Food addiction and lifetime alcohol and illicit drugs use in specific eating disorders ROMINA MIRANDA-OLIVOS1,2,3 , ZAIDA AGÜERA1,2,3,4p , ROSER GRANERO1,3,5 , RHIANNA R. VERGEER6, Journal of Behavioral CARLOS DIEGUEZ1,7 , SUSANA JIMÉNEZ-MURCIA1,2,3,8 , Addictions ASHLEY N. GEARHARDT9 and 11 (2022) 1, 102–115 FERNANDO FERNÁNDEZ-ARANDA1,2,3,8pp DOI: 10.1556/2006.2021.00087 1 CIBER Fisiopatología Obesidad y Nutrición (CIBERobn), Instituto de Salud Carlos III, Madrid, © 2021 The Author(s) Spain 2 Department of Psychiatry, Hospital Universitari de Bellvitge, L’Hospitalet de Llobregat, Spain 3 Psychoneurobiology of Eating and Addictive Behaviors Group, Neurosciences Programme, Bellvitge Biomedical Research Institute (IDIBELL), L’Hospitalet de Llobregat, Spain 4 Department of Public Health, Mental Health and Perinatal Nursing, School of Nursing, University FULL-LENGTH REPORT of Barcelona, L’Hospitalet de Llobregat, Spain 5 Departament de Psicobiologia i Metodologia de les Ciències de la Salut, Universitat Autònoma de Barcelona, Barcelona, Spain 6 Laboratory on Human Psychopharmacology, National Institute on Alcohol Abuse and Alcoholism, Bethesda, MD, USA 7 Instituto de Investigaciones Sanitarias (IDIS), Department of Physiology (CIMUS), Universidad de Santiago de Compostela, Santiago de Compostela, Spain 8 Department of Clinical Sciences, School of Medicine and Health Sciences, University of Barcelona, Barcelona, Spain 9 Department of Psychology, University of Michigan, Ann Arbor, MI, USA Received: June 7, 2021 • Revised manuscript received: October 21, 2021; December 9, 2021 • Accepted: December 16, 2021 Published online: January 13, 2022 ABSTRACT Background and aims: Food addiction (FA) and substance use (SU) have frequently been reported in patients with eating disorders (EDs). Our study aimed to assess the prevalence rates of FA and/or lifetime problematic alcohol and illicit drug use among patients with specific ED, such as: bulimia nervosa (BN), binge eating disorder (BED), and other specified feeding and eating disorder (OSFED). We sought to identify clinical, psychopathological, and personality profiles involved in these addictive behavior-based phenotypes. Methods: The total sample was 527 patients (176 BN, 115 BED, and 236 OSFED). FA was assessed through the Yale Food Addiction Scale 2.0. To determine lifetime SU, a semi- structured clinical interview was carried out. Results: Patients with BN had the highest rates of FA both with and without SU. No gender differences were obtained for the prevalence of current FA and/or lifetime SU. Patients reporting at least one addictive-related behavior exhibited increased clinical severity compared to those who reported none. Increased impulsivity (such as high lack of premedi- tation, sensation seeking, and positive urgency) and low self-directedness were differentiating factors for presenting one or two addictive behaviors. Discussion and Conclusions: Overall, patients presenting with p at least one addictive-like behavior reported a poorer clinical status than those without. Also, patients Corresponding author. E-mail: zaguera@ub.edu with FA and SU exhibited a more dysfunctional profile characterized by high impulsivity and low self- directedness. These findings would support the need for targeted treatments to reduce impulsivity and pp Corresponding author. increase self-directedness, especially in patients with any addictive-related behavior, as a step towards E-mail: ffernandez@bellvitgehospital. improving their treatment outcome. cat KEYWORDS eating disorders, food addiction, impulsivity, personality, substance use Journal of Behavioral Addictions 11 (2022) 1, 102–115 103 INTRODUCTION Michopoulos, 2018; Hebebrand et al., 2014) and behavioral addictions (Um, Whitt, Revilla, Hunton, & Cyders, 2019). It has even been proposed that interventions for other SUDs Eating disorders (EDs) are mental illnesses characterized may prove useful in the treatment of FA (Blumenthal & mainly by maladaptive eating behaviors with significant Gold, 2010). Similar to substances of abuse, palatable foods physical and psychosocial impairments. Based on the appear to have reinforcing effects characteristic of addictive Diagnostic and Statistical Manual of Mental Disorders processes (Bonder, Davis, Kuk, & Loxton, 2018; Volkow, (DSM-5) (American Psychiatric Association, 2013), the EDs Wang, Fowler, Tomasi, & Baler, 2011). FA is highly preva- include bulimia nervosa (BN), binge eating disorder (BED), lent in patients with binge eating symptomatology such as anorexia nervosa (AN), and other specified eating disorders BN, BED, and OSFED (Gearhardt, White, & Potenza, 2011; (OSFED), among others. The first two diagnostic categories Granero et al., 2014; Jimenez-Murcia et al., 2019; Romero (BN and BED) are mainly characterized by the presence of et al., 2019). Patients with ED and comorbid FA had binge eating episodes (i.e., excessive food intake in a short increased psychopathology, more severe ED symptom- period accompanied by a sense of loss of control) as the atology, and higher impulsivity traits (Gearhardt, Boswell, & main symptom. The major difference between these two White, 2014; Jimenez-Murcia et al., 2019; Khalil et al., 2020). diagnoses is that the former exhibits compensatory behav- In addition, a recent prospective study identified that FA was iors (such as fasting and/or purging behaviors including self- associated with poor treatment outcome in patients with induced vomiting; laxative or diuretic abuse), and the latter BED (Romero et al., 2019). Furthermore, recent studies does not engage in these behaviors. AN is defined by re- show increasing interest in the relationship between FA and striction of food intake leading to weight loss or failure to AN because of the strikingly high prevalence of FA in these gain weight. On the other hand, OSFED is a formal diag- patients (reporting rates from 47% to 84%) (Fauconnier nostic category that includes heterogeneous nosological en- et al., 2020; Granero et al., 2018; Tran et al., 2020). The study tities that do not meet the full diagnostic criteria of other of Tran et al. (2020) suggests that the comorbid FA may EDs but are no less severe. represent a more severe variant of AN. However, concep- The comorbidity of EDs and substance use disorder tualizing AN-R with FA is somewhat debatable. Patients (SUD) has frequently been reported with rates ranging from with AN, especially the restrictive subtype (AN-R), show 21% to 50%, and varying among the ED diagnostic types (Bahji et al., 2019; Fouladi et al., 2015). Most studies have food deprivation and seem to be the opposite of FA (Fau- reported that comorbid SUD is most prevalent in BN, fol- connier et al., 2020). In this vein, Mallorquí-Bagué et al. lowed by BED, AN, and lastly, eating disorders not other- (2020) suggested that patients with AN showed successful wise specified (EDNOS)/OSFED (Bahji et al., 2019; Fouladi down-regulation of food craving, despite the presence of FA et al., 2015). Overall, comorbid SUD is less common in symptomatology. The question of whether patients with AN patients with restrictive behaviors [namely restrictive AN actually show such high FA rates or, rather, reflect a sub- (AN-R)] compared to those with binge-purging symptom- jective fear of losing control with high-caloric food is atology (Anzengruber et al., 2006; Krug et al., 2009; Root controversial so far. Therefore, the current literature sug- et al., 2010). Although Baker, Mitchell, Neale, and Kendler gests the need for further research to explore how patients (2010) were not able to find differences in the SUD preva- with ED, and especially restrictive AN, can interpret FA lence between women with AN and BN, these authors items (Albayrak et al., 2017). described differences in the chronology of onset of both EDs A general population-based study found that women and SUD. In this regard, while women with BN reported with FA and women with SUD shared similar psychological that their ED symptomatology preceded the development of characteristics, reporting more depression, emotion dysre- SUD, for patients with AN it was the opposite (Baker et al., gulation, lack of impulse control, and difficulties in goal- 2010). Patients with binge-purging behaviors often exhibit directed behaviors when compared with women without high impulsivity and low inhibitory control (Bogusz et al., addiction (Hardy et al., 2018). Similarly, the co-occurrence 2021). These impulsive traits have also been related to food of FA and other addictive-like behaviors has been related to cravings, similar to substance cravings in patients with SUD poorer emotional and psychological states (Canan, Karaca, (Polk, Schulte, Furman, & Gearhardt, 2017). In addition, Sogucak, Gecici, & Kuloglu, 2017; Jiménez-Murcia et al., both pathologies present difficulties in emotion regulation, 2017; Tinghino et al., 2021). Most studies have separately which acts as a common underlying mechanism that im- explored SUD or FA among patients with EDs so far pacts impulse control and predisposes to the development of (Jimenez-Murcia et al., 2019; Lozano-Madrid et al., 2020; addictive patterns (Gregorowski, Seedat, & Jordaan, 2013; Romero et al., 2019). However, it could be interesting to Lozano-Madrid et al., 2020). analyze the co-occurrence of different addictive behaviors Similarly, in the last years, there has been a growing given the common risk factors observed at psychological, interest in studying the comorbidity of food addiction (FA) genetic, and neuronal levels, which could potentially have an in EDs. Although FA has not yet been recognized as a important epidemiological relevance (Kotyuk et al., 2020). In diagnostic entity per se, it seems to share some clinical EDs, there are hardly any studies examining the co-occur- characteristics and neurobiological mechanisms (e.g. reward rence of both addictive behaviors (i.e., SUD and FA) among system and the dopaminergic pathways) with other addic- EDs so far. A study conducted with men with heroin use tions, both substance-related (Hardy, Fani, Jovanovic, & disorder reported a high co-occurrence of BED and FA 104 Journal of Behavioral Addictions 11 (2022) 1, 102–115 which, in turn, was associated with increased craving and different ED-related psychopathological characteristics, such more suicide attempts (Canan et al., 2017). Likewise, a as: drive for thinness, body dissatisfaction, bulimia, ineffec- recent study observed overlapping risk factors linked to tiveness, perfectionism, interpersonal distrust, interoceptive impulsive personality traits in youth with FA or alcohol awareness, maturity fears, asceticism, impulse regulation, and misuse (Minhas et al., 2021). Further, concurrent FA and social insecurity. For the current sample, the internal consis- alcohol use disorder can predict binge eating in women from tency was excellent for the EDI-2 total score (a 5 0.95), and a community sample (Levallius, Monell, Birgegård, Clinton, also satisfactory for all the subscales, ranging from a 5 0.70 & ForsénMantilla, 2020). for asceticism to a 5 0.87 for body dissatisfaction (see Thus, given this background and the above-mentioned Cronbach’s alphas of all subscales in supplementary Table S1). gaps in the literature, the objectives of this study were Symptom Checklist-90 Items-Revised (SCL-90-R) (Dero- fourfold: 1) to assess the prevalence of current FA and/or gatis, 1990); Spanish validation (Derogatis, 1994). The SCL- lifetime problematic SU among a large sample of patients 90-R assesses nine scales on general psychopathology: so- with different ED diagnostic types (based on DSM-5 matization, obsessive-compulsive, interpersonal sensitivity, criteria); 2) to examine gender-related differences among depression, anxiety, hostility, phobic anxiety, paranoid patients with ED with or without both current FA and ideation, and psychoticism. In addition, it assesses three lifetime problematic SU; 3) to investigate whether clinical, global indices of psychological distress: Global Severity In- psychopathological, impulsive, and personality characteris- dex (GSI), Positive Symptom Total (PST), and Positive tics are associated with FA and/or SU phenotypes; and 4) to Symptom Distress Index (PSDI). The internal consistency identify relevant factors related to the presence of current FA was excellent in our sample for the GSI score (a 5 0.98), and and/or lifetime problematic SU in patients with EDs. also satisfactory for all the subscales, ranging from good (a 5 0.75 for paranoid ideation) to excellent (a 5 0.90 for depression) (see Cronbach’s alphas of all subscales in sup- MATERIAL AND METHODS plementary Table S1). Temperament and Character Inventory–Revised (TCI-R) Participants (Cloninger, 1999); Spanish validation (Gutiérrez-Zotes et al., 2004). The TCI-R contains 240 items for measuring per- The whole sample consisted of 527 patients with specific sonality traits structured into seven personality dimensions: EDs (176 BN, 115 BED, and 236 OSFED) consecutively four temperamental (novelty seeking, harm avoidance, admitted to the EDs Unit of the Bellvitge University Hos- reward dependence, and persistence) and three characters pital (Barcelona, Spain) between May 2016 and October (self-directedness, cooperativeness, and self-transcendence) 2020. The diagnoses were made according to DSM-5 criteria dimensions. Cronbach’s alpha for the current sample was (American Psychiatric Association, 2013) by experienced good (a 5 0.79 for novelty-seeking) to excellent (a 5 0.90 clinical psychologists and psychiatrists specialized in EDs. for harm avoidance and persistence) (see Cronbach’s alphas Figure S1 (supplementary material) includes a flowchart of all subscales in supplementary Table S1). summarizing the sampling process used in the inclusion of Yale Food addiction Scale 2.0 (YFAS 2.0) (Gearhardt, participants. Corbin, & Brownell, 2016); Spanish validation (Granero et al., 2018). This is a self-reported scale to assess FA based Assessment on the 11 substance dependence-related symptoms adapted Sociodemographic and clinical data were obtained by means to the context of food consumption. The YFAS 2.0 consists of a face-to-face semi-structured interview based on the of 35 items and produces two measurements: (1) a contin- SCID-5 (First, Williams, Karg, & Spitzer, 2015). During this uous symptom count score that reflects the number of ful- clinical interview, data on lifetime SU were also retrieved filled diagnostic criteria (ranging from 0 to 11), and (2) a from specific questions based on Module E of the SCID-5 binary measurement (present versus absent) based on the (First et al., 2015). Having problematic SU was defined as number of symptoms (at least 2) and the self-reported patterns of substance use (alcohol and/or illicit drugs) in an clinically impairment or distress. Additionally, it gives amount, frequency, or circumstance that is potentially severity cut-offs: mild (2–3 symptoms), moderate (4–5 physically, mentally, and socially harmful to the individual, symptoms), and severe (6–11 symptoms). The internal regardless of whether the individual met full diagnostic consistency of our sample was excellent (a 5 0.965). criteria for SUD. Due to the high prevalence tobacco use was The UPPS-P Impulsivity Scale (UPPS-P) (Whiteside, excluded from the category of “problematic SU” used in this Lynam, Miller, & Reynolds, 2005); Spanish validation (Ver- study and, therefore, from all subsequent analyses, because dejo-García, Lozano, Moya, Alcázar, & Pérez-García, 2010). its use was typically continuous, and because there is a high This questionnaire includes 59 items developed for assessing prevalence of smoking in the population of EDs. five impulsivity factors: lack of perseverance, lack of pre- Additionally, a comprehensive battery composed of meditation, sensation seeking, negative urgency, and positive commonly applied questionnaires in the field of EDs was urgency. The internal consistency in our sample was satis- administered: factory for all the subscales, ranging from a 5 0.83 for lack of Eating Disorders Inventory-2 (EDI-2) (Garner, 1991); perseverance to a 5 0.91 for positive urgency (see Cronbach’s Spanish validation (Garner, 1998). The EDI-2 assesses alphas of all subscales in supplementary Table S1). Journal of Behavioral Addictions 11 (2022) 1, 102–115 105 Statistical analyses reported the oldest age, the latest onset of the ED, and the longest duration of the disorder. As expected, the highest Statistical analysis was carried out with Stata16 for Windows BMI mean was registered for BED, followed by BN and (Stata-Corp & StataCorp, 2019). Comparison between the OSFED. groups was done with chi-square tests (c2) for categorical variables and with analysis of variance (ANOVA) for quantitative variables. Effect size for the mean differences in Prevalence of current FA and lifetime problematic SU the psychometric tools was estimated through the stan- A total of 414 patients screened positive for FA (FAþ) dardized Cohen’s-d coefficient, considering small effect as (prevalence 5 78.6%). Lifetime problematic SU (SUþ) was |d| > 0.20, mild-moderate as |d| > 0.50, and large-high as identified in n 5 144 patients (prevalence 5 27.3%) [alcohol |d| > 0.80. Correction in Type-I error due to the multiple use was reported by n 5 57 (10.8%) and other illicit drugs by statistical significance tests was based on Finner’s method, a n 5 124 (23.5%)]. The condition of FAþ and SUþ was familywise error procedure which has proved higher power present in n 5 125 participants (23.7%). than classical Bonferroni method (Finner & Roters, 2001). Table 2 includes the prevalence of FA and lifetime Logistic regression was used to obtain a predictive model problematic SU within the ED subtypes and the comparison for the presence of current FA with lifetime SU (excluding between the groups. FAþ screening score achieved the tobacco), considering as potential predictors sociodemo- highest prevalence among BN (93.2%) followed by BED graphic variables (gender, age, and marital status), age of (82.6%), and OSFED (65.7%). No statistical differences were onset of the ED, ED severity (EDI-2 total), general psy- obtained for SUþ (alcohol or illicit drugs) comparing the chopathology (SCL-90R GSI), impulsivity levels (UPPS-P ED diagnoses. Further, patients with FAþ were more likely scales), and personality traits (TCI-R scales). In addition, to present with SUþ (30.2%) than patients without FA two predictive models were conducted: the first one within (16.8%) (c2 5 8.00, P 5 0.005). The FA and SU the SUþ subsample (considering the presence of FAþ as phenotype was more likelihood within OSFED patients criterion) and the next one within the FAþ subsample (28.0%), followed by BED (16.5%) and BN (5.1%). The FAþ (considering the presence of SUþ as criterion). A stepwise and SUþ phenotype obtained the highest likelihood for BN procedure was used to automatically select the predictors (31.8%), followed by BED (21.7%) and OSFED (18.6%). No achieving statistical significance (P < 0.05). Goodness-of-fit significant differences between both BED and OSFED were was tested with Hosmer-Lemeshow test (adequate fitting observed. The OSFED subtype displayed the lowest pro- was considered for P > 0.05). Global predictive capacity was portion for FAþ and SUþ, but the highest proportion of measured with Nagelkerke’s pseudo-R2 coefficient and FA and SUþ. global discriminative accuracy with the area under the ROC curve (AUC). Association between gender with current FA and lifetime problematic SU Ethics Within the subsample of women, FA and SU was The study procedures were carried out in accordance with observed in n 5 84 (17.2%), FAþ and SU in n 5 273 the Declaration of Helsinki. The Clinical Research Ethics (55.8%), FA and SUþ in n 5 18 (3.7%), and FAþ and Committee of Bellvitge University Hospital approved the SUþ in n 5 114 (23.3%). Within the male subsample, FA study. Written informed consent was obtained from all and SU was observed in n 5 10 (26.3%), FAþ and SU in participants before taking part in the study. n 5 16 (42.1%), FA and SUþ in n 5 1 (2.6%), and FAþ and SUþ in n 5 11 (28.9%). No differences between gender were obtained for the distribution of FA and lifetime SU RESULTS (P 5 0.328). Sample description Comparison of clinical profiles Most participants in the study were female (n 5 489, 92.8%), Table 3 displays the distribution of the ED symptomatology single (n 5 368, 69.8%), employed (n 5 291, 55.3%), and (EDI-2 scales), the psychopathological state (SCL-90R had completed primary or secondary education (n 5 445, scales), the impulsivity levels (UPPS-P scales), and the per- 84.4%). Mean age was 32.1 years old (SD 5 12.6), mean age sonality traits (TCI-R scales), as well as the comparison of onset of ED was 20.4 years old (SD 5 9.9), and mean between the groups defined by the presence-absence of FA duration of the ED was 12.0 years (SD 5 10.1). Table 1 and lifetime SU. Overall, the condition characterized by contains the descriptive for the ED subtypes and the com- FA and SU was associated with the lowest scores on ED parison between the groups. Statistical differences were severity and psychopathology, impulsivity, harm avoidance, observed for marital status and employment status: the and the highest self-directedness. The condition FA and highest proportion of single participants was observed for SU also presented lower novelty seeking compared with OSFED followed by BN and BED, whereas the highest FAþ and SUþ. The comorbid condition of FAþ and SUþ proportion of employed individuals was reported by BED, obtained higher mean scores on the EDI-2 impulse regula- followed by BN and OSFED. Patients in the BED group also tion subscale, on the UPPS-P scales (except for lack of 106 Table 1. Descriptive of the sample Total BN BED OSFED BN versus BN versus BED versus N 5 527 N 5 176 N 5 115 N 5 236 BED OSFED OSFED n % n % n % n % P P C-V P C-V P C-V Gender Female 489 92.8% 167 94.9% 102 88.7% 220 93.2% 0.129 0.051 0.094 0.484 0.035 0.148 0.077 Male 38 7.2% 9 5.1% 13 11.3% 16 6.8% Marital status Single 368 69.8% 122 69.3% 56 48.7% 190 80.5% 0.001* 0.001* 0.222y 0.032* 0.130y 0.001* 0.337y Married - partner 106 20.1% 34 19.3% 44 38.3% 28 11.9% Divorced - separated 53 10.1% 20 11.4% 15 13.0% 18 7.6% Education Primary 221 41.9% 67 38.1% 52 45.2% 102 43.2% 0.248 0.281 0.093 0.070 0.095 0.844 0.031 Secondary 224 42.5% 73 41.5% 47 40.9% 104 44.1% University 82 15.6% 36 20.5% 16 13.9% 30 12.7% Journal of Behavioral Addictions 11 (2022) 1, 102–115 Employment Unempl. 104 19.7% 37 21.0% 26 22.6% 41 17.4% 0.001* 0.046* 0.146y 0.001* 0.195y 0.001* 0.301y Student 132 25.0% 34 19.3% 10 8.7% 88 37.3% Employed 291 55.3% 105 59.7% 79 68.7% 107 45.3% Mean SD Mean SD Mean SD Mean SD P P |d| P |d| P |d| Age (years-old) 32.06 12.60 31.73 11.31 39.35 12.70 28.75 12.05 0.001* 0.001* 0.63y 0.013* 0.25 0.001* 0.86y Onset ED (years-old) 20.35 9.90 19.53 9.24 24.41 11.87 18.97 8.79 0.001* 0.001* 0.51y 0.566 0.06 0.001* 0.52y Duration ED (years) 11.97 10.09 12.58 9.38 15.39 11.23 9.86 9.55 0.001* 0.018* 0.27 0.006* 0.29 0.001* 0.53y BMI (kg m2) 28.06 9.49 26.46 6.16 40.28 9.42 23.31 5.71 0.001* 0.001* 1.74y 0.001* 0.53y 0.001* 2.18y Note. BN: bulimia nervosa. BED: binge eating disorder. ED: eating disorder. OSFED: other specified feeding or eating disorder. SD: standard deviation. C-V: Cramer-V coefficient. *Bold: significant parameter. y Bold: effect size into the range mild-moderate to large-high. Journal of Behavioral Addictions 11 (2022) 1, 102–115 107 premeditation and positive urgency), and on the TCI-R 0.189y 0.125y 0.130y 0.123y Note. FA: food addiction. FA-: food addiction negative screening score. FAþ: food addiction positive screening score. SU: lifetime substance use. SU: lifetime substances use absent. SUþ: 0.176 0.026 0.010 0.033 0.037 C-V novelty seeking, as well as lower scores on the TCI-R self- BED versus OSFED directedness and cooperativeness scales. In addition, Table S2 (supplementary material) shows the 0.006* 0.019* 0.015* 0.021* results of the comparison between the different types of 0.001 0.624 0.848 0.531 0.493 P problematic SU (i.e., alcohol and other illicit drugs) among patients with FAþ. Overall, patients with FAþ and lifetime consumption of both alcohol and other illicit drugs pre- y 0.331y 0.293y 0.142y 0.113y 0.152y 0.325 0.093 0.076 0.073 C-V sented higher ED symptomatology (EDI-2 total) and general lifetime substances use present. BN: bulimia nervosa. BED: binge eating disorder. OSFED: other specified feeding or eating disorder. C-V: Cramer-V coefficient. BN versus psychopathology (SCL-90R GSI), as well as higher scores on OSFED positive urgency and novelty-seeking, and lower scores on self-directedness than those who had used a single type of 0.001* 0.001* 0.001* 0.004* 0.022* 0.002* 0.058 0.123 0.138 substance (i.e., alcohol or illicit drugs). P Predictive model for the presence of current FA and lifetime problematic SU y 0.206y 0.189y 0.110y 0.165 0.097 0.083 0.095 0.005 0.035 C-V BN versus The upper part of Table 4 shows the results of the stepwise BED logistic regression obtained for the whole sample (n 5 527) considering the presence of FAþ and SUþ. The likelihood Table 2. Prevalence of FA and/or lifetime problematic SU in EDs 0.006* 0.001* 0.041* 0.933 0.550 0.005 0.052 0.158 0.073 of FAþ and SUþ was increased for unmarried or older P patients or those with higher scores in novelty-seeking or lower scores in self-directedness. This model achieved adequate fitting (P 5 0.522 in the Hosmer-Lemeshow test), 0.001* 0.001* 0.069 0.204 0.145 P P but the global predictive capacity and the discriminative capacity were low (Nagelkerke’s-R2 5 0.112 and AUC 5 0.684). 9.3% 6.4% 65.7% 25.0% 22.0% 28.0% 47.0% 18.6% The middle part of Table 4 shows the result of an % % n 5 236 additional stepwise logistic regression obtained in the sub- OSFED sample of patients who reported SUþ (n 5 144) considering as criterion the presence of FAþ. Results showed that FAþ 66 15 44 155 59 22 52 111 n n increased likelihood for patients with higher levels in ED severity, negative urgency and cooperativeness, and lower levels in reward dependence and self-directedness. This 8.7% 0.9% 82.6% 22.6% 19.1% 16.5% 60.9% 21.7% model also achieved adequate goodness-of-fit (P 5 0.957 in % % n 5 115 BED the Hosmer-Lemeshow test), and good global predictive capacity and discriminative capacity (Nagelkerke’s-R2 5 0.334 and AUC 5 0.843). 95 26 10 22 19 70 25 1 n n The lower part of Table 4 shows a final stepwise logistic model, obtained within the subsample that reported the Italic font: pairwise comparison for each level of the variable. Bold: effect size into the range mild-moderate to large-high. 5.1% 1.7% 93.2% 33.5% 14.2% 28.4% 61.4% 31.8% presence of FAþ (n 5 414, considering as the criterion the % % n 5 176 presence of SUþ). The predictors increasing the likelihood BN for the consumption of substances were marital status (being unmarried), higher levels of novelty-seeking, and lower 9 3 25 50 56 164 59 108 n n levels of cooperativeness. Adequate fitting was achieved for this model (P 5 0.502 in Hosmer-Lemeshow test), but low predictive capacity (Nagelkerke’s-R2 5 0.079) and low 3.3% 78.6% 27.3% 10.8% 23.5% 54.8% 23.7% 17.8 % % discriminative capacity (AUC 5 0.640). N 5 527 Total *Bold: significant parameter. 57 94 19 414 144 124 289 125 DISCUSSION n n The current study aimed to investigate the prevalence of Alcohol or drugs current FA and/or lifetime problematic SU (except for to- FA and SU FAþ and SU FA and SUþ FAþ and SUþ FA and/or SU bacco) among patients with BN, BED, and OSFED. Likewise, illicit drugs the clinical, psychopathological, and personality profiles Alcohol exhibited by the different addictive behaviors based on FAþ SU phenotypes were also examined. The main findings derived y 108 Journal of Behavioral Addictions 11 (2022) 1, 102–115 Table 3. Comparison of the clinical profiles in the study FA and SU FA and SU FAþ or SUþ FA and SU FAþ or SUþ FAþ and SUþ versus versus versus n 5 94 n 5 308 n 5 125 FAþ or SUþ FAþ and SUþ FAþ and SUþ Mean SD Mean SD Mean SD P |d| P |d| P |d| ED severity (EDI-2) Drive for thinness 11.19 6.55 15.27 4.80 15.60 4.99 0.001* 0.71y 0.001* 0.76y 0.545 0.07 Body dissatisfaction 14.79 8.65 18.60 6.98 18.38 7.19 0.001* 0.49y 0.001* 0.45 0.770 0.03 Interoceptive awareness 8.16 6.44 13.26 6.51 14.06 7.17 0.001* 0.79y 0.001* 0.87y 0.258 0.12 Bulimia 3.65 4.18 8.96 5.15 9.29 5.23 0.001* 1.13y 0.001* 1.19y 0.543 0.06 Interpersonal distrust 4.91 4.56 6.21 5.13 6.30 4.83 0.027* 0.27 0.041 0.30 0.855 0.02 Ineffectiveness 8.33 6.58 12.53 7.28 12.98 7.60 0.001* 0.61y 0.001* 0.65y 0.557 0.06 Maturity fears 7.96 4.66 8.60 5.75 9.24 6.39 0.339 0.12 0.102 0.23 0.296 0.10 Perfectionism 4.76 3.90 6.36 4.49 6.39 4.52 0.002* 0.38 0.007 0.39 0.952 0.01 Impulse regulation 4.61 4.98 6.91 5.63 9.06 7.29 0.001* 0.43 0.001* 0.71y 0.001* 0.33 Ascetism 5.20 3.74 7.90 3.77 8.32 4.41 0.001* 0.72y 0.001* 0.76y 0.317 0.10 Social insecurity 6.47 5.11 8.39 5.13 8.92 5.26 0.002 0.38 0.001* 0.47 0.333 0.10 Total score 80.0 39.0 113.0 37.9 118.6 42.6 0.001* 0.86y 0.001* 0.94y 0.183 0.14 Psychop. (SCL-90R) Somatization 1.31 0.88 1.95 0.86 2.09 0.92 0.001* 0.73y 0.001* 0.86y 0.130 0.16 Obsessive-compulsive 1.45 0.94 2.12 0.80 2.11 0.81 0.001* 0.77y 0.001* 0.76y 0.951 0.01 Interpersonal sensitivity 1.56 1.01 2.15 0.85 2.18 0.95 0.001* 0.63y 0.001* 0.64y 0.686 0.04 Depressive 1.67 0.94 2.43 0.83 2.51 0.88 0.001* 0.86y 0.001* 0.93y 0.387 0.09 Anxiety 1.14 0.82 1.80 0.83 1.96 0.95 0.001* 0.79y 0.001* 0.92y 0.070 0.19 Hostility 0.79 0.84 1.38 0.91 1.56 1.04 0.001* 0.68y 0.001* 0.82y 0.071 0.18 Phobic anxiety 0.66 0.77 1.14 0.90 1.30 1.04 0.001* 0.58y 0.001* 0.70y 0.101 0.16 Paranoia 1.15 0.78 1.53 0.84 1.59 0.88 0.001* 0.46 0.001* 0.53y 0.478 0.07 Psychotic 0.85 0.65 1.46 0.71 1.51 0.80 0.001* 0.89y 0.001* 0.90y 0.484 0.07 GSI 1.26 0.72 1.89 0.66 1.98 0.77 0.001* 0.90y 0.001* 0.97y 0.193 0.13 PST 51.6 20.32 67.3 14.65 68.2 16.26 0.001* 0.89y 0.001* 0.90y 0.616 0.06 PSDI 2.06 0.58 2.46 0.52 2.53 0.58 0.001* 0.72y 0.001* 0.80y 0.213 0.13 Impulsivity (UPPS-P) Lack premeditation 22.7 4.93 24.0 5.93 25.2 5.89 0.060 0.23 0.002* 0.45 0.058 0.20 Lack perseverance 21.7 5.43 23.5 5.14 24.8 5.85 0.004* 0.34 0.001* 0.55y 0.025* 0.23 Sensation seeking 26.1 8.22 25.6 8.05 27.7 7.86 0.568 0.07 0.165 0.19 0.016* 0.26 Positive urgency 26.8 8.33 28.8 8.57 31.1 9.73 0.047* 0.25 0.001* 0.48y 0.014* 0.25 Negative urgency 30.1 7.23 34.5 6.35 35.7 5.82 0.001* 0.64y 0.001* 0.85y 0.066 0.20 Personality (TCI-R) Novelty seeking 97.2 14.39 100.1 15.41 105.9 17.51 0.121 0.19 0.001* 0.54y 0.001* 0.35 Harm avoidance 110.8 20.81 119.8 19.93 121.7 20.05 0.001* 0.44 0.001* 0.53y 0.372 0.10 Reward dependence 102.8 16.81 100.0 16.09 98.9 17.01 0.154 0.17 0.085 0.23 0.524 0.07 Persistence 109.0 19.81 108.6 21.39 105.6 21.66 0.884 0.02 0.252 0.16 0.189 0.14 Self-directedness 125.6 19.20 113.6 20.31 107.0 19.75 0.001* 0.61y 0.001* 0.96y 0.002* 0.33 Cooperativeness 133.0 17.47 134.3 16.63 129.9 16.48 0.500 0.08 0.170 0.18 0.012* 0.27 Self-transcendence 63.7 15.64 66.1 15.48 66.2 15.22 0.204 0.15 0.236 0.16 0.910 0.01 Note. FA: food addiction negative screening score. FAþ: food addiction positive screening score. SU: lifetime substances use disorders absent. SUþ: lifetime substances use disorders present. *Bold: significant parameter. y Bold: effect size into the range mild-moderate (|d| > 0.50) to large-high (|d| > 0.80). from this study suggest that patients with at least one of the FA and SUþ groups (because of the high rates of FA addictive-related behavior exhibited a more severe clinical in the ED population). state compared to those who reported none. Interestingly, First, our results confirmed the high prevalence of FA in the patients with the SUþ and FAþ phenotype did not show a sample of patients with EDs (78.6%) (Fauconnier et al., greater clinical severity than those with a single addiction, 2020; Jimenez-Murcia et al., 2019; Khalil et al., 2020; although they did show greater impulsivity and less self- Romero et al., 2019). In agreement with previous literature directedness compared to the latter. Therefore, our findings (Gearhardt et al., 2014; Granero et al., 2018; Jimenez-Murcia suggest that, although presenting an addictive behavior is et al., 2019), our results confirmed the higher prevalence of associated with greater ED symptomatology and general FAþ in the BN group, followed by the BED and OSFED psychopathology, an additive effect of both addictive-like groups. In this regard, a prior study suggested that FA may behaviors could not be evidenced. However, these findings reflect the increased psychopathology of BN patients should be interpreted cautiously due to the low sample size (Gearhardt et al., 2014). Regarding lifetime problematic SU, Journal of Behavioral Addictions 11 (2022) 1, 102–115 109 Table 4. Predictive model for the presence of FAþ and SUþ: stepwise logistic regression Complete sample (n 5 527) B SE P OR 95%CI OR Marital status (married) 0.898 0.323 0.005 0.407 0.216 0.767 Age (years-old) 0.018 0.009 0.048 1.018 1.000 1.037 TCI-R novelty seeking 0.022 0.007 0.001 1.023 1.009 1.036 TCI-R self-directedness 0.022 0.006 <0.001 0.978 0.967 0.989 Fitting indexes Hosmer and Lemeshow test: c2 5 7.138; df 5 8; P 5 0.522 Nagelkerke's pseudo-R2: 0.112 Area under the ROC curve: 0.684 (95%CI: 0.631–0.738) Subsample: SUþ (n 5 144) B SE P OR 95%CI OR ED severity 0.025 0.009 0.007 1.025 1.007 1.044 UPPS-P negative urgency 0.118 0.050 0.018 1.125 1.020 1.241 TCI-R reward dependence 0.042 0.020 0.036 0.959 0.923 0.997 TCI-R cooperativeness 0.052 0.023 0.023 1.053 1.007 1.101 TCI-R self-directedness 0.041 0.020 0.038 0.960 0.923 0.998 Fitting indexes Hosmer and Lemeshow test: c2 5 2.595; df 5 8; P 5 0.957 Nagelkerke's pseudo-R2: 0.334 Area under the ROC curve: 0.843 (95%CI: 0.751–0.934) Subsample: FAþ (n 5 414) B SE P OR 95%CI OR Marital status (married) 0.764 0.300 0.011 0.466 0.259 0.840 TCI-R novelty seeking 0.022 0.007 0.002 1.022 1.008 1.036 TCI-R cooperativeness 0.014 0.007 0.038 0.986 0.974 0.999 Fitting indexes Hosmer and Lemeshow test: c2 5 7.326; df 5 8; P 5 0.502 Nagelkerke's pseudo-R2: 0.079 Area under the ROC curve: 0.640 (95%CI: 0.582–0.697) Note. SE: standard error. OR: odds ratio. SUþ: lifetime substances use disorders present. Predictors entered in the stepwise procedure: sociodemographic (gender, marital status), age, age of onset of the ED, ED severity (EDI-2 total), psychopathology distress (SCL-90R GSI), impulsivity level (UPPS-P scales), and personality traits (TCI-R scales). our results were not able to find statistical differences across related to gender-related differences should be cautiously ED diagnostic types for neither alcohol nor illicit drugs. This considered because there is an underrepresentation of finding is inconsistent with previous studies describing males in current literature, due to the low proportion of higher frequencies of SUþ in patients with BN compared to them in routine clinical practice (Bahji et al., 2019). other ED diagnoses (Fouladi et al., 2015; Nøkleby, 2012). Therefore, future studies focusing on gender-related dif- However, it is in line with other studies suggesting that in- ferences should consider the inclusion of a larger sample of dividuals presenting binge eating and/or purging behaviors males to get a better understanding. are more susceptible to engage in SUþ, independent of ED Concerning clinical and psychopathological characteris- diagnostic type (Bahji et al., 2019; Lozano-Madrid et al., tics, our findings support prior evidence suggesting that the 2020). Further, our results showed a prevalence of present- presence of addictive-related behaviors (i.e., FAþ or SUþ or ing with both addictive-like behaviors (i.e., FAþ and SUþ) both) is associated with more severe ED symptomatology, in almost 24% of the ED participants. In addition, our general psychopathology, and emotional dysregulation in findings suggest that patients with FAþ may be more likely the general (Hardy et al., 2018; Wiedemann, Carr, Ivezaj, & to present with SUþ than those without FA. Barnes, 2021) and clinical populations (Romero et al., 2019; On the other hand, patients with ED and FAþ and/or Tinghino et al., 2021). Specifically for FA, several studies SUþ did not show significant gender-related differences. have identified the presence of FA as a severity marker in Similarly, previous studies screening FA were not able to ED, in patients with binge eating and/or purging behaviors find gender differences among participants with ED (Brewerton, 2017; Carter, Van Wijk, & Rowsell, 2019; (Granero et al., 2014; Wolz, Granero, & Fernández-Ara- Jimenez-Murcia et al., 2019; Romero et al., 2019), but also in nda, 2017). In the same vein, a study in general pop- those with AN-R (Fauconnier et al., 2020; Tran et al., 2020). ulations failed to find sex differences in the prevalence of On the other hand, our results were not able to show that FAþ (Wu, Zimmer, Munn-Chernoff, & Baker, 2020). patients with both addictive-related behaviors (i.e., FAþ and Likewise, when exploring gender differences in patients SUþ) presented with a more severe clinical condition than with ED and SUþ, our findings are in line with other patients with a single comorbid addiction. Although we studies that failed to find gender differences (Lozano- expected to find that the FAþ and SUþ phenotype would Madrid et al., 2020). However, Becker and Grilo (2015) also be the most severe, we were not able to determine the observed a higher proportion of men with BED and co- additive effect of FA and lifetime SU. However, these find- morbid SUD compared to those without SUD. Findings ings reinforce the need to explore and treat the comorbidity 110 Journal of Behavioral Addictions 11 (2022) 1, 102–115 of addictive behaviors in these patients, regardless of seeking and lower cooperativeness (i.e., more self-centered whether they have one or more addictive behaviors. and less willing to have social relationships with other In terms of personality and impulsivity, patients with at people). These personality traits have previously been least one comorbid addictive condition (i.e., FAþ, SUþ, or identified as possible risk factors for the development of both) were more likely to have a dysfunctional personality addictive behaviors (Del Pino-Gutiérrez et al., 2017; Krug profile characterized by higher harm avoidance and lower et al., 2009; Steingrimsson, Carlsen, Lundström, Lundström, self-directedness, as well as greater impulsivity than patients & Nilsson, 2020) as well as with the presence of more than without addictive behaviors. These impairments in impulse one addictive behavior (Dash et al., 2019). control in patients with EDs and comorbid addictive be- haviors had already been described by previous studies Limitations and strengths (Dawe & Loxton, 2004; Minhas et al., 2021). Specifically, negative urgency, defined as the tendency to rashly act under The present study should be considered within the context strong emotions (Cyders & Smith, 2008), was the impulsive of some limitations. First, our results should be interpreted trait most strongly related to the presence or absence of at with caution, keeping in mind that they refer to patients least one addiction-related behavior. This result is consistent with a history of problematic SU and not a full diagnosis of with a previous study suggesting that negative urgency may SUD. Second, the FA measure refers to current FA but the be a predictor of FAþ in patients with EDs (Wolz et al., problematic SU measure refers to lifetime SU. Further 2017) and that it may play a key role as a risk factor to studies should disentangle current vs. lifetime use for treatment success in patients with SUD (Halcomb, Argyriou, problematic SU, as well as assess lifetime FA. Third, our & Cyders, 2019). participants were recruited from a hospital setting and, in Furthermore, our findings identified that patients with addition, patients with AN were not included because of the both addictive behaviors (i.e., FAþ and SUþ) showed a existing controversy as to whether they actually present with lower self-directedness and a greater exacerbation of the FA or their responses to the YFAS 2.0 items reflect their impulsive traits (namely, higher lack of perseverance, subjective fear of losing control with food. Therefore, our sensation seeking, and positive urgency) even than those sample is not representative of the general ED population. with only FAþ or SUþ. Hence, this relevant finding suggests Four, in this study we compared the characteristics of pa- that these impulsivity traits and the lack of self-determina- tients presenting at least one addictive behavior (i.e., FAþ or tion could act as a differentiating risk factor for developing SUþ) versus those presenting neither or both. It would be more than one co-morbid addictive-related behavior in pa- interesting for further research to analyze whether there are tients with EDs. differences between the FAþ and SU- and the FA- and SUþ When we specifically considered the types of problematic groups, which could not be analyzed in this study due to the SU (i.e., alcohol or other illicit drugs) in patients with FAþ, low sample size of the FA- and SUþ groups. Also, further it was observed that patients with FA who had consumed studies with large sample sizes might include tobacco use in both alcohol and other illicit drugs in their lifespan exhibited the analysis of substance addiction. Finally, the cross- more ED-related and psychopathological severity, higher sectional design of our study does not allow us to draw in- positive urgency and novelty-seeking traits, and lower self- ferences regarding causality. directedness than those who had used a single type of sub- Notwithstanding these limitations, the study also has stance. This finding is consistent with previous research several strengths that should be noted. To the best of our studies suggesting an additive effect of multiple addictive knowledge, this is the first study exploring FA and lifetime patterns associated with a poor clinical state and increased problematic SU, both jointly and separately, in a large impulsivity (Jimenez-Murcia et al., 2021; Martínez-Loredo, sample of adult patients, including women as well as men Macipe, ErrastiPérez, & Al-Halabí, 2021). with EDs. Finally, some sociodemographic variables such as being older and unmarried were associated with a higher likeli- hood of presenting with the FAþ and SUþ phenotype. CONCLUSIONS These findings are in line with previous studies carried out in other populations that have addictions (Del Pino- In conclusion, our findings indicate that the presence of one Gutiérrez et al., 2017). This may suggest a protective effect of or more addictive behaviors in patients with BN, BED, and marriage in substance use and abuse (Choe, Yoo, JeKarl, & OSFED was associated with increased ED severity. The Kim, 2018; Jimenez-Murcia et al., 2017; Kendler, Lönn, difference between engaging in one or more lifetime Salvatore, Sundquist, & Sundquist, 2016). Furthermore, as addictive-like behaviors (i.e. FA and/or alcohol and/or or expected, high scores on novelty seeking and low scores on illicit drug use or both) was only reflected in personality self-directedness were strongly associated with the FAþ and traits, with greater impulsivity and less self-directedness in SUþ phenotype. Specifically, for patients with SUþ, the the FAþ and SUþ phenotype. Although further longitudinal presence of comorbid FAþ was associated with increased studies assessing whether these addictive phenotypes may ED severity and negative urgency, and less reward depen- influence treatment outcome are needed, these findings may dence and self-directedness. Likewise, for patients with FAþ, suggest the need for treatments to reduce impulsivity and the presence of SUþ was associated with higher novelty increase self-directedness, especially in patients with this Journal of Behavioral Addictions 11 (2022) 1, 102–115 111 greater addictive phenotype, as a step toward improving Blumenthal, D. M., & Gold, M. S. (2010). Neurobiology of food their treatment outcome. addiction. Current Opinion in Clinical Nutrition and Metabolic Care, 13(4), pp. 359–365. https://doi.org/10.1097/MCO. 0b013e32833ad4d4. Funding sources: We thank CERCA Programme/Generalitat Bogusz, K., Kopera, M., Jakubczyk, A., Trucco, E. M., Kucharska, de Catalunya for institutional support. This research was K., Walenda, A., & Wojnar, M. (2021). Prevalence of alcohol supported by Instituto de Salud Carlos III (ISCIII; grant use disorder among individuals who binge eat: A systematic number: PI17/01167 and PI20/00132), by PERIS (Generalitat review and meta-analysis. Addiction, 116(1), 18–31. https://doi. de Catalunya, SLT006/17/00246), and co-funded by FEDER org/10.1111/add.15155. funds/European Regional Development Fund (ERDF), a way Bonder, R., Davis, C., Kuk, J. L., & Loxton, N. J. (2018). Compulsive to build Europe. CIBERobn is an initiative of ISCIII. Funding “grazing” and addictive tendencies towards food. European Eating sources have not any role in the design, collection, analysis, Disorders Review, 26(6), 569–573. https://doi.org/10.1002/erv.2642. and interpretation of data; in the writing of the report; and in Brewerton, T. (2017). Food addiction as a proxy for eating disorder the decision to submit the article for publication. and obesity severity, trauma history, PTSD symptoms, and comorbidity. Eating and Weight Disorders: EWD, 22(2), 241– Authors’ contribution: R.M-O, Z.A, R.R.V, A.N.G, and F.F-A 247. https://doi.org/10.1007/S40519-016-0355-8. designed the study. R.M-O conducted literature research Canan, F., Karaca, S., Sogucak, S., Gecici, O., & Kuloglu, M. (2017). and provided summaries of previous studies. R.G. conducted Eating disorders and food addiction in men with heroin use the statistical analysis. R.M-O, Z.A, and R.G wrote the first disorder: A controlled study. Eating and Weight Disorders, draft of the manuscript. All authors revised, contributed to, 22(2), 249–257. https://doi.org/10.1007/s40519-017-0378-9. and approved the final manuscript. Carter, J., Van Wijk, M., & Rowsell, M. (2019). Symptoms of “food addiction” in binge eating disorder using the Yale Food Conflicts of interest: Fernando Fernández-Aranda reports Addiction Scale version 2.0. Appetite, 133, 362–369. https://doi. consultation fees from Novo Nordisk and editor-in-Chief org/10.1016/J.APPET.2018.11.032. honorarium from Wiley. No other conflicts of interest are Choe, S. A., Yoo, S., JeKarl, J., & Kim, K. K. (2018). Recent trend declared by the authors. and associated factors of harmful alcohol use based on age and gender in Korea. Journal of Korean Medical Science, 33(4). https://doi.org/10.3346/jkms.2018.33.e23. Cloninger, C. R. (1999). The temperament and character inventory– REFERENCES revised. Center for Psychobiology of Personality, Washington University. Albayrak, Ö., Föcker, M., Kliewer, J., Esber, S., Peters, T., de Zwaan, Cyders, M. A., & Smith, G. T. (2008). Emotion-based dispositions M., & Hebebrand, J. (2017). Eating-related psychopathology to rash action: Positive and negative urgency. Psychological and food addiction in adolescent psychiatric inpatients. Euro- Bulletin, 134(6), 807–828. https://doi.org/10.1037/a0013341. pean Eating Disorders Review, 25(3), 214–220. https://doi.org/ Dash, G. F., Slutske, W. S., Martin, N. G., Statham, D. J., Agrawal, 10.1002/erv.2509. A., & Lynskey, M. T. (2019). Big five personality traits and American Psychiatric Association (2013). APA DSM-5. Diagnostic alcohol, nicotine, cannabis, and gambling disorder comorbidity. and statistical manual of mental disorders. Psychology of Addictive Behaviors, 33(4), 420. https://doi.org/10. Anzengruber, D., Klump, K. L., Thornton, L., Brandt, H., Crawford, 1037/adb0000468. S., Fichter, M. M., . . . Bulik, C. M. (2006). Smoking in eating Dawe, S., & Loxton, N. J. (2004). The role of impulsivity in the disorders. Eating Behaviors, 7(4), 291–299. https://doi.org/10. development of substance use and eating disorders. Neurosci- 1016/j.eatbeh.2006.06.005. ence and Biobehavioral Reviews, 28(3), 343–351. Pergamon. Bahji, A., Mazhar, M. N., Hudson, C. C., Nadkarni, P., MacNeil, https://doi.org/10.1016/j.neubiorev.2004.03.007. B. A., & Hawken, E. (2019). Prevalence of substance use Del Pino-Gutiérrez, A., Jimenez-Murcia, S., Fernández-Aranda, F., disorder comorbidity among individuals with eating disor- Agüera, Z., Granero, R., Hakansson, A., . . . Menchón, J. M. ders: A systematic review and meta-analysis. Psychiatry (2017). The relevance of personality traits in impulsivity-related Research, 273, 58–66. https://doi.org/10.1016/j.psychres. disorders: From substance use disorders and gambling disorder 2019.01.007. to bulimia nervosa. Journal of Behavioral Addictions, 6(3), 396– Baker, J. H., Mitchell, K. S., Neale, M. C., & Kendler, K. S. (2010). 405. https://doi.org/10.1556/2006.6.2017.051. Eating disorder symptomatology and substance use disorders: Derogatis, L. R. (1990). SCL-90-R. Administration, scoring and Prevalence and shared risk in a population based twin sample. procedures manual. Clinical Psychometric Research. International Journal of Eating Disorders, 43(7), 648–658. Derogatis, L. R. (1994). SCL-90-R Cuestionario de 90 Síntomas - https://doi.org/10.1002/eat.20856. Manual. TEA Editorial. Becker, D. F., & Grilo, C. M. (2015). Comorbidity of mood and Fauconnier, M., Rousselet, M., Brunault, P., Thiabaud, E., Lam- substance use disorders in patients with binge-eating disorder: bertm, S., Rocher, B., . . . Grall-Bronnec, M. (2020). Food Associations with personality disorder and eating disorder pa- addiction among female patients seeking treatment for an thology. Journal of Psychosomatic Research, 79(2), 159–164. eating disorder: Prevalence and associated factors. Nutrients, https://doi.org/10.1016/j.jpsychores.2015.01.016. 12(6). https://doi.org/10.3390/NU12061897. 112 Journal of Behavioral Addictions 11 (2022) 1, 102–115 Finner, H., & Roters, M. (2001). On the false discovery rate and Hebebrand, J., Albayrak, Ö., Adan, R. A. H., Antel, J., Dieguez, C., expected type I errors. Biometrical Journal, 43(8), 985. https:// de Jong, J., . . . Dickson, S. L. (2014). “Eating addiction”, rather doi.org/10.1002/1521-4036(200112)43:8<985::AID-BIMJ985>3. than “food addiction”, better captures addictive-like eating 0.CO;2-4. behavior. Neuroscience and Biobehavioral Reviews, 47, 295–306. First, M. B., Williams, J. B. W., Karg, R. S., & Spitzer, R. L. (2015). https://doi.org/10.1016/j.neubiorev.2014.08.016. Structured clinical interview for DSM-5 research version. Jimenez-Murcia, S., Agüera, Z., Paslakis, G., Munguia, L., Granero, Washington D.C: American Psychiatric Association. R., Sánchez-González, J., . . . Fernández-Aranda, F. (2019). Fouladi, F., Mitchell, J. E., Crosby, R. D., Engel, S. G., Crow, S., Hill, Food addiction in eating disorders and obesity: Analysis of L., . . . Steffen, K. J. (2015). Prevalence of alcohol and other clusters and implications for treatment. Nutrients, 11(11). substance use in patients with eating disorders. European https://doi.org/10.3390/nu11112633. Eating Disorders Review, 23(6), 531–536. https://doi.org/10. Jimenez-Murcia, S., Giménez, M., Granero, R., Masuet-Aumatell, 1002/erv.2410. C., Ramón, J. M., Agüero, F., . . . Fernández-Aranda, F. (2021). Garner, D. M. (1991). Eating disorder inventory-2. Psychological Gambling disorder seeking treatment patients and tobacco use Assessment Resources. in relation to clinical profiles. Addictive Behaviors, 114, 106723. Garner, D. M. (1998). Inventario de Trastornos de la Conducta https://doi.org/10.1016/j.addbeh.2020.106723. Alimentaria (EDI-2). TEA ediciones. Jimenez-Murcia, S., Granero, R., Wolz, I., Baño, M., Mestre-Bach, Gearhardt, A. N., Boswell, R. G., & White, M. A. (2014). The as- G., Steward, T., . . . Fernández-Aranda, F. (2017). Food addic- sociation of “food addiction” with disordered eating and body tion in gambling disorder: Frequency and clinical outcomes. mass index. Eating Behaviors, 15(3), 427–433. https://doi.org/ Frontiers in Psychology, 8(473). https://doi.org/10.3389/fpsyg. 10.1016/j.eatbeh.2014.05.001. 2017.00473. Gearhardt, A. N., Corbin, W. R., & Brownell, K. D. (2016). Jiménez-Murcia, S., Tremblay, J., Stinchfield, R., Granero, R., Fer- Development of the Yale food addiction scale version 2.0. nández-Aranda, F., Mestre-Bach, G., . . . Menchón, J. M. (2017). Psychology of Addictive Behaviors: Journal of the Society of The involvement of a concerned significant other in gambling Psychologists in Addictive Behaviors, 30(1), 113–121. https://doi. disorder treatment outcome. Journal of Gambling Studies, org/10.1037/adb0000136. 33(3), 937–953. https://doi.org/10.1007/s10899-016-9657-z. Gearhardt, A. N., White, M., & Potenza, M. (2011). Binge eating Kendler, K. S., Lönn, S. L., Salvatore, J., Sundquist, J., & Sundquist, disorder and food addiction. Current Drug Abuse Reviews, 4(3), K. (2016). Effect of marriage on risk for onset of alcohol use 201–207. https://doi.org/10.2174/1874473711104030201. disorder: A longitudinal and co-relative analysis in a Swedish Granero, R., Hilker, I., Agüera, Z., Jimenez-Murcia, S., Sauchelli, S., national sample. American Journal of Psychiatry, 173(9), 911– Islam, M. A., . . . Fernández-Aranda, F. (2014). Food addiction 918. https://doi.org/10.1176/appi.ajp.2016.15111373. in a Spanish sample of eating disorders: DSM-5 diagnostic Khalil, R. B., Sleilaty, G., Richa, S., Seneque, M., Iceta, S., Rodgers, subtype differentiation and validation data. European Eating R., . . . Guillaume, S. (2020). The impact of retrospective Disorders Review: The Journal of the Eating Disorders Associa- childhood maltreatment on eating disorders as mediated by tion, 22(6), 389–396. https://doi.org/10.1002/erv.2311. food addiction: A cross-sectional study. Nutrients, 12(10), 2969. Granero, R., Jimenez-Murcia, S., Gerhardt, A. N., Agüera, Z., https://doi.org/10.3390/NU12102969. Aymamí, N., Gómez-Peña, M., . . . Fernández-Aranda, F. Kotyuk, E., Magi, A., Eisinger, A., Király, O., Vereczkei, A., Barta, C., (2018). Validation of the Spanish version of the Yale Food . . . Demetrovics, Z. (2020). Co-occurrences of substance use and Addiction Scale 2.0 (YFAS 2.0) and clinical correlates in a other potentially addictive behaviors: Epidemiological results sample of eating disorder, gambling disorder, and healthy from the Psychological and Genetic Factors of the Addictive control participants. Frontiers in Psychiatry, 9(208). https://doi. Behaviors (PGA) Study. Journal of Behavioral Addictions, 9(2), org/10.3389/fpsyt.2018.00208. 272–288. https://doi.org/10.1556/2006.2020.00033. Gregorowski, C., Seedat, S., & Jordaan, G. P. (2013). A clinical Krug, I., Pinheiro, A. P., Bulik, C. M., Jimenez-Murcia, S., Granero, approach to the assessment and management of co-morbid eating R., Penelo, E., . . . Fernández-Aranda, F. (2009). Lifetime sub- disorders and substance use disorders. BMC Psychiatry, 13(1), 1– stance abuse, family history of alcohol abuse/dependence and 12. BioMed Central. https://doi.org/10.1186/1471-244X-13-289. novelty seeking in eating disorders: Comparison study of eating Gutiérrez-Zotes, J. A., Bayón, C., Montserrat, C., Valero, J., Labad, disorder subgroups. Psychiatry and Clinical Neurosciences, A., Cloninger, C. R., & Fernández-Aranda, F. (2004). 63(1), 82–87. https://doi.org/10.1111/j.1440-1819.2008.01908.x. Temperament and Character Inventory Revised (TCI-R). Levallius, J., Monell, E., Birgegård, A., Clinton, D., & Forsén Standardization and normative data in a general population Mantilla, E. (2020). Binge eating and addictive-like behaviours sample. Actas Espanolas de Psiquiatria, 32(1), 8–15. in males and females. Psychological Reports. https://doi.org/10. Halcomb, M., Argyriou, E., & Cyders, M. A. (2019). Integrating 1177/0033294120971750. preclinical and clinical models of negative urgency. Frontiers in Lozano-Madrid, M., Clark Bryan, D., Granero, R., Sánchez, I., Psychiatry, 10(324). https://doi.org/10.3389/fpsyt.2019.00324. Riesco, N., Mallorquí-Bagué, N., . . . Fernández-Aranda, F. Hardy, R., Fani, N., Jovanovic, T., & Michopoulos, V. (2018). Food (2020). Impulsivity, emotional dysregulation and executive addiction and substance addiction in women: Common clinical function deficits could be associated with alcohol and drug characteristics. Appetite, 120, 367–373. https://doi.org/10.1016/ abuse in eating disorders. Journal of Clinical Medicine, 9(6), j.appet.2017.09.026. 1936. https://doi.org/10.3390/jcm9061936. Journal of Behavioral Addictions 11 (2022) 1, 102–115 113 Mallorquí-Bagué, N., Testa, G., Lozano-Madrid, M., Vintró- FODRAT study (FOod addiction, DRugs, alcohol and to- Alcaraz, C., Sánchez, I., Riesco, N., . . . Fernández-Aranda, F. bacco): First data on food addiction prevalence among pa- (2020). Emotional and non-emotional facets of impulsivity in tients with addiction to drugs, tobacco and alcohol. Eating eating disorders: From anorexia nervosa to bulimic spectrum and Weight Disorders, 26(2), 449–455. https://doi.org/10. disorders. European Eating Disorders Review, 28(4), 410–422. 1007/s40519-020-00865-z. https://doi.org/10.1002/erv.2734. Tran, H., Poinsot, P., Guillaume, S., Delaunay, D., Bernetiere, M., Martínez-Loredo, V., Macipe, V., Errasti Pérez, J. M., & Al-Halabí, Bégin, C., . . . Iceta, S. (2020). Food addiction as a proxy for S. (2021). Clinical symptoms and personality traits predict anorexia nervosa severity: New data based on the Yale food subpopulations of treatment-seeking substance users. Journal of addiction scale 2.0. Psychiatry Research, 293. https://doi.org/10. Substance Abuse Treatment, 125, 108314. https://doi.org/10. 1016/j.psychres.2020.113472. 1016/j.jsat.2021.108314. Um, M., Whitt, Z. T., Revilla, R., Hunton, T., & Cyders, M. A. Minhas, M., Murphy, C. M., Balodis, I., Acuff, S. F., Buscemi, J., (2019). Shared neural correlates underlying addictive disorders Murphy, J. G., & MacKillop, J. (2021). Multidimensional ele- and negative urgency. Brain Sciences, 9(2). https://doi.org/10. ments of impulsivity as shared and unique risk factors for food 3390/brainsci9020036. addiction and alcohol misuse. Appetite, 159, 105052. https://doi. Verdejo-García, A., Lozano, Ó., Moya, M., Alcázar, M. A., & Pérez- org/10.1016/j.appet.2020.105052. García, M. (2010). Psychometric properties of a Spanish version Nøkleby, H. (2012). Comorbid drug use disorders and eating dis- of the UPPS-P impulsive behavior scale: Reliability, validity and orders — a review of prevalence studies. Nordic Studies on association with trait and cognitive impulsivity. Journal of Alcohol and Drugs, 29(3), 303–314. https://doi.org/10.2478/ Personality Assessment, 92(1), 70–77. https://doi.org/10.1080/ v10199-012-0024-9. 00223890903382369. Polk, S. E., Schulte, E. M., Furman, C. R., & Gearhardt, A. N. Volkow, N. D., Wang, G. J., Fowler, J. S., Tomasi, D., & Baler, R. (2017). Wanting and liking: Separable components in prob- (2011). Food and drug reward: Overlapping circuits in human lematic eating behavior? Appetite, 115, 45. https://doi.org/10. obesity and addiction. Current Topics in Behavioral Neurosci- 1016/J.APPET.2016.11.015. ences, 11, 1–24. https://doi.org/10.1007/7854_2011_169. Romero, X., Agüera, Z., Granero, R., Sánchez, I., Riesco, N., Whiteside, S. P., Lynam, D. R., Miller, J. D., & Reynolds, S. K. Jimenez-Murcia, S., . . . Fernández-Aranda, F. (2019). Is food (2005). Validation of the UPPS impulsive behaviour scale: A addiction a predictor of treatment outcome among patients four-factor model of impulsivity. European Journal of Person- with eating disorder? European Eating Disorders Review, 27(6), ality, 19, 559–574. https://doi.org/10.1002/per.556. 700–711. https://doi.org/10.1002/erv.2705. Wiedemann, A., Carr, M., Ivezaj, V., & Barnes, R. (2021). Exam- Root, T. L., Pinheiro, A. P., Thornton, L., Strober, M., Fernández- ining the construct validity of food addiction severity specifiers. Aranda, F., Brandt, H., . . . Bulik, C. M. (2010). Substance use Eating and Weight Disorders: EWD, 26(5), 1503–1509. https:// disorders in women with anorexia nervosa. International doi.org/10.1007/S40519-020-00957-W. Journal of Eating Disorders, 43(1), 14–21. https://doi.org/10. Wolz, I., Granero, R., & Fernández-Aranda, F. (2017). A compre- 1002/eat.20670. hensive model of food addiction in patients with binge-eating Stata-Corp, & StataCorp (2019). Stata statistical software: Release. symptomatology: The essential role of negative urgency. StataCorp LLC, 16(16.1). Comprehensive Psychiatry, 74, 118–124. https://doi.org/10. Steingrimsson, S., Carlsen, H., Lundström, E., Lundström, S., & 1016/j.comppsych.2017.01.012. Nilsson, T. (2020). Problematic alcohol and drug use is associated Wu, Y. K., Zimmer, C., Munn-Chernoff, M. A., & Baker, J. H. with low self-directedness and cooperativeness. European Addic- (2020). Association between food addiction and body dissatis- tion Research, 26(6), 326–334. https://doi.org/10.1159/000506473. faction among college students: The mediating role of eating Tinghino, B., Lugoboni, F., Amatulli, A., Biasin, C., Bramani expectancies. Eating Behaviors, 39, 101441. https://doi.org/10. Araldi, M., Cantiero, D., . . . Nizzoli, U. (2021). The 1016/j.eatbeh.2020.101441. 114 Journal of Behavioral Addictions 11 (2022) 1, 102–115 APPENDIX Supplementary material Table S1. Internal consistency in the sample (Cronbach's alpha) TCI-R Novelty seeking 0.786 EDI-2: Drive for thinness 0.801 SCL-90R Somatization 0.891 TCI-R Harm avoidance 0.898 EDI-2: Body dissatisfaction 0.870 SCL-90R Obsessive 0.851 TCI-R Reward dependence 0.836 EDI-2: Interoceptive awareness 0.798 SCL-90R interpersonal Sensitivity 0.857 TCI-R Persistence 0.898 EDI-2: Bulimia 0.778 SCL-90R Depression 0.901 TCI-R Self-directedness 0.863 EDI-2: Interpersonal distrust 0.809 SCL-90R Anxiety 0.871 TCI-R Cooperativeness 0.848 EDI-2: Ineffectiveness 0.864 SCL-90R Hostility 0.851 TCI-R Self-transcendence 0.849 EDI-2: Maturity fears 0.772 SCL-90R Phobic anxiety 0.826 UPPS-P Lack premeditation 0.844 EDI-2: Perfectionism 0.714 SCL-90R Paranoid ideation 0.750 UPPS-P Lack perseverance 0.825 EDI-2: Impulse regulation 0.783 SCL-90R Psychoticism 0.811 UPPS-P Sensation seeking 0.869 EDI-2: Ascetism 0.701 SCL-90R GSI 0.976 UPPS-P Positive urgency 0.909 EDI-2: Social insecurity 0.773 SCL-90R PST 0.976 UPPS-P Negative urgency 0.828 EDI-2: Total score 0.945 SCL-90R PSDI 0.976 YFAS-2 0.965 Table S2. Clinical profiles of patients with FA and SU FA & AL FA & DR FA & AL & DR FA & AL FA & AL FA & DR vs FA & AL vs FA & AL N 5 18 N 5 77 N 5 30 vs FA & DR & DR & DR Mean SD Mean SD Mean SD P |d| P |d| P |d| ED severity (EDI-2) Drive for thinness 14.44 5.91 15.38 5.13 16.87 3.77 0.475 0.17 0.104 0.51y 0.166 0.33 Body dissatisfaction 20.00 7.34 17.64 7.62 19.30 5.79 0.211 0.32 0.744 0.11 0.284 0.25 Interoceptive awareness 15.56 7.49 13.13 7.12 15.53 6.93 0.196 0.33 0.992 0.00 0.120 0.34 Bulimia 10.17 5.01 8.21 5.25 11.53 4.59 0.142 0.38 0.368 0.28 0.003* 0.67y Interpersonal distrust 5.22 2.90 6.01 4.99 7.70 5.17 0.530 0.19 0.086 0.59y 0.105 0.33 Ineffectiveness 15.44 6.72 11.82 7.52 14.50 7.89 0.067 0.51y 0.674 0.13 0.100 0.35 Maturity fears 10.56 7.56 8.52 5.78 10.30 7.06 0.225 0.30 0.893 0.03 0.197 0.28 Perfectionism 6.17 4.44 5.94 4.29 7.70 5.01 0.844 0.05 0.254 0.32 0.070 0.38 Impulse regulation 10.83 7.36 7.60 6.63 11.77 8.08 0.084 0.46 0.660 0.12 0.007* 0.56y Ascetic 9.44 4.68 7.78 4.35 9.03 4.33 0.150 0.37 0.754 0.09 0.187 0.29 Social insecurity 8.44 4.31 8.60 5.30 10.03 5.64 0.912 0.03 0.313 0.32 0.207 0.26 Total score 126.3 42.85 110.6 42.00 134.3 40.23 0.154 0.37 0.522 0.19 0.010* 0.57y Psychop. (SCL-90R) Somatization 2.34 0.91 1.94 0.92 2.34 0.88 0.092 0.44 0.995 0.00 0.040* 0.45 Obsessive-compulsive 2.36 0.97 1.97 0.71 2.34 0.87 0.060 0.46 0.929 0.02 0.031* 0.47 Interpersonal sensitivity 2.22 1.02 2.09 0.96 2.41 0.90 0.609 0.13 0.493 0.20 0.118 0.35 Depressive 2.78 0.90 2.37 0.86 2.70 0.87 0.079 0.46 0.764 0.09 0.084 0.38 Anxiety 2.18 0.98 1.78 0.96 2.29 0.84 0.102 0.42 0.693 0.12 0.012* 0.57y Hostility 1.74 1.20 1.55 1.04 1.49 0.94 0.486 0.17 0.420 0.23 0.787 0.06 Phobic anxiety 1.52 1.13 1.23 1.03 1.36 0.99 0.276 0.27 0.591 0.16 0.560 0.13 Paranoia 1.61 0.95 1.44 0.86 1.97 0.80 0.446 0.19 0.161 0.41 0.005* 0.64y Psychotic 1.66 0.88 1.37 0.81 1.79 0.67 0.170 0.34 0.569 0.17 0.015* 0.57y GSI 2.17 0.84 1.85 0.76 2.21 0.69 0.109 0.40 0.855 0.05 0.029* 0.50y PST 70.00 19.09 65.73 15.81 73.43 14.66 0.311 0.24 0.474 0.20 0.028* 0.51y PSDI 2.69 0.59 2.45 0.58 2.64 0.56 0.111 0.42 0.763 0.09 0.128 0.33 Impulsivity (UPPS-P) Lack premeditation 24.83 5.22 24.92 5.62 25.97 6.98 0.954 0.02 0.522 0.18 0.414 0.16 Lack perseverance 24.94 5.08 24.65 6.39 25.07 4.91 0.849 0.05 0.945 0.02 0.743 0.07 Sensation seeking 25.78 5.04 27.34 8.14 29.63 8.33 0.448 0.23 0.101 0.56y 0.176 0.28 Positive urgency 30.50 8.26 29.82 9.52 34.93 10.38 0.786 0.08 0.123 0.47 0.014* 0.51y Negative urgency 34.94 5.60 35.44 5.99 36.87 5.54 0.745 0.09 0.271 0.35 0.258 0.25 (continued) Journal of Behavioral Addictions 11 (2022) 1, 102–115 115 Table S2. Continued FA & AL FA & DR FA & AL & DR FA & AL FA & AL FA & DR vs FA & AL vs FA & AL N 5 18 N 5 77 N 5 30 vs FA & DR & DR & DR Mean SD Mean SD Mean SD P |d| P |d| P |d| Personality (TCI-R) Novelty seeking 102.1 21.81 104.3 17.53 112.1 13.09 0.625 0.11 0.054 0.56y 0.038* 0.50y Harm avoidance 118.8 22.17 121.5 19.88 123.7 19.63 0.610 0.13 0.423 0.23 0.624 0.11 Reward dependence 98.7 17.81 99.2 17.73 98.1 15.04 0.900 0.03 0.917 0.03 0.766 0.07 Persistence 103.6 22.95 105.6 21.65 106.9 21.57 0.722 0.09 0.618 0.15 0.796 0.06 Self-directedness 103.4 21.68 110.7 20.22 99.6 14.83 0.149 0.35 0.511 0.20 0.008* 0.63y Cooperativeness 131.9 22.10 129.6 15.29 129.4 16.08 0.599 0.12 0.611 0.13 0.949 0.01 Self-transcendence 68.1 16.87 64.2 13.02 70.4 18.66 0.326 0.26 0.617 0.13 0.061 0.38 Note. FA: food addiction present. AL: Alcohol disorder. DR: drugs disorder. SD: standard deviation. *Bold: significant parameter. yBold: effect size into the range mild-moderate (|d| > 0.50) to large-high (|d| > 0.80). Fig. S1. Flow-chart with the sampling Note. AN: anorexia nervosa. BN: bulimia nervosa. BED: binge eating disorder. OSFED: other specified feeding or eating disorder Open Access. This is an open-access article distributed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (https:// creativecommons.org/licenses/by-nc/4.0/), which permits unrestricted use, distribution, and reproduction in any medium for non-commercial purposes, provided the original author and source are credited, a link to the CC License is provided, and changes – if any – are indicated.