The clinical identification of regression phenomena in ASD lacks specific biological or laboratory criteria and is often based on family history and highly subjective observations by clinicians. The present study aimed to investigate the potential role of plasma clusterin (CLU), very long-chain fatty acids (VLCFA), and carnitine as biomarkers of neurodegeneration in children with autism spectrum disorder (ASD) with and without regression. By exploring these biomarkers, we sought to provide insights into mitochondrial dysfunction, glial activation, and lipid metabolism, which may contribute to the pathophysiology of ASD and aid in the early diagnosis and intervention of regression phenomena in ASD. Ninety children aged 2-6 years were included: 30 with autism spectrum disorder (ASD), 30 with regressive ASD, and 30 healthy controls. Psychiatric assessments were conducted using DSM-5 criteria, CARS, ABC, RBS-R, and ASSQ scales. Regression in ASD was evaluated retrospectively using a modified ADI-R questionnaire. Fasting blood samples were collected, and plasma clusterin (CLU), VLCFA, and carnitine levels were measured. Statistical analyses were performed using MANOVA to assess the effect of group differences on dependent biochemical variables. Serum clusterin and carnitine levels showed no significant differences between groups. However, C22 VLCFA levels were significantly higher in both autism groups compared to controls (p = 0.04), with post hoc analysis indicating the difference between the non-regressive and control groups (p = 0.02). Serum carnitine was positively correlated with stereotypic behaviors subscale scores (r = 0.37, p = 0.004) and total scores (r = 0.35, p = 0.006) of RBS-R. Our study provides insights into the complexities of biomarker research in autism spectrum disorder (ASD), highlighting the challenges in identifying consistent biological markers for regression and non-regression phenotypes. Although no significant findings were observed, further biomarker studies are essential to distinguish possible endophenotypes, improve early diagnosis, and uncover potential therapeutic targets in ASD.