Starvation hepatitis and refeeding syndrome in anorexia nervosa: a retrospective cohort study

Anorexia nervosa is associated with severe malnutrition-related complications, including starvation hepatitis, defined by elevated transaminases in the context of advanced protein-energy malnutrition, and refeeding syndrome (RS), defined as a clinically significant decrease in plasma phosphate, potassium, and/or magnesium within five days of resuming caloric intake. Both carry a non-negligible impact on mortality in this population. Although starvation hepatitis has been proposed as a potential link to RS, no study had yet specifically examined this association. This study aimed to assess the prevalence and characteristics of starvation hepatitis, and to examine its association with the risk of developing RS during nutritional rehabilitation in patients hospitalised for severe anorexia nervosa. This was a retrospective, single-centre cohort study conducted at French University Hospital between January 2020 and December 2025, including adult patients hospitalised for more than five days with a diagnosis of anorexia nervosa and no pre-existing liver disease. Clinical, anthropometric, and biological data were collected at admission and on day 5 of refeeding. Univariate analyses and a multivariate logistic regression were performed to identify independent predictors of RS. Sixty-two patients were included (95% female; mean age 28.9 ± 10.6 years; mean BMI 13.47 ± 2.98 kg/m²). Starvation hepatitis was present in 31 patients (50%) at admission. RS occurred in 19 patients (30.6%) overall and was significantly more frequent in patients with starvation hepatitis than in those without (48.4 vs. 12.9%; p = 0.006). On multivariate analysis, starvation hepatitis was the only independent predictor of RS (adjusted OR 3.6; 95% CI 1.08–12.2; p = 0.038), after adjustment for BMI and initial caloric intake. This is the first study to demonstrate a significant and independent association between starvation hepatitis at admission and the subsequent risk of RS in patients with anorexia nervosa. These findings suggest that liver enzyme abnormalities, a simple routinely available marker, could help identify patients at particularly high risk of refeeding complications, warranting enhanced biological monitoring and more cautious nutritional rehabilitation. Anorexia nervosa is a serious eating disorder that can lead to severe malnutrition. When patients with anorexia nervosa are hospitalised and start eating again, two complications can occur in succession: liver injury caused by starvation (called starvation hepatitis), where blood liver enzyme levels become abnormally high, and then refeeding syndrome, where the restart of nutrition causes dangerous drops in important electrolytes such as phosphate, potassium, and magnesium in the blood. Both complications can be life-threatening. In this study, we examined whether patients with pre-existing liver injury on arrival at hospital were more likely to develop refeeding syndrome during nutritional treatment. We analysed the medical records of 62 adult patients hospitalised for anorexia nervosa at French University Hospital between 2020 and 2025. We found that half of all patients had starvation hepatitis when they arrived at hospital. Among those patients, nearly half went on to develop refeeding syndrome, more than 3-fold the rate observed in patients without liver injury. After accounting for other factors such as body weight and the amount of food given initially, liver injury at admission remained the strongest predictor of refeeding syndrome. These results suggest that a simple blood test measuring liver enzyme, which is already routinely performed in hospitals, could help to identify which patients with anorexia nervosa are at the highest risk of refeeding complications. These patients may benefit from even more careful nutritional rehabilitation and closer monitoring of their blood tests.