OPEN Citation: Nutition & Diabetes (2014) 4, e140; doi:10.1038/nutd.2014.37 2014 Macmillan Publishes Limited All ights eseved 2044-4052/14 www.natue.com/nutd ORIGINAL ARTICLE Eating behavio taits and sleep as deteminants of weight loss in oveweight and obese adults M-L Filiatault 1,2, J-P Chaput 3, V Dapeau 2,4 and A Temblay 1,2 OBJECTIVE: To examine the associations between eating behavio taits and weight loss accoding to sleep quality and duation in adults enolled in common weight-loss inteventions. METHODS: Paticipants included oveweight and obese men and women (n = 150) (mean ± s.d. age, 38.8 ± 8.6 yeas; mean ± s.d. body mass index (BMI), 33.3 ± 3.5 kg m 2 ) who wee subjected to a dietay intevention ove a peiod of 12 16 weeks. Anthopometic measuements, eating behavio taits (Thee-Facto Eating Questionnaie), sleep quality (total Pittsbugh Sleep Quality Index (PSQI) scoe) and sleep duation (hous pe night, self-epoted fom the PSQI) wee assessed at both baseline and post intevention. Linea egession analysis was used to quantify the elationships between eating behavio taits and changes in anthopometic makes fo all subjects and by sleep categoies (shot sleep: o7 h pe night vs ecommended sleep: 7 h pe night; poo sleep quality: 5 PSQI scoe vs good sleep quality: o5 PSQI scoe). We adjusted fo age, sex and baseline BMI in analyses. RESULTS: Baseline eating behavio taits wee modest pedictos of weight-loss success, but they wee all significantly associated with thei changes ove the weight-loss intevention (Po0.01). The diet intevention induced significant changes in eating behavio taits and even moe fo those having a non-favoable eating behavio pofile at baseline. We obseved that changes in flexible contol and stategic dieting behavio wee constantly negatively associated with changes in body weight and fat mass (Po0.05) fo ecommended duation sleepes. The change in situational susceptibility to disinhibition was positively associated with the change in fat mass and body weight fo those having healthy sleeping habits (Po0.05). Fo poo quality sleepes, the change in avoidance of fattening foods was negatively associated with changes in adiposity (Po0.05). CONCLUSION: Eating behavio taits and sleep may act togethe to influence the outcome of weight-loss pogams. Nutition & Diabetes (2014) 4, e140; doi:10.1038/nutd.2014.37; published online 20 Octobe 2014 INTRODUCTION A lage body of evidence shows that sleeping habits have an influence on body weight egulation. 1 Moe specifically, studies epot that shot sleep duation and poo sleep quality ae both associated with weight gain and obesity. 2 Results of ecent studies also suggest that lack of sufficient sleep can impact the success of weight-loss inteventions. Fo example, Nedeltcheva et al. 3 conducted a cossove study in oveweight adults andomly assigned to eithe 5.5 o 8.5 h pe night fo 2 weeks in addition to modeate caloic estiction in a laboatoy setting. They obseved that sleep estiction led to less body fat loss and geate hunge compaed with the expeimental condition with adequate sleep. We ecently eplicated these findings in a eal-life setting and showed that both sleep duation (hous pe night) and sleep quality (total Pittsbugh Sleep Quality Index (PSQI) scoe) pedicted fat mass loss duing common dietay inteventions in oveweight and obese adults. 4 In paticula, we obseved that a 1-h change in sleep duation was associated with a 0.7 kg change in fat mass afte adjustment fo covaiates. Although the latte study also suggests that poo sleeping habits can undemine dietay effots to educe adiposity, the inteaction between sleep and eating behavios with egad to the magnitude of body weight, fat mass and waist cicumfeence (WC) loss is unknown. A bette undestanding of the connections between sleep and eating behavios is needed if we want to bette tailo weight-loss pogams, avoid patients disappointment and maximize success. The objective of the pesent study was theefoe to examine the elationship between eating behavios, sleep quality/quantity and loss of body weight, fat mass and WC in adults involved in dietbased weight-educing pogams. We hypothesized that having good sleeping habits togethe with adequate eating behavio taits would pomote geate weight and fat loss in adults undegoing dietay inteventions. SUBJECTS AND METHODS Subjects Fou weight-educing inteventions pefomed at Laval Univesity (Quebec City, Canada) wee pooled togethe to maximize powe. Details about these studies have been peviously published. 5 8 Biefly, all the studies included in the pesent aticle had included a standadized diet aimed at pomoting a negative enegy balance while subjects wee eithe taking a placebo o a supplement of an active agent. In the pesent study, only data fom paticipants enolled in the contol (placebo) goup of these studies wee consideed fo statistical analyses to avoid the influence of the active supplements on the pimay outcome measues (i.e., weight, fat mass and WC losses). Oveall, 150 healthy oveweight and obese men and women 1 Depatment of Kinesiology, Laval Univesity, Quebec City, Quebec, Canada; 2 Quebec Heat and Lung Reseach Institute, Laval Hospital, Quebec City, Quebec, Canada; 3 Healthy Active Living and Obesity Reseach Goup, Childen s Hospital of Easten Ontaio Reseach Institute, Ottawa, Ontaio, Canada and 4 Depatment of Physical Education, Laval Univesity, Quebec City, Quebec, Canada. Coespondence: Pofesso A Temblay, Depatment of Kinesiology (PEPS), Faculty of Medicine, Laval Univesity, 2300 de la Teasse, Quebec City, Quebec G1V 0A6, Canada. E-mail: angelo.temblay@kin.ulaval.ca Received 24 July 2014; evised 22 August 2014; accepted 4 Septembe 2014
2 Eating behavios, sleep and weight loss aged between 20 and 55 yeas wee subjected to a diet intevention supevised by a dietician. In most of these inteventions, 5 7 each individual eceived a pesonalized diet plan tageting 2085 2919 kj d 1 (500 700 kcal d 1 ) decease in enegy intake. In the last study, 8 paticipants wee assigned a contol diet based on the Canada s Food Guide to pomote a negative enegy balance. 9 Depending on the study, the length of the intevention vaied between 12 and 16 weeks. The following inclusion citeia wee also consideed: absence of pegnancy, beast-feeding o menopause (detemined by the cessation of menstuation), stable body weight, body mass index (BMI) between 27 and 42 kg m 2, o3 peiods of 20 min of physical activity pe week, no use of medication and/o all kinds of supplements that could intefee with the study s objective, without comobidities, consumption of 10 alcoholic beveages pe week o no moe than 2 dinks pe day depending on the study, and consumption of 5 cups of coffee pe day. Evey 2 weeks (fom the beginning until the completion of the study), paticipants met thei assigned dietician. Compliance was veified by compaing the diet pescibed (total daily enegy intake and maconutient composition) to the actual diet composition of the paticipants which was checked at evey visit using 24 h food ecalls. Study potocols wee appoved by the Laval Univesity Ethics Committee. Witten infomed consent was obtained fom all paticipants. Anthopometic and body composition measuements Height was measued to the neaest 0.1 cm (without shoes and light clothes) using a standad stadiomete and body weight was measued to the neaest 0.1 kg using eithe a standad beam scale o a digital panel indicato scale. BMI was calculated as body weight divided by height squaed (kg m 2 ). WC was measued with an anthopometic tape at midpoint between the lowe bode of the last ib and the uppe bode of the iliac cest. Body fat mass was eithe measued by dual-enegy X-ay absoptiomety 6 8 (GE Medical Systems Luna, Diegem, Belgium) o by the undewate weighing technique 5 with the use of the Sii fomula 10 to estimate the pecentage of body fat fom body density. All measuements wee pefomed in the same way and in accodance with standadized pocedues at both baseline and at the end of the dietay inteventions. Sleep assessment At both baseline and at the end of the intevention, each paticipant had to complete the PSQI, a self-ated questionnaie that assesses sleep quality and distubances ove the peceding 1-month time peiod. 11 This questionnaie contains 19 individual items and geneates seven component scoes: subjective sleep quality, sleep latency, sleep duation, habitual sleep efficiency, sleep distubances, use of sleeping medication and daytime dysfunction. The sum of the component scoes yields one total scoe with a maximum of 21. A total PSQI scoe geate than 5 is highly sensitive and specific in distinguishing good fom poo sleepes and has been validated in a numbe of populations. 12 In the context of ou study, both sleep quality (total PSQI scoe) and sleep duation (self-epoted fom the PSQI) wee used fo statistical analyses. Shot sleep duation was defined as o7 h pe night and ecommended sleep duation as 7 h pe night accoding to the median value in this sample and in ageement with the liteatue on the topic. 1,2 Futhemoe, poo sleepes wee those having a PSQI scoe 5 and good sleepes those having a PSQI scoe o5. Eating behavio taits assessment Paticipants had to complete a Fench vesion of the 51-item Thee-Facto Eating Questionnaie at baseline and at the end of the intevention. 13 This questionnaie measues thee dimensions of human eating behavio, that is, cognitive dietay estaint (intent to contol food intake), disinhibition (oveconsumption of food in esponse to cognitive o emotional cues) and susceptibility to hunge (food intake in esponse to feelings and peceptions of hunge). It is also possible to distinguish two types of cognitive dietay estaint behavios: flexible contol and igid contol of eating. 14 Othe moe specific eating behavios have been identified as subscoes in each of the thee main dimensions, which offes a moe complex pictue of eating behavio in humans. 15 The Thee-Facto Eating Questionnaie is a epoducible and valid tool 13,16,17 and the Fench vesion has been peviously validated. 18 We examined each tait (baseline and change), including thei espective subscoes, as it has been poposed that measuing domain-specific factos ae moe likely to help in claifying the complexity of eating behavios. 15 Diet assessment At both baseline and at the end of the intevention, paticipants had to fill out a 3-day food ecod fo 2 weekdays and 1 weekend day. A dietician explained to each paticipant how to accuately complete the dietay ecod including how to quantify the foods consumed. This method of dietay assessment has been shown to povide a elatively eliable measue of diet in this population. 19 Mean daily enegy intake was estimated by a dietician using a computeized vesion of the Canadian Nutient File. 20 Statistical analysis Using egession analysis, we fist quantified the elationships between baseline eating behavio taits and changes in body weight, fat mass and WC. Secondly, we veified the pevious elationships by sleep categoies (shot sleep: o7 h pe night vs ecommended sleep: 7 h pe night; poo sleep quality: 5 PSQI scoe vs good sleep quality: o5 PSQI scoe). The same analyses wee futhe examined by using the changes in eating behavio taits. Models wee all adjusted fo age, sex and baseline BMI. Multiple analysis of vaiance was used to monito the effect of the dietay intevention on anthopometic makes, eating behavio taits and sleep. Linea egession analysis was used to quantify the elationships between baseline eating behavio taits and changes in eating behavio taits in esponse to the diet intevention. The same analysis was caied out fo sleep quality and duation. Adiposity changes by tetiles of change in sleep (delta PSQI scoe (post-pé); Δ+: impovement, 0: no change, Δ : deteioation) and anthopometic makes wee pefomed using analysis of vaiance adjusting fo age, sex and baseline BMI. A Tukey post hoc test was pefomed to detemine which goups wee significantly diffeent. A two-tailed P value of less than 0.05 was consideed to indicate statistical significance. Data ae pesented as means ± s.d. All statistical analyses wee pefomed using the JMP vesion 9 pogam (SAS Institute, Cay, NC, USA). RESULTS Desciptive chaacteistics of paticipants ae pesented in Table 1. Paticipants lost on aveage 3.7 ± 3.1 kg of body weight, 2.8 ± 2.8 kg of fat mass and 5.0 ± 3.5 cm of WC ove the intevention peiod. Changes in enegy intake wee veified with the 3-day food ecods and wee of 522 ± 465 kcal pe day on aveage. Paticipants complied at 78% with thei enegy-esticted diet, as assessed by the diffeence between total enegy intake actually consumed and the pesciption. The aveage length of the intevention was 14.0 ± 1.7 weeks. Table 2 shows that body weight, fat mass and WC wee all significantly educed in esponse to the dietay intevention (Po0.01). Sleep quality was significantly impoved in esponse to the weight-loss pogam, wheeas post-intevention sleep duation was not significantly diffeent fom baseline. As shown in Table 2, the diet intevention had significant effects on eating behavio taits fo all subjects and on sleep quality (Po0.01). Baseline eating behavio taits wee all significantly associated with thei changes ove the intevention fo all subjects (Table 3). Howeve, as shown in Table 3, when pefoming the same analysis by sleep categoies, no significant elationship was obseved between both baseline disinhibition and emotional susceptibility to disinhibition with the changes in these behavios fo the poo sleep quality categoy. Likewise, no significant association was obseved fo baseline situational susceptibility to disinhibition with its change duing weight loss fo those having inadequate sleeping habits (Table 3). Afte adjusting fo covaiates, baseline situational susceptibility to disinhibition was negatively associated with changes in body weight and WC fo all subjects ( = 0.21; = 0.18; Po0.05). Fo good and poo quality sleepes, baseline susceptibility to hunge was associated with fat and weight loss, espectively ( = 0.33; = 0.29; Po0.05) (Table 4a). Fo good quality sleepes, baseline flexible contol and extenal locus of hunge wee significantly associated with the change in fat mass ( = 0.06; = 0.27; Po0.05) (Table 4a). Fo those within the ecommended sleep duation categoy, baseline situational susceptibility to Nutition & Diabetes (2014) 1 8 2014 Macmillan Publishes Limited
Eating behavios, sleep and weight loss 3 Table 1. Baseline chaacteistics of paticipants Vaiables All subjects Poo sleep quality (PSQI scoe 5) Good sleep quality (PSQI scoe o5) Shot sleep (o7 h pe night) Recommended sleep ( 7 h pe night) Mean ± s.d. Mean ± s.d. Mean ± s.d. Mean ± s.d. Mean ± s.d. Age (yeas) 38.8 ± 8.6 40.3 ± 8.8 37.8 ± 8.3 42.1 ± 8.0 37.6 ± 8.6** Sex (n) Men 83 38 45 34 49 Women 67 24 43 16 51 Body weight (kg) 94.4 ± 15.4 94.2 ± 13.0 94.9 ± 16.8 94.6 ± 11.1 94.2 ± 16.5 BMI (kg m 2 ) 33.3 ± 3.5 33.2 ± 3.8 33.6 ± 3.4 33.3 ± 3.3 33.4 ± 3.6 Waist cicumfeence (cm) 106.0 ± 11.1 106.3 ± 10.2 106.2 ± 11.8 106.4 ± 8.8 105.8 ± 11.8 Fat mass (kg) 36.6 ± 8.8 36.34 ± 9.5 37.3 ± 8.6 35.9 ± 8.8 37.0 ± 8.8 Delta weight (kg) 3.7 ± 3.14 4.5 ± 2.8 3.2 ± 3.3* 4.0 ± 2.3 3.6 ± 3.4 Delta fat mass (kg) 2.8 ± 2.8 3.6 ± 2.2 2.4 ± 3.0* 3.1 ± 2.3 2.7 ± 2.9 Delta waist cicumfeence (cm) 5.0 ± 3.5 5.6 ± 3.6 4.7 ± 3.4 4.8 ± 2.8 5.0 ± 3.7 Sleep duation (h) 7.2 ± 0.9 6.7 ± 0.9 7.6 ± 0.7** 6.0 ± 0.5 7.6 ± 0.6** PSQI (total scoe) 4.6 ± 2.6 7.1 ± 2.2 2.9 ± 1.0** 6.8 ± 2.9 3.9 ± 2.1** TFEQ Cognitive dietay estaint (0 21) 7.4 ± 3.8 7.0 ± 3.8 7.5 ± 3.8 6.9 ± 3.7 7.5 ± 3.9 Flexible contol (0 7) 2.1 ± 1.6 1.7 ± 1.6 2.2 ± 1.5 1.8 ± 1.5 2.2 ± 1.6 Rigid contol (0 7) 2.4 ± 1.7 2.6 ± 1.5 2.3 ± 1.7 2.6 ± 1.7 2.4 ± 1.7 Stategic dieting behavio (0 4) 0.6 ± 1.0 0.4 ± 0.9 0.7 ± 1.0 0.4 ± 0.8 0.7 ± 1.0 Attitude to self-egulation (0 5) 2.5 ± 1.3 2.7 ± 1.3 2.5 ± 1.3 2.6 ± 1.2 2.6 ± 1.3 Avoidance of fattening foods (0 4) 2.0 ± 1.2 1.8 ± 1.2 2.1 ± 1.2 1.8 ± 1.1 2.0 ± 1.2 Disinhibition (0 16) 8.4 ± 3.1 8.9 ± 2.9 8.1 ± 3.2 8.6 ± 2.8 8.3 ± 3.1 Habitual susceptibility (0 5) 1.6 ± 1.4 1.6 ± 1.3 1.6 ± 1.5 1.4 ± 1.2 1.6 ± 1.5 Emotional susceptibility (0 3) 1.7 ± 1.2 1.9 ± 1.2 1.6 ± 1.3 1.9 ± 1.2 1.7 ± 1.2 Situational susceptibility (0 5) 3.4 ± 1.3 3.5 ± 1.2 3.3 ± 1.3 3.4 ± 1.1 3.3 ± 1.3 Susceptibility to hunge (0 14) 6.1 ± 3.4 6.7 ± 3.2 5.7 ± 3.4 5.7 ± 3.1 6.2 ± 3.4 Intenal locus of hunge (0 6) 2.2 ± 1.7 2.4 ± 1.6 2.1 ± 1.8 1.9 ± 1.6 2.3 ± 1.7 Extenal locus of hunge (0 6) 2.8 ± 1.6 3.1 ± 1.7 2.6 ± 1.6 2.7 ± 1.6 2.8 ± 1.7 Enegy intake (kcal pe day) 2524 ± 584 2535 ± 536 2512 ± 586 2631 ± 587 2478 ± 564 Abbeviations: BMI, body mass index; PSQI, Pittsbugh Sleep Quality Index; TFEQ, Thee-Facto Eating Questionnaie. Mean values wee significantly diffeent among sleep categoies *Po0.05, **Po0.01. Analysis of compaison among sleep categoies wee pefomed using analysis of vaiance. disinhibition was negatively associated with changes in body weight and WC (Po0.05) (Table 4b). Fo all subjects, we obseved that the changes in cognitive dietay estaint, flexible contol and stategic dieting behavio wee constantly significantly negatively associated with changes in weight and fat mass fo all subjects (Po0.05) (Table 5a). The change in situational susceptibility to disinhibition was positively associated with changes in weight and fat mass fo all subjects (Po0.05) (Table 5a). By pefoming the analysis by sleep categoies, we obseved within the sleep quality categoy that the change in situational susceptibility to disinhibition was positively associated with changes in body weight and fat mass fo good sleepes (Po0.05) (Table 5b). Fo poo sleepes, the change in avoidance of fattening foods was negatively associated with changes in fat mass and WC (Po0.05) (Table 5b). Within the ecommended sleep duation subcategoy, we obseved that changes in flexible contol and stategic dieting behavio wee constantly negatively associated with changes in body weight and fat mass (Po0.05) (Table 5c). Relationships between the change in flexible contol with the change in WC was not influenced by sleep duation (Po0.05) (Table 5c). The change in situational susceptibility to disinhibition was positively associated with the change in fat mass and body weight within the ecommended sleep duation subcategoy (Po0.05) (Table 5c). Finally, we compaed the changes in anthopometic makes by tetiles of change in sleeping habits (thee goups wee fomed; impovement, no change and deteioation). Thee was a significant elationship in those who impoved and deteioated thei sleep quality (Po0.05) with the change in WC. Adjusted WC means wee 5.9 cm vesus 3.7 cm in those who impoved and deteioated thei sleep quality, espectively (Figue 1a). DISCUSSION Results fom the pesent study show that eating behavio taits ae elated with weight-loss outcome and that sleep quality and duation seem to influence these elationships. Cognitive dietay estaint and its elated subscoes all inceased in esponse to the diet intevention. Disinhibition and its elated subscoes all deceased in esponse to the weight-loss pogam. Those vaiations in eating behavio taits ae associated with weightloss success and maintenance. 21 Susceptibility to hunge and its elated subcoes all diminished as a teatment effect. Those changes benefit weight loses because highe levels of susceptibility to hunge cause a poblem fo subsequent oveconsumption of enegy and ae positively associated with the level of obesity. 22,23 Ou esults ae in ageement with those of Teixeia et al. 24 because baseline eating behavio taits wee modest pedictos of weight-loss success. This emphasizes the need to conside the effect of the dietay intevention on eating behavio taits to develop stategies fo long-tem teatment of obesity. Indeed, baseline situational susceptibility to disinhibition was negatively associated with weight-loss outcomes and the change in the latte evealed an opposite tend. One possible explanation might be that individualized dietay advice poposed by the dietician to highly disinhibited individuals at baseline favoed highe contol 2014 Macmillan Publishes Limited Nutition & Diabetes (2014) 1 8
4 Eating behavios, sleep and weight loss ove food intake esulting in bette compliance to the diet plan. Indeed, the diet intevention induced significant changes in eating behavio taits in the expected diection with weight loss fo all paticipants. Thee wee significant impovements in disinhibition taits fo all paticipants and even moe fo those having a nonfavoable eating behavio pofile at baseline. As fo cognitive dietay estaint, even though this eating behavio tait inceased in esponse to the intevention, only flexible contol pedicted adiposity changes. By looking at changes in eating behavio taits by sleep categoies, esults of the pesent study show that impovement in flexible contol of eating was associated with geate body weight, Table 2. Teatment effects on anthopometic makes, eating behavio taits and sleeping habits Vaiables Befoe Afte Anthopometic makes Body weight (kg) 94.4 ± 15.4 89.4 ± 15.3** Fat mass (kg) 36.7 ± 8.8 33.1 ± 9.5** Waist cicumfeence (cm) 106.0 ± 11.1 99.89 ± 11.5** Eating behavio taits Cognitive dietay estaint 7.37 ± 3.86 12.27 ± 4.03** Flexible contol 2.07 ± 1.56 4.14 ± 1.74** Rigid contol 2.40 ± 1.66 3.52 ± 1.74** Stategic dieting behavio 0.62 ± 0.97 1.83 ± 1.34** Attitude to self-egulation 2.54 ± 1.29 3.47 ± 1.32** Avoidance of fattening foods 1.95 ± 1.21 2.84 ± 1.04** Disinhibition 8.38 ± 3.06 6.41 ± 2.79** Habitual susceptibility 1.57 ± 1.41 0.97 ± 1.00** Emotional susceptibility 1.70 ± 1.23 1.19 ± 1.24** Situational susceptibility 3.37 ± 1.27 2.39 ± 1.44** Susceptibility to hunge 6.13 ± 3.39 3.43 ± 2.84** Intenal locus of hunge 2.19 ± 1.73 1.03 ± 1.38** Extenal locus of hunge 2.84 ± 1.64 1.45 ± 1.40** Sleeping habits Sleep quality 4.59 ± 2.61 3.74 ± 2.13** Sleep duation 7.23 ± 0.90 7.22 ± 1.36 Values ae means ± s.d. Mean values wee significantly diffeent fom befoe teatment *Po0.05, **Po0.01. Teatment effects wee analyzed using multiple analysis of vaiance. fat mass and WC losses in individuals having good sleep duation. In paticula, we obseved that a positive incement of one unit in the level of flexible contol was associated with a eduction of 0.57 kg in body weight afte adjustment fo covaiates. Moe flexible contol was also associated with WC eduction fo shot duation and poo quality sleepes. Inceased level of flexible contol ove eating is associated with lowe body weight and is pedictive of long-tem weight-loss outcomes. 22 Although speculative, one explanation could be that eating behavio tait dominated the influence of sleeping habits on thei effect on visceal adiposity. An incease in stategic dieting behavio (a subscoe of cognitive dietay estaint) was associated with body weight and fat loss solely within the ecommended sleep duation subcategoy. In a longitudinal study quantifying the associations between the change in eating behavios and body weight changes ove 6 yeas, women elied moe on stategic dieting behavio and avoided moe fattening foods as a way to diminish food intake. 25 Although the pesent study was not sufficiently poweed to detect diffeences between men and women, futue studies should investigate whethe the associations between eating/sleeping behavios and body weight, fat and WC loss ae gende specific. An incease in avoidance of fattening foods (a subscoe of cognitive dietay estaint) was associated with moe body fat mass and WC loss. The late elationship was only obseved in the poo quality sleepes subcategoy. Moeove, positive associations between the changes in situational susceptibility to disinhibition and body weight and fat loss wee solely seen in those having healthy sleeping habits. Among eating behavio taits that wee egulaly significantly associated with body weight and adiposity loss, situational susceptibility to disinhibition seems to have the most ponounced impact on body weight in ou cohot. Indeed, a positive incement of one unit in the level of this vaiable was associated with an incease of 0.67 kg of body weight afte adjustment fo covaiates among the ecommended sleep duation subcategoy. On the othe hand, a positive incement of one unit in the levels of stategic dieting behavio and flexible contol was associated with a decease of 0.63 kg and 0.57 kg of body weight, espectively, afte adjustment fo covaiates also among the ecommended sleep duation subcategoy. This is in ageement with othe eseach that has identified the disinhibition subscale to be the tait most pesistently associated with obesity and highe enegy consumption. 26 Table 3. Associations between baseline eating behaviou taits and changes of these vaiables in esponse to the diet intevention in all subjects and subjects classified accoding to sleep quality o sleep duation Vaiables All subjects Poo sleep quality (PSQI scoe 5) Good sleep quality (PSQI scoe o5) Shot sleep (o7 h pe night) Recommended sleep ( 7 h pe night) Cognitive dietay estaint 0.47** 0.57** 0.40** 0.50* 0.46** Flexible contol 0.48** 0.58** 0.34** 0.48* 0.48** Rigid contol 0.47** 0.49** 0.46** 0.24* 0.47** Stategic dieting behavio 0.52** 0.53** 0.51** 0.55** 0.51** Attitude to self-egulation 0.51** 0.65** 0.45** 0.68** 0.46** Avoidance of fattening foods 0.63** 0.71** 0.58** 0.63** 0.61** Disinhibition 0.50** 0.22 0.60** 0.53** 0.48** Habitual susceptibility 0.71** 0.58** 0.76** 0.85** 0.68** Emotional susceptibility 0.38** 0.27 0.42** 0.43* 0.36** Situational susceptibility 0.34** 0.06 0.52** 0.35 0.35** Susceptibility to hunge 0.59** 0.49** 0.69** 0.62** 0.60** Intenal locus of hunge 0.63** 0.55** 0.71** 0.77** 0.62** Extenal locus of hunge 0.64** 0.58** 0.72** 0.55** 0.66** Abbeviation: PSQI, Pittsbugh Sleep Quality Index. Relationship was statistically significant *Po0.05, **Po0.01. and P values ae fo the elationship between baseline eating behavio taits and changes of these in esponse to the diet intevention. The associations between baseline eating behavio taits and changes of these in esponse to the diet intevention wee tested using linea egession analysis. Nutition & Diabetes (2014) 1 8 2014 Macmillan Publishes Limited
Eating behavios, sleep and weight loss Table 4. Associations between baseline eating behavio taits and changes in mophological vaiables in esponse to the diet intevention in subjects classified accoding to sleep quality (a) o sleep duation (b) 5 (a) Baseline eating behavio taits Delta fat mass Delta body weight Delta WC PSQI 5 PSQIo5 PSQI 5 PSQIo5 PSQI 5 PSQIo5 Cognitive dietay estaint 0.06 0.09 0.09 0.09 0.04 0.11 Flexible contol 0.13 0.25* 0.13 0.21 0.12 0.22 Rigid contol 0.12 0.01 0.001 0.03 0.16 0.03 Stategic dieting behavio 0.16 0.02 0.12 0.04 0.10 0.14 Attitude to self-egulation 0.01 0.02 0.02 0.09 0.17 0.06 Avoidance of fattening foods 0.15 0.03 0.14 0.03 0.04 0.01 Disinhibition 0.19 0.01 0.20 0.06 0.01 0.07 Habitual susceptibility 0.11 0.03 0.15 0.02 0.03 0.14 Emotional susceptibility 0.04 0.004 0.002 0.09 0.12 0.16 Situational susceptibility 0.20 0.15 0.25 0.19 0.01 0.21 Susceptibility to hunge 0.20 0.33* 0.29* 0.14 0.20 0.16 Intenal locus of hunge 0.19 0.24 0.27 0.004 0.02 0.09 Extenal locus of hunge 0.15 0.27* 0.21 0.11 0.14 0.05 (b) Baseline eating behavio taits Delta fat mass Delta body weight Delta WC o7 h pe night 7 h pe night o7 h pe night 7 h pe night o7 h pe night 7 h pe night Cognitive dietay estaint 0.21 0.11 0.28 0.01 0.02 0.04 Flexible contol 0.17 0.07 0.28 0.05 0.25 0.003 Rigid contol 0.03 0.01 0.04 0.04 0.02 0.01 Stategic dieting behavio 0.17 0.05 0.15 0.09 0.13 0.07 Attitude to self-egulation 0.31 0.03 0.28 0.05 0.18 0.05 Avoidance of fattening foods 0.001 0.13 0.11 0.10 0.07 0.06 Disinhibition 0.14 0.04 0.25 0.08 0.08 0.04 Habitual susceptibility 0.14 0.05 0.23 0.02 0.22 0.12 Emotional susceptibility 0.01 0.04 0.18 0.07 0.11 0.15 Situational susceptibility 0.13 0.20 0.01 0.22* 0.01 0.23* Susceptibility to hunge 0.05 0.13 0.14 0.002 0.13 0.01 Intenal locus of hunge 0.12 0.07 0.24 0.06 0.10 0.06 Extenal locus of hunge 0.02 0.13 0.21 0.01 0.18 0.07 Abbeviations: PSQI, Pittsbugh Sleep Quality Index; WC, waist cicumfeence. Relationship was statistically significant *Po0.05, **Po0.01. and P values ae fo the elationship between baseline eating behavio taits and anthopomety makes accoding to sleep quality o duation categoies. The associations between baseline eating behavio taits and anthopometic makes accoding to sleep categoies wee tested using linea egession analysis adjusting fo age, sex and baseline body mass index. Bold indicates significant esults. The mechanisms that can explain why insufficient sleep may impede the success of dietay inteventions ae lagely unknown. Poposed explanations ae that highe ghelin concentations geneally obseved in shot sleepes may facilitate the etention of fat. 27,28 Additionally, the inceased dive to eat that accompanies lack of sleep can cetainly compomise adheence to caloic estiction because of the fact that dieting also inceases hunge. Killgoe et al. 29 showed that shot sleep duation might make people moe vulneable to hedonic food consumption. Thee was a significant positive elation between the change in situational susceptibility to disinhibition and the change in total enegy intake within all sleep categoies in the pesent study. Geate scoes in situational susceptibility to disinhibition wee elated to moe enegy consumed. Inteestingly, the change in this paticula tait esulted in food intake modifications and thus weight and fat mass change uniquely among those having healthy sleeping habits. A ecent study by Dweck et al. 30 examined the ole of stess and emotional eating (as assessed by the Dutch Eating Behavio Questionnaie) 31 in shot and nomal sleepes. They obseved that poo sleep quality was associated with high scoes of emotional and extenal eating but not of dietay estaint. Shot sleep duation was also associated with highe food consumption among stessed emotional eates vesus nomal sleepes. The authos poposed the same conclusion as Chaput et al. 32 in that the effects of sleep depivation on food consumption ae dependent on individual chaacteistics. Anothe study indicated that poo sleep quality was associated with an incease in hunge, disinhibition, dietay estaint and emotional susceptibility to disinhibition wheeas sleep duation was not. 33 Thus, poo sleep quality could be an indicato o make of poblematic eating behavios. Ou findings agee with esults obtained in pevious studies showing that poo sleeping habits could attenuate dietay effots 2014 Macmillan Publishes Limited Nutition & Diabetes (2014) 1 8
6 Eating behavios, sleep and weight loss Table 5. Associations between changes in eating behavio taits and changes in mophological vaiables in esponse to the diet intevention fo all paticipants (a) and in paticipants classified accoding to thei sleep quality (b) o sleep duation (c) (a) Delta eating behavio taits Delta fat mass Delta body weight Delta WC Cognitive dietay estaint 0.24* 0.23* 0.18 Flexible contol 0.25* 0.32** 0.30** Rigid contol 0.06 0.03 0.01 Stategic dieting behavio 0.22* 0.28** 0.19* Attitude to self-egulation 0.06 0.02 0.02 Avoidance of fattening foods 0.20* 0.18 0.03 Disinhibition 0.11 0.16 0.02 Habitual susceptibility 0.10 0.15 0.05 Emotional susceptibility 0.03 0.03 0.01 Situational susceptibility 0.24* 0.22* 0.09 Susceptibility to hunge 0.07 0.17 0.08 Intenal locus of hunge 0.06 0.11 0.03 Extenal locus of hunge 0.11 0.21* 0.05 (b) Delta eating behavio taits Delta fat mass Delta body weight Delta WC PSQI 5 PSQIo5 PSQI 5 PSQIo5 PSQI 5 PSQIo5 Cognitive dietay estaint 0.21 0.18 0.21 0.15 0.13 0.11 Flexible contol 0.29 0.12 0.30 0.19 0.32* 0.12 Rigid contol 0.05 0.09 0.10 0.04 0.01 0.02 Stategic dieting behavio 0.23 0.14 0.23 0.21 0.19 0.10 Attitude to self-egulation 0.09 0.07 0.07 0.01 0.15 0.04 Avoidance of fattening foods 0.32* 0.14 0.24 0.09 0.30* 0.05 Disinhibition 0.11 0.18 0.11 0.22 0.24 0.13 Habitual susceptibility 0.02 0.16 0.06 0.18 0.16 0.01 Emotional susceptibility 0.05 0.05 0.05 0.09 0.14 0.08 Situational susceptibility 0.18 0.34* 0.16 0.31* 0.09 0.21 Susceptibility to hunge 0.08 0.06 0.19 0.15 0.05 0.01 Intenal locus of hunge 0.08 0.07 0.12 0.10 0.04 0.06 Extenal locus of hunge 0.11 0.11 0.23 0.18 0.01 0.07 (c) Delta eating behavio taits Delta fat mass Delta body weight Delta WC o7 h pe night 7 h pe night o7 h pe night 7 h pe night o7 h pe night 7 h pe night Cognitive dietay estaint 0.30 0.21 0.39 0.18 0.31 0.02 Flexible contol 0.18 0.26* 0.32 0.31* 0.45* 0.26* Rigid contol 0.33 0.02 0.35 0.03 0.25 0.01 Stategic dieting behavio 0.02 0.23* 0.20 0.28* 0.32 0.16 Attitude to self-egulation 0.42* 0.002 0.32 0.03 0.16 0.02 Avoidance of fattening foods 0.11 0.22* 0.16 0.15 0.35 0.11 Disinhibition 0.05 0.09 0.33 0.11 0.05 0.004 Habitual susceptibility 0.19 0.07 0.34 0.09 0.25 0.01 Emotional susceptibility 0.13 0.03 0.26 0.003 0.13 0.05 Situational susceptibility 0.03 0.27* 0.17 0.23* 0.03 0.14 Susceptibility to hunge 0.04 0.09 0.02 0.17 0.09 0.06 Intenal locus of hunge 0.02 0.08 0.04 0.11 0.02 0.04 Extenal locus of hunge 0.14 0.12 0.20 0.19 0.07 0.02 Abbeviations: PSQI, Pittsbugh Sleep Quality Index; WC, waist cicumfeence. Relationship was statistically significant *Po0.05, **Po0.01. and P values ae fo the elationship between changes in eating behavio taits and anthopomety makes accoding to sleep quality o duation categoies. The associations between changes in eating behavio taits and anthopometic makes accoding to sleep categoies wee tested using linea egession analysis adjusting fo age, sex and baseline body mass index. Bold indicates significant esults. Nutition & Diabetes (2014) 1 8 2014 Macmillan Publishes Limited
Eating behavios, sleep and weight loss 7 Figue 1. Associations between changes in sleep quality (a) and sleep duation (b) with changes in mophological vaiables in esponse to the diet intevention. The elationship was statistically significant at *Po0.05 and **Po0.01. P values ae fo the elationship between tetiles of change in sleeping habits and anthopometic makes. Values ae expessed as least squaes means. The associations between tetiles of change in sleeping habits and anthopometic makes wee pefomed using analysis of vaiance adjusting fo age, sex and baseline body mass index. A Tukey post hoc test was pefomed to detemine which goups wee significantly diffeent. to educe adiposity. 3,4,34 Indeed, we obseved that having healthy sleeping habits favoed eating behavio taits that wee significantly associated with body weight and adiposity loss. Additionally, we obseved that impovement and deteioation in sleeping habits wee constantly associated with moe and less body weight and fat loss, espectively. Inteestingly, thee was a significant decease in WC in those who impoved thei sleep quality. Poo sleep quality ends up in oveall sleep loss and ou finding is concodant with ecent epidemiological and laboatoy evidence that confim the association between sleep loss and obesity. 35 A longitudinal study by Chaput et al. 36 demonstated that switching fom a shot to a ecommended sleep duation was associated with less visceal fat mass gain ove time. Additionnally, Chaput et al. 37 showed that shot sleep duation favos abdominal obesity in childen. Moe eseach is needed to veify the effects of sleeping habits impovement in the teatment of obesity. This study is the fist one to simultaneously examine the elationships between eating behavios, sleep and adiposity loss. Although eplication studies ae clealy needed to confim the obseved findings, ou esults suggest that eating and sleeping may inteact togethe to pedict the outcome of weight-loss pogams. One must nevetheless keep in mind that most significant associations epoted in the pesent study had coefficients of coelation anging between 0.2 and 0.3, suggesting that they explain a small pat of the vaiability in the elated vaiables. Howeve, this is geneally the case in the fields of obesity and nutition because many factos ae at play. Despite small effect sizes, we believe that moe attention should be paid to eating behavios and sleep hygiene duing weight loss to maximize success. Ou esults need to be intepeted in light of the following limitations. Fist, the small sample size and exploatoy natue of this investigation peclude any definitive conclusion about the inteconnections between eating behavios, sleeping habits and fat loss. Second, the esults come fom a pool of fou diffeent weight-loss inteventions with slightly diffeent aims and inteventions; the idea exploed in the pesent study was thus a etospective analysis. Fo example, the pe-post study design does not include a contol goup (i.e., individuals eviewed on a egula basis but with no instuctions povided). Thid, we have to keep in mind the limitations of self-epoted measues as well as the possibility of esidual confounding by unmeasued vaiables. Fouth, the pesent study only allows fo the identification of associations and does not pemit causal infeences. Fifth, othe dimensions of sleep (e.g., timing, achitectue, consistency and continuity) have not been assessed and could be elated to eating behavios and adiposity loss. Likewise, obstuctive sleep apnea diagnosis o daytime sleepiness wee not available and could have povided elevant infomation. Sixth, the PSQI coves sleep quality ove the month pio to administation, suggesting that the PSQI pefomed at the end of the weight-loss pogam coves sleep duing pat of the intevention. Seventh, we did not contol fo stess, pesonality taits and depession, which might be petinent tagets fo health-cae pofessionals as they seem to be impotant deteminants of weight-loss success. 30,34,38,39 Finally, the extenal genealizability is limited to the sample studied. In conclusion, sleep may modulate eating behavio in adults involved in diffeent dietay inteventions and may act togethe to influence the outcome of weight-loss pogams. Although eplication studies ae needed to confim ou peliminay findings, one must emembe that multiple factos can impact the success of weight-loss pogams. In paticula, dietay pactitiones should exploe sleep hygiene and eating behavio taits as pat of thei oveall assessment befoe pescibing any dietay plan if they want to impove the effectiveness of thei intevention. CONFLICT OF INTEREST The authos declae no conflict of inteest. 2014 Macmillan Publishes Limited Nutition & Diabetes (2014) 1 8
8 ACKNOWLEDGEMENTS Eating behavios, sleep and weight loss We expess ou gatitude to the subjects fo thei paticipation and the staff of the Physical Activity Sciences Laboatoy at Laval Univesity fo thei contibution to this study. None of the authos had a pesonal inteest o potential conflict of inteest with the oganizations sponsoing the pojects. JPC holds a Junio Reseach Chai in Healthy Active Living and Obesity Reseach. AT is patly funded by the Canada Reseach Chai in Envionment and Enegy Balance. REFERENCES 1 Chaput JP, Temblay A. Insufficient sleep as a contibuto to weight gain: an update. Cu Obes Rep 2012; 1: 245 256. 2 Chaput JP. Sleep pattens, diet quality and enegy balance. Physiol Behav 2014; 134: 86 91. 3 Nedeltcheva AV, Kilkus JM, Impeial J, Schoelle DA, Penev PD. Insufficient sleep undemines dietay effots to educe adiposity. Ann Inten Med 2010; 153: 435 441. 4 Chaput JP, Temblay A. Sleeping habits pedict the magnitude of fat loss in adults exposed to modeate caloic estiction. Obesity Facts 2012; 5: 561 566. 5 Majo GC, Doucet E, Jacqmain M, St-Onge M, Bouchad C, Temblay A. Multivitamin and dietay supplements, body weight and appetite: esults fom a coss-sectional and a andomised double-blind placebo-contolled study. B J Nut 2008; 99: 1157 1167. 6 Majo GC, Alaie F, Doe J, Phouttama S, Temblay A. Supplementation with calcium+vitamin D enhances the beneficial effect of weight loss on plasma lipid and lipopotein concentations. Am J Clin Nut 2007; 85: 54 59. 7 Sanchez M, Daimont C, Dapeau V, Emady-Aza S, Lepage M, Rezzonico E et al. Effect of Lactobacillus hamnosus CGMCC1.3724 supplementation on weight loss and maintenance in obese men and women. B J Nut 2013; 3: 1 13. 8 Dapeau V, Aguin H, Temblay A. Impact of satiating diet on satiety quotient and body weight in obese men. Appl Physiol Nut Metab 2010; 35: 373. 9 Eating Well with Canada s Food Guide. Ottawa, Canada, 2007. Available fom http://www.hc-sc.gc.ca/fn-an/food-guide-aliment/index-eng.php (accessed 17 August 2014). 10 Sii WE. The goss composition of the body. Adv Biol Med Phys 1956; 4: 239 280. 11 Buysse DJ, Reynolds CF, Monk TH, Beman SR, Kupfe DJ. The Pittsbugh Sleep Quality Index: a new instument fo psychiatic pactice and eseach. Psychiaty Res 1989; 28: 193 213. 12 Capente JS, Andykowski MA. Psychometic evaluation of the Pittsbugh Sleep Quality Index. J Psychosom Res 1998; 45: 5 13. 13 Stunkad A, Messick S. The thee facto eating questionnaie to measue dietay estaint, disinhibition and hunge. J Psychosom Res 1985; 29: 13. 14 Westenhoefe J, Stunkad AJ, Pudel V. Validation of the flexible and igid contol dimensions of dietay estaint. Int J Eat Disod 1999; 26: 53 64. 15 Bond MJ, McDowell AJ, Wilkinson JY. The measuement of dietay estaint, disinhibition and hunge: an examination of the facto stuctue of the Thee Facto Eating Questionnaie (TFEQ). Int J Obes Relat Metab Disod 2001; 256: 900 906. 16 Laessle RG, Tuschl RJ, Kotthaus BC, Pike KM. A compaison of the validity of thee scales fo the assessment of dietay estaint. J Abnom Psychol 1989; 98: 504 507. 17 Allison DB, Kalinsky LB, Goman BS. A compaison of the psychometic popeties of thee dimensions of dietay estait. Psychol Assessment 1992; 4: 8. 18 Lluch A. Identification des conduites alimentaies pa appoches nutitionnelles et psychométiques: implications théapeutiques et péventives dans l'obésité humaine (Identification of food intake behavious by nutitional and psychometics means: implications fo pevention and teatment of human obesity). PhD thesis. 1995, Univesity of Nancy: Fance. 19 Yang YJ, Kim MK, Hwang SH, Ahn Y, Shim JE, Kim DH. Relative validities of 3-day food ecods and the food fequency questionnaie. Nut Res Pact 2010; 42: 142 148. 20 Govenment of Canada. The Canadian Nutient File. Health and Welfae Canada: Ottawa, Canada, 2010. 21 Keanen AM, Savolainen MJ, Reponen AH, Kujai ML, Lindeman SM, Bloigu RS et al. The effect of eating behavio on weight loss and maintenance duing a lifestyle intevention. Pev Med 2009; 49: 32 38. 22 Elfhag K, Rossne S. Who succeeds in maintaining weight loss? A conceptual eview of factos associated with weight loss maintenance and weight egain. Obes Rev 2005; 6: 67 85. 23 Povenche V, Dapeau V, Temblay A, Despes JP, Lemieux S. Eating behavios and indexes of body composition in men and women fom the Quebec family study. Obes Res 2003; 11: 783 792. 24 Teixeia PJ, Going SB, Sadinha LB, Lohman TG. A eview of psychosocial pe-teatment pedictos of weight contol. Obes Rev 2005; 6: 43 65. 25 Dapeau V, Povenche V, Lemieux S, Despés JP, Bouchad C, Temblay A. Do 6-y changes in eating behavios pedict changes in body weight? Results fom the Quebec Family Study. Int J Obes Relat Metab Disod 2003; 27: 808 814. 26 Byant EJ, King NA, Blundell JE. Disinhibition: its effects on appetite and weight egulation. Obes Rev 2008; 9: 409 419. 27 Rodiguez A, Gomez-Ambosi J, Catalan V, Gil MJ, Beceil S, Sainz N et al. Acylated and desacyl ghelin stimulate lipid accumulation in human visceal adipocytes. Int J Obes (Lond) 2009; 33: 541 552. 28 Tschop M, Smiley DL, Heiman ML. Ghelin induces adiposity in odents. Natue 2000; 407: 908 913. 29 Killgoe WD, Kahn-Geene ET, Lipizzi EL, Newman RA, Kamimoi GH, Balkin TJ. Sleep depivation educes peceived emotional intelligence and constuctive thinking skills. Sleep Med 2008; 9: 517 526. 30 Dweck JS, Jenkins SM, Nolan LJ. The ole of emotional eating and stess in the influence of shot sleep on food consumption. Appetite 2014; 72: 106 113. 31 van Stien T, Fijtes JER, Beges GPA, Defaes PB. The Dutch Eating Behavio Questionnaie (DEBQ) fo assessment of estained, emotional, and extenal eating behavio. Int J Eat Disode 1986; 5: 295 315. 32 Chaput JP, Despés JP, Bouchad C, Temblay A. The association between shot sleep duation and weight gain is dependent on disinhibited eating behavio in adults. Sleep 2011; 34: 1291 1297. 33 Kilkus JM, Booth JN, Bomley LE, Daukhanavala AP, Impeial JG, Penev PD. Sleep and eating behavio in adults at isk fo type 2 diabetes. Obesity (Silve Sping) 2012; 20: 112 117. 34 Elde CR, Gullion CM, Funk KL, Deba LL, Lindbeg NM, Stevens VJ. Impact of sleep, sceen time, depession and stess on weight change in the intensive weight loss phase of the LIFE study. Int J Obes (Lond) 2012; 36: 86 92. 35 Beccuti G, Pannain S. Sleep and obesity. Cu Opin Clin Nut Metab Cae 2011; 14: 402 412. 36 Chaput JP, Bouchad C, Temblay A. Change in sleep duation and visceal fat accumulation ove 6 yeas in adults. Obesity (Silve Sping) 2014; 22: E9 E12. 37 Chaput JP, Temblay A. Does shot sleep duation favo abdominal adiposity in childen? Int J Pediat Obes 2007; 2: 188 191. 38 Fench SA, Epstein LH, Jeffey RW, Blundell JE, Wadle J. Eating behavio dimensions. Associations with enegy intake and body weight. A eview. Appetite 2012; 59: 541 549. 39 Bekke MH, van de Meeendonk C, Molleus J. Effects of negative mood induction and impulsivity on self-peceived emotional eating. Int J Eat Disode 2004; 36: 461 469. This wok is licensed unde a Ceative Commons Attibution- NonCommecial-NoDeivs 4.0 Intenational License. The images o othe thid paty mateial in this aticle ae included in the aticle s Ceative Commons license, unless indicated othewise in the cedit line; if the mateial is not included unde the Ceative Commons license, uses will need to obtain pemission fom the license holde to epoduce the mateial. To view a copy of this license, visit http:// ceativecommons.og/licenses/by-nc-nd/4.0/ Nutition & Diabetes (2014) 1 8 2014 Macmillan Publishes Limited