The search for gastrointestinal inflammation in autism: a systematic review and meta-analysis of non-invasive gastrointestinal markers.

Very high heterogeneity in the calprotectin meta-analysis (I² = 89.07%), meaning the pooled estimate is imprecise and the null result is not definitive. Most included studies were small and underpowered, so a true subgroup difference in calprotectin could easily be missed. Many studies failed to screen non-autistic control participants for GI conditions, which could inflate apparent differences or mask real ones. The majority of participants were male, and no study included enough autistic girls to examine sex differences. No studies included adult participants, so findings cannot be generalised beyond children and adolescents. Age was not adequately controlled in most studies, which matters because calprotectin is physiologically higher in children under four. For 32% of the pooled autistic sample (191 children), it was unclear whether participants with limited verbal or cognitive abilities were included.

Funded by the Australian Government Department of Education, Skills and Employment, Neuroscience Research Australia, the Schizophrenia Research Institute, and a National Health and Medical Research Council Investigator Grant. No industry funding or commercial conflicts declared.

Do not read the null result as proof that GI inflammation never occurs in autism. The calprotectin meta-analysis had very high heterogeneity (I² = 89.07%), most included studies were small, and the authors themselves note a subgroup with elevated calprotectin may have been missed by underpowered studies. A single negative faecal calprotectin test does not rule out GI inflammation in an individual autistic child.