The incidence and mortality of refeeding syndrome in older hospitalized patients, based on three different diagnostic criteria: A longitudinal study.

Background & Aims: Refeeding syndrome (RFS) lacks both a global definition and diagnostic criteria. Different diagnostic criteria are used; serum phosphate (traditional criterion (TC)), the Friedli consensus recommendations, and the ASPEN. We investigated the incidence of RFS in older hospitalized patients and the mortality rates in patients with or without RFS using these three different diagnostic criteria.

Methods: This is a longitudinal study with data originating from a randomized controlled trial conducted between March 2017 and August 2019. A total of 85 malnourished hospitalized patients at risk of RFS according to the National Institute for Health and Clinical Excellence tool for detecting patients at risk of RFS, were included. All patients were provided with enteral tube feeding, and electrolytes were measured daily during the intervention period. Friedli and ASPEN included phosphate, magnesium, and potassium in their definitions, but used different cut-off values. Incidences were recorded, and Kaplan-Meier estimates were used to determine whether mortality was more prevalent in patients with RFS. Regression analysis was used to test for confounders regarding the association between RFS and death, and Kappa was used to test for agreement between the three diagnostic criteria.

Results: The mean (SD) age of the patients was 79.8 (7.4) years, and the mean (SD) BMI was 18.5 (3.4) kg/m. The mean (SD) kcal/kg/day was 19 (11) on day one and 26 (15) on day seven. The incidences of RFS differed with the criteria used; 12.9% (TC), 31.8% (Friedli), and 65.9% (ASPEN). Mortality was high, with 36.5% (n = 31) and 56.5% (n = 48) of patients dead at three-month and one-year follow-up, respectively. In the TC, 8/11 (72.7%) with RFS vs. 40/74 (54.1%) without RFS died within one-year, in Friedli 15/27 (55.5%) with RFS vs. 33/58 (56.9%) without RFS died, and in ASPEN 32/56 (65.9%) with RFS, vs. 16/29 (55.2%) without RFS died within one-year. There was no statistically significant difference in mortality between patients with or without RFS regardless of which criteria were used. Age was the only variable associated with death at one-year. The Kappa analysis showed very low agreement between the categories.

Conclusion: Our results show that using different diagnostic criteria significantly impacts incidence rates. However, regardless of criteria used, the mortality was not significantly higher in the group of patients with RFS compared to the patients without RFS. Furthermore, none of the criteria showed a significant association with death at one-year. This supports the need for a global unified diagnostic criterion for RFS. This study was registered in ClinicalTrials.gov (identifier NCT03141489).