Soc. for Advancement of Quality Assurance in Medical Laboratories,Germany
Title: Neurochemical dementia diagnostics - Interlaboratory variation of analysis, reference ranges and interpretations in the European CSF Survey of INSTAND
Hansotto Reiber is Consulting of Neurolife (CSF laboratoy) in Rio de Janeiro, theory of blood-CSF barrier function and CSF flow. Dynamics of brain and blood-derived proteins in CSF. Neuroimmunology, and basic research in chronic neurological diseases.Particular research in CSF analysis in tropical Neurology. Basic research on nonlinear dynamics of biological processes, self organization concepts in biological, medical and social sciences. (e.g., immunological network and Neurosciences). Analysis of aqueous humor for diagnosis of eye diseases.
In the external quality assessment system for dementia marker analysis in Cerebrospinal fluid (EQAS) of INSTAND, Germany, participate 49 laboratories from nine European countries and USA. Stabilized CSF samples with physiologically relevant data pattern are distributed to allow, together with the numerical data, a reference range - related evaluation and a differential diagnostic interpretation of combined parameters (total Tau protein, hyperphosphorylated Tau protein and Amyloid-ß-peptide Aß42). Results: The inter-laboratory variation for the three parameters from N= 6 surveys are shown to be a problem of the analytical methods: A 10-fold variation between lower and upper concentrations of the values for Aß42 (69-771 pg/ml) with up to 31% outliers or a 4-fold variation for Tau (315-1292 pg/ml) with 13-15% outliers needs urgent reaction of the supplier. Only pTau results with less then twofold variation (53-83 pg/ml, Table 1) and only 3-11% outliers corresponding to success quota of 89-97% were of acceptable performance. For the evaluation of these numerical data (normal /pathological/border line) the participants used own cut off values with a huge variation: Tau (150-540, median 450 pg/ml), pTau (35-85, median 61 pg/ml) and Aß1-42 (205-600, median 500 pg/ml). These data represent the dilemma of this analytical approach as shown in the examples: In case of a sample with normal median values (e.g. Tau = 381 pg/ml and Aß= 748 pg/ml) 45% of participants regarded the value as pathological with a stunning interpretation of the combined data of Tau and Aß1-42: 29% of the participants found this data combination compatible with an Alzheimer’s disease, definitely a false pathological result. 29% report this as a normal sample, and still 42 % regard an integrating interpretation as not possible. In case of pathological values (921/66/508 pg/ml) 30% of participants evaluated Aß1-42 data as normal, 30% as decreased and 40% as borderline. In spite of this ambivalence in reference range-related evaluation 74% of the participants regarded their results compatible with the diagnosis of AD. Conclusion: Up to 31% outliers are a source of serious diagnostic errors. The unacceptable large variation of the reference ranges (cut off values) used in daily practice with false negative and false positive diagnostic interpretations question the practical relevance of statistically founded differences between different dementive diseases as reported in the literature. Improvement of the assay performance with reliable control material for interlaboratory compatibility and a knowledge based consensus about cut off values are the precondition to avoid the confusing interpretations of combined biomarker data. The calculation of mathematical formulas or ratios are not improving the discriminative sensitivity due to the error propagation in mathematical functions with uncoupled analytical data.