Assessing Energy Expenditure in Cancer Patients: A Pilot Validation of a New Wearable Device

By Cereda, Emanuele Turrini, Mauro; Ciapanna, Denis; Marbello, Laura; Pietrobelli, Angelo; Corradi, Ettore

ABSTRACT. Background: Nutrition problems are common in cancer patients and are frequently due to metabolic derangements. Thus, accurately assessing energy expenditure (EE) is important in planning adequate nutrition support. Indirect calorimetry (IC) represents the gold standard method but is not always available or applicable to all settings. The purpose of this study was to preliminary compare a new wearable device, the SenseWear armband (SWA), to IC in cancer patients. Methods: Ten (6 M, 4 F) subjects (mean +- SD: 56.6 +- 13.3 years) affected by newly diagnosed acute myelogenous leukemia, undergoing induction chemotherapy, were prospectively enrolled. Resting EE (REE) was measured simultaneously by SWA and IC on admission (day 0) and at discharge (end). Total daily EE (TDEE) was determined by SWA 4 times during the stay (days 0, 7, 14, and end) and predicted values were calculated according to IC REE estimates (TDEE = IC x correction factor 1.2). Results: Mean length of stay was 27.1 +- 6.2 days. Bland-Altman plots revealed no significant differences between overall REE estimates (day 0 + end) performed by IC and SWA (mean +- SD; 1645 +- 282 vs 1705 +- 278 kcal/d) and the correlation was high (r = 0.84; p < .0001). SWA TDEE showed a progressive reduction during the stay. No bias was detected between overall SWA TDEE (1799 +- 153 kcal/d) and IC predicted TDEE (1974 +- 176 kcal/d), but there was a wide 95% confidence interval (-672; +321 kcal/d). Moreover, the correlation between these values was significant (r = 0.68; p = .001). Conclusions: SWA seems to provide accurate and reliable estimation of REE and useful information on TDEE also in cancer patients. Its use appears promising. Validation studies on larger samples and different cancer types should be considered. (Journal of Parenteral and Enteral Nutrition 31:502-507, 2007) Cancer is characterized by various nutrition problems, ranging from localized effects, induced by the tumor on the organ involved, to systemic actions caused by metastases or humoral factors.1,2 These may lead to adverse prognosis.3,4 Moreover, nutrition needs are also increased due to treatment regimens (eg, surgery, chemotherapy, radiotherapy).5-7 Nutrition support should be started when necessary, and the therapeutic goal should be the improvement of function and outcome by (1) preventing and treating undernutrition, (2) enhancing antitumor treatment effects, (3) reducing adverse effects of antitumor therapies and (4) improving quality of life.8 In reaching these targets, accurately determining rates of resting energy expenditure (REE) is important because it determines adequate nutrition support, particularly in cancer patients who are candidates for an enteral or a parenteral approach.9

Although indirect calorimetry represents the gold standard method to assess REE, it is not always available or easily applicable to all settings. Practical predieting equations have been validated for more accurate estimates of REE, but these usually lead to a considerable number of over- and underestimations.8,10,11 Moreover, stress factors9 and assumptions8 have been proposed to calculate total daily energy expenditures (TDEE). A new wearable device for the assessment of REE and TDEE, the Sense Wear system armband (SWA; BodyMedia, Inc, Pittsburgh, PA), has recently been introduced and validated in healthy subjects.12″15 The portability, acceptability to the patient, low cost, and the basic skill required to the personnel are the strength of SWA in clinical practice.14’16 Thus, its use may be of benefit in cancer patients. However, no information about the validity and the feasibility of this new method in such a population has been provided yet. The present study aims to provide a preliminary validation of SWA by comparing it to indirect calorimetry (IC). Another focus will also be the evaluation of its feasibility over time when considering a group of difficult- to-manage patients such as those affected by acute myelogenous leukemia (AML) undergoing cytoreductive induction course.

MATERIALS AND METHODS

Ten patients (6 men, 4 women; age, mean +- SD: 56.6 +- 13.3 years, range: 27-66 years), with a new diagnosis of AML were enrolled in the study. The study was appropriately reviewed and approved by the Niguarda-Ca Granda Hospital’s ethical committee, and written informed consent was obtained before all the measurements were performed. Subjects were studied by anthropometry (weight, height, upper midarm circumference, triceps skinfold thickness), biochemistry (albumin, prealbumin, and transferrin), and bioimpedance analysis (Human-IM Scan Plus; DS Medigroup, Milan, Italy) according to standard procedures. Body composition data were derived according to the specific formula for healthy subjects provided by the manufacturer. BMI and mid-arm muscle area were calculated and 3 months’ previous weight loss was obtained from usual-weight recall.

REE was assessed by IC (Sensor Medics Vmax-29N; Anheim, CA) on admission and at discharge. The measurements were executed between 7:30 and 9:30 AM and after 12-hour fast, with the patients resting supine in bed. A rigid, transparent, ventilated canopy was placed over the head. Oxygen consumption and carbon dioxide production, standardized for temperature, barometric pressure, and humidity, were measured continuously and averaged at 1-minute intervals for 30 minutes after a 15-minute period of equilibration. Before each measurement, the system was calibrated with a standard gas mixture. Automatic calculation of REE was provided by the metabolic cart software according to Weir formula. REE, as well as TDEE, was also assessed by SWA. This instrument collects and memorizes a variety of physiologic data through multiple specific sensors: (1) movements by a 2-axis microelectronic mechanical sensor, (2) heat flux by a thermocouple array, (3) skin and near-body ambient temperature by temperature sensor, and (4) galvanic skin response (an indicator of evaporative heat loss) by 2 hypoallergenic stainless steel electrodes. Through the use of manufacturer software, physiologic body signals are then integrated to accelerometry data and used, in combination with free-living activity recognition patterns, to calculate energy expenditure (EE) according to specific algorithms. Daily physical activity is also monitored.17 As recommended by the manufacturer, the device was placed over the triceps muscle halfway between the acromion and olecranon processes of the dominant upper arm 15 minutes before IC was started. The device was then removed after 24 hours. Data were analyzed using the specific software developed by the manufacturer (Innerview Research Software 5.1; BodyMedia, Inc) after entering necessary demographic characteristics (gender, age, height, weight, smoking habit). To compare IC and SWA measurements, the exact time of IC was recorded and the appropriate 30-minute time window of SWA registration considered. However, TDEE were extrapolated from all the periods of registration. For further comparison, REE estimates were derived using the equation of Harris and Benedict (REE-HB)16 and predicted TDEEs were calculated as 120% (correction factor = 1.2) of the REE measured by IC (TDEE-IC).9

Baseline evaluations (day O) were performed before chemotherapy was started. IC was performed only on admission and at hospital discharge. Time-course anthropometric, biochemical, body composition, and SWA measurements were collected throughout hospital stay (4 times all during the hospital stay, at days O, 7, 14, and at discharge).

Patients underwent standard induction chemotherapy, and erythrocyte/platelet transfusions were given as necessary.

All statistical analyses were performed by MedCalc for Windows, version 7.301 (MedCalc Software, Mariakerke, Belgium). Wilcoxon nonparametric test was used to compare paired data sets within each bout of metabolic measurements and between 2 time-course groups of values. Kruskal-Wallis nonparametric test was performed to assess differences between more than 2 groups. The Bland-Altman plot analysis was conducted to assess the difference/agreement between (1) overall (day O + discharge) SWA and IC measurements of REE and (2) SWA evaluations of TDEE and TDEE-IC estimates. The correlation between parameters was assessed by simple correlation model. Statistical significance was set to a p value < .05.

Finally, the acceptability of this new device was assessed through the use of a simple question: “How did you find it?” Patients were asked to select acceptable with no discomfort, acceptable with discomfort, undecided, and unacceptable, and when discomfort was reported the subject was asked to clarify.

RESULTS

Mean length of stay was 27.1 +- 6.2 days. Baseline and time- course anthropometrie, biochemical, and body composition characteristics of the study group are shown in Table I. Statistically significant changes, between day O and end, were observed only for body weight (p < .05), weight loss (p < .05), percentage of weight loss (p < .05), arm circumference (p < .01), and triceps skinfold thickness (p < .05). A significant difference was also detected in arm circumference values between day O and both day 7 and day 14 (p < .05). With regard to biochemical parameters, albumin concentration on day 14 showed a difference with the concentration on day O (p < .05). No significant changes were observed in body composition variables, with the exception of total body water (TBW) at end when compared with day O (p < .05). None of the patients included had or developed edema during the stay, and only a slight increase in TBW was recorded on day 7. Data concerning the metabolic evaluations performed during recovery are presented in Table II. Kruskal-Wallis analysis revealed no difference in comparing data sets on days O, 7, 14, and end. Within the single course of assessments, a significant difference was detected between predicted REE-HB and both REE-SWA and REE-IC on day O (p < .01 andp < .05, respectively) and at discharge (p < .05). However, on day O REE-IC and REE-SWA were similar (p > .05). End evaluations revealed a significant difference between REE-HB and both REE-SWA and REE-IC and between REE-IC and REE-SWA (p < .05), which was probably due to a low standard deviation in mean Delta values (76 +- 90 kcal/d; see Table II). TABLE I

Anthropometric, biochemical, and body composition parameters during recovery

Similar levels of significance (p < .05) were recorded when assessing differences in REE-SWA and in REE-IC between day O and end.

When considering data on TDEEs, no significant differences were detected between TDEE-IC and TDEESWA on day O and at discharge (p > .05). On day O TDEE-SWA was not statistically different from both REE-IC and REE-SWA; however, at discharge the difference became significant (p < .01 and p < .05, respectively). In addition, we observed a progressive reduction in measured TDEE-SWA (significant at discharge when compared with day O; p < .05). However, TDEE-SWA expressed as kilocalories per kilogram of body weight appeared as a U-shaped curve, with a significant reduction on day 7 (p < .01). Similarly, a significant reduction on day 14 (p < .005; compared with day 0) and subsequent rise at discharge (p < .05; compared with day 14) were recorded in physical activity (number of steps).

TABLE II

Metabolic measurements during recovery

Bland-Altman plot showed that SWA was able to assess REE with sufficient accuracy (95% confidence interval: -162; +283 kcal/d) when compared with IC (Figure 1). Although no bias was seen between TDEE-SWA and TDEE-IC, we detected a wide 95% confidence interval (- 672; +321 kcal/d; Figure 2).

Simple correlation model revealed a significant association between overall SWA and IC measurements of REE (r = 0.84; p < .0001) and between overall TDEE-SWA and TDEE-IC estimates (r = 0.68; p = .001).

FIGURE 1. Bland-Altman bias plots to describe the difference for average values between overall (admission + discharge) resting energy expenditures (REE) measured by SenseWear Armband (SWA) and indirect calorimetry (IC). The middle solid horizontal line represents the mean difference between the methods, and the 2 dotted lines represent the 95% limits of agreement.

Finally, with regard to SWA acceptability, all the patients reported it as acceptable, 9 with no discomfort, and 1 with minor discomfort (slight pruritus in the application site, but without skin reaction).

DISCUSSION

The present pilot study shows the potential reliability and validity of the SWA in the assessment of EEs in cancer patients.

This multiple sensor array was initially introduced to obtain estimates of energy consumption and to quantify the duration of physical activities in a fitness setting. Since then, few studies have been designed to validate it in comparison with the clinical reference standard (IC),12-14 and only 1 has evaluated the concordance with the gold standard method (doubly labeled water).15 However, its use has only scantly been suggested in presence of diseases: (1) for the evaluation of REE in cardiac rehabilitation patients18; (2) for quantifying daily life physical activity in diabetic and COPD patients.17,19 No application of this device has been proposed in cancer yet.

FIGURE 2. Bland-Altman bias plots to describe the difference for average values between overall total daily energy expenditures (TDEE) measured by SenseWear Armband (SWA) and estimated according to REE-IC corrected for “correction factor” (TDEE-IC = REE-IC x 1.2). The middle solid horizontal line represents the mean difference between the methods, and the 2 dotted lines represent the 95% limits of agreement.

The first purpose of this pilot analysis was to investigate the validity of SWA in the assessment of REE.

Measuring EE in every cancer patient during baseline nutrition assessment is controversial, and frequently the choice is left to the physician on an individual case basis. In general, energy requirements might be considered similar to those predicted. However, an increase or decrease of REE in relation to predictions has been reported but the direction of this error (from – 10% to +10%) cannot be established.1,2,8,10,16 The type of tumor, its site, and the presence of a systemic disease (metastatic cancer) have been considered among the factors involved in this process in terms of cytokine-induced metabolic alterations and systemic inflammatory response.1,2,8 Moreover, different therapeutic approaches may be provided, and the overlapping detrimental effects of surgery, chemotherapy, and irradiation must be considered.1,6,7,11

Overall, REE-SWA measurements presented good agreement with those performed by IC, although a slight difference was detected at discharge. However, predictions by HB equations showed a general trend to underestimate REE, which could be considered a limitation to their use in clinical practice, particularly when facing both the frequent increase in energy requirements and the difficulties in reaching proteincalorie targets. This was suggested by the slight decrease, from day O to end, in measured REE (IC and SWA). This reduction is probably due to the effect of chemotherapy on the disease,6,10,11 although in part it may be reasonably ascribed to weight loss.5 Thus, measuring EEs would be of profit to the planning of adequate nutrition support. However, in individual cases, REE cannot be evaluated by gold standard technique. To this purpose, portable IC devices have been studied and partly validated.20 It is in this sense that the feasibility of SWA may be emphasized, particularly when considering AML patients. AML subjects are candidates for intensive cytoreductive regimens, inducing immunodepression and frequently mucositis, with increased risk of infections and decreased oral intake.4’7 In some cases, bone marrow transplantation represents the therapy end point whose success depends also on the preexistent nutrition status.4 Thus, total-body irradiation is used in combination with chemotherapy leading to a significant effect on metabolism, nutrition status, and energy requirements.7

These considerations at once underscore the need for personalized nutrition support, starting with safe, accurate, and acceptable patient assessment of daily energy requirements.

In this regard, we discuss the feasibility of SWA in time-course evaluations. SWA was well accepted by the patients and appeared applicable in a timely manner. Unfortunately, although alterations of REE have been extensively studied, literature about TDEE in cancer patients is less abundant. Moreover, in the present study the lack of an independent reference standard for TDEE is an obvious limitation. In fact, although in SWA sufficient concordance has been demonstrated in healthy subjects by the doubly labeled water method,15 the same findings cannot be generalized to cancer patients. TDEE-SWA baseline measurements were not significantly higher than both IC and SWA measured REE (about +7% and +4%, respectively), suggesting the former as the major factor responsible for the TDEE and leaving a small contribution of physical activity- and diet-induced thermogenesis to TDEE. Total energy expenditure involves REE (approximately 70%), voluntary energy expenditure (25%), and energy expenditure in digestion (5%).21 In the presence of neoplastic disorders, voluntary energy expenditure may be decreased due to apathy, fatigue, and depression, with food intake often being inadequate.2,21 In addition, serial assessments described a mild and progressive reduction hi TDEE, significant only at discharge. With regard to our experience, the decrease in TDEE may be considered the sum of 3 main factors: (1) the effect of chemotherapy, (2) the reduction in body weight (and so in REE), and (3) the effect of physical activity on energy consumption. TDEE expressed as kcal per kg of body weight appeared as a U-shaped curve. On day 7, we observed a slight increase in body weight, related to hyperhydration after chemotherapy (increase in TBW; Table II). However, during the second week (7-10 days after chemotherapy), the patient usually develops mucositis, with possible decrease of both energy intake due to hyporexia and eventual malabsorption. Finally, it is reasonable that near discharge the patient starts feeling better and increases both oral intake and physical activity. A delayed chemotherapy-related stress has also been suggested in the increase of EE.5 This seems to be confirmed by the significant difference detected at discharge between TDEE-SWA and both REE-IC and REE-SWA. Though there is a lack of gold standard methods suitable for clinical setting, it is difficult to assert the accuracy of SWA in the assessment of TDEE. However, the discrepancy found with IC estimates lets us hypothesize that SWA, having some measurement of physical activity, may be more accurate than correcting REE-IC by an arbitrary factor.

The data presented are only preliminary, and firm conclusions are far from being made. However, particularly considering the selected population investigated, new insights on the nutrition approach of the cancer patient have been revealed. This technique also overcomes the limitations of other objective energy expenditure assessment tools. More must be done to generalize the present findings, according to the etherogeneity of cancer types, settings, or even of the stages of the same disease. Portability and the acceptability to the patient, low cost, and the basic skill required by the personnel are the points of strength of SWA in clinical practice. Moreover, suggestions for future analyses and comparisons have been provided, and nutrition treatment studies might be hypothesized to confirm the reliability of the measurements. ACKNOWLEDGMENTS

The work was supported by Bodymedia International and Sensor Medics (Milan, Italy), only in terms of instruments provided. We thank Gian Piero Babbi and Roberta Perissin for their assistance.

REFERENCES

1. Barrera R. Nutritional support in cancer patients. JPEN J Parenter Enteral Nutr. 2002;26(5 Suppl):S63-S71.

2. Inui A. Cancer anorexia-cachexia syndrome: current issues in research and management. CA Cancer J Clin. 2002;52: 72-91.

3. Chute CG, Greenberg ER, Baron J, Korson R, Baker J, Yates J. Presenting conditions of 1539 population-based lung cancer patients by cell type and stage in New Hampshire and Vermont. Cancer. 1985;56:2107-2111.

4. Eriksson KM, Cederholm T, Palmblad JE. Nutrition and acute leukemia in adults: relation between nutritional status and infectious complications during remission induction. Cancer. 1998;82:1071-1077.

5. Garcia-Peris P, Lozano MA, Velasco C, et al. Prospective study of resting energy expenditure changes in head and neck cancer patients treated with chemoradiotherapy measured by indirect calorimetry. Nutrition. 2005;21:1107-1112.

6. Lerebours E, Tilly H, Rimbert A, Delarue J, Piguet H, Colin R. Change in energy and protein status during chemotherapy in patients with acute leukemia. Cancer. 1988;61: 2412-2417.

7. Muscaritoli M, Grieco G, Capria S, Iori AP, Rossi Fanelli F. Nutritional and metabolic support in patients undergoing bone marrow transplantation. Am J Clin Nutr. 2002;75:183-190.

8. Arends J, Bodoky G, Bozzetti F, et al. ESPEN Guidelines on enterai nutrition: non-surgical oncology. Clin Nutr. 2006;25:245- 259.

09. Barak N, Wall-Alonso E, Sitrin MD. Evaluation of stress factors and body weight adjustments currently used to estimate energy expenditure in hospitalized patients. JPEN J Parenter Enteral Nutr. 2002;26:231-238.

10. Delarue J, Lerebours E, Tilly H, et al. Effect of chemotherapy on resting energy expenditure in patients with non- Hodgkin’s lymphoma: results of a sequential study. Cancer. 1990:65:2455-2459.

11. Reeves MM, Battistutta D, Capra S, Bauer J, Davies PS. Resting energy expenditure in patients with solid tumors undergoing anticancer therapy. Nutrition. 2006;22:609-615.

12. Jakicic JM, Marcus M, Gallagher KI, et al. Evaluation of the SenseWear Pro armband to assess energy expenditure during exercise. Med Sci Sports Exerc. 2004;36:897-904.

13. Fruin ML, Rankin JW. Validity of a multi-sensor armband in estimating rest and exercise energy expenditure. Med Sci Sports Exerc. 2004;36:1063-1069.

14. Malavolti M, Pietrobelli A, Dugoni M, et al. A new device for measuring resting energy expenditure (REE) in healthy subjects. Nutr Metab Cardiovasc Dis. 2007;17:338-343.

15. St-Onge M, Mignault D, Allison DB, Rabasa-Lhoret R. Evaluation of a portable device to measure daily energy expenditure in free-living adults. Am J Clin Nutr. 2007;85:742-749.

16. Frankenfield DC, Muth ER, Rowe WA. The Harris-Benedict studies of human basal metabolism: history and limitation. J Am Diet Assoc. 1998:98:439-445.

17. Mignault D, St-Onge M, Karelis AD, Allison DB, Rabasa-Lhoret R. Evaluation of the portable Health Wear armband: a device to measure total daily energy expenditure in free-living type 2 diabetic individuals. Diabetes Care. 2005;28:225-227.

18. Cole PJ, LeMura LM, Klinger TA, Strohecker K, McConnell TR. Measuring energy expenditure in cardiac patients using the Body Media armband versus indirect calorimetry: a validation study. J Sports Med Phys Fitness. 2004;44:262-271.

19. Pitta F, Troosters T, Probst VS, Spruit MA, Decramer M, Gosselink R. Quantifying physical activity in daily life with questionnaires and motion sensors in COPD. Bur Respir J. 2006;27: 1040-1055.

20. Stewart CL, Goody CM, Branson R. Comparison of two systems of measuring energy expenditure. JPEN J Parenter Entend Nutr. 2005;29:212-217.

21. Fouladiun M, Korner U, Bosaeus I, Daneryd P, Hyltander A, Lundholm KG. Body composition and time course changes in regional distribution of fat and lean tissue in unselected cancer patients on palliative care: correlations with food intake, metabolism, exercise capacity, and hormones. Cancer. 2005;103:2189-2198.

Emanuele Cereda, MD*; Mauro Turrini, MD[dagger]; Denis Ciapanna, MD[dagger]; Laura Marbello, MD[dagger]; Angelo Pietrobelli, MD[double dagger]; and Ettore Corradi, MD[section]

From the * International Center for the Assessment of Nutritional Status (ICANS), University of Milan, Italy; [dagger] Division of Haematology, Department of

Oncology, Niguarda Ca’ Grande Hospital, Milano, Italy; [double dagger] Paediatric Unit, Verona University Medical School, Verona, Italy; and the [section] Division of

Clinical Nutrition, Department of Medical Area, Niguarda Ca’ Granda Hospital, Milano, Italy

Received for publication January 4, 2007.

Accepted for publication May 11, 2007.

Correspondence: Emanuele Cereda, International Center for the Assessment of Nutritional Status (ICANS, Director Professor Giulio Testolin), University of Milan, via Botticelli 21, 20133 Milan, Italy. Electronic mail may be sent to emanuele.cereda@virgilio.it.

Copyright American Society for Parenteral and Enteral Nutrition Nov/ Dec 2007

(c) 2007 JPEN, Journal of Parenteral and Enteral Nutrition. Provided by ProQuest Information and Learning. All rights Reserved.