This prospective controlled community-based 30-year follow-up study of adolescent-onset AN aimed at examining perinatal status, psychiatric and physical health in the offspring of mothers with a history of AN. In addition, perinatal status of the children was investigated. At least some data were available on 90.2% of the offspring in the AN group and all of the offspring in the COMP group. Birth weight and length, head circumference and ponderal index were significantly reduced in the offspring of the mothers in the AN group. After the perinatal period, BMI did not differ between the children in the AN and COMP groups. Parental interviews (DAWBA, MINI) indicated a higher prevalence of current psychiatric disorders among the offspring of mothers with a history of AN. The SDQ assessment did however not confirm the presence of more psychiatric symptoms among the children in the AN group. Furthermore, an overrepresentation of endocrinological, immune and metabolic disorders was observed among the offspring in the AN group.
As we had hypothesized, offspring in the AN group exhibited worse perinatal outcomes (lower birth weight and length, lower ponderal index and smaller head circumference) compared with offspring in the COMP group. This finding is consistent with other studies in terms of lower birth weight and smaller head circumference [8, 14]. We did not find an elevated risk of preterm birth or lower Apgar scores in the AN group, which is in agreement with some other reports [15, 51]. The observed reduced birth size in the offspring signals that children of mothers with lifetime AN constitute a risk group for a worse perinatal outcome. Stringer and Furber emphazise the importance that midwifes identify women with ED early in pregnancy and tailor interventions appropriately [52]. We suggest that maternity care routines to identify women with ongoing or previous ED should be established.
Moreover, we had hypothesized that the offspring in the AN group would be at an increased risk of developing psychiatric disorders throughout childhood. Current psychiatric symptoms were not more common in the offspring in the AN group compared with the offspring in the COMP group, using the SDQ. This is in line with findings by Stein et al.[26]. On the other hand, Micali et al. [23] and Barona et al. [24], who also used the SDQ, found an increased prevalence of emotional and conduct disorders in the offspring of mothers with a history of AN. These two studies comprised larger samples than the present study, and we cannot exclude the possibility that our results are due to our sample being underpowered.
The diagnostic interviews concerning current psychiatric disorders in the offspring indicated that psychiatric disorders in general were overrepresented among the offspring in the AN group. The results are consistent with findings by Micali and colleagues [23], showing an overrepresentation of emotional disorders, including anxiety disorders, among children of mothers with EDs. Ten percent of the offspring in the AN group fulfilled the criteria for an anxiety disorder, although the difference failed to reach statistical significance compared with 1.4% of the offspring of the COMP group. Extending the groups to include offspring of all ages (combining register data with interview data), lifetime psychiatric disorders did not differ between the AN and the COMP group.
There was a discrepancy between the diagnostic interview results (DAWBA, MINI) and the SDQ data. The former showed an overrepresentation of current psychiatric disorders among the offspring of the AN group, while the latter did not signal more psychiatric symptoms among the children in the AN group. The difference could reflect the different age ranges of the children involved in the questionnaires versus the interviews; the children rated with the SDQ were younger (mean age: 11.1, age range: 2–17 years) than the offspring in the diagnostic interviews (AN group mean age: 14.8, range: 5–25 years; COMP group mean age: 13.2, range: 5–24 years).
In our sample, six children in the AN group and one adolescent in the COMP group were diagnosed with a lifetime ED, the difference was not statistically significant (p = 0.061). An overrepresentation of EDs among the offspring of mothers with AN, as has been observed by other groups [28, 53]. Our sample included children in a wide range of ages and a considerable minority of the children in the present study had not yet reached puberty, when the onset of AN and other EDs usually takes place [54]. Although EDs have a genetic component explaining the likelihood of a higher prevalence of EDs in children of mothers with EDs, environmental factors have been put forward as crucial mediators in the expression of underlying predisposition [55]. For instance, nutritional intake during pregnancy may influence the DNA methylation in genes related to growth, metabolism and appetite regulation [56]. Further, the literature implies that a parent with ED impacts on her/his child’s feeding and eating patterns. Stein and colleagues found that during mealtime interactions between mothers with EDs and their children, the mothers expressed more intrusive behaviors and negative emotions towards the infant compared with the interaction between healthy control mothers and their offspring [32]. Our sample was too small to investigate the impact of environmental factors and the mediation of gene expression.
We observed a much higher prevalence of endocrinological, immune and metabolic diseases among the offspring in the AN group. The results need to be replicated using larger samples. The findings are, however, in line with previous reports of an association between EDs and various autoimmune disorders [57, 58]. Furthermore, recent findings have indicated that parental mental illness is associated with an elevated risk of autoimmune diseases in the offspring [59]. In addition, eating disorders have been linked to Paediatric acute-onset neuropsychiatric syndrome (PANS) as one defining feature of some cases with sudden onset of eating restrictions [60]. In individuals with PANS a family history of inflammatory or autoimmune diseases has been reported to be common [61]. Contrary to our expectations, we found no support for impaired growth after the perinatal period in the children born to mothers with a history of AN. Although the ponderal index was lower at birth among the offspring in the AN group, the majority of these children had a normal BMI later on.
Strengths And Limitations
This is the only study that has followed a group of individuals with adolescent-onset AN prospectively, along with a matched comparison group, for as long as 30 years. The sample is community-based and half of the AN cases constituted a total birth cohort. The individuals have been examined on five occasions after their onset of AN. Some larger cohort studies examining psychiatric health in the offspring of mothers with ED have relied on parental questionnaires when studying the offspring’s development e.g. [24]. Our group has, however, interviewed the mothers in person using a semi-structured interview, and therefore the interview data can be considered as reliable [62]. The parental interviews resulted in psychiatric diagnoses in the children according to the DSM-IV, the DSM-5 and the ICD-10. Having access to register data enabled us to obtain complementary information regarding the children. Combining different data sources enabled us to obtain at least some information on physical and mental health concerning 90.2% and 100% of the offspring in the AN and COMP group, respectively. The perinatal data covered 88% and 79% of the children in the AN and COMP group, respectively. To the best of our knowledge, this is the first study examining several aspects of health, including perinatal status, psychiatric and somatic morbidity, in the offspring of mothers with a history of AN.
The current study has some obvious limitations. The sample size was relatively small, which is in contrast to some larger register-based studies that have been performed in this area [22, 24, 25]. The small sample size also prevented us from performing gender-stratified analyses. This must be considered a limitation since a previous study found gender differences between children when parents with EDs completed the SDQ [24]. The children were not examined by the researchers. Instead, we had to rely on the parental reports using established and well-validated interviews. Although parent reported weight and height can be considered reasonably accurate regarding children [63], reports of anthropometric measures in adolescent and adult offspring might be less reliable. In order to gain further information, personal assessments of the offspring would have been valuable. Regarding the diagnostic interviews, the offspring in the COMP group were significantly younger than the offspring in the AN group. The children in the former group had therefore had less time to develop psychiatric disorders, including EDs, than the children in the latter group. Another limitation was the wide age range of the offspring groups (from newborn to young adulthood). Further, we had no access to information on psychiatric history of the children’s fathers which limits our understanding of the familial transmission. The phenomenon of nonrandom mating patterns, both within and across psychiatric disorders, might additionally increase the risk of psychiatric disorders in offspring of women with a history of AN as suggested by Nordsletten et al. [64].