The ASD Phenotype in Neurofibromatosis Type 1: Evidence From a Two-Phase Population-Based Epidemiological Study

Background:  Neurofibromatosis Type 1 (NF1) is a common autosomal dominant disorder (birth incidence > 1 in 2,700), caused by mutation on chromosome 17 (17q11.2), with a well elucidated neuro-pathological pathway involving the RASMapkinase intracellular signalling pathway. Although noted for somatic manifestations, the main morbidity from NF1 is actually in cognitive, social and behavioural difficulties. Characterisation of the behavioural phenotype in NF1 has lacked well-designed epidemiological studies on representative samples, with previous studies on clinic-referral populations.

Objectives:  To estimate the population prevalence of Autism Spectrum Disorders (ASD) in NF1 using rigorous methodology.

Methods:  A two-phase population-based design using standardised assessment. In a phase 1 screening stage all parents of children aged 4-16 from a whole population NF1 Genetics Register in the North-West of England were invited to participate. 109/207 (53%) completed the Social Responsiveness Scale (SRS) to estimate autism symptoms, along with other questionnaire measures. Invitation to the Phase 2 in-depth phenotyping stagewas stratified by random number generation of cases within three SRS groups. We assessed 97% of the clinical-range SRS scorers (T scores >75); 72% of the subclinical scorers (T score 60-75); and 31% of the normal-range scorers (T score <60). The Autism Diagnostic Interview-Revised (ADI-R), the Autism Diagnostic Observation Schedule (ADOS-G) and the Verbal IQ items of the Wechsler Abbreviated Scale of Intelligence (WASI) were delivered by experienced administrators. High inter-rater reliability on ADOS scores was confirmed on 20% of the sample.  Collaborative Program of Excellence in Autism (CPEA) criteria (Lainhart et al. 2006) were used to combine IQ, ADI-R and ADOS-G scores into ASD and ‘Broad ASD’ (partial features) categories. Statistical analysis used weighting of this phase 2 CPEA categorisation in order to estimate the representative population prevalence of ASD and Broad ASD.

Results:  Responders and non-responders in both phases did not differ on key characteristics. In Phase 1 28.7% (31/109) of children scored in the severe, clinical autism range on the SRS (T score >75); 29.9% (29/109) in the mild to moderate range (T score 60-75) and 44% of children in the normal range (T score <60). In Phase 2, 29.8% of the sample met CPEA ASD criteria and a further 27.6% showed ‘Broad ASD’ with partial features. Mean verbal IQ was 95.56(16.61) and did not differ across these groups.  20/47 (42.5%) met ADOS-R ASD criteria and 12/47 (25.5%) met ASD criteria on 3 ADI-R domains. ASD was not associated with the extent of physical illness, age, SES, or familial inheritance of NF1.

Weighting these results by the probability of responding to phase 1 gives a whole-population prevalence for ASD of 24.9% (95%CI 13.1%, 42.1%) and for Broad ASD of 20.8% (95%CI: 10.0%, 38.1%).

Conclusions: This two-phase epidemiological study yields an ASD population prevalence in NF1 of 25%, with a further 21% showing "broad ASD" with partial features. Because the pathogenesis of NF1 is known in detail and there are interventions theoretically impacting on the pathway, this makes NF1 an important single gene model of autism with considerable potential for future research in illuminating the nature of the disorder.