Determination of Total Galactose from Dried Blood Spots — Extensive Assay Evaluation of a CE-Marked Test-Kit

Most newborn screening laboratories use CE-marked or FDA-approved test-kits, like in routine clinical chemistry. National regulations require only minimal evaluation from the customer, if the test-kits are used as specified by the manufacturer. The microtiter-based kit-concept is often based on the perception, that the laboratory always processes whole microtiter plates. However, in the daily routine, this is rather a rare exception, which leads to much higher costs per newborn, compared to the costs per assay in the test-kits. In addition, the amount of wasted resources is quite high. Performance of the Neonatal Total Galactose kit from Perkin Elmer was tested. We have determined specificity, limit of detection (LOD), limit of quantitation (LOQ), intra and inter assay variation, recovery, stability of measuring signal and reagents. Results were also compared with the Astoria Pacific Spot Check System. In addition, we had (by chance) the opportunity to test 2 kits, which were already expired for more than 3 years. LOD was 165 306 μmol/L and LOQ 475 703 μmol/L, depending on the definition of LOD/LOQ. Mean recovery was 112.8%, intra assay CVs were 11.3, 7.3, 4.0, and 3.0, and inter assay CVs 28.7, 15.9, 7.8, and 9.3, at 220, 590, 1200, and 2060 μmol/L respectively. Reconstituted and mixed reagents must be used within some hours, and were unstable even if stored at −20 ̊C. However, if the reconstituted galactose substrate reagent and galactose oxidase reagent were only mixed according to the daily requirements, and the rest stored separately at −20 ̊C, they were stable for at least 12 days. The performance of the expired test-kits did not differ from the others. The performance of the Total Galactose kit is comparable to other tests used for newborn screening. However, we could significantly reduce the costs per newborn and reduce unnecessary production of waste, by thorough validation and modification of the assay procedures.


Introduction
Newborn Screening for Galactosemia was introduced in the 1970s [1].While the classical "Beutler-Test" can only detect newborns with classical galactosaemie [2], the determination of total galactose [3,4] will also detect newborns with galactokinase deficiency and uridine diphosphate galactose-4-epimerase deficiency.
Most newborn screening laboratories use CE-marked or FDA-approved test-kits, like in routine clinical chemistry.National regulations require only minimal evaluation from the customer if the test-kits are used as specified by the manufacturer.The microtiter-based kit-concept is often based on the perception, that the laboratory always processes whole microtiter plates.However, in the daily routine of newborn screening laboratories, this is rather a rare exception, which leads to much higher costs per newborn, compared to the costs per assay in the test-kits.In addition, the amount of wasted resources is quite high.We have thoroughly tested the performance of the Neonatal Total Galactose kit from Perkin Elmer.We have determined specificity, LOD, LOQ, intra-and inter-assay variation, recovery, signal stability of the fluorescent reaction product, and stability of reagents.Results were also compared with the Astoria Pacific Spot Check System.In addition, we had (by chance) the opportunity to test 2 kits, which were already expired for more than 3 years.

Materials and Methods
Total Galactose was determined from whole blood, dried on filterpaper, with the Neonatal Total Galactose test-kit from Perkin Elmer.For the determination of intra and inter assay variation various kit controls were used (concentrations up to 726 µmol/L).In addition, dried blood samples with galactose concentrations above 1000 µmol/L were prepared by adding galactose to whole blood of a healthy adult volunteer.After thorough mixing, the blood was spotted onto filterpaper and let dry for several hours.Control samples were stored at −18˚C.The Galactose-Oxidase/Substrate-Reagent has to be mixed from reconstituted Galactose-Oxidase-Reagent and Galactose-Substrate-Reagent.One vial of lyophilized Galactose-Oxidase-Reagent has to be dissolved with 6 mL of reconstitution buffer and one vial of Galactose-Substrate-Reagent has to be dissolved with 2 mL of reconstitution buffer.This results in 8 mL of Galactose-Oxidase/Substrate-Reagent which is sufficient for one microtiterplate.Reconstituted and mixed reagents must be used within one hour, and were unstable even if stored at −20˚C.Due to the instability of reconstituted Galactose-Oxidase/ Substrate-Reagent, we used a modified assay procedure.After reconstitution of Galactose-Oxidase-Reagent and Galactose-Substrate-Reagent only the necessary amount of Galactose-Oxidase/Substrate-Reagent is mixed.A whole vial of Galactose-Oxidase-Reagent and Galactose-Substrate-Reagent for a whole microtiterplate, and x * 750 µL of reconstituted Galactose-Oxidase-Reagent and x * 375 µL of reconstituted Galactose-Substrate-Reagent.With x being the number of rows on the microtiterplate that contain samples.Then all mixed Galactose-Oxidase/Substrate-Reagent is pooled for the whole daily batch.With this procedure one calibration curve on the first plate is sufficient for the daily batch, in contrast to the kit instructions, which call for a calibration curve on each plate.Residual reagents can be stored at −20˚C, and will be used the next working day.
The performance of the PE-kit for routine newborn screening was compared with the results of the Astoria Pacific Spot Check Kit.

Results
When the Total Galactose Kit was used according to our modified assay procedure, reagents are stable for at least 12 days (Figure 1).The fluorescent signal of the reaction product, after stopping the reaction, was stable for more than 3 hours (Figure 2).
Values for LOD, LOQ are shown in Table 1, and values for intra-and inter-assay-variation are shown in Table 2, at various concentration levels.Mean recovery was 112.8%.The correct determination of LOD and LOQ is difficult.Whole blood samples without endogenous galactose are hard to obtain and the measurement of plain white filterpaper gives significantly higher values than kit-calibrator A (n = 30, p < 0.001).This is also reflected by the very high values of the statistically achieved LOD 4 , LOQ 1 , and LOQ 2 .
Our data show, that the use of the unweighted linear regression algorithm, as proposed by the manufacturer, results in an underestimation of the galactose concentrations above 1000 µmol/L (Figure 3).The use of a spline function for calibration curves gives much more reliable results.
Comparison of the Total Galactose kit from PE with the Astoria Pacific Spot Check Kit under routine conditions showed no perfect correlation.The best curve fit was achieved with a quadric equation (Figure 4).
The performance of the test-kits, that were expired for more than 3 years, showed no difference to still valid test  LOB 1 = limit of blanc; defined as the 95th centile of a distribution of blank samples.LOD 2 = limit of detection; according to NCCLS document EP17-A.LOQ 3 = limit of quantitation; defined as the lowest concentration with a total CV equal to or less than 20%.LOD 4 = limit of detection; defined as the mean of a sample without analyte + 3 SD (blanc filterpaper).LOQ 1 = limit of quantitation; defined as the mean of a sample without analyte + 6 SD (blanc filterpaper).LOQ 2 = limit of quantitation; defined as the mean of a sample without analyte + 10 SD (blanc filterpaper).LOQ 3 = limit of quantitation; defined as the lowest concentration with a total CV equal to or less than 20%.

Discussion
Newborn screening laboratories have to measure all screening samples on the day of arrival in the laboratory.Therefore it is extremely rare that the total sample number of a daily routine, including calibrators and controls, will exactly (or nearly) fill a whole microtiterplate, or two, three, etc.But since the reagents in the Total Galactose kit from Perkin Elmer are packed for one plate, respectively, sticking to the procedures outlined in the kit-insert will result in a significant waste of reagents, and hence increase of costs per newborn.This can be up to 50% subject to the number of samples per day.In addition, the stated expiry date seems to be unnecessarily short, which could also lead to unnecessary waste of reagents in small screening laboratories.
Our data show that a thorough in house validation even of CE-marked test-kits can significantly improve the overall performance of a test, including economical aspects.

Table 2 . Intra-and inter-assay variation at various galactose concentrations.
galactose with the PE-kit is not influenced by EDTA up to 100 mM.Mean value of 18 control samples with EDTA concentrations from 0 -100 mM: 719 µmol/L, c.v. 9.8%.
*Determination of total kits (data not shown).