Research design and patients

A matched cross-sectional study was conducted in the Pediatric Nephrology and HD Unit of a tertiary children’s hospital between October 2024 and October 2025. The study included 42 cases; all eligible patients were enrolled consecutively in the case group, and 42 matched healthy individuals were included as the control group. Children aged 1–16 years receiving regular maintenance HD for ≥ 6 months were eligible. Exclusion criteria included acute HD, known neuromuscular disorders, limb deformities affecting measurement, or refusal to participate.

Healthy controls were recruited from children attending routine pediatric outpatient visits or those admitted to the hospital for acute conditions. The control group was selected using individual age- and sex-matching. For each child receiving maintenance hemodialysis, a healthy control of the same sex and within ± 1 year of age was recruited whenever feasible. Inclusion criteria for the control group were age- and sex-matching with the patient group and the absence of chronic illness. Exclusion criteria included chronic systemic disease, obesity, known nutritional disorders, use of medications affecting body composition, and acute illnesses associated with dehydration, systemic inflammation, or reduced oral intake.

Sample size: a convenience sample was used due to the limited number of pediatric HD patients.

The sample size was determined as follows: the total number of dialysis cases referred to the pediatric HD unit was 42; therefore, all cases were included in the study as a consecutive sample for the case group, and 42 matched healthy individuals were included as the control group.

Ethics approval: following Institutional Review Board approval (ZU-IRB#751/20-Oct-2024), this study was approved by the Pediatrics Department at Zagazig University Hospitals. Informed consent was obtained from the legal guardians of all participants. This research was conducted in accordance with the ethical principles of the Declaration of Helsinki for research involving human participants.

Biochemical measurements

Serum creatinine, blood urea nitrogen (BUN), uric acid (UA), cholesterol, total protein, albumin, calcium, and phosphorus (Pi) were measured using an automated chemical analyzer. Blood samples were collected immediately before the hemodialysis session (predialysis sampling) to minimize the effect of dialysis-related biochemical fluctuations. Biochemical analyses were performed in the central laboratory of the study institution using standardized automated laboratory methods [18]. These parameters were included to describe the clinical characteristics and were not used as diagnostic criteria for nutritional risk.

Nutritional status assessment and definition of nutritional risk

Nutritional risk was defined using a composite assessment based on anthropometric indices and dietary intake according to Pediatric Renal Nutrition Taskforce (PRNT) recommendations; this composite definition was used as a proxy measure of nutritional risk rather than a formal diagnosis of protein-energy wasting, as full criteria application requires longitudinal data not available in this study.

Parents or patients were asked to complete a retrospective 3-day food intake recall. This involved recording everything the child drank and ate, including specific amounts, with a focus on daily energy and protein intake. The results were then compared with the recommended intake based on estimated energy requirements (EER) calculations, which accounted for sex, weight, age, and height, using the 50th percentile height-for-weight. Daily caloric and protein intake was compared with the Pediatric Renal Nutrition Taskforce (PRNT)–suggested dietary intake (SDI), which varies by age and sex [19]. Nutritional assessment and dietary adequacy were evaluated according to Pediatric Renal Nutrition Taskforce (PRNT) clinical practice recommendations [4, 7].

Anthropometric data were collected in accordance with WHO growth standards, along with dietary assessment. Calibrated scales were used to measure weight and height. To avoid edema-related weight bias, post-dialysis dry weight was estimated. Weight was measured using a sensitive platform scale, with the children wearing minimal clothing. Anthropometric measurements included weight, height, BMI, BMI-for-height-age, mid-upper arm circumference (MUAC), waist circumference, and skinfold thickness.

Given the high prevalence of short stature in children receiving HD, BMI interpretation was performed with consideration of height-age rather than chronological age when appropriate, using WHO growth standards. Nutritional risk was defined using conventional criteria, including anthropometric indices (BMI-for-height-age z-score <  − 2 SD and MUAC <  − 2 SD) and inadequate dietary intake based on PRNT recommendations [4]. US-derived muscle thickness was evaluated as an adjunct marker and was not used as a standalone diagnostic criterion.

US measurements

A low-frequency curvilinear-array transducer (3.5 MHz) and a high-frequency linear-array transducer (7 MHz) were used with a Siemens ACUSON X300 ultrasound machine. All examinations were performed after completion of the hemodialysis session to minimize the influence of fluid overload and fluid shifts. Participants were positioned supine or slight lateral decubitus position as appropriate, and all measurements were performed on the right side.

Muscle thickness measurements

Quadriceps muscle thickness was assessed for the rectus femoris (RF), vastus intermedius (VI), vastus medialis, and vastus lateralis muscles. Children were examined in the supine position with the leg semiextended and relaxed (Supplementary Fig. 1). A transverse (cross-sectional) approach was used at standardized anatomical landmarks. Muscle thickness was measured as the distance between the superficial and deep fasciae of each muscle at the midpoint and at the junction between the upper two-thirds and lower third of the distance from the anterior superior iliac spine to the upper pole of the patella [12].

Fat thickness measurements

Fat distribution was evaluated by measuring visceral fat thickness (VFT), anterior abdominal subcutaneous fat thickness (AASFT), and subcutaneous thigh fat thickness (STFT). VFT was measured in the epigastric region using the 3.5-MHz transducer with the participant in the supine position and the knees slightly flexed. The distance from the peritoneal line to the posterior aspect of the rectus abdominis muscle was recorded. AASFT was measured using the 7-MHz transducer placed perpendicular to the skin in the midline above the umbilicus, and the hypoechoic layer between the skin surface and the anterior rectus sheath was recorded. STFT was measured at the mid-thigh level with the participant in the supine position and the leg extended, as the distance from the skin surface to the superficial fascia of the quadriceps muscle group.

Data recording and analysis

All US measurements were performed by a single trained radiologist. All measurements were taken in triplicate, and the mean value was used for analysis. Care was taken to maintain a consistent probe pressure and angle to avoid measurement bias. The assessor was blinded to the clinical data to reduce measurement bias.

Statistical analysis

Following data collection, Microsoft Excel 2016 for Windows, part of the Microsoft Office 2016 package (Microsoft Corporation, USA), was used to code the data into a spreadsheet. The data were then analyzed using IBM Statistical Package for the Social Sciences (SPSS) software, version 27 (IBM, USA). The normality of continuous data distribution was assessed using the Kolmogorov–Smirnov test. Data were presented as numbers and percentages for qualitative variables; means, standard deviations, and ranges for quantitative variables with a parametric distribution; and medians and interquartile ranges (IQRs) for quantitative variables with a nonparametric distribution.

For comparisons between two independent groups, Student’s t-test was used for normally distributed variables, and the Mann–Whitney U test was used for nonnormally distributed variables. The chi-square test was applied to assess associations between categorical variables. Correlation analysis was performed using Spearman’s correlation coefficient (r) to evaluate associations between quantitative variables. Receiver operating characteristic (ROC) curve analysis was used to determine the diagnostic performance of US measurements. A p value ≤ 0.05 was considered statistically significant [20]. The primary purpose of our exploratory ROC analysis was to evaluate whether ultrasound-derived muscle thickness could accurately identify children already classified as being at nutritional risk according to conventional clinical and dietary assessment methods.

Nutritional risk was defined using a composite reference standard based on anthropometric indices (BMI-for-height-age z-score <  − 2 SD and/or MUAC <  − 2 SD) and inadequate dietary intake, as defined by PRNT recommendations. The area under the curve (AUC), optimal cutoff values, sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) were calculated. A p value ≤ 0.05 was considered statistically significant.

A small proportion of data points was missing for laboratory and dietary variables: hemoglobin (2/42), albumin (1/42), and daily protein intake (3/42). Single imputation using mean substitution was applied because of the limited sample size. Missing values were assumed to be missing at random. Although multiple imputation is generally recommended, the very small proportion of missing data and the limited sample size reduced the anticipated benefit of more complex imputation approaches.

Ultrasound-derived muscle and fat thickness measurements were not included in the composite nutritional reference standard used to classify nutritional risk for ROC analysis.