The association between fat free mass and physical performance in overweight and obese older adults

(1)

The association between fat

free mass and physical

performance in overweight and

obese older adults.

Katya de Groot (500661073) 16-5-2016

Thesis number: 2016201 Version: 1

Client: Dr. jr. P.J.M Weijs, Research Group Weight Management Bachelor Thesis: Nutrition and Dietetics

(2)

2 K. S. de Groot

Student number: 500661073 katya.de.groot@hva.nl

Bachelor Thesis: Nutrition and Dietetics

HBO Nutrition and Dietetics

Amsterdam University of Applied Science Client: Dr. jr. P.J.M Weijs

Mentor: Ir. Amely Verreijen

Examinator: Dr. Ir. Mariëlle Engberink Graduation period: February to June 2016 Thesis number: 2016201

(3)

3

Acknowledgment

In this thesis you will read about my research on the association between fat-free mass and physical performance in overweight and obese older adults. It is written for the Bachelor of Nutrition and Dietetics at the Amsterdam University of Applied Sciences. The research data used for this thesis were derived from the baseline data of three studies performed in the Amsterdam Nutritional Assessment Center (PROBE, WelPrex, and the Muscle Preservation Study). The thesis was written in a period of 20 weeks and was commissioned by Dr. jr. P.J.M Weijs.

I would like to thank the following people for their help during the production of my thesis. First of all, I would like to thank my mentor, Amely Verreijen. She helped me understand how to interpret my statistical analysis, and gave me advice during the writing period. Furthermore, I would like to thank Teresa Miguel for sharing her results from the ESPEN abstract that my thesis is based on. I would also like to thank Robert Memelink, coordinator of the PROBE study, for his support

throughout the thesis period. Finally, many thanks to the PROBE team for letting me bounce ideas and questions off them.

Katya de Groot

(4)

4

Abstract

Rationale: Obesity worsens age-related decline in physical performance. A high BMI, however, is not only related to a higher fat mass (FM), but also to a higher fat-free mass (FFM). For dietetic

treatment it is relevant to know how body composition, specifically FFM, is related to physical performance in the growing population of older overweight and obese adults.

Method: We included 246 overweight and obese adults (55-80 years old) in a cross-sectional analysis and studied the association between FFM parameters (FFM in kg, FFM index in kg/m2 (FFMi) and FFM%) and physical performance (handgrip strength (HGS), 4m gait speed (4mGS), 400mGS and time to perform 5 chair stands (CS)). FFM and FM were measured by air displacement plethysmography. Linear regression analysis was performed with determinant FFM in kg, FFMi, and FFM% using physical performance measures as outcome variables. Adjustments were made for sex (and height for CS). Because age was an effect modifier for HGS and 4mGS in the three FFM models, analyses were stratified for age (younger: 55-65 years old vs. older: 66-80 years old).

Results: The mean age of the subjects was 64±5 years with a BMI of 33±5 kg/m2. 43% were men. FFM in kg and FFMi show no significant association with physical performance measures. FFM% was significantly associated with all physical performance measures. For all subjects, in the crude model, an increase of 1% in FFM% was associated with +1.6 kg in HGS, +0.01m/s 4mGS, +0.01m/s 400mGS and decrease in the 5x chair stand -0.1s CS (p<0.01). The adjusted model (sex and height CS) shows an increase of 1% in FFM% was associated with +0.6kg HGS, +0.01s/m 4mGS, +0.01s/m 400mGS (p<0.01). The adjusted ‘younger’ group shows that an increase of 1% in FFM% was significantly associated with +0.01s/m 440mGS (p<0.01). The adjusted ‘older’ group shows an increase of 1% in FFM% was associated with +1.4 kg HGS, +0.02s/m 4mGS, +0.02s/m 400mGS (p<0.01).

Conclusion: A higher FFM% is significantly associated with better physical performance in older obese adults, while the absolute amount of FFM in kg and FFMi showed no significant association. In conclusion: FFM% seems to be the better parameter to predict physical performance in overweight and obese ‘older’ elderly, than ‘younger’ elderly people. These findings support that weight loss treatment should focus on FFM preservation and FM loss in overweight and obese older adults. This means that dieticians should promote physical activity throughout the treatment for overweight and obese elderly people.

Disclosure of Interest: None

(5)

5

Content

Acknowledgment ... 3

Abstract ... 4

1. Introduction ... 7

2. Subjects and methods ... 9

2.1 Subjects ... 9 2.2 Body composition ... 9 2.3 Physical performance ... 10 2.4 Statistics... 10 3. Results ... 11 4.Discussion ... 16 5. References ... 18

Appendix I: Espen abstract ... 20

(6)

(7)

7

Introduction

In recent decades, the prevalence of obesity has been rising. In 2014, more than 1.9 billion adults were overweight and 600 million were obese. Obesity is a risk factor for cardiovascular disease, diabetes, and even some cancers1. Life expectancy has increased over the years in most countries. Although people generally live longer, this does not mean that the years spent in good health have increased. It is estimated that by the year 2035 over 25% of Europeans will be aged 65 and over2. The prevalence of obesity in the older population is growing progressively in all layers in all socio-economic groups of developed countries in Europe3. This means that the number of obese older adults is also rising.

The body mass index (BMI) is a simple tool that is used to determine the classification of overweight and obesity. BMI is a widely used method for both younger and older adults. However, it does not take into account age-dependent changes in body composition in elderly people. Height loss

beginning at the age 30 until the age of 70 shows an average total decline of 3 cm for men, and 5 cm for women. Beyond the age of 80, a higher rate in loss of height is seen in men and women. This may induce a false BMI increase, despite minimal changes in weight3.

The relationship between BMI and mortality in elderly people shows a ‘U-shape’. It seems that mortality in elderly people is lower between the BMI of 25-28 kg/m2, and increases when BMI shifts to higher values. Overweight elderly people appear to somehow be ‘protected’ against mortality3,4. A higher fat mass could, however, have the opposite effect, causing a rise in mortality. Visceral fat that is stored around the organs, is especially associated with increased risk of cardiovascular diseases, resistance to insulin, hypertension, and some types of cancer5. The question that arises is whether higher fat free mass (FFM), and not fat mass(FM), is the cause of lower mortality in obese elderly people, since obese subjects in general have a higher fat free mass6. Graf et al. (2015) found that a low FFM Index (< 19.5 kg/m2) for older men is more strongly associated with mortality than a low BMI. This relation was not clear for women7.

Aging also causes a change in body composition. Beginning around the age of 40-50, there is a gradual decrease of lean mass, referred to as the process of sarcopenia8. With increasing age, the amount of fat mass significantly rises, even without a change in the body weight and BMI3. Excessive body fat may cause a decrease in mobility and physical function and could affect performance of daily activities9,10.

Because of the uncertainties of the health benefits of a lower BMI, controversy exists around weight loss with calorie restriction in the elderly11,12. Rapid unintentional weight loss in the elderly is usually a sign of underlying health problems and conditions. If weight loss is intentional, not only do the elderly lose fat mass, but also muscle. The loss of muscle mass can be up to more than 25% of the total weight-loss in elderly people11. This could cause a decrease in physical

performance12. However, studies show that voluntary weight loss with exercise leads to fat mass loss and an improvement of physical functioning13.

The strength of muscles and balance in elderly people are qualities that are directly linked to their health. Decline in physical performance exposes elderly people to frailty, dependence, and the inability to perform daily tasks. The changes in body composition of the elderly, such as redistribution of body fat and loss of lean muscle mass, influence the physical performance of day-to-day life14.

(8)

8 Therefore, the aim of this study is to investigate the association between FFM and physical

performance in overweight and obese older adults. The findings of this study can be used by dieticians to improve body composition, enhancing physical performance in this age group.

(9)

9

2. Subjects and methods

2.1 Subjects

For this thesis, baseline data of three different weight-loss trials were included for data analysis. Subjects were between 55 and 88 years old. They had a BMI >25 kg/m2 and a waist circumference >88 cm (women) or >102 cm (men). Table 1 displays the inclusion criteria for the three studies. Subjects were excluded if they had any malignant diseases during the last five years and if they used insulin or corticosteroids for systemic use, kidney failure, or liver failure. Detailed information about the inclusion and exclusion criteria of the three studies is available online in the Dutch Trial Register (http://www.trialregister.nl with numbers NTR2751; NTR4556; NTR4497).

The three studies were 10 to 13-week weight loss trials in which the effect of a higher protein diet (with or without supplementation) and/or resistance training on preservation of lean mass was investigated. All studies were performed at the Amsterdam University of Applied Science in the Netherlands at the Department of Nutrition and Dietetics between 2011 and 2015. Written informed consent was obtained from all subjects and ethical approval was obtained for the three studies. Table 1: inclusion criteria research

Study

name: Muscle Preservation Study Welprex Study Probe Study

Inclusion 1.Age between 55 and 85 years;

2. BMI > 30 kg/m2 or BMI > 28 kg/m2 in combination with waist circumference > 88 cm (women) or > 102 cm (men)

1. Age 55 and over 2. BMI >28 kg/m2 and/or BMI > 25 kg/m2 in combination with a waist circumference > 88 cm (women) or > 102 cm (men)

1. Age 55-85 years 2. BMI >27 kg/m2 in combination with a waist circumference > 88 cm for women or > 102 cm for men

3. Ambulant type 2 diabetes patients (verified by used medication for diabetes). In the event no medication is used HbA1c should be >53 mmol/mol (>7.0 %)

2.2 Body composition

The body composition (FFM and FM) was measured using air displacement plethysmography (Bodpod, Life Measurement Inc., Concord (CA), USA). This device was also used to measure the weight of the subjects. The Bodpod system was calibrated before each measurement using the precision 50.384-L cylinder provided by the instrument manufacturer. The scale of the Bodpod system was calibrated using two 10kg weights. The measurements were performed in a fasting state, wearing only underwear. Subjects underwent measurements in the Bodpod 2 to 3 times in case the measurements deviated by more than 150ml in body volume. The Siri equation was used to estimate the percentage of body fat from the measured density15,16. FM and FFM were calculated using body fat percentage and body weight

.

Height was measured using a wall-mounted stadiometer (Seca 222, Seca, Hamburg, Germany).

(10)

10

2.3 Physical performance

The physical tests were conducted in all three studies using a standardized protocol. Handgrip strength (HGS) (JAMAR 5030J1, Sammons Preston Rolyan, Bollingbrook, CA) was measured in kilograms. The tests were performed three times for both the right and left hand in a sitting position with the arm at a 90-degree angle. The first trail was to make the participant comfortable with the equipment. The next two trails were recorded to the nearest 0.1kg. The sum of the maximum value of both hands was calculated.

The 4m gait-speed (4mGS) was performed on a pre-set course using a starting and ending line 4 meters apart, with an additional >0.5 m at each end. The participant was asked to walk at regular walking speed for a distance of 4 meters. The regular walking speed test was performed twice and the time was measured using a stopwatch. Time was stopped when one of the participants’ feet completely crossed the end line. The fastest of two repetitions was used for the analyses.

The 5 chair stands (CS) were performed using an armless chair placed against a wall. During the test, the subject’s arms were folded across their chest. A researcher first demonstrated the movement before the subject was asked to copy the movement. A single chair stand was performed to assess the subject’s ability to complete the test. After this, 5 chair stands were timed using a stopwatch. The time started on the commend “stand” and ended upon the 5th fully-extended knee angle.

The 400m walking speed (400mGS) was performed on a pre-set 20m course. A cone was placed on the second and last line of the course to indicate the turning point. The course was repeated 10 times and for the last round the participant was asked to walk over the starting line. The time was monitored using a stopwatch. The number of rounds were counted and written down on paper. The time stopped when one of the participant’s feet completely crossed the finish line.

2.4 Statistics

The data from the baseline of the three studies were statistically analyzed using SPSS Software (version 22.0, IBM). The data was analyzed using linear regression analysis using determinants absolute amount of Fat Free Mass in kilograms (FFMkg), FFMi, and FFM%, and HGS, 4mGS, 5x CS, and 400mGS as outcome variables.

Confounders were determined using linear regression analysis. A >10% change in the beta for the FFM parameter determined that it was a confounder. The models for each test were adjusted for sex (and additionally adjusted for height for CS). Effect modification for age and gender was examined using linear regression analysis. An alpha of 10% was used to determine the effect modification. Effect modification was found for age and analyses were stratified for the effect modifier age. The two lowest tertiles of age were compared to the highest tertile of age (younger group: 55-65 years old vs older group: 66-80 years old). An alpha of 5% was used to determine statistical significance.

(11)

11

3. Results

In this analysis, 246 elderly adults were included (56.9% female) with a mean age of 64±5 years old and a BMI of 33±5kg/m2. For the analysis, three parameters for FFM were used in relation to measures of physical performance. The variables for FFM were the absolute amount of FFM in kg, FFMi, and FFM%. Physical performance outcome variables were HGS (kg), 4mGS (m/s), 5x CS (s), and 400mGS (m/s). Sex was related to all three FFM parameters in relation to physical performance measures. The models were adjusted for sex as a confounder.

The absolute amount of FFM in kg was not significantly associated with physical performance measures, except for HGS. An increase of 1 kg in FFM was associated with +1.4 kg HGS (p<0.001) in the crude model. When adjusted for sex an increase of 1 kg in FFM was associated with +0.8 kg HGS (p<0.001). FFM and performance in the 4mGS and 400mGS were significantly associated in the crude model, but not in the adjusted model. The association between FFM in kg and physical performance is shown in Table 2. The R2of the models displays how much of the variability in the outcome variable is predicted by the model.

Table 2: association of absolute amount of FFM in kg with physical performance measures

All Subjects (n=246)

Crude models determinant: FFM in kg Beta (SE) P R2 Hand grip strength (kg) 1.356 (0.075) <0.001 0.582

4m gait speed (m/s) 0.003 (0.001) 0.027 0.021

Timed chair stand 5x (s) -0.012 (0.021) 0.581 0.001 400m gait speed (m/s) 0.005 (0.001) <0.001 0.080 Adjusted models determinant: FFM in kg* Beta (SE) P R2 Hand grip strength (kg) 0.808 (0.121) <0.001 0.632

4m gait speed (m/s) 0.004 (0.002) 0.135 0.021

Timed chair stand 5x (s) 0.025 (0.045) 0.581 0.025

400m gait speed (m/s) 0.003 (0.002) 0.124 0.087

*adjusted models for sex (and height for 5 chair stands) FFM= fat free mass

Table 3 shows the association between FFMi and the physical performance variables. The crude model shows that a 1 unit increase of FFMi was associated with +5.0 kg HGS (p<0.001) and in the adjusted model +1.8 kg HGS (p<0.001). There were no significant associations between physical performance and FFMi in the crude model and the adjusted model.

Table 3: association of FFMi with physical performance measures

Crude models determinant: FFMi Beta (SE) P R2

Hand grip strength (kg) 5.030 (0.401) <0.001 0.402

4m gait speed (m/s) 0.004 (0.006) 0.541 0.002

Timed chair stand 5x (s) -0.076 (0.093) 0.418 0.003

400m gait speed (m/s) 0.009 (0.005) 0.114 0.011

Adjusted models determinant: FFMi* Beta (SE) P R2 Hand grip strength (kg) 1.761(0.463) <0.001 0.587

4m gait speed (m/s) -0.006 (0.009) 0.472 0.014

Timed chair stand 5x (s) 0.074 (0.128) 0.565 0.025 400m gait speed (m/s) -0.014 (0.007) 0.053 0.092 *adjusted models for sex (and height for 5 chair stands) FFMi = fat free mass index

(12)

12 Table 4 shows that FFM% presented significant associations with all measures for physical

performance. Unadjusted, a 1% increase of FFM% shows an increase of +1.6 kg in HGS, +0.01m/s in 4mGS, +0.01m/s in 400mGS and a decrease in the 5x CS of -0.1s CS (p<0.01). The adjusted model shows a significant +0.6kg in HGS, +0.01m/s in 4mGS, +0.01m/s in 400mGS (p<0.01), an increase per 1% in FFM% and a significant decrease of 0.1s in CS.

Table 4: association of FFM% with physical performance measures

Crude models determinant: FFM% Beta (SE) P R2

Hand grip strength (kg) 1.616 (0.119) <0.001 0.441 4m gait speed (m/s) 0.007 (0.002) <0.001 0.062 Timed chair stand 5x (s) -0.073 (0.028) 0.010 0.028 400m gait speed (m/s) 0.011 (0.002) <0.001 0.180 Adjusted models determinant: FFM%* Beta (SE) P R2 Hand grip strength (kg) 0.600 (0.151) <0.001 0.589 4m gait speed (m/s) 0.011 (0.003) <0.001 0.075 Timed chair stand 5x (s) -0.092 (0.042) 0.029 0.044 400m gait speed (m/s) 0.012 (0.002) <0.001 0.182 *adjusted models for sex (and height for 5 chair stands) FFM= fat free mass

Effect modification was found for age. Using p<0.10 in the effect modification analysis. The relation for age was found between FFM% and 400mGS, and HGS. Analysis was split in two age categories. The ‘young’ elderly (55-65 years old) and the ‘old’ elderly (66-80 years old). The ‘young’ group contained 161 subjects, and the ‘old’ group contained 85 subjects. This division of participants into 'young' and 'old' categories was done by comparing the highest tertile with the lowest two age-tertiles. The absolute amount of FFM in kg, and FFMi was not significantly associated with the physical performance measures, and was not used for analyses per age category. 400mGS and HGS were most strongly associated with the physical performance measures in the two age categories. The analysis was split into the two age categories in Figure 1 and Figure 2. The 4mGS and CS were less strongly associated with the two age categories. Therefore, the analysis was not spilt into the two groups in Figure 3 and Figure 4.

(13)

13

Figure 1: Fat Free Mass % and 400m gait speed in the younger and older age group

Figure 1 shows the crude association between FFM% and 400mGS in the different age groups. For the adjusted model, a 1% increase of FFM% shows a +0.01 m/s increase in the younger group and a +0.02 m/s increase in the older group. Both have a p <0.001.

(14)

14 Figure 2:Fat Free Mass % and hand grip strength in the younger and older age group

Figure 2 shows the crude association between HGS and FFM% in the different age groups. In the adjusted model, a 1% increase of FFM% shows a +0.2 kg increase in the younger group and a +1.4 increase in the older age group. Both have a p<0.001.

Figures 3 and 4 show the crude association between 4mGS, CS and FFM%. The adjusted models show that for the younger group’s 4mGS, a 1% increase of FFM% was associated with +0.01m/s (p0.05) and the older group shows a +0.02m/s (p<0.001) increase. The adjusted model for the CS shows that an increase of 1% of FFM% was associated with a -0.05s (p0.35) decrease in CS in the younger group and a -0.17s (p0.02) decrease in the older group.

(15)

15 Figure 3: Fat Free Mass % and 4 meter gait speed in the younger and older age group

(16)

16

4.Discussion

The main inquiry of this study is to investigate the association of body composition with physical performance in overweight and obese older adults. The results show that FFM% is the parameter that is most strongly associated with FFM’s effect on physical functioning. FFM in kg and FFMi show no significant association with physical performance parameters, except for in the HGS. The strong association of FFM% could be because FFM% also reflects FM%. A higher FM% is associated with poorer physical performance17. The study of Shin et al. (2014) found that FM is negatively associated with physical performance measurements. Reduced physical performance could lead to falls and limitations in other daily physical performances17,18. FFM in kg and FFMi do not have this counterpart in their model. FFM in kg and FFMi do not show the distribution of the FM. However, an unexpected association was found between FFMi and an increase in waist circumference. Further research should be done to investigate the cause of this association. Even when adjusting the models FFM in kg and FFMi for FM, no significant association was found.

FFM in kg and FFMi show no significant association with physical performance measures, except for HGS, where the association was significant. This association is most likely due to the fact that HGS measures the muscle strength19. Stevens et al. found a significant association between grip strength and physical performance in men and women20. When looking at the amount of FFM, it could be that the association with physical performance is not very strong and that muscle strength is a better predictor of FFM21. This study did not look at the quality of the muscle, and intramuscular fat could be a confounding factor. The use of HGS in a dietician’s practice could be managed. It is a low cost and non-invasive method to use. However, Tieland et al.(2015) found that the HGS is not a valid measurement for change in muscle strength over a period of time. They performed a 24-week resistance training intervention and, although the psychical performance did significantly improve, it was not translated to results in the HGS22. Besides the monitoring of body composition of elderly people during weight loss, perhaps the focus should also be on physical performance. Research into the use of physical performance measures in dietician practices could be done to determine the correct tool for dieticians to use.

The confounders in the regression analysis were sex and height (for CS only). Overall, men have more skeletal muscle mass (in absolute terms in kilograms) than women23. Because men have more lean mass, it is likely that there is a difference in performance in the two groups. Results show that sex is a confounder for mostly absolute amount of FFM in kg and FFMi. The effect on FFM% is less strong. That is why it was not used as a confounder in the ESPEN abstract that this thesis is based on. When tested if sex was an effect modification, there were positive results but with a less significant effect, and weaker associations compared to age. Because of this, it was not included in the analyses of the baseline data. A separate analysis could be done using sex as an effect modification, where the analysis is split into groups of men and women. Such analysis was outside the scope of this study. The analyses were divided into two groups. One included the ‘younger’ elderly (55-65 years old) and one included the ‘older’ elderly (66-80 years old). The results show a difference between the

‘younger’ and the ‘older’ elderly. The older group shows a stronger association between FFM% and physical performance measurements compared to the younger group. Sallinen et al. (2011) also found a difference in age groups in the relationship between body composition and walking speed24. They speculate that changes in body composition, such as loss of lean mass and fat distribution in the body in older elderly people, could be the cause of the difference between the age groups. They speculate that the health issues of the older age group may override the effect of high FM on physical performance3,24. A different study found that FFM and appendicular skeletal muscle mass stay stable until the age of 60, after which a more accelerated change was seen25. However, another

(17)

17 study found that FFM begins to decline at the fourth decade of life26. It is possible that, although FFM loss starts earlier in life, the amount of lean mass lost is not yet significantly associated with physical performance. This would suggest that the loss of FFM in the older group is more related to physical performance because of a more accelerated loss of lean mass. This would mean that dieticians need to monitor FFM during treatment of elderly people. Research into which tool is most practical, cost efficient, and valid to use in a dietician practice to measure FFM% in the elderly is advised.

One of the strengths of this study is the use of a valid instrument to determine body composition. The use of air displacement plethysmography (by the BodPod device) gives a reliable outcome of body composition27. The physical performance measurements were all conducted in the same location, using the same measurement protocols for body composition, handgrip strength and physical performance. This study did not investigate the quality of the muscle. This factor could also influence the performance of physical functions. Further research could be done to look at the quality of the muscle compared to physical performance. Additionally, FFM contains muscle mass and other lean mass. This is less precise than muscle mass alone. The use of an MRI-scan to

determine the amount of muscle mass was not available for this study. A future study could focus on absolute amount of muscle mass in association with physical performance instead of FFM alone. A higher FFM% is significantly associated with better physical performance in older obese adults while absolute amount of FFM in kg and FFMi showed no significant association. In conclusion, FFM% appears to be the better parameter to predict physical performance in overweight and obese ‘older’ elderly people, than ‘younger’ elderly people. These findings support the notion that weight loss treatment should focus on FFM preservation and FM loss in overweight and obese older adults. This means that dieticians should promote physical activity throughout treatment for overweight and obese elderly people. The use of physical performer measures, such as CS and 4mGS, could be used to monitor the FFM. However, further research on use of these parameters for physical performance in dietician practice should be explored over a period of time. Collaboration between dieticians, fitness centers and physiotherapists to encourage physical activity in obese elderly people could help preserve FFM and promote the health of this age group. This is a venture that should be pursued by all parties because they can assist and help each other to make the most complete and safe program for elderly clients to follow.

(18)

18

5. References

1. WHO. World Health Organization. Obesity and overweight, factsheet N311.2015. http://www.who.int/mediacentre/factsheets/fs311/en/

2. Mathus-Vliegen EM; Obesity Management Task Force of the European Association for the Study of Obesity. Prevalence, pathophysiology, health consequences and treatment options of obesity in the elderly: a guideline. Obes Facts. 2012;5(3):460-83. 3. Zamboni M, Mazzali G. Health consequences of obesity in the elderly: a review of

four unresolved questions. Int J Obes (Lond). 2005 Sep;29(9):1011-29

4. Flicker L, McCaul KA, Hankey GJ, Jamrozik K, Brown WJ, Byles JE, Almeida OP. Body mass index and survival in men and women aged 70 to 75. J Am Geriatr Soc. 2010 Feb;58(2):234-41

5. Patel P, Abate N. Body fat distribution and insulin resistance. Nutrients. 2013 Jun 5;5(6):2019-27.

6. Forbes GB, Welle SL. Lean body mass in obesity. Int J Obes. 1983;7(2):99-10 7. Graf CE, Karsegard VL, Spoerri A, Makhlouf AM, Ho S, Herrmann FR, Genton L.

Body composition and all-cause mortality in subjects older than 65 y. Am J Clin Nutr. 2015 Apr;101(4):760-7.

8. Breen L, Phillips SM. Skeletal muscle protein metabolism in the elderly: Interventions to counteract the 'anabolic resistance' of ageing. Nutr Metab (Lond). 2011 Oct 5;8:68.

9. Woo J, Leung J, Kwok T. BMI, body composition, and physical functioning in older adults. Obesity (Silver Spring). 2007 Jul;15(7):1886-94.

10. Porter Starr KN, Pieper CF, Orenduff MC, McDonald SR, McClure LB, Zhou R,

Payne ME, Bales CW. Improved Function With Enhanced Protein Intake per Meal: A Pilot Study of Weight Reduction in Frail, Obese Older Adults. J Gerontol A Biol Sci Med Sci. 2016 Jan 18.

11. Porter Starr KN, Bales CW. Excessive Body Weight in Older Adults: Concerns and Recommendations. Clinics in geriatric medicine. 2015;31(3):311-326.

12. Miller SL, Wolfe RR. The danger of weight loss in the elderly. J Nutr Health Aging. 2008 Aug-Sep;12(7):487-91.

13. Villareal DT, Chode S, Parimi N, Sinacore DR, Hilton T, Armamento-Villareal R, Napoli N, Qualls C, Shah K. Weight loss, exercise, or both and physical function in obese older adults. N Engl J Med. 2011 Mar 31;364(13):1218-29.

14. Silva Nde A, Pedraza DF, de Menezes TN. Physical performance and its

association with anthropometric and body composition variables in the elderly. Cien Saude Colet. 2015 Dec;20(12):3723-32.

(19)

19 15. Ginde SR, Geliebter A, Rubiano F, Silva AM, Wang J, Heshka S, Heymsfield SB.

Air displacement plethysmography: validation in overweight and obese subjects. Obes Res. 2005 Jul;13(7):1232-7.

16. Wagner DR, Heyward VH, Gibson AL. Validation of air displacement

plethysmography for assessing body composition. Med Sci Sports Exerc. 2000 Jul;32(7):1339-44.

17. Shin H, Liu PY, Panton LB, Ilich JZ. Physical performance in relation to body composition and bone mineral density in healthy, overweight, and obese postmenopausal women. J Geriatr Phys Ther. 2014 Jan-Mar;37(1):7-16. 18. Broadwin J, Goodman-Gruen D, Slymen D. Ability of fat and fat-free mass

percentages to predict functional disability in older men and women. J Am Geriatr Soc. 2001 Dec;49(12):1641-5.

19. Roberts HC, Denison HJ, Martin HJ, Patel HP, Syddall H, Cooper C, Sayer AA. A review of the measurement of grip strength in clinical and epidemiological studies: towards a standardised approach. Age Ageing. 2011 Jul;40(4):423-9. 20. Stevens PJ, Syddall HE, Patel HP, Martin HJ, Cooper C, Aihie Sayer A. Is grip

strength a good marker of physical performance among community-dwelling older people? J Nutr Health Aging. 2012;16(9):769-74.

21. Kim YH, Kim KI, Paik NJ, Kim KW, Jang HC, Lim JY. Muscle strength: A better index of low physical performance than muscle mass in older adults. Geriatr Gerontol Int. 2016 May;16(5):577-85.

22. Tieland M, Verdijk LB, de Groot LC, van Loon LJ. Handgrip strength does not

represent an appropriate measure to evaluate changes in muscle strength during an exercise intervention program in frail older people. Int J Sport Nutr Exerc

Metab. 2015 Feb;25(1):27-36.

23. Janssen I, Heymsfield SB, Wang ZM, Ross R. Skeletal muscle mass and

distribution in 468 men and women aged 18-88 yr. J Appl Physiol (1985). 2000 Jul;89(1):81-8. Erratum in: J Appl Physiol (1985). 2014 May 15;116(10):1342. 24. Sallinen J, Stenholm S, Rantanen T, Heliöaara M, Sainio P, Koskinen S. Effect

of age on the association between body fat percentage and maximal walking speed. J Nutr Health Aging. 2011 Jun;15(6):427-32.

25. Kyle UG, Genton L, Hans D, Karsegard L, Slosman DO, Pichard C. Age-related differences in fat-free mass, skeletal muscle, body cell mass and fat mass between 18 and 94 years. Eur J Clin Nutr. 2001 Aug;55(8):663-72.

26. Nair KS. Aging muscle. Am J Clin Nutr. 2005 May;81(5):953-63.

27. Bosy-Westphal A, Mast M, Eichhorn C, Becker C, Kutzner D, Heller M, Müller MJ. Validation of air-displacement plethysmography for estimation of body fat mass in healthy elderly subjects. Eur J Nutr. 2003 Aug;42(4):207-16.

(20)

20

Appendix I: Espen abstract

A higher fat free mass percentage is associated with better physical performance in overweight and obese older adults

Teresa S. Miguel1,2, Katya de Groot1, Amely M. Verreijen1, Mariëlle F. Engberink1, Robert G. Memelink1, Peter J.M. Weijs1,3

1

Research Group Weight Management, Faculty of Sports and Nutrition, Amsterdam University of

Applied Sciences, 2University CEU-San Pablo of Madrid, 3VU University Medical Center, Amsterdam,

Netherlands

Rationale: Obesity worsens the age related decline in physical performance. A high BMI however is not only related to a higher fat mass (FM), but also to a higher fat free mass (FFM). For dietetic treatment it is relevant to know how body composition (FFM%) is related to physical performance in the growing population of older overweight and obese adults.

Methods: We included 246 overweight and obese adults (55-80y) in a cross-sectional analysis and studied the association between FFM% and physical performance (handgrip strength (HGS), 4m gait speed (4mGS), 400mGS and time to perform 5 chair stands (CS)). FFM and FM were measured by

air

displacement plethysmography

. Linear regression analysis was performed with determinant FFM% and physical performance measures as outcome variables. Adjustments were made for age (and height for CS). Because age was an effect modifier for HGS and 4mGS, analyses were stratified for age (younger: 55-65y vs. older: 66-80y).

Results: Mean age of the subjects was 64±5y with a BMI of 33±5kg/m2 and 43% were men. FFM% was significantly associated with all physical performance measures. For all subjects an increase of 1% in FFM was associated with +1.6kg HGS, +0.01m/s 4mGS , +0.01m/s 400mGS and -0.1s CS (all: P<0.01). An increase of 1% in FFM was associated with a +2.0kg for HGS (P<0.01) and +0.02m/s for 4mGS (P<0.01) in the older subjects. In the younger, associations of +1.5kg (P<0.01) and +0.01m/s (P=0.02) were shown.

Conclusion: A higher FFM% is significantly associated with better HGS, 4mGS, 400mGS and CS in older obese adults. These findings support that weight loss treatment should focus on FFM preservation and FM loss in overweight and obese older adults.

Disclosure of Interest: None

(21)

Appendix II: Syntax

* descriptives age, sex, and BMI.

DESCRIPTIVES VARIABLES=AGE

/STATISTICS=MEAN STDDEV MIN MAX. FREQUENCIES VARIABLES=SEX

/ORDER=ANALYSIS.

DESCRIPTIVES VARIABLES=BMI_V1 /STATISTICS=MEAN STDDEV MIN MAX.

*crude models FFM%.

REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER FFMpct. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Time5Stands_V1 /METHOD=ENTER FFMpct. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT WALKSPEED400_V1 /METHOD=ENTER FFMpct. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN

/DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER FFMpct.

RECODE NAGE (3=1) (1 thru 2=0) INTO young_old. EXECUTE.

*crude models FFMI.

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER FFMi. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER FFMi. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Time5Stands_V1 /METHOD=ENTER FFMi. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT WALKSPEED400_V1 /METHOD=ENTER FFMi. *crude models ffmkg. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT WALKSPEED400_V1 /METHOD=ENTER BP_LWkg_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER BP_LWkg_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER BP_LWkg_V1. REGRESSION /MISSING LISTWISE

(22)

22 /CRITERIA=PIN(.05) POUT(.10)

/NOORIGIN

/DEPENDENT Time5Stands_V1 /METHOD=ENTER BP_LWkg_V1.

*crude models bmi.

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER BMI_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER BMI_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Time5Stands_V1 /METHOD=ENTER BMI_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT WALKSPEED400_V1 /METHOD=ENTER BMI_V1. *crude models FM. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER BP_FATkg_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER BP_FATkg_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Time5Stands_V1 /METHOD=ENTER BP_FATkg_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT WALKSPEED400_V1 /METHOD=ENTER BP_FATkg_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER BP_FATpct_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER BP_FATpct_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Time5Stands_V1 /METHOD=ENTER BP_FATpct_V1. REGRESSION /MISSING LISTWISE

/DEPENDENT WALKSPEED400_V1 /METHOD=ENTER BP_FATpct_V1.

*effect modificatie sex.

COMPUTE effct_sex=FFMpct * SEX. EXECUTE.

/DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER FFMpct SEX effct_sex.

(23)

23 REGRESSION

/MISSING LISTWISE

/DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER FFMpct SEX effct_sex. REGRESSION

/MISSING LISTWISE

/DEPENDENT Time5Stands_V1

/METHOD=ENTER FFMpct SEX effct_sex. REGRESSION

/MISSING LISTWISE

/DEPENDENT WALKSPEED400_V1 /METHOD=ENTER FFMpct SEX effct_sex.

*effectmodificatie age.

COMPUTE effct_age=FFMpct * young_old. EXECUTE.

/DEPENDENT HGSsumRLmax_V1

/METHOD=ENTER FFMpct young_old effct_age. REGRESSION

/MISSING LISTWISE

/DEPENDENT GaitSpeed_V1fastest

/MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10)

/NOORIGIN

/DEPENDENT WALKSPEED400_V1

/METHOD=ENTER FFMpct young_old effct_age.

*effectmodification FFMi.

COMPUTE effect_sex_ffmi=FFMi * SEX. EXECUTE.

COMPUTE effect_age_ffmi=FFMi *young_old. EXECUTE.

/DEPENDENT HGSsumRLmax_V1

/METHOD=ENTER FFMi SEX effect_sex_ffmi. REGRESSION

/MISSING LISTWISE

/DEPENDENT GaitSpeed_V1fastest

/MISSING LISTWISE

/DEPENDENT WALKSPEED400_V1

/MISSING LISTWISE

/DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER FFMi young_old effect_age_ffmi.

(24)

24 /DEPENDENT GaitSpeed_V1fastest

/METHOD=ENTER FFMi young_old effect_age_ffmi.

/DEPENDENT Time5Stands_V1 /METHOD=ENTER FFMi young_old effect_age_ffmi.

/DEPENDENT WALKSPEED400_V1 /METHOD=ENTER FFMi young_old effect_age_ffmi.

*effectmodification FFMkg.

COMPUTE effect_sex_ffmkg=BP_LWkg_V1 *SEX. EXECUTE. COMPUTE effect_age_ffmkg=BP_LWkg_V1 *young_old. EXECUTE. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER BP_LWkg_V1 SEX effect_sex_ffmkg. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER BP_LWkg_V1 SEX effect_sex_ffmkg. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Time5Stands_V1 /METHOD=ENTER BP_LWkg_V1 SEX effect_sex_ffmkg. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT WALKSPEED400_V1 /METHOD=ENTER BP_LWkg_V1 SEX effect_sex_ffmkg. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER BP_LWkg_V1 young_old effect_age_ffmkg. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER BP_LWkg_V1 young_old effect_age_ffmkg. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Time5Stands_V1 /METHOD=ENTER BP_LWkg_V1 young_old effect_age_ffmkg. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT WALKSPEED400_V1 /METHOD=ENTER BP_LWkg_V1 young_old effect_age_ffmkg. *confounder age. REGRESSION /MISSING LISTWISE

/DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER FFMi AGE. REGRESSION

/MISSING LISTWISE

(25)

25 /CRITERIA=PIN(.05) POUT(.10)

/NOORIGIN

/DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER FFMi AGE. REGRESSION

/MISSING LISTWISE

/DEPENDENT Time5Stands_V1 /METHOD=ENTER FFMi AGE. REGRESSION

/MISSING LISTWISE

/DEPENDENT WALKSPEED400_V1 /METHOD=ENTER FFMi AGE. REGRESSION

/MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER FFMpct AGE. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER FFMpct AGE. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Time5Stands_V1 /METHOD=ENTER FFMpct AGE. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT WALKSPEED400_V1 /METHOD=ENTER FFMpct AGE. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER BP_LWkg_V1 AGE. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER BP_LWkg_V1 AGE. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Time5Stands_V1 /METHOD=ENTER BP_LWkg_V1 AGE. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT WALKSPEED400_V1 /METHOD=ENTER BP_LWkg_V1 AGE. *confounder sex. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER BP_LWkg_V1 SEX. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER BP_LWkg_V1 SEX. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Time5Stands_V1 /METHOD=ENTER BP_LWkg_V1 SEX. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10)

(26)

26 /NOORIGIN /DEPENDENT WALKSPEED400_V1 /METHOD=ENTER BP_LWkg_V1 SEX. REGRESSION /MISSING LISTWISE

/DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER FFMi SEX. REGRESSION

/MISSING LISTWISE

/DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER FFMi SEX. REGRESSION

/MISSING LISTWISE

/DEPENDENT Time5Stands_V1 /METHOD=ENTER FFMi SEX. REGRESSION

/MISSING LISTWISE

/DEPENDENT WALKSPEED400_V1 /METHOD=ENTER FFMi SEX. REGRESSION

/MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER FFMpct SEX. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER FFMpct SEX. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Time5Stands_V1 /METHOD=ENTER FFMpct SEX. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT WALKSPEED400_V1 /METHOD=ENTER FFMpct SEX. *confounder height. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER BP_LWkg_V1 HEIGHT_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER BP_LWkg_V1 HEIGHT_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Time5Stands_V1 /METHOD=ENTER BP_LWkg_V1 HEIGHT_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT WALKSPEED400_V1 /METHOD=ENTER BP_LWkg_V1 HEIGHT_V1. REGRESSION /MISSING LISTWISE

/DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER FFMi HEIGHT_V1. REGRESSION

/MISSING LISTWISE

(27)

27 /DEPENDENT GaitSpeed_V1fastest

/METHOD=ENTER FFMi HEIGHT_V1. REGRESSION

/MISSING LISTWISE

/DEPENDENT Time5Stands_V1 /METHOD=ENTER FFMi HEIGHT_V1. REGRESSION

/MISSING LISTWISE

/DEPENDENT WALKSPEED400_V1 /METHOD=ENTER FFMi HEIGHT_V1. REGRESSION

/MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER FFMpct HEIGHT_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER FFMpct HEIGHT_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Time5Stands_V1 /METHOD=ENTER FFMpct HEIGHT_V1. REGRESSION /MISSING LISTWISE

/DEPENDENT WALKSPEED400_V1 /METHOD=ENTER FFMpct HEIGHT_V1.

*confounder physical activity level.

/DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER FFMi PAL_V1.

/DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER FFMi PAL_V1. REGRESSION

/MISSING LISTWISE

/DEPENDENT Time5Stands_V1 /METHOD=ENTER FFMi PAL_V1. REGRESSION

/MISSING LISTWISE

/DEPENDENT WALKSPEED400_V1 /METHOD=ENTER FFMi PAL_V1. REGRESSION

/MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER FFMpct PAL_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER FFMpct PAL_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Time5Stands_V1 /METHOD=ENTER FFMpct PAL_V1. REGRESSION /MISSING LISTWISE

(28)

28 /DEPENDENT WALKSPEED400_V1

/METHOD=ENTER FFMpct PAL_V1. REGRESSION

/MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER BP_LWkg_V1 PAL_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER BP_LWkg_V1 PAL_V1. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT Time5Stands_V1 /METHOD=ENTER BP_LWkg_V1 PAL_V1. REGRESSION /MISSING LISTWISE

/DEPENDENT WALKSPEED400_V1 /METHOD=ENTER BP_LWkg_V1 PAL_V1.

*analyze confounders.

SORT CASES BY young_old.

SPLIT FILE SEPARATE BY young_old.

*ffmi.

DATASET ACTIVATE DataSet1. REGRESSION

/MISSING LISTWISE

/DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER FFMi SEX. REGRESSION

/MISSING LISTWISE

/DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER FFMi SEX. REGRESSION

/MISSING LISTWISE

/METHOD=ENTER FFMi SEX HEIGHT_V1. REGRESSION

/MISSING LISTWISE

/DEPENDENT WALKSPEED400_V1 /METHOD=ENTER FFMi SEX.

*ffm%.

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER FFMpct SEX. REGRESSION /MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER FFMpct SEX. REGRESSION /MISSING LISTWISE

/METHOD=ENTER FFMpct SEX HEIGHT_V1. REGRESSION

/MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT WALKSPEED400_V1 /METHOD=ENTER FFMpct SEX. *ffm KG. REGRESSION /MISSING LISTWISE

/DEPENDENT HGSsumRLmax_V1 /METHOD=ENTER BP_LWkg_V1 SEX.

(29)

29 REGRESSION

/MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT GaitSpeed_V1fastest /METHOD=ENTER BP_LWkg_V1 SEX. REGRESSION /MISSING LISTWISE

/METHOD=ENTER BP_LWkg_V1 SEX HEIGHT_V1. REGRESSION

/MISSING LISTWISE

/STATISTICS COEFF OUTS R ANOVA /CRITERIA=PIN(.05) POUT(.10) /NOORIGIN /DEPENDENT WALKSPEED400_V1 /METHOD=ENTER BP_LWkg_V1 SEX. *graph. * Chart Builder. GGRAPH /GRAPHDATASET NAME="graphdataset" VARIABLES=FFMi WALKSPEED400_V1 young_old MISSING=LISTWISE

REPORTMISSING=NO

/GRAPHSPEC SOURCE=INLINE. BEGIN GPL

SOURCE: s=userSource(id("graphdataset")) DATA: FFMi=col(source(s), name("FFMi")) DATA: WALKSPEED400_V1=col(source(s), name("WALKSPEED400_V1"))

DATA: young_old=col(source(s), name("young_old"), unit.category()) GUIDE: axis(dim(1), label("FFMi"))

GUIDE: axis(dim(2), label("Regular walking speed 400m baseline_m/s")) GUIDE: legend(aesthetic(aesthetic.color.exterior), label("young_old")) ELEMENT: point(position(FFMi*WALKSPEED400_V1), color.exterior(young_old)) END GPL. * Chart Builder. GGRAPH /GRAPHDATASET NAME="graphdataset" VARIABLES=BP_LWkg_V1 WALKSPEED400_V1 young_old MISSING=LISTWISE REPORTMISSING=NO /GRAPHSPEC SOURCE=INLINE. BEGIN GPL SOURCE: s=userSource(id("graphdataset")) DATA: BP_LWkg_V1=col(source(s), name("BP_LWkg_V1")) DATA: WALKSPEED400_V1=col(source(s), name("WALKSPEED400_V1")) DATA: young_old=col(source(s), name("young_old"), unit.category())

GUIDE: axis(dim(1), label("BodPod Fat free mass baseline_kg"))

GUIDE: axis(dim(2), label("Regular walking speed 400m baseline_m/s")) GUIDE: legend(aesthetic(aesthetic.color.exterior), label("young_old")) ELEMENT: point(position(BP_LWkg_V1*WALKSPEED400_V1), color.exterior(young_old)) END GPL. * Chart Builder. GGRAPH /GRAPHDATASET NAME="graphdataset" VARIABLES=BP_LWkg_V1 WALKSPEED400_V1 young_old MISSING=LISTWISE REPORTMISSING=NO /GRAPHSPEC SOURCE=INLINE. BEGIN GPL SOURCE: s=userSource(id("graphdataset")) DATA: BP_LWkg_V1=col(source(s), name("BP_LWkg_V1")) DATA: WALKSPEED400_V1=col(source(s), name("WALKSPEED400_V1")) DATA: young_old=col(source(s), name("young_old"), unit.category())

GUIDE: axis(dim(1), label("BodPod Fat free mass baseline_kg"))

GUIDE: axis(dim(2), label("Regular walking speed 400m baseline_m/s")) GUIDE: legend(aesthetic(aesthetic.color.exterior), label("young_old")) ELEMENT: point(position(BP_LWkg_V1*WALKSPEED400_V1), color.exterior(young_old)) END GPL. * Chart Builder. GGRAPH /GRAPHDATASET NAME="graphdataset"

VARIABLES=FFMpct WALKSPEED400_V1 young_old MISSING=LISTWISE

REPORTMISSING=NO

/GRAPHSPEC SOURCE=INLINE. BEGIN GPL

SOURCE: s=userSource(id("graphdataset")) DATA: FFMpct=col(source(s), name("FFMpct")) DATA: WALKSPEED400_V1=col(source(s), name("WALKSPEED400_V1"))

(30)

30 DATA: young_old=col(source(s),

name("young_old"), unit.category()) GUIDE: axis(dim(1), label("FFMpct"))

GUIDE: axis(dim(2), label("Regular walking speed 400m baseline_m/s")) GUIDE: legend(aesthetic(aesthetic.color.exterior), label("young_old")) ELEMENT: point(position(FFMpct*WALKSPEED400_V1), color.exterior(young_old)) END GPL.