Peripheral blood inflammatory and metabolic biomarkers in autism spectrum disorder: A systematic review and...

Peripheral-blood studies in autism spectrum disorder (ASD) are often interpreted through an immunometabolic lens, yet inflammatory and metabolic findings have not been benchmarked within a common quantitative framework. We tested whether published case-control evidence is more consistent with cross-domain convergence or inflammatory-dominant asymmetry. In a preregistered systematic review (INPLASY202620085), 264 studies were eligible and 132 contributed 355 effect sizes across 27 biomarkers. Multilevel random-effects models estimated biomarker effects, and a preregistered domain-contrast model compared inflammatory and metabolic domains; because participant-level cross-biomarker covariance was unavailable, this contrast describes evidence structure rather than individual immune-metabolic coupling. The inflammatory domain exceeded the metabolic domain (Δg = 0.616, 95% CI 0.290-0.942, p = 3.15 × 10⁻⁴). The inflammatory summary was positive (g = 0.463, 95% CI 0.300-0.626), whereas the metabolic summary was near null (g = -0.087, 95% CI -0.199-0.025). Eleven of 14 inflammatory markers had positive estimates; TNF-α and IL-6 passed inflammatory-family FDR, whereas no metabolic marker reached nominal significance. Biomarker-level inference was nevertheless bounded: prediction intervals crossed zero, neither TNF-α nor IL-6 survived global FDR, GRADE certainty was Moderate for TNF-α and Very low for IL-6, and fasting-restricted analyses attenuated both below nominal significance. The evidence therefore favors inflammatory-dominant asymmetry over broad immunometabolic disturbance, but only as a group-level, non-clinical signal. Individual diagnosis, stratification, disease specificity, and participant-level metabolic co-dysregulation remain unestablished. Prospective standardized panels should replicate the inflammatory signal and test metabolic markers within defined subgroups.