Article

Specifications for trueness and precision of a reference measurement system for serum/plasma 25-hydroxyvitamin D analysis

Laboratory for Analytical Chemistry, Faculty of Pharmaceutical Sciences, Ghent University, Harelbekestraat 72, B-9000 Ghent, Belgium.
Clinica chimica acta; international journal of clinical chemistry (Impact Factor: 2.76). 07/2009; 408(1-2):8-13. DOI: 10.1016/j.cca.2009.06.027
Source: PubMed

ABSTRACT The divergence in analytical quality of serum/plasma 25-hydroxy-vitamin D analysis calls for defining specifications for a reference measurement system.
Fundamentally, in a reference measurement system, there should be a relationship between the analytical specifications for higher- (reference) and lower-order (routine) measurements. Therefore, when setting specifications, we started with limits for routine imprecision (CV(rou)) and bias (B(rou)) using 4 models: (1) the misclassifications in diagnosis, (2) biological variation data (reference interval (RI) and monitoring), (3) expert recommendations, and (4) state-of-the-art performance. Then, we used the derived goals to tailor those for reference measurements and certified reference materials (CRMs) for calibration by setting the limits for CV(ref) at 0.5 CV(rou), B(ref) at 0.33 B(rou)(,) max. uncertainty (U(max)) at 0.33 B(ref).
The established specifications ranged between CV(rou)<or=22%, B(rou)<or=10%, CV(ref)<or=11%, B(ref)<or=3.3%, U(max) 1.1% (model 3) and CV(rou)<or=4%, B(rou)<or=2.6%, CV(ref)<or=2%, B(ref)<or=0.9%, U(max) 0.3% (model 2, monitoring).
Model 2 (monitoring) gave the most stringent goals, model 3, the most liberal ones. Accounting for state-of-the-art performance and certification capabilities, we used model 2 (RI) to recommend achievable goals: for routine testing, CV(rou)<or=10%, B(rou)<or=5%, for reference measurements, CV(ref)<or=5%, B(ref)<or=1.7%, and for CRMs, U(max) 0.6%.

1 Follower
 · 
89 Views
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: OBJECTIVES: To evaluate the newly developed Roche MODULAR Analytics E170 Total Vitamin D and the Siemens ADVIA Centaur® Vitamin D Total assays. MATERIALS AND METHODS: Assays were evaluated using the Clinical and Laboratory Standards Institute protocols. Split patient samples were compared with LC-MS/MS and DiaSorin LIAISON assays (n=79 including 15 specimens with detectable endogenous 25-OH vitamin D(2)). Assay accuracy was also evaluated using the Vitamin D External Quality Assessment Scheme (DEQAS) samples. RESULTS: The ADVIA Centaur and E170 assays demonstrated maximum total CVs of 14.1% and 5.9%, respectively. Both showed excellent linearity (R(2)>0.99). The ADVIA Centaur assay demonstrated interference with bilirubin at 800μmol/L, hemolysis at 1.25g/L, and triglycerides at 2.8mmol/L. Compared to LC-MS/MS, the ADVIA Centaur assay demonstrated a R(2) value of 0.893, average bias of -8.8%; the E170 assay an R(2) value of 0.872, average bias of 14.3% with underestimation of 25-OH vitamin D(2). Compared to the LIAISON assay, the ADVIA Centaur assay demonstrated an R(2) value of 0.781, average bias of -17.3%; the E170 assay an R(2) value of 0.823, average bias of 11.4%. The ADVIA Centaur and E170 assays demonstrated a biases of <20% in 10/10 and 8/10 DEQAS samples, respectively. CONCLUSIONS: The ADVIA Centaur and E170 vitamin D assays demonstrated acceptable linearity, imprecision, and accuracy. The E170 assay demonstrated consistent underestimation of 25-OH vitamin D(2) levels. Compared with LC-MS/MS, the ADVIA Centaur assay demonstrated a higher R(2) value and a smaller average bias than the E170 assay.
    Clinical biochemistry 06/2012; 45(16-17). DOI:10.1016/j.clinbiochem.2012.06.002 · 2.23 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Vitamin D testing is becoming increasingly important with recent research demonstrating a correlation between vitamin D insufficiency and metabolic diseases, immunodeficiencies and other diseases. However, existing 25-hydroxyvitamin D (25OHD) assays lack comparability to the candidate reference method, causing difficulties in diagnosis and monitoring of vitamin D deficiency. We looked at the accuracy of 3 automated assays (Roche Diagnostics Elecsys® Total 25OHD assay, Abbott Architect® Total vitamin D assay, Advia Centaur® vitamin D Total assay) and Diasorin® Radioimmunoassay (RIA) compared to a routine laboratory Liquid Chromatography-Tandem Mass Spectrometry (LC-MS/MS). The correlation based on Passing Bablok regression was good with the slopes between 0.95 and 1.31 and the intercepts between -3.24 and 3.68. However, a significant positive bias was observed using the Abbott Architect and the Diasorin RIA. Using published analytical goals of coefficient of variation (CV) <10% and bias <5%, most methods did not meet these criteria. Using measurement of uncertainty of 9%, most methods were able to meet criteria using quality control materials but not patient samples. Inadequacies of these assay performances are contributed by differences in method of extraction of vitamin D from vitamin D binding protein, cross-reactivities to 25OHD(2), 25OHD(3) and other vitamin D metabolites, matrix interferences and a lack of standardization.
    Clinica chimica acta; international journal of clinical chemistry 03/2012; 413(13-14):1127-34. DOI:10.1016/j.cca.2012.03.009 · 2.76 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: Premesse L’utilizzo di sistemi di cromatografia liquida ultra veloce, non accoppiata a un rivelatore di massa, consentirebbe il dosaggio accurato delle differenti forme di vitamina D, proprio dei metodi cromatografici, in tempi rapidi, con produttività simile a quella dei metodi immunometrici, senza richiedere investimenti iniziali elevati generalmente tipici dei sistemi LC/MS/MS. In questo studio, utilizzando plasma e siero come matrici, abbiamo confrontato un nuovo metodo cromatografico liquido ultraveloce (UPLC) con un metodo immunometrico e uno cromatografico tradizionale (HPLC). Metodi 40 campioni di plasma e 50 campioni di siero sono stati valutati tramite kit ClinRep 25-OH-vitamina D2/D3® Recipe (BSN Srl, Castelleone, CR, Italia) su strumentazione UPLC Nexera Prominence LC20 (Shimadzu Italia Srl, Milano, Italia). I campioni di plasma sono stati analizzati con kit LIAISONsr 25-OH-Vitamin D Total (DiaSorin SpA, Saluggia, VC, Italia) su strumentazione Liaison Diagnostic System, mentre i campioni di siero sono stati misurati con kit 25-OH Vitamin D3/D2 by HPLC® Bio-Rad Laboratories Srl (Segrate, MI, Italia) su strumentazione Agilent Technologies (Santa Clara, CA, USA) HPLC serie 1200. Per la valutazione dell’imprecisione e dell’accuratezza sono stati utilizzati un controllo a bassa (LC) e alta concentrazione (HC) e un pool di plasma (PN). Risultati Il metodo UPLC presenta una buona precisione per entrambe le forme di vitamina D (5–7%). Sebbene per livelli bassi di vitamina D l’inaccuratezza si attesta a valori del 2%, per concentrazioni nell’intervallo di normalità (>30 ng/ml) essa è pari a 6,5-8,5%. Per il confronto DiaSorin-UPLC su plasma, l’analisi di Bland-Altman ha mostrato una differenza media percentuale pari a 74,5%, con valori circa doppi per il metodo UPLC rispetto al metodo DiaSorin, come confermato dall’analisi di regressione di Passing-Bablok (UPLC = DiaSorin*2.38). Per il confronto HPLC-UPLC su siero, l’analisi di Bland-Altman ha mostrato una differenza media percentuale pari a 11,6%, con valori più elevati per il metodo UPLC. L’analisi di regressione di Passing-Bablok ha evidenziato una relazione UPLC = HPLC*1,09+0,27. Il numero di previsioni non corrette, relativamente agli intervalli <10, 10-30, 30-100, >100 ng/ml, non supera il 20%. Conclusioni Il metodo UPLC presenta valori di vitamina D sostanzialmente più elevati rispetto al metodo DiaSorin, non risultando intercambiabile a meno di una modifica degli intervalli di riferimento. Pur presentando un’inaccuratezza superiore al limite desiderabile suggerito dalla recente letteratura, il metodo UPLC fornisce risultati di vitamina D sostanzialmente sovrapponibili al metodo HPLC tradizionale.
    Rivista Italiana della Medicina di Laboratorio 09/2012; 8(3). DOI:10.1007/s13631-012-0054-z