«Marianne Udnæseth Master Thesis in Clinical Nutrition Supervisors: Christine Helle Lene Frost Andersen Department of Nutrition School of Medicine ...»
Dietary assessment of athletes:
validation of a four-day weighed
diet record and a physical activity
Master Thesis in Clinical Nutrition
Lene Frost Andersen
Department of Nutrition
School of Medicine
UNIVERSITY OF OSLO
The present master thesis in clinical nutrition was planned during spring 2008, after an inquiry from the Department of Sports Nutrition at the Norwegian Olympic Sports Centre, Olympiatoppen. The study was conducted at the Department of Nutrition, University of Oslo, from autumn 2008 to spring 2009.
I would like to thank my supervisors, Christine Helle and Lene Frost Andersen, for valuable advice and guidance during the planning and implementation of the study.
To Christine and Olympiatoppen, thank you for giving me the opportunity to get better known with the field of sports nutrition. To Lene, thank you for shearing your great knowledge and experiences in nutritional methodology and research with me.
I would also like to address a special thanks to Anders Aandstad at the Defence Institute at the Norwegian School of Sport Sciences for lending me the physical activity monitors used in this study. Without your contribution, this study would not have been realized.
To all the athletes who participated in the study, you deserve a big thank-you for devoting your time and for showing a great positivism and an admirably go-ahead spirit.
To all my wonderful fellow students, I will miss our daily lunches, discussions, and all the laughter and fun we have sheared the last five years.
And last but not least I would like to give my deepest thanks to my family and my dearest Tom-Erik for your support, encouragement, patience and love and for always having faith in me.
Oslo, May 2009 Marianne Udnæseth Summary Background: The ability to assess the energy intake (EI) and the energy needs of an individual athlete or a group of athletes is of vital importance in the field of sports nutrition. Optimal nutrition may contribute to enhanced performance and recovery from exercise whereas inadequate EI relative to energy expenditure (EE) compromises performance and negates the benefits of training. Therefore, meeting energy needs have become a nutrition priority for athletes. However, the majority of studies looking at athletes’ energy balance have found limited agreement between self-reported EI and EE measured using the doubly labelled water (DLW) technique.
The discrepancies between EI and EE cast doubt on the validity of self-reported dietary data, which is often used as a basis for the dietary assessment of athletes.
Objective: To investigate the validity of a four-day weighed diet record developed for assessing the dietary intake of Norwegian elite athletes by comparing EI with EE measured with the physical activity monitor SenseWear Pro2 Armband (SWA) in a group of male endurance athletes. In addition, EE estimated from a four-day physical activity record was validated against EE measured with SWA.
Design: The participants completed simultaneously a four-day weighed diet record and a four-day physical activity record. During the same four days, they also measured EE with SWA.
Subjects: Thirty-five Norwegian male athletes being members of National teams in summer sports (rowing, kayaking, orienteering, middle- and long-distance running, cycling and race walking) volunteered to participate and completed the study.
However, two participants were excluded due to acute sickness during the monitoring period and defect SWA-measurements. Thus, only 33 athletes were included in the study.
Results: The EI was on average 7.6% (1.3 MJ/d, P = 0.017) lower than the EE measured with SWA (EESWA). The 95% confidence limits of agreement in a BlandAltman plot for EI and EESWA varied from -7.1 to 4.6 MJ/d. The Pearson correlation coefficient between reported EI and EESWA was 0.58 (P 0.001). Nineteen athletes (58%) were classified into the same third for both EI and EESWA whereas two athletes (6%) were grossly misclassified.
In the comparison of EE estimated from the activity record (EErecord) and EESWA, the EErecord was on average 13.5% (2.3 MJ/d, P 0.001) lower than EESWA. The 95% confidence limits of agreement in a Bland-Altman plot for EErecord and EESWA varied from -5.7 to 1.2 MJ/d. The Pearson correlation coefficient between estimated and measured EE was 0.86 (P 0.001). Twenty-two participants (67%) appeared in the same third with both the activity record and SWA. There were no grossly misclassified participants.
Conclusion: The data showed that the four-day weighed diet record and the four-day physical activity record under-estimated the average EI and EE respectively.
Moreover, there was substantial variability in the accuracy of the diet record and the activity record at the individual level. The ability to rank individuals according to self-reported EI and EE were found to be good with both methods.
Norsk sammendrag Bakgrunn: Vurdering av energiinntaket og energibehovet til en individuell idrettsutøver eller en gruppe idrettsutøvere er en sentral problemstilling i fagfeltet idrettsernæring. Et optimalt kosthold kan bidra til å forbedre prestasjon og restitusjon etter trening, mens et utilstrekkelig energiinntak i forhold til energiforbruk vil ha negativ innvirkning på prestasjon og redusere effekten av trening. Å dekke energibehovet har derfor blitt en av prioritetene for idrettsutøvere. Likevel har de fleste studier som har sett på energibalansen til idrettsutøvere, funnet lite samsvar mellom selvrapportert energiinntak og energiforbruk målt med dobbelt merket vann (DLW) metoden. Den observerte differansen mellom energiinntak og energiforbruk sår tvil om validiteten til selvrapporterte kostholdsdata, som ofte blir brukt som grunnlag for den individuelle kostveiledningen av idrettsutøvere.
Formål: Formålet med denne studien var å undersøke validiteten til en fire-dagers veid kostregistrering utviklet for å vurdere kostinntaket til norske toppidrettsutøvere, ved å sammenligne energiinntak med energiforbruk målt med aktivitetsmonitoren SenseWear Pro2 Armband (SWA) i en gruppe mannlige utholdenhetsutøvere. I tillegg ble energiforbruk estimert fra en fire-dagers aktivitetsregistrering validert mot energiforbruk målt med SWA.
Design: Deltakerne gjennomførte en fire-dagers veid kostregistrering og en firedagers aktivitetsregistrering samtidig. I de samme fire dagene målte de også energiforbruket sitt med SWA.
Deltakere: Trettifem norske landslagsutøvere i sommeridretter (roing, kajakk, orientering, mellom- og langdistanse løping, sykling og kappgang) deltok frivillig og fullførte datainnsamlingen. To utøvere måtte imidlertid ekskluderes på grunn av akutt sykdom i løpet av registreringsperioden og feil på SWA-målingen av energiforbruk.
Totalt ble 33 idrettsutøvere inkludert i studien.
Resultater: Det gjennomsnittlige energiinntaket for hele gruppen var 7,6 % (1,3 MJ/d, P = 0,017) lavere enn energiforbruket målt med SWA (EESWA). Øvre og nedre grenseverdi for grad av overensstemmelse mellom de to metodene (gjennomsnittlig differanse ± 2 SD) i et Bland-Altman plott var -7,1 og 4,6 MJ/d.
Pearson korrelasjonskoeffisient mellom rapportert energiinntak og EESWA var 0,58 (P 0,001). Nitten idrettsutøvere (58 %) ble klassifisert i den samme tredjedelen for både energiinntak og EESWA mens to utøvere (6 %) ble grovt misklassifisert.
I sammenligningen mellom energiforbruk estimert fra aktivitetsregistreringen (EErecord) og EESWA var gruppegjennomsnittet for EErecord 13,5 % (2,3 MJ/d, P 0,001) lavere enn EESWA. Et Bland-Altman plott viste at gjennomsnittlig differanse ± 2 SD var -5,7 og 1,2 MJ/d. Pearson korrelasjonskoeffisient mellom de to metodene var 0,86 (P 0,001). Tjueto deltakere (67 %) ble klassifisert i samme tredjedel med både aktivitetsregistreringen og SWA. Ingen deltakere ble grovt misklassifisert.
Konklusjon: Resultatene fra denne studien viste at fire-dagers veid kostregistrering og fire-dagers aktivitetsregistrering underestimerte henholdsvis energiinntak og energiforbruk. Det var stor variasjon i nøyaktigheten til kostregistreringen og aktivitetsregistreringen på individnivå. Begge metodene viste imidlertid god evne til å rangere individer ut ifra selvrapportert energiinntak og energiforbruk.
Table of contents ACKNOWLEDGEMENTS
TABLE OF CONTENTS
LIST OF TABLES
LIST OF FIGURES
LIST OF APPENDIXES
ABBREVIATIONS AND DEFINITIONS
1.1.1 Energy needs of athletes
1.1.2 Dietary assessment of athletes
1.1.3 The search for an ideal method to assess the energy intake of athletes
1.2 VALIDATION OF DIETARY ASSESSMENT METHODS
1.2.1 Definition of validity
1.3 HOW TO VALIDATE A METHOD
1.4 VALIDATION OF ENERGY INTAKE
1.5 METHODS FOR VALIDATION OF ENERGY INTAKE
1.5.1 The doubly labelled water technique
1.5.2 Indirect calorimetry
1.5.3 Heart rate monitors
1.5.5 SenseWear Pro2 Armband
1.5.6 Physical activity questionnaires
1.5.7 Physical activity records
2. OBJECTIVE AND RESEARCH QUESTIONS
2.2 RESEARCH QUESTIONS
3. SUBJECTS AND METHODS
3.1.1 Criteria for participation
3.2.2 Body measurements
3.2.3 The test methods
Four-day weighed diet record
Four-day physical activity record
3.2.4 The reference method
SenseWear Pro2 Armband
3.2.5 Statistical methods
4.2 COMPARISON OF ENERGY INTAKE ESTIMATED FROM THE FOUR-DAY WEIGHED DIET RECORD ANDENERGY EXPENDITURE MEASURED WITH THE SENSEWEAR PRO2 ARMBAND
Accuracy in reporting
Comparison of macronutrient intake
Comparison of intake of selected foods and food groups
5.1.1 Participation rate
5.1.2 The representativity of the sample
5.2 THE REFERENCE METHOD
5.2.1 The validity of SenseWear Pro2 Armband
5.3 THE TEST METHODS
5.3.1 The validity of the four-day weighed diet record
5.3.2 The use of the EI:EE ratio
5.3.3 Possible reasons for under-estimation of energy intake
5.3.4 Characteristics of under-reporters
5.3.5 The validity of the four-day physical activity record
5.3.6 Possible reasons for under-estimation of energy expenditure
List of tables Table 1.1: Why an athlete’s energy intake is of great importance (p. 16) Table 3.1: Approximate energy costs expressed as multiples of basal metabolic rate (BMR) for six activity categories (p. 34) Table 3.2 Example of calculation of total daily energy expenditure for one day based on the four-day activity record (p. 35) Table 4.1: Number of participants from different sports (p. 38) Table 4.2: Physical characteristics of the participants (p. 38) Table 4.3: Energy intake (EI) estimated from the weighed diet record, energy expenditure measured with the SenseWear Pro2 Armband (EESWA) and the relationship between EI and EESWA (p. 39) Table 4.4: Distribution of participants in thirds with regard to energy intake (EI) estimated from the weighed diet record and energy expenditure measured with the SenseWear Pro2 Armband (EESWA) (p. 42) Table 4.5: Physical characteristics, energy intake (EI) estimated from the weighed diet record, energy expenditure measured with the SenseWear Pro2 Armband (EESWA) and the relationship between EI and EESWA among under-reporters (UR), acceptable-reporters (AR) and over-reporters (OR) (p. 43) Table 4.6: Absolute intake of macronutrients and intake of macronutrients as percentage of energy estimated from the weighed diet record among under-reporters (UR), acceptable-reporters (AR) and over-reporters (OR) (p. 44) Table 4.7: Absolute intake of selected food items and food groups estimated from the weighed diet record among under-reporters (UR), acceptable-reporters (AR) and over-reporters (OR) (p. 45) Table 4.8: Energy expenditure measured with the SenseWear Pro2 Armband (EESWA) and estimated from the activity record (EErecord) and the relationship between EE from the two methods (p. 46) Table 4.9: Distribution of participants in thirds with regard to energy expenditure estimated from the activity record (EErecord) and measured by the SenseWear Pro2 armband (EESWA) (p. 47) List of figures Figure 1.1: The SenseWear Pro2 Armband (SWA) (p. 26) Figure 4.1: Comparison of energy intake (EI) estimated from the weighed diet record and energy expenditure measured with the SenseWear Pro2 Armband (EESWA) within and between the different sports (p. 40) Figure 4.2: The difference between energy intake (EI) estimated from the weighed diet record and energy expenditure measured with the SenseWear Pro2 Armband (EESWA) plotted against the mean of the two methods (p. 41) Figure 4.3: Energy intake (EI) estimated from the weighed diet record plotted against energy expenditure measured with the SenseWear Pro2 Armband (EESWA) (p. 41) Figure 4.4: The difference between energy expenditure estimated from the activity record (EErecord) and energy expenditure measured with the SenseWear Pro2 Armband (EESWA) plotted against the mean of the two methods (p. 46) Figure 4.5: Energy expenditure estimated from the activity record (EErecord) plotted against energy expenditure measured with the SenseWear Pro2 Armband (EESWA) (p. 47) List of appendixes Appendix 1: Approvals from the National Committees for Research Ethics in Norway (p. 72) Appendix 2: Approval from Norwegian Social Science Data Services (p. 75) Appendix 3: Written invitation to participate in the study (p. 77) Appendix 4: The record notebook (p. 81) Appendix 5: Written information about the SenseWear Pro2 Armband given to the participants (p. 98) Appendix 6: Control questions to the diet- and activity record (p. 101) Appendix 7: An example of the feedback the participants received (p. 105) Abbreviations and definitions Anteroposterior From front to back