Hunger on the Three-Factor Eating Questionnaire is Associated with Increased Ad Libitum Energy Intake in Different Conditions

Abstract

Objective:

To assess how food-related behavioral traits, examined with the Three-Factor Eating Questionnaire (TFEQ) and the Perceived Stress Scale (PSS), influence ad libitum intake after fasting and during and after cold exposure compared to thermoneutral conditions, beyond what is explained by energy expenditure alone.

Methods:

Healthy participants (n=46; 30 males; age 36.9±10.5y; Body Mass Index (BMI) 32.3±8.7 kg/m²) completed the TFEQ, PSS, and Dual-energy X-ray absorptiometry (DXA) scan. The Hunger subscale of the TFEQ was further divided into internal and external hunger based on the methodology of Bond et al. Participants ate ad libitum from a food intake paradigm on five separate occasions, each over 24-h: (1) after a 36-hour fast at thermoneutral temperature, (2) during cold exposure (19.0 °C) in the chamber, (3) during thermoneutral conditions (23.5 °C) in the chamber, (4) after cold exposure, and (5) after thermoneutral exposure.

Results:

Total hunger score predicted intake across all five conditions (β=0.58 MJ/day, p=0.02). Internal and external subscales also predicted intake in separate models (β=1.04 MJ/day p=0.02; β=1.46, p=0.01). Participants ate more during the cold condition than during thermoneutrality (mean = 1.69 MJ/day, or 404 kcal/day; p = 0.009). Higher external hunger scores correlated with greater intake (partial r = 0.38; p = 0.01), but internal hunger scores did not (partial r = 0.22; p = 0.16), after accounting for sex, race, fat mass, and fat-free mass. Restraint, disinhibition, and PSS were not associated with intake.

Conclusion:

Increased susceptibility to hunger cues, particularly external cues, was associated with higher food intake during cold stress. These findings suggest that behavioral approaches, such as reducing responsivity to food cues, are important targets for weight loss interventions.

INTRODUCTION

Overweight and obesity have become increasingly prevalent in the United States and globally (1, 2). Weight gain is a consequence of energy intake in excess of energy expenditure. Although the mechanisms controlling eating behaviors are multifactorial and complex (including genomic, metabolic, lifestyle, and social factors), psychological factors have an important role in food intake (3). Cold exposure is a common environmental condition that affects energy expenditure and eating behavior, making it relevant to population-level weight gain. Even small, sustained increases in ad libitum energy intake can outweigh the modest expenditure stimulus and contribute to a positive energy balance on a large scale. Cold exposure increases ad libitum energy intake, and evidence suggests that concurrent changes in energy expenditure alone cannot fully explain this increase (4). This motivates the evaluation of behavioral traits that may influence eating habits in response to environmental challenges.

The Three-Factor Eating Questionnaire (TFEQ) assesses restraint, disinhibition, and hunger (5). TFEQ-Restraint measures a person’s concern and actions regarding weight control; TFEQ-Disinhibition measures a person’s tendency to eat opportunistically; and TFEQ-Hunger measures the extent to which feelings of perceived hunger lead to food intake (6, 7). Prior research has found a positive association between TFEQ-Hunger and TFEQ-Disinhibition and greater propensity for overweight and obesity, with mixed results on the role of TFEQ-Restraint in increasing BMI (6, 8, 9). Furthermore, higher energy expenditure has been shown to be associated with lower restraint, higher disinhibition, and greater susceptibility to hunger cues, suggesting a relationship between different dimensions of the TFEQ and energy demand that alter appetite (10). Since its initial development, the TFEQ has been restructured and researchers have called for subscales of hunger, restraint, and disinhibition to measure aspects embedded within each of these factors. Bond et al. (2001) identified two hunger subscales: internal (TFEQ-Internal Hunger) and external locus for hunger (TFEQ-External Hunger) where external locus describes a hunger sensation that is triggered by an external cue (e.g., seeing a delicacy makes them want to eat), and internal hunger describes hunger that is driven by internal cues. While there is strong evidence for associations between the TFEQ and overweight/obesity, the underlying behavioral mechanism in this relationship has not been widely studied.

Stress has also been shown to be associated with emotional eating and overweight and obesity (11). The Perceived Stress Scale (PSS) measures the degree to which situations in one’s life are perceived as stressful (12). Nonsmokers have reported eating more when perceived stress is high (13) and the PSS has been shown to be associated with more uncontrolled and emotional eating (14).

Cold exposure has been shown to increase ad libitum intake, mostly attributed to an increase in energy expenditure (15). However, we recently reported that acute, short-term cold exposure increases ad libitum energy intake compared to thermoneutral conditions, though this increase is unrelated to changes in energy expenditure. This finding suggests that other factors, independent of energy expenditure (e.g., behavioral traits), play a role in food intake in response to cold exposure. The impact of disinhibited eating, susceptibility to hunger cues, and cognitive dietary restraint on intake after cold exposure and fasting is still unclear. Previous research has shown that physiological stress from a cold-pressor task, in which participants’ hands are submerged in ice water for two minutes, is correlated with a desire to binge eat (16), and that acute psychological stress can increase eating in the absence of hunger (17). Rutters et al. demonstrated that individuals more prone to disinhibited eating, as identified by the TFEQ, are more impacted by stress induced by an unsolvable arithmetic task (17). Therefore, cold exposure can serve as a physiological challenge to elucidate how eating behavior changes in relation to self-report questionnaires (e.g., PSS and TFEQ).

In this study, we conceptualized cold exposure as a physiological challenge rather than a psychological stressor and investigated whether psychosocial factors such as the TFEQ and PSS are associated with increased ad libitum energy intake beyond energy expenditure (a) during cold exposure, (b) during thermoneutral conditions, (c) after cold exposure with a eucaloric diet, (d) after a thermoneutral condition with a eucaloric diet, and (e) after 36-h of fasting in thermoneutral conditions.

METHODS

Research Design

The study was registered at clinicaltrials.gov (NCT02939404) and was approved by the Institutional Review Board of the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK). Participants were recruited from the Phoenix, Metropolitan area to the NIDDK clinical research unit (CRU) in Phoenix, Arizona; no direct recruitment was done at Phoenix Indian Medical Center or any Indian Health Service facility. All participants provided written informed consent prior to beginning the study. Participants were between age 18–55, non-pregnant (verified by urine pregnancy test), and healthy as determined by medical history, physical examination, and routine screening labs, and had stable weight within the past six months by volunteer report. Volunteers with recent use of nicotine-related products and drugs of abuse were excluded based on urine cotinine and drug screen (e.g., for opiates, amphetamines). Complete inclusion and exclusion criteria are described in the Supplemental Methods.

After screening, eligible volunteers were admitted to the CRU. An overview of the study procedures is shown in Figure 1. Volunteers were prescribed a daily weight-maintaining diet (50% carbohydrate, 30% fat, and 20% protein; food quotient 0.87) using equations developed specifically for the CRU based on weight and sex (18). Each participant had a different caloric maintenance value. Initially, energy requirements were estimated based on body weight. Then, the research dietitian adjusted them to maintain a constant body weight within 1% of the admission weight. A computerized software package (CBORD Diet Analyzer, ESHA Database, Ithaca, NY) was used to calculate energy content and macronutrient composition. The weight maintenance diets were the same inside and outside the chamber.

Figure 1.

Study Diagram of Clinical Trial.

On the second day, participants had body composition measured by DXA scan (iDXA, GE Lunar Healthcare). On the third day after admission, participants completed behavioral questionnaires including the TFEQ and PSS. After four days of the weight-maintaining diet, participants had a 75-g oral glucose tolerance test (OGTT) and were withdrawn from the study if they had diabetes based on American Diabetes Association criteria for fasting and 2-h plasma glucose concentrations (19).

After the OGTT, participants completed five separate ad libitum eating periods using a reproducible vending machine paradigm (details discussed below in Ad Libitum Food Intake as Measured by the Vending Machine Paradigm) (20). Volunteers ate ad libitum for 24-h on two separate occasions while inside a respiratory chamber (21, 22). The ambient air temperature inside the respiratory chamber is controlled by its own heating, ventilation, and air conditioning system. Ad libitum intake was measured inside a respiratory chamber during cold exposure over 24-h (thermostat set to 19 °C, mean temperature: 19.2 °C ± 0.3 °C) compared with a similar 24-h period inside the respiratory chamber during thermoneutral conditions (thermostat set to 23.5 °C, mean temperature: 23.5 °C ± 0.2 °C). Participants wore standardized clothing consisting of a hospital gown, pants, and ankle-length socks (23). From 11 p.m. to 6 a.m., three blankets were provided so participants could create a comfortable sleep environment. If they began to shiver, they notified the nurse and used the blankets until the shivering ceased. Movement space in the chamber was limited, and exercise was not permitted. Reading, writing, phone calls, and watching television were permitted inside the chamber.

To examine the residual effects of cold exposure on ad libitum intake, participants were fed a eucaloric diet during the cold exposure inside the respiratory chamber for 24-h (mean temperature 19.2 °C ± 0.4 °C) followed by ad libitum intake over 24-h after exiting the chamber (i.e., returning to the thermoneutral conditions of the CRU), compared with a similar 24-h ad libitum intake period after exiting a eucaloric respiratory chamber with thermoneutral conditions (mean temperature 23.5 °C ± 0.2 °C). For these two chambers with eucaloric diet provided inside the chamber, the prescribed energy intake (50% carbohydrate, 30% fat, 20% protein) was calculated from an equation based on sex, height, and weight developed in the CRU that provides approximately 20% fewer calories than weight-maintenance requirements on the ward to account for decreased physical activity in the respiratory chamber (24). Mealtimes and food items were the same for eucaloric chambers during cold exposure and thermoneutral conditions (Table S1). The final 24-h ad libitum eating period occurred after exiting a fasting 24-h respiratory chamber at thermoneutral conditions (mean temperature 23.4 °C ± 0.3 °C). The total duration of the fast was 36-h when including the 12-h fasting period since the last meal on the ward.

Ad Libitum Food Intake as Measured by the Vending Machine Paradigm

Ad libitum food intake on the ward (i.e., after exiting the respiratory chambers) was measured by a vending machine paradigm previously described, validated, and tested for reproducibility (20). The refrigerated vending machines (model 3007; U-Select-It) were stocked with 40 items. To determine which food items were made available in the vending machine, each participant was asked to complete a food preference questionnaire which consisted of a list of 80 food items presented in random order. Individuals were asked to assign each food item a hedonic rating with a 9-point Likert scale (1=dislike, 5=neutral, 9=like extremely). Food items made available to participants in the vending machine consisted of food that the participants rated as intermediate (between 4 and 8) on the food preference questionnaire (77 of the items from the food preference questionnaire were used to stock the vending machine due to three being unavailable for purchase). The list of food items is available in Table S2.

Participants were assigned to a single vending machine and had unrestricted access to the vending machine for 23.5 h/day. Participants were instructed to eat only in the vending room, whatever they wished, whenever they desired, and to return the unconsumed food portions to the metabolic kitchen for calculation of actual energy consumed obtained using a food database. For the ad libitum food intake occurring inside the respiratory chamber, the paradigm was similar though modified for the small room. Inside the chamber, the 40 food items were placed in a small refrigerator instead of the vending machine. After each meal, leftover foods were collected by kitchen staff and weighed to actual food consumed.

Daily energy, protein, fat, and carbohydrate intakes were calculated from the actual weights of food and condiments consumed with the use of the Food Processor SQL Edition (ESHA, version 10.0.0; ESHA Research, Salem, OR) modified to reflect the nutrient content of specific food items as indicated by the manufacturer.

Perceptive Response to Cold and Fasting

Hunger, fullness, coldness, and stress scores were assessed before entering the chamber, at 06:30 p.m while inside the chamber (10.5 hours into the chamber), and upon exiting the chamber stay using a visual analog scale (VAS) with scores ranging from 0 (“not at all”) to 100 (“extremely”).

Behavioral Questionnaires

The TFEQ (5) consists of 51 items assessing three factors of eating behavior: susceptibility to hunger, dietary disinhibition, and cognitive dietary restraint. The hunger subscale was further divided into internal (e.g., “I often feel so hungry that I just have to eat something”) and external loci (e.g., “Being with someone who is eating often makes me hungry enough to eat also”) of hunger based on the methodology described by Bond et al. (25). The questions that make up each subscale can be found in Table S3.

The Perceived Stress Scale (12) consists of 14 items that assess the stress domains of unpredictability, lack of control, burden overload, and stressful life circumstances in the last 28 days.

Statistical Analysis

Statistical analyses were performed using SAS, version 9.4 (SAS Institute Inc., Cary, NC, USA). Alpha was set at 0.05 and two-sided p-values were reported. Categorical variables are displayed as counts; continuous variables are displayed as means and standard deviations (SD).

To compare VAS scores for the five different conditions, during and after each chamber stay, repeated measures mixed model analyses with compound symmetry covariance structure were used. In these mixed models, VAS scores from each exposure were compared with the appropriate control condition via planned comparisons as described in the following section. VAS scores during cold exposure with ad libitum food intake were compared with scores during the thermoneutral condition with ad libitum food intake (control). The VAS scores during cold exposure with eucaloric food intake and from the fasting, thermoneutral condition were compared with the scores during the eucaloric, thermoneutral chamber stay (control). Pearson correlations were used to quantify the association between VAS scores and ad libitum intake during and after chambers.

The associations between psychometric tests (TFEQ, PSS) and ad libitum energy intake during the five conditions were quantified by linear mixed models using compound symmetry covariance structure. Additionally, mixed models were performed with interaction terms to assess whether the association between behavioral questionnaire scores and intake were different across the five conditions. However, the interaction terms were not significant for any model and were removed. As ad libitum food intake was not measured beyond 24 hours for fasting and cold eucaloric conditions, only the first 24-h ad libitum intake period after the eucaloric thermoneutral chamber was used in these mixed models for the main analysis. In sensitivity analysis, the second and third ad libitum food intake days after the eucaloric thermoneutral chamber were included. External and internal hunger β coefficients from separate models were compared utilizing the following Z test (26) where the β coefficients and standard errors (SE) were derived from the separate mixed models for internal and external hunger described above.

$$Z=\frac{B_1-B_2}{\sqrt{({SE{B}_1)}^2-{\left(SE{B}_2\right)}^2}}$$

Pearson correlations were also used to quantify the associations between psychometric tests and the change in ad libitum energy intake from the appropriate control condition to the exposure (i.e., during cold, after cold, and after fasting). Partial Pearson correlations were adjusted for sex, fat mass, fat-free mass, and Native American race.

RESULTS

Of 116 people screened, 69 people met eligibility criteria and were admitted. Of the 69, 23 people were excluded (Figure S1). Thus, 46 people were analyzed. Participant characteristics are reported in Table 1 along with mean test scores for the TFEQ and PSS.

Table 1.

Participant characteristics and response to behavioral questionnaires.

Variable	Female (n=16)	Male (n=30)	Total (n=46)
Age (years)	32.9 ± 10.4 (19.4, 49.6)	39.0 ± 10.1 (18.9, 55.4)	36.9 ± 10.5 (18.9, 55.4)
Race/ethnicity	Native American 12 Black 1 Hispanic 1 White 0 Unspecified 2	Native American 15 Black 5 Hispanic 4 White 4 Unspecified 2	Native American 27 Black 6 Hispanic 5 White 4 Unspecified 4
BMI (kg/m2)	38.1 ± 7.8 (23.5, 52.8)	29.2 ± 7.5 (17.6, 46.5)*	32.3 ± 8.7 (17.6, 52.8)
Fat-free mass (kg)	50.9 ± 7.8 (39.7, 66.1)	60.6 ± 8.2 (46.7, 75.0)*	57.2 ± 9.2 (39.7, 75.0)
Fat-mass (kg)	51.9 ± 16.2 (21.7, 79.4)	28.2 ± 14.7 (9.2, 64.2)*	36.4 ± 18.9 (9.2, 79.4)
Height (cm)	163.9 ± 5.4 (153.5, 173.0)	174.7 ± 7.4 (163.0, 188.0)*	170.9 ± 8.5 (153.5, 188.0)
Weight (kg)	102.8 ± 23.1 (61.4, 145.5)	88.8 ± 21.3 (61.2, 139.2)*	93.7 ± 22.7 (61.2, 145.5)
PSS	25.4 ± 6.4 (14, 37)	21.7 ± 7.3 (7, 38)	23.0 ± 7.2 (7, 38)
TFEQ
Restraint	8.4 ± 2.9 (4, 16)	8.3 ± 5.0 (0, 19)	8.3 ± 4.4 (0, 19)
Disinhibition	5.1 ± 2.5 (2, 11)	4.3 ± 3.0 (1, 12)	4.6 ± 2.8 (1, 12)
Hunger, Total	4.7 ± 3.3 (0, 12)	4.0 ± 3.4 (0, 13)	4.2 ± 3.4 (0, 13)
Internal Hunger	1.9 ± 1.8 (0, 5)	1.6 ± 1.9 (0, 6)	1.7 ± 1.9 (0, 6)
External Hunger	1.7 ± 1.2 (0, 5)	1.3 ± 1.4 (0, 5)	1.4 ± 1.4 (0, 5)

Mean ± standard deviation (minimum, maximum); BMI, body mass index; PSS, Perceived Stress Scale; TFEQ, Three-Factor Eating Questionnaire.

^*Significant differences between sexes (p <0.05).

Ad Libitum Intake

As previously reported (4), participants ate more overall during the cold exposure (Table 2). On average, participants ate 404 kcal or 1.69 MJ (p=0.009) more during the cold exposure chamber compared to the thermoneutral chamber. Average ad libitum energy intake after cold exposure did not change compared with after the thermoneutral chamber (p=0.35, Table 2). Average ad libitum energy intake also did not increase after 36-h fasting compared with after the eucaloric thermoneutral chamber (p=0.54, Table 2), similar to a previous report in a smaller sample (27).

Table 2.

Ad Libitum Energy Intake Measurements.

Condition	Ad Libitum Intake, MJ (kcal)	Mean Change, MJ (kcal)
Inside Respiratory Chamber
During thermoneutral chamber (23.5 °C)	17.80 ± 5.22 (4254 ± 1249)	Reference
During cold chamber (19 °C)	19.49 ± 6.06 (4658 ± 1449)	1.69 ± 4.20 (404 ± 997) *
After Respiratory Chamber
After eucaloric thermoneutral chamber (23.5 °C)	16.41 ± 5.51 (3923 ± 1316)	Reference
After eucaloric cold chamber (19 °C)	16.95 ± 5.83 (4052 ± 1392)	0.54 ± 3.84 (129 ± 917)
After fasting thermoneutral chamber (23.5 °C)	16.88 ± 5.46 (4034 ± 1305)	0.35 ± 3.82 (85 ± 914)

^*p=0.009

Ad libitum intake in each condition with pairwise t-test between cold or fasting condition and the reference condition (thermoneutral). Data represented as mean ± standard deviation.

Perceptive Ratings (Visual Analog Scale)

Cold.

As expected, participants felt colder overall during the cold exposure inside the respiratory chamber when prescribed a eucaloric diet compared to the thermoneutral eucaloric chamber (β=30.3 VAS points, p<0.0001) and during the ad libitum cold exposure chamber compared to the ad libitum thermoneutral chamber (β=29.3 VAS points, p<0.0001, Figure 2A; VAS averages across timepoints and conditions are shown in Figure S2).

Figure 2.

Visual Analog Scores during chamber for (A) cold, (B) stress, (C) hunger, (D) fullness. Individual scores depicted by dots and bars represent means with 95% confidence intervals. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. ns, not significant.

Stress.

On average, participants also felt more stressed in the ad libitum cold exposure chamber compared to the ad libitum thermoneutral chamber (β=7.0 VAS points, p=0.0095), but not during the fasting chamber compared to the eucaloric thermoneutral chamber (Figure 2B).

Hunger.

As expected, participants felt hungrier during the fasting chamber compared to the eucaloric thermoneutral chamber (β=30.0 VAS points, p<0.0001), but there was no difference in hunger ratings between the ad libitum cold exposure chamber and the ad libitum thermoneutral chamber (Figure 2C). Despite the 20% reduction in caloric content during the eucaloric chamber stays, participants did not report greater hunger, as VAS ratings of hunger were similar between the eucaloric cold and thermoneutral conditions.

Fullness.

As expected, participants were fuller during the eucaloric chamber compared to the fasting chamber (β =27.0 VAS points, p<0.0001) (Figure 2D) but perceptions of fullness were similar during the ad libitum cold exposure chamber and during the ad libitum thermoneutral chamber (Figure 2D).

VAS ratings during and after the chambers were not associated with ad libitum food intake (all p’s > 0.09).

Behavioral Questionnaires and Ad Libitum Intake

The correlations between the behavioral questionnaires (i.e., TFEQ and PSS) and the five 24-h ad libitum energy intake conditions are shown in Figure S3. In a mixed model, TFEQ-Hunger was associated with ad libitum energy intake independent of chamber (β=140 kcal/day/1 point difference in TFEQ-Hunger or 0.58 MJ/day/1 point difference in TFEQ-Hunger, p=0.02). TFEQ-Internal Hunger and TFEQ-External Hunger predicted ad libitum energy intake in separate models (β=248 kcal/day/1 point difference in TFEQ-Internal Hunger, 1.04 MJ/day/1 point difference in TFEQ-Internal Hunger, p=0.02; β=350 kcal/day/1 point difference in TFEQ-External Hunger, 1.46 MJ/day/1 point difference in TFEQ-External Hunger, p=0.01). Although the slope for external hunger was qualitatively higher than that for internal hunger, the difference was not statistically significant (Z = 0.58, p = 0.56). There was no interaction (i.e., testing whether the association between TFEQ and intake is different across the five conditions) between eating condition and TFEQ-Hunger, TFEQ-External Hunger, and TFEQ-Internal Hunger on intake. TFEQ-Disinhibition, TFEQ-Restraint, and PSS were not associated with ad libitum energy intake in separate models.

Behavioral Questionnaires and Change in Ad libitum Food Intake to Cold

TFEQ-Hunger, TFEQ-External Hunger, and TFEQ-Internal Hunger were not associated with the increase in ad libitum energy intake during the cold exposure (Table S4). However, after adjusting for sex, fat mass, fat-free mass, and Native American race, higher TFEQ-Hunger and TFEQ-External Hunger were correlated with greater increases in intake (partial r= 0.33, p=0.03; partial r=0.38, p=0.01, Figure 3A–B), but TFEQ-Internal Hunger was not (partial r=0.22, p=0.16, Figure 3C).

Figure 3.

Scatterplot of residuals of change in ad libitum intake between the ad libitum cold exposure chamber and the ad libitum thermoneutral chamber and (A) TFEQ-Hunger, (B) TFEQ-External Hunger, and (C) TFEQ-Internal Hunger. Models adjusted for sex, fat mass, fat-free mass, and Native American race.

To determine if the prior associations extend to the day after the cold chamber, we performed the same analysis to examine the difference in ad libitum intake the day after the eucaloric chambers (thermoneutral and cold). There was no significant correlation between TFEQ and change in intake between the day after the eucaloric cold chamber and the day after the thermoneutral chamber (Table S4).

Behavioral Questionnaires and Change in Ad libitum Food Intake to Fasting

There was no significant difference in ad libitum energy intake between the day after the fasting chamber and the day after the eucaloric thermoneutral chamber (85 ± 914 kcal or 0.35 ± 3.82 MJ, p=0.54). TFEQ-Hunger, TFEQ-Internal Hunger, and TFEQ-External Hunger were associated with ad libitum intake after fasting (r=0.30, p=0.047; r=0.30, p=0.048; r=0.36, p=0.01; Figure S3). However, these TFEQ hunger scores were not associated with the difference in intake between ad libitum intake the day after the fasting chamber and ad libitum intake the day after the eucaloric thermoneutral chamber (all p’s > 0.32 Table S4).

DISCUSSION

This study investigated the role of the PSS, TFEQ-Hunger, TFEQ-Disinhibition, and TFEQ-Restraint in observed ad libitum eating behavior in five conditions including after a fast and during and after cold exposure. TFEQ-Hunger was broadly associated with ad libitum intake across the different conditions. Furthermore, we found that TFEQ-Hunger and TFEQ-External Hunger, but not TFEQ-Internal Hunger, were associated with the increase in ad libitum energy intake from thermoneutral to cold conditions.

That TFEQ-External Hunger was associated with increased food intake overall and in response to cold indicates that heightened food cue responsivity may be contributing to food consumption. In general, food cues may include the sight and smell of real food, advertisements, or cues associated with memories of food (e.g., association of holiday family gatherings with food) (28). Schachter’s externality hypothesis proposes that individuals with obesity are more responsive to external cues as opposed to internal (29) and may be highly relevant in the current obesogenic environment where the ability of food-related marketers to target individuals with food cues is expanding (e.g., digital advertisements through smartphones). Food cue reactivity was found to be associated with increased eating and subsequent weight gain in a meta-analysis of 45 prior reports (30). Consistent with this meta-analysis and Schachter’s externality hypothesis, we found that external hunger cues were contributors to ad libitum food intake. Our findings further suggest that external food cues as identified by the TFEQ may apply to other situations such as that provoked by cold. Thus, the current study extends the importance of food cue responsivity to a cold stressor. This distinction likely reflects that VAS hunger ratings measure short-term appetite sensations, while the TFEQ assesses a stable behavioral tendency to respond to hunger cues, which more strongly predicts ad libitum intake.

The association between intake and external hunger cues suggests that food cue responsivity may be a useful target for weight loss interventions. External cues and cravings can substantially impact eating behavior (30). Cue-based interventions have been developed to reduce responsiveness to food cues through self-monitoring and exposure to foods for craving habituation. These interventions have been shown to be effective at reducing intake and overeating (31, 32, 33), stabilizing BMI (31) and resulting in greater weight loss than lifestyle interventions (32). Appetite awareness may be a useful target for internal hunger cues, as appetite awareness interventions have also been found to reduce intake and improve emotional eating, external eating, and food responsiveness (34).

The pathways by which increased susceptibility to hunger cues leads to increased intake overall and during a cold stressor are unclear. Others have proposed that the hedonic reward system involving dopamine have a role in food cue reactivity and may be altered in obesity (35) and cold (36). Other possible hormones include cortisol and ghrelin. Both hormones have been previously shown to increase food intake (11, 37, 38, 39) and increase with cold stimuli (16, 40).

Although participants reported higher hunger scores on the VAS from short-term fasting, overall participants did not eat more after fasting. Prior literature has shown mixed results on intake after a period of fasting, with some reporting that that fasting does not result in significant changes in total energy consumption (41, 42, 43) while others report increased intake after a period of fasting (44). As previously suggested by Unlu et al., 36-h of fasting may be an insufficient amount of time for there to be a compensatory increase in ad libitum energy intake (27).

Both the PSS and perception of current stress as assessed through the VAS were not associated with ad libitum intake. This may be because our sample had a low momentary stress score, even under cold exposure conditions. In addition, the PSS measures perceptions of stress within the past month which may not accurately portray perception of stress in the cold conditions. Stress has been found to both increase and decrease food intake (11) and prior research has shown that only those that are high cortisol reactors increase snack intake upon increased number of hassles (45). Participants may not have been stressed enough during cold exposure to elucidate a response or that they were not enough high cortisol reactors to establish an association between stress and intake.

The strengths of this study include multiple objective measures of ad libitum energy intake that are highly reproducible within individuals (20) and perceptive responses to cold exposure. Furthermore, this study assessed stress in multiple ways including momentary stress with the VAS and perceptions of stress over the longer term with the PSS. This study is unique because participants were subjected to 24h of cold exposure and their ad libitum intake response was measured both during and after this cold exposure for 24 hours. Furthermore, all procedures were conducted in a single inpatient visit in highly controlled conditions. However, this study had several limitations. First, although participants felt colder during cold exposure, there was only a limited difference in self-reported psychological stress during cold exposure. As the cold stress was mild (19°C or 66°F), it is unclear if results would have differed with a stronger cold stimulus. Second, participants ate ad libitum after 36-h of fasting and this may not have been enough time to induce compensatory overeating. Although ad libitum intake was higher inside the chambers, we do not believe this significantly impacted the findings of this paper. In the chambers, food was accessible within two feet, unlike in the post-chamber setting, where participants had to enter another room. This, combined with potential boredom during chamber stays, may have contributed to the increased intake. However, since all comparisons were made within the same context (chamber-to-chamber and post-chamber-to-post-chamber), the relative differences remain valid. It would be interesting to elucidate this finding in a further paper.

Due to collinearity with body composition covariates and limited power for interaction tests, we did not evaluate BMI as a modifier of the effects of behavioral traits on intake. The potential for BMI to moderate these associations should be addressed in larger studies.

Lastly, the absence of a measurable increase in 24-hour energy expenditure during mild cold exposure suggests that increased ad libitum intake does not merely compensate for thermogenesis. This finding indicates the presence of alternative mechanisms. Cold afferent signals reach hypothalamic integrators that recruit orexigenic pathways, including the neuropeptide Y (NPY) system, the agouti-related peptide (AgRP) system, and the orexin system. These pathways promote feeding even when total energy expenditure remains unchanged. Cold exposure may also increase the salience of food cues and enhance reward valuation, aligning with the observed association between external hunger and intake. Additionally, acute changes in peripheral signals, such as ghrelin or thyroid axis elements, could transiently increase appetite. The chamber environment provided immediate access to food, which can interact with these biological drivers. Together, these pathways provide testable mechanisms through which cold exposure can increase intake without a concurrent rise in energy expenditure.

In conclusion, this study showed that perceived hunger, particularly that elicited by external cues as measured by the TFEQ-Hunger and TFEQ-External Hunger, is associated with higher ad libitum intake. Food cue responsivity may be a useful target for interventions to reduce energy intake. Additionally, these treatments may include teaching coping strategies to deal with external stressors such as cold.

STUDY IMPORTANCE.

What is already known?

• Increased disinhibition and hunger on the Three-Factor Eating Questionnaire (TFEQ) are associated with obesity.

• Cold exposure increases ad libitum energy intake, though this has primarily been attributed to increased energy expenditure.

What does this study add?

• Identification of a behavioral trait, TFEQ-Hunger, that helps explain the association between the TFEQ and obesity.

• The results highlight that external hunger is a greater contributor to ad libitum intake than internal hunger during a cold stressor.

How might these results change the direction of research or the focus of clinical practice?

• This study provides further support for interventions that address cue responsivity and appetite awareness to reduce overeating and food responsiveness.

• Suggests that interventions related to overeating should teach coping strategies for external stressors and promote appetite awareness to reduce susceptibility to internal hunger cues.