Does the 48-hour BH4 loading test miss responsive PKU patients?

ABSTRACT

BACKGROUND: Phenylketonuria (PKU) is an inborn error of phenylalanine (Phe) metabolism. Besides dietary treatment, some patients are responsive to and treated with tetrahydrobiopterin (BH4). Our primary objective was to examine whether the 48-hour BH4 loading test misses BH4-responsive PKU patients. Secondary, we assessed if it would be beneficial to 1) use a cut-off value of 20% Phe reduction instead of commonly used 30%, and 2) extend the loading test to 7 days. METHODS: 24 patients with a 20-30% decrease of blood Phe levels during their initial 48-hour BH4 loading test or at least one mutation associated with long-term BH4 responsiveness, were invited to participate. 22 of them underwent the 7-day BH4 loading test. During the BH4 loading test, BH4 was administered orally once daily for 7 days (20 mg/kg/day). Blood samples on filter paper were collected at 13 time points. Potential BH4 responders (≥20% decrease in blood Phe concentrations at ≥1 moment within the first 48 h or ≥30% at ≥1 moment during the entire test) underwent a treatment trial to assess true long-term responsiveness (≥30% decrease of Phe levels compared to baseline and/or ≥50% increase in natural protein tolerance in accordance with the Dutch guidelines before 2017). The duration of the treatment trial varied from 2 to 18 months. RESULTS: Of the 22 patients who completed the 7-day BH4 loading test, 2 were excluded, 8 had negative tests and 12 were considered to be potential BH4 responders. Of these 12 potential BH4-responsive PKU patients, 5 turned out to be false positive, 6 true-responder and 1 was withdrawn. CONCLUSION: Even though the 48-hour BH4 loading test has proven its efficacy in the past, a full week may be necessary to detect all responders. So, if blood Phe concentrations during the 48-hour BH4 test shows a clear tendency, but not sufficient decrease, a full week (with only measurements each 24 h) could be offered. A threshold of ≥20% decrease within 48 h is not useful for predicting true BH4 responsiveness.