.5% in 3 months or >10% in 6 months was reported; • There was no or negligible oral intake for >7 days OR < 50% of estimated energy requirement for >5 days during an acute illness or injury OR < 50% of estimated energy requirement for >1 month; • There were moderately/significantly low levels of potassium, phosphorus, magnesium, or minimally low or normal levels and recent low levels necessitating significant or multiple-dose supplementation; • There was evidence of severe subcutaneous fat loss; • There was evidence of severe muscle loss; • In presence of higher-risk comorbidities (severe disease).

Diagnosis of RS
The diagnosis of RS was then confirmed according to the above-mentioned ASPEN criteria [6], which are as follows:

•
A decrease in serum phosphorus, potassium, and/or magnesium levels by 10-20% (mild RS), 20-30% (moderate RS), or >30%, and/or organ dysfunction resulting from a decrease in any of these and/or due to thiamin deficiency (severe RS).

•
The decrease occurs within 5 days of reinitiating or substantially increasing energy provision.

Outcomes Measures
The primary outcome was the incidence of RS risk according to ASPEN 2020 criteria. The secondary outcomes were the diagnosis of RS after the initiation of nutritional support and its impact on mortality, LOS, and hospital readmission within 30 days.

Sample Size Calculation
According to a recent systematic review, the incidence of RS widely varied from 0% to 62% across the studies [3]. However, a previous study identified an incidence of RS risk up to 54% and an RS diagnosis rate of 8% of patients admitted to an internal medicine department [8]. With a margin of error of 6% and a confidence interval of 95%, between 79 and 158 patients should have been enrolled to intercept the above-mentioned incidences (percentages). Considering a dropout rate of 10%, our sample size was set at 176 patients.

Data Collection and Statistical Analysis
Data were collected using a specific Excel © spreadsheet. Data are shown using descriptive statistical methods. The following measures were used as quantitative variables: minimum, maximum, range, mean and standard deviation. The qualitative variables were summarized in tables of absolute and percentage frequencies. The possible normality of continuous distributions wasexamined by applying the Kolmogorov-Smirnov test.
The primary objective was reached by calculating the cumulative incidence of risk of developing RS in the enrolled patients. The secondary objectives were reached through the cumulative incidence of overt RS development in the enrolled at-risk patients, and the development of a Cox regression model for the evaluation of mean hospital stay, with log-rank analysis to highlight differences between the groups that develop RS or not. To analyze other secondary outcomes, univariate logistic models were created, obtaining odds ratios (OR).
A p < 0.05 was set as statistically significant. All statistical analyses were carried out with STATA (version 13, Stata Corporation; College Station, TX, USA).

Baseline Patients' Characteristics
Two hundred and three patients were enrolled during 11 months of observation. The clinical and demographical data are presented in Table 1. The mean age was 66.1 ± 14.1 years; there were 127 male patients (62.6%) and 68.5% of the patients (n = 139) were admitted from the emergency department (ED). The mean Charlson's Comorbidity Index (CCI) was 3.0 ± 2.4.  (Figure 1). tritional risk screening (NRS-2002), 70 patients (34.5%) were at risk of malnutrition, while, according to the Malnutrition Universal Screening Tool (MUST), 99 patients (48.7%) had the same risk. Twenty-four patients (11.8%) underwent specialist nutritional evaluation during their hospital stay, and 74 (36.5%) were treated with medical nutrition within 48 h from admission, in particular, 63 (64.3%) with oral nutritional supplementation (ONS) and 13 (13.3%) with parenteral nutrition (PN).

Refeeding Syndrome
The risk of RS was identified in 98 (48.3%) patients; of these, 44 (21.7%) were at medium risk and 54 (26.6%) were at high risk of developing RS. Thirty-eight patients (18.7% of the entire cohort) developed RS (Figure 1).

Length of Hospital Stay
The mean LOS was 8.2 ± 5.8 days. Patients diagnosed with RS had a mean LOS of 12.5 ± 7.9 days. Patients without RS had a mean LOS of 7.1 ± 4.2 days (p < 0.0001) (Figure 2).  6), and medical nutritional support within 48 h from admission (OR 0.12; 95% CI 0.02-0.56) were associated with mortality, even if these associations were different (Table 4). Thirteen patients (6.4%) were re-admitted to hospital within 30 days from discharge; however, none of the tested variables were associated with re-admission (Supplementary Files Table S1).   6), and medical nutritional support within 48 h from admission (OR 0.12; 95% CI 0.02-0.56) were associated with mortality, even if these associations were different (Table 4). Thirteen patients (6.4%) were re-admitted to hospital within 30 days from discharge; however, none of the tested variables were associated with re-admission (Supplementary Files Table S1).

Discussion
This observational prospective cohort study consecutively enrolled 203 inpatients admitted to an Internal Medicine and Gastroenterology Ward of an Italian Tertiary Referral Center; most of these patients (68.5%) were admitted from the ED. According to the ASPEN criteria [6], 98 patients (48.3%) were at risk of developing RS and 38 patients (18.7% of the whole cohort) developed RS. During the hospital stay, nine patients died; RS was associated