- Email: [email protected]
Validation of the Actiheart for estimating physical activity related energy expenditure in pregnancy
e-SPEN Journal 7 (2012) e5ee10
Contents lists available at SciVerse ScienceDirect
e-SPEN Journal journal homepage: http://www.elsevier.com/locate/clnu
Original article
Validation of the Actiheart for estimating physical activity related energy expenditure in pregnancy K. Melzer a, M. Lazzeri b, S. Armand c, M. Boulvain d, Y. Schutz e, B. Kayser b, * a
Swiss Federal Institute of Sports, Magglingen, Switzerland Institute of Movement Sciences and Sports Medicine, Faculty of Medicine, University of Geneva, Switzerland c Willy Taillard Laboratory of Kinesiology, Geneva University Hospitals and Faculty of Medicine, Geneva, Switzerland d Department of Obstetrics and Gynaecology, University Hospitals of Geneva, Switzerland e Department of Physiology, University of Lausanne, Switzerland b
a r t i c l e i n f o
s u m m a r y
Article history: Received 24 November 2011 Accepted 15 December 2011
Background & aims: The objective of this study was to assess the validity of the individually calibrated Actiheart (a combined heart rate and movement sensor device), in pregnant women against indirect calorimetry in a laboratory setting. Methods: Ten healthy pregnant women (aged 32.9 3.2 yrs, pre-pregnancy BMI ¼ 21.0 2.4 kg/m2, 36.9 2.4 weeks of gestation) walked at 3, 4, 5, and 6 km/h on a treadmill, cycled at 25 W and 50 W on an ergometer and stepped on and off a 15 cm high step. During each routine, AEE was measured simultaneously with the Actiheart (AEEa) and indirect calorimetry (AEEcalo). AEE measurements were compared with paired Student’s t-test, and their agreement with Bland and Altman plots. Results: The mean AEEcalo was not significantly different from AEEa for any activity except for cycling at 50 W (45 J/kg/min, p ¼ 0.01). Cumulated AEEa and AEEcalo, combining all activities, were not different (p ¼ 0.9). All data points (100%) fell within 2SD for all activities except for walking at 6 km/h (89% of data points). All data points fell within 2SD for the sum of all speeds of walking (3, 4, 5, and 6 km/h). Conclusions: The Actiheart can be used as a valid method for AEE estimation in pregnant women. Ó 2011 European Society for Clinical Nutrition and Metabolism. Published by Elsevier Ltd. All rights reserved.
Keywords: Exercise Heart rate Movement Acceleration Pregnant women
1. Introduction Regular moderately intense physical activity during pregnancy results in marked benefits for mother and foetus.1e5 A decrease or complete cessation of habitual physical activity negatively influences the maternal cardiovascular system,6,7 glucose tolerance,1 body weight8,9 and the capacity to sustain the efforts required during labour and delivery.7,10,11 In order to further clarify the association of physical activity levels with particular health outcomes, precise measurement of physical activity related energy expenditure is necessary. However, activity related energy expenditure is difficult to measure precisely, especially in a free-living environment. The methods currently used for activity energy expenditure estimations have their limitations. Self-report methods, like questionnaires or
* Corresponding author. Institute of Movement Sciences and Sports Medicine, Faculté de médecine, Université de Genève, 10 rue du Conseil Général, 1205 Genève, Switzerland. Tel.: þ41 22 3790028; fax: þ41 22 3790035. E-mail address: [email protected] (B. Kayser).
interviews, are limited in their reliability and validity due to misreporting or miscoding of activities, inaccurate estimation of activity intensity or duration, and differences in body mass.12 The doubly labelled water method13 gives an average daily metabolic rate over several days and does not provide specific information on physical activity patterns, such as duration and intensity of activity performed. Indirect calorimetry, a method of reference by which the energy expenditure is estimated from measurements of oxygen consumption and carbon dioxide production, is not well suited for monitoring for prolonged periods outside the laboratory setting (bulky device, use of a mask or a mouthpiece).14 Accelerometry and heart rate recording have their own limitations when used alone. While accelerometry is unable to account for increases in activity when stepping, cycling, changing grade during walking or load-bearing activities, heart rate measurements are affected by other factors, such as training state, mental stress, dehydration or high ambient temperature.15 A combination of heart rate and movement sensor device was reported to give precise estimates of activity energy expenditure and physical activity patterns among men and non-pregnant women.16
2212-8263/$36.00 Ó 2011 European Society for Clinical Nutrition and Metabolism. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.eclnm.2011.12.008
e6
K. Melzer et al. / e-SPEN Journal 7 (2012) e5ee10
This device is commercially available (Actiheart, CamNtech Ltd., Cambridge, UK) and validated against indirect calorimetry during a wide range of activities (from low, through moderate and high activities) in both laboratory17 and field settings.18 The reliability and validity of the Actiheart have been confirmed elsewhere.15,19e21 Although the Actiheart gives accurate estimations of activity energy expenditure and activity patterns in non-pregnant subjects, there are at present no studies that validated the accuracy of the instrument in pregnant women. The Actiheart has been used to measure energy expenditure during pregnancy,11,22,23 but pregnancy induced physiological changes (alterations in heart rate, weight, oxygen uptake) might influence the estimates. Further research is thus needed to explore the Actiheart measurements against other reference techniques in the pregnant state. The objective of this study was to assess the validity of the Actiheart in pregnant women against indirect calorimetry in a laboratory setting.
(22e23 C) for 10e20 min. After that period, a mask was positioned _ 2) and carbon over their mouth and nose, and oxygen uptake (VO _ dioxide production (VCO 2) were measured for 20 min on a breathby-breath basis, averaged over 5 s epochs, with a metabolic cart (MedGraphics, CardiO2, USA). The first 5 min of data were eliminated as acclimatization artefact. From the remaining 15 min a segment of 5 consecutive minutes of measures showing <10% _ _ 2 and VCO coefficient of variation in VO 2 was considered as steady _ _ state. VO2 and VCO2 were then used to calculate RMR using the abbreviated Weir equation and expressed in J/kg/min.24 Calibration of the metabolic cart was carried out with a 3 L syringe (flow transducer) and gases of known concentration (gas analysers) before each measurement. After the RMR measurements, women took a light breakfast (consisting of a croissant and juice) and rested for around 30 min before the calibration and activity energy expenditure measurements started.
2. Materials and methods 2.1. Subjects During routine prenatal consultation at the maternity unit of the University Hospitals of Geneva, Switzerland, participation in the study was proposed to healthy women in their third trimester of uncomplicated singleton pregnancy. Criteria for exclusion were heart disease or treatment that may alter cardiovascular conditioning or any other health problems. Seventeen women were recruited for the study. Informed, written consent was obtained from each participant. All participating subjects were in good health and took no medications. The institutional ethics committee approved the study. 2.2. Procedures We measured the agreement between Actiheart and indirect calorimetry measurements during walking, cycling and stepping. The women were asked to come to the laboratory early in the morning, after an overnight fast, by car or public transport in order to prevent any vigorous efforts. Body weight was measured to the nearest 0.1 kg on a calibrated beam scale (Seca, Hamburg, Germany) and body height to the nearest 0.5 cm with a height rod (Seca) with the subjects in light clothing and without shoes. Gestational age was assessed based on the last menstrual period, or on a first trimester ultrasound measurement if a difference of 1 week between the 2 estimates was detected. After cleaning and light abrasion preparation of the skin, ECG electrodes (Red Dot 2560, 3M) were applied the left side of the chest, just below the apex of the sternum.20 The Actiheart, a lightweight (10 g) combined heart rate (HR) and movement (uniaxial accelerometer oriented to measure acceleration (ACC) along the body’s longitudinal axis) sensor was clipped on the electrodes, and worn constantly on the chest day and night for 2 consecutive days. The laboratory measurements were performed on day 1. The Actiheart was then left attached overnight for sleeping heart rate (SHR) recordings. The SHR was defined as the highest value among the 60 lowest HR recordings during a 24-h period. Mean SHR was determined as the average value of the 2 daily SHR recordings. 2.3. The laboratory measurements 2.3.1. Resting metabolic rate Women arrived at the hospital in the morning, after an overweight fast, avoiding any strenuous physical effort. Before the resting metabolic rate (RMR) measurements, the subjects relaxed on a bed in a thermoneutral environment
2.3.2. Calibration of the Actiheart As pregnancy changes the HR and activity energy expenditure (AEE) relationship (Lotgering et al., 1991), the device was calibrated each time for each individual using a standardized step test, an inbuilt function of the Actiheart software, as described previously.21 The women stepped up and down a 15 cm high step, progressively increasing step frequency from 15 to 32.5 body lifts per minute (rate of change: 2.5 body lifts per min2). They were advised to stop the test if they felt uncomfortable. The mass-specific lift work rate, that is, mechanical power of the step test, was calculated as 9.81 m/s2 step height (m) lift frequency (number of body weight lifts per min) and expressed in J/min per kg. Linear regression was used to model the relationship between the power and HR during stepping. This produced individual calibration parameters, net HR above sleep (HRaS ¼ HRSHR) slope and intercept (denoted by bstep and astep, respectively) of the regression line. The step test derived calibration parameters were then introduced into an HReenergy expenditure (EEHR) equation19:
EEHR ¼ 2:9HRaS þ 2:9 HRaS bstep þ 1:3astep 75: The individually adjusted EEHR relationship was then introduced in the branched equation model for AEE estimation provided by the Actiheart software. The branched combined model uses activity (acceleration) and heart rate (HR) for AEE estimation. The model was presented in detail elsewhere.22 The Actiheart was configured to record ACC and ECG derived heart rate (32 and 128 Hz sampling, respectively), averaged in 30-s epochs. 2.3.3. Activity energy expenditure The women were asked to perform the following activities: walking at 3, 4, 5 and 6 km/h on a motorized treadmill set in a horizontal position (HP Cosmos Mercury Med 4.0, Germany), cycling at 25 and 50 W (at 60 rpm) on a mechanically braked ergocycle (Monark Ergomedic 829E, Vardberg, Sweden) and stepping at 15 up to 32.5 body lifts per minute for a total of up to 8 min on a 15 cm high step (Reebok, UK). The calibration protocol for walking and cycling consisted of 5-min intervals at each exercise intensity; data obtained after reaching steady state (discarding the first 2 min of each exercise level) was used for analysis. The calibration protocol for stepping exercise consisted of the same procedure as used for the individual calibration of the Actiheart. There were 5e10 min breaks between the activities. Gas exchange _ 2 data was was measured for the duration of each activity. The VO used to calculate total energy expenditure, from which the RMR was subtracted in order to obtain activity energy expenditure (AEE). The AEE was expressed in J/kg/min by assuming an energetic value
K. Melzer et al. / e-SPEN Journal 7 (2012) e5ee10
of 1 L of O2 ¼ w20.35 kJ25 assuming that the energy was derived equally from fat and carbohydrate. 2.4. Statistical analysis
Table 2 Mean activity energy expenditure data of 10 pregnant women in their third trimester of pregnancy (36.9 2.4 weeks) measured using indirect calorimetry (AEEcalo) and Actiheart (AEEa) (means SD). Activity
Data were tested for normality using the KolmogoroveSmirnov test and for homogeneity of variance with Levene’s test. Comparison of activity energy expenditure measurements was assessed with paired Student’s t-test analysis. The statistical significance level was set at p < 0.05. Bland and Altman plots were used to assess the agreement between two measurement types during different activities. The limits of agreement were defined as the mean difference 2SD. Statistical analyses were performed with SPSS version 17.0 for PC (SPSS, Inc. Chicago, USA). The sample size was calculated to be sufficient to show a difference in continuous measurements of 1 SD between methods, with a power of 90 and 95% confidence intervals. 3. Results A total of 17 women were recruited for the study. After enrolment 3 women dropped out at the beginning of the exercise testing being afraid that it would harm the foetus and/or the mother in this late stage of pregnancy. Data analyses for 4 women had to be excluded due to technical problems in data recording. Final analyses were performed on data obtained in 10 pregnant women. There are two missing data points: one woman had difficulties in finalizing the walking test at 6 km/h, and in another the Actiheart became detached during cycling. The characteristics of the women are presented in Table 1. The laboratory results of the calorimetry AEE (AEEcalo) and Actiheart AEE (AEEa) measurements were normally distributed according to the KolmogoroveSmirnov test (p > 0.05). The means, standard deviations and significant differences (p < 0.05) between AEEcalo and AEEa are presented in Table 2. The mean AEEcalo measurements for any activity were not significantly different from AEEa for that activity, except for AEEa during cycling at 50 W, where AEEa significantly underestimated energy expenditure in comparison to the AEEcalo (p ¼ 0.01). When the cycling results of two loads (25 W and 50 W) were analysed together, AEEa was still significantly different compared to AEEcalo (p ¼ 0.027). The mean AEEaAEEcalo difference of all activities taken together (walking at 3, 4, 5, and 6 km/h, cycling at 25 W and 50 W and stepping), that can represent a surrogate of a variety of activities encountered in daily life, did not yield a significant difference (p ¼ 0.9). Mean difference between AEEcalo and AEEa, limits of agreement and standard error of the limits of agreement are presented in Table 3, according to the Bland and Altman method.26,27 While treadmill walking at 3 km/h showed the lowest mean AEEaAEEcalo difference (2 J/kg/min), cycling at 50 W showed the highest mean AEEaAEEcalo difference (46 J/kg/min). The individual differences between AEEa and AEEcalo against their mean values are plotted for each activity according to the Table 1 Characteristics of the subjects. Characteristics
Women (n ¼ 10) mean SD (range)
Age (years) Weight (kg) Height (cm) Pre-pregnancy BMI (kg/m2) Gestation (weeks) Ethnicity Caucasian Asian
32.9 3.2 (26e39) 70.7 8.9 (57e83) 167.0 5.9 (155e173) 21.0 2.4 (18.7e25.5) 36.9 2.4 (31e39) 9 1
e7
AEEcalo (J/kg/min)
Treadmill 3 km/h 4 km/h 5 km/h 6 km/hb Cycling 25 Wb 50 Wb Step test a b
119 142 189 252
AEEa (J/kg/min)
21 26 36 37
121 151 194 253
157 41 219 58 194 12
14 14 17 34
137 20 174 24a 200 17
Significant difference between AEEcalo and AEEa (p < 0.05). n ¼ 9.
Bland and Altman method in Fig. 1. The difference between AEEa and AEEcalo is plotted against the means of the two measurements. It is recommended that at least 95% of the data points should lie within 2SD of the mean difference.26e28 In Fig. 1(aeg), 10 out of 10 points (100%) fell within 2SD for any of the activities (walking at 3, 4, 5 km/h, cycling at 25 W and 50 W, and stepping) except for walking at 6 km/h, where 8 out of 9 (89%) data points fell within 2SD, while 1 out of 9 (11%) fell above 2SD. When the difference between AEEa and AEEcalo is plotted against the means of the two measurements for the sum of all speeds of walking (3, 4, 5, and 6 km/h), all data points fell within 2SD. 4. Discussion This study assessed the validity of the Actiheart, a combined ACC and HR monitor, to estimate physical activity related energy expenditure against indirect calorimetry in 10 pregnant women during walking, cycling and stepping. Our results show that the Actiheart estimates AEE of walking, cycling and stepping during pregnancy with good reliability. The Actiheart can thus be used, after individual calibration of the device, for estimation of physical activity related energy expenditure in pregnant women as well. The only disagreement between the AEEa the AEEcalo measurements was shown during walking at 6 km/h where 89% of data points fell within 2SD, instead of 95%. Conversely, the AEEa and AEEcalo showed good agreement, with 100% of data points falling within 2SD, when the mean value for 4 velocities (3, 4, 5, and 6 km/h) was taken into account. As for the cycling at 50 W, the
Table 3 Limits of agreement between activity energy expenditure data of 10 pregnant women in their third trimester of pregnancy (37 2 weeks) measured using indirect calorimetry (AEEcalo) and Actiheart (AEEa). Activity
Treadmill 3 km/h 4 km/h 5 km/h 6 km/ha Cycling 25 Wa 50 Wa Step test Mean a
n ¼ 9.
Mean difference AEEa-AEEcalo (J/kg/min)
Limits of agreement (mean difference 2SD) (J/kg/min)
Standard error for the limits of agreement (95%CI for the bias) (J/kg/min) Lower limits
Upper limits
42 26 52 86
7 to 46 2 to 44 6 to 61 16 to 89
2 9 5 1
48 30 58 91
20 45
85 to 45 127 to 36
87 to 8 123 to 31
32 to 42 60 to 32
6
121 to 133
108 to 29
17 to 120
to to to to
52 48 68 93
to to to to
11 16 16 18
e8
K. Melzer et al. / e-SPEN Journal 7 (2012) e5ee10
a
b
c
d
e
f
g
Fig. 1. BlandeAltman plots describing agreement between the AEEa and AEEcalo during walking at 3 km/h (a), 4 km/h (b), 5 km/h (c) and 6 km/h (d), cycling at 25 W (e) and 50 W (f), and stepping (g).
AEEa and AEEcalo showed good agreement, although mean AEEa was significantly underestimated (45 J/kg/min, p ¼ 0.01) compared to the mean AEEcalo (Table 3). This finding could be expected since stationary cycling on an ergometer represents an activity with little effect on acceleration. Calibration procedures for methods estimating physical activity related energy expenditure should reflect the activities most
commonly engaged in by the studied population.15 In this study, the agreement between the Actiheart and calorimetry measurements was tested during the activities (walking and stepping) considered to be the most commonly performed in late pregnancy. In addition, we also looked at light intensity cycling, even though not a common means of mobility for a majority of women in their last trimester of pregnancy. Nevertheless, the AEEa and AEEcalo
K. Melzer et al. / e-SPEN Journal 7 (2012) e5ee10
provide similar values for the mean of all activities (walking at 3, 4, 5, and 6 km/h, cycling at 25 W and 50 W and stepping) (p ¼ 0.9). As an internal validation of the calibration procedure we used the calibrated Actiheart to estimate energy expenditure during stepping exercise done in the same fashion as during the calibration and found good agreement (Fig. 1g). The good agreement of the measurements can be explained by the fact that we calibrated the Actiheart for each individual (individual calibration step tests were _ 2 performed). This means that the individually developed HR-VO regression lines had largely taken into account physiological changes which the women experience during pregnancy period. The physiological changes induced by pregnancy are primarily developed to meet increased metabolic demands of mother _ 2) and foetus. Changes in submaximal oxygen uptake (VO during pregnancy depend on the type of exercise performed. During maternal rest or submaximal weight bearing exercise _ 2 (e.g. walking, stepping, treadmill exercise), absolute maternal VO (L/min) is significantly increased compared to the non-pregnant state.29,30 The magnitude of change is approximately proportional to maternal weight gain. At the same speed or grade of _ 2 expressed in ml/kg/min are walking or running, the values for VO thus similar or only slightly higher during pregnancy compared to the non-pregnant state.29,31e33 When pregnant women perform submaximal weight-supported exercise on land (e.g. level cycling), the findings are contradictory. _ 2,29,32,34 Some studies reported significantly increased absolute VO 30,32,35e41 while many others reported unchanged or only slightly _ 2 compared to non-pregnant state. The latter increased absolute VO findings may be explained by the fact that the metabolic demand of cycle exercise is largely independent of the maternal body mass, _ 2 alteration. resulting in no absolute VO Few studies that directly measured changes in maternal _ 2max) showed no difference in the VO _ 2max (L/ _ 2 (VO maximal VO min) between pregnant and non-pregnant subjects in cycling, swimming or weight bearing exercise. Efficiency of work during _ 2 and work exercise, i.e. the slope of the relationship between VO rate, appears to be unchanged during pregnancy in non-weight bearing exercise.30,36,42 During weight bearing exercise, the work efficiency was shown to be improved in both, exercising women and those who exercised but stopped exercising during pregnancy.43,44 When adjusted for weight gain, the increased efficiency is maintained throughout the pregnancy with the improvement being greater in exercising women.43 The Actiheart was intentionally programmed to measure energy expenditure using 30-s epochs. This registration mode allows measurements of 5 days (heart rate, acceleration and minimal and maximal inter-beat-interval measurements) to 20 days (heart rate and acceleration measurements). As physical activity engagement can vary over the days, shorter periods of measurement might not be sufficient to provide reliable data on overall physical activity. A 15-sec epoch registration would give even more precise measurements, but the recording time would be too short for precise measurement of AEE in large population groups (two and half days for heart rate, acceleration and minimal and maximal inter-beatinterval recording). A limitation of the use of the Actiheart in pregnant women is that the placement, using ECG electrodes on the participant’s chest just below the sternum, is difficult due to the enlarged breasts and womb in this late stage of pregnancy. Although it was shown that the HR recordings are less prone to error if the electrodes are positioned on that level (in comparison to the upper chest placement),19,45 recording of the HR might be easier for this study group if a belt could be used instead of the electrodes. In conclusion, the results of the study show that the Actiheart can be used as a valid method for physical activity estimation in
e9
pregnant women. The data serve as valuable information for improving the assessment of activity energy expenditure in pregnancy, which can clarify its association with particular health outcomes. Author contribution All authors participated in the study design. KM and ML performed the recruitment of the subjects, the data collection, the statistical analysis, and drafted the manuscript. SA, YS, MB, BK contributed to data interpretation and drafting of the manuscript. All authors approved the final version. None of the authors had any conflict of interest in connection to the study. Conflict of interest There are no conflicts of interest. Acknowledgements The study was financed by a competitive grant attributed to Drs Boulvain and Kayser by the CRC (Centre de Recherche Clinique), and further supported by the Faculty of Medicine of the University of Geneva, and the University Hospitals of Geneva, Switzerland. We thank all the pregnant women who agreed to participate in this study. References 1. Artal R. Exercise: the alternative therapeutic intervention for gestational diabetes. Clin Obstet Gynecol 2003;46:479e87. 2. Clapp III JF, Lopez B, Harcar-Sevcik R. Neonatal behavioral profile of the offspring of women who continued to exercise regularly throughout pregnancy. Am J Obstet Gynecol 1999;180:91e4. 3. Clapp III JF. Exercise during pregnancy. A clinical update. Clin Sports Med 2000;19:273e86. 4. Poudevigne MS, O’Connor PJ. Physical activity and mood during pregnancy. Med Sci Sports Exerc 2005;37:1374e80. 5. Melzer K, Schutz Y, Boulvain M, Kayser B. Physical activity and pregnancy: cardiovascular adaptations, recommendations and pregnancy outcomes. Sports Med 2010;40:493e507. 6. Pivarnik JM, Ayres NA, Mauer MB, Cotton DB, Kirshon B, Dildy GA. Effects of maternal aerobic fitness on cardiorespiratory responses to exercise. Med Sci Sports Exerc 1993;25:993e8. 7. Söhnchen N, Melzer K, Tejada BM, Jastrow-Meyer N, Othenin-Girard V, Irion O, et al. Maternal heart rate changes during labour. Eur J Obstet Gynecol Reprod Biol 2011 Oct;158(2):173e8. 8. Special communications. Roundtable consensus statement. Impact of physical activity during pregnancy and postpartum on chronic disease risk. Med Sci Sports Exerc 2006;38:989e1006. 9. Melzer K, Schutz Y. Pre-pregnancy and pregnancy predictors of obesity. Int J Obes (Lond) 2010;34(Suppl. 2):S44e52. 10. Clapp III JF. The course of labor after endurance exercise during pregnancy. Am J Obstet Gynecol 1990;163:1799e805. 11. Söhnchen N, Melzer K, Tejada BM, Jastrow-Meyer N, Othenin-Girard V, Irion O, et al. Effects of recommended levels of physical activity on pregnancy outcomes. Am J Obstet Gynecol 2010;202:266. 12. Shephard RJ. Limits to the measurement of habitual physical activity by questionnaires. Br J Sports Med 2003;37:197e206. 13. Speakman JR. The history and theory of the doubly labeled water technique. Am J Clin Nutr 1998;68:932Se8S. 14. Andre D, Wolf DL. Recent advances in free-living physical activity monitoring: a review. J Diabetes Sci Technol 2007;1:760e7. 15. Brage S, Brage N, Franks PW, Ekelund U, Wong MY, Andersen LB, et al. Branched equation modeling of simultaneous accelerometry and heart rate monitoring improves estimate of directly measured physical activity energy expenditure. J Appl Physiol 2004;96:343e51. 16. Rennie K, Rowsell T, Jebb SA, Holburn D, Wareham NJ. A combined heart rate and movement sensor: proof of concept and preliminary testing study. Eur J Clin Nutr 2000;54:409e14. 17. Thompson D, Batterham AM, Bock S, Robson C, Stokes K. Assessment of low-tomoderate intensity physical activity thermogenesis in young adults using synchronized heart rate and accelerometry with branched-equation modeling. J Nutr 2006;136:1037e42. 18. Crouter SE, Churilla JR, Bassett Jr DR. Accuracy of the Actiheart for the assessment of energy expenditure in adults. Eur J Clin Nutr 2008;62:704e11.
e10
K. Melzer et al. / e-SPEN Journal 7 (2012) e5ee10
19. Brage S, Brage N, Franks PW, Ekelund U, Wareham NJ. Reliability and validity of the combined heart rate and movement sensor Actiheart. Eur J Clin Nutr 2005;59:561e70. 20. Brage S, Brage N, Ekelund U, Luan J, Franks PW, Froberg K, et al. Effect of combined movement and heart rate monitor placement on physical activity estimates during treadmill locomotion and free-living. Eur J Appl Physiol 2006;96:517e24. 21. Brage S, Ekelund U, Brage N, Hennings MA, Froberg K, Franks PW, et al. Hierarchy of individual calibration levels for heart rate and accelerometry to measure physical activity. J Appl Physiol 2007;103:682e92. 22. Melzer K, Schutz Y, Boulvain M, Kayser B. Pregnancy-related changes in activity energy expenditure and resting metabolic rate in Switzerland. Eur J Clin Nutr 2009;63:1185e91. 23. Melzer K, Schutz Y, Soehnchen N, Othenin Girard V, Martinez de Tejada B, Pichard C, et al. Prepregnancy body mass index and resting metabolic rate during pregnancy. Ann Nutr Metab 2010;57:221e7. 24. Weir JB. New methods for calculating metabolic rate with special reference to protein metabolism. Nutrition 1949;1990(6):213e21. 25. Consolazio C, Johnson R, Pecora L. Physiological measurements of metabolic functions in man. New York: McGraw Hill; 1963. 1963. 26. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307e10. 27. Bland JM, Altman DG. Applying the right statistics: analyses of measurement studies. Ultrasound Obstet Gynecol 2003;22:85e93. 28. Bland JM, Altman DG. Comparing methods of measurement: why plotting difference against standard method is misleading. Lancet 1995;346:1085e7. 29. Carpenter MW, Sady SP, Sady MA, Haydon B, Coustan DR, Thompson PD. Effect of maternal weight gain during pregnancy on exercise performance. J Appl Physiol 1990;68:1173e6. 30. Knuttgen HG, Emerson Jr K. Physiological response to pregnancy at rest and during exercise. J Appl Physiol 1974;36:549e53. 31. Artal R. Exercise and pregnancy. Clin Sports Med 1992;11:363e77. 32. Pivarnik JM, Lee W, Miller JF. Physiological and perceptual responses to cycle and treadmill exercise during pregnancy. Med Sci Sports Exerc 1991;23:470e5.
33. Artal R, Wiswell R, Romem Y, Dorey F. Pulmonary responses to exercise in pregnancy. Am J Obstet Gynecol 1986;154:378e83. 34. Pernoll ML, Metcalfe J, Schlenker TL, Welch JE, Matsumoto JA. Oxygen consumption at rest and during exercise in pregnancy. Respir Physiol 1975;25:285e93. 35. Bonen A, Campagna P, Gilchrist L, Young DC, Beresford P. Substrate and endocrine responses during exercise at selected stages of pregnancy. J Appl Physiol 1992;73:134e42. 36. Heenan AP, Wolfe LA, Davies GA. Maximal exercise testing in late gestation: maternal responses. Obstet Gynecol 2001;97:127e34. 37. O’Toole ML. Physiologic aspects of exercise in pregnancy. Clin Obstet Gynecol 2003;46:379e89. 38. Sady SP, Carpenter MW, Sady MA, Haydon B, Hoegsberg B, Cullinane EM. Prediction of VO2max during cycle exercise in pregnant women. J Appl Physiol 1988;65:657e61. 39. Sady SP, Carpenter MW, Thompson PD, Sady MA, Haydon B, Coustan DR. Cardiovascular response to cycle exercise during and after pregnancy. J Appl Physiol 1989;66:336e41. 40. Spinnewijn WE, Wallenburg HC, Struijk PC, Lotgering FK. Peak ventilatory responses during cycling and swimming in pregnant and nonpregnant women. J Appl Physiol 1996;81:738e42. 41. Wolfe LA, Mottola MF. Aerobic exercise in pregnancy: an update. Can J Appl Physiol 1993;18:119e47. 42. Spaaij CJ, van Raaij JM, de Groot LC, van der Heijden LJ, Boekholt HA, Hautvast JG. No changes during pregnancy in the net cost of cycling exercise. Eur J Clin Nutr 1994;48:513e21. 43. Clapp III JF. Oxygen consumption during treadmill exercise before, during, and after pregnancy. Am J Obstet Gynecol 1989;161:1458e64. 44. Lotgering FK, Struijk PC, van Doorn MB, Wallenburg HC. Errors in predicting maximal oxygen consumption in pregnant women. J Appl Physiol 1992;72:562e7. 45. Rautaharju PM, Park L, Rautaharju FS, Crow R. A standardized procedure for locating and documenting ECG chest electrode positions: consideration of the effect of breast tissue on ECG amplitudes in women. J Electrocardiol 1998;31:17e29.
Contents lists available at SciVerse ScienceDirect
e-SPEN Journal journal homepage: http://www.elsevier.com/locate/clnu
Original article
Validation of the Actiheart for estimating physical activity related energy expenditure in pregnancy K. Melzer a, M. Lazzeri b, S. Armand c, M. Boulvain d, Y. Schutz e, B. Kayser b, * a
Swiss Federal Institute of Sports, Magglingen, Switzerland Institute of Movement Sciences and Sports Medicine, Faculty of Medicine, University of Geneva, Switzerland c Willy Taillard Laboratory of Kinesiology, Geneva University Hospitals and Faculty of Medicine, Geneva, Switzerland d Department of Obstetrics and Gynaecology, University Hospitals of Geneva, Switzerland e Department of Physiology, University of Lausanne, Switzerland b
a r t i c l e i n f o
s u m m a r y
Article history: Received 24 November 2011 Accepted 15 December 2011
Background & aims: The objective of this study was to assess the validity of the individually calibrated Actiheart (a combined heart rate and movement sensor device), in pregnant women against indirect calorimetry in a laboratory setting. Methods: Ten healthy pregnant women (aged 32.9 3.2 yrs, pre-pregnancy BMI ¼ 21.0 2.4 kg/m2, 36.9 2.4 weeks of gestation) walked at 3, 4, 5, and 6 km/h on a treadmill, cycled at 25 W and 50 W on an ergometer and stepped on and off a 15 cm high step. During each routine, AEE was measured simultaneously with the Actiheart (AEEa) and indirect calorimetry (AEEcalo). AEE measurements were compared with paired Student’s t-test, and their agreement with Bland and Altman plots. Results: The mean AEEcalo was not significantly different from AEEa for any activity except for cycling at 50 W (45 J/kg/min, p ¼ 0.01). Cumulated AEEa and AEEcalo, combining all activities, were not different (p ¼ 0.9). All data points (100%) fell within 2SD for all activities except for walking at 6 km/h (89% of data points). All data points fell within 2SD for the sum of all speeds of walking (3, 4, 5, and 6 km/h). Conclusions: The Actiheart can be used as a valid method for AEE estimation in pregnant women. Ó 2011 European Society for Clinical Nutrition and Metabolism. Published by Elsevier Ltd. All rights reserved.
Keywords: Exercise Heart rate Movement Acceleration Pregnant women
1. Introduction Regular moderately intense physical activity during pregnancy results in marked benefits for mother and foetus.1e5 A decrease or complete cessation of habitual physical activity negatively influences the maternal cardiovascular system,6,7 glucose tolerance,1 body weight8,9 and the capacity to sustain the efforts required during labour and delivery.7,10,11 In order to further clarify the association of physical activity levels with particular health outcomes, precise measurement of physical activity related energy expenditure is necessary. However, activity related energy expenditure is difficult to measure precisely, especially in a free-living environment. The methods currently used for activity energy expenditure estimations have their limitations. Self-report methods, like questionnaires or
* Corresponding author. Institute of Movement Sciences and Sports Medicine, Faculté de médecine, Université de Genève, 10 rue du Conseil Général, 1205 Genève, Switzerland. Tel.: þ41 22 3790028; fax: þ41 22 3790035. E-mail address: [email protected] (B. Kayser).
interviews, are limited in their reliability and validity due to misreporting or miscoding of activities, inaccurate estimation of activity intensity or duration, and differences in body mass.12 The doubly labelled water method13 gives an average daily metabolic rate over several days and does not provide specific information on physical activity patterns, such as duration and intensity of activity performed. Indirect calorimetry, a method of reference by which the energy expenditure is estimated from measurements of oxygen consumption and carbon dioxide production, is not well suited for monitoring for prolonged periods outside the laboratory setting (bulky device, use of a mask or a mouthpiece).14 Accelerometry and heart rate recording have their own limitations when used alone. While accelerometry is unable to account for increases in activity when stepping, cycling, changing grade during walking or load-bearing activities, heart rate measurements are affected by other factors, such as training state, mental stress, dehydration or high ambient temperature.15 A combination of heart rate and movement sensor device was reported to give precise estimates of activity energy expenditure and physical activity patterns among men and non-pregnant women.16
2212-8263/$36.00 Ó 2011 European Society for Clinical Nutrition and Metabolism. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.eclnm.2011.12.008
e6
K. Melzer et al. / e-SPEN Journal 7 (2012) e5ee10
This device is commercially available (Actiheart, CamNtech Ltd., Cambridge, UK) and validated against indirect calorimetry during a wide range of activities (from low, through moderate and high activities) in both laboratory17 and field settings.18 The reliability and validity of the Actiheart have been confirmed elsewhere.15,19e21 Although the Actiheart gives accurate estimations of activity energy expenditure and activity patterns in non-pregnant subjects, there are at present no studies that validated the accuracy of the instrument in pregnant women. The Actiheart has been used to measure energy expenditure during pregnancy,11,22,23 but pregnancy induced physiological changes (alterations in heart rate, weight, oxygen uptake) might influence the estimates. Further research is thus needed to explore the Actiheart measurements against other reference techniques in the pregnant state. The objective of this study was to assess the validity of the Actiheart in pregnant women against indirect calorimetry in a laboratory setting.
(22e23 C) for 10e20 min. After that period, a mask was positioned _ 2) and carbon over their mouth and nose, and oxygen uptake (VO _ dioxide production (VCO 2) were measured for 20 min on a breathby-breath basis, averaged over 5 s epochs, with a metabolic cart (MedGraphics, CardiO2, USA). The first 5 min of data were eliminated as acclimatization artefact. From the remaining 15 min a segment of 5 consecutive minutes of measures showing <10% _ _ 2 and VCO coefficient of variation in VO 2 was considered as steady _ _ state. VO2 and VCO2 were then used to calculate RMR using the abbreviated Weir equation and expressed in J/kg/min.24 Calibration of the metabolic cart was carried out with a 3 L syringe (flow transducer) and gases of known concentration (gas analysers) before each measurement. After the RMR measurements, women took a light breakfast (consisting of a croissant and juice) and rested for around 30 min before the calibration and activity energy expenditure measurements started.
2. Materials and methods 2.1. Subjects During routine prenatal consultation at the maternity unit of the University Hospitals of Geneva, Switzerland, participation in the study was proposed to healthy women in their third trimester of uncomplicated singleton pregnancy. Criteria for exclusion were heart disease or treatment that may alter cardiovascular conditioning or any other health problems. Seventeen women were recruited for the study. Informed, written consent was obtained from each participant. All participating subjects were in good health and took no medications. The institutional ethics committee approved the study. 2.2. Procedures We measured the agreement between Actiheart and indirect calorimetry measurements during walking, cycling and stepping. The women were asked to come to the laboratory early in the morning, after an overnight fast, by car or public transport in order to prevent any vigorous efforts. Body weight was measured to the nearest 0.1 kg on a calibrated beam scale (Seca, Hamburg, Germany) and body height to the nearest 0.5 cm with a height rod (Seca) with the subjects in light clothing and without shoes. Gestational age was assessed based on the last menstrual period, or on a first trimester ultrasound measurement if a difference of 1 week between the 2 estimates was detected. After cleaning and light abrasion preparation of the skin, ECG electrodes (Red Dot 2560, 3M) were applied the left side of the chest, just below the apex of the sternum.20 The Actiheart, a lightweight (10 g) combined heart rate (HR) and movement (uniaxial accelerometer oriented to measure acceleration (ACC) along the body’s longitudinal axis) sensor was clipped on the electrodes, and worn constantly on the chest day and night for 2 consecutive days. The laboratory measurements were performed on day 1. The Actiheart was then left attached overnight for sleeping heart rate (SHR) recordings. The SHR was defined as the highest value among the 60 lowest HR recordings during a 24-h period. Mean SHR was determined as the average value of the 2 daily SHR recordings. 2.3. The laboratory measurements 2.3.1. Resting metabolic rate Women arrived at the hospital in the morning, after an overweight fast, avoiding any strenuous physical effort. Before the resting metabolic rate (RMR) measurements, the subjects relaxed on a bed in a thermoneutral environment
2.3.2. Calibration of the Actiheart As pregnancy changes the HR and activity energy expenditure (AEE) relationship (Lotgering et al., 1991), the device was calibrated each time for each individual using a standardized step test, an inbuilt function of the Actiheart software, as described previously.21 The women stepped up and down a 15 cm high step, progressively increasing step frequency from 15 to 32.5 body lifts per minute (rate of change: 2.5 body lifts per min2). They were advised to stop the test if they felt uncomfortable. The mass-specific lift work rate, that is, mechanical power of the step test, was calculated as 9.81 m/s2 step height (m) lift frequency (number of body weight lifts per min) and expressed in J/min per kg. Linear regression was used to model the relationship between the power and HR during stepping. This produced individual calibration parameters, net HR above sleep (HRaS ¼ HRSHR) slope and intercept (denoted by bstep and astep, respectively) of the regression line. The step test derived calibration parameters were then introduced into an HReenergy expenditure (EEHR) equation19:
EEHR ¼ 2:9HRaS þ 2:9 HRaS bstep þ 1:3astep 75: The individually adjusted EEHR relationship was then introduced in the branched equation model for AEE estimation provided by the Actiheart software. The branched combined model uses activity (acceleration) and heart rate (HR) for AEE estimation. The model was presented in detail elsewhere.22 The Actiheart was configured to record ACC and ECG derived heart rate (32 and 128 Hz sampling, respectively), averaged in 30-s epochs. 2.3.3. Activity energy expenditure The women were asked to perform the following activities: walking at 3, 4, 5 and 6 km/h on a motorized treadmill set in a horizontal position (HP Cosmos Mercury Med 4.0, Germany), cycling at 25 and 50 W (at 60 rpm) on a mechanically braked ergocycle (Monark Ergomedic 829E, Vardberg, Sweden) and stepping at 15 up to 32.5 body lifts per minute for a total of up to 8 min on a 15 cm high step (Reebok, UK). The calibration protocol for walking and cycling consisted of 5-min intervals at each exercise intensity; data obtained after reaching steady state (discarding the first 2 min of each exercise level) was used for analysis. The calibration protocol for stepping exercise consisted of the same procedure as used for the individual calibration of the Actiheart. There were 5e10 min breaks between the activities. Gas exchange _ 2 data was was measured for the duration of each activity. The VO used to calculate total energy expenditure, from which the RMR was subtracted in order to obtain activity energy expenditure (AEE). The AEE was expressed in J/kg/min by assuming an energetic value
K. Melzer et al. / e-SPEN Journal 7 (2012) e5ee10
of 1 L of O2 ¼ w20.35 kJ25 assuming that the energy was derived equally from fat and carbohydrate. 2.4. Statistical analysis
Table 2 Mean activity energy expenditure data of 10 pregnant women in their third trimester of pregnancy (36.9 2.4 weeks) measured using indirect calorimetry (AEEcalo) and Actiheart (AEEa) (means SD). Activity
Data were tested for normality using the KolmogoroveSmirnov test and for homogeneity of variance with Levene’s test. Comparison of activity energy expenditure measurements was assessed with paired Student’s t-test analysis. The statistical significance level was set at p < 0.05. Bland and Altman plots were used to assess the agreement between two measurement types during different activities. The limits of agreement were defined as the mean difference 2SD. Statistical analyses were performed with SPSS version 17.0 for PC (SPSS, Inc. Chicago, USA). The sample size was calculated to be sufficient to show a difference in continuous measurements of 1 SD between methods, with a power of 90 and 95% confidence intervals. 3. Results A total of 17 women were recruited for the study. After enrolment 3 women dropped out at the beginning of the exercise testing being afraid that it would harm the foetus and/or the mother in this late stage of pregnancy. Data analyses for 4 women had to be excluded due to technical problems in data recording. Final analyses were performed on data obtained in 10 pregnant women. There are two missing data points: one woman had difficulties in finalizing the walking test at 6 km/h, and in another the Actiheart became detached during cycling. The characteristics of the women are presented in Table 1. The laboratory results of the calorimetry AEE (AEEcalo) and Actiheart AEE (AEEa) measurements were normally distributed according to the KolmogoroveSmirnov test (p > 0.05). The means, standard deviations and significant differences (p < 0.05) between AEEcalo and AEEa are presented in Table 2. The mean AEEcalo measurements for any activity were not significantly different from AEEa for that activity, except for AEEa during cycling at 50 W, where AEEa significantly underestimated energy expenditure in comparison to the AEEcalo (p ¼ 0.01). When the cycling results of two loads (25 W and 50 W) were analysed together, AEEa was still significantly different compared to AEEcalo (p ¼ 0.027). The mean AEEaAEEcalo difference of all activities taken together (walking at 3, 4, 5, and 6 km/h, cycling at 25 W and 50 W and stepping), that can represent a surrogate of a variety of activities encountered in daily life, did not yield a significant difference (p ¼ 0.9). Mean difference between AEEcalo and AEEa, limits of agreement and standard error of the limits of agreement are presented in Table 3, according to the Bland and Altman method.26,27 While treadmill walking at 3 km/h showed the lowest mean AEEaAEEcalo difference (2 J/kg/min), cycling at 50 W showed the highest mean AEEaAEEcalo difference (46 J/kg/min). The individual differences between AEEa and AEEcalo against their mean values are plotted for each activity according to the Table 1 Characteristics of the subjects. Characteristics
Women (n ¼ 10) mean SD (range)
Age (years) Weight (kg) Height (cm) Pre-pregnancy BMI (kg/m2) Gestation (weeks) Ethnicity Caucasian Asian
32.9 3.2 (26e39) 70.7 8.9 (57e83) 167.0 5.9 (155e173) 21.0 2.4 (18.7e25.5) 36.9 2.4 (31e39) 9 1
e7
AEEcalo (J/kg/min)
Treadmill 3 km/h 4 km/h 5 km/h 6 km/hb Cycling 25 Wb 50 Wb Step test a b
119 142 189 252
AEEa (J/kg/min)
21 26 36 37
121 151 194 253
157 41 219 58 194 12
14 14 17 34
137 20 174 24a 200 17
Significant difference between AEEcalo and AEEa (p < 0.05). n ¼ 9.
Bland and Altman method in Fig. 1. The difference between AEEa and AEEcalo is plotted against the means of the two measurements. It is recommended that at least 95% of the data points should lie within 2SD of the mean difference.26e28 In Fig. 1(aeg), 10 out of 10 points (100%) fell within 2SD for any of the activities (walking at 3, 4, 5 km/h, cycling at 25 W and 50 W, and stepping) except for walking at 6 km/h, where 8 out of 9 (89%) data points fell within 2SD, while 1 out of 9 (11%) fell above 2SD. When the difference between AEEa and AEEcalo is plotted against the means of the two measurements for the sum of all speeds of walking (3, 4, 5, and 6 km/h), all data points fell within 2SD. 4. Discussion This study assessed the validity of the Actiheart, a combined ACC and HR monitor, to estimate physical activity related energy expenditure against indirect calorimetry in 10 pregnant women during walking, cycling and stepping. Our results show that the Actiheart estimates AEE of walking, cycling and stepping during pregnancy with good reliability. The Actiheart can thus be used, after individual calibration of the device, for estimation of physical activity related energy expenditure in pregnant women as well. The only disagreement between the AEEa the AEEcalo measurements was shown during walking at 6 km/h where 89% of data points fell within 2SD, instead of 95%. Conversely, the AEEa and AEEcalo showed good agreement, with 100% of data points falling within 2SD, when the mean value for 4 velocities (3, 4, 5, and 6 km/h) was taken into account. As for the cycling at 50 W, the
Table 3 Limits of agreement between activity energy expenditure data of 10 pregnant women in their third trimester of pregnancy (37 2 weeks) measured using indirect calorimetry (AEEcalo) and Actiheart (AEEa). Activity
Treadmill 3 km/h 4 km/h 5 km/h 6 km/ha Cycling 25 Wa 50 Wa Step test Mean a
n ¼ 9.
Mean difference AEEa-AEEcalo (J/kg/min)
Limits of agreement (mean difference 2SD) (J/kg/min)
Standard error for the limits of agreement (95%CI for the bias) (J/kg/min) Lower limits
Upper limits
42 26 52 86
7 to 46 2 to 44 6 to 61 16 to 89
2 9 5 1
48 30 58 91
20 45
85 to 45 127 to 36
87 to 8 123 to 31
32 to 42 60 to 32
6
121 to 133
108 to 29
17 to 120
to to to to
52 48 68 93
to to to to
11 16 16 18
e8
K. Melzer et al. / e-SPEN Journal 7 (2012) e5ee10
a
b
c
d
e
f
g
Fig. 1. BlandeAltman plots describing agreement between the AEEa and AEEcalo during walking at 3 km/h (a), 4 km/h (b), 5 km/h (c) and 6 km/h (d), cycling at 25 W (e) and 50 W (f), and stepping (g).
AEEa and AEEcalo showed good agreement, although mean AEEa was significantly underestimated (45 J/kg/min, p ¼ 0.01) compared to the mean AEEcalo (Table 3). This finding could be expected since stationary cycling on an ergometer represents an activity with little effect on acceleration. Calibration procedures for methods estimating physical activity related energy expenditure should reflect the activities most
commonly engaged in by the studied population.15 In this study, the agreement between the Actiheart and calorimetry measurements was tested during the activities (walking and stepping) considered to be the most commonly performed in late pregnancy. In addition, we also looked at light intensity cycling, even though not a common means of mobility for a majority of women in their last trimester of pregnancy. Nevertheless, the AEEa and AEEcalo
K. Melzer et al. / e-SPEN Journal 7 (2012) e5ee10
provide similar values for the mean of all activities (walking at 3, 4, 5, and 6 km/h, cycling at 25 W and 50 W and stepping) (p ¼ 0.9). As an internal validation of the calibration procedure we used the calibrated Actiheart to estimate energy expenditure during stepping exercise done in the same fashion as during the calibration and found good agreement (Fig. 1g). The good agreement of the measurements can be explained by the fact that we calibrated the Actiheart for each individual (individual calibration step tests were _ 2 performed). This means that the individually developed HR-VO regression lines had largely taken into account physiological changes which the women experience during pregnancy period. The physiological changes induced by pregnancy are primarily developed to meet increased metabolic demands of mother _ 2) and foetus. Changes in submaximal oxygen uptake (VO during pregnancy depend on the type of exercise performed. During maternal rest or submaximal weight bearing exercise _ 2 (e.g. walking, stepping, treadmill exercise), absolute maternal VO (L/min) is significantly increased compared to the non-pregnant state.29,30 The magnitude of change is approximately proportional to maternal weight gain. At the same speed or grade of _ 2 expressed in ml/kg/min are walking or running, the values for VO thus similar or only slightly higher during pregnancy compared to the non-pregnant state.29,31e33 When pregnant women perform submaximal weight-supported exercise on land (e.g. level cycling), the findings are contradictory. _ 2,29,32,34 Some studies reported significantly increased absolute VO 30,32,35e41 while many others reported unchanged or only slightly _ 2 compared to non-pregnant state. The latter increased absolute VO findings may be explained by the fact that the metabolic demand of cycle exercise is largely independent of the maternal body mass, _ 2 alteration. resulting in no absolute VO Few studies that directly measured changes in maternal _ 2max) showed no difference in the VO _ 2max (L/ _ 2 (VO maximal VO min) between pregnant and non-pregnant subjects in cycling, swimming or weight bearing exercise. Efficiency of work during _ 2 and work exercise, i.e. the slope of the relationship between VO rate, appears to be unchanged during pregnancy in non-weight bearing exercise.30,36,42 During weight bearing exercise, the work efficiency was shown to be improved in both, exercising women and those who exercised but stopped exercising during pregnancy.43,44 When adjusted for weight gain, the increased efficiency is maintained throughout the pregnancy with the improvement being greater in exercising women.43 The Actiheart was intentionally programmed to measure energy expenditure using 30-s epochs. This registration mode allows measurements of 5 days (heart rate, acceleration and minimal and maximal inter-beat-interval measurements) to 20 days (heart rate and acceleration measurements). As physical activity engagement can vary over the days, shorter periods of measurement might not be sufficient to provide reliable data on overall physical activity. A 15-sec epoch registration would give even more precise measurements, but the recording time would be too short for precise measurement of AEE in large population groups (two and half days for heart rate, acceleration and minimal and maximal inter-beatinterval recording). A limitation of the use of the Actiheart in pregnant women is that the placement, using ECG electrodes on the participant’s chest just below the sternum, is difficult due to the enlarged breasts and womb in this late stage of pregnancy. Although it was shown that the HR recordings are less prone to error if the electrodes are positioned on that level (in comparison to the upper chest placement),19,45 recording of the HR might be easier for this study group if a belt could be used instead of the electrodes. In conclusion, the results of the study show that the Actiheart can be used as a valid method for physical activity estimation in
e9
pregnant women. The data serve as valuable information for improving the assessment of activity energy expenditure in pregnancy, which can clarify its association with particular health outcomes. Author contribution All authors participated in the study design. KM and ML performed the recruitment of the subjects, the data collection, the statistical analysis, and drafted the manuscript. SA, YS, MB, BK contributed to data interpretation and drafting of the manuscript. All authors approved the final version. None of the authors had any conflict of interest in connection to the study. Conflict of interest There are no conflicts of interest. Acknowledgements The study was financed by a competitive grant attributed to Drs Boulvain and Kayser by the CRC (Centre de Recherche Clinique), and further supported by the Faculty of Medicine of the University of Geneva, and the University Hospitals of Geneva, Switzerland. We thank all the pregnant women who agreed to participate in this study. References 1. Artal R. Exercise: the alternative therapeutic intervention for gestational diabetes. Clin Obstet Gynecol 2003;46:479e87. 2. Clapp III JF, Lopez B, Harcar-Sevcik R. Neonatal behavioral profile of the offspring of women who continued to exercise regularly throughout pregnancy. Am J Obstet Gynecol 1999;180:91e4. 3. Clapp III JF. Exercise during pregnancy. A clinical update. Clin Sports Med 2000;19:273e86. 4. Poudevigne MS, O’Connor PJ. Physical activity and mood during pregnancy. Med Sci Sports Exerc 2005;37:1374e80. 5. Melzer K, Schutz Y, Boulvain M, Kayser B. Physical activity and pregnancy: cardiovascular adaptations, recommendations and pregnancy outcomes. Sports Med 2010;40:493e507. 6. Pivarnik JM, Ayres NA, Mauer MB, Cotton DB, Kirshon B, Dildy GA. Effects of maternal aerobic fitness on cardiorespiratory responses to exercise. Med Sci Sports Exerc 1993;25:993e8. 7. Söhnchen N, Melzer K, Tejada BM, Jastrow-Meyer N, Othenin-Girard V, Irion O, et al. Maternal heart rate changes during labour. Eur J Obstet Gynecol Reprod Biol 2011 Oct;158(2):173e8. 8. Special communications. Roundtable consensus statement. Impact of physical activity during pregnancy and postpartum on chronic disease risk. Med Sci Sports Exerc 2006;38:989e1006. 9. Melzer K, Schutz Y. Pre-pregnancy and pregnancy predictors of obesity. Int J Obes (Lond) 2010;34(Suppl. 2):S44e52. 10. Clapp III JF. The course of labor after endurance exercise during pregnancy. Am J Obstet Gynecol 1990;163:1799e805. 11. Söhnchen N, Melzer K, Tejada BM, Jastrow-Meyer N, Othenin-Girard V, Irion O, et al. Effects of recommended levels of physical activity on pregnancy outcomes. Am J Obstet Gynecol 2010;202:266. 12. Shephard RJ. Limits to the measurement of habitual physical activity by questionnaires. Br J Sports Med 2003;37:197e206. 13. Speakman JR. The history and theory of the doubly labeled water technique. Am J Clin Nutr 1998;68:932Se8S. 14. Andre D, Wolf DL. Recent advances in free-living physical activity monitoring: a review. J Diabetes Sci Technol 2007;1:760e7. 15. Brage S, Brage N, Franks PW, Ekelund U, Wong MY, Andersen LB, et al. Branched equation modeling of simultaneous accelerometry and heart rate monitoring improves estimate of directly measured physical activity energy expenditure. J Appl Physiol 2004;96:343e51. 16. Rennie K, Rowsell T, Jebb SA, Holburn D, Wareham NJ. A combined heart rate and movement sensor: proof of concept and preliminary testing study. Eur J Clin Nutr 2000;54:409e14. 17. Thompson D, Batterham AM, Bock S, Robson C, Stokes K. Assessment of low-tomoderate intensity physical activity thermogenesis in young adults using synchronized heart rate and accelerometry with branched-equation modeling. J Nutr 2006;136:1037e42. 18. Crouter SE, Churilla JR, Bassett Jr DR. Accuracy of the Actiheart for the assessment of energy expenditure in adults. Eur J Clin Nutr 2008;62:704e11.
e10
K. Melzer et al. / e-SPEN Journal 7 (2012) e5ee10
19. Brage S, Brage N, Franks PW, Ekelund U, Wareham NJ. Reliability and validity of the combined heart rate and movement sensor Actiheart. Eur J Clin Nutr 2005;59:561e70. 20. Brage S, Brage N, Ekelund U, Luan J, Franks PW, Froberg K, et al. Effect of combined movement and heart rate monitor placement on physical activity estimates during treadmill locomotion and free-living. Eur J Appl Physiol 2006;96:517e24. 21. Brage S, Ekelund U, Brage N, Hennings MA, Froberg K, Franks PW, et al. Hierarchy of individual calibration levels for heart rate and accelerometry to measure physical activity. J Appl Physiol 2007;103:682e92. 22. Melzer K, Schutz Y, Boulvain M, Kayser B. Pregnancy-related changes in activity energy expenditure and resting metabolic rate in Switzerland. Eur J Clin Nutr 2009;63:1185e91. 23. Melzer K, Schutz Y, Soehnchen N, Othenin Girard V, Martinez de Tejada B, Pichard C, et al. Prepregnancy body mass index and resting metabolic rate during pregnancy. Ann Nutr Metab 2010;57:221e7. 24. Weir JB. New methods for calculating metabolic rate with special reference to protein metabolism. Nutrition 1949;1990(6):213e21. 25. Consolazio C, Johnson R, Pecora L. Physiological measurements of metabolic functions in man. New York: McGraw Hill; 1963. 1963. 26. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307e10. 27. Bland JM, Altman DG. Applying the right statistics: analyses of measurement studies. Ultrasound Obstet Gynecol 2003;22:85e93. 28. Bland JM, Altman DG. Comparing methods of measurement: why plotting difference against standard method is misleading. Lancet 1995;346:1085e7. 29. Carpenter MW, Sady SP, Sady MA, Haydon B, Coustan DR, Thompson PD. Effect of maternal weight gain during pregnancy on exercise performance. J Appl Physiol 1990;68:1173e6. 30. Knuttgen HG, Emerson Jr K. Physiological response to pregnancy at rest and during exercise. J Appl Physiol 1974;36:549e53. 31. Artal R. Exercise and pregnancy. Clin Sports Med 1992;11:363e77. 32. Pivarnik JM, Lee W, Miller JF. Physiological and perceptual responses to cycle and treadmill exercise during pregnancy. Med Sci Sports Exerc 1991;23:470e5.
33. Artal R, Wiswell R, Romem Y, Dorey F. Pulmonary responses to exercise in pregnancy. Am J Obstet Gynecol 1986;154:378e83. 34. Pernoll ML, Metcalfe J, Schlenker TL, Welch JE, Matsumoto JA. Oxygen consumption at rest and during exercise in pregnancy. Respir Physiol 1975;25:285e93. 35. Bonen A, Campagna P, Gilchrist L, Young DC, Beresford P. Substrate and endocrine responses during exercise at selected stages of pregnancy. J Appl Physiol 1992;73:134e42. 36. Heenan AP, Wolfe LA, Davies GA. Maximal exercise testing in late gestation: maternal responses. Obstet Gynecol 2001;97:127e34. 37. O’Toole ML. Physiologic aspects of exercise in pregnancy. Clin Obstet Gynecol 2003;46:379e89. 38. Sady SP, Carpenter MW, Sady MA, Haydon B, Hoegsberg B, Cullinane EM. Prediction of VO2max during cycle exercise in pregnant women. J Appl Physiol 1988;65:657e61. 39. Sady SP, Carpenter MW, Thompson PD, Sady MA, Haydon B, Coustan DR. Cardiovascular response to cycle exercise during and after pregnancy. J Appl Physiol 1989;66:336e41. 40. Spinnewijn WE, Wallenburg HC, Struijk PC, Lotgering FK. Peak ventilatory responses during cycling and swimming in pregnant and nonpregnant women. J Appl Physiol 1996;81:738e42. 41. Wolfe LA, Mottola MF. Aerobic exercise in pregnancy: an update. Can J Appl Physiol 1993;18:119e47. 42. Spaaij CJ, van Raaij JM, de Groot LC, van der Heijden LJ, Boekholt HA, Hautvast JG. No changes during pregnancy in the net cost of cycling exercise. Eur J Clin Nutr 1994;48:513e21. 43. Clapp III JF. Oxygen consumption during treadmill exercise before, during, and after pregnancy. Am J Obstet Gynecol 1989;161:1458e64. 44. Lotgering FK, Struijk PC, van Doorn MB, Wallenburg HC. Errors in predicting maximal oxygen consumption in pregnant women. J Appl Physiol 1992;72:562e7. 45. Rautaharju PM, Park L, Rautaharju FS, Crow R. A standardized procedure for locating and documenting ECG chest electrode positions: consideration of the effect of breast tissue on ECG amplitudes in women. J Electrocardiol 1998;31:17e29.