Sleep Quality and Biomarkers in Adolescent Girls With Anorexia Nervosa

Anorexia nervosa (AN) is a severe psychiatric disorder defined by restrictive food intake, significant weight loss, distorted body image, and an intense fear of weight gain, resulting in a three times higher mortality rate among individuals with AN than in the general population. The peak incidence occurs at 14 years of age, with a prevalence of approximately 1.7% among adolescent girls. While early intervention during adolescence is associated with improved outcomes, chronic illness persisting into adulthood often results in recovery rates of only 20-30%.

Sleep disturbances are a clinically significant yet systematically underinvestigated aspect of AN. Almost 50% of individuals with AN report sleep difficulties, irrespective of disorder subtype. A 2024 meta-analysis demonstrated that patients with AN exhibit changes in sleep architecture resulting in significantly shorter total sleep time (TST), reduced sleep efficiency (SE), prolonged wake after sleep onset (WASO), increased N1 non-REM sleep, and reduced REM sleep compared to healthy controls, and especially reduced slow-wave sleep (SWS). Reduced quality of sleep is associated with poorer treatment compliance, increased therapy discontinuation, and a reduced probability of recovery. Notably, weight restoration alone does not regularly normalize sleep architecture, indicating that sleep disturbances require targeted clinical attention independent of nutritional rehabilitation.

The mechanisms underlying sleep disturbances in AN remain unclear. Malnutrition induces hormonal alterations, including elevated ghrelin and orexin-A, decreased leptin, and increased cortisol, that, combined, promote wakefulness and disrupt circadian rhythms. AN is uniquely associated with a morning chronotype, in contrast to most other psychiatric conditions, which are linked to eveningness; this distinction has recently been confirmed by Mendelian randomization studies. Comorbid depression and anxiety, which are highly prevalent in AN, further induce sleep difficulties via overlapping neurobiological pathways.

The majority of current literature has focused on adult populations. Adolescent-specific data remain limited and methodologically inconsistent, with most studies relying on questionnaire-based measures or laboratory polysomnography, both of which are subject to first-night effects.

To date, no published study has combined objective home-based sleep measurement with comprehensive longitudinal biomarker assessment specifically in early-to-mid-adolescent girls with AN. The increasing frequency of pre-menarchal presentations in clinical practice further indicates the necessity for pediatric-focused evidence. In addition to sleep, peripheral biomarkers in AN remain incompletely characterized in adolescents. Data from adult cohorts indicate elevated inflammatory cytokine levels, low-grade systemic inflammation, altered thyroid and prolactin function, dyslipidemia, and reduced 25-hydroxyvitamin D. The neutrophil-to-lymphocyte ratio (NLR), a readily available marker of physiological stress and systemic inflammation, is dysregulated in adult AN. However, the applicability to the adolescent population and their evolution over time with treatment remain insufficiently investigated.

This study identifies two primary gaps: the absence of objective home-based sleep data and the lack of longitudinal biomarker trajectories specifically in adolescent girls with AN undergoing structured psychiatric treatment.

The study will be conducted at the Day Hospital for Children and Adolescents, Department of Psychiatry, University Hospital of Split (KBC Split), Croatia. Participants will be consecutive female patients aged 10-18 years diagnosed with anorexia nervosa (ICD-10: F50.0, F50.1, F50.9) referred to and treated by a child and adolescent psychiatrist at KBC Split.

At enrollment, a detailed history will be collected, including developmental, family, and menstrual history (menarche status and menstrual irregularities), and social background. Parental heteroanamnestic data will be collected separately. Physical examination will include vital signs (heart rate, blood pressure, body temperature) and anthropometric measures (body weight, height, BMI).

All assessments are conducted at two time points:

T0 - Baseline (study enrollment, prior to or at initiation of Day Hospital treatment) T1 - 12-month follow-up (after one year of structured multimodal treatment)

At both time points, participants will undergo:

OBJECTIVE SLEEP ASSESSMENT Home polysomnography using the NOX A1 portable wireless device (Nox Medical). The NOX A1 is a full-channel ambulatory polysomnographic system that eliminates the discomfort associated with traditional in-laboratory PSG and enables sleep recording under naturalistic home conditions. The device records electroencephalography (EEG), electrooculography (EOG), chin electromyography (EMG), electrocardiography (ECG), respiratory effort (thoracic and abdominal bands), nasal airflow, pulse oximetry, body position, and actigraphy. Sleep staging will follow AASM 2023 scoring criteria. Participants will self-apply the device at home following standardized written and verbal instructions. Derived variables: TST (minutes), SE (%), SOL (minutes), WASO (minutes), N1%, N2%, N3% (SWS), REM%, REM latency (minutes), arousal index.

SUBJECTIVE SLEEP AND SLEEPINESS ASSESSMENT Pittsburgh Sleep Quality Index (PSQI) - 19-item self-report assessing sleep quality over the previous month; global score >5 indicates clinically poor sleep quality.

Epworth Sleepiness Scale for children and adolescents (ESS-CHAD) - 8-item self-report measure of daytime sleepiness; score >10 indicates excessive daytime sleepiness; participant version completed by the adolescent; ESS-CHAD for parent/caregiver version completed independently by parent or legal guardian (Croatian validated version) Sleep habits questionnaire - structured assessment of sleep timing, duration, screen use, and sleep behavior

EATING DISORDER PSYCHOPATHOLOGY Eating Disorder Examination Questionnaire (EDE-Q) - 28-item self-report; global score and four subscales (Restraint, Eating Concern, Shape Concern, Weight Concern); higher scores indicate greater eating disorder psychopathology EMOTIONAL AND PSYCHOLOGICAL ASSESSMENT Depression Anxiety Stress Scale - 21 items (DASS-21) - self-report measure of depression, anxiety, and stress symptom frequency and severity over the preceding week; Subscale scores doubled per standard scoring to yield scores comparable to the full DASS-42 Avoidance and Fusion Questionnaire for Youth - 8 items (AFQ-Y8) - self-report measure of psychological inflexibility, experiential avoidance, and cognitive fusion in children and adolescents

QUALITY OF LIFE KIDSCREEN-52 - 52-item health-related quality of life instrument for children and adolescents (10 dimensions: physical well-being, psychological well-being, moods and emotions, self-perception, autonomy, parent relations, social support and peers, school environment, social acceptance, financial resources); both participant self-report version and parent proxy version were completed independently

LABORATORY BIOMARKER ASSESSMENT Blood samples collected under standardized fasting conditions (minimum 8-hour fast) at both time points: Inflammatory markers: complete blood count with differential, neutrophil-to-lymphocyte ratio (NLR) and C-reactive protein (CRP, mg/L).

Hormonal and nutritional markers: free thyroxine (fT4), free triiodothyronine (fT3), prolactin, 25-hydroxyvitamin D (25OH-D).

Metabolic markers: total cholesterol, LDL cholesterol, HDL cholesterol, triglycerides TREATMENT DOCUMENTATION

For each participant, the following will be systematically recorded throughout the 12-month treatment period:

• Pharmacotherapy: agent, dose, duration, and any modifications
• Psychotherapy: modality, number of sessions attended, adherence
• Physical and psychiatric status at each clinical contact
• Adverse events and side effects
• Treatment discontinuation and reasons TREATMENT PROGRAMME

All participants are enrolled in the structured Day Hospital Program for Children and Adolescents at the Department of Psychiatry, KBC Split. The program comprises:

1. Acceptance and Commitment Therapy (ACT) combined with Cognitive Behavioral Therapy (CBT) - group and individual format targeting cognitive distortions, body image, emotions, and values-based behavior change
2. Supportive individual psychotherapy - addressing motivation for recovery, self-esteem, and interpersonal functioning
3. Family-based work and parent sessions - psychoeducation, family communication, and caregiver support
4. Psychoeducational interventions - nutrition, body development, sleep hygiene, and emotion regulation
5. Pharmacotherapy - prescribed where clinically indicated by the treating child and adolescent psychiatrist; type, dose, and duration recorded as covariates

STATISTICAL ANALYSIS PLAN

All statistical analyses will be performed using IBM SPSS Statistics (version 29.0 or later) and R (version 4.3.0 or later). A two-tailed significance level of p < 0.05 will be applied throughout. Given the exploratory and hypothesis-generating nature of this study, both uncorrected and Bonferroni-corrected p-values will be reported for multiple comparisons. Continuous variables: mean ± standard deviation (SD) for normally distributed data; median and interquartile range (IQR) for non-normally distributed data. Normality will be assessed using the Shapiro-Wilk test. Categorical variables: frequencies and percentages.

WITHIN-SUBJECT CHANGE (T0 TO T1) Paired-samples t-test (parametric) or Wilcoxon signed-rank test (non-parametric) will be used to assess the change in continuous outcome measures from baseline to 12 months follow-up. Effect sizes will be reported as Cohen's d (parametric) or rank-biserial correlation r (non-parametric). Spearman's rank correlation will be used to examine associations among sleep parameters, biomarker levels, psychopathological scores, and quality-of-life measures at both time points and for change scores (ΔT1-T0). Multivariable linear regression will be applied to identify independent predictors of: (1) sleep architecture parameters at baseline; (2) change in EDE-Q global score at 12 months. Covariates will include age, BMI z-score, illness duration, depression score (DASS-21), and pharmacotherapy status. The assumptions of linearity, independence, homoscedasticity, and the absence of multicollinearity will be verified prior to modeling. Longitudinal changes in biomarker panels will be examined using repeated-measures analysis or mixed-effects models, as applicable, with pharmacotherapy as a between-subject factor.

SAMPLE SIZE CONSIDERATIONS This is a single-center prospective cohort study with consecutive enrollment. A formal a priori power calculation is limited by the absence of comparable pediatric polysomnographic data in AN. Based on estimated annual referral rates at the KBC Split Day Hospital Program. A sample of 30-50 eligible participants is anticipated over the enrollment period. The study is, therefore, appropriately characterized as exploratory and hypothesis-generating, with findings intended to inform the design of adequately powered future multi-center studies. The study will be conducted in accordance with the Declaration of Helsinki and applicable Croatian and European Union regulations on clinical research and data protection (GDPR).

Ethical approval will be obtained from the Ethics Committee of the University Hospital of Split (KBC Split) and the Ethics Committee of the University of Split School of Medicine prior to enrollment of any participant. Written informed consent will be obtained from each participant's parent or legal guardian. Written assent will be obtained from each participant. Participation is voluntary, and withdrawal at any time will not affect the quality of clinical care provided. All data will be pseudonymized and stored in accordance with institutional data protection protocols.