Participants

In 1985, all 4291 individuals born in 1970 and attending eighth grade in schools in Gothenburg, Sweden, were screened for AN. The eight-graders completed an ED symptom questionnaire and underwent a physical examination. One of the authors (MR) scrutinized the growth charts and ED questionnaires of all 4291 children and examined in person each case with a suspicion of AN. In total, 25 individuals (23 females, two males) met the criteria for AN in the 1970 birth cohort. One of the girls refused an in-depth examination, leaving 22 females and two males to constitute the population-based group. A population screening group was also recruited for the study and consisted of 27 adolescents with AN (26 females, 1 male) born in 1969 and in 1971–1977. The population screening group was reported to the researchers by the school health service. The groups were combined to constitute the AN group, consisting of 51 subjects (48 women, 3 men) (see [34, 35] for details). All 51 adolescents met the DSM-III-R [36] and the DSM-IV [37] criteria for AN. A comparison group (COMP group) was recruited, the individuals were selected by the school health nurses and were matched for gender, age and school. In the original study physical and psychiatric symptoms were examined in all individuals in the AN and COMP group, and their mothers were interviewed by a psychiatrist concerning family situation, early development, temperament, personality, physical and mental symptoms. The adolescents of the COMP group had no history of ED. Concerning psychiatric comorbidity, at the time of the first examination, 86% percent of the adolescents in the AN group had an axis I diagnosis (other than AN) or depressed mood, compared with 22% of the individuals in the COMP group. The majority of the diagnoses were mood disorders [34].

Previous examinations of the sample

The 51 individuals in the AN group and the 51 individuals in the COMP group have previously been examined on four occasions (AN study 1, AN study 2, AN study 3 and AN study 4), at a mean age of 16, 21, 24 and 32 years [16, 34, 35, 38, 39]. There has been no attrition in any of the previous follow-up studies.

The present study

All 102 participants in the AN and the COMP groups were contacted and invited to participate in AN study 5, 30 years after the onset of AN, at a mean age of 44 years. Four individuals in the AN group (two women, two men) declined participation, leaving 47 participants. All 51 individuals in the COMP group agreed to participate. Data on the male participants were excluded due to their small number, leaving 46 and 48 participants in the AN and COMP group respectively to answer “the general questions about reproduction and offspring” (described in detail below). Participants with offspring were asked if they would agree to take part in interviews regarding their offspring’s health status. Thirty-three females in the AN group and 40 women in the COMP group agreed to participate. In addition, complementary data were obtained from the Swedish Medical Birth Register (SMBR) (see supplementary figure S1).

Outcome measuresPsychiatric health of the mothers

The MINI International Neuropsychiatric interview [40] was used to assess psychiatric disorders, including EDs, in the mothers. The Eating disorder module of the Structured Clinical Interview for DSM-IV-TR Axis I Disorders (SCID-I) [41] and a DSM-5 [42] criteria checklist for feeding and eating disorders was also used to assign ED diagnoses. The Global Assessment of Functioning (GAF) Scale [37] was applied to measure general outcome. Weight and height were measured on the day of the interviews [43].

General questions about reproduction and offspring

All subjects were asked three brief questions about reproduction and offspring: (1) whether the mother had any biological/adopted or stepchildren (and, if so, the gender, age and number of children born); (2) whether the participant had any first-degree relatives with underweight or overweight, any eating-related problems, including EDs, and (3) whether their offspring had any severe somatic or psychiatric disorder.

The results presented below is based on data regarding the biological children in the AN and COMP group.

Perinatal status measures

The perinatal outcome measures were based on data from the SMBR. The following perinatal variables were selected: gestational age, birth weight, birth length, head circumference, Apgar score (Appearance, Pulse, Grimace response, Activity, Respiration) [44] at 1, 5 and 10 min, and perinatal death (death within six days of birth). The ponderal index (PI), a standard measure for reporting body mass in newborns (kg/m3) was calculated. Preterm birth was defined as a dichotomous variable (birth at < 259 days of gestation). Maternal weight and height were recorded when the mother was enrolled at the maternal health care center, usually between week 6–10 of pregnancy, and were used as a measure of pre-pregnancy body mass index.

Measures of health and psychiatric problems in the offspring

For 0–4-year-old children, a brief semi-structured interview was conducted, targeting developmental problems in infancy, including sleep, colic, breastfeeding, selective eating, parental concern regarding weight and psychomotor development. The interview was used for the first time in AN Study 4, the 18-year follow-up [16]. The interview has not been tested for validity or reliability. The mothers completed the Strengths and Difficulties Questionnaire (SDQ) regarding children 2–17 years old. The SDQ is a brief well-established, well-validated questionnaire used for general mental health screening in children [45] that has been translated into more than 50 languages. The SDQ comprises 25 items that are divided into five subscales: emotional symptoms, conduct problems, hyperactivity, peer problems and prosocial behavior. Each of the items is rated on a three-point Likert scale: “not true”, “somewhat true”, or “certainly true”. The score of each subscale ranges from 0–20; the higher the score, the more symptoms. The prosocial behavior subscale is inverted. The SDQ is completed by parents reporting on the child’s behavior in the past six months. The Swedish adaptation of the questionnaire has been validated, and the following cut-off scores for values above the 90th percentile have been proposed: total difficulties ≥ 14; emotional symptoms ≥ 5; conduct problems ≥ 4; hyperactivity ≥ 6; peer problems ≥ 5; prosocial behavior ≤ 6 [46]. Regarding children younger than 5 years, cut-off scores suggested by Dahlberg, Fält, Ghaderi, Sarkadi, & Salari [47] were used (total difficulties ≥ 12; emotional symptoms ≥ 3; conduct problems ≥ 4; hyperactivity ≥ 5; peer problems ≥ 2; prosocial behavior ≤ 6).

In the study of psychiatric health in offspring at or above age 18 years, the same interview procedure as for the assessment of psychiatric disorders in the mothers was used, i.e. the mother answered the questions concerning their adult offspring. The parent-reported ADHD Rating Scale IV [48] was administered to investigate Attention-Deficit/Hyperactivity Disorder for children at or above age 5 years. The developmental and well-being assessment (DAWBA), a well-validated parental interview, was used to assess psychiatric health in 5–17-year-old children. The instrument generates ICD-10 and DSM-IV diagnoses. The DAWBA has separate sections covering individual emotional, behavioral and hyperactivity disorders, including a section for diagnosing EDs. A web-based version (www.dawba.net) [49] was administered. The timeframe of the interview is the present and recent past; for instance, anxiety disorders are based on the past four weeks while conduct disorders are based on the previous twelve months.

Interview instruments were selected according to the age of the offspring (see Fig. 1). The interviewer was blinded to the participant’s group status. Before the interview the participant was informed that she should not reveal whether she had a history of AN or not. The interviewer was blinded until all interviews were completed. The majority of the offspring interviews were performed face-to-face, although online video conferences/telephone interviews were performed with twelve of the mothers (AN group, n = 3; COMP group, n = 9).

Fig. 1

Interview instruments administered to the mothers in the AN and the COMP group regarding the offspring’s health

As a complement to the interviews, the Swedish National Patient Register (NPR) was searched for any psychiatric diagnoses coded according to the International Statistical Classification of diseases and Related health problems, 10th edition [50], codes F00-F99. The NPR covers diagnoses of inpatient care from 1973 and specialized outpatient care from 2001. We had access to data from January 1, 1996, until December 31, 2015. Of the children below 18 years of age, 59 out of 67 (88%) in the AN group, and 57 out of 74 (77%) in the COMP group were found in the NPR.

Physical health and body mass index of the offspring

Physical disorders were identified in the NPR using ICD-10 codes for in- and outpatient visits. All mothers of 5–25-year-old offspring were asked to report weight and height of their children, and body mass index (BMI) was calculated and converted to standard deviation scores (SDS) based on the specific age and gender of each child. Four BMI categories were derived: underweight (< -1 SD), normal weight (± 1 SD), overweight (> + 1SD) and obese (> + 2SD).

Ethics

Data were collected from May 2015 through November 2016. The study was approved by the Regional Ethical Review Board at the University of Gothenburg (398–14). The mothers participated voluntarily after giving their informed consent. Research based on offspring register data was approved for children below age 18 years. For this reason, no register data were extracted for offspring who had turned 18 at the time of the examination of the mother. Informed consent was not collected from the offspring, as they were not participating in the study.

Statistical analyses

Statistical analysis was carried out by using SPSS, version 28.0. Group comparisons between the mothers in the AN and COMP group were tested mainly with non-parametric tests, due to the data not being normally distributed. The Mann–Whitney test was used for continuous variables and Chi-squared test or Fisher’s Exact test for dichotomous variables. Offspring data were mainly analyzed with Generalized Estimating Equations (GEE) for dichotomous variables, and for continuous variables, a mixed model was used to account for clusters among the women as some contributed to the dataset with multiple births. The significance threshold was set at a p value of < 0.05.

Attrition

Total offspring interview attrition applied to seven mothers of 19 children in the AN group and two mothers of four children in the COMP group, leaving 33 and 38 mothers in the AN and COMP group respectively to answer the questions. Of the participating mothers, four children in the AN and three in the COMP group were evaluated only with the SDQ. One of the mothers in the AN group who did not complete the offspring interviews had a current ED. The seven mothers in the AN group who declined the offspring interviews did not differ significantly from the participating mothers with regard to age, BMI and ED duration (mean age: non-participants: 45.3, SD:1.4; participants: 44.4, SD:1.8, p = 0.11; mean BMI: non-participants: 21.4, SD: 1.6; participants: 22.8, SD: 4.6, p = 0.53; mean ED duration in years: non-participants: 6.6, SD: 3.3; participants: 9.7, SD: 7.6, p = 0.30). Additional register data gave us complementary information on the offspring’s status. Combining all sources, at least some data regarding physical and mental health were available on 90.2% and 100% of the offspring in the AN and COMP group, respectively. Thirty-six of the women in the AN group and 32 of the mothers in the COMP group were found in the SMBR, providing data on 82 (88%) and 68 (79%) neonates, respectively. Participants living abroad/giving birth abroad were not found in the SMBR and this applied to four mothers in the AN group and six mothers in the COMP group. Two other women in the COMP group were missing in the SMBR, one had not yet given birth at the time the register data were obtained and one was missing for unknown reasons.