Table 1.
Characteristics and findings of reviewed studies.
| References | Country | N | Cohort | Age (M ± SD) | Gender | Follow-up | ED diagnosis | ED diagnostic threshold | Findings | Quality |
|---|---|---|---|---|---|---|---|---|---|---|
| Ackard et al. [40] | United States | 2,516 | Project EAT-II; 31 schools | 20.4 ± 0.8 at follow-up | 45% male | 5 years | Self-report questions re: ED criteria (DSM-IV) | Full threshold ED cases | 2.2% males and females with body image disturbance developed BED or BN. 0.9% of males and 3.7% of females with binge eating or compensatory behaviors developed BED, or BED and BN, respectively. | Poor |
| Allen et al. [41] | Australia | 1,597 | Raine study; birth cohort | 14.1 ± 0.1 at follow-up | 55% male | 14 years | Adapted child EDE and EDEQ (DSM-IV) | Full, partial, and “at risk” ED cases | A restricted diet predicted later ED cases. | Good |
| Beato-Fernandez et al. [42] | Spain | 1,076 | 22 schools | 12–13 at baseline | 46% male | 2 years | EAT-40 (DSM-III-R/IV) | Full (AN, BN) and EDNOS cases | Body dissatisfaction predicted later ED diagnosis. Those who developed an ED had more suicidal and self-harm tendencies, worse depressive symptoms, and body dissatisfaction at baseline. | Poor |
| Herle et al. [43] | United Kingdom | 4,760 | ALSPAC | 1.3 at baseline | 52% male | 8.7 years | Data from YRBSSQ and DEBQ (DSM-5) | Full threshold ED cases | Childhood overeating, persistent undereating, and persistent fussy eating were predictive of adolescent binge eating disorder, anorexia nervosa (girls only), and anorexia nervosa (whole sample), respectively. | Good |
| Johnson et al. [44] | United States | 726 | Families from two NY counties | 13.8 ± 2.6 at baseline | 49% male | 10 years | DISC-I (DSM-III-R) | Full Threshold ED cases and ED behaviors | Depressive and anxiety disorders were predictive of EDs. Disruptive and substance use disorders were not predictive of EDs. | Good |
| Killen et al. [45] | United States | 887 | 4 schools | 12.4 ± 0.7 at baseline | All female | 3 years | Self-report questions (diagnostic tool not reported) | Full and partial ED cases | Elevated weight concerns associated with onset of partial EDs. Earlier drive for thinness, body dissatisfaction, perfectionism, and restraint most characteristic of those who later develop partial EDs | Poor |
| Killen et al. [46] | United States | 825 | 4 schools | 14.9 at baseline | All female | 4 years | Interview and adapted EDE (DSM-III-R) | Full and partial ED cases | Elevated weight concerns associated with onset of partial EDs. Earlier drive for thinness, bulimia, body dissatisfaction, ineffectiveness, interoceptive awareness, temperament (distress, fear), dieting, restraint, and alcohol use most characteristic of those who later develop partial EDs. | Good |
| Kotler et al. [47] | United States | 976 | Families from two NY counties | 6.1 at baseline | 50% male | 17 years | DISC (DSM-III-R/IV) | Full threshold ED cases and ED behaviors | Eating conflicts, struggles with food and unpleasant meals predicted AN. | Good |
| Liechty and Lee [48] | United States | 14,322 | Add Health; 80 high schools | 15.9 ± 1.8 at baseline | 49% male | 7 years | Self-report (ever told have an ED? yes/no) | Implied full threshold ED cases | Depression, body image dissatisfaction (males only), and extreme weight loss behaviors (females only) was associated with unspecified ED diagnosis. | Good |
| Marchi and Cohen [49] | United States | 659 | 9 schools | ~6 years at baseline | 51% male | 10 years | DISC (DSM-III-R) | Full threshold ED cases | Risks in early childhood for subsequent symptoms of anorexia nervosa include picky eating and digestive problems. Early symptoms of AN/BN predictive of later diagnosis and pica predictive of BN. |
Good |
| Neumark-Sztainer et al. [50] | United States | 2,516 | Project EAT-II; 31 schools | 12.8/15.8 ± 0.8 at baseline | 45% male | 5 years | Self-report (ever told have an ED? yes/no) | Implied full threshold ED cases and ED behaviors | In females, dieting significantly associated with later EDs. | Poor |
| Nicholls and Viner [51] | United Kingdom | 16,567 | BCS70; birth cohort | From birth | 56% male | 30 years | Self-report (ever told have an ED? yes/no) | Implied full threshold ED cases | Infant feeding problems, under eating, and increased exercise (at 10 years) associated with later AN. | Poor |
| Nicholls et al. [52] | United Kingdom | 16,567 | BSC70; birth cohort | From birth | 49% male | 30 years | Self-report (ever told have an ED? yes/no) | Implied full threshold ED cases | Eating, sleep problems, and overeating (5 years) associated with later BN/BED. | Poor |
| Patton et al. [53] | Australia | 1,947 | 45 schools | 14.5 ± 0.5 at baseline | 47% male | 3 years | BET (DSM-IV) | Full and partial syndrome ED cases | In females, severe and moderate dieting increased risk of future ED. Psychiatric morbidity (independent of dieting) predicted onset of EDs. | Poor |
| Penas-Lledo et al. [54] | Sweden | 615 | TChAD; Twin study | 16–17 years at baseline | All female | 3 years | Self-report (ever had BN or AN? Yes/no) | Implied full threshold ED cases | Drive for thinness predicted later BN. Interaction between drive for thinness and anxious/depressed mood predicted risk of both BN and AN. | Poor |
| Ranta et al. [55] | Finland | 3,278 | Finnish Adolescent Mental Health Cohort Study | 15.5 ± 0.4 at baseline | 51% male | 2 years | Self-report questions re: ED criteria (DSM-IV-TR) | Implied full threshold ED cases | Depression predicted AN, while social phobia and depression both predicted BN. However, controlling for initial comorbidity, Eds, and socioeconomic factors removed this effect. | Poor |
| Schaumberg et al. [56] | United Kingdom | 7,767 | ALSPAC | 10 years | Not reported | 6 years | Data from YRBSSQ & DEBQ (DSM-5) | Full threshold ED cases | Physical anxiety symptoms predicted BN, while worries (e.g., about the future) predicted AN. | Fair |
| Stice et al. [57] | United States | 496 | 8 schools | 13.5 ± 0.7 at baseline | All female | 2 years | EDE (DSM-IV) | Full and partial threshold ED cases | Depressive symptoms (but not substance abuse) predicted onset of subthreshold BN. | Good |
| Stice et al. [58] | United States | 496 | 8 schools | 15.4 ± 0.7 at baseline | All female | 5 years | EDDI (DSM-IV) | Full and partial threshold ED cases | Fasting (not eating for 24 hours) for weight control more predictive of future sub/full BN than dietary restraint. | Good |
| Stice et al. [59] | United States | 496 | 8 schools | 13.5 ± 0.7 at baseline | All female | 8 years | EDDI (DSM-IV) | Full and partial threshold ED cases | Body dissatisfaction predictive of ED onset. Those with high or low body dissatisfaction and elevated depressive or dieting associated with increased ED onset respectively. | Good |
| Stice and Van Ryzin [60] | United States | 496 | 8 schools | 13.5 ± 0.7 at baseline | All female | 8 years | EDDI (DSM-IV) | Full and partial threshold ED cases | Growth curve models showed perceived pressure to be thin and/or thin-ideal internalization, before showing onset of disorder-predictive levels of body dissatisfaction, before showing onset of disorder-predictive levels of dieting and/or negative affect, before showing onset of the ED. | Good |
| Wilkinson et al. [61] | United Kingdom | 945 | Roots study | 14 at baseline | 47% male | 3 years | K-SADS-PL (DSM-IV) | Full threshold ED cases | Recurrent nonsuicidal self-injury predicted onset of EDs. | Good |
Abbreviations: ALSPAC, UK Avon Longitudinal Study of Parents and Children; AN: anorexia nervosa; BCS70, 1970 British Cohort Study; BED: binge eating disorder; BET, Branched Eating Disorders Test; BN: bulimia nervosa; DEBQ, Dutch Eating Behavior Questionnaire; DISC-II, Diagnostic Interview Schedule for Children, Version 2; DSM (-III-R/-IV/-5), Diagnostic and Statistical Manual of Mental Disorders, Third (Revised)/Fourth/Fifth Edition; EAT-40: Eating Attitudes Test; ED, eating disorder; EDDI: Eating Disorder Diagnostic Interview; EDE, Eating Disorder Examination; EDE-Q, Eating Disorder Examination Questionnaire; EDI: Eating Disorder Inventory; EDNOS, Eating Disorder Not Otherwise Specified; KSADS-PL, Kiddie Schedule for Affective Disorders and Schizophrenia, Present and Lifetime version; Project EAT-II, Eating Among Teens; TChAD, Twin study of Child and Adolescent Development; YRBSSQ, Youth Risk Behavior Surveillance System Questionnaire.