Eating Behaviors 15 (2014) 291–297
Contents lists available at ScienceDirect
Disordered eating partly mediates the relationship between poor sleep quality and high body mass index Shin-Shyuan Sally Yeh ⁎, Rhonda Frances Brown 1 Research School of Psychology, The Australian National University, Canberra, ACT 2601, Australia
a r t i c l e
i n f o
Article history: Received 21 August 2013 Received in revised form 31 January 2014 Accepted 11 March 2014 Available online 4 April 2014 Keywords: Sleep quality Body mass index Overweight Obesity Binge eating Night-eating
a b s t r a c t Objective: We evaluated the relationship between poor sleep quality and high body mass index (BMI) in a community-derived sample. In addition, we explored the premise that disordered eating (i.e. eating late at night and/or binge eating, which can occur at night) may partly explain the relationship. Method: An online survey asked 330 participants about their height and weight, recent sleep quality, and recent experiences of binge-eating and night-time eating. Results: Using multiple regression analyses, high BMI was shown to be related to shorter sleep duration, increased sleep latency, use of sleeping medications and worse binge-eating, whereas worse sleep quality was related to worse night-eating, after controlling for depression and demographics. Using mediational analyses, bingeeating was shown to partly mediate the relationship between worse sleep quality to higher BMI, whereas night-eating mediated the reverse association of high BMI to worse sleep quality. Discussion: The results suggest that night- and/or binge-eating may partly explain the observed relationship between worse sleep quality and overweight/obesity. Thus, the relationship may simply reﬂect that overweight people are more likely to binge-eat while they wait for sleep to come, and this may contribute to weight gain over time. In addition, the results may indicate that eating rather than weight gain or obesity may be responsible for causing the sleep deﬁcits in overweight people. © 2014 Elsevier Ltd. All rights reserved.
1. Introduction A number of recent studies have reported the existence of a relationship between speciﬁc sleep deﬁcits and overweight/obesity, or its proxy measure of high body mass index (BMI). For example, insufﬁcient sleep has been shown to be associated with overweight/obesity (Vorona et al., 2005). In particular, sleeping less than 6 h/night and falling asleep after midnight are related to an increased risk of obesity (Shigeta, Shigeta, Nakazawa, Nakamura, & Yoshikawa, 2001). Similar results have been obtained in relation to sleep problems and long-term weight gain (Patel, Malhotra, White, Gottlieb, & Hu, 2006). In addition, a recent large meta-analysis of sleep data obtained from children and adults found a signiﬁcant relationship between obesity and shorter sleep duration (Cappuccio et al., 2008). However, the relationship between overweight/obesity and other sleep parameters is less clear. For example, overweight/obesity has been shown to predict worse sleep problems (Patel & Hu, 2008),
⁎ Corresponding author. Tel.: +61 405 469 506. E-mail addresses: [email protected]
(S.-S.S. Yeh), [email protected]
(R.F. Brown). 1 Tel.: +61 2 6125 0635.
http://dx.doi.org/10.1016/j.eatbeh.2014.03.014 1471-0153/© 2014 Elsevier Ltd. All rights reserved.
although the sleep disturbance appears to decrease as a person's weight decreases, suggesting that weight gain may contribute to sleep disturbance (Dixon, Schachter, & O'Brien, 2001). In contrast, Owens and Matthews (1998) failed to ﬁnd any association between waking in the middle of the night or earlier than intended and high BMI, in a sample of healthy women. Similarly, the relationship between longer sleep latency (i.e., time to fall sleep) and overweight/obesity is unclear, with one study showing similar sleep latencies in obese and control subjects (Resta et al., 2003), and another showing that obese people have trouble falling asleep relative to non-obese controls (Pearson, Johnson, & Nahin, 2006). Other than these sleep parameters, few aspects of sleep have been evaluated in relation to overweight/obesity. However, excessive daytime sleepiness has been observed in obese people with sleepbreathing problems (Resta et al., 2003) and obese women have been shown to use fewer sleeping pills than lean older women (60–64 years) (Asplund & Åberg, 1995). Thus, there is a lack of clarity as to which aspects of sleep are impaired in overweight/obese people, although a robust link has been reported between shorter sleep duration and overweight/obesity or weight gain (Cappuccio et al., 2008; Patel et al., 2006; Vorona et al., 2005). Therefore, a broader examination of a variety of sleep parameters is certainly required. In this study, we used the Pittsburgh Sleep Quality Index (PSQI), a well-validated self-report measure of sleep that assesses
S.-S.S. Yeh, R.F. Brown / Eating Behaviors 15 (2014) 291–297
seven subjective domains of sleep quality (Buysse, Reynolds, Monk, Berman, & Kupfer, 1989). Furthermore, few studies have sought to explain the relationship between sleep deﬁcits and overweight/obesity, other than by referring to sleep apnea (Owens & Matthews, 1998). Sleep apnea is the repeated interruption of airﬂow through the nose and mouth for at least 10 s on at least 30 occasions during REM and non-NREM sleep (Guilleminault, Tilkian, & Dement, 1976). It is a relatively infrequent condition occurring in about 5–10% of the population (Tishler, Larkin, Schluchter, & Redline, 2003), although poor sleep quality is much more common at 10–40% (Simon & VonKorff, 1997; Ustun et al., 1996). Thus, only a proportion of overweight people who suffer from poor sleep are likely to meet diagnostic criteria for sleep apnea; and there are likely to be other possible explanations of their documented sleep deﬁcits. However, no other potential clinical or statistical mediators appear to have been investigated in this regard, despite a recent review suggesting that the relationship is likely to be mediated by other factors (Owens & Matthews, 1998). Thus, in this study we explored the premise that night-time eating might mediate the relationship between worse sleep quality and high BMI, in a community-derived sample. For example, a delayed sleep onset and shorter sleep duration may simply provide an overweight person with more time in which to eat, thereby, contributing to an increase in their weight over time. Such an assertion is corroborated by the results of recent studies indicating that nighttime eating is associated with later weight gain in obese people (Andersen, Stunkard, Sørensen, Petersen, & Heitmann, 2004). Nighteating is also related to less weight loss during weight-loss programs (Gluck, Geliebter, & Satov, 2001), suggesting that a person's excessive eating at night may partly explain their weight gain over time. In addition, excessive night-eating is well known to be related to sleep problems (e.g. reduced sleep efﬁciency, increased sleep latency) (Crispim et al., 2011). In fact, sleep problems are a key symptom of night-eating syndrome (NES), a disorder characterized by the delayed consumption of food during the day resulting in hyperphagia (i.e., increased appetite and food consumption) in the evening or after waking at night on three or more occasions/week (Stunkard, Grace, & Wolff, 1955). NES has a reported prevalence of about 1.5% in the population (Rand, Macgregor, & Stunkard, 1997), although several clinical studies suggest that overweight/obese people may be at an increased risk of developing NES (Stunkard et al., 1955). In this study, we operationalized night-eating in terms of the above-described eating patterns seen in NES patients. Binge eating might also potentially explain the relationship between poor sleep quality and overweight/obesity, since binge-eating can occur at night, and it is known to be related to nocturnal eating (i.e. eating after awakening at night then returning to sleep) (Striegel-Moore et al., 2010). Consistent with this interpretation, obese women who are binge-eaters have been shown to experience worse insomnia than those who are not binge-eaters (Bulik, Sullivan, & Kendler, 2002). However, the relationship between binge eating disorder (BED) and overweight/obesity is less clear. For example, one study found that people with BED were more likely to have a BMI ≥ 40 than those without BED (Hudson, Hiripi, Pope, & Kessler, 2007), whereas in another, the prevalence of BED was similar in overweight/obese and normal weight women (Kinzl, Traweger, Trefalt, Mangweth, & Biebl, 1999). BED is characterized by impaired control over binge-eating (e.g., eating very quickly until excessively full) that can occasion signiﬁcant distress (e.g., depression), without any compensatory behavior (e.g., purging, over-exercising) to offset the binge-eating, once or more times weekly for three months or more (American Psychiatric Association, 2013). The person must also exhibit three or more of the following symptoms: eating hastily till uncomfortably full, consuming large amounts of food even when they are not physically hungry, eating alone due to self-embarrassment or self-disgust, and feeling depressed and/or guilty after binge-eating (Lundgren, Rempfer, Brown, Goetz, & Hamera, 2010). The lifetime prevalence estimate of BED is reported to be about 2.8% in the population (Hudson et al., 2007).
Finally, regarding potential confounders, depression is a well-known correlate of overweight/obesity, binge- and night-eating, and sleep disturbance. For example, overweight/obesity has been shown to be related to worse clinically-relevant depression (Bulik et al., 2002); binge-eating is associated with a higher lifetime risk of major depression (Specker, de Zwaan, Raymond, & Mitchell, 1994) and worse depression (Bulik et al., 2002); night-eaters are reported to experience more severe depression than non-night eaters (Gluck et al., 2001); and, depression is reliably linked to sleep disturbance (Owens & Matthews, 1998; Taylor, Lichstein, Durrence, Reidel, & Bush, 2005). Thus, we controlled for depression in all the planned regression analyses in this study. In summary, the abovementioned sleep and weight problems associated with NES and BED are consistent with the premise that disordered eating at night (i.e. night-eating and binge-eating) may explain the relationship between poor sleep quality and high BMI. In accordance with the limited available literature, we expected that: (1) shorter sleep duration and more binge- and night-eating will be associated with high BMI; (2) higher BMI and more binge- and night-eating will be associated with worse total sleep quality; (3) binge-and night-eating will mediate the relationship between total sleep quality to high BMI; and, (4) binge- and night-eating will mediate the relationship between high BMI to worse sleep quality. 2. Method 2.1. Participants Potential participants were recruited via advertisements placed at the Australian National University (ANU) campus and on a number of online platforms (e.g., ANU Notices webpage, Facebook event and pages, and health webpages). Participants were encouraged to email snowball the imbedded study link to their friends and colleagues. First-year psychology students at ANU who participated in the study could receive 30 min course credit for their participation, although they had a large number of studies to choose from. No reimbursements were provided to other participants. Study inclusion criteria were an age of 18 years or older and a BMI of 18.5 or more. The study took approximately 30 min to complete. Six-hundred-seventy-eight people entered the study webpage by clicking on the study URL, of whom 312 immediately exited (i.e. nonresponders), leaving 366 participants (i.e. response rate = 54%). Of these, 35 cases were eliminated due to a BMI less than 18.5 and one case was eliminated due to an age of 17 years, leaving 330 participants. Of these, 107 (32.4%) were males and 223 (67.6%) were females. The mean age of the participants was 27.42 years (SD = 10.36, range: 18– 87). Most participants were never married (N = 238, 72.2%) and the remainder were married/living de facto (N = 64, 19.4%), divorced/ separated (N = 12) or other (N = 16). Most participants held a university/college degree (N = 137, 41.5%), postgraduate degree (N = 59, 17.9%) or trade certiﬁcate/diploma (N = 35, 10.6%), and the remainder had completed part (N = 7) or all of high school (N = 92). Nearly onehalf of the participants were university students (N = 161, 48.4%) or they worked full-time (N = 97, 29.4%) or part-time (N = 46, 13.9%), with the remainder being unemployed (N = 16), retired (N = 2), volunteers (N = 2), unable to work (N = 3) or completing home duties (N = 3). 2.2. Procedure Interested people were asked to access the online study by clicking on the imbedded URL in the advertisement (https://anupsych.us. qualtrics.com/SE/?SID=SV_1HZrb18I7ZcrxPe). They were asked to read the information page which explained the study, its purpose and format. They were asked if they were 18 years or older, and if not,
S.-S.S. Yeh, R.F. Brown / Eating Behaviors 15 (2014) 291–297
they were kindly thanked for their time. If they answered yes, they were directed to the informed consent page to which they were asked to tick the box to indicate their consent. They were then asked to complete the short online study questionnaire that was delivered via Qualtrics survey software. 2.3. Materials Demographics included age, sex, relationship status, education and employment status. Participants were asked to provide their height (in meters) and weight (in kilograms) so that their BMI could be computed. BMI is an individual's weight in kilograms over their height in meters squared; with a BMI of ≥25 indicating a person is overweight and ≥ 30 indicating that one is obese (World Health Organization, 2000). Participants were then asked about their recent experience of sleep quality, binge eating and night-eating using the following scales. Seven sleep parameters were assessed over the past month using the Pittsburgh Sleep Quality Index (PSQI, 19-item). The PSQI assesses seven subjective domains of sleep quality including sleep duration, sleep disturbance, sleep latency, daytime dysfunction, habitual sleep efﬁciency, subjective sleep quality, and the use of sleep medications (Buysse et al., 1989). Item scores were combined to give a total sleep quality score out of 21, ranging from 0 (no sleeping difﬁculty) to 21 (severe sleeping difﬁculty across all seven components). Individual sleep component scores ranged from 0 to 3, with 0 representing no sleep difﬁculty and 3 indicating a severe speciﬁc sleep difﬁculty. A total score N 5 indicates that a person has moderate to severe sleep difﬁculties, and it distinguishes well between poor and good sleepers (Buysse et al., 1989). Total sleep quality score has high internal consistency reliability with a Cronbach's α of .83 (Smith & Wegener, 2003). In this study, the scale showed adequate internal consistency with a Cronbach's alpha of .76. Disordered eating was operationalized as: (i) binge eating, as indexed by the Binge Eating Scale (Gormally, Black, Daston, & Rardin, 1982), and (ii) night-time eating, as indexed by the Night-Eating Questionnaire (Striegel-Moore, Franko, & Garcia, 2009). Binge eating and its severity were assessed using the 16-item Binge Eating Scale (BES) (Gormally et al., 1982). Participants were asked to select the responses that best described their current feelings and/or tendencies. The items were rated on scales from 0 (no binge eating) to 3 (severe binge eating), with total scores ranging from 0 to 46, and high scores indicating worse binge-eating. Using established cut-offs for the scale, participants who scored ≥ 27 were considered to be binge-eaters (Marcus et al., 1990). The scale has high internal consistency, with Cronbach's alphas above .8 (Gormally et al., 1982). In this study, the scale had high internal consistency with a Cronbach's alpha of .92. Night-time-eating was assessed using the 14-item Night Eating Questionnaire which screens for potential night-eating syndrome (NES) cases (Striegel-Moore et al., 2009). Participants were asked to answer items asking about the presence, frequency and/or amount of the behavior, using Likert type scales ranging from 0 to 4. A total NEQ score (score range: 0–52) was computed by summing the item scores (except item 13). A recommended cut-off score of 25 is reported to identify possible NES cases. The scale has adequate convergent validity and internal consistency reliability, with a Cronbach's alpha of .70 (Allison et al., 2008). In this study, the NEQ showed adequate internal consistency with a Cronbach's alpha of .73. Depression over the past week was assessed using the Depression Anxiety Stress Scales (DASS-42) (Lovibond & Lovibond, 1995). This scale asks participants to rate the presence and severity of negative emotions, using 4-point scales ranging from 0 (did not apply to me at all) to 3 (applied to me very much/most of the time), with the items summed to give a total score, and high scores indicating worse depression. Signiﬁcant depression is indicated by a recommended cut-off score of 12 or more. The scale has high internal consistency, with Cronbach's α for the depression subscale above .9 (Lovibond & Lovibond, 1995). In
this study, the depression subscale had high internal consistency with a Cronbach's α of .97. 2.4. Statistical analyses PASW/SPSS statistical software (version 20) was used to perform all statistical analyses. A planned hierarchical multiple regression analysis examined the association between total sleep quality to BMI, and another examined the association of the PSQI subscales to BMI, controlling for depression and demographics at step 1. Binge-eating and night-eating were included at step 2 in these regression models. Two planned hierarchical regression analyses evaluated the reverse association of BMI to sleep quality, separately for total sleep score and the PSQI subscales, along with binge-eating and night-eating. Finally, four mediational analyses were conducted using the methods of Preacher and Hayes (2004) including Sobel tests. Two analyses tested whether bingeeating and night-eating mediated the relationship between total sleep quality to BMI, and two evaluated the reverse association of BMI to total sleep quality. An online tool (Preacher & Leonardelli, 2001) was used to compute Sobel test's values in each analysis. 3. Results 3.1. Data screening and cleaning Univariate outliers were detected in total sleep quality, binge-eating, night-eating and BMI, but no cases were excluded since they all represented clinically-relevant cases of sleep disturbance, disordered eating and obesity. In addition, the 5% trimmed means were similar to the original means. Total sleep quality, sleep duration, habitual sleep efﬁciency, use of sleeping medication, binge-eating, night-eating, BMI, age and depression score were skewed and the Kolmogorov–Smirnov and Shapiro–Wilk normality tests were signiﬁcant (p b .001), indicating that they had non-normal distributions. Data transformation did not normalize the score distributions, and the same proﬁle of results was obtained using square-root-transformed and untransformed data (data not shown). Thus, only the untransformed data is presented in this study. No signiﬁcant multicollinearity was evident between the variables, except between total sleep quality and the seven sleep subscales. Thus, total sleep quality and the sleep subscales were placed in separate regressions when assessing their association with BMI. Means and standard deviations of the key study variables and the correlations between them are presented in Table 1. Binge-eating and night-eating were shown to be strongly correlated, suggesting that a proportion of binge-eaters were also night-eaters, and vice-versa. Using the established PSQI cut-off score of N5 (Buysse et al., 1989), 169 out of 330 (51%) participants were deemed to have poor sleep quality. Using the established BES cut-off score of ≥27 (Greeno, Marcus, & Wing, 1995), 39 out of 330 (12%) participants were shown to be clinically-relevant binge-eaters. Using the established NEQ cut-off score of ≥ 25 (Allison et al., 2008), 31 out of 330 (9%) participants were shown to be clinically-relevant night-eaters. Using the above BES and NEQ cut-off scores, 16 of the 39 binge eaters were also found to be night-eaters (41.0%), and 16 of the 31 night-eaters were also bingeeaters (51.6%). Finally, using BMI cut-offs of ≥ 25 and b30 for overweight and ≥30 for obese (World Health Organization, 2000), 57 out of 330 (17%) participants were considered to be overweight and 59 (18%) were obese. 3.2. Hierarchical multiple regression analyses In the ﬁrst regression analysis, depression score and demographics that were signiﬁcantly correlated with BMI were entered at step 1, and total sleep quality and binge- and night-eating were entered at step 2, see Table 2. At step 1, the variables accounted for 15% (R2 = .148) of the variance in BMI. When total sleep quality and the disordered eating
S.-S.S. Yeh, R.F. Brown / Eating Behaviors 15 (2014) 291–297
Table 1 Means, standard deviations and correlationsa for the key study variables.
Age Dep Durat Distb Laten Dayd HSE Qual Meds PSQI BES NEQ BMI
27.42 12.27 0.51 1.14 1.35 1.32 0.51 1.19 0.35 6.37 13.78 14.60 26.08
(10.36) (11.71) (0.83) (0.54) (1.01) (0.84) (0.85) (0.78) (0.82) (3.66) (10.25) (6.87) (8.55)
.22⁎⁎ .35⁎⁎ .21⁎⁎
.00 .38⁎ .38⁎⁎ .30⁎⁎
−.02 .56⁎⁎ .22⁎⁎ .27⁎⁎ .26⁎
.13⁎ .24⁎⁎ .53⁎⁎ .24⁎⁎ .55⁎ .17⁎⁎
.15⁎⁎ .38⁎⁎ .46⁎⁎ .38⁎⁎ .50⁎ .42⁎⁎ .41⁎⁎
.11⁎ .22⁎⁎ .16⁎⁎ .18⁎⁎ .29⁎ .14⁎ .20⁎⁎ .18⁎⁎
.15⁎⁎ .53⁎⁎ .67⁎⁎ .52⁎⁎ .77⁎ .55⁎⁎ .71⁎⁎ .75⁎⁎ .49⁎⁎
−.05 .59⁎⁎ .20⁎⁎ .22⁎⁎
.34⁎⁎ .17⁎⁎ .21⁎⁎ .22⁎⁎
.31. .39⁎⁎ .19⁎⁎ .31⁎⁎ .22⁎⁎ .41⁎⁎
.54⁎⁎ .33⁎⁎ .40⁎⁎ .42⁎ .41⁎⁎ .33⁎⁎ .47⁎⁎ .25⁎⁎ .58⁎⁎ .59⁎⁎
.03. .22⁎⁎ .05. .22⁎⁎ .21⁎⁎ .25⁎⁎ .23⁎⁎ .21⁎⁎
Notes. N = 330. Dep = Depressive Symptoms Score. Durat = sleep duration. Distb = sleep disturbances. Laten = sleep latency. Dayd = daytime dysfunction. HSE = habitual sleep efﬁciency. Qual = subjective sleep quality. Meds = use of sleep medication. PSQI = PSQI global score = total sleep quality score. BES = binge eating scale score. NEQ = night eating questionnaire score. BMI = body mass index. a 2-Tailed Pearson correlation in SPSS. ⁎ p b .05. ⁎⁎ p b .01
variables were entered at step 2 this increased the BMI variance to 19% (R2 = .189), R2change = .041, F3,323 = 5.404, p = .001. In this analysis, high BMI was shown to be associated with older age and binge-eating, but not total sleep quality or night-eating. Depression score was not signiﬁcantly associated with BMI after sleep quality was added to the regression model. In the second analysis, the same variables were entered at step 1 and the seven sleep subscales and disordered eating variables were entered at step 2, see Table 3. The variables at step 1 accounted for 15% (R2 = .148) of the variance in BMI, and those at step 2 increased the BMI variance to 25% (R2 = .248), R2change = .100, F9,317 = 4.681, p = .000. In this analysis, high BMI was shown to be associated with older age, shorter sleep duration, longer sleep latency, greater use of sleeping medication and binge-eating, with a trend towards worse daytime dysfunction. Depression score was not associated with BMI after individual sleep parameters were added to the model. In a third analysis, the reverse association of high BMI to worse sleep quality was examined. Depression score and demographics that were signiﬁcantly correlated with total sleep score were entered at step 1, and BMI and night- and binge-eating were entered at step 2, see Table 4. The variables at step 1 accounted for 33% (R2 = .333) of the variance in total sleep score, and those at step 2 signiﬁcantly increased the variance to 45% (R2 = .450), R2change = .117, F3,322 = 22.811, p = .000. In this analysis, worse sleep quality was associated with older age, female gender, worse depression and night-time eating, but not BMI or binge-eating.
3.3. Mediational analyses Finally, four mediational analyses and associated Sobel tests evaluated whether binge-eating and/or night-eating mediated the association of worse sleep quality to high BMI, and the reverse association of high BMI to worse sleep quality. Depression score and signiﬁcant demographics were controlled at step 1 in all the analyses. First, binge-eating and night-eating were tested as mediators of the relationship between worse sleep quality to high BMI, see Fig. 1. As indicated, the sleep quality to BMI relationship was weakened after controlling for binge- and nighteating. Binge-eating partly mediated between worse sleep quality to high BMI, and the mediation was signiﬁcant using Sobel's test (t = 2.17, p = .03). Night-time eating also partly mediated between worse sleep quality to high BMI, but the mediation was not signiﬁcant using Sobel's test (t = 1.79, p = .07), although the result was close to signiﬁcance. We then tested the reverse association of high BMI to worse total sleep quality, separately for binge-eating and night-eating, see Fig. 2. The BMI to total sleep quality relationship was weakened after controlling for binge- and night-eating. As indicated, binge-eating partly mediated between high BMI and worse sleep quality, but Sobel's test result was marginally non-signiﬁcant (t = 1.76, p = .08). However, night-time
Table 3 Summary of hierarchical multiple regression analyses predicting BMI. Variable Step 1
Table 2 Summary of hierarchical multiple regression analyses predicting BMI. Variable Step 1
Age Relationship status Depression Age Relationship status Depression Total sleep quality Night-eating Binge-eating
Notes. N = 330. ⁎ p b .01. ⁎⁎ p b .005.
0.27 0.42 0.13 0.27 0.43 −0.01 0.22 0.06 0.15
0.43 0.44 0.04 0.04 0.43 0.05 0.16 0.09 0.06
0.33 0.05 0.17 0.32 0.05 −0.01 0.10 0.05 0.18
6.30⁎⁎ 0.96 3.38⁎⁎ 6.08⁎⁎ 0.99 −0.14 1.45 0.67 2.67⁎
Age Relationship status Depression Age Relationship status Depression Sleep duration Sleep disturbances Sleep latency Daytime dysfunction Habitual sleep efﬁciency Subjective sleep quality Use of sleeping medication Night-eating Binge-eating
Notes. N = 330. ⁎ p b .01. ⁎⁎ p b .005.
0.27 0.42 0.13 0.24 0.45 −0.04 1.63 1.05 −1.11 1.23 −1.23 0.08 1.39 0.04 0.15
0.04 0.44 0.04 0.04 0.43 0.05 0.64 0.91 0.56 0.63 0.65 0.72 0.54 0.09 0.06
0.33 0.05 0.17 0.29 0.05 −0.06 0.16 0.07 −0.13 0.12 −0.12 0.07 0.13 0.03 0.18
t 6.30⁎⁎ 0.96 3.38⁎⁎ 5.34⁎⁎ 1.07 −0.79 2.55⁎ 1.16 −1.99⁎ 1.96 −1.89 1.11 2.56⁎ 0.44 2.63⁎
S.-S.S. Yeh, R.F. Brown / Eating Behaviors 15 (2014) 291–297 Table 4 Summary of hierarchical multiple regression analyses predicting total sleep quality. Variable Step 1
Age Relationship status Gender Depression Age Relationship status Gender Depression BMI Night-eating Binge-eating
0.06 −0.03 1.34 0.16 0.05 −0.12 1.13 0.10 0.04 0.22 −0.02
0.02 0.17 0.36 0.01 0.02 0.15 0.33 0.02 0.02 0.03 0.02
0.17 −0.01 0.17 0.51 0.14 −0.03 0.15 0.31 0.08 0.40 −0.05
3.67⁎⁎ −0.17 3.75⁎⁎ 11.14⁎⁎ 3.11⁎⁎ −0.79 3.42⁎⁎ 5.87⁎⁎ 1.78 7.47⁎⁎ −0.90
Notes. N = 330. ⁎⁎ p b .005.
eating partly mediated between high BMI and worse sleep quality, and Sobel's test result was signiﬁcant (t = 2.64, p = .01) indicating that it signiﬁcantly mediated this association. 4. Discussion This is the ﬁrst study to evaluate binge- and night-eating as a possible mediator of the relationship between high BMI and worse sleep quality. About one-half of the community-derived sample reported experiencing some degree of clinically-relevant sleep impairment, one-third was overweight/obese and one-tenth met caseness for binge-eating disorder (BED) and/or night-eating syndrome (NES). Consistent with hypothesis 1, shorter sleep duration, longer sleep latency, and greater use of sleep medications were shown to be associated with high BMI, after controlling for depression and demographics (i.e. age, marital status). These results are consistent with those of a recent meta-analysis indicating that obesity is related to shorter sleep duration (Cappuccio et al., 2008), and other studies indicating that insufﬁcient sleep or falling asleep after midnight was related to an increased risk of overweight/obesity (Shigeta et al., 2001; Vorona et al., 2005). In addition, the results are consistent with those of Pearson et al. (2006) who found that obese people had trouble falling asleep at night, but they are inconsistent with those of Resta et al. (2003) who reported similar sleep latencies in obese and control subjects, and Asplund and Åberg (1995) who found that obese women used fewer sleeping pills than lean older women. No other sleep parameters (e.g. total sleep quality, sleep disturbance, daytime dysfunction, habitual sleep efﬁciency, subjective sleep quality) (Buysse et al., 1989) were shown to be related to higher BMI in this study. Only daytime sleepiness (Resta et al., 2003) and sleep disturbance have previously been shown to be related to high BMI in the
literature, and then only inconsistently so (Owens & Matthews, 1998; Patel & Hu, 2008). Taken together, the results suggest that the sleep deficits experienced by overweight/obese people are restricted to a delay in falling asleep, a reduction in the time spent sleeping, and a tendency to use sleep medications. In relation to hypothesis 1, BMI was shown to be positively related to binge-eating but not night-eating. The results suggest that binge-eaters are more likely to be overweight/obese than non-binge-eaters (Hudson et al., 2007), although some studies report similar estimates of BED in overweight/obese and normal weight people (e.g. Kinzl et al., 1999). In addition, the results are inconsistent with those of Stunkard et al. (1955) who found that overweight/obese people were at an increased risk of developing NES (Stunkard et al., 1955). Hypothesis 2 evaluated the reverse proposition that higher BMI and more binge- and night-eating will predict worse total sleep quality. Contrary to expectation, BMI and binge-eating did not predict worse total sleep quality, after controlling for depression and demographics (i.e. age, gender, marital status), although worse sleep quality was related to worse night-eating and depression. These results are consistent with a small literature indicating that night-time food consumption is related to less sleep efﬁciency and longer sleep latency (Crispim et al., 2011), and depressed individuals experience poorer sleep quality (e.g. insomnia) (Taylor et al., 2005). However, the results are inconsistent with Bulik et al. (2002), who reported an association between bingeeating and poor sleep quality, although levels of binge-eating were low in this study. In hypotheses 3 and 4 we evaluated the premise that binge- and night-eating might mediate the relationship between total sleep quality and high BMI, and the reverse association of high BMI and worse sleep quality. In the mediational analyses, we found that binge-eating partly mediated the relationship between worse sleep quality and higher BMI, whereas night-time eating partly mediated between high BMI and worse sleep quality. However, night-eating was a marginally nonsigniﬁcant predictor of sleep quality to BMI, and binge-eating was a marginally non-signiﬁcant predictor BMI to sleep quality. These results are consistent with the small literature indicating that excessive nighteating is related to sleep problems (Crispim et al., 2011; Stunkard et al., 1955), and obese women who are binge-eaters report worse insomnia than those who are not binge-eaters (Bulik et al., 2002). Taken together, the results suggest that eating late at night (e.g. after the evening meal or during sleep) may explain some of the observed relationship between insufﬁcient sleep to higher BMI. For example, if an overweight person ﬁnds it difﬁcult to fall asleep at night, they may binge-eat as they wait for sleep to come, and this may cause them to gain weight over time. That is, chronic delayed sleep onset and reduced sleep times may simply provide an overweight person with more time in which to eat.
Fig. 1. Standardized regression coefﬁcients for the relationship between total sleep quality and BMI as potentially partially mediated by binge eating and night-eating. Standardized regression coefﬁcients between total sleep quality and BMI when night-eating or binge eating was for controlled for are in parentheses. *p b .05. **p b .01. ***p b .005. ****p b .001.
S.-S.S. Yeh, R.F. Brown / Eating Behaviors 15 (2014) 291–297
Fig. 2. Standardized regression coefﬁcients for the relationship between BMI and total sleep quality as potentially partially mediated by binge eating and night-eating. Standardized regression coefﬁcients between BMI and total sleep quality when night-eating or binge eating was for controlled for are in parentheses. *p b .05. **p b .01. ***p b .005. ****p b .001.
Additionally, the results suggest that excessive night-time eating may interfere with sleep onset in high BMI individuals. Eating whether it occurs during the day or night is associated with an increase in body temperature (e.g., 0.8–1.5 °C increase in liver temperature for 60–90 min in normal-weight unrestrained eaters) (Westernterp-Plantenga, Wouters, & ten Hoor, 1990). Thus, if a person eats food late at night, this will lead to an elevation in their nocturnal body temperature. Such body temperature elevations are well-known to interfere with the ability of people to fall asleep at night and they disturb sleep and impair sleep quality (Lack, Gradisar, Van Someren, Wright, & Lushington, 2008; Murphy & Campbell, 1997). Thus, it is possible that the participants who ate late at night in this study had experienced sleep deﬁcits via an increase in their nocturnal body temperature. However, the results of this study should be interpreted in light of the following study limitations. First, most of the sample was well-educated, female and of normal weight, suggesting that the generalizability of these ﬁndings to overweight people may be limited. Nonetheless, half of the sample experienced poor sleep quality, one-third was overweight/obese and a proportion met caseness for BED and/or NES. Second, participants were required to have internet access to participate in this study and this access is known to be more frequent in younger and better educated adults (Howard, Rainie, & Jones, 2001). Thus, the results may be less generalizable to older and less well educated people. Third, BMI is not an ideal measure of overweight/obesity, as it disregards the contribution of body fat (Lee, Huxley, Wildman, & Woodward, 2008), but it was an expedient measure of overweight/obesity in this online study. Fourth, subjective sleep quality was assessed using the Pittsburgh Sleep Quality Index, rather than objective sleep data obtained in a sleep laboratory. Fifth, self-report measures of binge-eating are problematic and the behavior is thought to be better measured by means of an interview (Fairburn & Beglin, 1994). Sixth, the cross-sectional nature of this study precludes any causal inferences being made, although the ﬁndings may help build theoretical models that can be used to guide the design of longitudinal studies. Finally, there was considerable overlap (r = .59) in the night- and binge-eating scores which could not be attributed solely to similarities or overlap in the scale items (e.g., item 9, option 3 of the BES: “I have a regular habit of overeating during the night, It seems that my routine is not to be hungry in the morning but overeat in the evening” (Gormally et al., 1982, p. 54)). Thus, the moderate correlation between binge- and night-eating scores may reﬂect a degree of functional overlap between the two disorders, at least in obese people (Adami, Meneghelli, & Scopinaro, 1999; Kuldau & Rand, 1996), although Stunkard et al. (1996) reported a smaller degree of overlap between BED and NES (0–21%) in obese patients. Nonetheless, BED and nighteating are thought to represent independent clinical and behavioral entities (Adami et al., 1999).
5. Conclusion This study explored the premise that late-night and binge-eating may partly explain the relationship between poor sleep quality and high body mass index (BMI). In a series of regression analyses, high BMI was shown to be related to shorter sleep duration, longer sleep latency, greater use of sleep medications and worse binge-eating, after controlling for depression and demographics. In contrast, worse sleep quality was related to worse night-eating but not higher BMI or binge-eating. Finally, binge-eating was shown to partly mediate the relationship between worse sleep quality and higher BMI, whereas night-eating mediated the reverse association of high BMI and worse sleep quality. Taken together, the results suggest that the sleep deﬁcits experienced by overweight/obese people are restricted to a delay in falling asleep, reduction in the time spent sleeping, and a tendency to use sleep medications. Clinically, the results suggest that eating late at night may partly explain the tendency of overweight/obese people to experience sleep problems. First, an overweight person with insomnia may binge-eat while they wait for sleep to come and this may cause them to gain weight over time. Second, excessive eating at night may interfere with sleep onset, possibly due to the propensity of eating to transiently increase body temperature (Westernterp-Plantenga et al., 1990). Thus, the results suggest that there may be some merit in including sleep hygiene training or other sleep treatments in existing obesity treatment programs. In addition, the results suggest that obese patients should avoid eating late at night, since this may disrupt their ability to sleep. Role of funding sources No funding was provided for this study.
Contributors Shin-Shyuan Sally Yeh and Dr. Rhonda Frances Brown designed the study, wrote the protocol, conducted literature searches, provided summaries of previous research studies, and conducted the statistical analysis. Shin-Shyuan Sally Yeh wrote the ﬁrst draft of the manuscript and all authors contributed to and have approved the ﬁnal manuscript.
Conﬂict of interest All other authors declare that they have no conﬂicts of interest.
References Adami, G. F., Meneghelli, A., & Scopinaro, N. (1999). Night eating and binge eating disorder in obese patients. International Journal of Eating Disorders, 25(3), 335–338. http:// dx.doi.org/10.1002/(SICI)1098-108X(199904)25:3b335::AID-EAT12N3.0.CO;2-1. Allison, K. C., Lundgren, J.D., O'Reardon, J. P., Martino, N. S., Sarwer, D. B., Wadden, T. A., et al. (2008). The Night Eating Questionnaire (NEQ): Psychometric properties of a measure of severity of the Night Eating Syndrome. Eating Behaviors, 9(1), 62–72. http://dx.doi.org/10.1016/j.eatbeh.2007.03.007.
S.-S.S. Yeh, R.F. Brown / Eating Behaviors 15 (2014) 291–297 American Psychiatric Association (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Arlington, VA: American Psychiatric Publishing. Andersen, G. S., Stunkard, A. J., Sørensen, T. I., Petersen, L., & Heitmann, B.L. (2004). Night eating and weight change in middle-aged men and women. International Journal of Obesity, 28, 1338–1343. http://dx.doi.org/10.1038/sj.ijo.0802731. Asplund, R., & Åberg, H. E. (1995). Body mass index and sleep in women aged 40 to 64 years. Maturitas, 22(1), 1–8. http://dx.doi.org/10.1016/0378-5122(95)00908-4. Bulik, C. M., Sullivan, P. F., & Kendler, K. S. (2002). Medical and psychiatric morbidity in obese women with and without binge-eating. International Journal of Eating Disorders, 32(1), 72–78. http://dx.doi.org/10.1002/eat.10072. Buysse, D. J., Reynolds, C. F., III, Monk, T. H., Berman, S. R., & Kupfer, D. J. (1989). The Pittsburgh Sleep Quality Index: A new instrument for psychiatric practice and research. Psychiatry Research, 28(2), 193–213. http://dx.doi.org/10.1016/0165-1781(89) 90047-4. Cappuccio, F. P., Taggart, F. M., Kandala, N. -B., Currie, A., Peile, E., Stranges, S., et al. (2008). Meta-analysis of short sleep duration and obesity in children and adults. Sleep, 31(5), 619–626. Crispim, C. A., Zimberg, I. Z., dos Reis, B. G., Diniz, R. M., Tuﬁk, S., & de Mello, M. T. (2011). Relationship between food intake and sleep pattern in healthy individuals. Journal of Clinical Sleep Medicine, 7(6), 659–664. http://dx.doi.org/10.5664/jcsm.1476. Dixon, J. B., Schachter, L. M., & O'Brien, P. E. (2001). Sleep disturbance and obesity: Changes following surgically induced weight loss. Archives of Internal Medicine, 161(1), 102–106. http://dx.doi.org/10.1001/archinte.161.1.102. Fairburn, C. G., & Beglin, S. J. (1994). Assessment of eating disorders: Interview or selfreport questionnaire? International Journal of Eating Disorders, 16, 363–370. Gluck, M. E., Geliebter, A., & Satov, T. (2001). Night eating syndrome is associated with depression, low self-esteem, reduced daytime hunger, and less weight loss in obese outpatients. Obesity Research, 9(4), 264–267. http://dx.doi.org/10.1038/oby.2001.31. Gormally, J., Black, S., Daston, S., & Rardin, D. (1982). The assessment of binge-eating severity among obese persons. Addictive Behaviors, 7(1), 47–55. http://dx.doi.org/10. 1016/0306-4603(82)90024-7. Greeno, C. G., Marcus, M.D., & Wing, R. R. (1995). Diagnosis of Binge eating Disorder: Discrepancies between a questionnaire and clinical interview. International Journal of Eating Disorders, 17(2), 153–160. http://dx.doi.org/10.1002/1098-108X(199503)17: 2b153::AID-EAT2260170208N3.0.CO;2-V. Guilleminault, C., Tilkian, A., & Dement, W. C. (1976). The sleep apnea syndromes. Annual Review of Medicine, 27(1), 465–484. http://dx.doi.org/10.1146/annurev.me.27. 020176.002341. Howard, P., Rainie, L., & Jones, S. (2001). Days and nights on the Internet: The impact of a diffusing technology. American Behavioral Scientist, 45(3), 383–404. http://dx.doi.org/ 10.1177/0002764201045003003. Hudson, J. I., Hiripi, E., Pope, H. G., Jr., & Kessler, R. C. (2007). The prevalence and correlates of eating disorders in the National Comorbidity Survey Replication. Biological Psychiatry, 61(3), 348–358. http://dx.doi.org/10.1016/j.biopsych.2006.03.040. Kinzl, J. F., Traweger, C., Trefalt, E., Mangweth, B., & Biebl, W. (1999). Binge eating disorder in females: A population-based investigation. International Journal of Eating Disorders, 25(3), 287–292. http://dx.doi.org/10.1002/(SICI)1098-108X(199904)25:3b287::AIDEAT6N3.0.CO;2-8. Kuldau, J. M., & Rand, C. S. (1996). The night eating syndrome and bulimia in the morbidly obese. International Journal of Eating Disorders, 5, 143–149. Lack, L. C., Gradisar, M., Van Someren, E. J. W., Wright, H. R., & Lushington, K. (2008). The relationship between insomnia and body temperatures. Sleep Medicine Reviews, 12(4), 307–317. http://dx.doi.org/10.1016/j.smrv.2008.02.003. Lee, C. M. Y., Huxley, R. R., Wildman, R. P., & Woodward, M. (2008). Indices of abdominal obesity are better discriminators of cardiovascular risk factors than BMI: A metaanalysis. Journal of Clinical Epidemiology, 61(7), 646–653. http://dx.doi.org/10.1016/ j.jclinepi.2007.08.012. Lovibond, S. H., & Lovibond, P. F. (1995). Manual for the Depression Anxiety Stress Scales (2nd ed.). Sydney: Psychology Foundation. Lundgren, J.D., Rempfer, M. V., Brown, C. E., Goetz, J., & Hamera, E. (2010). The prevalence of night eating syndrome and binge eating disorder among overweight and obese individuals with serious mental illness. Psychiatry Research, 175(3), 233–236. http://dx. doi.org/10.1016/j.psychres.2009.10.027. Marcus, M.D., Wing, R. R., Ewing, L., Kern, E., Gooding, W., & McDermott, M. (1990). Psychiatric disorders among obese binge eaters. International Journal of Eating Disorders, 9(1), 69–77. http://dx.doi.org/10.1002/1098-108X(199001)9:1b69::AIDEAT2260090108N3.0.CO;2-K. Murphy, P. J., & Campbell, S. S. (1997). Nighttime drop in body temperature: A physiological trigger for sleep onset? Sleep, 20(7), 505–511 (Retrieved from http://www. journalsleep.org). Owens, J. F., & Matthews, K. A. (1998). Sleep disturbance in healthy middle-aged women. Mauritas, 30(1), 41–50. http://dx.doi.org/10.1016/S0378-5122(98)00039-5.
Patel, S. R., & Hu, F. B. (2008). Short sleep duration and weight gain: A systematic review. Obesity, 16(3), 643–653. http://dx.doi.org/10.1038/oby.2007.118. Patel, S. R., Malhotra, A., White, D. P., Gottlieb, D. J., & Hu, F. B. (2006). Association between reduced sleep and weight gain in women. American Journal of Epidemiology, 164(10), 947–954. http://dx.doi.org/10.1093/aje/kwj280. Pearson, N. J., Johnson, L. L., & Nahin, R. L. (2006). Insomnia, trouble sleeping, and complementary and alternative medicine: Analysis of the 2002 National Health Interview Survey data. Archives of Internal Medicine, 166(16), 1775–1782. http://dx.doi.org/10. 1001/archinte.166.16.1775. Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures for estimating indirect effects in simple mediation models. Behavior Research Methods, Instruments, & Computers, 36(4), 717–731. http://dx.doi.org/10.3758/BF03206553. Preacher, K. J., & Leonardelli, G. J. (2001). Calculation for the Sobel test: An interactive calculation tool for mediation tests. Retrieved from. http://quantpsy.org/sobel/ sobel.htm Rand, C. S. W., Macgregor, A.M. C., & Stunkard, A. J. (1997). The night eating syndrome in the general population and among postoperative obesity surgery patients. International Journal of Eating Disorders, 22(1), 65–69. http://dx.doi.org/10.1002/ (SICI)1098-108X(199707)22:1b65::AID-EAT8N3.0.CO;2-0. Resta, O., Foschino Barbaro, M. P., Bonﬁtto, P., Giliberti, T., Depalo, A., Pannacciulli, N., et al. (2003). Journal of Internal Medicine, 253(5), 536–543. http://dx.doi.org/10.1046/j. 1365-2796.2003.01133.x. Shigeta, H., Shigeta, M., Nakazawa, A., Nakamura, N., & Yoshikawa, T. (2001). Lifestyle, obesity, and insulin resistance. Diabetes Care, 24(3), 608. http://dx.doi.org/10.2337/ diacare.24.3.608. Simon, G. E., & VonKorff, M. (1997). Prevalence, burden, and treatment of insomnia in primary care. American Journal of Psychiatry, 154(10), 1417–1423 (Retrieved from http://ajp.psychiatryonline.org/journal.aspx?journalid=13). Smith, M. T., & Wegener, S. T. (2003). Measures of Sleep: The Insomnia Severity Index, Medical Outcomes Study (MOS) Sleep Scale, Pittsburgh Sleep Diary (PSD), and Pittsburgh Sleep Quality Index (PSQI). Arthritis Care & Research, 49(5S), S184–S196. http://dx.doi.org/10.1002/art.11409. Specker, S., de Zwaan, M., Raymond, N., & Mitchell, J. (1994). Psychopathology in subgroups of obese women with and without binge eating disorder. Comprehensive Psychiatry, 35(3), 185–190. http://dx.doi.org/10.1016/0010-440X(94)90190-2. Striegel-Moore, R. H., Franko, D. L., & Garcia, J. (2009). The validity and clinical utility of night eating syndrome. International Journal of Eating Disorders, 42(8), 720–738. http://dx.doi.org/10.1002/eat.20721. Striegel-Moore, R. H., Rosselli, F., Wilson, G. T., Perrin, N., Harvey, K., & DeBar, L. (2010). Nocturnal eating: Association with binge-eating, obesity, and psychological distress. International Journal of Eating Disorders, 43(6), 520–526. http://dx.doi.org/10.1002/ eat.20725. Stunkard, A. J., Grace, W. J., & Wolff, H. G. (1955). The night-eating syndrome: A pattern of food intake among certain obese patients. American Journal of Medicine, 19(1), 78–86 (Retrieved from: http://www.amjmed.com). Stunkard, A., Berkowitz, R., Wadden, T., Tanrikut, C., Reisse, E., & Young, L. (1996). Binge eating disorder and the night-eating syndrome. International Journal of Obesity, 20(1), 1–6. Taylor, D. J., Lichstein, K. L., Durrence, H. H., Reidel, B. W., & Bush, A. J. (2005). Epidemiology of insomnia, depression, and anxiety. Sleep, 28(11), 1457–1464 (Retrieved from http://www.journalsleep.org). Tishler, P. V., Larkin, E. K., Schluchter, M.D., & Redline, S. (2003). Incidence of sleepdisordered breathing in an urban adult population: The relative importance of risk factors in the development of sleep-disordered breathing. Journal of the American Medical Association, 289(17), 2230–2237. http://dx.doi.org/10.1001/jama.289.17. 2230. Ustun, T., Privett, M., Lecrubier, Y., Weiller, E., Simon, G., Korten, A., et al. (1996). Form, frequency and burden of sleep problems in general health care: A report from the WHO Collaborative Study on Psychological Problems in General Health Care. European Psychiatry, 11, 5s–10s. http://dx.doi.org/10.1016/0924-9338(96)80462-7. Vorona, R. D., Winn, M. P., Babineau, T. W., Eng, B. P., Feldman, H. R., & Ware, J. C. (2005). Overweight and obese patients in a primary care population report less sleep than patients with a normal body mass index. Archives of Internal Medicine, 165(1), 25–30. http://dx.doi.org/10.1001/archinte.165.1.25. Westernterp-Plantenga, M. S., Wouters, L., & ten Hoor, F. (1990). Deceleration in cumulative food intake curves, changes in body temperature and diet-induced thermogenesis. Physiology & Behavior, 48(6), 831–836. http://dx.doi.org/10.1016/0031-9384(90) 90235-V. World Health Organization (2000). Part I: The problem of overweight and obesity. Retrieved from World Health Organization website. http://whqlibdoc.who.int/trs/ WHO_TRS_894_(part1).pdf