Autism (AUT) is a complex neurodevelopmental disorder that, together with Asperger's syndrome and Pervasive Developmental Disorder-Not Otherwise Specified (PDD-NOS), comprises the expanded classification of autistic spectrum disorder (ASD). The heterogeneity of ASD underlies the need to identify biomarkers or clinical features that can be employed to identify meaningful subtypes of ASD, define specific etiologies, and inform intervention and treatment options. Previous studies have shown that disordered porphyrin metabolism, manifested principally as significantly elevated urinary concentrations of pentacarboxyl (penta) and coproporphyrins, is commonly observed among some children with ASD. Here, we extend these observations by specifically evaluating penta and coproporphyrins as biological indicators of ASD among 76 male children comprising 30 with validated AUT, 14 with PDD-NOS, and 32 neurotypical (NT) controls. ASD children (AUT and PDD-NOS) had higher mean urinary penta (P < 0.006) and copro (P < 0.006) concentrations compared with same-aged NT children, each characterized by a number of extreme values. Using Receiver Operating Characteristic curve analysis, we evaluated the sensitivity and specificity of penta, copro, and their combined Z-scores in ASD detection. The penta sensitivity was 30% for AUT and 36% for PDD-NOS, with 94% specificity. The copro sensitivity was 33% and 14%, respectively, with 94% specificity. The combined Z-score measure had 33% and 21% sensitivity for AUT and PDD-NOS, respectively, with 100% specificity. These findings demonstrate that porphyrin measures are strong predictors of both AUT and PDD-NOS, and support the potential clinical utility of urinary porphyrin measures for identifying a subgroup of ASD subjects in whom disordered porphyrin metabolism may be a salient characteristic. Autism Res2012,••: ••–••. © 2012 International Society for Autism Research, Wiley Periodicals, Inc.