Early neurodevelopmental problems and risk for avoidant/restrictive food intake disorder (ARFID) in 4-7-year-old children: A Japanese birth cohort study

Abstract

Background: An overrepresentation of neurodevelopmental problems (NDPs) has been observed in individuals with avoidant/restrictive food intake disorder (ARFID). Previous studies on the association between ARFID and NDPs have been limited by cross-sectional data from clinical samples of small size. This study aimed to extend previous research by using prospectively collected data in a non-clinical child cohort. We examined the occurrence of early NDPs in 4-7-year-old children with suspected ARFID and how predictive early NDPs are of ARFID.

Methods: Data were collected via parent-report a sub-sample of the Japan Environment and Children's Study (JECS) including 3728 children born 2011-2014 in Kochi prefecture. NDPs were assessed biannually between 0.5 and 3 years of age with the Ages and Stages Questionnaire-3, at age 2.5 years with the ESSENCE-Q, and at age 1 and 3 years via parent-reported clinical diagnoses. ARFID was identified cross-sectionally (at age 4-7 years) using a newly developed screening tool. Logistic regressions were used to test association of (1) a composite early NDP risk score, (2) specific early NDPs, and (3) neurodevelopmental trajectories over time with ARFID.

Results: Children in the highest risk percentiles of the NDP risk score had roughly three times higher odds of having suspected ARFID; the absolute risk of later ARFID for children above the 90th percentile was 3.1%. Early NDPs (excluding early feeding problems) were more predictive of later ARFID than were early feeding problems. Specific NDPs predictive of ARFID were problems with general development, communication/language, attention/concentration, social interaction, and sleep. Neurodevelopmental trajectories of children with and without suspected ARFID started to divert after age 1 year.

Conclusions: The results mirror the previously observed overrepresentation of NDPs in ARFID populations. In this non-clinical child cohort, early feeding problems were common and rarely developed into ARFID; however, our findings imply that they should be monitored closely in children with high NDP risk to prevent ARFID.

Keywords: Japan Environment and Children's Study (JECS); attention‐deficit/hyperactivity disorder; autism spectrum disorder; eating disorder; neurodevelopmental disorders.

FIGURE 1

Overview of assessments, predictors, and outcomes in the current study. Predictors were derived as follows. 1. Neurodevelopmental risk score. Score below the referral cut‐off on an Ages and Stages Questionnaire‐3 (ASQ‐3) domain: 1 risk point (6 time points of measurement on 5 ASQ‐3 domains: max. 30 points) + ESSENCE‐Q items (except feeding problems), rated with “yes”: 1 risk point (10 items: max. 10 risk points) + parent‐reported diagnoses of developmental delay (DD; at age 1 year), motor delay (MD), intellectual disability (ID) or autism spectrum disorder (ASD; at age 3 years: 1 risk point per diagnosis (4 diagnoses: max. 4 risk points) = theoretical range 0–44. 2. Specific Neurodevelopmental problems. Each of 10 ESSENCE‐Q items and each ASQ‐3 domain that was failed 2 or more out of 6 time points of measurement. 3. ASQ‐3 score. Score below monitoring cut‐off: 1 risk point/domain, score below referral cut‐off: 2 risk points/domain, × 5 domains = 0–10 per time point

FIGURE 2

Ages and Stages Questionnaire (ASQ‐3) risk score across age by ARFID group. The x‐axis represents age in years. The y‐axis represents the mean ASQ‐3 risk score in the ARFID group (dashed line) versus the no‐ARFID group (solid line) with a theoretical range of 0–10 points per measurement point (below the monitoring cut‐off on an ASQ‐3 domain: 1 risk point; below the referral cut‐off on an ASQ‐3 domain: 2 risk points; aggregated over the five ASQ‐3 domains). Error bars represent the standard error of the mean