Study design and participants
This naturalistic, retrospective observational study was conducted within the framework of an observational research that aimed to investigate the utilization of psychopharmacological treatments in a tertiary-level Italian Regional Center specialized in FED in Children and Adolescents. It received approval from the local ethical committee under the code NPI-DAPSIFA2020. The study adhered to the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines both during the planning and implementation phases [
16]. This study did not receive any sponsorship or funding from external companies.
The research was carried out in July 2022 and involved a retrospective analysis of patients who had been assessed at the study Center during the period from January 1, 2016, to December 31, 2020, and had undergone at least one hospitalization due to FED at the same Center. Hospitalization, in this context, refers to either inpatient or day hospital treatment. Notably, the day-hospital treatment program for patients with FED is designed to be equally structured and intensive as inpatient treatment. The hospital’s treatment protocol involves a comprehensive, multidisciplinary approach that encompasses psychological, psychopharmacological, and nutritional interventions, conformed to established clinical international guidelines. The treatment protocol adopted in the management of the patients here included has been discussed in two previous papers published by our group, reporting the endocrinological [
17] and psychopathological [
18] features of the population admitted to our Center; the same population was assessed to be included in this research, with study-specific inclusion and exclusion criteria.
For patients diagnosed with AN, the Center provides various nutritional therapies, including oral dietary plans, nutritional supplements, nasogastric tube feeding (NGT), and parenteral nutrition. The oral diet emphasizes balanced meals for weight management, while nutritional supplements are offered to those who may have difficulty meeting their nutritional requirements. NGT is employed in cases of severe undernourishment or when oral feeding presents challenges. In circumstances where alternative approaches are unfeasible, parenteral nutrition is administered intravenously. These interventions are personalized to each patient’s specific needs and are the result of careful assessments and consultations with the multidisciplinary team.
The Center serves a dual role, providing psychiatric hospital care for AN and acting as the primary healthcare facility for individuals with severe underweight conditions and metabolic impairment. While the primary approach is to avoid parenteral nutrition, this intervention is reserved for exceptional cases when alternative interventions are impractical or insufficient. To provide a naturalistic depiction of RS incidence and severity, it is important to note that parenteral nutrition is considered a potential risk factor. Despite its impact on electrolyte balance, we intentionally included such cases to comprehensively assess challenges and outcomes, offering a holistic understanding of the real-world clinical scenario.
Concerning daily caloric intake, at our Centre clinicians usually started refeeding with 15–20 kcal/Kg/day in the first 24 h for patients at intermediate risk, followed by a progressive increase every 2–3 days based on plasma electrolyte levels and clinical circumstances. To reduce the risk of underfeeding and refeeding syndrome, high-risk patients consumed 5–10 kcal/kg/day of calories on the first days, increasing to a maximum of 20 kcal/kg/day in the first weeks. More rapid, AN-specific, protocols for renutrition have been published after patients enrolled in this study were hospitalized [
5].
To be eligible for inclusion in this study, individuals had to meet the following criteria: (a) a diagnosis of AN based on the criteria outlined in the DSM-5; (b) the provision of informed consent; (c) a documented history of hospitalization at our Center during the relevant time frame; and (d) two blood samples, including all the three RS-involved electrolytes (phosphorus, potassium, magnesium). The exclusion criteria in this study were (a) the absence of sufficient clinical documentation, (b) a first blood sample taken more than 1 day after the start of the treatment program, and (c) a second blood sample taken more than 1 month after the first blood sample or occurring more than 5 days of reinitiating or substantially increasing energy provision.
Patients with evidence of RS were then included in the RS group according to the ASPEN criteria [
2]. Individuals who did not develop RS were included in the no-RS group. The selection of the 2 groups was performed including all the patients consecutively undergoing the same hospital treatment during the selected period, to provide an unbiased and naturalistic observation. Given the naturalistic nature of the study, missing data were not replaced.
Assessment methods
Patients underwent a thorough assessment for FED, covering psychopathological, nutritional, and biochemical screenings upon hospital admission. Collected data included demographic information, clinical factors (AN subtype, comorbidities, duration of untreated illness, use of nasogastric tube feeding, and parenteral nutrition), and anthropometric parameters (admission and discharge %BMI and BMI). The reference values for calculating %BMI (BMI/median BMI for age and gender × 100) were taken from the World Health Organization’s tables [
19].
AN diagnoses and comorbidities were determined by specialists using DSM-5 criteria. Standardized laboratory tests, including blood counts and electrolyte assessments, were conducted throughout hospitalization.
Detailed data on RS were recorded, including energy intake (kcal/kg/day) and electrolyte levels, before and during RS. Electrolyte changes were measured ad a percentage between initial and subsequent samples. The laboratory methods used for blood measurements included UV photometry for phosphorus, indirect potentiometry for potassium, and colorimetric kinetic method for magnesium. Laboratory reference values were as follows: phosphorus, 2.5–4.5 mg/dl; potassium, 3.5–5.3 mmol/L; and magnesium, 1.6–2.6 mg/dl.
Variables related to psychopharmacological treatments were assessed by thoroughly reviewing the clinical documentation, which included details about the timing and duration of treatment, initial and maximum dosages, any reasons for discontinuation of treatment, and potential adverse drug reactions (ADR) that may have arisen. It is essential to note that all patients received a multidisciplinary hospital-based treatment for AN, which encompassed both individual and group psychotherapeutic interventions.
Statistical analysis
The enrolled patients were categorized into two distinct groups based on the occurrence of RS during their respective hospitalizations. Descriptive analyses were conducted for the entire study population as well as for the two specified groups. A significance level of 0.05 was established for all statistical tests, and two-tailed tests were employed. To evaluate the normality of data distribution and the homogeneity of variance, Shapiro–Wilk’s and Levene’s tests were utilized. The two groups were subsequently compared using the Chi-Square test for categorical variables (with the Fisher exact test being employed when necessary due to small sample sizes) and the Student’s T-test for continuous variables (Mann–Whitney U Test was used for non-normally distributed data). The percentage of variation for the ASPEN-relevant electrolytes involved in RS (phosphorus, potassium, magnesium) was compared between patients treated and untreated with olanzapine. To account for the potential confounding effect of baseline electrolyte levels, this analysis was conducted by performing analyses of covariance (ANCOVA) corrected for baseline electrolytes (phosphorus, potassium, and magnesium, respectively). Finally, the potential admission-discharge modifications in %BMI were compared between patients treated and untreated with olanzapine, with a baseline %BMI-corrected ANCOVA.
The determination of the sample size was based on the number of individuals enrolled during the study period. Given the retrospective design of this study focusing on the naturalistic portrayal of real-world data, missing data were not imputed or replaced. This approach was employed to maintain the originality and unaltered nature of the dataset, aligning with the study’s aim to authentically reflect the clinical scenario under investigation. All the statistical analyses were carried out using JASP (Jeffrey’s Statistical Program), version 17.1 for Windows.