Characteristics of Children with ASD Who Improve with Fever: Insights from the Simons Simplex Collection

Background: Anecdotal evidence suggests that a subset of children with ASD may show improvements during episodes of fever. The only prospective study on this topic (Curran et al, 2007) found that 83% of children with ASD showed significant improvements on at least one domain of the Aberrant Behavior Checklist (ABC) during an episode of fever. Overall, little is known about the broader behavioral profiles of children with ASD who are reported to improve with fever.

Objectives:  The purpose of this work is to explore the behavioral characteristics of children whose parents report improvements during episodes of fever.

Methods: Using data from the Simons Simplex Collection (SSC), parents of 2,152 children between the ages of 4 and 18 (mean=8.9, SD=3.6) provided information about whether and in what areas their child improved during fever. Children who were reported to improve during fever (n=362; Improve Group) and those who reportedly do not improve during fever (n=1690; No Improve Group) were randomly assigned into discovery or replication samples (See Figure 1). The discovery sample consisted of 850 children (50%) from the No Improve group and 183 children (51%) from the Improve Group while the replication sample consisted of 840 children (50%) from the No Improve Group and 179 children (49%) from the Improve Group. Improve and No Improve groups in the discovery sample were compared on 1) demographic variables (age, sex, race), 2) NVIQ, 3) language level (per ADOS module), 4) parent report measures of medical history, ASD symptoms, repetitive behavior, adaptive skills, 5) clinical observation of ASD symptoms, and 6) presence of an ASD-associated de novo genetic mutation. Results from the discovery analyses that met the statistical threshold of α < .05 were compared in the replication sample to confirm results.

Results: 362 children (17%) were reported to improve during fever across a range of domains (Temper/Behavior: 59%, Communication: 49%, Social Interaction: 41%, Repetitive Behaviors: 38%, and Cognition/Learning: 21%). Discovery analyses revealed that children in the Improve Group had significantly lower NVIQ [t(1031) = 3.70, p < .001] and language level [χ²(2, N = 1033) = 6.79, p = .03], as well as more restricted, repetitive behaviors across several domains on the Repetitive Behavior Scale-Revised and ABC. These results were consistent during replication analyses. In discovery analyses, children in the Improve Group were statistically more likely to have breathing problems during sleep [χ²(1, N = 1024) = 7.94, p< .01], but this result was not significant in the replication sample. Groups did not differ in age, race, sex, current parent-reported symptoms of ASD or in the proportion of children with de novo mutations.

Conclusions: Parents of a subset of children with ASD report improvements during episodes of fever. These children have lower language skills, lower NVIQ, and more restricted, repetitive behaviors. Understanding the behavioral profiles of these children may provide insights into the etiology of ASD and lead to innovative treatments.