Relationship Between Oxytocin Levels and Brain Volume in Anorexia Nervosa Compared to Healthy Controls

Abstract

Background: A disruption in food motivation pathways has been described in females with anorexia nervosa (AN), a psychiatric disorder characterized by food restriction despite low weight. We have shown that in AN and healthy controls (HC), levels of oxytocin (OXT), a hormone involved in lactation, social behavior and weight regulation, decrease after a meal. In HC, we identified a relationship between postprandial change in OXT and subjective appetite, yet this association was absent in females with AN, suggesting a disconnect between OXT and appetite regulation in AN. Prior studies have shown that gray matter volume of the amygdala and hippocampus, areas rich in OXT receptors, correlate with OXT levels in HC. Furthermore, these regions play a central role in food reward and decreased volume has been reported in AN. We hypothesized that the relationship between postprandial change in OXT and amygdala and hippocampal gray matter volume would differ between AN and HC. Methods: We performed a cross-sectional study of 51 females (23 restrictive AN; 28 HC). We drew blood for OXT levels fasting and 60 min after a standard meal and performed T1-weighted MRI scans of the brain in the fasted state. MRI data was quality controlled and processed with FreeSurfer. Average gray matter brain volumes were extracted from the bilateral amygdala and hippocampus for each subject. Linear regression models were used to determine differences between AN and HC of postprandial percent change in OXT on amygdala and hippocampus gray matter volume. Results: Median [IQR] age was higher in females with AN (20.6 years [19.3, 21.5]) than HC (18.8 years [IQR 17.6, 20.3], p=0.02), and percentage of ideal body weight was lower in AN (75.5%) than HC (97.4%, p<0.01). Right hippocampus volume, adjusted for age and total intracranial volume, was significantly lower in AN (estimated difference -188 dm3 [95%-CI -360, -17], p=0.04). Percent change in OXT was not different (p=0.5) but there was a trend for a positive interaction effect (p=0.08) for AN and percent change in OXT on right hippocampus volume. Posthoc exploratory analysis indicated a positive correlation in AN (R2=0.41, p=0.02) and no correlation in HC (R2=0.17,p=0.4) between percent change in OXT and right hippocampus volume. There was no significant between group difference in volume nor postprandial change in OXT for the bilateral amygdala or left hippocampus between groups. Discussion: Our results indicate smaller right hippocampus volume and a trend towards a positive association with postprandial change in OXT in AN compared to HC. Future studies will be important to better define the relationships between OXT secretion and food motivation brain regions and the impact on eating behavior in AN.