Older individuals experience alterations in physical activity, body composition, appetite and food intake that may lead to a dysregulation of energy balance(Reference Roberts and Rosenberg1). Typically, there is an increase in daily fullness(Reference Morley2) and a tendency to decrease energy consumption(Reference Wilson, Purushothaman and Morley3, Reference Vellas, Hunt and Romero4), leading to anorexia of ageing, lower body weight and sarcopenia. However, most older adults over-consume energy relative to their need, resulting in weight gain(Reference Roberts and Rosenberg1). It is important to investigate exercise- and diet-related strategies that might help older adults effectively manage body weight because 0·7 and 2·4 % of Americans aged 60–69 and 70+ years, respectively, are underweight (BMI < 18·5 kg/m2) and 75·5 and 65·8 % of older persons aged 60–69 and 70+ years are overweight and obese (BMI ≥ 25·0 kg/m2)(Reference Flegal, Graubard and Williamson5). Furthermore, sarcopenia and obesity cost the USA approximately 18·5(Reference Janssen, Shepard and Katzmarzyk6) and 110·5 billion dollars a year(Reference Tsai, Williamson and Glick7), respectively.
Older adults expend less energy than younger adults due to sarcopenia and lower levels of physical activity. One common treatment to combat sarcopenia is resistance training (RT), which increases muscle strength, muscle mass and resting energy expenditure(8, Reference Villareal, Apovian and Kushner9). Limited research in young men has suggested that acute resistance exercise may reduce hunger and ghrelin concentration(Reference Broom, Batterham and King10), but the impact of RT on fasting and postprandial appetite and related hormones has been undocumented in older adults.
RT may have an impact on the dietary response of older adults to nutritional supplementation. When sedentary (Sed), frail, elderly men and women consumed a nutritional supplement in beverage form, they compensated for this energy intake by reducing their habitual food intake(Reference Fiatarone, O'Neill and Ryan11). Alternately, when RT individuals consumed the dietary supplement, energy compensation was reduced, leading to increased energy intake. These findings suggest that RT may alter ingestive behaviour in older adults.
Food form is also known to influence energy regulation(Reference Bray, Nielsen and Popkin12, Reference Wharton and Hampl13). Specifically, beverages elicit reduced satiety compared with solid foods in some(Reference Tieken, Leidy and Stull14–Reference DiMeglio and Mattes16) but not all studies(Reference Almiron-Roig, Flores and Drewnowski17). Previous research(Reference Stull, Apolzan and Thalacker-Mercer18) from our laboratory examined appetitive sensations and energy intake following consumption of isoenergetic beverage v. solid foods in older adults (age range 50–80 years)(Reference Stull, Apolzan and Thalacker-Mercer18). Beverage meal replacement products resulted in greater postprandial hunger and a 13·4 % higher energy intake at the next eating occasion compared with isoenergetic solid meal replacement products(Reference Stull, Apolzan and Thalacker-Mercer18). Knowledge regarding the mechanisms explaining the differential food form responses is limited, but alterations of postprandial hormone concentrations (i.e. insulin, ghrelin and cholecystokinin (CCK)) and energy expenditure responses have been posited(Reference Tieken, Leidy and Stull14, Reference Santangelo, Peracchi and Conte19, Reference Peracchi, Santangelo and Conte20). In the present study, we critically evaluated the acute effects of isoenergetic- and macronutrient-matched beverage and solid supplements on postprandial appetite sensations, endocrine responses, energy expenditure and satiation in Sed v. RT older adults. We hypothesised the postprandial appetite responses (decreased hunger, desire to eat and increased fullness) and endocrine responses (increased glucose, insulin, CCK and decreased ghrelin) would be greater following the solid food form and these differential responses would be enhanced with RT.
Methods
Screening and subjects
Potential participants responded to newspaper advertisements and flyers recruiting RT and Sed individuals. A phone interview was conducted to estimate physical activity patterns and weight stability. Inclusion criteria for all potential subjects were the following: age ≥ 60 years; BMI 20–29 kg/m2; < 2 kg weight change during the previous 6 months; consistent physical activity patterns during the previous 6 months; consume breakfast and lunch; non-smoking; clinically normal blood profile; clinically normal heart function based on resting electrocardiogram; no osteoporosis based on self-report; fasting plasma glucose ≤ 1100 mg/l; no diabetes mellitus; not taking medications known to influence appetite or metabolism; acceptability of test foods. Further inclusion criteria were to be included in the Sed group: participants could not have engaged in RT in the previous 6 months. The RT group was required to have engaged in RT ≥ 2 times/week during the previous 6 months.
Initially, nineteen Sed and seventeen RT subjects were in the study, and eighteen (nine males and nine females) and sixteen (seven males and nine females) completed the study, respectively. The present study was conducted according to the guidelines laid down in the Declaration of Helsinki, and all participants were given verbal and written explanations about the study, provided signed informed consent and received a monetary stipend. The study was approved by the Purdue University Biomedical Institutional Review Board and registered at www.clinicaltrials.gov (NCT00798668).
Experimental design and protocol
A randomised, mixed-model, cross-over design study, incorporating food form (solid v. beverage) as a within-subject factor and Sed v. RT as a between-subject factor, was performed. Each subject participated in 5 d of testing. The first day of testing (baseline testing) was used to assess subject characteristics (Table 1). The randomised second and third days were used to assess the effects of food form on most of the study's dependent variables (hunger, fullness, desire to eat, glucose, insulin, ghrelin, CCK, glucagon-like peptide-1 (GLP-1), energy expenditure and RER). These days were separated by 48 h. On these days, participants came to the research laboratory after a 12 h overnight fast. A venous catheter was appropriately placed, and appetite sensations, endocrine responses and energy expenditure were measured at specified times (Fig. 1). During the second week of testing, again volunteers came in on two different days separated by at least 48 h after a 12 h overnight fast. These two randomised days were used to assess the effects of food form on satiation.
(Mean values with their standard errors)
RT, resistance training; VAI, vigorous activity index; FFM, fat-free mass.
Mean values were significantly different between the groups: *P < 0·05, **P ≤ 0·001.
† One-way ANOVA comparing sedentary v. RT.
‡ Estimated from the Yale Physical Activity Questionnaire.
§ Measured from the Caltrac™ Activity Monitor (Body Flex X-Max).
∥ Maximum strength: sum of one-repetition maximum-seated row, seated chest press, leg extension, leg curl and leg press exercises divided by kg of FFM.
Timeline of nutritional supplement tests. Percentage of total energy intake is basal energy need (computed using the Harris–Benedict equation) multiplied by an activity factor of 1·5. A, appetite assessment (hunger, fullness and desire to eat); B, blood draw (glucose, insulin, ghrelin, cholecystokinin, glucagon-like peptide-1); REE, fasting resting energy expenditure; PPEE, postprandial REE (thermic effect of feeding).
Baseline testing
Height was measured to ± 0·1 cm using a wall-mounted stadiometer (Holtain Limited, Crymych, Wales, UK). Body weight and body composition were measured by air displacement plethysmography (Bod Pod; Life Measurement, Inc., Concord, CA, USA)(Reference McCrory, Gomez and Bernauer21). Fat mass and fat-free mass were estimated from body density using the two-compartment Siri equation(Reference Siri22). BMI was calculated as weight divided by height squared (kg/m2). Maximum strength (one-repetition maximum) was assessed on five pieces of resistance exercise equipment (Keiser Sports Health Equipment Company, Fresno, CA, USA). Lower body (leg extension, seated leg curl and leg press), upper body (upper back (seated row) and seated chest press) and total strength were computed to be the sum of these maximal strength values and are reported as total kg lifted divided by kg of fat-free mass.
The Yale Physical Activity Questionnaire was used to estimate hours of habitual physical activity and a vigorous activity index(Reference Dipietro, Caspersen and Ostfeld23). The vigorous activity index was determined by multiplying a frequency score (not at all, 0; 1–3 times/month, 1; 1–2 times/week, 2; 3–4 times/week, 3; 5+ times/week, 4) by a duration score (not applicable, 0; 10–30 min, 1; 31–60 min, 2; 60+ min, 3) and multiplying again by a weighting factor (vigorous, 5; leisurely, 4; moving, 3; standing, 2; sitting, 1)(Reference Dipietro, Caspersen and Ostfeld23). On 3 d (two weekdays and one weekend day), each subject's energy expenditure as physical activity (kJ/d) was assessed using a Caltrac™ Activity Monitor (Body Flex X-Max, Van Nuys, CA, USA) worn during waking hours(Reference Melanson and Freedson24).
Also during baseline testing, each subject completed a taste test of the nutritional supplements, rating the palatability (pleasantness) of the solid and beverage using a scale from 1 to 9 (1, extremely unpleasant; 9, extremely pleasant).
Nutritional supplement feeding response tests
Each participant's total energy need was calculated to equal 1·5 times their estimated resting energy expenditure(Reference Shetty, Henry and Black25), which was determined using the sex-specific Harris–Benedict equations(Reference Harris and Benedict26). Previously, RT has been shown not to increase daily energy requirements(Reference Campbell, Kruskall and Evans27) compared with Sed older individuals. On study days 2, 3, 4 and 5, each subject consumed test supplements that contained 12·5 % of their total energy need (1·08 (se 0·03); 0·84–1·42 MJ) in either solid (hardness 1012 g, Texture Analyzer (TA-TX2; Texture Technologies Corporation, Scarsdale, NY, USA) or beverage form (viscosity 21.5 cP s, Brookfield Rheometer (RVDV); Brookfield Corporation, Middleboro, MA, USA) with approximately 89 ml (3 oz) of water for each treatment. After the beverage was consumed, the participants were instructed to rinse the bottle with approximately 89 ml of water and to consume the rinse. By design, the non-commercially available test supplements contained comparable energy and macronutrients (Table 2). A baseline (fasting) blood sample was taken (Table 3) and an appetite questionnaire was completed. At time 0, the subjects began to consume the test supplement simultaneous with the blood draw. The participants were given 15 min to consume each test supplement.
(Mean values with their standard errors for the sedentary and resistance trained groups combined, n 34)
(Mean values with their standard errors)
CCK, cholecystokinin; GLP-1, glucagon-like peptide-1.
* Mean values were significantly different between the groups (P < 0·05).
† Nine males and nine females.
‡ Seven males and nine females.
§ The lower anchor for the 13-point category scale was ‘not at all’ (1) and the upper anchor was ‘extremely’ (13).
∥ For statistics, see the Statistical analyses section.
Appetite
At the time points corresponding with each blood draw (Fig. 1), the following appetite-related questions(Reference Rogers and Blundell28) were asked; ‘How strong is your: feeling of hunger; feeling of fullness; desire to eat?’ Each response was recorded on a paper that had a 13-point category scale for each question(Reference DiMeglio and Mattes16). The subject circled the dash that corresponded with their perception at the moment. The lower anchor(Reference Roberts and Rosenberg1) was ‘not at all’ and the upper anchor(Reference Wharton and Hampl13) was ‘extremely’.
Endocrine testing
During each testing period, eleven blood samples were taken (Fig. 1) and immediately placed into blood collection tubes containing potassium EDTA (Becton, Dickinson and Company, Franklin Lakes, NJ, USA). Tubes were kept on ice until they were centrifuged at 4°C for 15 min at 3000 g. Aliquots of plasma were stored at − 80°C until thawed for analyses. Plasma glucose concentration was measured by enzymatic colorimetry, using an oxidase method on a COBAS Integra 400 analyser (Roche Diagnostic Systems, Indianapolis, IN, USA). Plasma insulin concentration was measured by an electrochemiluminescence immunoassay method on the Elecsys 2010 analyser (Roche Diagnostic Systems). Total plasma ghrelin, CCK26–33 and GLP-17–36 were analysed through enzyme immunoassay techniques, following the manufacturer's standard protocol (Phoenix Pharmaceuticals, Inc., Burlingame, CA, USA). All samples were run in duplicate and each individual's samples were analysed on the same day within the same assay.
Metabolic testing
Indirect calorimetry was used to measure resting energy expenditure in the fasting and postprandial states (MedGraphics Cardiopulmonary Diagnostics Systems; MedGraphics Corporation, St Paul, MN, USA). Fasting-state energy expenditure was measured for 30 min before supplement consumption. Following the supplement, postprandial energy expenditure was periodically measured for three time intervals (+15 to +120 min; +150 to +180 min; +210 to +240 min). Non-protein energy expenditure was estimated using the Weir equation(Reference Weir29) and non-protein RER was calculated.
Food intake
On study days 4 and 5, volunteers were seated 60 min before consuming the test supplements. The protocol was designed to be similar to days 2 and 3. At − 50, − 40 and − 30 min, three baseline appetite questionnaires (see above) were given. Then while receiving the test supplement, the participants completed an appetite questionnaire. Questionnaires were also completed at 15, 30, 60, 90, 120, 150 and 180 min. At 120 min after consuming the supplement, the participants were presented with a bowl of hot oatmeal. They were asked to consume the oatmeal to a ‘comfortable level of fullness’. Each bowl contained rolled oats (120 g), 2 % reduced fat milk (75 g), brown sugar (24 g), salt (1 g) and water (550 g). This represented three commercial servings. The total amounts consumed (weighed to the nearest 0·1 g) were evaluated as an index of satiation.
Diversionary task
To minimise bias and not declare the true purpose of the study, several mental diversionary tasks were included in the protocol. They were performed after the appetite questionnaires were completed at − 50, 15, 90 and 150 min. At − 50 (practice) and 90 min, participants were asked to take eye–hand coordination and memory tests using an online computer game, Escapa(Reference Escapa30). At 90 min, they were given twelve optical illusions and asked to document what was observed first. Finally, at 150 min, participants had 45 s to circle as many of the letter ‘S’ as possible on two pages with a random combination of letters.
Statistical analyses
All values are reported as means with their standard errors. For subject characteristic data, differences between the Sed and RT groups were assessed using one-way ANOVA. After adjusting the postprandial responses for the corresponding fasting values (i.e. expressing the data as a change from baseline), area under the curve (AUC) was calculated using the trapezoidal method. After all preliminary calculations, if needed, based on the Shapiro–Wilk test, data were normalised using a log or square root transformation to approximate a normal distribution. Statistical evaluation of subject characteristic data revealed differences between groups for age and BMI (Table 1). Also, since the initiation of this project, a growing body of emerging research has suggested that sex influences appetitive and endocrine responses, especially insulin and ghrelin(Reference Carroll, Kaiser and Franks31–Reference Hagobian, Sharoff and Stephens35). Therefore, sex, age and BMI were included in the statistical analyses. Note that a priori hypotheses were not generated based on age, BMI and sex. Repeated measures with random subject effects were performed using the PROC MIXED model. Fixed effects included the following: food form; RT status; sex; time; age; BMI. Interactions were initially examined among food form, RT status and sex. Then, a backward approach was taken to determine the smallest (best) Akaike information criterion for the model. Least-square mean was used to determine treatment effects. The Tukey–Kramer test was used for multiple comparisons. The correlations between fat-free mass and ghrelin concentrations were established using the partial correlation coefficient controlling for RT status, sex and age, and performed according to the previous studies by Tai et al. (Reference Tai, Visvanathan and Hammond36) and Bertoli et al. (Reference Bertoli, Magni and Krogh37). All data are presented without statistical transformation, but the statistics are reported on the transformed data. The criterion for statistical significance was set at P < 0·05. Statistical evaluations were performed using SAS Statistical Discovery Software (version 9.1; SAS Institute, Cary, NC, USA).
Results
Day 1: subject characteristics/hedonics
Subjects ranged in age from 62 to 84 years. The RT group was younger than the Sed group (P < 0·01; Table 1) and had lower body weight (P < 0·05) and BMI (P < 0·05). Height, body fat percentage and fat-free mass were not different between the groups. Total body strength was higher in the RT v. Sed groups (P < 0·01). The RT group also had greater energy expenditure due to physical activity (P < 0·01) and a greater amount of vigorous physical activity (P < 0·01) compared with the Sed group. Training status of the subjects did not have an impact on palatability, but the solid supplement was more palatable than the beverage (solid 7 (se 0), beverage 6 (se 0); P < 0·05).
Days 2 and 3: food form and appetitive, endocrine and metabolic responses
Appetite
Baseline appetite values are shown in Table 3. The changes in appetite sensations during the 4 h period are shown in Fig. 2 and Table 4. No training status × food form interactions were observed for the appetitive responses. Postprandial fullness AUC was not different with regard to training status, but was higher in the solid v. beverage treatment (Fig. 2(a); P < 0·01). No differences in postprandial hunger or desire-to-eat AUC were observed with respect to training status or food form (Fig. 2(B) and 2(C)).
Appetitive sensations and plasma glucose and endocrine responses for the sedentary and resistance trained (RT) males and females after beverage and solid supplement consumption. Values are means for eighteen sedentary and sixteen RT subjects for appetitive sensations and eighteen sedentary and fifteen RT subjects for glucose and endocrine responses, with standard errors represented by vertical bars. (a) RT status did not influence postprandial fullness. Postprandial fullness was lower after the beverage v. solid was consumed (218 (se 94) v. 475 (se 96) arbitrary units (AU) × 240 min; P < 0·01). (b) No RT status or food form differences were seen with hunger. (c) No RT status or food form differences were seen with desire to eat. 
, Solid (trained); 
, beverage (trained); 
, solid (sedentary); 
, beverage (sedentary).
(Mean values with their standard errors)
AU, arbitrary units; DE, desire to eat; CCK, cholecystokinin; GLP-1, glucagon-like peptide-1.
Mean values were significantly different between the treatment groups: *P < 0·05, **P ≤ 0·01.
Mean values were significantly different between the training status groups (†P < 0·05).
‡ For the appetite sensations, eighteen sedentary (nine males and nine females) and sixteen resistance trained (seven males and nine females).
§ For the endocrine responses, eighteen sedentary (nine males and nine females) and fifteen resistance trained (seven males and nine females).
Endocrine testing
The fasting concentrations of glucose and hormones are shown in Table 3. The RT group had a higher fasting plasma CCK concentration (P < 0·05). The changes in glucose and endocrine responses during the 4 h period are shown in Fig. 3 (a)–(e). No training status × food form interactions were observed with postprandial glucose or endocrine responses. No training status differences were seen with postprandial glucose or insulin (Table 4). Postprandial glucose and insulin were higher following the solid v. beverage test supplement (P < 0·01 and < 0·01, respectively). No difference in postprandial ghrelin was observed with training status or food form. The CCK concentration over the 4 h period was elevated in the RT v. Sed individuals (P < 0·01) and higher following the solid v. beverage test supplement (P < 0·05). Postprandial GLP-1 AUC did not differ between the training status groups or food form stimuli.
Plasma glucose and endocrine responses for the sedentary and resistance trained (RT) males and females after beverage and solid supplement consumption. Values are means for eighteen sedentary and fifteen RT subjects, with standard errors represented by vertical bars. (a) RT status did not affect plasma glucose. Beverages decreased plasma glucose area under the curve (AUC) v. solids ( − 2660 (se 2400) v. 5030 (se 2920) mg/l × 240 min; P < 0·01). (b) No training status effects were observed with plasma insulin. Beverages decreased plasma insulin AUC v. solids (12 978 (se 1244) v. 19 522 (se 1897 pmol/l × 240 min; P ≤ 0·001). (c) No training status or food form differences were observed with ghrelin. (d) Training increased cholecystokinin (CCK) AUC v. sedentary (28·97 (se 12·70) v. 12·16 (se 7·73) ng/ml × 240 min; P < 0·01) and beverages decreased CCK AUC levels compared with solids (6·19 (se 8·91) v. 33·42 (se 8·86) ng/ml × 240 min; P < 0·05). (e) RT status and food form did not affect glucagon-like peptide-1 AUC. , Solid (trained); , beverage (trained); , solid (sedentary); , beverage (sedentary).
Metabolic testing
No training status × food form interactions were observed with postprandial energy expenditure or RER responses. Postprandial energy expenditure over the 4 h period was not affected by training status or food form (data not shown; RT 33·22 (se 27·61) kJ/min × 240 min (7·94 (se 6·60) kcal/min × 240 min) v. Sed 43·63 (se 26·56) kJ/min × 240 min (10·43 (se 6·35) kcal/min × 240 min); solid 34·60 (se 28·86) kJ/min × 240 min (8·27 (se 6·90) kcal/min × 240 min) v. beverage 46·19 (se 23·63) kJ/min × 240 min (11·04 (se 5·65) kcal/min × 240 min). Postprandial RER AUC was unaffected by training status, but lower following the solid (data not shown; 0·04 (se 0·01) × 240 min) v. beverage test supplement (0·05 (se 0·01) × 240 min; P < 0·001).
Fat-free mass and ghrelin correlations
Fat-free mass was not associated with fasting or postprandial ghrelin (fasting r − 0·070; postprandial r 0·004).
Days 4 and 5: food form and satiation
Appetite
No training status × food form interactions were observed for the appetitive responses. Postprandial fullness, hunger and desire to eat were unaffected by training status (data not shown; fullness, RT 276 (se 89) v. Sed 198 (se 49); hunger, RT − 277 (se 68) v. Sed − 202 (se 43); desire to eat, RT − 239 (se 71) v. Sed − 178 (se 51)) but fullness was higher in the solid v. beverage (302 (se 47) v. 167 (se 56), respectively; P < 0·01) and hunger and desire to eat were lower in the solid v. beverage treatments (hunger, solid − 260 (se 47) v. beverage − 186 (se 53); desire to eat, solid − 237 (se 44) v. beverage − 175 (se 47); P < 0·05).
Food intake
No training status × food form interactions were observed with food intake. Satiation was not affected by training status or food form (data not shown; RT 333·7 (se 26·3) v. Sed 303·9 (se 25·4) g; solid 329·8 (se 22·9) v. beverage 311·5 (se 20·4) g).
Discussion
The present study compared the appetitive, metabolic and endocrine responses between RT and Sed older individuals following the consumption of energy- and macronutrient-matched beverage and solid nutritional supplements. Contrary to our hypothesis, RT did not influence postprandial appetitive, metabolic or endocrine responses to food form (i.e. there was no RT status × food form interactions). However, the RT group had reduced fasting hunger and desire to eat and increased fasting CCK. The findings that the nutritional supplement in beverage form elicited lower fullness along with decreased glucose, insulin and CCK responses compared with the solid food form coincides with most(Reference Tieken, Leidy and Stull14–Reference DiMeglio and Mattes16, Reference Mourao, Bressan and Campbell38, Reference Mattes39), but not all(Reference Almiron-Roig, Flores and Drewnowski17) published research indicating that beverages elicit weaker satiety sensations than solid foods. However, in contrast to our hypothesis, the findings also suggest no differential food form effect over 4 h on hunger, desire to eat or GLP-1. Recently, Mourao et al. (Reference Mourao, Bressan and Campbell38) also suggested that solid food form resulted in lower postprandial feelings of fullness compared with beverages but did not affect feelings of hunger. The findings that food form affected fullness but not hunger are plausible since these are different dimensions of appetite (i.e. previously CCK has been shown to reduce meal size, which demonstrates satiation but not satiety). Overall, the weaker dietary compensation previously observed with beverages(Reference DiMeglio and Mattes16, Reference Stull, Apolzan and Thalacker-Mercer18, Reference Mourao, Bressan and Campbell38) may be attributable to the differential glucose and endocrine responses they elicited, but in these older individuals, no food intake differences were observed. We consider the provision of the nutritional supplement, 12·5 % of each subject's estimated energy need and not matched for volume, as practically important because this amount corresponds with approximately a single serving of supplement: one bar (solid) and one drink (beverage).
Our finding of no differential response for postprandial energy expenditure in RT v. Sed older adults is consistent with findings in younger adults(Reference Gilbert and Misner40). The present finding that food form did not influence postprandial energy expenditure contrasts with the report that postprandial energy expenditure was approximately 85 % higher when young men consumed a 2·6 MJ (approximately 615 kcal) meal as whole foods (solid–liquid), compared with when the same food items were homogenised with water and consumed as a viscous suspension(Reference Peracchi, Santangelo and Conte20). Both of these studies contrast with the observation that postprandial energy expenditure was 54 % higher after eight healthy, normal-weight young men consumed a 2·1 MJ (approximately 500 kcal) solid meal compared with a liquid meal of similar macronutrient composition(Reference Habas and Macdonald41). Furthermore, the interpretation of these disparate results is complicated by multifarious factors, including energy content and macronutrient distribution of the test meals; glycaemic and insulinaemic responses; gastrointestinal transit time and absorption; and subjects' sex and age(Reference Roberts and Rosenberg1, Reference Granata and Brandon42–Reference Tataranni, Larson and Snitker46). Peracchi et al. (Reference Peracchi, Santangelo and Conte20) did not evaluate RER. In the present study, RER was lower after solid consumption, suggesting higher fat oxidation, whereas beverage consumption resulted in higher carbohydrate oxidation and thus lower fat oxidation.
The higher postprandial glucose, insulin and CCK AUC following the solid v. beverage nutritional supplement might be the result of increased (slower) gastric transit time. Specifically, solid foods appear to elicit a slower gastrointestinal transit time than beverages(Reference v Schonfeld, Evans and Goebell47–Reference Hinder and Kelly49), which may result in a different absorption profile. Within physiological ranges, higher CCK and GLP-1 inhibit gastric emptying(Reference Little, Horowitz and Feinle-Bisset50, Reference Holst51). Our data follow this pattern since solids elicited a greater CCK response than beverages. GLP-1 secretion did not vary between solid and beverage test supplements. Previous research provided mixed results regarding gastrointestinal transit time and ageing(Reference Horowitz, Maddern and Chatterton52–Reference Kupfer, Heppell and Haggith55) and suggested that, in younger individuals, beverages have faster gastric emptying time(Reference Notivol, Carrio and Cano56, Reference Moore, Christian and Coleman57) and orocaecal transit time than solids(Reference Bennink, Peeters and Van den Maegdenbergh58).
RT increases fat-free mass due to muscle hypertrophy in older adults(Reference Frontera, Meredith and O'Reilly59), and fat-free mass and appendicular muscle mass were shown to be positively correlated with fasting ghrelin in healthy younger, older and elderly subjects(Reference Tai, Visvanathan and Hammond36, Reference Bertoli, Magni and Krogh37). In contrast to these findings, our RT and Sed subjects did not exhibit a significant correlation between fat-free mass and ghrelin. Tai et al. (Reference Tai, Visvanathan and Hammond36) studied adults of all ages, 22–82 years, while the present study and Bertoli et al. (Reference Bertoli, Magni and Krogh37) studied older adults. The 60-year age range by Tai et al. may have resulted in a broader distribution of data contributing to a significant correlation. Differences in the measurement of fat-free mass might also have contributed to the apparently disparate findings among studies. Tai and Bertoli et al. measured appendicular muscle mass using dual-energy X-ray absorptiometry(Reference Tai, Visvanathan and Hammond36, Reference Bertoli, Magni and Krogh37), while the present study measured whole-body fat-free mass using plethysmography. Appendicular muscle mass is a strong predictor of fasting ghrelin concentration(Reference Tai, Visvanathan and Hammond36, Reference Bertoli, Magni and Krogh37). This parameter was not measured in the present study.
One of the strengths of the present study was the recruitment of older people who habitually performed RT. Currently, approximately 11 % of adults aged 65 years and older in the USA(60) engage in this mode of exercise because it is encouraged for older people to retain and enhance muscle mass, strength, physical function and health indices associated with the metabolic syndrome. The 1·7 kg/m2 BMI difference between the RT and Sed groups may be considered a weakness, but this subtle difference was accounted for statistically. The appetite ratings appear relatively low for the fasted state, which may be due to the use of the equal interval appetite scale (compared with the labelled magnitude scale)(Reference Green, Dalton and Cowart61). Also, appetite questionnaires were examined across groups (between subject) for RT. It is not possible to determine whether all subjects had similar responses at the various intensities(Reference Bartoshuk, Duffy and Hayes62, Reference Bartoshuk, Duffy and Green63) so caution is warranted when interpreting these results. Lastly, not matching the beverage and solid nutritional supplements for weight or volume might also be considered a weakness. However, we chose to administer the products comparably with how they are consumed commercially. It is perhaps important to note that while the volume of the beverage was greater than the solid, the appetitive responses were consistent with lower satiety.
Conclusions
Findings from the present study suggest that RT and food form independently, but not synergistically, affect appetitive, metabolic or endocrine responses in older adults. However, the RT effect was limited to fasting and postprandial CCK. None of the results suggest that the beverage was more satiating than the solid. Some, but not all, endocrine responses found the solid to be more satiating than the beverage. The differential glucose, insulin, ghrelin and CCK responses between supplement treatments implicate food form as an important factor influencing energy homeostasis and indicate that energy- and macronutrient-matched nutritional supplements in the solid v. beverage form are not equivalent. Although the beverage supplement altered appetitive, endocrine and metabolic responses, the beverage food form did not alter subsequent food intake in this acute laboratory setting. RT and food form should be considered when recommending a weight management strategy to older adults, although they may not affect dietary energy intake.
Acknowledgements
The authors have no conflict of interest. The authors thank the volunteers for their time and dedication: Arthur Rosen, MD, for providing medical coverage, Douglas Maish, EMT-P, for inserting the catheters and providing other clinical services, Kristin Duke, BS, for assisting with clinical services, and Eunjung Lim, MS, for consulting on statistical analyses. W. W. C. and R. D. M. originated the study, obtained funding and provided supervision of the research. J. W. A. and H. J. L. were involved with the development of the study purpose and design. J. W. A. performed the clinical testing and data analysis. J. W. A. drafted the manuscript and all co-authors contributed to the data interpretation and finalisation of the manuscript. The study was supported by a NIH R01 AG021911 and the study supplements were provided by The Solae Company. Support for H. J. L. was provided in part by an Ingestive Behavior Research Center post-doctoral fellowship from Purdue University, and support for J. W. A. was provided in part by a Purdue University Bilsland Fellowship.
