Changes in anthropometrical status and body composition in children with cancer during initial chemotherapy

Abstract Children with cancer require adequate nutritional support to prevent malnutrition. This study investigated the impact of chemotherapy on anthropometrical status and body composition during the first six months of treatment. Anthropometrical status and body composition were measured at diagnosis, utilizing standardized protocols and validated S10 InBody bio-electrical impedance (BIA) measurements and compared to subsequent consecutive monthly follow-up measurements to plot changes over time during the first six months. Statistical significance was defined as p < 0.05. Forty-three newly diagnosed children (median age 4 years, IQR: 2.0-7.6; male-female ratio 1:0.9; 53% haematological malignancies and 47% solid tumors) were included. Prevalence of malnutrition varied, with under-nutrition 14% (mid-upper arm circumference (MUAC)/body mass index (BMI)), over-nutrition 9.3% (BMI) and stunting 7% at diagnosis. MUAC (14%) identified fewer participants with underlying muscle store depletion than BIA (41.8%). Chemotherapy exposure acutely exacerbated existing nutritional depletion during the first two months after diagnosis for all variables except fat mass (FM), with contrary effects on cancer type. Haematological malignancies had rapid increases in weight, BMI and FM. All patients had an acute loss of skeletal muscle mass. Nutritional improvement experienced by all cancer types during month two to three of treatment resulted in catch-up growth, with a significant increase in weight (chi2=40.43, p < 0.001), height (chi2=53.79, p < 0.001), BMI (chi2=16.32, p < 0.005), fat free mass (chi2=23.69, p < 0.003) and skeletal muscle mass (chi2=24.19, p < 0.001) after six months. Monthly nutritional assessments, including advanced body composition measurements, are essential to provide timely nutritional interventions to overcome the acute decline in nutritional reserves observed during the first two months of chemotherapy exposure.


Introduction
Children with cancer are faced with many nutritional challenges along their treatment journey that may result in the development of malnutrition, whose profound negative impact on their quality of life, growth and development and clinical outcomes from as early as diagnosis has been well established. [1][2][3] The importance of maintaining an optimal nutritional status through timely nutritional interventions has recently been highlighted with a positive correlation demonstrated between the patient's health related quality of life (HRQL) and 5-year survival and their ability to overcome malnourished states during the first year of treatment. 2,3 The prevalence and degree of malnutrition has traditionally been quantified by anthropometrical status. However, such assessments are merely indicators of body size and not body composition, failing to distinguish between fat and functional tissue (fat free mass). 2,4,5 Although alterations in nutritional status frequently occur during the initial months of treatment, little is known about the exact timing and degree of impact on nutritional status and body composition since it cannot be identified by conventional anthropometrical assessments alone. 5,6 It is pertinent to investigate interactions between cancer drugs and the alteration of said body compartments, 7 as alterations of the metabolically active fat free mass may lead to changes in drug metabolism (volume of distribution, drug absorption, protein binding, decreased oxidative metabolism and glomerular filtration) with detrimental effects on HRQL and chemotoxicity throughout treatment. 2,[8][9][10] The understanding and interpretation of the effects of cancer-driven mechanisms and treatments on body composition through detailed, advanced body composition analysis used in conjunction with anthropometrical assessments 2,8 is thus essential to help navigate and implement prompt, patient specific nutritional interventions that will ensure the best outcome in these patients -throughout all stages of treatment. 2,[6][7][8]11 This study investigated the impact of initial chemotherapy on anthropometrical status and body composition and its subsequent pattern of change during the first six months of treatment to determine when this patient population is most vulnerable for nutritional depletion during this time frame.

Study design
This prospective, descriptive, cohort study was conducted at the Pediatric Oncology Unit (POU) at Tygerberg Hospital, Stellenbosch University, Cape Town, South Africa according to the principles of the Declaration of Helsinki, after obtaining ethical approval from the University of Stellenbosch`s Health and Research Ethics Committee (S18/04/050). Written consent and assent were provided by all caregivers and children older than seven years. Recruited patients were followed up from diagnosis for a maximum of six consecutive monthly visits during the study period from April 2019 until June 2020.

Data collection
Medical information and demographic data were collected at diagnosis. Baseline assessments were performed within 72 hours of diagnosis prior to the initiation of chemotherapy and compared to subsequent monthly visits to plot change over time. Anthropometrical measurements included weight measured to nearest 0.1 kg (0.01 kg for infants; baby scale, Seca 354 and 874), height measured to the nearest 0.1 cm (stadiometer, Seca217) and mid-upper arm circumference (MUAC) using a non-stretchable tape to the nearest 0.1 cm at the mid-point of the upper-arm. Recumbent length was used in children <2 years or those unable to stand. Body compartments assessed included fat mass (FM), fat free mass (FFM), skeletal muscle mass (SMM) and phase angle (PA), measured by a single, calibrated, multi-frequency portable segmental body composition analyzer (Inbody S10, InBody Co Ltd, Korea). The device is validated against dual energy x-ray absorptiometry (Dexa), has proven sensitivity for detecting fluid status in the clinical setting and suitable for pediatric use. [12][13][14] Four touch-type electrodes (sticky-adhesive type electrodes in infants) were placed on both hands and feet and patients were required to remain static in supine position for two minutes without jewelry or metal-containing clothing. This was performed prior to the initiation of hyper-hydration as a precaution to eliminate the potential confounding impact of fluid balance on its readings. Individualized nutritional support (counselling, oral supplements, tube feeds and total parenteral nutrition) was implemented from diagnosis for both in-and out-patients according to the dietician's existing nutrition support protocols.

Data analysis
Anthropometrical status was defined according to age and gender specific World Health Organization (WHO) and Mramba et al. (MUAC for children older than 5 years) classifications 15,16 as under-nutrition (weight-for-age <-2 z-score, body mass index (BMI)-for-age or weight-for-length <-2 z-score, MUAC-for-age <-2 z-score), over-nutrition (weight-for-age > +2 z-score, BMI-for-age > +2 z-score) and stunting (height-for-age <-2 z-score). A z-score <-3 and > +3 was considered severe. Body stores were categorized as high or low according to the pre-programmed InBody S10 specifications as no population standards for healthy children exist in South Africa. In the absence of such 'normal' values for comparative purposes, the monitoring and description of changes over time in relation to baseline values were considered relevant and most appropriate. The absolute values of all continuous variables were used to assess and describe trends in actual monthly measurements. This allowed for insights into underlying body compartment modifications in relation to changes in absolute anthropometrical values over time.
Anthropometrical status and body composition at baseline, as well as changes and associations between baseline and follow-up measurements, were identified utilizing both inferential and basic descriptive statistics. Trends in change over time were expressed in absolute values (median; IQR), percentage changes and the appropriate classification for anthropometrical and body composition status as per z-scores and bio-electrical impedance analysis (BIA). The Friedman rank test (non-parametric) compared median variable values between diagnosis and the end of follow-up and per cancer group. Data were analyzed using the STATISTICA (version 13, TIBCO Software Inc. (2018)) data analysis software system. p < 0.05 indicated statistical significance for all variables.

Results
Forty-three of the forty-seven consecutive newly diagnosed patients aged three months to 15 years with haematological malignancies and solid tumors were recruited and received chemotherapy ( Figure 1). Four patients requiring exclusive radiotherapy or surgical interventions were excluded. Patients who died (n = 5) during the first six months or who required less than six months of chemotherapy (n = 6) were included in the study up to the last of their respective measurements. The male-female ratio was 1:0.9 with a median age 4 years (IQR: 2.0-7.6) ( Table 1). The most common haematological malignancy (n = 23, 53%) was acute lymphoblastic leukemia (n = 13, 56.5%) and nephroblastoma (n = 5, 25%) was the most frequent solid tumor (n = 20, 47%).

Trends in change over time
Several changes occurred for all anthropometrical variables over time of which the acute decrease of median weight (300 g, 2%), BMI (0.4 kg/m 2 , 2.6%) and MUAC (0.7 cm, 4.4%) for all cancer types during the first month after chemotherapy was the most notable (Table 4). Height remained static during this period. This acute decline improved during months two and three to surpass baseline values for all classifications based on S10-inbody pre-programmed interpretation.
variables except MUAC, resulting in a significant increase in weight (chi 2 =40.43, p < 0.001), height (chi 2 =53.79, p < 0.001) and BMI (chi 2 =16.32, p < 0.005) between diagnosis and month five. An improvement in severe wasting, stunting, muscle depletion and overweight was seen by month five, but the number of wasted patients remained the same as at diagnosis and obesity doubled. Cancer sub-groups responded differently to chemotherapy: The solid tumor group experienced an acute decline in month one in weight (800 g, 4.8%), BMI (0.5 kg/m 2 , 3.1%) and MUAC (0.5 cm, 3.2%) in contrast to the gradual monthly increase in weight (100-500 g) and height (0.5-1.5 cm) seen in the haematological malignancy group. Over time the solid tumor group experienced a significant increase in both weight (800 g, 4.8%, chi 2 =14.93, p < 0.01) and height (6 cm Table 4). The initial decrease in low muscle levels involved 62% of the study population after the first month of treatment, affecting both cancer groups. Body store repletion only improved from month three onwards with a subsequent significant increase from baseline values in FFM (700 g, 5.4%; chi 2 =23.69, p < 0.003) and SMM (300 g, 5.5%; chi 2 =24.19, p < 0.001) for all cancer types by the end of follow-up. At the end of the study, 13 (40.6%) of the remaining 32 patients still had a low SMM, of whom seven had a haematological malignancy and six a solid tumor (Figure 2, Supplementary Table S1).The initial decrease in median SMM (1.35 kg, 20.6%) and PA (0.3 degrees, 6.9%) was more pronounced in the solid tumor group, but only the haematological malignancies showed an overall improvement in phase angle (0.4 degrees, 10.8%) and significant increase in SMM (650 g, 13.3%); chi 2 =11.32, p = 0.045) by month five. In contrast to the significant decrease in FM (350 g, 11.5%, chi 2 =0.79, p = 0.98) for the solid tumor group, the steady increase in FM for all cancer types (600 g, 24%, chi 2 =2.17, p = 0.82) and haematological malignancies (1.2 kg, 48%, chi 2 =5.01, p = 0.41) was not statistically significant when compared to baseline values (Table 4). All the patients with a high FM at the end of follow-up had haematological malignancies (Figure 2, Supplementary Table S1).

Discussion
Our findings demonstrate that anthropometrical status could not identify the extent of underlying body store depletion from diagnosis throughout follow-up. Chemotherapy exposure furthermore acutely exacerbated the existing malnutrition and body store depletion during the first two months of treatment with contrary effects on cancer groups. The haematological group was prone to a continuous gain in weight, fat mass and over-nutrition, ending with a significant increase in weight, height, FFM, SMM and BMI by the end of the study. The solid tumor group by contrast experienced a rapid decline in all measured variables during the first two months of chemotherapy to end Table 4. median change in variables over time.    with a significant increase in weight, height and FFM alone. The only similar trend over time between groups was the initial loss of functional tissues (MUAC, SMM, PA) during the first two months of treatment, albeit to varying degrees of severity. Compared to the national South African rate of malnutrition for stunting (< 5 yr: 27.4%, 5-19 yr: 12.9%), wasting (< 5 yr: 2.5%, 5-19 yr: 5%), severe wasting (< 5 yr: 1%) and overweight (< 5 yrs: 13.3%, 5-19 yr: 25%), only the level of baseline wasting (including severe wasting) was worse in our participants than the national prevalence. [17][18][19] These findings are in contrast with higher diagnosis prevalence rates of malnutrition (12-65%) as reported in recent South African publications regarding newly diagnosed children with nephroblastoma, [20][21][22] which may be attributed to probable advanced disease in these studies or the various cancer diagnoses included in our study. Late presentation at healthcare centers with advanced disease were also a cited reason for a higher prevalence and severity of malnutrition in those with solid tumors living in low-middle income countries (LMIC). 3,23,24 We did however find comparative rates of under-and over-nutrition as reported by various LMIC and high income countries studies despite their use of weight and BMI z-scores alone. 2,[23][24][25][26] A much larger proportion of patients suffered from decreased fat and muscle stores at diagnosis, which may be attributed to the low levels of physical activity and appetite reported upon initial presentation at POUs 2,27 or the onset of the disease itself. 24 The early identification of depleted muscle stores prior to treatment initiation is essential since it results in a loss of functional tissues, strength, immune-and pulmonary function 4,5 and may serve as a poor prognostic marker. 28,29 Our findings concur with those of Yaprak et al. since MUAC and BMI identified more acute malnutrition than weight alone. 24 However, the inability of these variables to identify the severity of muscle depletion in our study highlights the need for incorporating advanced body composition analysis as part of routine assessments to avoid overlooking underlying body store depletion from as early as diagnosis. 4,5,28 Examining trends in change over time revealed an acute decline of nutritional stores during the first two months after diagnosis. Similar longitudinal studies investigating changes in anthropometrical status and body composition over time employed larger, follow-up assessment intervals (usually three monthly) and a variety of methods for body composition measurements that include BIA, deuterium solutions and Dexa. 5,28,[30][31][32] Despite differing methodologies, the current study findings also describe the first three months of chemotherapy as the period that presented with the worst impact on anthropometrical status 28,31,33 and body compartments. 5,28,32 Severe acute changes in anthropometry (weight loss >5% or BMI z-score <1.5) during the first three months of initial chemotherapy may affect between 46-63.7% of children with cancer, resulting in profoundly detrimental effects on HRQL, overall patient survival and infectious outcomes in a variety of malignancies. 2,28,33,34 The acute weight loss seen in our population during the first month of treatment fell short of the 5% cutoff proposed by the American Society for Parenteral and Enteral Nutrition for severe acute weight loss, 35 but the large difference seen in patterns of change among cancer types might have diluted the severity of the overall effect. This illustrated contrast between cancer groups is consistent with reported findings describing those with solid tumors as prone to develop acute malnutrition upon initial chemotherapy, whilst the haematological group remained most affected by chronic malnutrition (stunting) and over-nutrition. 26,28,31 Contrary to the results presented by Brinksma et al. 28 the rapid increase in weight and BMI did not translate to an equally rapid increase in the overweight classification of our participants. Rather, current findings illustrate the value of monitoring both absolute values and z-score classifications in relation to baseline measurements to anticipate nutritional derangements, as a large variance in absolute values (≥10%) was required before a change in z-score classification was seen.
The most critical acute change in body compartments during the first two months after diagnosis impacts stores (SMM, FFM, PA) related to functionality, quality of life and prognosis, 2,9,28,36 affecting both cancer groups albeit to varying degrees of severity. The simultaneous increase in FFM and decrease in SMM may be explained by the haematological malignancy group's exposure to steroids rather than hyper-hydration, 37 whilst the initial decrease in FFM as experienced by the solid tumor group might be attributed to an overall loss of weight or inflammatory processes associated with tumor activity. 28 The profound loss of functional tissues may furthermore be obscured by the large cumulative increase in weight and FM as demonstrated in our study. Similar trends (high FM and the low FFM and SMM) within the haematological group have been well described and may be induced by the use of steroid therapy and its resulting increased appetite, increased levels of reported fatigue with lower activity levels than those of their peers, diminished height velocity during active treatment, active tube feeding and taste changes. 2,5,20,28,32 As a measure of cell integrity and functionality, PA values may change in line with a patient`s response to treatment, the presence of malnutrition or infection. [38][39][40] The varied degree of malnutrition among cancer type may thus explain the difference noted in PÀs patterns of change among our cancer sub-groups. 39,40 The lack of pediatric-specific PA reference values makes the interpretation of results challenging. A range between 4.5°-5.6° has been linked to an improved prognosis in adults with a variety of cancer diagnoses, 39,40 whilst a range below 2.8° and above 5.1° indicated either increased or improved morbidity and mortality in critically ill pediatric and adults patients respectively. 38,41 More studies are therefore needed to establish a threshold value for the use of PA as a proxy for functional capacity within the pediatric cancer population. Until such time, the patterns of change from diagnosis may be a valuable measure of growth and the patient's response to nutritional interventions. 42 The limited sample size secondary to the relative rarity of the underlying cancer diagnosis prevented effective stratification of the research population's response to chemotherapy according to age group, sex, and disease-risk categories. Future research within a larger study population will assist healthcare practitioners to better understand the impact of these factors on the child with cancer`s response to chemotherapy. Our interpretation of catch-up growth was also limited as we did not investigate the relationship between changes in nutritional stores and overall survival due to the short time span (six months) of the study period. We furthermore acknowledge that the pre-programmed BIA classifications utilized within this study are not validated for all populations. Results should therefore be interpreted with caution in conjunction with patterns of change over time according to the individual`s response to treatment.
This study adds to the growing body of knowledge describing the negative impact of chemotherapy exposure on nutritional reserves from as early as diagnosis. The consequent acute changes in anthropometrical status and body composition for all cancer types should therefore be assessed by differentiating between body size and body composition in pursuit of a more thorough representation of the child's overall growth patterns and the body's nutritional reserves. 4,28 Additionally, we illustrated a trend toward nutritional store improvement from as early as two (anthropometrical status) and three months (body composition) post chemotherapy initiation, therefore minimum monthly assessments are warranted to avoid such severe, acute changes requiring immediate nutritional support. This study utilized BIA due to the lack of access to Dexa or computerized tomography scans as an accessible, child-friendly alternative 9,28,43 with demonstrated improved accuracy in the estimation of hydration status than single frequency devices. 36 Our findings regarding nutritional repletion attest to implementing prompt nutritional interventions at the patient's bedside the moment nutritional derangements become apparent, as facilitated by our mobile BIA device. However, MUAC`s ability to mimic the patterns of change in functional tissues over time demonstrates its value for use in the absence of advanced body composition assessments, especially in resource constraint settings. Small changes in trends over time should be carefully considered and interpreted as they may represent larger underlying changes in functional tissues requiring immediate intervention. We therefore conclude that frequent follow-up assessments based on an evaluation of patterns of change in comparison to the patient's unique baseline results may be more appropriate to determine outcomes than large assessment intervals based on z-score and pre-programmed classifications alone.

Conclusion
Children with cancer requiring initial chemotherapy are most vulnerable for nutritional depletion during the first two months of treatment, with a demonstrated contrary effect between haematological and solid malignancies. Anthropometrical status, including MUAC, failed to identify the extent of underlying body store depletion and should accompany advanced body composition techniques to thoroughly assess nutritional reserves. Early identification of malnutrition through a minimum of monthly assessments from diagnosis may result in timely nutritional interventions and nutritional repletion. The first two months of treatment exposure therefore becomes the window of opportunity to serve as basis for timely nutritional interventions that preempt, rather than react to the presence of malnutrition in order to pave the way for improved HRQL throughout their cancer journey onto survival.