Appetitive traits as targets for weight loss: The role of food cue responsiveness and satiety responsiveness

Abstract

Individuals with overweight or obesity (OW/OB) are at increased risk for significant physical and psychological comorbidities. The current treatment for OW/OB is behavioral weight loss, which provides psychoeducation on nutrition and physical activity, as well as behavior therapy skills. However, behavioral weight loss is not effective for the majority of the individuals who participate. Research suggests that overeating, or eating past nutritional needs, is one of the leading causes of weight gain. Accumulating evidence suggests that appetitive traits, such as food cue responsiveness and satiety responsiveness, are associated with overeating and weight in youth and adults. The following review presents the current literature on the relationship between food cue responsiveness, satiety responsiveness, overeating, and OW/OB. Research suggests that higher food cue responsiveness and lower satiety responsiveness are associated with overeating and OW/OB cross-sectionally and longitudinally. Emerging data suggest that food cue responsiveness and satiety responsiveness may exist along the same continuum, and can be targeted to manage overeating and reduce weight. We have developed a treatment model targeting food cue responsiveness and satiety responsiveness, to reduce overeating and weight and have preliminary feasibility, acceptability, and efficacy data, with testing currently being conducting in larger trials. Through programs targeting appetitive traits we hope to develop an alternative weight loss model to assist individuals with a propensity to overeat.

1. Introduction

Obesity is highly prevalent in the United States as approximately one third of children and two thirds of adults have overweight or obesity (OW/OB). (1, 2) OW/OB is associated with cardiovascular disease, type 2 diabetes, cancer, osteoarthritis, psychological impairment, poor quality of life and all-cause mortality. (3, 4) Medical care for individuals with OB costs nearly $150 billion across the United States, (5) and these costs are expected to rise by $48–66 billion per year by the year 2030. (6) Considering the high prevalence rates, rising medical costs, and significant comorbidities, it is essential that more potent models are developed to treat OW/OB effectively.

The current empirically supported treatment for adults with OW/OB is behavioral weight loss (BWL). (7–12) BWL includes dietary and caloric recommendations, guidelines for physical activity, and behavior therapy skills to adhere to treatment recommendations. Although BWL is effective and provides clinically significant weight loss for some adults, (10) BWL is not effective for all individuals. Up to 50% of participants in BWL treatment programs fail to achieve meaningful weight loss. (11) Moreover, BWL is even less effective in providing long-term weight-loss maintenance, as 65% of participants no longer meet the initial 7% weight-loss goal 4 years post-treatment. (12) These results could be due to heterogeneity and a lack of consideration for individual characteristics that can impact treatment response. (13, 14) Thus, it is imperative to understand the individual characteristics that predict treatment outcomes in the short and long term, and to develop novel treatments to address mechanisms associated with non-response.

Eating past nutritional needs, or overeating, is one of the most proximal drivers of OB rates (15, 16) and is considered more important than metabolic changes. (17–19) Rates of overeating are especially high in OW/OB samples, with up to 80% of adults with OW/OB endorsing overeating regularly. (20) Overeating is a complex process which is influenced by individual behavior, the environment, genes, physiology, and neural processes. However, societal advances have created an “obesogenic” environment that encourages excess energy intake and discourages energy expenditure. (21–24) Calorically dense foods are easily available, highly variable, tasty, relatively inexpensive, and portable. (25–29) Overeating in today’s environment is incredibly easy, especially for those who have a propensity to overeat.

2. Behavioral susceptibility to overeating

Interestingly, not everyone who lives in this environment has OW/OB. In fact, within the same family there can be weight discordant children. It is possible that some individuals have a behavioral susceptibility to overeat. Behavioral susceptibility was first described by Stanley Schachter who hypothesized that individuals with OW/OB, compared to individuals with a healthy weight, are more reactive to external cues to eat and less sensitive to internal satiety signals. (30, 31) Jane Wardle and colleagues extended this theory and developed the behavioral susceptibility theory (BST), (32–35) which proposes that genetically determined appetitive traits interact with the environment and lead to overeating and weight gain in individuals with these risk factors. The BST focuses on two important aspects of appetite, eating onset and eating offset. Within the context of BST, food cue responsiveness (FR) is considered the primary driver of eating onset, while satiety responsiveness (SR) is the primary driver of eating offset. FR conceptually relates to biological, cognitive and emotional responses to food cues, and includes concepts such as cravings, emotional eating, and reward-based eating. SR refers to perceiving and stopping eating based on satiety cues. The following review of the literature takes the perspective of the BST and briefly summarizes data to date on genetics, appetitive traits, interactions with the environment, overeating and body weight.

The BST states that genetic risk factors interact with appetitive traits and the environment to influence overeating. (35) Decades of research have shown that human body weight is highly heritable, with estimates ranging from 32% (36) to 90%, (37) with a median estimate of 73%. (38) The Collaborative Project of Development of Anthropometrical Measures in Twins (CODATwins), which includes over 200,000 twin pairs from 22 countries, (39) showed that the influence of genetic factors on BMI is lowest at age 4 years (40%) and highest at age 19 years (75%). (39, 40) Interestingly, the influence of the shared environment is not observed after age 15. These results suggest that over time, the genetic influence on BMI strengthens and the shared environment weakens, pointing to the importance of an individual’s interaction with the obesogenic environment. (35) Although the BST is an important theory, it does not specifically describe how these appetitive traits develop and more importantly, how the environment interacts with individual level variables. Theoretically, by decreasing FR (eating onset) and improving SR (eating offset), individuals could learn how to reduce overeating in the current obesogenic food environment.

3. Assessments of Food Cue Responsiveness (FR) and Satiety Responsiveness (SR)

A variety of measures exist that assess FR and SR. Jane Wardle and colleagues developed the Child Eating Behavior Questionnaire (CEBQ) (41) several years before the BST was formally articulated. The CEBQ was initially designed to be a parent report of children’s eating, and includes 8 subscales, FR, SR, emotional over-eating, enjoyment of food, desire to drink, slowness in eating, emotional under-eating and food fussiness. The FR subscale includes four items that assess desires to eat outside of physiological hunger and the SR subscale includes five items that evaluate a child’s sensitivity to stop eating or choosing not to eat in response to feeling full. The initial validation of the CEBQ was with 4–5 year olds in the US. (42) Since the initial publication, the CEBQ has been adapted to assess eating behaviors in babies (Baby Eating Behavior Questionnaire (43)) and adults (Adult Eating Behavior Questionnaire (44)).

FR can be assessed using neuroimaging techniques, such as MRI. Typically a food cue, either as a picture or as a taste, is presented and Blood Oxygenated Level Dependent (BOLD) response either in anticipation of the cue or delivery of the cue is assessed. Neural areas associated with the motivation to eat past nutritional requirements are the dopamine reward/motivation circuitry involving striatal limbic and cortical substrates. (45, 46) Key components of this hedonic pathway are located in the cortico-limbic areas of the brain and include the nucleus accumbens and caudate nucleus (dopaminergic reward pathways which govern anticipation and motivation); amygdala and hippocampus (learning); anterior insula (sensory processing); and orbitofrontal cortex (reward value appraisal, executive control, and decision-making). (47) Consumption of palatable food (those high in fat and sugar) releases dopamine in the ventral and dorsal striatum. The release of dopamine in the dorsal striatum is proportional to the self-reported level of pleasure gained by eating the food. (48) High fat and high sugar foods may differentially affect these brain reward regions. One study compared high sugar milkshakes and high fat milkshakes on BOLD response and showed high fat milkshakes show higher levels of BOLD responses in caudate and somatosensory regions, but no significant bilateral insular changes while high sugar milkshakes showed higher levels of BOLD responses in putamen and gustatory regions and increased bilateral insula. (49)

Data suggest that these reward circuitry dysregulations differentiated those who have OW/OB and those who have a healthy weight. (50–52) For instance, children with OW/OB compared to those with healthy weight show higher levels of BOLD responses in the bilateral insula and bilateral amygdala in response to taste of sugar compared to water. (53) Another study compared children with OW/OB to children with healthy weight and showed greater BOLD responses in right insula, operculum, bilateral precuneus, and posterior cingulate cortex following milkshake consumption. (54) Studies also demonstrate greater increase in fronto-striatal circuitry activation during anticipation of high-caloric foods as compared with those with healthy weight. (55, 56) After eating a 500 calorie meal, participants with obesity, compared to participants with healthy weight, show greater activation in the medial prefrontal cortex, superior frontal gyrus, caudate, and hippocampus. (57) In addition, over time, individuals with obesity, compared to individuals with healthy weight, demonstrated greater activation in the corticolimbic regions (lateral orbital frontal cortex, caudate, anterior cingulate), suggesting that individuals with OB have sustained responses in brain regions implicated in reward, even after eating (58). Importantly, for individuals with OW/OB, greater activation of the hedonic pathways in response to food images has been shown to predict short-term weight loss (59, 60) and was associated with successful maintenance of ≥ 13.6 kg (30 lbs.) weight loss over 3 years or more. (61) Hypothesized mechanisms of action include impulsivity or lack of self-control, which when coupled with higher levels of reward, may contribute to the drive to overeat. (62, 63)

In addition to using the CEBQ and brain imaging, concepts related to FR and SR can also be assessed with a variety of other questionnaires and tasks. The eating in the absence of hunger (EAH) paradigm (64) and the EAH questionnaire (65, 66) assess eating when exposed to food when physically satiated. The EAH behavioral paradigm has been the focus of considerable research to understand appetitive mechanisms. Children with higher FR and lower SR have higher caloric intake during this paradigm. (42, 67) Furthermore, greater caloric consumption during EAH has been linked with weight gain in children and adults. (68, 69) Attention bias, or the direction and strength of attention associated with a cue, can also be considered an assessment of FR. Attentional bias can be measured by reaction time, eye movements, or brain activity in response to salient stimuli, using event related potentials (ERPs; see (70) for review).

The remaining measures are self-report questionnaires that assess constructs closely related to FR. The Power of Food scale (PFS) (71) assesses appetite for high-palatable foods, and includes three levels of proximity, food available, food present and food tasted. The Food Craving Questionnaire (FCQ) (72, 73) assesses cravings using a multidimensional approach, with one subscale that queries about cues that may trigger eating. One study showed that individuals with higher FR, as assessed via physiological responsivity to a food cue, indicated higher subjective craving rating of those foods. (74) The Intuitive Eating Scale-2 (IES-2) assesses eating when physically hungry and stopping when full, with one subscale that focuses on reliance on hunger and satiety cues. The water load test (75, 76) evaluates how much water is consumed until perceived satiation and then again until maximum fullness. More research is needed to understand how these assessments are interrelated and what aspects of FR and SR are measured.

4. Development of Food Cue Responsiveness (FR)

Changes in FR over time is considered a product of both genetic and environmental influences. Emerging quantitative genetic modeling suggests that FR may be up to 75% heritable, (77) with the most evidence to date focusing on the “high risk” FTO alleles at rs9939609 in children. (78) Presence of this FTO minor allele at rs9939609 was associated with greater consumption during an EAH task compared with children without any risk allele in multiple samples. (79–81) As these studies were conducted in children, how the potential genetic risk for high FR changes over the life course remains to be fully elucidated.

Beyond genetic susceptibility, overeating develops through basic learning processes, including Pavlovian conditioning and operant conditioning. (82, 83) In today’s obesogenic food environment there are a plethora of opportunities to overeat through associations of cues in the environment with food, and over time, these cues can trigger responding (i.e., FR). Through Pavlovian conditioning, these food cues become directly associated with food intake and can elicit arousal, cravings, expectancies, thoughts, urges, motivation to eat. (84) There are also opportunities for operant conditioning, where the association of food seeking actions or eating are paired with the reinforcing effects of eating. (85) Importantly, these two learning processes act in concert (86) and the presentation of Pavlovian food cues can increase the strength or of operant eating or food-seeking through Pavlovian Instrumental Transfer. (87, 88)

Food cues can also acquire secondary reinforcing properties through their direct association with food. (89) Discriminative stimuli, those stimuli that are present when operant actions are reinforced, can increase operant responding by “setting the occasion” for the action–outcome relationship rather than eliciting or motivating behavior through their simple direct association with food. (90) Finally, with repetition and practice, stimuli associated with operant responding can eventually elicit the operant behavior directly. (91–93)

Increased FR may also provide opportunities for basic learning processes to take place. Ferriday and Brunstrom (94) suggest that one mechanism by which FR can increase consumption is by individuals planning to consumer larger amounts, resulting in overeating. Another potential mechanism linking FR and consumption is the association between FR and attention bias towards food cues, as there is some evidence that greater FR is associated with an increased attention bias toward foods. (95, 96) Increased attention to food cues may provide more opportunities for basic learning processes to occur, and lead to increased number of cues leading to consumption.

5. Development of Satiety Responsiveness (SR)

Conceptually, individuals will continue to eat in response to conditioned food-related cues until the intake is terminated by interoceptive satiety cues, (97) suggesting that SR is an inhibitory mechanism. SR develops early in life and is also considered an interaction of both genetic influences and factors in the environment. Several studies suggest that SR could be up to 63–72% heritable. (77, 98) Genetic mutations (such as in the leptin gene and melanocortin 4 receptor (MC4R) gene) are believed to contribute to OW/OB by disrupting satiety signaling. (99) One study evaluated the mechanisms by which a 16p11.2 deletion impacts BMI and found that altered SR preceded the development of obesity. (100) Research on FTO in children showed that the polymorphism rs9939609 is associated with reduced SR and that SR may mediate the relationship between FTO and BMI. (101) A more recent study (102) using participants from the same sample created a polygenetic risk score using 28 of 34 known obesity single-nucleotide polymorphisms (SNPs) identified in meta-analyses (103, 104) of children and adults. Results showed the polygenetic risk score to be positively associated with adiposity and negatively associated with SR, with SR mediating the association between the polygenetic risk score and adiposity. It is possible that decreased SR is a mechanism through which genetic risk accelerates weight gain and contributes to the development of OW/OB.

Some youth are born with this genetic risk factor of poorer SR. Beyond genetics, some data suggest that heightened SR during infancy may be promoted through breastfeeding. Research showed that breastfeeding during the first year of life (of at least 6 weeks in duration) is related to greater SR at 18–24 months (105) and that baby-led weaning compared to standard weaning is also related to greater SR at 18–24 months. (106) A study utilizing a sample from Amsterdam demonstrated that exclusively breastfed infants in the first four months have greater SR at age 5 than those who were introduced to solid foods during the first four months. (107) There is some conflicting research regarding breastfeeding, SR, and BMI such that one study found that breastfeeding exposure is related to higher SR at age 3–6, but not related to BMI change; (108) yet another study found breastfeeding intensity is related to obesity risk but not SR. (109)

It is also possible that factors in the environment can impact SR. Davidson and colleagues outline how interaction with the obesogenic environment can interfere with associative mechanisms that underlie the learned control of energy regulation. (110) Consistent with this analysis, overeating and the resulting excess weight gain is considered the result of physiological inhibitory signals (i.e. hormone cholecystokinin, leptin, Glucagon-like peptide-1) failing to suppress the capacity of cues in the environment (sight, smell, perceived palatability of a desirable food) from continuing to stimulate appetitive behaviors. (97) Excess energy intake and body weight gain is triggered by reducing the ability to inhibit appetitive and consummatory responses to food-related cues. (111–114) Practically, consumption of a westernized diet that is high in sugar and saturated fat may promote higher FR because a reduction in inhibitory stimulus control by satiety signals can enable food cues to be more excitatory within a Pavlovian framework. (110)

A number of factors in the environment can contribute to this mechanism, although a full review is available elsewhere. (110) Research in animals showed that consumption of non-caloric sweeteners can reduce the validity of sweet tastes as predictors of post-ingestive outcomes; (114, 115) however, data in humans tends to be more mixed. (116) Data suggest that SR depends, at least in part, on the hippocampus. (117) The hippocampus is involved in the encoding and retrieval of spatial relations among objects in the environment and the formation and recall of memories about events and facts. (118) However, more recently, the hippocampus has been associated with the utilization of hunger signals (119) and resolving “predictable ambiguities” when a single stimulus signals different outcomes dependent on other cues. (120) Data suggest that a diet that is high in sugar and saturated fat impacts the function of the blood brain barrier and the hippocampus itself. (121–123) Thus, the consumption of a western diet may promote overeating by affecting the hippocampus and other brain substrates that are involved in Pavlovian learning, ultimately reducing the ability of satiety cues to be inhibitory. (123, 124)

6. Relationship between appetitive traits, eating behavior, and body weight

In both children and adults, studies suggest that FR is associated with greater food consumption, while SR is associated with less food consumption. In children, higher FR is related to faster eating rate and increased energy intake during an EAH paradigm, while higher SR is associated with lower energy intake in the EAH paradigm and lower overall energy intake measured across 5 days. (42) The Gemini Birth Cohort collected data from over 1000 families with children 16 to 21 months old, and revealed that higher FR is associated with more frequent meals, but not larger meals. (125) Furthermore, it suggested that SR is inversely related to meal size (125) and is associated with consumption and post-meal satiety. (126) A study employing ecological momentary assessment found a significant positive relationship between food cues and snacking behavior, with FR moderating the relationship. (127)

Studies among toddlers show an inverse relationship between SR and energy intake during a lunch meal; but not during subsequent snack intake (EAH). (128) Similarly, adolescents classified as having greater SR (based on latent profile analyses) consume less food in an EAH paradigm. (129) There is also some evidence that SR, as measured by the Satiety Quotient, is related to subsequent ad libitum eating and self-reported food intake. (130, 131) These data provide a wide breadth of literature supporting the positive association between FR and eating behaviors and negative association between SR and eating behaviors across the lifespan.

FR is also consistently associated with higher body weight, while SR is inversely associated with BMI. (34, 109, 128, 132–135) Higher FR and lower SR are related to higher standardized child BMI, (136) with these associations demonstrated in multiple samples, including school aged Latino children, (137) preschool children of families of low socioeconomic status, (138) and Dutch children aged 6–7. (139) Similarly higher FR is associated with higher BMI in adults, (140) while SR in adults is inversely associated with BMI. (44, 141) Further, self-reported intuitive eating is consistently inversely associated with BMI. (142–149) Longitudinal investigations of these associations are more limited but reflect the same effect. Higher FR in children at the age of 6 is a significant predictor of a greater increase in BMI from age 6 to 8. (135) The same study showed that higher BMIz at age 4 predicted lower SR at age 6, and the increase in BMIz from age 4 to age 6 predicted decreased SR between ages 6 and 8. (135) In a sample children ages 7 to 9, one year changes in BMI were significantly predicted by scores on a relative reinforcing value of food task at baseline. (150) In infants, SR at age 3 months is inversely related to weight SD scores at age 15 months independent of weight at 3 months. (151) One study demonstrated that SR at age 2 is negatively associated with BMIz at age 4 controlling for birth weight Z score. (128) A separate study found that SR at 5–6 years is negatively associated with BMI at age 7–8 while controlling for birth weight. (152) Taken together, sufficient evidence suggests that SR is inversely related to BMI, while higher FR is associated with higher BMI. Longitudinal associations suggest the predictive power of FR and SR on BMI but additional studies are warranted.

Furthermore, data suggest that adults with OW/OB, compared to those with healthy weight, differentially respond to external food cues, with increases in both subjective ratings (increased desire to eat) and physiological responses (increased salivary response) to food cues. (153, 154) Our group demonstrated a stronger conditioned salivary response to innocuous food cues for individuals with obesity compared to those with healthy weight. (155) Several studies have shown that individuals with obesity, and those with binge eating, compared to those with healthy weight, display greater attention bias to food words and pictures. (156–164)

Emerging research also suggests that SR is related to weight loss. (165) Our group conducted a secondary data analysis among 150 children enrolled in family-based treatment for weight loss. (166) Latent class analyses revealed 3 trajectories of appetitive traits: high SR group (47.4%), high FR group (34.6%), and high emotional eating group (18.0%). (167) Interestingly, children in all three trajectories lost weight during treatment, however, only the children in the high SR group maintained their weight loss at the 12- and the 24-month follow-ups. Similarly, a study in adult men demonstrated that those with poorer SR lose less weight than those with high SR immediately following a 16-week randomized control trial of a satiating diet (higher protein; lower carbohydrate) to control diet (lower protein, higher carbohydrate) irrespective of group assignment. (165)

More recently, emerging research suggests that FR and SR may exist along the same continuum. Data in children and adults consistently show a significant negative relationship between FR and SR. (44, 168, 169) Additionally, more recent understandings suggest that the brain circuitry involved in both energy homeostasis and hedonic eating overlap and are less distinct than previously considered. (170) Thus, it is possible that these two appetitive traits may influence the impact of each other on overeating. This is similar to models in addiction research between reward (FR) and inhibition (SR). Considering this profile of appetitive traits, there may be intervention opportunities related to decreasing FR and increasing SR.

In summary, appetitive traits are associated with overeating and body weight. The BST outlined the interaction between appetitive traits, genetics and the environment. As part of this paper, we extended the BST by elucidating pathways showing how the environment interacts with individual appetitive characteristics. Figure 1 presents our extension of the original BST (solid lines) with the addition of these pathways that further explain the mechanisms of how genetics and environment interact with appetitive traits.

Figure 1:

Extension of the Behavioral Susceptibility Theory which elaborates on the mechanisms of the development of higher food cue responsiveness (FR) and lower satiety responsiveness (SR) and relation to overeating and weight gain (adapted from Llewellyn & Fildes, 2017 (35))

7. Treatment development of a program designed to target FR and SR

Based on the theory and existing literature, we believe that FR and SR are emerging, important variables related to overeating and OW/OB and could be potential mechanisms for the development of a novel model for weight loss. We developed a treatment program that specifically targets both FR and SR, called Regulation of Cues (ROC). To target improvement in SR, we adapted Appetite Awareness Training (AAT). AAT focuses on rating hunger and satiety and learning to stop eating when physically full, and has been tested in children and adults. (171–173) To target decreases in FR, we developed a cue-exposure treatment for food (CET-Food) to reduce eating in response to food cues while sated. CET-Food involves exposure to food without consuming it while sated. CET-Food can teach people to resist eating when they are not physically hungry even though food is present by improving inhibitory learning. (174–177) ROC incorporates these two skill sets with psychoeducation, coping skills, experiential learning and parenting skills (when applied with children). Our pilot studies to date have shown feasibility, acceptability and initial efficacy with children (178, 179) and adults. (180) Large randomized control trials are needed to establish ROC as an evidence-based treatment for obesity, overeating, and/or binge eating. Two large studies among adults in our lab are currently underway (NCT02516839, NCT03678766 (181)) and will help elucidate whether ROC is an effective weight-loss treatment targeting the mechanisms of FR and SR. To date, it is unclear whether ROC can be a stand-alone treatment or if it can be used to improve the potency of BWL. It is also unclear how to optimize inhibitory learning with food cues. (175–177)

8. Conclusions

In summary, we believe that FR and SR are important variables related to overeating and weight gain, and it is possible that these two appetitive characteristics may exist along the same continuum. We have outlined the methods for the development of these appetitive traits and how they contribute to overeating in the current food environment. In particular, we have outlined options for intervening with these two mechanisms, by focusing on improving SR by training participants to respond to their appetite and on changing responses to FR by training inhibitory learning. We have developed the ROC program which targets these two mechanisms and are testing this program in larger trials. We believe that by targeting mechanisms of overeating, we may be able to develop more potent and durable interventions to decrease overeating and weight.