Longitudinal Sex Differences in ASD Symptoms and Diagnostic Stability in Early Childhood

Background: Research has demonstrated sex differences in the early screening of Autism Spectrum Disorder (ASD) in toddlers; however, studies examining the early clinical presentation of ASD have shown minimal sex differences at this age. Given research showing sex differences later in life, it is important to understand the potential developmental trajectory of sex differences and how this may impact early detection and diagnosis.

Objectives: The current study examined the relationship between sex and the developmental trajectory of clinical features and diagnostic stability of ASD in a sample of children initially evaluated based on screen positive status (i.e., at risk) on the Modified Checklist for Autism in Toddlers(-Revised) with Follow-Up (M-CHAT(-R/F) and re-evaluated two years later.

Methods: The sample included 424 (N_male=318) children who were initially evaluated as toddlers (M_age=26.31 months, SD=4.46) and re-evaluated two years later (M_age=49.88 months, SD=7.19). Evaluations included autism diagnostic measures (ADOS-2), cognitive testing (Mullen Scales of Early Learning), and parent report of history and adaptive behavior (Vineland Adaptive Behavior Scales(-II)). Notably, ADOS-2 Calibrated Severity Scores (CSS) for Overall Total, Social Affect and Restricted, Repetitive, Behavior (RRB), as well as algorithm domain scores were utilized for analyses given minimal research examining the domain-specific CSS as an adequate measure of longitudinal changes in symptoms.

Results: Sex differences emerged in diagnostic stability. Although a similar percentage of males (8%) and females (12%) who were not diagnosed with ASD at the initial evaluation gained a diagnosis at the time of re-evaluation (X²(1,N=238)=.562, p=.453), girls initially diagnosed with ASD were almost twice as likely (38%) as males (20%) to lose their ASD diagnosis at re-evaluation (X²(1,N=270)=3.647, p=.05). Subsequent analyses on the subsample of children who lost an ASD diagnosis at re-evaluation (N=51) indicated a significant interaction between time of evaluation and sex (F(1,47)=4.38, p=.043, partial η²=.085) for RRB algorithm domain scores; males demonstrated higher scores (M_male=3.09, SD=2.01) compared to females (M_female=1.50, SD=1.71) at the initial evaluation, but similar levels at re-evaluation (M_male=.85, SD=1.18; M_female=.63, SD=.81;). Toddlers with stable ASD diagnosis– ASD confirmed at both time points (N=219; N_male=177) – did not evidence differential ASD symptoms, developmental skills, or adaptive skills based on sex across time (ps>.05).

Conclusions: Results are consistent with previous literature demonstrating minimal sex differences in young children with ASD and extend previous findings by showing no sex differences in the developmental trajectory of the clinical features of ASD from toddlerhood to the preschool years. Nevertheless, sex differences emerged in diagnostic stability such that females were more likely to lose an initial diagnosis of ASD from age 2 to 4. While females demonstrated low rates of RRBs at both time-points, males demonstrated a significant decline in RRBs over time. This suggests that in females diagnosed with ASD in toddlerhood, fewer RRBs may be predictive of losing a diagnosis at age 4. Results call for future research examining early compensatory factors in females that may facilitate symptom reduction that lead them to lose their initial ASD diagnosis by preschool.