[The ways of harmonization of clinical laboratory measurement techniques].

The results of implementation of different clinical laboratory techniques are to be equal in clinically significant limits to be optimally applied in diagnostics of diseases and treatment of patients. When the results of laboratory tests are not standardized and harmonized for the very same clinical assay the results can be expressed by unmatched numbers. Unfortunately, in some handbooks the values are presented based on the results of application of specific laboratory techniques without considering possibility or likelihood of differences between various techniques. When this is a case, accumulation of data of diferent clinical research studies and working out of clinical handbooks on this basis will be inconsistent. Inadequate understanding of issue that the results of laboratory tests are not standardized and harmonized can lead to incorrect clinical, financial, managerial or technical decisions. The standardization of clinical laboratory techniques was applied to many measurands related to primary referent techniques (standard specimen of pure substance) or/and developed referent measurement techniques. However, harmonization of clinical laboratory techniques for those measurands which are not related any developed measurement techniques is quite problematic due to inadequate determination of measurand, its inadequate analytical specificity, insufficient attention to commutability of referent materials and poor systematic approach to harmonization. To overcome these issues an infrastructure is to be developed to support systematic approach to identification and prioritization of measurands which are to be harmonized on the basis of clinical importance and technical applicability. The management of technical implementation harmonization process for specific measurands.

Validation of 5 routine assays for serum free testosterone with a candidate reference measurement procedure based on ultrafiltration and isotope dilution-gas chromatography-mass spectrometry.

BACKGROUND/METHOD: The analytical validity of free testosterone (FTe) analog immunoassays is subject to much controversy. We revisited the validation of 4 analog assays and 1 FTe calculation procedure with a metrologically traceable reference measurement procedure (RMP) based on ultrafiltration and isotope dilution-mass spectrometry for direct measurement of Te in the ultrafiltrate. To this end, we performed split-sample measurements of 40 male sera. RESULTS: Deming regression showed that 3 of the immunoassays had moderate to good correlation (0.8474 < or = r < or = 0.9241) with the RMP; however, the slope was markedly below 1. The FTe calculation procedure was in good agreement with this result. The Sy/x values for all assays were higher than the combined imprecision values, which indicate their susceptibility to matrix-related effects. CONCLUSIONS: The study demonstrated substantial differences in analytical quality of FTe assays; however, the results suggested that after extending the validation with a larger variety of samples, recalibration of some analog assays might be worth further investigation.

Trueness verification in European external quality assessment schemes: time to care about the quality of the samples.

An external quality assessment (EQA) survey on 14 fresh-frozen, single-donation sera assigned with reference measurement procedure (RMP) values revealed a mean bias of + 5.2% and + 3.7% for the cholesterol oxidase and the photometric glucose oxidase procedure groups, respectively. Conversely, on lyophilized sera, the same procedure groups showed almost bias-free results, the differences from the RMP values being only -0.8% for cholesterol and + 0.7% for glucose. These data, which are in fairly good agreement with the literature, suggest the existence of artificial matrix effects in processed materials. Therefore they indicate that, currently, assessment of trueness is hampered in many European EQA schemes, as most of them use lyophilized sera. This approach may give a false impression about the trueness of laboratory results as well as carrying the risk that laboratories calibrated on the RMP values of the survey samples could make errors in patient testing. Consequently, if European EQA is willing to fulfil a post-market vigilance function of the performance of in vitro diagnostic medical devices, then the time has come to tackle the problem of the quality of the survey samples. EQA organizers urgently need to make an effort to seek out materials that analytically behave like authentic clinical specimens. In the meantime, alternative approaches should be used. Although not ideal, the special survey described in this article is one of the possibilities. Naturally, it implies logistic problems and increased costs for the individual EQA schemes. However, both can be overcome with the cooperation of the predominantly nationally organized schemes.

The ionized fraction of serum total magnesium in hemodialysis patients: is it really lower than in healthy subjects?

AIM OF THE STUDY: Based on data in the literature, it remains unclear whether the ionized fraction of serum total magnesium (Mg) is lower in chronic hemodialysis (HD) patients compared to healthy subjects. PATIENTS AND METHODS: The ionized fraction of serum total Mg was investigated in 49 HD patients, pre- and post-dialysis, and compared to 30 healthy controls. The quality of the analytical performance of the Mg measurements has been emphasized by applying a reference method and/or rigorous internal quality control (IQC). In addition, the ionized fraction of serum total calcium (Ca) was measured in both populations, because the results for Mg should be related to those of Ca. RESULTS: In HD patients, the ionized fraction of serum total Mg was on average 65% (pre-dialysis 64.2% and post-dialysis 66.2%). In healthy controls, the ionized fraction was 64.9%. When the analytical variability was taken into account, no significant differences (p > 0.05) were observed between pre- and post-dialysis samples and controls. For Ca, an ionized fraction of 55.3% was found in HD patients, which was not significantly different from the fraction obtained in the control group (55.7%). CONCLUSION: The present study demonstrates that, compared to healthy controls, the ionized fraction of serum total Mg is not different in hemodialysis patients.

Strategies for determination of insulin with tandem electrospray mass spectrometry: implications for other analyte proteins?

Using human insulin (MW 5808 Da) as a model compound, the possible strategies towards optimization of sensitivity and selectivity of measurement by electrospray ionization with a standard triple quadrupole mass spectrometer were investigated. For measurement in selected ion-monitoring (SIM) mode, these strategies involved systematic variation of instrumental parameters and spray pH. In this investigation four different operating modes were used corresponding to positive/negative ionization modes with acidic/basic sprays and pH reversed (hereafter termed 'wrong-way-round' operation); the cone voltage was optimized for each mode of operation. When collision-activated dissociation (CAD) is employed, two additional operation modes are possible: namely, low collision energies (10-35 eV, CAD-l) for the generation of sequence-specific fragments and high collision energies (>80 eV, CAD-h) for the generation of nonspecific fragments. Overall, this results in twelve different modes of operation. Loop-injection of aqueous insulin standards were run for each of the twelve operating modes and measurements made for five different charge states (n = 2-6) observable with our instrument that has an upper mass limit of m/z 4000. The signal/noise (S/N) ratio was optimized for each charge state, resulting in 60 measurements. The best S/N ratios (20 000) were achieved under positive SIM conditions with charge state 6 (m/z 969) and under 'wrong-way-round' negative SIM conditions with charge state 3 (m/z 1935). Lower S/N ratios were observed under positive CAD-h conditions with charge state 5 (m/z 1163, S/N 15 000) and positive CAD-l conditions with charge state 6 (m/z 969, S/N 10 000). All other operating modes gave maximum S/N ratios of 4000. For measurement of insulin standards, the results obtained show SIM to give the best S/N ratio. However, for samples in complex matrices, our general experience suggests CAD to be the preferable operating mode. Consequently, for the development of a quantitative method for proteins in general, it might be advocated that all of the twelve operating modes and all relevant charge states be investigated to find the optimum S/N ratio.

Quantitative analysis of urinary C-peptide by liquid chromatography-tandem mass spectrometry with a stable isotopically labelled internal standard.

We describe the first results of a quantitative LC-tandem mass spectrometry method for urinary C-peptide with the use of [2H14]C-peptide as internal standard. LC was based on gradient elution of a Hypersil PEP C18 column. Mass spectrometry was performed in the negative electrospray ionization mode and by monitoring of the transitions at m/z 1514/1334 ([2H14]C-peptide) and 1507/1320 (C-peptide). For sample preparation, we applied ultrafiltration. The analytical performance of the method in terms of measurement precision gave an RSD of <2% (n=10). The overall imprecision was investigated from independent analysis of two urine samples in six-fold and resulted in an RSD<5%. The limit of detection, expressed as signal-to-noise ratio 3, was approximately 0.15 ng C-peptide injected. Analysis of 10 random urine samples from laboratory volunteers showed interference-free ion chromatograms at a signal-to-noise ratio of approximately 75 on average. The C-peptide concentrations calculated from quantification by the bracketing calibration technique ranged from 32 to 165 ng/ml.

Evaluation of intrinsic and routine quality of serum total magnesium measurement.

We investigated the intrinsic (as delivered by the manufacturer) and routine quality of four systems for measurement of serum total magnesium (t-Mg(2+)) by method comparison with an ion chromatography reference method. The results of the study were interpreted on the basis of analytical quality specifications derived from the biological variation of t-Mg(2+), expanded by the analytical uncertainty of the reference measurements. This resulted in limits for systematic error of 2.1% and for total error of 4.3%. The study demonstrated that those limits were challenging for all routine systems. Most of them met the total error criterium just borderline and one showed a considerable systematic error (-5.2%). Concerning the measurement quality in the routine laboratories, the study showed that many were unable to preserve the intrinsic quality of the respective manufacturer. Consequently, loss of system performance in the routine laboratory mostly led to violation of the analytical specifications. Most strikingly, the study revealed enormous quality differences between routine laboratories. This indicates that, still, many routine laboratories do not make adequate use of currently available internal and external quality control tools. Moreover, some laboratories considerably expanded the high end of the reference interval, thereby reducing the diagnostic potential of t-Mg(2+).

pH dependency of serum ionised calcium.

Measurement by ion selective electrode showed that the pH dependency of serum ionised calcium is better described by an inversely S-shaped third-degree function than by the conventionally used logarithmic function.

Isotope dilution-gas chromatography/mass spectrometry and liquid chromatography/electrospray ionization-tandem mass spectrometry for the determination of triiodo-L-thyronine in serum.

Isotope dilution-gas chromatography/mass spectrometry (ID-GC/MS) and isotope dilution-liquid chromatography/tandem mass spectrometry (ID-LC/MS/MS) methods have been developed for an determination of triiodothyronine (T3) in serum and their potential as candidate reference methods investigated. In both methods, (13)C(9)-T3 was used as internal standard. Sample pretreatment consisted of deproteinization, extraction and high performance liquid chromatography (HPLC) purification. Conversion of serum thyroxine (T4) to T3 was controlled by adding (13)C(6)-T4. For GC/MS, T3 and (13)C(9)-T3 were converted to the N,O-di-heptafluorobutyryl (HFB) methyl ester derivatives and monitored at m/z 844 and 853. For LC/MS with electrospray ionization, the transitions m/z 652/661 to 606/614 were monitored. For use of the methods as candidate reference methods, special attention was paid to the calibration and the measurement protocol (duplicate analysis of each sample on three occasions). Evaluation of the ID-GC/MS and ID-LC/MS/MS methods showed the absence of interference by reverse T3 and T4, a limit of detection of 100 pg (GC/MS) and 18 pg (LC/MS), a recovery of 100 +/- 1.5% (95% confidence interval) and good precision (the total coefficient of variation, n = 6, was typically 1.5%). In addition to the recovery study, the accuracy of the methods was proven by method comparison (ID-GC/MS vs. LC/MS/MS) on three sera, showing a maximum deviation of 1.1%. Finally, the ID-GC/MS method was applied for measurement of 10 different human sera with a T3 concentration range from 0.6 to 7.3 ng/mL.

Survey of serum potassium reference measurements.

We compared the quality of reference measurements for serum potassium in four reference laboratories from three different European countries, using a panel of 60 native patients' samples. The reference methods were based on either ion chromatography (one laboratory) or flame atomic emission spectrometry (three laboratories). Performance specifications for serum potassium measurements were defined as a maximum overall coefficient of variation (CV) of 1.5%, a maximum bias of 0.65% and a maximum total error of 3.0%. The overall imprecision for all laboratories was in the range of 0.7 to 1.3%, and was thus below the proposed specification of 1.5%. However, two laboratories reported 12 and 13 quadruplicates with CVs exceeding this limit. The mean bias (expressed as deviation from the overall mean of all laboratories) for all reference laboratories was < 0.65%. In the lower concentration range, however, one laboratory exceeded this limit. No laboratory measured samples with a total error above 3.0%. From these results, it can be concluded that the reference measurements, and, thus, also the reference methodologies, based on ion chromatography and flame atomic emission spectrometry were equivalent, and able to satisfy current analytical specifications for serum potassium measurements.