Research Article Received: 29 December 2008, Revised: 16 March 2009, Accepted: 16 March 2009 Published online in Wiley Interscience: 2009 (www.interscience.wiley.com) DOI 10.1002/bmc.1250 Biomed. Chromatogr. 2009 Copyright © 2009 John Wiley & Sons, Ltd. 1 John Wiley & Sons, Ltd. The external quality assurance of phylloquinone (vitamin K 1 ) analysis in human serum The external quality assurance of phylloquinone analysis David J. Card, a * Martin J. Shearer, a Leon J. Schurgers b and Dominic J. Harrington a ABSTRACT: The vitamin K external quality assurance scheme (KEQAS) aims to assist in the harmonization of phylloquinone (vitamin K 1 ) analysis in order to improve the comparability of clinical and nutritional studies. Serum samples were despatched to 17 groups from eight countries during 2000–2006. Using pilot data (1996–1999), an analytical performance target of 20% absolute difference from the all-laboratory trimmed mean (ALTM) was assigned and formed the basis for interlaboratory comparison. Assay specificity, analytical bias and assay performance were evaluated. From 21 batches of samples distributed, 414 results were reported of which 2.7% were outliers. The mean interlaboratory absolute difference from the ALTM was 21.7% with 47% of groups consistently meeting the performance target. The mean interlaboratory coefficient of variation was 29.6%. The false positive rate for phylloquinone depleted samples was high at 35%. Bias was found to be independent of HPLC-detector type (fluorescence vs electrochemical). Assay characteristics for the measurement of phylloquinone in human serum compare favourably with methods for analytes at equivalent concentrations. The high proportion of false positive results suggest that poor assay specificity at low phylloquinone concentrations is a common problem, which in the clinical setting could lead to underreporting of vitamin K deficiency. Copyright © 2009 John Wiley & Sons, Ltd. Keywords: vitamin K; HPLC; quality assurance Introduction The first method capable of accurate measurements of circulat- ing phylloquinone [vitamin K 1 (20); abbreviated K 1 ] at physiological concentrations was published in the early 1980s (Shearer et al., 1982). This advance followed the development of reversed-phase HPLC during the late 1970s that, with improved microparticulate packing materials and sensitive in-line UV detectors, allowed the separation and detection of structurally similar compounds at relatively low concentrations. Subsequently, several groups developed methodologies to measure endogenous K 1 concentra- tions using electrochemical and fluorescence detection methods that are both more sensitive and selective than those using UV detection (Table 1). Vitamin K is an essential micronutrient that, in the hydro- quinone form (K 1 H 2 ), is a cofactor for γ-glutamyl carboxylase during the post-translational γ-carboxylation of specific peptide bound glutamic acid residues to γ-carboxyglutamic acid (Gla) (Berkner and Runge, 2004). The presence of Gla residues within the Gla-domain of vitamin K-dependent proteins confers metal binding properties. These proteins are involved in a range of processes that include haemostasis, bone mineralization, arterial calcification, apoptosis, phagocytosis, growth control, chemot- axis and signal transduction (Berkner and Runge, 2004). K 1 is the predominant dietary source of vitamin K in humans. Circulatory levels of K 1 are a well-recognized indicator of vitamin K status and have been shown to reflect dietary intake (Booth et al., 1999; McKeown et al., 2002) and to correlate with the γ-carboxylation status of vitamin K-dependent proteins such as osteocalcin and prothrombin (Sokoll et al., 1996; Binkley et al., 2000, 2002) and with the urinary excretion of Gla (Sokoll et al., 1997). K 1 H 2 is readily converted to K 1 during pre-analytical sample preparation through exposure to atmospheric air (Haroon et al., 1986); therefore all current assays for K 1 quantify the sum of K 1 H 2 and K 1 . Measurement of K 1 is challenging and is carried out by a relatively small number (<40) of laboratories worldwide. The low endogenous levels of K 1 is one factor that complicates its measurement; with typical serum concentration ranges in UK healthy adults of 0.15–1.55 μg/L (median 0.53 μg/L) and 0.17– 0.68 μg/L (median 0.37 μg/L) in the nonfasting and fasting states respectively (Shearer et al., 1988). These levels are far lower than the other fat-soluble vitamins which, in fasting subjects, are present at concentrations that are approximately 50-fold (25- OH-vitamin D 3 ), 2000-fold (retinol) and 25,0000-fold (α-tocopherol) higher than K 1 . * Correspondence to: D. J. Card, The Centre for Haemostasis and Thrombosis, Guy’s and St Thomas’ NHS Foundation Trust, London SE1 7EH, UK. E-mail: david.card@gstt.nhs.uk a The Centre for Haemostasis and Thrombosis (Nutristasis Unit), Guy’s and St Thomas’ NHS Foundation Trust, London, UK b Cardiovascular Research Institute (CARIM) and Vitak, University of Maas- tricht, Maastricht, The Netherlands Abbreviations used: ALTM, all laboratory trimmed mean; EQA, external quality assurance; CV, coefficient of variation; Gla, γ-carboxyglutamic acid; K 1 , phylloquinone or vitamin K1(20); K 1 H 2 , vitamin K 1 hydroquinone; K 1 (25), vitamin K 1 (25); KEQAS, the Vitamin K External Quality Assurance Scheme; UV, ultraviolet.