ABSTRACT: A need exists to expand the characterization of tetrahydrobiopterin (BH(4)) responsiveness in patients with phenylketonuria (PKU), beyond simply evaluating change in blood phenylalanine concentrations. The clinical interpretation of BH(4) responsiveness should be evaluated within the context of phenylalanine hydroxylase (PAH) genotype.This investigation seeks to use a modified version of a previously developed PAH genotype severity tool, the assigned value (AV) sum, to assess the molecular basis of responsiveness in a clinical cohort and to explore the tool's ability to differentiate BH(4) responsive groups.BH(4) response was previously clinically classified in 58 patients with PKU, with three response groups emerging: definitive responders, provisional responders, and non-responders. Provisional responders represented a clinically ambiguous group, with an initial decrease in plasma phenylalanine concentrations, but limited ability to improve dietary phenylalanine tolerance. In this retrospective analysis, mutations in the PAH gene were identified in each patient. PAH genotype was characterized through the AV sum approach, in which each mutation is given an AV of 1, 2, 4, or 8; the sum of both mutations' AV corresponds to genotype severity, with a lower number representing a more severe phenotype. An AV sum cutoff of 2 (indicative of the most severe genotypes) was used to dichotomize patients and predict BH(4) responsiveness. Provisional responders were classified with the definitive responders then the non-responders to see with which group they best aligned.In 17/19 definitive responders, at least one mutation was mild or moderate in severity (AV sum>2). In contrast, 7/9 provisional responders carried two severe or null mutations (AV sum=2), suggesting little molecular basis for responsiveness. Non-responders represent a heterogeneous group with 15/25 patients carrying two severe mutations (AV sum=2), 5/25 patients carrying one moderate or mild mutation in combination with a severe or null mutation (AV sum>2), and the remaining five patients carrying an uncharacterized mutation in combination with a severe mutation. Predictive sensitivity of the AV sum was maximized (89.5% vs. 67.9%) with limited detriment to specificity (79.4% vs. 80.0%), by classifying provisional responders with the non-responders rather than with the definitive responders.In our clinical cohort, the AV sum tool was able to identify definitive responders with a high degree of sensitivity. As demonstrated by both the provisional responder group and the substantial number of non-responders with AV sums>2, a potential exists for misclassification when BH(4) response is determined by relying solely on change in plasma phenylalanine concentrations. PAH genotype should be incorporated in the clinical evaluation of BH(4) responsiveness.
