Principal Pathogenetic Components in Autism Spectum Disorder: A Validation Study

Background: Applying principal component analysis (PCA), we previously described four pathogenetic components: I) “circadian and sensory dysfunction”, II) "immune dysfunction”, III) “neurodevelopmental delay", and IV) "stereotypic behavior” (Sacco et al., 2010). These four pathogenetic components subsequently defined four patient clusters: a) ICS (immune abnormalities and some circadian and sensory issues); b) CS (major circadian and sensory dysfunction, with little or no dysimmunity); c) S (prominent stereotypies) and (d) M (patients with a mixture of all four components associated with developmental delay).

Objectives: To replicate and to extend these results using an independent sample.

Methods: PCA followed by correlation analysis was performed on 19 clinical and family history variables in an independent sample of 192 ASD patients admitted to our Unit subsequently to the previous recruiting period (Sacco et al., 2010). This sample includes 163 males and 29 females, mean age 10.6 years (SD ± 6.8), belonging to 163 simplex and 29 multiplex Italian families.

Results: PCA revealed a component structure similar, but more complex compared to the previous structure, now encompassing five components which altogether explain 50% of the total variance. These include: 1) a “developmental delay” component, including various developmental milestones (non verbal and verbal language development, social smile onset, independent walking, acquisition of bladder and bowel control). These variables were also related to a positive history of obstetric complications and/or repeated spontaneous abortions in the mother; 2) a “lower functioning” component including intellectual disability, motor and/or verbal stereotypies, self-injurious behaviors; 3) a “mixed” component including sleep disorders, abnormal pregnancy duration,  history of regression and hypotonia; 4) an “allergic” component  including allergies in the patient, history of obstetric complications and/or repeated spontaneous abortions in the mother, level of verbal language development, and hyperactivity; 5) an “autoimmune” component including a family history positive for autoimmune diseases, history of any infectious disease at autism onset (i.e., recurrent otitis, upper airway infections, etc), excessive pain tolerance.

Conclusions: This validation study confirms the existence of  the “Developmental delay” and “Immune” components, which seemingly represent consistent patterns in the ASD clinical phenotype in addition to the “low functioning” component. Interestingly, the former “Immune” component from Sacco et al. (2010) splits here into two components, one “allergic” and one “autoimmune”, suggesting two different pathways underlying the dysimmune endophenotype in ASD. Despite an extreme and long-recognized clinical heterogeneity, the autistic phenotype can be partly, yet consistently dissected on the basis of simple clinical and family history variables, allowing future investigations over patient clustering, biomarker identification, developmental trajectories, therapeutic outcomes and strategies.