Use of body composition and phase angle analysis for the assessment of nutritional status and clinical outcomes in critically ill children

Background Malnutrition is highly prevalent in critically ill children in the pediatric intensive care unit .We aimed to investigate the eciency of bioelectrical impedance analysis (BIA) measurements and phase angle (PhA) analysis for the assessment of nutritional risk and clinical outcomes in critically ill children. Methods This single-center observational study included patients admitted to the Pediatric Intensive Care Unit (PICU) of Chengdu Women’s and Children’s Central Hospital. All patients underwent anthropometric measurement in the rst 24 h of admission and underwent BIA measurements within 3 days after the admission. The patients were classied into different groups based on body mass index (BMI) for age. Electronic hospital medical records were reviewed to collect clinical data for each patient. All the obtained data were analyzed by the statistics method. We found that BMI, arm muscle circumference, fat mass, and %body fat were lower in the group with poorer nutritional status (P < 0.05). Evident differences in the score of the Pediatric Risk of Mortality and the duration of mechanical ventilation (MV) among the three groups with different nutritional statuses were observed (P < 0.05). Patients in the severely malnourished group had the longest duration of MV. In the MV groups, there were signicant differences (P < 0.05) in albumin level, PhA, and extracellular water/total body water (ECW/TBW ratio). The ECW/TBW ratio and the time for PICU stay had a weak degree of correlation (Pearson correlation coecient = 0.375). PhA showed a weak degree of correlation with the duration time of medical ventilation (coecient of correlation = 0.398).


Introduction
Malnutrition is highly prevalent in critically ill children in the pediatric intensive care unit (PICU). Critically ill children usually experience total body water redistribution, systemic metabolic status changes, and rapid lean body mass loss, thereby leading to their body composition changes dramatically. However, weight loss, body mass index (BMI), and biochemical indicators do not accurately re ect the nutritional status of critically ill children. Nutritional therapy for critically ill children must be based on a precise assessment of the changes in body composition and uid status. Thus, adopting proper nutritional assessment is signi cantly important.
Bioelectrical impedance analysis (BIA) has been used to assess body composition by measuring the amount of water in and out of cells and fat and muscle mass. Previous studies have shown that BIA can be useful for assessing the prognosis of critically ill patients [1][2][3]. Moreover, BIA is a portable, easy-touse, non-invasive, and low-cost method. For this reason, coupled with children's comfort and cooperation, BIA is signi cantly appropriate for children in daily clinical practice. BIA estimates body composition by measuring the impedance to an applied current while passing through the body. Impedance consists of two components: resistance and reactance. Tissues have different contents of uid and conducting electrolytes; thus, the resistance of tissues is different. Cell membrane and tissue interfaces can affect capacitive reactance. BIA measurement of fat mass (FM) and fat-free mass (FFM) can reliably re ect body composition and is associated with patient's nutritional statuses. BIA-derived phase angle (PhA) is calculated using the following equation: . PhA can be used when assessing the integrity of the cellular membrane and when evaluating cell mass and hydration status. In addition, PhA has been a good predictor of morbidity and mortality in different clinical situations. The European Society for Clinical Nutrition and Metabolism recommends PhA as a screening tool to identify patients at risk of deteriorating nutritional status and functionality [4,5], but studies assessing PhA among critically ill children are scarce.
This study aimed to explore the e ciency of BIA measurements for the assessment of nutritional risk and clinical outcomes in critically ill children and demonstrate whether PhA or other BIA-derived indexes were associated with adverse clinical outcomes.

Study Population
This study comprised patients admitted to the PICU of Chengdu Women's and Children's Central Hospital.
This study was conducted between March 2019 and February 2021 and recruited patients aged 1-14 years. The participants underwent BIA within 3 days after admission. The exclusion criteria were as follows: patients with any amputation, with skin injury on the area where the electrodes of the BIA instrument would be placed, those who underwent dialysis 2 h before BIA, those with congenital chromosomal abnormalities and hereditary metabolic diseases, and those with serious errors in their BIA results. The children's legal guardians provided written informed consent for their children's participation in the study.

Anthropometric measurement and classi cation of nutritional status
Anthropometric measurement was performed in the rst 24 h of admission. To reduce the possibility of errors, all measurements were performed by a trained PICU physician. Weight was measured using a scale that was calibrated for accuracy before each use. Infants were weighed using a scale accurate to 5 g. Children who could not be weighed standing were held by an adult. The child's weight was obtained by subtracting the weight of the adult from the total weight of the child and adult. Children aged 3 years or younger assumed the supine position for the measurement. Length was measured using a pediatric anthropometer with an accuracy of 0.1 cm. In children aged older than 3 years, height was measured using a stadimeter with an accuracy of 0.1 cm. In children whose condition prevented the use of conventional measuring techniques (e.g. patients who were mechanically ventilated or taking vasoactive drugs and above 1 m), the ulna length was measured using a pediatric anthropometer. Height and length prediction was extrapolated by Gauld et al. [6]. Arm muscle circumference (AMC) was measured using a metric tape marked in 0.5-cm increments. Measurements were taken at the midpoint of the distance between the acromion and olecranon with the arm extended along the body.
BMI was calculated using the following equation: BMI/A = W (kg)/H 2 (m). Nutritional status was classi ed based on BMI for age using the World Health Organization (WHO) growth charts as the reference. Patients were categorized as the non-malnourished group, moderately malnourished group, and severely malnourished group, de ned by weight for age 0 to − 2 standard deviation (SD),−2 to − 3 SD, and less than − 3 SD of the WHO growth charts, respectively.

Demographic, clinical, and biochemical data
Electronic hospital medical records were reviewed to collect data for each patient, including age, diagnosis, height, weight, Pediatric Critical Illness Score, Pediatric Risk of Mortality (PRISM) score, length of hospital stay, length of PICU stay, duration of mechanical ventilation (MV), and other notable characteristics. Blood was drawn within 24 h of admission. Serum albumin level, total lymphocyte count (TLC), and hemoglobin level were also collected from the electronic medical records.
Bioelectrical impedance analysis data InBodyS10 (Biospace, Seoul, South Korea) was used for the measurements of medical purpose. Body composition was measured using BIA with the patient in the supine position. The patients' arms were separated from the trunk, and both legs were separated from each other. Subsequently, surface electrodes were placed on the patient's thumbs and middle ngers and two sides of the ankles. The BIA data gathered included intracellular water, extracellular water (ECW) and total body water (TBW), ECM/TBW, body cell mass, bone mineral content, skeletal muscle mass, FM, %body fat (%BF), TBW/FFM, proteins, and minerals. Upon resistance and reactance results, the PhA score was calculated.

Statistical analysis
Statistical analysis was performed using the Statistical Package for the Social Sciences version 20.0 software. Variables are expressed as number (%), mean ± SD, or median (interquartile range). Continuous variables were compared and analyzed based on the independent sample t-test and analysis of variance. Pearson's correlation analysis was used to evaluate the association between the BIA results and duration of MV, length of ICU stay, and length of hospital stay. P value < 0.05 was considered statistically signi cant.

Results
We enrolled 204 patients. All these patients were divided into three groups according their BMI. Among them, 138 were non-malnourished, 27 were moderately malnourished, and 39 were severely malnourished. There were signi cant differences in ages and BMI among the three groups (P < 0.05).We also found signi cant differences in the score of PRISM and the duration of MV in each group (P < 0.05). In the severely malnourished group, patients had the longest duration of MV (Table 1). Table 2 provides a comparison of the biochemical, BIA data, and PhA among the three different nutritional statuses.There were no signi cant differences in all biochemical test result (including albumin, TLC, and hemoglobin), but BMI, AMC, FM, and BF% were lower in the group with poorer nutritional status (P < 0.05). PhA showed no signi cant difference among these three groups. The MV group had signi cantly lower albumin levels compared with the non-MV group. We also found signi cant differences (P < 0.05) of PhA and ECW/TBW ratio between the two groups ( Table 3). Table 4 shows the correlations between the BIA results and duration of MV, length of PICU stay, and length of hospital stay. A weak positive correlation was found between the duration of MV and PhA, the length of PICU stay, and ECM/TBW. The length of hospital stay had no signi cant correlation with any BIA indicators. The analyses were adjusted for age.

Discussion
Malnutrition is common in critically ill children, with a prevalence rate of 32.4% (66/204) in our study. Compared with critically ill adults, critically ill children are at a higher risk of experiencing malnutrition due to their lower nutritional status when they are admitted in the PICU. Hypermetabolic states and enhanced catabolic processes easily lead to exacerbation of nutritional status during hospitalization [7]. Meanwhile, malnutrition is always associated with an increased morbidity and mortality [8,9]. Thus, early identi cation of the high risk of malnutrition and proper assessment of nutritional status are crucial to improve the patients' outcome.
Traditional anthropometric measurements might not accurately re ect body composition changes in lifethreatening disease states. Overall weight loss or decreased BMI does not differentiate FM from muscle mass. Compared with dual-energy X-ray absorptiometry, BIA can be considered a reliable method for assessing body composition in children [10]. A recent systematic literature review [11] has con rmed that BIA predicted body composition and other body composition tools. Recently, sarcopenia has become an important concern when evaluating patients' nutritional status. Moreover, depleted muscle mass is associated with infectious complications, prolonged duration of MV, longer hospitalization, greater need for rehabilitation care after hospital discharge, and higher mortality [12]. In our study, the malnourished group had longer duration of MV than the normal nutritional status group (P < 0.05). As demonstrated by Jaitovich et al. [13] and Weijs et al. [14], low muscle mass is an independent risk factor for mortality in critically ill patients [15]. Thus, early identi cation of low muscle mass may lead to more timely nutritional support, which may bene t the prognosis of patients. Computed tomography (CT) scan is the gold standard tool used to evaluate sarcopenia. However, the routine use of CT images in detecting low muscle mass in clinical practice is limited by radiation hazard. Therefore, both BIA and mid-upper arm circumference (MUAC) might be well suited for the routine assessment of muscle mass because they are widely available, non-expensive, and relatively easy to perform. In our study, we found that AMCs signi cantly changed different nutritional statuses (P < 0.5). Chiabi et al. [16] observed that the MUAC was a more sensitive mortality predictor compared with weight-for-length/height z-score in children with severe acute malnutrition. Notably, MUAC measurements are in uenced by the variability of observers' measurement technique and cannot discriminate adipose from non-adipose tissue. BIA showed high speci city at detecting low muscle mass in patients [17]. A recent study by Looijaard et al. [18] also showed that BIA and CT identi ed the same critically ill population with low skeletal muscle area on CT scan. Signi cant correlations have been detected for different BIA-derived muscle mass equations and CT-derived measurements (correlations ranging between 0.64 and 0.834). However, BIA also has some limitations. BIA measurements may have underestimated the presence of low muscle mass due to abnormal uid redistribution in critically ill patients [19]. Among younger infants (especially aged less than 6 months), BIA may provide little bene t over anthropometry-based prediction equations.
In this study, lower albumin level, lower PhA values, and high ECW/TBW ratios were observed among patients who needed MVs. Albumin is the most abundant protein in the serum. Leite et al.'s [20] study showed that hypoalbuminemia at PICU admission is associated with a higher 60-day mortality, longer duration of MV, and lower probability of ICU discharge. Our study was consistent to this nding: patients with medical ventilation had lower serum albumin level. Possible causes of decreased serum albumin levels include in ammation and illness, increased vascular permeability, and decreased hepatic synthesis, not necessarily from poor nutrition. We also found that there was no statistical signi cance among the different nutritional status group. Additionally, the half-life of albumin is approximately 20 days, suggesting that in some cases, even if the blood albumin level is normal, the total body protein reserves have begun to decrease, but the extent of this decline has not yet affected the maintenance of albumin. Consequently, measuring serum albumin may not provide clinicians an accurate picture of the patient's nutritional status [21].
Fluid overload and positive cumulative uid balance are common in life-threatening disease states and are positively correlated with adverse clinical outcomes [22]. Volume status assessment is signi cantly important in clinical practice to avoid inappropriate uid therapy. The current methods used to evaluate uid status have limitations. Chung et al.'s [23] study showed that BIA and ECW/TBW ratios are useful and convenient tools used to assess the volume status of patients. According to the result of previous studies, the cut-off value for the assessment of edema was 0.39 [24][25][26]. However, in our study, we measured the ECW/TBW ratio in all patients, and the median ECW/TBW ratio was greater than or equal to 0.39 in the different nutritional status groups. In fact, most of our patients did not display any edema symptoms. We hypothesize that children having higher water content than adults may be associated with a higher ECW/TCW ratio. We also found that the ECW/TCW ratio was higher in the MV group than that in the non-MV group, and the ECW/TBW ratio and the length of PICU stay had a weak degree of correlation (Pearson's correlation coe cient = 0.375). Slobod et al.'s [27] study also showed that a higher ECW/TBW ratio within 24 h of admission was correlated with a longer duration of MV. Therefore, the ECW/TBW ratio can be a good prognostic factor for diseases or as a guide for uid management in critically ill patients [28,29]. BIA has intrinsic limitations in its ability to accurately distinguish between intravascular and interstitial volume in the extracellular compartment. Some recent studies have shown that bioelectrical impedance vector analysis, using the resistance and reactance values obtained directly from BIA, could be superior to any other parameter of the BIA for evaluating the hydration of critically ill patients in the ICU [30,31].
BIA-derived PhA re ects the integrity of cell membranes and hydration status and is in uenced by acute illness and general health. A low PhA always indicates cell membrane breakdown and decreased ability to store energy and complete metabolic functions [32]. Thus, it has been considered a prognostic, health, functional, and nutrition indicator [33-35]. Some previous studies have shown that PhA is a relatively sensitive indicator for monitoring and evaluating the nutritional status of patients. Sometimes, it is better than anthropometry and blood biochemical analysis [36][37][38][39][40]. However, PhA showed no signi cant difference among different nutritional status groups in our study. A systematic review has concluded that PhA is not an independent indicator of malnutrition [40] because it is affected by factors, such as age, sex, level of physical activity, uid status, and body composition [41]. To account for such confounders, the calculation of a standardized PhA (SPhA) was proposed to solve this problem. An SPhA can be calculated as a z-score, which may be based on established population reference values strati ed by a combination of age, sex, BMI, or ethnicity [42][43][44]. This method of reducing bias in BIA parameters in critically ill children requires further validation. Some studies have suggested that PhA revealed a predictive power for mortality than severity scoring systems in an ICU [45]. In our study, we found that patients who received MV had lower PhA (mean, 2.5 [SD 0.3]). Moreover, it showed a weak degree of correlation with the duration of MV (coe cient of correlation = 0.398). Another study has reported similar ndings. Zamberlan's study noted that [46] patients with PhA > 2.8° compared with patients with PhA ≤ 2.8° (P < .0001) showed higher survival rate, and children with lower PhA values were more likely to remain in the PICU. Stapel's [47] study found that a PA < 4.8° at ICU admission was an independent predictor of 90-day mortality (adjusted odds ratio = 3.65, con dence interval: 1.34-9.93, P = 0.011). Compared with critically ill adults, critically ill children had lower PhA. More studies are needed to con rm this nding. Until now, there is no clear PhA reference cut-off values, and it is di cult to associate it with disease and nutritional therapy. All these factors could affect the utility of BIA technology for pediatric clinical use.

Conclusion
Overall, BIA can be considered an alternative way to assess the nutritional status of critically ill children. PhA has a limitation in predicting adverse clinical outcomes, but has a potential role in determining the severity of illness in the PICU. Thus, more high-quality studies are needed to obtain more integrated and clear conclusions.

Con ict of Interest
Con ict of Interest All authors declare that they have no con ict of interests.

Availability of data and materials
The datasets used and analyzed during the current study are not publicly available due to limitations of ethical approval involving the patient data and anonymity but are available from the corresponding author on reasonable request.