Objectives: Determine if response to treatment and weight gain following 8 weeks of risperidone treatment is associated with baseline serum levels and changes in cytokine levels.
Methods: We used data of the plasma levels of 27 different cytokines from 41 children who had an initial Aberrant Behavior Checklist Irritability (ABC-I) subscale rating of ≥18. Subjects were uptitrated to a daily total of 1.5 mg of risperidone over 8 weeks. Outcome measures were % decrease in ABC-I (%ABC-I) scores and Clinical Global Impression-Improvement (CGI-I) score. Subjects were overall responders to treatment if they had a decrease in %ABC-I>.25 and a CGI-I of “very much improved” or “much improved”. Cytokine analysis was performed using multiplex assays (Millipore) according to manufacturer's recommendation and read on Luminex 100TMplatform. Weight gain was represented through change in z-scores of anthropomorphically-adjusted BMI distributions. We assessed normality of the data and identified outliers and did not adjust p-values for multiple testing for this exploratory analysis. Correlations were determined using Spearman’s rank correlation coefficient, differences between baseline and post-treatment values were determined using Wilcoxin sign rank tests and changes between responder and nonresponder groups were determined using Mann-Whitney tests.
Results: D %ABC-I (p<.00001) was significantly decreased following 8 weeks of treatment. Plasma levels of Eotaxin (p=.0005) and MCP1 (p=.02) was significantly decreased. There are no significant correlations between D%ABC-I and change in cytokines levels. Baseline IL-15 is a significant predictor of ABC-I response only (p=.04) with an adjusted (sex, age) odds ratio of 3.5. The median values of DIL-5 (p=.006) are significantly higher in the overall responder group compared to nonresponders and DIL-15 (p=.07) is marginally higher. In comparing the highest responders (decrease %ABC-I>.60) to nonresponders, median values of DIL-15 (p=.04) are significantly higher and DIL-5 (p=.07) is marginally higher. There are no differences in baseline measure.There is no association between weight gain and response. Weight gain is negatively correlated with DIL-5 (p=.02) and DINFa2 (p=.01). We found no baseline predictors of weight gain.
Conclusions: Responders had a greater increase in IL-5 compared to nonresponders. Greater overall improvement is associated with significant increases in IL-5 and IL-15. Those with higher levels of baseline IL-15 are more likely to have a reduction in irritability symptoms alone. There is a significant decrease in Eotaxin and MCP1 levels following 8 weeks of risperidone treatment. Previous studies report increased MCP-1 and Eotaxin levels in children with autism when compared to controls and are associated with higher ABC-I scores. The effects on cytokines due to weight gain and baseline cytokine levels are not predictive biomarkers of weight gain.