Reference(s):

1. Main Reference: Wade, T. D., Tiggemann, M., Bulik, C. M., Fairburn, C. G., Wray, N. R., & Martin, N. G. (2008). Shared temperament risk factors for anorexia nervosa: a twin study. Psychosom Med, 70(2), 239-244.

Article(s):

1. Wade Shared Temperament Risk Factors for AN - a Twin Study.pdf

Discussion:

Thanks to Beth for a great presentation. Here are some highlights (mostly from Beth - Thanks, Beth!) from the paper and conversation:

INTRO

• The goal of the paper is to determine the link between dimensional temperament and risk factors for AN
• Past work does not distinguish between chicken and egg (i.e. are traits risk factors for AN or could they be a result of living with AN?)
• Authors argue that if certain characteristics are more present in unaffected relatives of AN pts than in unaffected relatives of controls, that the characteristic is a risk factor for AN rather than something caused by AN
  • One question about this is couldn’t the pt with AN have affected the temperament of the unaffected relative?  Adoption studies might be better suited to this question

• Authors would like to answer 2 questions: 1) what traits are associated with AN? And 2) what traits are transmitted with AN and may be seen as risk factors?

METHODS

• Australian twin registry was established from 1980 to 1982 and participants have been bombarded with multiple waves of data collection since then.  Current study is drawn from Wave 3 female twin pairs.
• Mean age at time of data collection = 35 +/- 2 years.  Included 348 complete pairs (226 MZ and 122 DZ) and 360 incomplete pairs (170 MZ and 190 DZ)

Measures

• Self-report measures used for temperament: some data was from initial wave of collection, and two measures were added.

• Eysenck: extraversion, psychoticism and neuroticism

• Tridimensional Personality Scale: measure response to environmental stimuli: novelty seeking, harm avoidance, and reward dependence

• Interpersonal Sensitivity Measure: developed to assess hypersensitivity to interpersonal rejection, a suggested trait of depression-prone personality

• Barratt Impulsiveness: unidimensional model of impulsiveness usually included as a part of a larger groups of personality pre-dispositions such as extraversion, sensation seeking, and a lack of inhibitory behavioural controls.  Further research led Barratt to classify impulsivity in three main aspects: motor (acting without thinking), cognitive (quick decisions), and non-planning (present orientation)

• Frost Multidimensional: developed at Smith College to measure concern about mistakes, personal standards, doubts about actions, organization (used to include parental criticism and parental standards)

• Telephone interview for EDE

• They did not screen for comorbid mood disorders, something I think is particularly important given the fact that they are asking for self-report and recall of symptoms.  If someone has an underlying mood disorder, their perceptual distortion greatly influences the responses they might give and cause them to appear higher in areas of neuroticism or self-criticism, for example.

RESULTS

• 19 women met criteria for lifetime AN.  16 more met criteria for AN with the exception of amenorrhea and 8 met criteria for AN for amenorrhea status was unclear.

• Mean age of developing AN was 17 +/- 2.69 years.  None of the women met BMI criteria for AN at assessment but 1 had binge/purge behaviors and 1 had purging behaviors.

• The authors don’t comment on several things here. 

• What was the mean length of time these women were ill?
• What was the mean lowest BMI?

• What kind of treatment did these women have?
• What is the breakdown of AN-r vs. AN-p?

• These questions could lead to different hypotheses about whether severity of illness correlates with dimensional temperament.  For example, are women with higher scores on doubts about actions, personal standards more likely to have lower BMIs?  Do women with high scores in different dimensional traits respond differently to different types of therapy?  Do women with AN-r vs. AN-p have different dimensional traits?

• Within person Associations with AN

• Concern over mistakes, personal standards, organization and doubts about actions were associated with AN
• This (according to the authors) answers their first question, which was “What traits cosegregate with AN?”
• They interpret these traits as indicating higher levels of perfectionism

• Cross-Twin Associations with AN

• Remember, the idea here is to determine whether unaffected relatives of women with AN have higher scores on certain temperament traits than unaffected relatives of women without AN
• Personal standards, organization and reward dependence were increased compared to control probrands sisters

• Cross-Twin Associations within zygosity groups

DISCUSSION

• Replicates idea that AN is associated with perfectionist traits
• Authors argue that the cross twin associations of personal standards, organization and reward dependence indicate that these are RISK FACTORs for AN rather than results of or simply cosegragating factors.
• How can we use these risk factors, if indeed that is what they are, in the clinical setting?  Are we obligated to be screening young women for these traits and providing them with…what exactly? Education about eating disorders?  The idea of screening is that a test must accurately detect something at an early stage and that an intervention must be available to potentially ward off the bad outcome.  What intervention do we have in place?  Something a la Alice Waters and education about food?  More frequent pediatrician visits?  Family therapy and education about nutrition/cooking together/eating together in a pseudo-Maudsley fashion?

Also, if you weren't there, you missed the CUPCAKES!!!!

Graham