Gastrointestinal peptides, gastrointestinal motility, and anorexia of aging in frail elderly persons



This article is corrected by:

  1. Errata: Corrigendum Volume 28, Issue 1, 158, Article first published online: 21 December 2015

Address for Correspondence
Mateu Serra-Prat, Research Unit, Hospital de Mataró, Carretera de Cirera s/n, E-08304 Mataró, Barcelona, Spain.
Tel: +34 93 741 77 30; fax: +34 93 757 33 21;


Background  The mechanisms involved in anorexia in frail elderly people remain unclear. The objective of this study was to establish whether fasting and postprandial levels of gastrointestinal peptides, gastrointestinal motility, and hunger are modified by age and frailty.

Methods  Three groups of subjects were studied: (a) frail elderly (>70 years) persons, (b) non-frail elderly (>70 years) persons, and (c) healthy adults (aged 25–65 years). After an overnight fast, participants ingested a 400 Kcal liquid meal and appetite, hormonal, and gastrointestinal responses were monitored during early (0–60 min) and late (60–240 min) postprandial periods.

Key Results  Frail persons showed poor nutritional status, sarcopenia, and almost absence of hunger during fasting and postprandial periods. Older persons presented higher levels of glucose and insulin during fasting, enhanced postprandial CCK release in early postprandial period and postprandial hyperglycemia and hyperinsulinemia, but similar ghrelin levels than younger adults. Ultrasound scan showed that the fasting antral area was higher and antral compliance lower in old persons. The paracetamol absorption test showed enhanced postprandial gastric emptying in the frail. Non-gallbladder contractors showed no CCK peak in younger and non-frail groups, but the same high CCK peak as contractors in the frail.

Conclusions & Inferences  Frailty was associated with anorexia, risk of malnutrition, and sarcopenia. Frail persons showed impaired gastric motility (larger antral area at rest, impaired antral compliance, and enhanced postprandial emptying), impaired gallbladder motility, and fasting and/or postprandial alterations in CCK, glucose, and insulin release. Further studies are needed to determine if these factors may contribute to anorexia of aging in frail persons.


area under the curve




glucagon-like peptide-1


Mini-Nutritional Assessment


orocecal transit time


peptide YY




visual analog scale


Malnutrition is a highly prevalent clinical condition in older populations and has a major impact on morbidity and mortality.1–3 Causes of malnutrition in elderly persons are multiple and, among them, alterations in the gastrointestinal tract’s control of appetite are thought to play an important role. Changes in gastrointestinal motility and hormone release form the basis of anorexia of aging,1,4 an independent predictor of mortality in elderly patients.5

Although the mechanisms that regulate hunger and satiety are not well established, it is known that appetite is regulated at a central level in the hypothalamus.6,7 This central regulation is influenced by peripheral signals such as gut-derived gastrointestinal peptides including ghrelin, cholecystokinin (CCK), glucagon-like peptide-1 (GLP-1), or peptide YY (PYY); and blood levels of leptin, glucose, fatty acids, or insulin.4,8,9 Food ingestion releases these hormones and activates gastrointestinal motility with powerful effects on gastrointestinal physiology, in addition to the central function on appetite. Ghrelin is an orexigenic hormone in the control of food intake that enhances gastric emptying.4,7 Cholecystokinin is a major mediator of satiation at the central level and also delays gastric emptying. Some studies have found higher concentrations of CKK in the elderly than in younger persons.9,10 Recent studies suggest that two peptides released by the distal small intestine, GLP-1 and PYY, play a physiological role in the mediation of suppression of appetite by providing negative feedback to the stomach.7 Insulin has also been described as a satiety hormone4 and Leptin, secreted from the adipose tissue, has also shown an anorexigenic effect. Increased satiety hormones and reduced ghrelin release might be major causes of anorexia in the frail.11 On the motility side, knowledge of the aging GI tract and the mechanisms of impaired gastrointestinal motility in the elderly is also incomplete.12 Impaired relaxation of proximal stomach in the elderly might also cause accelerated antral filling and contribute to early satiation.13 Other studies found slow gastric emptying of solid foods, impaired gallbladder contraction, and delay in intestinal transit in older populations.14

Among the elderly, frailty is a specific syndrome of functional decline of multiple systems conferring high risk for falls, disability, hospitalization, and death.15 Nutritional status may be especially affected in the frail, but very little is known on the regulation of hunger and the changes in fasting and postprandial plasma levels of gut hormones or the patterns of gastrointestinal motility in frail elderly patients. We hypothesize that the physiological mechanisms of gastrointestinal hormone release and gastrointestinal motility that regulate appetite can be altered in older populations, especially frail ones, and might be responsible for anorexia which contributes to malnutrition.4,11,16 The aim of this study was to determine whether fasting and postprandial levels of hunger, orexigenic (ghrelin), and anorexigenic (CCK, insulin and GLP-1) gut peptides, fasting leptin and gastric, gallbladder, and intestinal postprandial motility responses are modified by age and frailty.

Subjects and methods


Three study groups were assessed and compared: (i) frail elderly (>70 years) persons (group A), (ii) non-frail elderly (>70 years) persons (group B), and (iii) non-obese [body mass index (BMI) <30 kg m−2] younger adults (25–65 years of age) (group C). According to Fried criteria, a person was considered frail if he/she had three or more of the following five conditions: (i) non-intentional weight loss (>4 kg or 5% of usual weight) or BMI < 19 kg m−2, (ii) self-reported exhaustion [usual energy <3 in a 0–10 cm visual analog scale (VAS)], (iii) poor muscle strength (hand grip measured by a hand held dynamometer <20th percentile: <7 kg in women and <14 kg in men), (iv) slow walking speed (≥7 s to walk 4.5 m), and (v) poor physical activity (no outdoor life or <0.5 h of outdoor walking daily).15 Persons older than 70 years were recruited from the Maresme Health Consortium medical centers (Barcelona, Spain), and younger healthy adults were volunteers recruited among hospital workers. Exclusion criteria for any study group were the following: dementia or severe psychiatric disorders, total or partial gastrectomy, cholecystectomy, active neoplasm, anemia (serum hemoglobin level<12 g dL−1), and severe dysphagia. The study protocol was approved by the institutional review board of the Consorci Sanitari del Maresme, Mataró, Barcelona (Spain), and all participants signed an informed consent form before inclusion.

Experimental design

After a 10 h overnight fast, all participants were given a 200 mL standard liquid preparation meal of 400 Kcal (T diet 20/2, Vegenat, Badajoz, Spain) containing 20.2 g proteins, 15.6 g fat, and 43.4 g carbohydrates. As in other published studies, the standard meal test was not adjusted by body weight because gastrointestinal peptide response depends on the total meal amount arriving in the gut. Together with this preparation, 1.5 g of paracetamol (Laboratorios Gelos, Barcelona, Spain) was given, as well as 10 g of lactulose (Duphalac, Abbot Laboratories, Madrid, Spain). A nurse ensured the complete ingestion of the preparation in less than 5 min. Just before meal intake (time 0) and at 15, 30, 45, 60, 90, 120, 180, and 240 min after meal ingestion, a 15 mL blood sample was obtained from a peripheral vein, and, at the same intervals, hunger was assessed, and gastric and gallbladder motility were studied by an abdominal ultrasound. Expiratory breath test samples were obtained at 15 min intervals since a positive result to assess orocecal transit time. Eating was not allowed during the 4-h postprandial period. Main study parameters were determined at all time intervals: (i) hunger perception by means of a 0–10 cm visual analog scale (VAS) ranging from 0 (no hunger at all) to 10 (maximum hunger),11 (ii) serum concentrations of ghrelin, CCK, GLP-1, insulin, glucose, and paracetamol, (iii) gallbladder volume, (iv) antral gastric area, and (v) expiratory H+ concentration. Fasting leptin was also determined.


All hormone measurements were determined using validated, commercialized kits. Total plasma ghrelin concentrations were measured with a human radioimmunoassay (RIA) kit (Linco Research Inc, St Charles, MO, USA). Plasma cholecystokinin (cholecystokinin 26–33) concentrations were measured with a validated commercial human RIA kit (Euro-Diagnostica, Malmö, Sweden). Active GLP-1 plasma levels [GLP-1 (7–36) amida] were determined by RIA (Millipore Corporation, Billerica, MA, USA). Plasma insulin levels were measured by chemiluminescence (Immulite 2000 DPC; Siemens Medical Solutions Diagnostics, Cornella, Spain) using sheep and mouse anti-insulin antibodies. Plasma leptin levels were measured by enzyme-linked immunosorbent assay method (DBC, Diagnostics Biochem Canada Inc., Dorchester, Ontario, Canada). Serum glucose was measured with the commercially available test kit Gluco-quant enzymatic Hexokinase (Roche Diagnostics GmbH, Mannheim, Germany). Paracetamol was determined using an enzymatic method and colorimetric determinations (Cobas Integra 400 Plus; Roche Diagnostics, Sant Cugat, Spain).

Gallbladder and gastric motility were assessed by repeated ultrasound scans (Hitachi EUB-6500 Digital Ultrasound Scanner; Hitachi Medical Co, Tokyo, Japan). The length, width, and height of the gallbladder were measured in each scan and its volume was estimated according to a validated method that considers the gallbladder to be an ellipsoid cylinder.17 Gastric compliance and emptying were assessed by two methods: (i) by monitoring antral area by ultrasound scanning; and (ii) by the paracetamol absorption test. Ultrasound has been widely used to asses liquid gastric emptying18,19 as it correlates well with scintigraphy20 and is a simple, non-invasive, and well-tolerated method useful in the geriatric population. To measure the antral area,21 the transducer was positioned to obtain a sagittal image of the antrum with the left lobe of the liver, the superior mesenteric vein and the aorta as landmarks.22,23 The first measurement was performed during fasting and the second was performed 5 min after onset of meal ingestion and was followed by images at the study intervals. To avoid inter-individual variability in antral size, postprandial areas were normalized to those at basal time (with a value of 100). Ultrasound T50 was defined as the time when antral area decreased to half of its maximum.22 The paracetamol absorption test is also a widely used technique based on the assumption that the paracetamol (acetaminophen) absorption rate represents the gastric emptying rate because it is rapidly absorbed from the small intestine, with little absorption from the stomach.24 This test has correlated well with scintigraphy when used with liquid meal tests.25 Previous studies have also correlated well between the half-time of gastric emptying of liquids and serum paracetamol concentrations at 30 (C30) and 60 (C60) min. The areas under the curve (AUC) are also considered to be a useful index for the rate of drug absorption. Finally, hydrogen breath test was performed at 15 min intervals (Gastrolyzer; Bedfont Scientific Ltd, Harrietsam, UK). Orocecal transit time (OCTT) reflecting integrated gastric and small bowel transit was defined as the time between ingestion and sustained increase (>5 ppm) in expiratory H+ concentration.26 High H+ expiratory values before 30 min were considered bacterial overgrowth. In those cases, breath test were continued at usual intervals and was considered positive when a second peak (increase >5 ppm) was observed.

Other study variables included socio-demographic characteristics, comorbidities, functional capacity assessed by the Barthel score, and quality of life assessed by the Euro-QoL-5D™ as a measure of health outcome. Nutritional assessment included anthropometric measurements and the short form of Mini-Nutritional Assessment (MNA), a questionnaire validated in the geriatric population to screen for malnutrition.27 A bioelectrical impedance analysis to assess body composition and basal metabolism was also performed (Bioimpedance Analyzer Model BIA101; Akern Srl, Pontassieve, Florence, Italy); and hand grip of the non-dominant hand was measured by a hand held dynamometer (Jamar Hand Dynamometer; Lafayette Instrument Co, Lafayette, IN, USA). To complete the nutritional study, some laboratory parameters such as hemoglobin (g dL−1), creatinine (mg dL−1), albumin (g L−1), cholesterol (mg dL−1), and C-reactive protein (mg dL−1) were collected.

Statistical analysis

Sample size was estimated considering paracetamol C60 as main outcome measure. Accepting an alpha risk of 0.05 and a beta risk of 0.2 in a two-sided test, 16 subjects are necessary in each group to recognize as statistically significant a difference greater than or equal to 5 μg mL−1 in paracetamol C60 between groups (a common standard deviation was assumed to be 5 μg mL−1).

For the purpose of analysis, three main periods were established: (i) fasting period, (ii) early postprandial phase (0–60 min), and (iii) late postprandial phase (60–240 min). The AUC during the two postprandial phases was calculated for all study variables. Area under the curve for hunger scores were calculated with respect to the zero value indicating total amount of hunger during this period. Incremental areas under or over the baseline value were calculated for hormone and motility response, indicating a change in relation to base value. For the paracetamol test, the following were considered: increase in paracetamol concentration at 30 min (C30), and AUC at 30 and 60 min. For ultrasound measurements of gastric antrum, absolute and normalized areas were used as well as T50. To explore the feedback loop of CCK release, persons were defined as gallbladder contractors if they reached a postprandial gallbladder volume less than 50% of its fasting volume.28 The OCTT was estimated considering only persons with a positive breath test during the 4-h study period. To assess the ileal brake, postprandial GLP-1 response was analyzed in those persons with a positive measurement of OCTT, to be sure of the arrival of nutritional load to the distal ileum within the study period.

An initial descriptive analysis of the main characteristics of the study sample was carried out using mean and SD for continuous variables and percentages for categorical ones. A description of the hunger pattern, gastrointestinal peptides, and gastric and gallbladder responses according to specified parameters was given for all groups. To assess the effect of age on these parameters, data from Groups A and B (elderly groups) were compared with Group C (younger group). To assess the effect of frailty, Group A was compared with Group B. The chi-square or the Fisher’s exact tests were used to compare proportions between groups. As most variables were not normally distributed and because of the small sample size, non-parametric tests were applied. The Mann–Whitney U test or the Kruskal–Wallis test was used to compare means between two or three groups, respectively. To compare the curves of repeated measures between study groups, the general lineal model (GLM) was applied. A P < 0.05 was considered statistically significant.


Demographics, health, and nutritional status

Fifty-three persons were recruited and distributed as follows: 14 in Group A (frail elderly), 20 in Group B (non-frail elderly), and 19 in Group C (younger adults). Table 1 shows main clinical and demographic characteristics of all groups. No elderly subject in the study showed cognitive impairment. No significant differences in main comorbidities were found between frail and non-frail persons, but significant differences were observed in functional capacity and nutritional status between these two groups of elderly persons. Frail elderly persons presented poorer muscle mass, strength, basal metabolism, functional capacity, nutritional status, and overall health status/quality of life than non-frail elderly persons. No subject in the study was a nursing home resident. On the other hand, the prevalence of heart disease, stroke, chronic bronchitis/chronic obstructive pulmonary disease (COPD), and diabetes was higher in the elderly groups as well as the use of benzodiazepines and omeprazole compared with the group of younger adults. Significant differences were also observed in the number of medications between young (0) and old (5.5) groups, but not between frail (5.4) and non-frail (5.5) groups. Regarding Ca-antagonists, 21.4% of frail subjects, 10% of non-frail subjects, and 0% of healthy young volunteers were taking this type of medication with no significant differences between groups. No other differences were observed in comorbidities and usual treatments between young adults and elderly groups.

Table 1. Main characteristics of the study groups
 Group A Frail Elderly = 14Group B Non-frail, Elderly = 20Group C Younger Adults = 19p A vs Bp A+B vs C
Age (years)84.5 ± 5.080.7 ± 8.438.4 ± 11.00.145<0.001
Gender (% women)11 (78.6)7 (35.0)12 (63.2)0.0120.472
Physical examination
 Weight (kg)
 Men67.5 ± 12.674.9 ± 13.579.0 ± 10.40.6420.284
 Women54.3 ± 10.868.3 ± 20.560.4 ± 8.20.0790.497
 Height (m)
 Men1.63 ± 0.0361.63 ± 0.0461.75 ± 0.0410.951<0.001
 Women1.51 ± 0.0351.54 ± 0.0681.63 ± 0.0770.510<0.001
 Men25.3 ± 4.027.8 ± 4.125.6 ± 3.40.4590.285
 Women23.8 ± 4.728.3 ± 6.322.6 ± 2.40.1600.176
 Walking speed8.7 ± 1.56.5 ± 2.5)4.6 ± 0.90.030<0.001
 Barthel score63.2 ± 17.396.0 ± 8.8100 ± 0<0.001<0.001
 Outdoor life (%)5 (35.7)18 (90.0)19 (100)0.0020.004
 Hand grip
 Men9.7 ± 9.112.4 ± 3.731.9 ± 8.30.416<0.001
 Women3.8 ± 4.07.1 ± 3.020.6 ± 6.80.032<0.001
 Nutritional status (MNA)
 Well-nourished (%)0 (0)17 (85.0)18 (94.7<0.01 
 At risk of malnutrition (%)14 (100)3 (15.0)1 (5.3)<0.001<0.05
 Euro-QoL (VAS)44.3 ± 24.766.0 ± 15.489.2 ± 7.50.007<0.001
Body composition
 % Fat mass36.1 ± 6.435.7 ± 7.226.2 ± 7.20.847<0.001
 % Muscular mass33.7 ± 4.837.2 ± 4.450.9 ± 6.80.044<0.001
 Basal metabolism (Kcal)1177.9 ± 131.91298.2 ± 310.01559.0 ± 221.10.006<0.001
 Creatinine1.0 ± 0.41.0 ± 0.40.9 ± 0.20.5060.133
 Cholesterol191.7 ± 39.5175.8 ± 35.6192.5 ± 23.80.2700.179
 Albumin4.0 ± 0.54.1 ± 0.54.5 ± 0.30.738<0.001
 C-reactive protein1.0 ± 1.01.9 ± 3.30.2 ± 0.31.000<0.001

Hunger scores

Fasting hunger scored 2.1 ± 2.2 in frail elderly persons, 4.6 ± 3.5 in non-frail elderly, and 5.3 ± 3.4 in younger adults. Frail elderly persons showed significantly less fasting hunger when compared with the other two study groups, but no significant differences between non-frail elderly and younger adults were found. Figure 1 shows the VAS curve for hunger in the three study groups. Younger adults experienced higher hunger recuperation than older groups and frail elderly showed poorer hunger recuperation in comparison to non-frail elderly persons. Significant differences were observed in the AUC60-240 min (hunger recuperation phase) between frail and non-frail (P = 0.032), and younger adults and elderly persons (P = 0.007) as well as in the overall AUC0-240min of hunger with P = 0.049 and P = 0.014, respectively. These results showed that frail elderly persons presented anorexia both in fasting and late postprandial periods.

Figure 1.

 Mean (SD) visual analog scale (VAS) hunger scores profile during fasting and 4 h after a 400 kcal meal test in the study groups (left). Integrated AUC for hunger score during the early (0–60 min) and late (60–240 min) postprandial periods (right). *< 0.05. AUC, area under the curve (in comparison with zero value).

Fasting leptin

Fasting leptin levels were some higher in elderly subjects in comparison to young adults (31.3 ± 28.3 ng mL−1vs 17.7 ± 16.4 ng mL−1; P = 0.068), but this difference completely disappeared when adjusting leptin levels for BMI in a multiple lineal regression model (P < 0.001 for BMI and P = 0.489 for age group). No significant differences in fasting leptin were observed between frail and non-frail subjects (24.6 ± 28.1 ng mL−1vs 36.3 ± 28.2 ng mL−1; P = 0.101), neither in the bivariate analysis nor when adjusting for BMI. Fasting leptin showed no significant correlation with fasting hunger (VAS) in any group, but showed a positive correlation in the frail and non-frail groups with BMI (rs = 0.745; = 0.002, and rs = 0.553; = 0.014, respectively). In frail, non-frail, and young groups, leptin was positively correlated with fat mass (rs = 0.802; = 0.001, rs = 0.571; = 0.011, and rs = 0.625; = 0.007, respectively) and negatively correlated with percentage of fat free mass (rs = −0.758; = 0.002, rs = −0.472; = 0.041, and rs = −0.600; = 0.011, respectively).

Postprandial gastrointestinal hormone and glucose response

Figure 2 shows the curves of ghrelin, CCK, and GLP-1 and Fig. 3 the curves of insulin and glucose for the three study groups.

Figure 2.

 Gastrointestinal hormone response. Ghrelin, CCK, and GLP-1 fasting levels and postprandial profiles (left); and integrated release during the early (0–60 min) and late (60–240 min) postprandial periods (right). For GLP-1, only persons with a positive breath test are considered. *< 0.05. AUC, area under the curve (as compared with baseline value).

Figure 3.

 Effect of age and frailty on fasting and postprandial glucose and insulin profiles (left) and on integrated release of insulin during the early (0–60 min) and late (60–240 min) postprandial periods (right). *< 0.05. AUC, area under the curve (as compared with baseline value).

Ghrelin  Fasting ghrelin levels were similar in all study groups, with mean values of 807 ± 317 pg mL−1 in frail persons, 681 ± 286 pg mL−1 in non-frail elderly, and 807 ± 294 pg mL−1 in younger adults. Ghrelin levels slightly decreased after meal ingestion in all groups, but statistically significant differences were not observed between groups.

CCK  Fasting CCK levels were similar in all three study groups (<1 pmol L−1). A common profile with an early postprandial peak of CCK 15 min after ingestion and a rapid normalization during the first postprandial hour was observed in all groups (Fig. 2). The GLM analysis and AUC0-60min showed no differences between frail and non-frail elderly persons, but there were significantly increased postprandial CCK levels in elderly persons when compared with younger adults (P = 0.013 and P = 0.006, respectively).

GLP-1  Fasting GLP-1 concentrations were 14.8 pmolL−1 (7.9) in the frail vs 11.9 (7.9) in the non-frail (= 0.184) and 13.1 (7.9) in the old groups vs 10.3 (6.4) in the younger group (P = 0.197). However, when considering those persons with a positive breath test within the study period, fasting and early postprandial GLP-1 levels were enhanced in the elderly in comparison to younger adults (Fig. 2). Moreover, younger adults presented a slight increase in GLP-1 in the late postprandial period, whereas GLP-1 release remained suppressed in the elderly. No differences in fasting or postprandial GLP-1 profiles were found between frail and non-frail persons.

Glucose and insulin  Elderly individuals showed higher fasting glucose levels compared with younger adults (109.4 ± 34.7 mg dL−1vs 87.5 ± 7.9 mg dL−1; = 0.005), and differences between frail and non-frail elderly were not observed (101.8 ± 31.9 mg dL−1vs 114.7 ± 36.3mg dL−1; = 0.151). Younger adults experienced a slight increase in serum glucose levels during the early postprandial period and a progressive decrease afterwards, maintaining postprandial levels below 110 mg dL−1 during the 4-h study period (Fig. 3). In contrast, elderly persons showed a higher increase in glucose levels during the early postprandial period and a sustained hyperglycemia (>150 mg dL−1) during the late postprandial phase. No differences in glucose response was observed between the frail and non-frail elderly groups, but very significant differences were found between younger adults and elderly persons in early and late postprandial periods. Elderly persons show significantly higher fasting insulin levels than younger adults (8.0 ± 7.5 mcUl mL−1vs 4.5 ± 5.3 mcUl mL−1; = 0.034), but no differences were observed between frail and non-frail elderly (7.9 ± 9.4 mcUl mL−1vs 8.1 ± 6.0 mcUl mL−1; P = 0.342). After ingestion, insulin values peaked during the first postprandial hour in all groups and then decreased progressively during the late postprandial period (60–240 min) in younger adults, but remained high in both groups of elderly persons. Insulin concentrations were significantly lower in younger adults than in elderly persons at 180 and 240 min, but no differences between the frail and non-frail elderly groups were found (Fig. 3).

Postprandial gastrointestinal motility

Gastric motility  Ultrasound showed reduced fasting antral area in younger adults (290.2 ± 118.7 mm2) when compared with elderly groups (379.5 ± 111 mm2, = 0.006). Significant differences in gastric antral area were also observed between frail and non-frail elderly (326 ± 91 mm2vs 418 ± 110 mm2; = 0.045, respectively). Following meal test ingestion, increase in antral area with respect to basal fasting state was higher in younger persons vs both groups of elderly in the early and late postprandial periods suggesting poor gastric compliance in older persons (Fig. 4). Ultrasound T50 was 28.8 ± 23 min in frail persons vs 35.8 ± 28 min in non-frail persons (= 0.323) and 32.9 ± 26 min in elderly groups vs 43.9 ± 29 min in younger adults (= 0.065). Increased paracetamol concentration at 30 min (C30) was 4.9 ± 3 μg mL−1 in the frail vs 3.2 ± 7 in non-frail (= 0.017) and 3.9 ± 6 in the elderly vs 2.8 ± 3 in the younger group (= 0.809). C60 was 9.0 ± 10 μg mL−1 in the frail vs 3.1 ± 4 in non-frail (= 0.008) and 5.5 ± 8 in the elderly vs 5.6 ± 2 in younger adults (= 0.303). These results, together with the AUC0-60 and AUC60-240 presented in Fig. 4, suggest an enhanced gastric emptying speed for liquids in the frail elderly in both early and late postprandial periods.

Figure 4.

 Gastrointestinal motility. (A) Gastric motility assessed by fasting and postprandial profiles of normalized antral areas; (B) Gastric emptying assessed by the paracetamol absorption test; and (C) Gallbladder emptying (% of initial volume) induced by the test meal. Right bars depict integrated motor responses during the early (0–60 min) and late (60–240 min) postprandial periods. *< 0.05. AUC, area under the curve (as compared with baseline value).

Gallbladder motility  No significant differences were observed in fasting gallbladder volume between groups: 38.4 ± 5 mL in frail elderly, 44.9 ± 3 mL in non-frail elderly, and 31.6 ± 2 mL in younger adults. The GLM analysis and comparisons of AUCs show no differences in the gallbladder emptying curve between groups. Prevalence of persons with reduced gallbladder emptying (non-contractors, <50% of emptying) was also similar between groups; 23.1% in the frail, 35.0% in non-frail, and 32.1% in younger adults. However, the mechanisms of impaired gallbladder contraction in these persons differed among groups. Younger adults and non-frail elderly ‘non-contractors’ showed a reduced CCK peak (1.4 ± 0.7 pmol L−1 and 4.0 ± 3 pmol L−1) when compared with that of ‘contractors’ of the same groups (5.7 ± 4 pmol L−1 and 9.1 ± 6 pmol L−1) suggesting impaired gallbladder contraction secondary to poor CCK release. In contrast, frail elderly persons, both ‘contractors’ and ‘non-contractors,’ presented higher and similar peaks of CCK (8.7 ± 8 pmol L−1 and 7.5 ± 6 pmol L−1, respectively) suggesting appropriate CCK release and impaired contraction caused by peripheral resistance to CCK in the gallbladder.

Octt  Time to a positive breath test (>5 ppm H+ increase) suggesting arrival of food to the cecum was similar among study groups; 114 ± 48 min in frail elderly, 125 ± 39 min in non-frail elderly, and 138 ± 42 min in younger adults. High H+ expiratory values in early stages with an early H+ peak suggesting bacterial overgrowth was observed in 50.0% of cases in the older groups and in 35.3% in the younger group.


Frail elderly patients presented anorexia characterized by diminished hunger during fasting and almost no hunger during the postprandial period. The pathophysiology of this anorexia in frail elderly patients might be related to increased gut satiety hormones and impaired gastrointestinal motility during fasting and postprandial periods. Frail elderly patients in this study were at risk of malnutrition with reduced weight and muscular mass and impaired physical function, fulfilling all criteria of the frail phenotype,15 which are associated with increased likelihood of institutionalization, illness, and increased mortality.29 To our knowledge, this is the first study to explore the two main groups of peripheral gastrointestinal mechanisms -hormones and motility- related to appetite control during fasting and postprandial periods in frail elderly patients. Gastrointestinal motility simultaneously monitored in studies assessing the postprandial profiles of these gut regulatory peptides allows a better interpretation of their effects on hunger and satiety and only low invasive methods such as ultrasonography or breath tests selected in this study should be used in these highly vulnerable patients.26 Non-frail older persons presented higher fasting and postprandial hunger values than frail persons, and younger adults showed greater hunger recuperation after ingestion in comparison to older persons, giving us the opportunity of analyzing three different hunger profiles in three phenotypes of patients differing by age and frailty. In relation to the differences observed between groups in body composition, our results agree with those in the scientific literature. A characteristic condition in aging is loss of lean body mass while fat mass is preserved or even increased. Age-related changes in body composition usually include an increase in the percentage of fat mass which can be attributed to an accelerated decrease in lean or muscle mass, reduced physical activity, reduced GH secretion, diminished sex hormones release, or decreased metabolic rate.30,31 All these factors are more accentuated in the frail.

In a previous study, we explored the effect of age and frailty on total ghrelin and CCK responses to a different meal test, and found low ghrelin concentrations in older persons and a loss of ghrelin postprandial response in frail elderly patients.11 In contrast, in this study, no significant differences have been observed between frail and non-frail and between younger and old adults regarding total ghrelin levels, suggesting that ghrelin does not play a major role in determining fasting and postprandial hunger. It’s known that only acylated ghrelin acts as a orexygenic signal. Total and acylated ghrelin curves are quite parallel over a 4-h postprandial period suggesting that acylated ghrelin remains a constant fraction over this short period of time. Thus, if no differences were found in total ghrelin between groups there will probably not be such differences in acylated ghrelin, either. Studies have compared fasting and postprandial ghrelin levels in aged and younger populations with no concordant results.9,11,32–36 Discrepancies may be due to the use of different designs, analytical methods, patient profile (age, gender, BMI, frailty), or to the use of different study meals. Ghrelin seems to be a long-term regulator of body weight rather than a short-term orexigenic signal.37,38 In contrast, CCK is widely accepted as a strong anorexigenic gastrointestinal hormone and a major mediator of anorexia of aging.9,10,39 In this study, a higher postprandial peak of CCK was observed in elderly persons in comparison to younger adults and a trend for even higher peak values of CCK was observed in frail elderly persons. The enhanced CCK release observed in older persons might mediate the anorexia in the early postprandial period, but not in fasting or the late postprandial phases and its anorexigenic effect does not seem to be mediated by a delay in gastric emptying as this was faster in frail patients. Cholecystokinin is mainly produced in the duodenum and proximal jejunum in response to the delivery of nutrients from the antrum, so enhanced gastric emptying may contribute to the higher peak of CCK observed in this study in frail elderly patients. As bile concentration in the duodenum strongly inhibits CCK closing a feedback loop regulatory mechanism, gallbladder emptying capacity is also one of the main determinants of postprandial plasma CCK levels.40 A study from Di Francesco9 reported elevated postprandial CCK levels that might inhibit hunger in a group of elderly persons showing impaired gallbladder contraction, a result very similar to that we found in our frail elderly patients. In this study, frail elderly patients with impaired gallbladder emptying presented a higher postprandial CCK peak showing decreased gallbladder sensitivity to CCK probably caused by a reduction in the population of CCK receptors.28,41 Low affinity CCK-A receptors located at vagal afferent fibers transmit the peripheral CCK signals to the CNS to mediate satiety, as CCK cannot penetrate the blood brain barrier.42 Taken together, these results suggest that high plasma CCK concentrations caused by enhanced gastric emptying and impaired gallbladder contraction and acting at vagal afferents might participate in the low early postprandial hunger values observed in frail older patients in our study. Hyperglycemia is another factor that could contribute to fasting and postprandial anorexia.4 Blood levels of glucose directly regulate hunger and satiety control centers and can also impair gastrointestinal motility.8 The increased insulin levels of elders might be related to insulin resistance related to an increase in fatty mass or adiposity in older population.43–45 In physiological conditions, increased insulin concentrations refer the presence of a positive energy balance to the hypothalamus and cause direct central satiety sensation and/or amplify leptin and inhibit ghrelin pathways.8 Leptin, which is secreted by peripheral fat cells, might also play a central role in these mechanisms of anorexia in elderly persons.1 Finally, we found a trend toward elevated GLP-1 levels in fasting and early postprandial periods in elderly patients. Glucagon-like peptide-1 is released by the distal intestine and increases in the late postprandial period in younger persons, as a part of the ileal brake. Glucagon-like peptide-1 is considered as a strong anorexigenic signal with its insulinotropic properties and direct action on satiety center in the brain and may also be involved in anorexia during these early phases in older persons8 and contribute to fasting increase in the antral area.46

Impaired gastric motility and disturbed gastric emptying might also play a relevant role in the anorexia of aging by several complex mechanisms. Firstly, enhanced fasting antral area –suggesting an increase in antral volume causing distention- might be involved in triggering fasting anorexia, as observed in patients with functional dyspepsia or diabetes.23 Secondly, impaired fundic NO synthesis leads to reduced adaptative relaxation in older persons causing satiation and also causing food to pass more rapidly from the fundus into the antrum and cause a more rapid antral filling.1 Thirdly, reduced antral compliance also enhances gastric emptying of liquids, contributing to early postprandial anorexia.1 And finally, delayed gastric emptying of solids may also cause prolonged postprandial satiety.8 It is thought that these gastric mechanisms are very relevant in the postprandial anorexia of older people, given that when nutrients are infused intraduodenally, the decrease in hunger in older persons is less pronounced.1 The three main mechanisms of impaired gastric motility in the frail elderly patients were enhanced fasting antral area, poor postprandial antral compliance, and enhanced gastric emptying rate during early postprandial period. Impaired antral accommodation might also contribute to satiation in our patients23 as well as enhanced gastric emptying rate that causes fast delivery of nutrients to the small intestine, also causing postprandial anorexia. Finally, although OCTT were similar in all three study groups, 50% of older persons presented intestinal bacterial overgrowth suggesting impaired intestinal clearance. This might be caused by decreased frequency of contractions after eating and reduced frequency of propagated clustered contractions described in older persons.47 The physiological significance of these intestinal findings in elders is uncertain, although they are similar to those observed in dyspeptic patients and patients with IBS with early satiety.48 Moreover, the high prevalence of bacterial overgrowth in the elderly might be related to the high use of PPI in this group. This study also presents some weaknesses such as a limited sample size and statistical power for some comparisons between groups, and the difficulty to adjust or control for a great number of possible socio-cultural, psychological, comorbidities, medications, and other factors that can also influence or determine hunger. Drugs were not considered as exclusion criteria to avoid a ‘hyper-selected’ and non-representative sample in aged groups, but some differences in medication were observed between young and old subjects in IPP or Ca-antagonists which may partially explain differences in gastrointestinal motility. Other limitations are the different sex distribution between frail and non-frail groups, the observer-dependent of ultrasound measurements and those limitations of the accuracy of breath test as a method to assess intestinal transit.

In summary, elderly subjects especially frail individuals present diminished hunger and also some alterations in gastrointestinal hormone release and motility. These alterations include enhanced postprandial gastric emptying for liquids and impaired gallbladder contraction in the frail and high levels of fasting glucose and insulin (and a trend for GLP-1) and increased fasting antral area, enhanced postprandial CCK release, postprandial hyperglycemia and hyperinsulinemia, and reduced antral compliance. Although all these changes in aged subjects may be related to the anorexia of aging, still further efforts are needed to translate the observed changes in peripheral gastrointestinal signals into therapeutic strategies to help prevent anorexia and under-nutrition during aging.


The authors would like to thank Dr. M. Cabré and Ms. R. Monteis for their collaboration in recruiting patients, Ms. E. Aguirre and C. Mas for technical assistance during the experiments, Dr. X Boquet and Dr R. Casamitjana for technical support during sample analysis, Dr Marta Pulido for editing the manuscript and editorial assistance, and Jane Lewis for reviewing the manuscript.


This work was supported by grants from the Spanish Ministry of Science and Innovation, Fondo de Investigación Sanitaria (PI 08/0478, INT 10/228), and the Agencia de Gestió d’Ajuts Universitaris i de Recerca (2009 SGR 708). Ciberehd is supported by the Instituto de Salud Carlos III, Barcelona, Spain.


None to be declared.

Author contribution

MSP and PC designed research (project conception, development of overall research plan, and study oversight); MSP and EM conducted research (hands-on conduct of the experiments and data collection); EP and MSP analyzed data or performed statistical analysis; MSP and PC wrote the manuscript (only authors who made a major contribution); MSP had primary responsibility for final content; all authors read and approved the final manuscript.