Table 3 Other studies looking for predictors of anorexia nervosa but not initially designated for this purpose
From: Systematic review of prospective studies assessing risk factors to predict anorexia nervosa onset
Authors | Year of pub | Country | Cohort size | Initial age | Follow-up period (years) | Initial cohort objective | Measuring risks factors tolls | ED Diagnoses tools | Results | Attrition bias | |||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Statistical method | Retained risk factors | Association with AN onset | Cut-off calculation | ||||||||||
Sanci L et al.[34] | 2008 | Australia | 1936 | 16 to 24 | 11 | Investigation of mental health, personality, behaviors, school, family, drug and alcohol use | CSA (childhood sexual aggression) before the age of 16 years (retrospectively ascertained at the age of 24 years) | Branched Eating Disorders Test, 9-item Adolescent Dieting Scale | Cross-sectional odds ratios, discrete-time proportional hazards models | CSA without physical contact and with physical contact | CSA without physical contact: 1 report: OR = 1.2 (0.19–7.7); ≥ 2 reports: OR = 2.4 (0.8–7.2) CSA with physical contact: 1 report: OR = 1.1 (0.17–7.5); ≥ 2 reports: OR = 1.6 (0.46–5.6) | No | Multiple imputation |
Buckner JD et al.[44] | 2010 | USA | 1709 | 16 | 14 | Epidemiology of depression | K-SADS, K-SADS-P | Structured clinical interview DSM-IV, non-patient version | Multivariate logistic regression, odds ratios | Obsessive–compulsive disorder (OCD) | OCD: OR = 49,26, 95% CI: 1.57 to 1541.47, p = .03 | No | Taken into account |
Meier SM et al.[35] | 2015 | USA | 1709 | 16 | 14 | Epidemiology of depression | K-SADS | SCID-for DSM-IV non-patient version | Multivariate logistic regression, odds ratios | Specific anxiety disorders | OCD: OR = 49.26, p = 0.03 | No | Take into account |
Copeland WE et al.[36] | 2015 | USA | 1420 | 9 to 16 | 10 | Child psychiatric epidemiology | Parent report and self-questionnaire on childhood bullying involvement | CAPA/YAPA eating disorder module assesses DSM-III-R and IV symptoms | Regression models with robust Variance, OR, MR | Bullying victims Bullies | Any AN symptom: OR = 4.6, 95% CI: 1.8–11.4, p = 0.001 OR = 1.9, 95% CI: 1.2–3.2, p = 0.01 Total AN symptom: no significant result | No | No |
Peñas-Lledó E et al.[37] | 2016 | Sweden | 615 | 19 to 20 | 3 | Investigation of Child and Adolescent Development | Psychometric questionnaires | Binary auto test response | K-means clustering | Drive for thinness and/or negative affect scores | Drive for Thinness-Anxiety/Depression cluster had the highest prevalence of AN at follow-up compared with all other clusters | No | No |
Ranta K et al. [38] | 2017 | Finland | 3278 | 15 to 16 | 2 | Adolescents’ mental health study | Self-reported SPIN, Self-reported Beck Depression Inventory | Self-reported questionnaires’ DSM IV criteria | Chi-square statistics, OR | Depression | OR = 3.6 1.3–10.1), p = 0.015 | No | Taken into account |
Schaumberg K et al.[39] | 2019 | England | 7767 | 10 | 6 | Global health study cohort | Parental symptoms Psychometric scores, BMI | Questionnaire using DSM-5 criteria, Parental report of AN symptoms at ages 14 and 16 | EFA, Geomin rotation, ESEM | Anxiety symptoms | OR = 1.10; p = 0.01 | No | Taken into account |
Yilmaz Z et al. [40] | 2019 | England | 1502 | 12 | 5 | Subslamp from a global health study cohort | BMI | Self-reported ED disorders based on the DSM 5 criteria, self-reported ED behaviors, parental reports of fear of fatness and restricting behaviors | Parametric and non-parametric analyses | BMI at age of 12 | Girls: lower BMIs by 4 years (B =  − 0.505, SE 0.22) Boys: lower BMIs by 2 years (B =  − 0.87, SE 0.38) | No | Take into account |
Lloyd EC et al.[41] | 2020 | England | Study 1: 14,882 Study 2: several cohorts[66, 68, 69] | 10 | Max 14 | 2 studies combined: global health study cohort + Mendelian randomization | Psychometric scores, parental report | Parental report of AN symptoms at ages 14 and 16. And Self child report at 14, 16, 18, 24 | Multiple logistics regression, multivariate, OR, Mendelian randomization | Anxiety symptoms | Study 1: Worry at age 10: OR = 1.60, 95% CI: 0.93 to 2.77, p = 0.09; anxiety disorder criteria at 10: OR = 2.85, 95% CI: 1.22 to 6.63, p = 0.02 Study 2: Worry increases AN risk: OR = 1.79, 95% CI: 1.25 to 2.55, p = 0.001 | No | Not clear |
Talmon A et Spatz Widom C [42] | 2021 | USA | 807 | 0 to 11 | Max 41 | Abused and/or neglected children followed into adulthood | Documented childhood maltreatment, psychometric scores | Diagnostic Interview Schedule, Version 4 | Logistic and linear regression, OR, adjusted OR | Neglect Physical abuse Sexual abuse | OR: 1.32, p = 0.10 OR: 1.10, p = 0.02 OR: 1.53, P = 0.01 | No | Taken into account |
Bufferd SJ et al.[43] | 2022 | USA | 609 | 3 to 6 | Max 12 | Risk factors for psychopathology | Psychometric questionnaires | DSM-IV anorexia nervosa (AN) and bulimia nervosa (BN) symptoms | Multivariate linear regression model | Perceptual sensitivity, Oppositional Defiant Disorder (ODD) and depressive disorder in early childhood | r correlation index from 0.11 to 0.15 (p < 0.05) | No | Take into account |