- Email: [email protected]
Subtyping undergraduate women along dietary restraint and negative affect
Appetite 51 (2008) 727–730
Contents lists available at ScienceDirect
Appetite journal homepage: www.elsevier.com/locate/appet
Short communication
Subtyping undergraduate women along dietary restraint and negative affect ˜ as-Lledo´ a,*, Katharine L. Loeb b,d, Rosa Puerto c, Thomas B. Hildebrandt d, Adria´n Llerena a Eva M. Pen a
University of Extremadura School of Medicine, Avda Elvas sn, 06071 Badajoz, Spain Fairleigh Dickinson University, 1000 River Road, T-WH1-01, Teaneck, NJ 07666, United States c University of Extremadura School of Education, Avda Elvas sn, 06071 Badajoz, Spain d Department of Psychiatry, Mount Sinai School of Medicine, One Gustave L. Levy Place, Box 1230, New York, NY 10029, United States b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 12 December 2007 Received in revised form 22 February 2008 Accepted 15 April 2008
Cluster-analytic studies of clinical populations of eating disordered women who binge eat have yielded two subtypes – pure dietary restraint and mixed dietary restraint-negative affect – consistent with etiologic and maintenance models of binge eating. This study aimed to replicate this subtyping scheme in university women. The scores of 623 undergrad females on the TFEQ-restraint and BDI scales were submitted to a cluster analysis and revealed three subtypes, ‘‘healthy’’ (47.4%), restrained (36.3%) and mixed (16.3%). In addition, comparisons between subtypes on bulimic behaviors showed that the mixed and restrained subtypes were characterized by greater likelihood than the healthy group to engage in fasting, purging and exercise to control weight, as well as in disinhibition of eating. The mixed subtype revealed higher scores than the restrained subtype on eating disinhibition and purging, and the restrained group was more likely than the mixed subtype to exercise to control weight. These findings provide further support for the reliability and validity of this subtyping scheme, in which the confluence of even mild levels of negative affect and dietary restraint differentiated a more ‘‘disturbed’’ group of undergraduate females. Findings also put into question the dietary restraint theory of eating pathology and suggest the need to control negative affect when studying eating behavior. ß 2008 Elsevier Ltd. All rights reserved.
Keywords: Dietary restraint Negative mood Binge eating Fasting Purging Exercise
Restrained eating has long been considered one of the most potent and consistent risk factors for binge eating and related bulimic pathology in young females. However, negative mood on its own or in interaction with restraint is also shown to predict binge eating and other related eating symptoms (Grilo, 2004; Peterson, Miller, Crow, Thuras, & Mitchell, 2005; Stice & Agras, 1999; Steiger et al., 2005; Stice et al., 2001; Van Strien, Engels, Van Leeuwe, & Snoek, 2005). Restrained eating refers to the volitional and sustained effort to restrict caloric intake as a weight control technique, either to lose weight or avoid weight gain (Laessle, Tuschl, Kotthaus, & Pirke, 1989). Dietary restraint is hypothesized to increase the risk for onset and maintenance of binge eating and bulimic symptoms (Polivy & Herman, 1985) via metabolic and psychological pathways (Lowe, 2002). This hypothesized restraintinduced loss of control over eating, in turn, may precipitate other compensatory behaviors to control weight such as exercise, fasting and purging. However, despite the vast evidence prospectively supporting this leading model on the link between dietary
* Corresponding author at: University of Extremadura School of Medicine and ˜ o de Salud, Hospital UniClinical Research Center (CICAB), Servicio Extremen versitario Infanta Cristina, Avda Elvas sn, E-06071 Badajoz, Spain. ˜ as-Lledo´). E-mail address: [email protected] (E.M. Pen 0195-6663/$ – see front matter ß 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.appet.2008.04.014
restraint and binge eating and related disordered symptoms (Leon, Fulkerson, Perry, Keel, & Klump, 1999; Santonastaso, Friederici, & Favaro, 1999), experimental studies do not replicate it (Lowe & Kleifield, 1988; Sysko, Walsh, & Wilson, 2007). Moreover, successful dietary restriction (high restraint in combination with a 10% decrease in BMI) has been shown to decrease binge eating and related psychopathology (Stice, Martinez, Presnell, & Groesz, 2006), suggesting that restraint may not be a necessary precursor of binge eating, as has been shown for a percentage of patients (Bulik, Sullivan, Carter, & Joyce, 1997), that restraint might in fact have a protective quality, and/or that restraint is a multidimensional construct with the potential for negative and positive effects. By extension, the validity of one of the most used measures to test this model, the Restraint scale of the Three Factor Eating Questionnaire (TFEQ-R, Stunkard & Messick, 1985), has been recently put into question (Stice, Cooper, Schoeller, Tappe, & Lowe, 2007; Stice, Fisher, & Lowe, 2004), since it is not negatively correlated with objective measures of short to long-term caloric intake. The latent construct measured by the TFEQ-R remains unknown. One possibility is that it measures intention to restrict rather than actual restrictive eating behavior. On the other hand, negative mood is emerging as an important factor to explain binge eating in both restrained (Steiger et al., 2005) and unrestrained eaters (Peterson et al., 2005). Furthermore,
728
E.M. Pen˜as-Lledo´ et al. / Appetite 51 (2008) 727–730
depression, which is frequently comorbid in eating disorder (ED) patients, has been the only consistent factor across studies associated with worse outcome (Berkman, Lohr, & Bulik, 2007) in bulimia nervosa (BN), an ED characterized by recurrent binge eating followed by maladaptive weight control methods such as purging and fasting (DSM-IV, American Psychiatric Association, 2000). In support of the key role of negative mood in binge eating syndromes, cluster-analytic studies have consistently yielded two reliable diagnostic subtypes, the pure restraint and the mixed restraint-negative mood, where the latter is representative of about one third of ED patients who binge eat (BN and Binge Eating Disorder) and signals a more severe variant of these disorders and a poorer outcome (Chen & Le Grange, 2007; Grilo, 2004; Stice & Agras, 1999; Stice et al., 2001). Because these diet-negative affect subtypes are consistent with etiologic and maintenance models of binge eating, they may capture not only clinical but also at-risk individuals. Therefore, the present study first examines if a large population of undergraduate females can be also categorized into three different subtypes: the two previously found in clinical samples of eating disordered women, the dietaryrestraint and the mixed dietary-restraint-negative mood, and a ‘‘healthy’’ one characterized by low scores on dietary restraint and negative affect. Secondly, it examines the validity of this subtyping scheme by comparing these subtypes on self-reported disinhibition of eating and weekly frequency of binge eating (BE) and weight control behaviors (fasting, purging and exercise). It was hypothesized that women in the restrained group will report significantly more concurrent eating disorder pathology than those in the healthier group, and that the mixed group would report greater severity of eating pathology than the other two groups.
Another item was included at the end of BITE severity scale to assess frequency of exercise behavior to control weight. Beck Depression Inventory (BDI; Beck, Rush, Shaw, & Emery, 1979) is a 21-item self-report inventory used to assess the severity of cognitive, behavioral, affective, and somatic symptoms related to negative mood. For each item, respondents select among four responses ranging from 0 to 3 (from absence to severity of symptoms), where a score between 10 and 18 indicates mild depression, and higher, moderate to severe. Data analyses cluster Participants’ scores on the TFEQ-R scale and the BDI scales were submitted to a cluster analysis as in previous studies (Stice et al., 2001). Cluster analysis groups cases on the basis of similarity in absolute levels on one or more variables. We performed a K-Means cluster analysis using the Quick Cluster algorithm from SPSS. The K-Means algorithm selects k participants, where k is the number of clusters requested. Given our a priori hypothesis, three clusters were specified. Squared Euclidean distances were used as the distance measure, and a nonhierarchical (iterative) clustering algorithm, rather than a hierarchical algorithm, because the latter does not detect and correct early misclassification and is more sensitive to the presence of outliers. A multivariate analysis of variance with Bonferroni corrections was conducted to compare the resulting subtypes on TFEQ-R and BDI scores, as well as on bulimic behaviors derived from the BITE severity scale, and the TFEQ-eating disinhibition. In addition, Fisher’s exact tests were used to compare differences between subtypes on categorical data (rate of women within each subtype that engaged in each bulimic behavior). Results
Methods Cluster analysis Participants Participants were volunteer undergraduate women from the University of Extremadura. There were no exclusion criteria. Six hundred twenty-three women between the ages of 18 and 28 (mean 20.7 2.1) participated in the study after giving their written informed consent. All were Caucasian. Reported mean body mass index was 21.4 3.0 (range 18.0–37.0) calculated from weight (kg)/ height (m2). This protocol was approved by the University of Extremadura ethics committee. Measures They completed Spanish translations of three self-report measures in the order listed below. All measures have demonstrated acceptable internal consistency and test–retest reliability (e.g., Laessle et al., 1989; Rivas, Bersabe´, & Jime´nez, 2004; Sanz & Vazquez, 1998), and in all cases higher scores reflect more pathological attitudes and behaviors. Three-Factor Eating Questionnaire (TFEQ; Stunkard & Messick, 1985) is a self-report measure with three subscales. Only the cognitive (dietary) restraint of eating, TFEQ-R, and the emotionally based disinhibition of eating subscales were included for the present study. Bulimic Investigatory Test, Edinburgh (BITE; Henderson & Freeman, 1987). For the present study only the BITE severity scale was used to determine the frequency of binge eating, fasting, and purging behaviors (use of vomiting, diet pills, laxatives and diuretics) to control weight. The BITE does not distinguish objectively large binge eating and subjectively large episodes.
Table 1 shows a three factor solution for this population: the first cluster (47.4%) reflected a ‘‘healthy’’ subtype characterized by low restraint and low negative mood, the second cluster reflected a ‘‘dietary-restraint’’ subtype with high restraint and low negative affect (36.3%), whereas the third cluster (16.4%) represented a ‘‘mixed’’ group with affective disturbances and moderate dietary restraint. Thus approximately half of the population (N = 328, 52.6%) presented with moderate to high dietary restraint in this study, 31% of which also presented with moderate negative mood. The mean TFEQ-R score for the restraint subtype (M = 12.5) is close to and within one standard deviation of the mean for dieters (M = 14.3) in the original validation study for this measure (Stunkard & Messick, 1985); the mean TFEQ-R score for the mixed subtype (M = 7.8) falls between the means for non-dieting group (M = 6.0) and dieting groups in the original study. The mean BDI score (M = 14.3) for the mixed subtype falls within the mild depression range (10–18; Beck et al., 1979). Comparison of the three subtypes on bulimic behaviors Table 1 also shows the rate of females (expressed in percentage) within each group that engaged in each of the bulimic behaviors. The rate of women who engaged in BE, fasting, and purging was similar across the mixed and dietary restraint subtypes, and significantly higher in these groups than in the ‘‘healthy’’ group, who also reported a high rate of individuals engaging in BE at least once a month. With regard to exercise, a higher rate of women in the pure restrained group reported using exercise to control weight, which was higher than the rate of women in the mixed
E.M. Pen˜as-Lledo´ et al. / Appetite 51 (2008) 727–730
729
Table 1 Means (M), standard deviations (SD) and differences (*p < .05; **p < .01) on the TFEQ-Restraint and BDI scales as well as on the TFEQ-Eating Disinhibition and bulimic behaviors (binge eating, fasting, purging and exercise) for the three subtypes (Healthy, H; Restraint, R and Mixed, M, which is the confluence of restraint and negative affect) derived from the cluster analysis of undergraduate females (N = 623)
TFEQ-Restraint BDI TFEQ-Disinhibition Rate
Healthy (H; N = 295)
Restraint (R; N = 226)
Mixed (M; N = 102)
M (SD)
M (SD)
M (SD)
4.10 (2.24) 3.68 (2.81) 3.71 (2.38)
12.5 (2.68) 4.43 (3.08) 4.68 (2.83)
7.77 (4.25) 14.3 (4.98) 6.36 (3.54)
Healthy (H; N = 295)
Rate
M (SD) BITE severity Binge eating Fasting Purging total Exercise
66% 11% 6.5% 55.1%
1.19 0.07 0.09 1.04
(1.15) (0.44) (0.44) (1.22)
Restraint (R; N = 226)
Rate
M (SD)
84.6% 27.9% 22% 85%
1.35 0.60 0.69 1.63
(0.97) (1.07) (1.97) (1.44)
subtype, that in turn was higher than the women in the healthy subtype. As predicted, differences emerged between the restrained and mixed subtypes on the reported frequency of bulimic behaviors. The mixed subtype reported greater eating disinhibition and frequency of use of purging behaviors than the restrained subtype, whereas the restrained subtype reported higher use of exercise to control weight than the mixed one. The mixed and the restrained subtypes compared to the ‘‘healthy’’ subtype were significantly more symptomatic on all assessed behaviors except for BE frequency, which differentiated only the mixed subtype versus controls. Discussion Evidence from restraint-affect subtyping studies have reliably found two subtypes: the restrained and the mixed restrainednegative affect (Chen & Le Grange, 2007; Grilo, 2004; Stice & Agras, 1999; Stice et al., 2001). However, these studies have focused only on clinical populations where restraint and affect were likely to be high, thus, potentially subject to selection bias. The present study was able to replicate these subtypes in a large non-clinical sample of undergraduate females, where the two restraint groups accounted for approximately half of the participants in the present study, of which approximately one-third represented the mixed subtype as expected and as seen (33–37%) in previous clinical studies (Chen & Le Grange, 2007; Grilo, Masheb, & Wilson, 2001; Stice et al., 2001). The remaining participants constituted a lowrestraint group with little associated pathology, suggesting that the restraint groups can effectively be differentiated from lowrestraint women. The identification of this low-restraint group also raises questions as to whether women move between these groups over time and what factors may contribute to this movement. Such research is likely to be important in assessing the predictive validity of this subtyping model in the onset of binge eating. The differences between subtypes were also similar to those found in previous studies. The rate of women who engaged in all bulimic behaviors assessed was higher in the restraint and mixed subtypes than in the ‘‘healthy’’ subtype. However, differences between the restraint and mixed subtypes showed that the mixed subtype reported more severe purging behaviors and disinhibited eating than the restraint subtype. The restrained subtype had the highest rate of women engaged in exercise behavior to control weight and the highest frequency of use. The present finding appears to suggest that the construct of cognitive dietary restraint from the TFEQ is primarily associated with the use of exercise to control
Mixed (M; N = 102)
Variance explained
MANOVA
63%** 57%** 10%**
H
Fisher’s exact
Variance explained
MANOVA
H < R**; H < M* H < R**; H < M** H < R**; H < M** H < M < R*
1.4%** 10%** 5.5%** 11%**
H
M (SD)
78% 34.3% 27.8% 72.7%
1.61 0.79 1.22 2.13
(1.26) (1.25) (2.95) (1.47)
weight but also with fasting, and to some extent with purging, particularly in confluence with negative mood. Thus, the TFEQ-R may assess the tendency to engage in more adaptive methods of weight and mood regulation, such as exercise, as well as maladaptive weight control methods (fasting or skipping meals). Furthermore, given the lower TFEQ-R scores in the mixed group relative to the restrained group, it appears that the experience of loss of control over eating (BE and TFEQ Disinhibition) may be driven by a function of high negative affect, as shown in non-clinical and clinical samples with binge eating disorders (Stice et al., 2001; Van Strien et al., 2005). Consequently, these findings provide some insights into the controversial results found in experimental studies of binge eating behavior where mood state was not adequately controlled and the link between dietary restraint and behavior was inconsistent (Lowe & Kleifield, 1988; Sysko et al., 2007). There are a number of limitations to the current study. The study relied on a cross-sectional design and self-report measures. Additionally, the BITE severity scale measures the subjective aspect of binge eating but not the objective one (i.e., size of eating episode) making it difficult to generalize results to other populations. This may be an advantage, however, as loss of control appears to be a key aspect of eating pathology in both large and small amounts of food (Latner, Hildebrandt, Rosewall, Chisholm, & Hayashi, 2007). The strength of this study was the size of the sample. These findings generally replicate those of previous studies testing Stice’s subtyping scheme. Future studies in non-clinical females should rely on longitudinal designs to provide evidence about the predictive validity of these subtypes as seen in studies with clinical samples (e.g., Stice & Fairburn, 2003) and more reliable interview measures for assessing in depth the full range of bulimic behaviors considering that the high rates of self-reported eating pathology may represent an overestimation. Notwithstanding these limitations, findings from this large population of undergraduate females suggest the importance of the confluence of dietary restraint and negative mood to explain severity of bulimic behaviors. In addition, results from this study suggest that future research utilizing college undergraduates should account for this observed heterogeneity in dietary restraint when studying eating behavior as it may have important implications for explaining discrepancies in laboratory studies. Acknowledgements This study was supported by a grant from Spanish Ministry of Health, Instituto Carlos III-FIS (PI06/1681) and European Union
730
E.M. Pen˜as-Lledo´ et al. / Appetite 51 (2008) 727–730
FEDER and the Network RETICS: Red de Enfermedades Mentales (RD06-0011REM-TAP).
References American Psychiatric Association (2000). Diagnostic and statistical manual of mental disorders (4th ed., text revision). Washington, DC: American Psychiatric Association. Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive therapy of depression: A treatment manual. New York: Guilford Press. Berkman, N. D., Lohr, K. N., & Bulik, C. M. (2007). Outcomes of eating disorders: A systematic review of the literature. International Journal of Eating Disorders, 40, 293–309. Bulik, C. M., Sullivan, P. F., Carter, F. A., & Joyce, P. R. (1997). Initial manifestations of disordered eating behavior: Dieting versus binging. International Journal of Eating Disorders, 22, 195–201. Chen, E. Y., & Le Grange, D. (2007). Subtyping adolescents with bulimia nervosa. Behavior Research and Therapy, 45, 2813–2820. Grilo, C. M. (2004). Subtyping female adolescent psychiatric inpatients with features of eating disorders along dietary restraint and negative affect dimensions. Behavior Research and Therapy, 42, 67–78. Grilo, C. M., Masheb, R. M., & Wilson, G. T. (2001). Subtyping binge eating disorder. Journal of Consulting and Clinical Psychology, 69, 1066–1072. Henderson, M., & Freeman, C. P. L. (1987). A self-rating scale for bulimia: The BITE. British Journal of Psychiatry, 150, 18–24. Laessle, R. G., Tuschl, R. J., Kotthaus, B. C., & Pirke, K. M. (1989). A comparison of the validity of three scales for the assessment of dietary restraint. Journal of Abnormal Psychology, 98, 504–507. Latner, J. D., Hildebrandt, T., Rosewall, J. K., Chisholm, A. M., & Hayashi, K. (2007). Loss of control over eating reflects eating disturbances and general psychopathology. Behavior Research and Therapy, 45, 2203–2211. Leon, G. R., Fulkerson, J. A., Perry, C. L., Keel, P. K., & Klump, K. L. (1999). Three to four year prospective evaluation of personality and behavioral risk factors for later disordered eating in adolescent girls and boys. Journal of Youth and Adolescence, 28, 181–196. Lowe, M. R. (2002). Dietary restraint and overeating. In C. G. Fairburn & K. D. Brownell (Eds.), Eating disorders and obesity: A comprehensive handbook (pp. 88–97). New York: The Guildford Press. Lowe, M. R., & Kleifield, E. I. (1988). Cognitive restraint, weight suppression, and the regulation of eating. Appetite, 10, 159–168.
Peterson, C. B., Miller, K. B., Crow, S. J., Thuras, P., & Mitchell, J. E. (2005). Subtypes of binge eating disorder based on psychiatric history. International Journal of Eating Disorders, 38, 273–276. Polivy, J., & Herman, C. P. (1985). Dieting and binging. A causal analysis. American Psychology, 40, 193–201. Rivas, T., Bersabe´, R., & Jime´nez, M. (2004). Fiabilidad y validez del test de investigacio´n bulı´mica de Edimburgo (BITE) en una muestra de adolescentes espan˜oles. Psicologı´a Conductual, 12, 447–461. Santonastaso, P., Friederici, S., & Favaro, A. (1999). Full and partial syndromes in eating disorders: A 1-year prospective study of risk factors among female students. Psychopathology, 32, 50–56. Sanz, J., & Vazquez, C. (1998). Fiabilidad, validez y datos normativos del inventario para la depresio´n de Beck. Psicothema, 10, 303–318. Steiger, H., Gauvin, L., Engelberg, M. J., Ying Kin, N. M., Israel, M., Wonderlich, S. A., et al. (2005). Mood- and restraint-based antecedents to binge episodes in bulimia nervosa: Possible influences of the serotonin system. Psychological Medicine, 35, 1553–1562. Stice, E., & Agras, W. S. (1999). Subtyping bulimic women along dietary restraint and negative affect dimensions. Journal of Consulting and Clinical Psychology, 67, 460– 469. Stice, E., Agras, W. S., Telch, C. F., Halmi, K. A., Mitchell, J. E., & Wilson, T. (2001). Subtyping binge eating-disordered women along dieting and negative affect dimensions. International Journal of Eating Disorders, 30, 11–27. Stice, E., Cooper, J. A., Schoeller, D. A., Tappe, K., & Lowe, M. R. (2007). Are dietary restraint scales valid measures of moderate- to long-term dietary restriction? Objective biological and behavioral data suggest not. Psychological Assessment, 19, 449–458. Stice, E., & Fairburn, C. G. (2003). Dietary and dietary-depressive subtypes of bulimia nervosa show differential symptom presentation, social impairment, comorbidity, and course of illness. Journal of Consulting and Clinical Psychology, 71, 1090–1094. Stice, E., Fisher, M., & Lowe, M. R. (2004). Are dietary restraint scales valid measures of acute dietary restriction? Unobtrusive observational data suggest not. Psychological Assessment, 16, 51–59. Stice, E., Martinez, E. E., Presnell, K., & Groesz, L. M. (2006). Relation of successful dietary restriction to change in bulimic symptoms: A prospective study of adolescent girls. Health Psychology, 25, 274–281. Sysko, R., Walsh, T. B., & Wilson, T. G. (2007). Expectancies, dietary restraint, and test meal intake among undergraduate women. Appetite, 49, 30–37. Stunkard, A. J., & Messick, S. (1985). The three-factor eating questionnaire to measure dietary restraint, disinhibition and hunger. Journal of Psychosomatic Research, 29, 71–83. Van Strien, T., Engels, R. C., Van Leeuwe, J., & Snoek, H. M. (2005). The Stice model of overeating: Tests in clinical and non-clinical samples. Appetite, 45, 205–213.
Contents lists available at ScienceDirect
Appetite journal homepage: www.elsevier.com/locate/appet
Short communication
Subtyping undergraduate women along dietary restraint and negative affect ˜ as-Lledo´ a,*, Katharine L. Loeb b,d, Rosa Puerto c, Thomas B. Hildebrandt d, Adria´n Llerena a Eva M. Pen a
University of Extremadura School of Medicine, Avda Elvas sn, 06071 Badajoz, Spain Fairleigh Dickinson University, 1000 River Road, T-WH1-01, Teaneck, NJ 07666, United States c University of Extremadura School of Education, Avda Elvas sn, 06071 Badajoz, Spain d Department of Psychiatry, Mount Sinai School of Medicine, One Gustave L. Levy Place, Box 1230, New York, NY 10029, United States b
A R T I C L E I N F O
A B S T R A C T
Article history: Received 12 December 2007 Received in revised form 22 February 2008 Accepted 15 April 2008
Cluster-analytic studies of clinical populations of eating disordered women who binge eat have yielded two subtypes – pure dietary restraint and mixed dietary restraint-negative affect – consistent with etiologic and maintenance models of binge eating. This study aimed to replicate this subtyping scheme in university women. The scores of 623 undergrad females on the TFEQ-restraint and BDI scales were submitted to a cluster analysis and revealed three subtypes, ‘‘healthy’’ (47.4%), restrained (36.3%) and mixed (16.3%). In addition, comparisons between subtypes on bulimic behaviors showed that the mixed and restrained subtypes were characterized by greater likelihood than the healthy group to engage in fasting, purging and exercise to control weight, as well as in disinhibition of eating. The mixed subtype revealed higher scores than the restrained subtype on eating disinhibition and purging, and the restrained group was more likely than the mixed subtype to exercise to control weight. These findings provide further support for the reliability and validity of this subtyping scheme, in which the confluence of even mild levels of negative affect and dietary restraint differentiated a more ‘‘disturbed’’ group of undergraduate females. Findings also put into question the dietary restraint theory of eating pathology and suggest the need to control negative affect when studying eating behavior. ß 2008 Elsevier Ltd. All rights reserved.
Keywords: Dietary restraint Negative mood Binge eating Fasting Purging Exercise
Restrained eating has long been considered one of the most potent and consistent risk factors for binge eating and related bulimic pathology in young females. However, negative mood on its own or in interaction with restraint is also shown to predict binge eating and other related eating symptoms (Grilo, 2004; Peterson, Miller, Crow, Thuras, & Mitchell, 2005; Stice & Agras, 1999; Steiger et al., 2005; Stice et al., 2001; Van Strien, Engels, Van Leeuwe, & Snoek, 2005). Restrained eating refers to the volitional and sustained effort to restrict caloric intake as a weight control technique, either to lose weight or avoid weight gain (Laessle, Tuschl, Kotthaus, & Pirke, 1989). Dietary restraint is hypothesized to increase the risk for onset and maintenance of binge eating and bulimic symptoms (Polivy & Herman, 1985) via metabolic and psychological pathways (Lowe, 2002). This hypothesized restraintinduced loss of control over eating, in turn, may precipitate other compensatory behaviors to control weight such as exercise, fasting and purging. However, despite the vast evidence prospectively supporting this leading model on the link between dietary
* Corresponding author at: University of Extremadura School of Medicine and ˜ o de Salud, Hospital UniClinical Research Center (CICAB), Servicio Extremen versitario Infanta Cristina, Avda Elvas sn, E-06071 Badajoz, Spain. ˜ as-Lledo´). E-mail address: [email protected] (E.M. Pen 0195-6663/$ – see front matter ß 2008 Elsevier Ltd. All rights reserved. doi:10.1016/j.appet.2008.04.014
restraint and binge eating and related disordered symptoms (Leon, Fulkerson, Perry, Keel, & Klump, 1999; Santonastaso, Friederici, & Favaro, 1999), experimental studies do not replicate it (Lowe & Kleifield, 1988; Sysko, Walsh, & Wilson, 2007). Moreover, successful dietary restriction (high restraint in combination with a 10% decrease in BMI) has been shown to decrease binge eating and related psychopathology (Stice, Martinez, Presnell, & Groesz, 2006), suggesting that restraint may not be a necessary precursor of binge eating, as has been shown for a percentage of patients (Bulik, Sullivan, Carter, & Joyce, 1997), that restraint might in fact have a protective quality, and/or that restraint is a multidimensional construct with the potential for negative and positive effects. By extension, the validity of one of the most used measures to test this model, the Restraint scale of the Three Factor Eating Questionnaire (TFEQ-R, Stunkard & Messick, 1985), has been recently put into question (Stice, Cooper, Schoeller, Tappe, & Lowe, 2007; Stice, Fisher, & Lowe, 2004), since it is not negatively correlated with objective measures of short to long-term caloric intake. The latent construct measured by the TFEQ-R remains unknown. One possibility is that it measures intention to restrict rather than actual restrictive eating behavior. On the other hand, negative mood is emerging as an important factor to explain binge eating in both restrained (Steiger et al., 2005) and unrestrained eaters (Peterson et al., 2005). Furthermore,
728
E.M. Pen˜as-Lledo´ et al. / Appetite 51 (2008) 727–730
depression, which is frequently comorbid in eating disorder (ED) patients, has been the only consistent factor across studies associated with worse outcome (Berkman, Lohr, & Bulik, 2007) in bulimia nervosa (BN), an ED characterized by recurrent binge eating followed by maladaptive weight control methods such as purging and fasting (DSM-IV, American Psychiatric Association, 2000). In support of the key role of negative mood in binge eating syndromes, cluster-analytic studies have consistently yielded two reliable diagnostic subtypes, the pure restraint and the mixed restraint-negative mood, where the latter is representative of about one third of ED patients who binge eat (BN and Binge Eating Disorder) and signals a more severe variant of these disorders and a poorer outcome (Chen & Le Grange, 2007; Grilo, 2004; Stice & Agras, 1999; Stice et al., 2001). Because these diet-negative affect subtypes are consistent with etiologic and maintenance models of binge eating, they may capture not only clinical but also at-risk individuals. Therefore, the present study first examines if a large population of undergraduate females can be also categorized into three different subtypes: the two previously found in clinical samples of eating disordered women, the dietaryrestraint and the mixed dietary-restraint-negative mood, and a ‘‘healthy’’ one characterized by low scores on dietary restraint and negative affect. Secondly, it examines the validity of this subtyping scheme by comparing these subtypes on self-reported disinhibition of eating and weekly frequency of binge eating (BE) and weight control behaviors (fasting, purging and exercise). It was hypothesized that women in the restrained group will report significantly more concurrent eating disorder pathology than those in the healthier group, and that the mixed group would report greater severity of eating pathology than the other two groups.
Another item was included at the end of BITE severity scale to assess frequency of exercise behavior to control weight. Beck Depression Inventory (BDI; Beck, Rush, Shaw, & Emery, 1979) is a 21-item self-report inventory used to assess the severity of cognitive, behavioral, affective, and somatic symptoms related to negative mood. For each item, respondents select among four responses ranging from 0 to 3 (from absence to severity of symptoms), where a score between 10 and 18 indicates mild depression, and higher, moderate to severe. Data analyses cluster Participants’ scores on the TFEQ-R scale and the BDI scales were submitted to a cluster analysis as in previous studies (Stice et al., 2001). Cluster analysis groups cases on the basis of similarity in absolute levels on one or more variables. We performed a K-Means cluster analysis using the Quick Cluster algorithm from SPSS. The K-Means algorithm selects k participants, where k is the number of clusters requested. Given our a priori hypothesis, three clusters were specified. Squared Euclidean distances were used as the distance measure, and a nonhierarchical (iterative) clustering algorithm, rather than a hierarchical algorithm, because the latter does not detect and correct early misclassification and is more sensitive to the presence of outliers. A multivariate analysis of variance with Bonferroni corrections was conducted to compare the resulting subtypes on TFEQ-R and BDI scores, as well as on bulimic behaviors derived from the BITE severity scale, and the TFEQ-eating disinhibition. In addition, Fisher’s exact tests were used to compare differences between subtypes on categorical data (rate of women within each subtype that engaged in each bulimic behavior). Results
Methods Cluster analysis Participants Participants were volunteer undergraduate women from the University of Extremadura. There were no exclusion criteria. Six hundred twenty-three women between the ages of 18 and 28 (mean 20.7 2.1) participated in the study after giving their written informed consent. All were Caucasian. Reported mean body mass index was 21.4 3.0 (range 18.0–37.0) calculated from weight (kg)/ height (m2). This protocol was approved by the University of Extremadura ethics committee. Measures They completed Spanish translations of three self-report measures in the order listed below. All measures have demonstrated acceptable internal consistency and test–retest reliability (e.g., Laessle et al., 1989; Rivas, Bersabe´, & Jime´nez, 2004; Sanz & Vazquez, 1998), and in all cases higher scores reflect more pathological attitudes and behaviors. Three-Factor Eating Questionnaire (TFEQ; Stunkard & Messick, 1985) is a self-report measure with three subscales. Only the cognitive (dietary) restraint of eating, TFEQ-R, and the emotionally based disinhibition of eating subscales were included for the present study. Bulimic Investigatory Test, Edinburgh (BITE; Henderson & Freeman, 1987). For the present study only the BITE severity scale was used to determine the frequency of binge eating, fasting, and purging behaviors (use of vomiting, diet pills, laxatives and diuretics) to control weight. The BITE does not distinguish objectively large binge eating and subjectively large episodes.
Table 1 shows a three factor solution for this population: the first cluster (47.4%) reflected a ‘‘healthy’’ subtype characterized by low restraint and low negative mood, the second cluster reflected a ‘‘dietary-restraint’’ subtype with high restraint and low negative affect (36.3%), whereas the third cluster (16.4%) represented a ‘‘mixed’’ group with affective disturbances and moderate dietary restraint. Thus approximately half of the population (N = 328, 52.6%) presented with moderate to high dietary restraint in this study, 31% of which also presented with moderate negative mood. The mean TFEQ-R score for the restraint subtype (M = 12.5) is close to and within one standard deviation of the mean for dieters (M = 14.3) in the original validation study for this measure (Stunkard & Messick, 1985); the mean TFEQ-R score for the mixed subtype (M = 7.8) falls between the means for non-dieting group (M = 6.0) and dieting groups in the original study. The mean BDI score (M = 14.3) for the mixed subtype falls within the mild depression range (10–18; Beck et al., 1979). Comparison of the three subtypes on bulimic behaviors Table 1 also shows the rate of females (expressed in percentage) within each group that engaged in each of the bulimic behaviors. The rate of women who engaged in BE, fasting, and purging was similar across the mixed and dietary restraint subtypes, and significantly higher in these groups than in the ‘‘healthy’’ group, who also reported a high rate of individuals engaging in BE at least once a month. With regard to exercise, a higher rate of women in the pure restrained group reported using exercise to control weight, which was higher than the rate of women in the mixed
E.M. Pen˜as-Lledo´ et al. / Appetite 51 (2008) 727–730
729
Table 1 Means (M), standard deviations (SD) and differences (*p < .05; **p < .01) on the TFEQ-Restraint and BDI scales as well as on the TFEQ-Eating Disinhibition and bulimic behaviors (binge eating, fasting, purging and exercise) for the three subtypes (Healthy, H; Restraint, R and Mixed, M, which is the confluence of restraint and negative affect) derived from the cluster analysis of undergraduate females (N = 623)
TFEQ-Restraint BDI TFEQ-Disinhibition Rate
Healthy (H; N = 295)
Restraint (R; N = 226)
Mixed (M; N = 102)
M (SD)
M (SD)
M (SD)
4.10 (2.24) 3.68 (2.81) 3.71 (2.38)
12.5 (2.68) 4.43 (3.08) 4.68 (2.83)
7.77 (4.25) 14.3 (4.98) 6.36 (3.54)
Healthy (H; N = 295)
Rate
M (SD) BITE severity Binge eating Fasting Purging total Exercise
66% 11% 6.5% 55.1%
1.19 0.07 0.09 1.04
(1.15) (0.44) (0.44) (1.22)
Restraint (R; N = 226)
Rate
M (SD)
84.6% 27.9% 22% 85%
1.35 0.60 0.69 1.63
(0.97) (1.07) (1.97) (1.44)
subtype, that in turn was higher than the women in the healthy subtype. As predicted, differences emerged between the restrained and mixed subtypes on the reported frequency of bulimic behaviors. The mixed subtype reported greater eating disinhibition and frequency of use of purging behaviors than the restrained subtype, whereas the restrained subtype reported higher use of exercise to control weight than the mixed one. The mixed and the restrained subtypes compared to the ‘‘healthy’’ subtype were significantly more symptomatic on all assessed behaviors except for BE frequency, which differentiated only the mixed subtype versus controls. Discussion Evidence from restraint-affect subtyping studies have reliably found two subtypes: the restrained and the mixed restrainednegative affect (Chen & Le Grange, 2007; Grilo, 2004; Stice & Agras, 1999; Stice et al., 2001). However, these studies have focused only on clinical populations where restraint and affect were likely to be high, thus, potentially subject to selection bias. The present study was able to replicate these subtypes in a large non-clinical sample of undergraduate females, where the two restraint groups accounted for approximately half of the participants in the present study, of which approximately one-third represented the mixed subtype as expected and as seen (33–37%) in previous clinical studies (Chen & Le Grange, 2007; Grilo, Masheb, & Wilson, 2001; Stice et al., 2001). The remaining participants constituted a lowrestraint group with little associated pathology, suggesting that the restraint groups can effectively be differentiated from lowrestraint women. The identification of this low-restraint group also raises questions as to whether women move between these groups over time and what factors may contribute to this movement. Such research is likely to be important in assessing the predictive validity of this subtyping model in the onset of binge eating. The differences between subtypes were also similar to those found in previous studies. The rate of women who engaged in all bulimic behaviors assessed was higher in the restraint and mixed subtypes than in the ‘‘healthy’’ subtype. However, differences between the restraint and mixed subtypes showed that the mixed subtype reported more severe purging behaviors and disinhibited eating than the restraint subtype. The restrained subtype had the highest rate of women engaged in exercise behavior to control weight and the highest frequency of use. The present finding appears to suggest that the construct of cognitive dietary restraint from the TFEQ is primarily associated with the use of exercise to control
Mixed (M; N = 102)
Variance explained
MANOVA
63%** 57%** 10%**
H
Fisher’s exact
Variance explained
MANOVA
H < R**; H < M* H < R**; H < M** H < R**; H < M** H < M < R*
1.4%** 10%** 5.5%** 11%**
H
M (SD)
78% 34.3% 27.8% 72.7%
1.61 0.79 1.22 2.13
(1.26) (1.25) (2.95) (1.47)
weight but also with fasting, and to some extent with purging, particularly in confluence with negative mood. Thus, the TFEQ-R may assess the tendency to engage in more adaptive methods of weight and mood regulation, such as exercise, as well as maladaptive weight control methods (fasting or skipping meals). Furthermore, given the lower TFEQ-R scores in the mixed group relative to the restrained group, it appears that the experience of loss of control over eating (BE and TFEQ Disinhibition) may be driven by a function of high negative affect, as shown in non-clinical and clinical samples with binge eating disorders (Stice et al., 2001; Van Strien et al., 2005). Consequently, these findings provide some insights into the controversial results found in experimental studies of binge eating behavior where mood state was not adequately controlled and the link between dietary restraint and behavior was inconsistent (Lowe & Kleifield, 1988; Sysko et al., 2007). There are a number of limitations to the current study. The study relied on a cross-sectional design and self-report measures. Additionally, the BITE severity scale measures the subjective aspect of binge eating but not the objective one (i.e., size of eating episode) making it difficult to generalize results to other populations. This may be an advantage, however, as loss of control appears to be a key aspect of eating pathology in both large and small amounts of food (Latner, Hildebrandt, Rosewall, Chisholm, & Hayashi, 2007). The strength of this study was the size of the sample. These findings generally replicate those of previous studies testing Stice’s subtyping scheme. Future studies in non-clinical females should rely on longitudinal designs to provide evidence about the predictive validity of these subtypes as seen in studies with clinical samples (e.g., Stice & Fairburn, 2003) and more reliable interview measures for assessing in depth the full range of bulimic behaviors considering that the high rates of self-reported eating pathology may represent an overestimation. Notwithstanding these limitations, findings from this large population of undergraduate females suggest the importance of the confluence of dietary restraint and negative mood to explain severity of bulimic behaviors. In addition, results from this study suggest that future research utilizing college undergraduates should account for this observed heterogeneity in dietary restraint when studying eating behavior as it may have important implications for explaining discrepancies in laboratory studies. Acknowledgements This study was supported by a grant from Spanish Ministry of Health, Instituto Carlos III-FIS (PI06/1681) and European Union
730
E.M. Pen˜as-Lledo´ et al. / Appetite 51 (2008) 727–730
FEDER and the Network RETICS: Red de Enfermedades Mentales (RD06-0011REM-TAP).
References American Psychiatric Association (2000). Diagnostic and statistical manual of mental disorders (4th ed., text revision). Washington, DC: American Psychiatric Association. Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive therapy of depression: A treatment manual. New York: Guilford Press. Berkman, N. D., Lohr, K. N., & Bulik, C. M. (2007). Outcomes of eating disorders: A systematic review of the literature. International Journal of Eating Disorders, 40, 293–309. Bulik, C. M., Sullivan, P. F., Carter, F. A., & Joyce, P. R. (1997). Initial manifestations of disordered eating behavior: Dieting versus binging. International Journal of Eating Disorders, 22, 195–201. Chen, E. Y., & Le Grange, D. (2007). Subtyping adolescents with bulimia nervosa. Behavior Research and Therapy, 45, 2813–2820. Grilo, C. M. (2004). Subtyping female adolescent psychiatric inpatients with features of eating disorders along dietary restraint and negative affect dimensions. Behavior Research and Therapy, 42, 67–78. Grilo, C. M., Masheb, R. M., & Wilson, G. T. (2001). Subtyping binge eating disorder. Journal of Consulting and Clinical Psychology, 69, 1066–1072. Henderson, M., & Freeman, C. P. L. (1987). A self-rating scale for bulimia: The BITE. British Journal of Psychiatry, 150, 18–24. Laessle, R. G., Tuschl, R. J., Kotthaus, B. C., & Pirke, K. M. (1989). A comparison of the validity of three scales for the assessment of dietary restraint. Journal of Abnormal Psychology, 98, 504–507. Latner, J. D., Hildebrandt, T., Rosewall, J. K., Chisholm, A. M., & Hayashi, K. (2007). Loss of control over eating reflects eating disturbances and general psychopathology. Behavior Research and Therapy, 45, 2203–2211. Leon, G. R., Fulkerson, J. A., Perry, C. L., Keel, P. K., & Klump, K. L. (1999). Three to four year prospective evaluation of personality and behavioral risk factors for later disordered eating in adolescent girls and boys. Journal of Youth and Adolescence, 28, 181–196. Lowe, M. R. (2002). Dietary restraint and overeating. In C. G. Fairburn & K. D. Brownell (Eds.), Eating disorders and obesity: A comprehensive handbook (pp. 88–97). New York: The Guildford Press. Lowe, M. R., & Kleifield, E. I. (1988). Cognitive restraint, weight suppression, and the regulation of eating. Appetite, 10, 159–168.
Peterson, C. B., Miller, K. B., Crow, S. J., Thuras, P., & Mitchell, J. E. (2005). Subtypes of binge eating disorder based on psychiatric history. International Journal of Eating Disorders, 38, 273–276. Polivy, J., & Herman, C. P. (1985). Dieting and binging. A causal analysis. American Psychology, 40, 193–201. Rivas, T., Bersabe´, R., & Jime´nez, M. (2004). Fiabilidad y validez del test de investigacio´n bulı´mica de Edimburgo (BITE) en una muestra de adolescentes espan˜oles. Psicologı´a Conductual, 12, 447–461. Santonastaso, P., Friederici, S., & Favaro, A. (1999). Full and partial syndromes in eating disorders: A 1-year prospective study of risk factors among female students. Psychopathology, 32, 50–56. Sanz, J., & Vazquez, C. (1998). Fiabilidad, validez y datos normativos del inventario para la depresio´n de Beck. Psicothema, 10, 303–318. Steiger, H., Gauvin, L., Engelberg, M. J., Ying Kin, N. M., Israel, M., Wonderlich, S. A., et al. (2005). Mood- and restraint-based antecedents to binge episodes in bulimia nervosa: Possible influences of the serotonin system. Psychological Medicine, 35, 1553–1562. Stice, E., & Agras, W. S. (1999). Subtyping bulimic women along dietary restraint and negative affect dimensions. Journal of Consulting and Clinical Psychology, 67, 460– 469. Stice, E., Agras, W. S., Telch, C. F., Halmi, K. A., Mitchell, J. E., & Wilson, T. (2001). Subtyping binge eating-disordered women along dieting and negative affect dimensions. International Journal of Eating Disorders, 30, 11–27. Stice, E., Cooper, J. A., Schoeller, D. A., Tappe, K., & Lowe, M. R. (2007). Are dietary restraint scales valid measures of moderate- to long-term dietary restriction? Objective biological and behavioral data suggest not. Psychological Assessment, 19, 449–458. Stice, E., & Fairburn, C. G. (2003). Dietary and dietary-depressive subtypes of bulimia nervosa show differential symptom presentation, social impairment, comorbidity, and course of illness. Journal of Consulting and Clinical Psychology, 71, 1090–1094. Stice, E., Fisher, M., & Lowe, M. R. (2004). Are dietary restraint scales valid measures of acute dietary restriction? Unobtrusive observational data suggest not. Psychological Assessment, 16, 51–59. Stice, E., Martinez, E. E., Presnell, K., & Groesz, L. M. (2006). Relation of successful dietary restriction to change in bulimic symptoms: A prospective study of adolescent girls. Health Psychology, 25, 274–281. Sysko, R., Walsh, T. B., & Wilson, T. G. (2007). Expectancies, dietary restraint, and test meal intake among undergraduate women. Appetite, 49, 30–37. Stunkard, A. J., & Messick, S. (1985). The three-factor eating questionnaire to measure dietary restraint, disinhibition and hunger. Journal of Psychosomatic Research, 29, 71–83. Van Strien, T., Engels, R. C., Van Leeuwe, J., & Snoek, H. M. (2005). The Stice model of overeating: Tests in clinical and non-clinical samples. Appetite, 45, 205–213.