Abstract
Introduction
Newborn screening for phenylketonuria (PKU) can reveal other conditions which lead to an increased blood spot phenylalanine (Phe) concentration. We have investigated the proportion of blood spot samples that gave a positive screen due to clinically significant conditions other than PKU, compared the positive predictive value (PPV) of our referral Phe cut-off with that recommended by the UK Newborn Screening Programme Centre (UKNSPC) (>210 and >240 μmol/L, respectively) and evaluated the effectiveness of reflex testing for galactosaemia using a lower blood spot Phe cut-off concentration of 130 μmol/L.
Methods
All blood spot samples that screened positive, for an increased Phe concentration, between April 2001 and March 2008, were identified from the records of the Sheffield Newborn Screening Laboratory and the diagnoses noted. In addition, all cases of galactosaemia detected in or notified to our screening laboratory within this time were also examined and the screened Phe concentrations compared.
Results
Out of 438,674 babies who were screened, 67 had Phe concentration >210 μmol/L (15 per 100,000). Of these, 40 had PKU or persistent hyperphenylalaninaemia with a Phe concentration identified by screening between 270 and 2350 μmol/L. A further 11 were diagnosed with another clinically significant disorder: galactosaemia (n = 8), biopterin defects (n = 2), tyrosinaemia Type 1 (n = 1). In addition, 16 had transient elevations in Phe. In total, nine cases of galactosaemia were identified, of whom, three had Phe concentrations <240 μmol/L with one asymptomatic individual having a concentration <210 μmol/L.
Conclusions
Adoption of the UKNSPC recommended cut-off (>240 μmol/L) will not affect the detection rate of classical PKU, but will improve the PPV from 76% to 80%. The use of a lower cut-off (130 μmol/L) for reflex galactosaemia testing enables the timely identification of asymptomatic cases that benefit particularly from early treatment, without prompting any unnecessary clinical referrals or delaying any referrals. This intervention may reduce mortality in this vulnerable group.
Introduction
Infants born in the UK are screened for phenylketonuria (PKU) between days five and eight by measuring phenylalanine (Phe) in dried blood spot samples using tandem mass spectrometry (MS/MS). However, several significant clinical conditions other than PKU may also lead to an elevated Phe concentration either as a primary effect, such as in biopterin defects or secondary to liver dysfunction, such as in galactosaemia. UK Newborn Screening Programme Centre (UKNSPC) national guidelines 1 currently recommend exclusion of such disorders after clinical referral of a positive screen. However, this department has, for many years, included reflex testing for galactosaemia, using the original blood spot sample prior to referral, in order to facilitate the early detection of this clinically urgent condition. 2 The implementation of this practice nationally is currently being considered under a review by the UKNSPC.
The National Guidelines for PKU screening 1 recommend a Phe concentration of 240 μmol/L as the cut-off, and individuals with values above this should be referred. Our local practice is to use a Phe concentration >210 μmol/L with an additional lower cut-off (Phe >130 μmol/L) for galactosaemia testing using thin layer sugar chromatography (TLC). 2 This method is a rapid, cost-effective, screening test for galactosaemia and is more robust than methods that rely on galactose-1-phosphate uridyl transferase detection.
We set out to examine the final diagnosis of those infants with an elevated Phe following newborn screening, who did not have PKU and determine the proportion of these with clinically significant conditions. We also aimed to compare the positive predictive value (PPV) of our referral cut-off for PKU (Phe concentration >210 μmol/L) with the nationally recommended one (Phe concentration >240 μmol/L), and evaluate the effectiveness of reflex testing for galactosaemia (currently being considered by the UKNSPC), using a lower Phe cut-off of 130 μmol/L.
Methods
The Sheffield Regional Newborn Screening Laboratory currently screens approximately 75,000 babies annually. The laboratory computer system, laboratory record books and patient referral records were searched. This included all records produced since the introduction of MS/MS technology and covered a seven-year period from April 2001 to March 2008. All individuals with elevated Phe concentrations (>210 μmol/L) were identified and causes noted. In addition, all cases of galactosaemia detected by or notified to the laboratory during this time were collated and the Phe concentrations at screening were compared.
Results
Outcome of cases with Phe concentration >210 μ mol/L
Out of 438,674 babies screened, 67 had a Phe concentration >210 μmol/L (15 per 100,000 screened). The outcomes are summarized in Table 1.
Summary of outcomes of cases with Phe >210 μmol/L
*PKU is defined in this report as those babies with a persistent Phe >600 μmol/L in whom dietary treatment is initiated
PKU, phenylketonuria; Phe, phenylalanine
†Hyperphenylalaninaemia is defined as those with a persistently raised Phe (>240 μmol/L) but remaining below the concentration at which treatment is initiated (600 μmol/L) on follow up
‡Based on dialogue with the referral clinician or diagnostic laboratory
Phe screening values in galactosaemia cases
Sugar TLC was carried out on 445 samples with Phe >130 μmol/L during the period studied. Nine samples (2%) tested positive for increased galactose due to galactosaemia. No secondary causes of increased galactose were reported during the period studied. The reported screening Phe concentration and the clinical condition at the time of reporting for the galactosaemia cases are summarized in Table 2. None of the galactosaemia cases had a screening Phe concentration below 130 μmol/L.
Screening Phe in galactosaemia cases
Phe, phenylalanine
*Indicates additional cases below national guidance cut-off (240 μmol/L)
†Indicates case below our in-house screen positive cut-off (210 μmol/L)
Discussion
Significance of Phe concentration >210 μ mol/L
A raised Phe concentration (>210 μmol/L) identified through newborn screening is most commonly due to classical PKU and hyperphenylalaninaemia (59.7%). However, 16.4% of elevations are due to other clinically significant conditions (biopterin defects, galactosaemia and tyrosinaemia type 1). The remaining 23.9% are not thought to be clinically significant, showing only a transient increase in Phe at the time of screening. This gives a PPV for the presence of any clinically significant condition (including classical PKU) of 76% using this cut-off. Had the recommended cut-off of 240 μmol/L been used to trigger any further investigations, three of the cases that we followed up would have been missed (two with galactosaemia and one sick baby who died before a repeat sample for Phe estimation could be collected, but which was thought to be a transient elevation). However, the PPV for any clinically significant condition would have increased from 76% to 80% overall, due to the exclusion of four transient/non-significant causes of the initial raised Phe value. None of the elevations between 210 and 240 μmol/L were due to PKU, so increasing the cut-off to 240 μmol/L would improve the overall PPV of the screening test but would not affect the detection rate for classical PKU.
Value of tiered lower cut-off to improve galactosaemia detection
Nine cases of galactosaemia out of the 445 samples with a screening Phe concentration >130 μmol/L were detected over the period of this study. Since this reflex test is undertaken at the same time as confirmatory tests on samples with an elevated Phe concentration, there was no delay in onward referral. Over the period 2005–2010, 53 referrals, identified through PKU screening have been made, 100% within two working days and 94% within one working day of sample receipt. This far exceeds the UKNSPC core standard to refer 100% cases within four working days and even its development standard of referral within three working days of sample receipt.
Adhering to the national guideline cut-off of 240 μmol/L as an indication for further investigation, would have led to three cases of galactosaemia, two of which were asymptomatic, being undetected by the screening process over the period of the study.
It is important to diagnose galactosaemia in the neonatal period since early detection can, on occasions, prevent death. A report from the British Paediatric Surveillance Unit suggested that newborn screening for galactosaemia may not be necessary as clinical vigilance provided a diagnosis in a sufficient time frame. 3 However, three of the cases included in that report were only suspected due to an elevated Phe concentration being obtained at screening. In addition, four cases were diagnosed relatively late on and so experienced a prolonged ‘diagnostic journey’ that could have been prevented had screening been undertaken. The report did conclude however that further investigation into reflex testing for galactosaemia as part of PKU screening may be warranted.
In conclusion, our experience shows that reflex testing for galactosaemia, undertaken as part of PKU screening before clinical referral, does not result in any additional unnecessary referrals or affect the timeliness of referral of cases with a Phe concentration above the cut-off. The use of an additional lower tier cut-off Phe concentration (130 μmol/L) for reflex galactosaemia testing improved the overall detection of galactosaemia cases from 67% to 100% in the population studied and reduced the diagnostic delay in two pre-symptomatic cases.
In view of these findings, we have adopted the cut-off for referring cases with a raised screening Phe of 240 μmol/L in line with the UKNSPC recommendation, but retained our lower tier trigger Phe concentration of 130 μmol/L for reflex galactosaemia testing.
DECLARATIONS