Pöördumine

Forensic Science International 340 (2022) 111438 Contents lists available at ScienceDirect Forensic Science International journal homepage: www.elsevier.com/locate/forsciint New evidence of high association between carbohydrate deficient transferrin (CDT) and alcohol-related road traffic accidents. A ]] ]]]]]] ]] retrospective study on 929 injured drivers ⁎ Nadia M. Porpiglia a, , Franco Tagliaro a,b, Rocco Micciolo c, Luisa Canal d, Giacomo Musile a, Federica Bortolotti a a Unit of Forensic Medicine, Department of Diagnostics and Public Health, University of Verona, Piazzale L. A. Scuro 10, 37134 Verona, VR, Italy b Institute Translational Medicine and Biotechnology, Sechenov First Moscow State Medical University, 2–4 Bolshaya Pirogovskaya Street, 119991 Moscow, Russia c Department of Psychology and Cognitive Sciences and Centre for Medical Sciences, University of Trento, Corso Bettini, 84, 38068 Rovereto, TN, Italy d Department of Psychology and Cognitive Sciences, University of Trento, Corso Bettini, 84, 38068 Rovereto, TN, Italy a r t i cl e i nfo a bstr ac t Article history: Background: It is well known that traffic injuries still represent one of the main causes of death and that Received 4 July 2022 high blood alcohol concentrations while driving significantly increase the occurrence of accidents. However, Accepted 18 August 2022 only limited literature on the correlation between chronic alcohol abuse and accident risk is available. The Available online 20 August 2022 aim of the present study was to investigate the hypothesis of an association between elevated con­ centrations of carbohydrate deficient transferrin (CDT) and the occurrence of alcohol-related traffic acci­ Keywords: dents. Alcohol-related road traffic accidents Carbohydrate deficient transferrin Methods: The analytical determinations of BAC and CDT were performed following certified methods in HS- Blood alcohol concentration GC-FID and HPLC, respectively. For BAC, 0.50 g/L was used as cut-off, whereas 2.0% was used for CDT, Forensic toxicology according to the standardisation proposed by IFCC. A total of 929 drivers, tested for BAC at the time of Traffic medicine hospital admission after a traffic accident, were classified into two groups: InjDr 1 (BAC ≤ 0.50 g/L) and InjDr 2 (BAC > 0.50 g/L); all drivers were also tested for CDT. Results: InjDr 1 included 674 individuals, only 2.5% showing a CDT above the cutoff, whereas InjDr 2 group consisted of 255 subjects, 28.6% testing positive for CDT (Odds Ratio 15.5). When subdividing the InjDr group into increasing classes of CDT, a steady increase in the percentage of BAC-positive drivers was ap­ preciated. Moreover, average BAC was found to parallel each class of CDT. Conclusions: The reported data strongly support the use of CDT as a biomarker of increased risk of alcohol- related traffic accidents in the procedures of re-granting of the driving license upon confiscation for “drink driving”. © 2022 Published by Elsevier B.V. 1. Introduction In fact, since the Grand Rapids Study [6] until present, a strong body of literature has piled highlighting that high blood alcohol con­ It is well known that road traffic accidents still represent one of centrations (BACs) drastically increase the likelihood of occurrence the main causes of death worldwide, also being the first cause of of road traffic accidents [2,4,7,8]. In particular, it has been estimated death in the population, aged 5–29 years [1]. The correlation be­ that for every 0.02 g/L of BAC increase, the risk of road traffic acci­ tween alcohol misuse and road traffic accidents has widely been dents doubles [5]. Such a vast and well-documented evidence-based investigated [2–5] and policies to contrast the phenomenon of consensus on the acute effect of alcohol on traffic safety is juxta­ “drunk driving” are adopted worldwide. A topic of general agree­ posed by a much weaker literature on the correlation between ment among authors is that concentrations of alcohol in blood above chronic alcohol abuse and the risk of traffic crashes. Nevertheless, 0.50 g/L strongly affect driving skills and compromise traffic safety. the need of verifying the existence of such an association is of paramount importance in the process of testing the fitness of in­ dividuals to hold the driving license or to release other certifications ⁎ attesting the suitability to safety-sensitive jobs, child custody and in Correspondence to: Piazzale L. A. Scuro 10, 37134 Verona, VR, Italy. E-mail address: [email protected] (N.M. Porpiglia). other cases of civil litigation. Evidence that chronic alcohol abusers https://doi.org/10.1016/j.forsciint.2022.111438 0379-0738/© 2022 Published by Elsevier B.V. N.M. Porpiglia, F. Tagliaro, R. Micciolo et al. Forensic Science International 340 (2022) 111438 are “predisposed” to road traffic accidents already appeared in early 5 test samples. The lower limit of quantitation (LOQ) of this tech­ studies carried out in the 1960 s [9], but the lack of objective para­ nique is 0.01 g/L. The laboratory follows an external quality control meters of this condition hindered the practical significance of these programme for ethanol provided by LGC Proficiency Testing (Bury, studies. In fact, the diagnosis of chronic alcohol abuse relied only on United Kingdom). clinical evidence, questionnaires, and non-specific biomarkers of alcohol-related organ damage (MCV, liver enzymes, etc.). Since the 2.2. CDT analysis 1980 s, however, the introduction of more specific biomarkers (CDT, FAEE, Peth, and EtG, just to mention a few) has offered better di­ CDT determination was performed on a gradient HPLC system agnostic tools and, consequently, the research in this field has (LC-10 VP series, Shimadzu Europe, Duisburg, Germany) equipped steadily increased [10,11]. Among these biomarkers, by far the most with a UV–vis detector (model SPD-10AV VP, Shimadzu) set at a studied is CDT, which has been widely applied to corroborate the wavelength of 460 nm. The analytical procedure was performed diagnosis of chronic abuse of alcohol [12–15] and, in clinical con­ following an edited version of the method reported by Jeppsson texts, this biomarker showed a diagnostic sensitivity and specificity et al. [24]. In brief, starting buffer (phase A) consisted of 10 mM Bis- of 70% and 98%, respectively [16]. Tris at pH 6.2, and phase B was made of phase A added with 500 mM The spreading of the diagnostic use of CDT in forensic and clinical NaCl at the same pH. The solutions were used to generate a linear contexts, as well as the availability of different analytical approaches gradient from 0% to 27% of phase B within 20 min. A solution of 2 M for its determination, encouraged the International Federation of NaCl was used for regeneration. Upon preparation, all solutions were Clinical Chemistry (IFCC) to establish a Working Group on degassed and filtered through 0.45 µm membrane filters. Samples Standardisation of CDT (WG-CDT). The final aim of their work was to were injected at a volume of 100 µL and flow rate was maintained at define and validate the CDT analyte, to select a reference method and 1 mL/min. The analytical anion-exchange column [65 × 4.6 mm (i.d.)] reference materials, and to create a network of laboratories [17–20]. was kept at a temperature of 40 °C during the separation and was This action has established a sound ground for the adoption of this part of a commercial kit named ClinRep® CDT in Serum provided by analyte as a biomarker of chronic alcohol abuse at international Recipe, Munich, Germany. Prior to HPLC analysis, sample preparation level, assuring an inter-laboratory comparability of the results. required ferric ion saturation and lipoprotein precipitation with On the other hand, the use of CDT (as well as any other alcohol CaCl2, included in the same reagent kit. The results obtained in terms abuse biomarker) to verify the fitness to drive (or to work), is only of %CDT were normalised according to the IFCC WG-CDT standar­ based on the generic assumption that a chronic drinker consumes disation and the adopted cut-off limit between alcohol abusers and alcohol compulsively and consequently is likely to drive (or to work) non-abusers was 2.00% [20]. The laboratory follows three external under the effect of such substance, representing a risk to public quality control programmes for CDT provided by ARVECON GmbH safety. However, considering that the refusal of the certification of (Walldorf, Germany), by IFCC Network on Standardisation of CDT fitness to drive (or to work) can limit an individual’s rights, this (Winterswijk, The Netherlands), and by Biologie Prospective (Villers- assumption needs to find specific validation and not merely lie in the Les-Nancy, France). realm of “plausibility”. Surprisingly enough, to date, the biomarkers of alcohol abuse 2.3. Subjects and specimen collection have rarely found a diagnostic validation in this very context, being their application simply been drawn from the experience in the The present study included a total of 929 subjects who were clinical diagnosis of alcohol misuse. admitted to the University Hospital of Verona after being involved in However, quite recently CDT has been studied in terms of its road traffic accidents and suffering injuries of different severity specific association with the occurrence of alcohol-related road during 2018–2019. The total of 929 individuals did not include those traffic accidents [21], proving more reliable than other traditional patients in which urine testing was positive for the presence of biomarkers such as MCV and GGT [22,23]. psychoactive drugs (i.e., cocaine, amphetamines and their metabo­ On these grounds, the present work was undertaken to in­ lites, opioids like morphine, methadone and the major metabolites, vestigate a possible association, more deeply, between the serum benzodiazepines, and cannabinoids). As drivers implicated in a road concentrations of CDT and the blood alcohol concentrations at the traffic crash, the subjects were mandatorily tested for alcohol and time of a car accident, with the intention of posing a ground of psychoactive drugs as required by the Italian Road Traffic Code, art. knowledge to support the diagnostic value of this chronic alcohol 180 and 181. In this context, no ethical approval was needed by law. abuse biomarker in the delicate context of traffic medicine. However, all the subjects, after being informed on the purpose of the sample collection, provided valid verbal consent. 2. Methods After venipuncture, the whole blood samples for BAC analysis were collected and stored in oxalate fluoride tubes at 4 °C until 2.1. BAC analysis analysed. Blood samples for CDT analysis were collected in test tubes without preservatives and anti-coagulants, let them clot and cen­ BAC measurement was performed by head-space gas chromato­ trifuged at 3500 rpm for 10 min. The resulting supernatant was graphy (HS-GC) equipped with a flame ionisation detector (on a collected and stored frozen at − 20 °C until analysis. model 6100 GC analyser purchased from Young Lin Instrument Co. Ltd., Anyang, Korea) with a method routinely used and compliant 3. Results and discussion with the accreditation criteria of our laboratory (Health Authority of Regione Veneto, Italy). In brief, sample preparation was carried out In the herein reported study, the 929 drivers were classified into by mixing whole blood (1:4 dilution) with an aqueous solution of two groups according to the measured BAC at the time of hospital tert-butyl alcohol (0.0975 g/L) as the internal standard (Carlo Erba admission: Reagents, Cornaredo, Italy). After equilibration of the sample vials at 60 °C, a total volume of 1250 µL of the head space was automatically • InjDr1 group: BAC ≤ 0.50 g/L (i.e., within the Italian legal limit) injected in the system at a temperature of 80 °C. Isotherm separation • InjDr2 group: BAC > 0.50 g/L (i.e., above the Italian legal limit) was carried out at 40 °C. An internal quality standard provided by ACQ Science GmbH (Rottenburg am Neckar, Germany) containing InjDr1 group included 674 individuals, whereas InjDr2 group ethanol at a concentration of 0.50 g/L was measured at an interval of consisted of 255 subjects, who were considered by law “under the 2 N.M. Porpiglia, F. Tagliaro, R. Micciolo et al. Forensic Science International 340 (2022) 111438 Table 1 Distribution of BAC negatives and positives according to selected classes of CDT. %CDT range Number of % individuals with % individuals with individuals BAC > 0.50 g/L BAC ≤ 0.50 g/L 0 – 0.5 20 0 (n = 0) 100 (n = 20) 0.5 – 1.0 319 12 (n = 39) 88 (n = 280) 1.0 – 1.5 431 23 (n = 100) 77 (n = 331) 1.5 – 2.0 69 62 (n = 43) 38 (n = 26) 2.0 – 4.0 56 75 (n = 42) 25 (n = 14) > 4.0 34 91 (n = 31) 9 (n = 3) influence of alcohol” at the time of the crash. Overall, the BAC range was between 0.01 and 4.63 g/L (average 0.56 ± 0.98 SD). On the other hand, CDT values ranged from 0.13% to 14.47%. In the InjDr1 group, CDT showed an average of 1.08% ( ± 0.49% SD), whereas in the InjDr2 group the average value was 2.21% ( ± 2.13% SD). When considering InjDr1 group, only 2.5% of the subjects showed a CDT value exceeding the cut-off of 2.0%, whereas 28.6% of the subjects belonging to InjDr2 tested positive for CDT (hence > 2.0%). Fig. 2. Estimated probability of being BAC positive in relation to CDT values. This high prevalence of CDT positives in the alcohol-positive injured drivers was already reported by our group [21], but in the present study, because of a larger group of drivers, we managed to better • CDT cut-off set at 1.5% - > odds ratio = 12.2 investigate quantitatively the association between BAC and CDT. • CDT cut-off set at 2.0% - > odds ratio = 15.5 By taking into consideration the entire InjDr group (n = 929), the • CDT cut-off set at 4.0% - > odds ratio = 31.0 whole range of CDT values was divided into six classes, and, within The odds ratio is the ratio between the odds of being positive at each class, subjects were classified as BAC positives or negatives, as BAC analysis when CDT is “positive” (i.e. above the selected cut-off) shown in Table 1. The graphical representation of the table (Fig. 1) and the odds of being positive at BAC analysis when CDT is “nega­ clearly shows a steady increase in the percentage of BAC positives in tive” (i.e. below or equal to the selected cut-off). For instance, con­ parallel with the increase of CDT in its different classes. sidering the currently used cut-off of CDT (i.e., 2.0%), the It is remarkable to notice how the percentage of BAC-positive corresponding odds ratio of 15.5 indicates that those individuals individuals increases along with the increase of CDT concentration. who, by law, cannot hold the driving license show a 15.5 times At CDT concentrations ≤ 0.5%, all the injured drivers showed BACs higher possibility to result positive to BAC determination with re­ below the Italian legal limit of 0.50 g/L. In contrast, the CDT class spect to subjects with CDT ≤ 2.0%. ranging from 1.5% to 2.0% (i.e., above the average CDT concentrations On the other hand, it is possible to consider the pattern of mean in the population but still within the current CDT cut-off [25]) in­ values of the BAC in the injured drivers in relation to different cut- cluded 62% of subjects testing positive for BAC at the time of the car offs of CDT. According to this approach, CDT values were subdivided crash. This percentage rapidly increased to 75% when considering into nine classes and the mean BAC was calculated within each class. the CDT range between 2.0% and 4.0%. A non-parametric evaluation The results are shown in Fig. 3. The strength of the association be­ of the correlation between the two variables (CDT and BAC) in all the tween the two variables was evaluated using the η2 coefficient, injured drivers, using the Spearman rho coefficient, provided a value which provided a value of 0.541. Looking at the fifth point in the of 0.455. figure, it is also clear that the class of CDT ranging from 1.25% to It was also noteworthy to interpolate a logistic model employing 1.50% already corresponds to an average BAC (0.58 g/L) which ex­ CDT as the independent variable and the probability to test positive ceeds the legal limit adopted in Italy and in most Western countries. to BAC as the dependent one. The graphical representation of the estimated model is shown in Fig. 2, where the CDT value that cor­ responds to a 50% probability of being BAC positive is 1.96, i.e. substantially the CDT cut-off in use in many laboratories, in­ cluding ours. Another statistical tool to evaluate the strength of the association between CDT and BAC is to assess the odds ratios corresponding to different CDT values candidate to become suitable cut-offs in the context of traffic safety. The results are summarised as follows: • CDT cut-off set at 1.0% - > odds ratio = 4.6 Fig. 3. Mean BAC values within nine selected classes of CDT. The classes were: ≤ 0.5, 0.5–0.75, 0.75–1, 1–1.25, 1.25–1.5, 1.5–1.75, 1.75–2, 2–4, > 4. CDT is plotted on a Fig. 1. Percentage of BAC negatives and positives according to selected classes of CDT. logarithmic scale. Bars represent standard errors. 3 N.M. Porpiglia, F. Tagliaro, R. Micciolo et al. Forensic Science International 340 (2022) 111438 This could suggest a substantial reduction of the adopted CDT cut-off [12] J.V. Sakran, A. Mehta, M.M. Matar, D.A. Wilson, A.J. Kent, R.F. Anton, S.M. Fakhry, for certifying the fitness-to-drive. However, this crucial decision The utility of carbohydrate-deficient transferrin in identifying chronic alcohol users in the injured patient: expanding the toolkit, J. Surg. Res. 257 (2021) should be grounded also on specific studies focused on the calcu­ 92–100, https://doi.org/10.1016/j.jss.2020.07.027 lation of the increased risk of alcohol-related road traffic accidents in [13] T. Arndt, Carbohydrate-deficient transferrin as a marker of chronic alcohol the population with CDT values below the currently used cut-off. abuse: a critical review of preanalysis, analysis, and interpretation, Clin. Chem. 47 (2001) 13–27 〈http://clinchem.aaccjnls.org/content/47/1/13.abstract〉. [14] F.J. Legros, V. Nuyens, M. Baudoux, K.Z. Boudjeltia, J.L. Ruelle, J. Colicis, 4. Conclusions F. Cantraine, J.P. Henry, Use of capillary zone electrophoresis for differentiating excessive from moderate alcohol consumption, Clin. Chem. 49 (2003) 440–449 〈http://clinchem.aaccjnls.org/content/48/12/2177.abstract〉. Only a few authors have investigated, retrospectively, the use­ [15] J. van Pelt, Carbohydrate-deficient transferrin: a new biochemical marker for fulness of CDT in the context of traffic safety, generally reporting a chronic excessive alcohol consumption, Ned. Tijdschr. Geneeskd. 141 (1997) correlation between CDT and BAC but mostly based on data not 773–777. [16] R.M. Nanau, M.G. Neuman, Biomolecules and biomarkers used in diagnosis of collected on the occasion of an occurred accident [26–34]. Also, the alcohol drinking and in monitoring therapeutic interventions, Biomol 5 (2015), research groups of Maenhout and Portman reported on CDT as a https://doi.org/10.3390/biom5031339 potential predictor of recidivism in drunk driving [35,36]. In con­ [17] J.O. Jeppsson, T. Arndt, F. Schellenberg, J.P.M. Wielders, R.F. Anton, J.B. Whitfield, trast, the present work, stemmed from previous articles from our A. Helander, Toward standardization of carbohydrate-deficient transferrin (CDT) measurements: I. Analyte definition and proposal of a candidate reference group on the association between CDT and elevated BAC in road method, Clin. Chem. Lab. Med 45 (2007) 558–562, https://doi.org/10.1515/CCLM. traffic crashes [21,23], and was aimed at studying this relationship 2007.107 on a quantitative basis, as a better confirmation of the reliability of [18] A. Helander, J.P.M. Wielders, J.O. Jeppsson, C. Weykamp, C. Siebelder, R.F. Anton, F. Schellenberg, J.B. Whitfield, Toward standardization of Carbohydrate-deficient CDT as a biomarker indicating the risk of being involved in a road transferrin (CDT) measurements: II. Performance of a laboratory network run­ accident while drink driving. The presented data clearly confirms ning the HPLC candidate reference measurement procedure and evaluation of a “quantitatively” the association of CDT with the occurrence of al­ candidate reference material, Clin. Chem. Lab. Med 48 (2010) 1585–1592, https://doi.org/10.1515/CCLM.2010.322 cohol-related road traffic accidents. Notwithstanding the inherent [19] C. Weykamp, J.P.M. Wielders, A. Helander, R.F. Anton, V. Bianchi, J.O. Jeppsson, limitations of a cross-sectional study, such as the present one, and C. Siebelder, J.B. Whitfield, F. Schellenberg, Toward standardization of carbohy­ considering the huge logistic and ethical problems of a longitudinal drate-deficient transferrin (CDT) measurements: III. Performance of native serum and serum spiked with disialotransferrin proves that harmonization of study, we believe that the overall information at present available on CDT assays is possible, Clin. Chem. Lab. Med 51 (2013) 991–996, https://doi.org/ CDT and traffic crashes supports once again with sound evidence the 10.1515/cclm-2012-0767 usefulness of this biomarker for the certification of the fitness-to- [20] F. Schellenberg, J. Wielders, R. Anton, V. Bianchi, J. Deenmamode, C. Weykamp, J. Whitfield, J.O. Jeppsson, A. Helander, IFCC approved HPLC reference mea­ drive, in case of a pre-existing alcohol related driving license con­ surement procedure for the alcohol consumption biomarker carbohydrate-de­ fiscation. These considerations could reasonably be extended also to ficient transferrin (CDT): Its validation and use, Clin. Chim. Acta 465 (2017) accidents occurring in other contexts, and primarily to the fields of 91–100, https://doi.org/10.1016/j.cca.2016.12.022 occupational accidents and safety-sensitive jobs. [21] F. Bortolotti, R. Micciolo, L. Canal, F. Tagliaro, First objective association between elevated carbohydrate-deficient transferrin concentrations and alcohol-related Last-but-not-least, the present data show also that an average traffic accidents, Alcohol. Clin. Exp. Res. 39 (2015) 2108–2114, https://doi.org/10. BAC above the common legal limit of 0.50 g/L can be found also in 1111/acer.12879 subjects showing CDT percentages below the currently adopted cut- [22] N.M. Porpiglia, F. Bortolotti, R.M. Dorizzi, R. Micciolo, F. Tagliaro, Critical eva­ luation of the association between elevated mean corpuscular volume and al­ off of 2.0%. This might suggest the opportunity of a re-evaluation of cohol-related traffic accidents: a retrospective study on 6244 car crash cases, the present cut-off established on the basis of clinical studies, far Alcohol. Clin. Exp. Res. 43 (2019) 1528–1532, https://doi.org/10.1111/acer.14046 from the field of traffic safety. Of course, this action would require [23] N.M. Porpiglia, F. Bortolotti, R. Micciolo, L. Canal, M. Murari, F. Gibelli, F. Tagliaro, CDT vs. GGT for the certification of the fitness to hold the driving license. A adequate field studies in a multicentre context. comparison based on the association of incremented values with the occurrence of alcohol-related road traffic accidents, Drug Alcohol Depend. 228 (2021) Declaration of interests 109088, https://doi.org/10.1016/j.drugalcdep.2021.109088 [24] J.O. Jeppsson, H. Kristensson, C. Fimiani, Carbohydrate-deficient transferrin quantified by HPLC to determine heavy consumption of alcohol, Clin. Chem. 39 We declare no competing interests. (1993) 2115–2120. [25] F. Bortolotti, G. De Paoli, J.P. Pascali, M.T. Trevisan, M. Floreani, F. Tagliaro, Analysis of carbohydrate-deficient transferrin: comparative evaluation of tur­ References bidimetric immunoassay, capillary zone electrophoresis, and HPLC, Clin. Chem. 51 (2005) 2368–2371. [1] WHO, Global status report on alcohol and health, 2018. [26] B.M.R. Appenzeller, S. Schneider, A. Maul, R. Wennig, Relationship between [2] J. Yao, R.B. Voas, J.H. Lacey, Drivers with alcohol use disorders and their risks of blood alcohol concentration and carbohydrate-deficient transferrin among dri­ crash involvement, Drug Alcohol Depend. 183 (2018) 210–216, https://doi.org/ vers, Drug Alcohol Depend. 79 (2005) 261–265, https://doi.org/10.1016/j. 10.1016/j.drugalcdep.2017.10.040 drugalcdep.2005.01.017 [3] J. de Carvalho Ponce, D.R. Muñoz, G. Andreuccetti, D.G. de Carvalho, V. Leyton, [27] B.M.R. Appenzeller, S. Schneider, M. Yegles, A. Maul, R. Wennig, Drugs and Alcohol-related traffic accidents with fatal outcomes in the city of Sao Paulo, chronic alcohol abuse in drivers, Forensic Sci. Int. 155 (2005) 83–90, https://doi. Accid. Anal. Prev. 43 (2011) 782–787, https://doi.org/10.1016/j.aap.2010.10.025 org/10.1016/j.forsciint.2004.07.023 [4] I. Evangelidis, The role of restraint omission in alcohol-related traffic fatalities, [28] B. Brinkmann, J. Beike, H. Köhler, A. Heinecke, T. Bajanowski, Incidence of alcohol Drug Alcohol Depend. 180 (2017) 423–426, https://doi.org/10.1016/j.drugalcdep. dependence among drunken drivers, Drug Alcohol Depend. 66 (2002) 7–10, 2017.09.003 https://doi.org/10.1016/S0376-8716(01)00177-6 [5] M.T. Gómez-Talegón, F.J. Alvarez, Road traffic accidents among alcohol-depen­ [29] H. Gjerde, J. Mørland, Concentrations of carbohydrate-deficient transferrin in dent patients: the effect of treatment, Accid. Anal. Prev. 38 (2006) 201–207, dialysed plasma from drunken drivers, Alcohol Alcohol 22 (1987) 271–276, https://doi.org/10.1016/j.aap.2005.09.006 https://doi.org/10.1093/oxfordjournals.alcalc.a044708 [6] R.F. Borkenstein, R.F. Crowther, R.P. Shumate, The role of the drinking driver in [30] R. Iffland, F. Grassnack, Epidemiologic study of CDT (carbohydrate-deficient traffic accidents (the Grand Rapids study), Blutalkohol 11 (1974) 131. transferrin) and other indicators of alcohol problems in high blood alcohol [7] WHO, WHO Expert Committee on Problems Related to Alcohol Consumption. German automobile drivers, Blutalkohol 32 (1995) 26–41. Second report., World Health Organ, Tech. Rep. Ser 1–53 (2007), pp. 55–57. [31] R. Iffland, M.P. Balling, G. Börsch, C. Herold, W. Kaschade, T. Löffler, [8] WHO, Global Status Report on Alcohol, World Health Organization. Department U. Schmidtmann, J. Stettner, Evaluation of an increased blood level of GGT, CDT, of Mental Health and Substance Abuse, Geneva, Switzerland, 2004. methanol, acetone and isopropanol in alcohol intoxicated automobile drivers. [9] J.A. Waller, H.W. Turkel, Alcoholism and traffic deaths. N. Engl. J. Med. 275 (1966) Alcoholism indicators instead of medical-psychological examination, 532–536, https://doi.org/10.1056/NEJM196609082751005 Blutalkohol 31 (1994) 273–314. [10] H. Andresen-Streichert, A. Müller, A. Glahn, G. Skopp, M. Sterneck, Alcohol [32] R. Jaster, R. Wegener, Diagnosis of chronic alcohol abuse in intoxicated drivers: biomarkers in clinical and forensic contexts, Dtsch. Arztebl. Int 115 (2018) carbohydrate-deficient transferrin (CDT) in combination with other parameters, 309–315, https://doi.org/10.3238/arztebl.2018.0309 Blutalkohol 30 (1993) 257–265. [11] J. Rainio, F. De Giorgio, F. Bortolotti, F. Tagliaro, Objective post-mortem diagnosis [33] H. Kristenson, J.O. Jeppsson, Drunken driver examinations. CD-transferrin is a of chronic alcohol abuse: a review of studies on new markers, Leg. Med. (Tokyo). valuable marker of alcohol consumption, Lakartidningen 95 (1998) 1425–1426 10 (2008) 229–235, https://doi.org/10.1016/j.legalmed.2008.01.006 1429–1430. 4 N.M. Porpiglia, F. Tagliaro, R. Micciolo et al. Forensic Science International 340 (2022) 111438 [34] P. Marques, S. Tippetts, J. Allen, M. Javors, C. Alling, M. Yegles, F. Pragst, F. Wurst, alcohol-impaired driving (ROAD) study, Addiction 109 (2004) 71–78, https://doi. Estimating driver risk using alcohol biomarkers, interlock blood alcohol con­ org/10.1111/add.12372 centration tests and psychometric assessments: initial descriptives, Addiction [36] M. Portman, A. Penttilä, J. Haukka, P. Eriksson, H. Alho, K. Kuoppasalmi, 105 (2009) 226–239, https://doi.org/10.1111/j.1360-0443.2009.02738.x Predicting DUI recidivism of male drunken driving: a prospective study of the [35] T.M. Maenhout, A. Poll, T. Vermassen, M.L. De Buyzere, J.R. Delanghe, Usefulness impact of alcohol markers and previous drunken driving, Drug Alcohol Depend. of indirect alcohol biomarkers for predicting recidivism of drunk-driving among 106 (2010) 186–192, https://doi.org/10.1016/j.drugalcdep.2009.08.018 previously convicted drunk-driving offenders: results from the recidivism of 5 Saatja: Krause, Karit <[email protected]> Saadetud: 26.06.2023 17:08 Teema: Kroonilise alkoholi liigtarbimise tuvastamiseks mõeldud markeri implementeerimine Eestis Manused: image001.png; PORPIGLIA aug22.pdf; IFCC approved HPLC RMP for CDT_Helander 2016.pdf; Preventing-Harmful-Alcohol-Use-Key-Findings-ESTONIA.pdf Tere! Umbes 2 aastat tagasi pöördusin Tervisearengu Instituudi poole seoses kroonilise alkoholi liigtarbimise skriininguga. Sealt suunati edasi Justiitsministeeriumi ja Sotsiaalministeeriumi kontaktide poole. Kahjuks ei ole selle teemaga minule teadaolevalt siiski edasi mindud. Nimelt on olemas võimalus kasutada vereseerumi markerit, mis on mujal maailmas laialt levinud nii kohtupraktikas, arsti abiliseks kroonilise alkoholi liigtarbimisega (ja/või muude haigustega) patsientide ravis kui ühiskonda teenindavate ametikohtade skriinimisel (juhilubade/relvalubade taotlemisel, piloodid, kaptenid, bussijuhid jne). Väga paljudes riikides on eelnevalt nimetatud juhtudel CDT-IFCC määramine ja skriining tavaprotseduur. Ka Eestis võiks sarnane määramine toimuda, et meie ühiskorda tervemaks ja turvalisemaks muuta. Eestis on olemas hetkel 3 laborit, kus määratakse standardiseeritud CDT-d, kuid, et puudub ühtne süsteem ja reeglistik on selle määramise hulk hetkel väga marginaalne. CDT-IFCC on Rahvusvahelise Kliinilise Keemia ja Laboratoorse Meditsiini Föderatsiooni poolt väljatöötatud kalibreeritud CDT väärtus, mis näitab alkoholi liigtarbimist möödunud 2 nädala kohta (sarnaselt glükeeritud hemoglobiiniga diabeediravis). Marker nimega CDT-IFCC ehk süsivesik-defitsiitne transferriin. Kalibreeritud süsteemidel, näiteks kapillaarelektroforeesil, on CDT-IFCC tulemused tõesed ja omavahel võrreldavad üle maailma. CDT on ainus IFCC poolt standardiseeritud ning kõige spetsiiflisem kroonilise alkoholi liigtarbimisega seotud marker, mis on kasutusel juba 2018ndast aastast. Lisasin manusesse mõne artikli CDT-IFCC kohta ning OECD hinnangu alkoholi liigtarbimisega seotud ennetustöö võimaluse kohta Eestis. Kas oskate soovitada, kelle poole peaksime pöörduma, et aidata määrustega ka Eesti ühikonda tervislikkuse ja ohutuse suunas sammuke edasi astuda? Ette tänades ja tagasisidet ootama jäädes! Parimatega / Best regards, Karit Krause Müügijuht / Sales Manager Laboritarbed ja -instrumendid / Laboratory Business T +372 6515 168 | GSM +372 56237 168 | [email protected] | www.mediq.ee | Mediq Eesti OÜ | Väljaotsa 2, 76505 Saue, Estonia | Please consider the environment before printing this email _________________________________________________________________________________________________ Disclaimer This e-mail is meant exclusively for the intended recipient. If you received this e-mail in error, any use, copying or distribution of this e-mail is strictly prohibited. Please notify the sender of the error by return e-mail and destroy/delete this e-mail, as well as any attachment. Mediq B.V. is registered in the chamber of commerce under number 56038062 Preventing Harmful Alcohol Use Alcohol can be a source of enjoyment and is an important part of the economy in many countries. Yet harmful alcohol use is a leading risk factor for premature death, injuries and many non- communicable diseases. Harmful alcohol use can also affect others, Estonia for example in the case of violence and foetal alcohol syndrome. These diseases and injuries, in turn, have an impact on healthcare budgets and the productivity of the labour force. People in Estonia consume 9.2 litres of pure alcohol per capita per year, roughly equivalent to 1.9 bottles of wine or 3.5 litres of beer per week per person aged 15 and over. In addition, in Estonia, some population groups are at higher risk than others; specifically: 42% of adults engage in binge drinking at least once a month. This Men consume 15.0 litres of pure corresponds to drinking more than alcohol per capita per year while 80% of a bottle of wine or 1.5 litres women consume 4.4 litres per of beer per occasion. capita per year. 25% of girls and 29% of boys Women are 6% more likely to aged 15 have been drunk at least binge drink monthly if they have twice in their life. Children who completed higher education never experienced drunkenness are 24% more likely to perform well at school. Life expectancy is 1.6 years lower over the next 30 years, due to diseases and injuries caused by drinking more than 1 drink per day for women and 1½ drinks per day for men, on average in the whole population, according to OECD simulations. Impact on life expectancy Impact on health expenditure ITA CAN JPN GBR USA OECD FRA EU DEU EST 0 4.0% -0.2 3.5% -0.4 3.0% % of health expenditure -0.6 2.5% -0.8 Years 2.0% -1 1.5% -1.2 -1.4 1.0% -1.6 0.5% -1.8 0.0% ITA JPN FRA CAN OECD EU GBR USA DEU EST Based on current consumption patterns in Estonia, OECD simulations estimate that diseases and injuries caused by drinking above 1-1½ drinks per day lead to treatment costs equal to 3.5% of health expenditure and a reduction in the workforce productivity. Consequently, Estonia’s GDP is estimated to be 3.4% lower on average between now and 2050, excluding any impact on the alcohol industry. Due to a lower GDP, and to maintain a constant public debt-to-GDP ratio, Estonia has to raise additional revenues equivalent to an increase in tax of EUR 170 per person per year.  Implementation level of policies to address harmful alcohol use in Estonia OECD analysis of WHO data reflects the implementation status across policy areas within Health services 1 2 3 4 the WHO’s Global Strategy to Reduce the Harmful School, Use of Alcohol. workplace 1 2 3 4 Estonia performs well in certain policy areas, but there are opportunities for further action. Policy Drink-driving 1 2 3 4 priorities could include: • Improving the implementation of screening Availability 1 2 3 4 and counselling within primary care services for people who drink heavily; Marketing 1 2 3 4 • Strengthening regulation of advertising on social media and new media, which are frequently used by younger people; Pricing 1 2 3 4 • Strengthening restriction on availability of Bar staff training, alcohol to vulnerable and high-risk groups, for 1 2 3 4 warning label instance by time or place; 1-lower level of implementation, 4-higher level. • Training servers on how to prevent, identify Countries with a maximum score can still and manage intoxicated drinkers, as for enhance policy implementation and enforcement. instance in Germany and Spain. OECD analyses looked at enhanced policy package to tackle harmful alcohol use. The package contains many of the policy priorities for Estonia, including: Strengthening sobriety checkpoints to Strengthening screening and counter drink-driving counselling in primary care Complete ban on alcohol advertising to Strengthening regulation on children via traditional and new media, alcohol advertising, sponsorships, sponsorships, branding and point-of- branding and point-of-sale displays sale displays Alcohol taxation Minimum unit pricing targeting cheap alcohol In Estonia, investing EUR 1.5 per person per year in the enhanced policy package to tackle harmful alcohol use will: • prevent 137 thousand non-communicable diseases and injuries by 2050; • save EUR 3.6 million per year in health costs; • increase employment and productivity by the equivalent of 1.5 thousand full-time workers per year. For every EUR 1 invested in the policy package, EUR 16 are returned in benefits, not considering any impact on the alcohol industry. Discover the OECD SPheP-NCD data explorer and the model’s documentation at http://oecdpublichealthexplorer.org/ Find the full OECD report Preventing Harmful Alcohol Use at oe.cd/alcohol2021 (This is a sample cover image for this issue. The actual cover is not yet available at this time.) This article appeared in a journal published by Elsevier. The attached copy is furnished to the author for internal non-commercial research and education use, including for instruction at the author's institution and sharing with colleagues. Other uses, including reproduction and distribution, or selling or licensing copies, or posting to personal, institutional or third party websites are prohibited. In most cases authors are permitted to post their version of the article (e.g. in Word or Tex form) to their personal website or institutional repository. Authors requiring further information regarding Elsevier's archiving and manuscript policies are encouraged to visit: http://www.elsevier.com/authorsrights  Author's Personal Copy Clinica Chimica Acta 465 (2017) 91–100 Contents lists available at ScienceDirect Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim IFCC approved HPLC reference measurement procedure for the alcohol consumption biomarker carbohydrate-deficient transferrin (CDT): Its validation and use☆ François Schellenberg a, Jos Wielders b,⁎,1, Raymond Anton c, Vincenza Bianchi d, Jean Deenmamode e, Cas Weykamp f, John Whitfield g, Jan-Olof Jeppsson h, Anders Helander i a Hôpital Trousseau, CHRU, Tours, France b Meander Medisch Centrum, Amersfoort, The Netherlands c Medical University of South Carolina, Charleston, SC, USA d SS. Antonio e Biagio Hospital, Alessandria, Italy e Homerton University Hospital, London, United Kingdom f Queen Beatrix Hospital, Winterswijk, The Netherlands g Queensland Institute of Medical Research, Brisbane, Australia h Skåne University Hospital, Malmö, Sweden i Karolinska Institutet, Karolinska University Laboratory, Stockholm, Sweden a r t i c l e i n f o a b s t r a c t Article history: Carbohydrate-deficient transferrin (CDT) is used as a biomarker of sustained high alcohol consumption. The cur- Received 21 December 2016 rently available measurement procedures for CDT are based on various analytical techniques (HPLC, capillary Accepted 21 December 2016 electrophoresis, nephelometry), some differing in the definition of the analyte and using different reference in- Available online 23 December 2016 tervals and cut-off values. The Working Group on Standardization of CDT (WG-CDT), initiated by the Internation- al Federation of Clinical Chemistry and Laboratory Medicine (IFCC), has validated an HPLC candidate reference Keywords: CDT measurement procedure (cRMP) for CDT (% disialotransferrin to total transferrin based on peak areas), demon- Disialotransferrin strating that it is suitable as a reference measurement procedure (RMP) for CDT. Presented is a detailed descrip- HPLC tion of the cRMP and its calibration. Practical aspects on how to treat genetic variant and so-called di-tri bridge Alcohol biomarker samples are described. Results of method performance characteristics, as demanded by ISO 15189 and ISO Standardization 15193, are given, as well as the reference interval and measurement uncertainty and how to deal with that in Reference Measurement Procedure (RMP) routine use. The correlation of the cRMP with commercial CDT procedures and the performance of the cRMP in a network of laboratories are also presented. The performance of the CDT cRMP in combination with previous- ly developed commutable calibrators allows for standardization of the currently available commercial measure- ment procedures for CDT. The cRMP has recently been approved by the IFCC and will be from now on be known as the IFCC-RMP for CDT, while CDT results standardized according to this RMP should be indicated as CDTIFCC. © 2016 Elsevier B.V. All rights reserved. 1. Introduction and scope oligosaccharide chains attached to Asn432 and Asn630 [1]. At least 38 ge- netic transferrin variants due to amino acid substitutions have been re- 1.1. General background ported [2]. Transferrin also occurs in different glycoforms due to variation in the oligosaccharide chains that can be bi-, tri-, and tetra- Transferrin is a glycoprotein synthesized in the hepatocytes that is antennary. Each antenna is usually terminated by a sialic acid residue, present in serum at a concentration of ~2.0–3.5 g/L [1]. Transferrin is the and the total number of sialic acids forms the basis for naming of the major Fe3+ transporter and consists of a single polypeptide chain of 679 glycoforms [3]. The transferrin glycoforms that are usually measurable in amino acids and has two binding sites for iron and two N-linked human serum are tetrasialotransferrin (~80%), pentasialotransferrin (~14%), trisialotransferrin (~4%), disialotransferrin (b 2%), and hexasialotransferrin (~1%) [4–6]. ☆ The IFCC Working Group on Standardization of Carbohydrate-Deficient Transferrin In the mid-1970s [7], the amount of disialotransferrin was shown to (WG-CDT). ⁎ Corresponding author. increase in response to chronic heavy alcohol consumption and de- E-mail address: [email protected] (J. Wielders). crease again on abstinence with a half-life of about 10 days. The elevated 1 Present chairman WG-CDT. value, which was accompanied by an increase in asialotransferrin, was http://dx.doi.org/10.1016/j.cca.2016.12.022 0009-8981/© 2016 Elsevier B.V. All rights reserved.  Author's Personal Copy 92 F. Schellenberg et al. / Clinica Chimica Acta 465 (2017) 91–100 termed “carbohydrate-deficient transferrin” (CDT) and suggested as a 2. Basis of CDT standardization specific biomarker for chronic heavy drinking [8]. A large number of studies have since reported on the utility of CDT as a more specific alco- 2.1. Definitions and terms hol biomarker [9–11] compared with traditional tests such as MCV, γ- GT, AST, and ALT. CDT levels related to various drinking levels, and iden- The HPLC cRMP is intended to quantify the relative concentration of tification of a few factors other than alcohol that may also elevate disialotransferrin in human serum, expressed as percentage of total disialotransferrin, have been documented. Today CDT is considered an transferrin, and calculated as the ratio of HPLC peak areas. Based on important objective indicator of sustained high alcohol consumption the IUPAC/IFCC format for designation of quantities in laboratory medi- and is used to support the diagnosis of alcohol abuse and dependence cine [34], the WG-CDT proposes to define transferrin in serum as the in medical and forensic settings [12–15], including driver's license with- system, and disialotransferrin as the component (analyte or substance drawal and reinstatement. of interest). The measurand is disialotransferrin measured by absor- Different measurement procedures have been used for quantifica- bance measurement of its iron-saturated glycan complex at 470 nm tion of CDT, including assays based on isoelectric focusing followed by [35] after separation of glycoforms by anion-exchange HPLC. The kind immunofixation [16], anion-exchange column chromatography follow- of property is “substance fraction”, as disialotransferrin is not an inde- ed by turbidimetry [17], anion-exchange HPLC with measurement of pendent molecule but a % fraction (the unit) of total serum transferrin. the iron-transferrin absorbance maximum at ~ 470 nm [18,19], latex The full measurand is called “P–Disialotransferrin/Transferrin; subst. particle-enhanced nephelometric immunoassay [20], and capillary elec- fraction”, according to IFCC C-NPU recommendations. trophoresis measuring the absorbance of the peptide bond at ~200 nm To distinguish between standardized and non-standardized CDT an- [21,22]. The transferrin glycoforms covered (i.e. the “CDT” analyte), and alytical procedures, values produced using standardized procedures (i.e. the way results are reported as either absolute or relative amounts, also calibrated against the HPLC cRMP using secondary calibrators approved differ between methods. Expressing CDT as percentage of total transfer- by the IFCC WG-CDT) are to be termed CDTIFCC. rin [19] was demonstrated to compensate for changes due to elevated (e.g. in iron deficiency) or lowered (e.g. in liver deficiency, iron over- 2.2. The standardization process load, and inflammatory diseases) serum transferrin concentrations, and is hence the current standard. Despite this, CDT results obtained Mass spectrometry (MS) has been used for characterization of trans- by different methods may differ significantly from each other, making ferrin glycoforms purified from human albumin and immunoglobulin- comparison with previous results or reference ranges error prone. Obvi- free serum by salt precipitation followed by affinity and ion-exchange ously, physicians, forensic experts, lawyers, and patients prefer directly chromatography [24]. This also provided information on the structure comparable results, irrespective of the method used. of transferrin glycoforms. An MS method for transferrin glycoform gly- This inconsistency prompted the International Federation of Clinical copeptides obtained by tryptic digestion has been suggested as an alter- Chemistry and Laboratory Medicine (IFCC) to form a working group on native to routine procedures for CDT measurement [36]. However, standardization of CDT measurement (WG-CDT). The CDT standardiza- because production and quantification of each transferrin glycoform tion work started in 2005 and aimed to define the analyte, select and val- with the needed purity is not yet solved, HPLC was proposed as the idate a reference measurement procedure (RMP) and reference materials, higher order available RMP [23]. The HPLC cRMP for CDT [19] fulfils and propose a reference interval using the RMP. In the first WG-CDT pub- the role as a RMP for certification of secondary calibrators and control lication [23], disialotransferrin (i.e. transferrin with one bi-antennary gly- materials, allowing for recalibration of all current CDT field methods can chain [24]) was recommended as the single analyte on which the to yield identical CDTIFCC results (see Section 7.4.). According to recent standardization is based, and the primary but not sole target for CDT mea- definitions [37], the combination of a cRMP and certified standards surement. It was further recommended that a well-defined qualitative fulfils the demands for method standardization. and quantitative HPLC method for transferrin glycoforms [19] was suit- able as a candidate RMP (cRMP), and that test results should be expressed 3. Principle, apparatus, chemicals and reagents as the relative concentration to total transferrin (% CDT). In a second pub- lication [25], the HPLC cRMP was demonstrated to produce reproducible 3.1. Principle of the HPLC candidate RMP results within a reference laboratory network, and a candidate reference material was found commutable and stable on storage. These findings The HPLC method is based on separation of the transferrin glycoforms were the basis for further development of the cRMP which should be suit- by anion-exchange chromatography and NaCl gradient elution [19]. The able for certification of secondary calibrators and control materials. In two global charge of each glycoform depends on the net charge of the amino subsequent WG-CDT publications [26,27], harmonization of currently acid chain [4], the number of iron molecules bound to the transferrin mol- employed CDT field methods was demonstrated possible and further ecule, and the number of negatively charged terminal sialic acid residues established, using the “toolbox” of the International Consortium for Har- on the glycan chains. The procedure is optimised for quantification of monization of Clinical Laboratory Results [28]. iron-saturated disialotransferrin in transferrin C homozygous subjects. Laboratory standards (e.g. ISO 15193 [29]) demand detailed infor- Accordingly, the transferrin glycoforms differ in net charge only based mation about the analytical procedure and performance characteristics on differences in sialic acid content. of a RMP. In this document, the HPLC cRMP for CDT [19] was further val- idated and demonstrated to fulfill the requirements of an IFCC RMP. 3.2. Apparatus 1.2. Scope of the HPLC candidate RMP The HPLC system should consist of a quaternary pump with a degasser for preparing the gradient, an autosampler suitable to inject The HPLC cRMP is designed to quantify CDT for clinical or forensic 200 μL, a specified commercially available anion-exchange column, a purposes in samples of human serum. Analytical interference may high-sensitivity UV–vis detector, and a data management system occur in the presence of genetic transferrin variants [30] (see Section allowing baseline peak integration with manual check. No special safety 5.3), but no interference of drugs therapies has been documented [31]. precautions are needed, other than usual for this type of equipment. It has also been reported that severe liver disease can obscure the mea- A typical set of equipment is obtained from Shimadzu (Kyoto, Japan) surement and interpretation of analytical results [32,33]. The HPLC containing an LC-10ADvp pump equipped with a low pressure gradient method is an optimised procedure and should be followed in every de- control valve FCV-10ALvp (mixing chamber 550 μL), a DGU-14A tail. The intended measurement interval is ~0.5%–16% disialotransferrin. degasser, and SIL-20AC autosampler. Absorbance detection at 470 nm  Author's Personal Copy F. Schellenberg et al. / Clinica Chimica Acta 465 (2017) 91–100 93 is performed with the UV–vis detector SPD-10AVvp. The UV–vis detec- [19]. Bilirubin and haemoglobin at respective maximum concentrations tor, which is critical, has the following specifications: cell path length of 150 μmol/L and 0.7 g/L do not cause marked interferences [19]. Hy- 10 mm, band width 10 nm, wave-length accuracy ±1 nm, wave-length perlipemia is not a source of potential interference, as the sample pre- precision ±0.5 nm, drift 0.1 mAU/h, and noise level 5–10 μAU. Other treatment includes delipidation. HPLC systems which have been found suitable are, for example, Agilent Serum samples are obtained by standard phlebotomy and standard 1100 and 1200, and Dionex UltiMate 3000. centrifugation procedure (e.g. 1500 g for 10 min). No special patient Separation of transferrin glycoforms is performed using a Source® preparation is needed. 15Q PE 4.6/100 anion-exchange chromatography column (GE Healthcare, Uppsala, Sweden) at room temperature (typically 20–25 °C). 4.2. Sample preparation 3.3. Chemicals Serum sample preparation includes complete iron saturation of transferrin using FeNTA, delipidation with dextran sulphate/CaCl2, and All chemicals are of analytical grade, and the water is of HPLC grade centrifugation. Iron saturation is obtained by mixing 100 μL serum (N18 MΏ). Chemical Abstract Service Registry Number (CAS), hazard with 20 μL FeNTA solution. Subsequently lipoproteins are precipitated class, H-, R- and S-phrases are reported where known. by adding 20 μL delipidation mixture. The sample is then gently mixed 1. Dextran sulphate sodium salt, Mw N 500.000, CAS 9011-18-1, harm- and stored at 4 °C for at least 4 h, or overnight. After centrifugation ful, R36-37-38, S 20-25-26-37-39 Sigma. (3500 g for 5 min), 100 μL of the clear supernatant is diluted with 400 μL water and transferred to HPLC sample vials. 2. Nitrilotriacetic acid trisodium salt monohydrate (NTA) (C6H6NNa3O6 • H2O), Mw = 275.1, CAS 18662-53-8, harmful, neither R- nor S- phrases. 4.3. Sample stability 3. 2-[Bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol (Bis-Tris) (C8H19NO5), Mw = 209.24, CAS 6975-37-0, harmful, nei- In centrifuged serum samples, the CDT level (disialotransferrin to ther R- nor S-phrases. total transferrin level in %) was stable for at least 5 days at room temper- 4. Calcium chloride dihydrate (CaCl2 • 2 H2O), Mw = 147.01, CAS ature [38]. At 4 °C, no significant change was observed after 10 weeks 10035-04-8, harmful, H319, R36. [36] or 3 months [27]. At − 20 °C and − 70 °C (Table 1), the stability 5. Iron(III) chloride hexahydrate (FeCl3 • 6 H2O), Mw = 270.3, CAS of frozen samples was 24 and 36 months, respectively [27], and proba- 10025-77-1, harmful, H302-315-318, R22-38-41. bly even much longer [39]. 6. Sodium hydroxide (NaOH), Mw = 40, CAS 12010-73-2, harmful, H314-318, R35. 5. Details of the procedure 7. Hydrochloric acid (HCl), Mw = 36.46, CAS 7647-01-0, harmful, H290-314-335, R34-37. 5.1. The HPLC method 8. Sodium chloride (NaCl), Mw = 58.44, CAS 7647-14-5, non harmful. 3.4. Reagents and HPLC mobile phases 5.1.1. Description of the HPLC method and data processing Separation of transferrin glycoforms is performed on a Source 15Q For separation of the transferrin glycoforms, column equilibration, PE 4.6/100 anion-exchange chromatography column at room tempera- and column cleaning, the following solutions are used [19]: ture (20–25 °C), using linear NaCl gradient elution in Bis-Tris buffer at a Mobile phase 1 (MP1) consists of 10 mmol/L Bis-Tris adjusted to flow rate of 1.0 mL/min. A 200-μL sample is injected on the column and pH 7.0 with 2 mol/L HCl. the column is rinsed after separation with 2 mol/L NaCl to eliminate any Mobile phase 2 (MP2) is 10 mmol/L Bis-Tris containing 200 mmol/L protein residues. After reconditioning with injection buffer, the system NaCl adjusted to pH 6.2 with 2 mol/L HCl. is ready for another sample injection. Detection and quantification of Mobile phase 3 (MP3) is 10 mmol/L Bis-Tris adjusted at pH 6.2 with the transferrin glycoforms is based on the absorbance maximum of 2 mol/L HCl. the iron-transferrin complex at 470 nm [35]. Typical chromatograms Mobile phase 4 (MP4) is a cleaning solution consisting of 2 mol/L for a low and a chronic heavy alcohol user are shown in Fig. 1. NaCl. HPLC peak integration and data processing are performed according After preparation, all mobile phases are filtered through a 0.45-μm to standard routines. The amount of each transferrin glycoform is quanti- filter and degassed. The mobile phases are stable at 4 °C for 3 months. fied separately by peak integration using a baseline mode from the start of Serum transferrin is iron-saturated using a FeNTA solution [19], pre- the disialotransferrin peak, or monosialotransferrin if visible, to the end of pared by dissolving 275 mg NTA and 270 mg FeCl3 in 90 mL water. The the hexasialotransferrin peak [19]. The transferrin glycoforms with more pH is adjusted to 7.0 with 1.0 mol/L NaOH, and water is added to a final than six sialic acid residues (heptasialo- and octasialotransferrin) are volume of 100 mL. The FeNTA solution is stored at 4 °C and is stable for present at very low concentrations, so the summed peak area for all 1 year. glycoforms from asialo- to hexasialotransferrin is directly proportional The delipidation mixture is prepared by mixing equal volumes of to the total transferrin concentration. When detectable, asialotransferrin 20 g/L dextran sulphate and 1.0 mol/L CaCl2. The solution is stored at is integrated as an isolated fraction (Fig. 1B). This calculation method im- 4 °C and is stable for at least 1.5 months [19]. plies that the molar absorptivity of all iron-saturated glycoforms should be the same (see Section 5.2). The disialotransferrin (CDT) result is 4. Sample expressed as % of total transferrin. According to the NEN 1047 norm and 4.1. Sample type Table 1 CDT (% disialotransferrin) values by the HPLC cRMP after prolonged storage of a sample at The HPLC method is intended for measurement of transferrin two different temperatures (* p b 0.05 vs time0). glycoforms in human serum samples. Heparin plasma also seems suit- Storage temperature (°C) Storage time (months) able [19, unpublished results] but would need additional validation. 0 3 6 15 24 36 EDTA plasma is not suitable, due to the existence of an additional asym- metric peak with a retention time similar to that of asialotransferrin, −20 °C 2,50 2,47 2,38 2,42 2,40 2,22* −70 °C 2,50 2,46 2,46 2,44 2,40 2,44 and citrate plasma causes a shoulder on the tetrasialotransferrin peak  Author's Personal Copy 94 F. Schellenberg et al. / Clinica Chimica Acta 465 (2017) 91–100 Fig. 1. HPLC chromatograms of serum from a healthy, light drinking control subject (A), the same sample after immunosubtraction of transferrin (A, insert), and from a chronic, heavy alcohol user (B) (data from [19]). the European Committee for Standardization, the rounding should be equations showing an almost perfect agreement between expected 0.01%. and measured concentrations (batch 1: y = 1.0049x–0.0507, r2 = 0.9998; batch 2: y = 0.9982x–0.0454, r2 = 0.9996). The slope of the re- 5.1.2. Structure of analytical series gression equation was not significantly different from 1.00, and the in- In order to verify the performance of the HPLC equipment, control tercept was not significantly different from the origin. This proves the samples should be analyzed at the beginning and end of each analytical suitability of using a relative concentration (% CDT). series. A matrix-free ready-to-inject test solution (e.g. ClinTest® CDT Using secondary calibrators: Purified transferrin glycoforms obtained test solution from Recipe, Munich, Germany) allows checking the reten- by preparative chromatography were assessed by the HPLC cRMP and tion time of all glycoforms, peak shapes, and concentrations, without a MS analysis [24]. Experimental secondary calibrators were prepared by possible bias due to improper sample preparation. Analytical or reagent spiking disialotransferrin at five levels to a human serum pool with a blank samples are not needed. disialotransferrin level of 1.1%, resulting in final concentrations of 1.1– The analysis can start, when all glycoforms are correctly separated, 9.0% disialotransferrin. The HPLC analysis of disialotransferrin of these elute at the expected retention times, and peak shapes (or calculated spiked sera (Fig. 3) showed a high degree of correlation with the expected number of plates) are satisfactory. Using the described sequence and (calculated) values (y = 0.941x–0.19, r2 = 0.999). gradient of mobile phases ensures no carry-over between samples. These two experiments provide indirect yet very strong evidence For routine determination of % disialotransferrin by the HPLC cRMP, that the different iron-saturated transferrin glycoforms have the same no replicate analysis is needed due to the low imprecision; the effect of molar absorption coefficients. Therefore, the relative peak area of each single measurement on the imprecision is taken into account in the item fraction corresponds to the relative amount of this glycoform to total devoted to measurement uncertainty. For measurement of reference transferrin and indicates that the HPLC cRMP does not require primary materials, using three replicates is advised. calibrators. Peak areas of iron-saturated serum transferrin glycoforms can be used as the measurand. Consequently, target values of secondary 5.1.3. Quality control calibrators derived from peak area ratios can be assigned, by analysis in At least one control sample is required to verify the performance of the reference laboratory network using the cRMP. disialotransferrin concentration measurement. However, a two-level set of internal quality control samples is recommended, of which one 5.3. Limitations of the HPLC method – special cases with a disialotransferrin concentration near the upper limit of the refer- ence interval. CDT controls are commercially available products or High levels of bilirubin (N 150 μmol/L) and haemolysis (Hb N 0.7 g/L), home-made serum pools stored at − 20 °C or below. Participation in alone or in combination with a low serum transferrin concentration, can an external quality assurance scheme (EQAS) is mandatory, according cause severe HPLC baseline sloping and loss of accuracy of peak area in- to ISO 15189 [40]. tegration for the minor transferrin glycoforms. Difficulties in the separa- tion of glycoforms, and consequently in peak integration, may occur in 5.2. Calibration of the HPLC method by gravimetry genetic transferrin variant samples, and in cases of acquired changes in transferrin such as severe liver disease. Elevated CDT levels may Using primary calibrators: Isolated disialo- and trisialotransferrin occur in rare cases of inherited metabolic diseases known as congenital and a mixture of tetrasialo- and pentasialotransferrin [24] were obtain- disorders of glycosylation (CDG), and in fructosemia and galactosemia ed from Bio-Rad Laboratories GmbH (Munich, Germany). A set of pri- [41], whereas common disorders (e.g. diabetes) [42] or medications mary calibrators was prepared using transferrin-free serum (obtained (e.g. antiepileptic drugs) [31] do not interfere. by immunosubtraction [19]) spiked with the isolated glycoforms up to a total transferrin concentration of 3.0 g/L. These primary calibrators 5.3.1. Genetic variants of transferrin contained a fixed amount of trisialotransferrin (3–4% of total transfer- In Caucasians, transferrin-C is the wild type allele with a prevalence rin) and six equidistant concentrations (0–10%) of disialotransferrin. of ~95%. The subtypes C1 (wt) and C2 (Pro570Ser) have allele frequen- The mixtures were prepared by gravimetry to increase precision. cies of about 75 and 19%, respectively [43], resulting in a frequency of Two batches of these six primary calibrators were analyzed by the about 59% for the C1-1 and 23% for the C2-1 phenotypes as the main HPLC cRMP in two independent laboratories, using different lot num- forms. Heterozygous transferrin-BC and transferrin-CD forms are pres- bers of the anion-exchange column and buffer batches (Fig. 2). A linear ent in ~ 1% of Caucasian populations [30,44]; these heterozygotes are, relationship between the target (gravimetry) and measured (HPLC) however, more common in certain Asian [45,46], African, and South disialotransferrin concentrations was observed, the regression American populations [44].  Author's Personal Copy F. Schellenberg et al. / Clinica Chimica Acta 465 (2017) 91–100 95 Fig. 2. Measurement of CDT (% disialotransferrin) by the HPLC cRMP in two batches of primary calibrators at six levels of disialotransferrin in the same mixture of purified pentasialo-, tetrasialo- and trisialotransferrin prepared by weighing (by courtesy of Bio-Rad Laboratories GmbH, Munich, Germany). The HPLC cRMP is optimised for measurement of homozygous trans- 5.3.2. High trisialotransferrin and “di-tri bridging” ferrin-C subjects. Heterozygous variants like transferrin-BC and trans- Some serum specimens show an elevated amount (N 6%) of ferrin-CD present “abnormal” HPLC profiles with a superimposed trisialotransferrin. The concentration of this glycoform is not dependent double pattern of glycoforms [30], resulting respectively from transfer- on alcohol intake and the increase is typically accompanied by a mea- rin B (elutes after C in the HPLC method) and C, or transferrin C and D surable quantity of monosialotransferrin [51]. (elutes before C) glycoform chromatographic separation. Exact In other samples, baseline separation between disialo- and quantification of “total” disialotransferrin is not possible for transfer- trisialotransferrin is not achieved, due to the presence of transferrin rin-BC or transferrin-CD variants, due to overlay of the individual al- fractions eluting in-between the two (Fig. 4). This so-called “di-tri lele peak patterns. Generally the presence of a genetic variant is bridging” phenomenon is not caused by the presence of a genetic vari- easily recognized by a double peak of the major tetrasialotransferrin ant and has been described for C1-C1, C1-C2, and C1-C3 allele combina- form (e.g. one originating from the C allele and one from the B or D tions [32]. The di-tri bridging is typically associated with severe liver allele). Transferrin-C subtypes, C1, C2, and most C1-C2 are suitable disease (e.g. cirrhosis) [32,33], and often occurs together with an elevat- for quantification with the HPLC cRMP. However, for a recently de- ed trisialotransferrin level. The di-tri bridging hampers accurate quanti- scribed C2 subtype [47], HPLC proved to be not suitable. Transferrin fication of disialotransferrin by the HPLC cRMP, and is due to increased mutations can affect also the oligosaccharide-chain binding sites at branching and fucosylation of the oligosaccharide chains [52]. Asn432 and Asn 630. Thus, a mutation at one of these sites abolishes N-glycosylation and the mutated transferrin molecule can bear only one oligosaccharide chain, which increases the proportion of 5.3.3. CDG syndrome disialotransferrin. This has been described both for Asn432 [48,49] In CDG, mutations in the genes coding for enzymes involved in the and Asn630 [49,50]. synthesis of glycans (oligosaccharide chains) lead to a variety of Fig. 3. Measurement of CDT (% disialotransferrin) by the HPLC cRMP in a five-level secondary CDT reference material (determined; y axis) prepared by spiking a low % disialotransferrin serum with weighted amounts of purified disialotransferrin (data from [24]), compared to the target values (x axis).  Author's Personal Copy 96 F. Schellenberg et al. / Clinica Chimica Acta 465 (2017) 91–100 of the HPLC cRMP is considered sufficient to avoid any clinical or foren- sic consequence related to the quantification of CDT. 6.2. Linearity Linearity and serum transferrin concentration: CDT was measured in a serum sample containing a very high transferrin concentration (~ 5 g/L) at different dilution ratios (10–75% serum) with water [19], and the measured sum of peak areas were compared to the calculated value. The correlation was linear (r2 = 0.999) and the relative amount of disialotransferrin unaffected by dilution. The HPLC cRMP is therefore considered linear over a transferrin concentration range of at least 0.5– 5 g/L, thereby well covering the reference interval for serum transferrin (2.0–3.5 g/L). This enables use of the HPLC cRMP also in samples with an abnormally low transferrin concentration (b1.4 g/L). In addition, using variable sample injection volumes in the range 50– 250 μL did not affect the relative quantification of disialotransferrin (% CDT). Even a 10-μL injection volume has been used in CDG testing on newborns, where the serum sample volume may be limited [51]. Linearity and disialotransferrin content: The linearity of the HPLC cRMP was assessed by measuring CDT in three sets of serum samples. Each set was constituted by mixing (from 9:1 to 1:9, v/v) a high CDT sam- ple (16.30%, 16.28%, and 15.43% disialotransferrin, respectively) with a low CDT sample (1.02%, 1.02%, and 1.07%, respectively). The correlation Fig. 4. Transferrin glycoform pattern in serum of a cirrhotic patient showing a loss of between calculated and measured values was high (r2 = 0.999), and chromatographic resolution of disialotransferrin from trisialotransferrin (“di/tri the regression equation (CDT[measured] = CDT[calculated] × 0.997 [95%CI bridging”) due to fucosylation of transferrin. Genotyping showed a usual homozygous 0.985–1.010] + 0.017 [95%CI −0.096–0.131]) (Fig. 5) showed the equiv- C1 type (data from [32]). alence between expected and measured concentrations. abnormal glycosylation patterns of transferrin, depending on the en- 6.3. Imprecision zyme affected. The transferrin glycoform profiles obtained in HPLC anal- ysis are typical of these diseases, often leading to highly elevated The reproducibility of the HPLC cRMP was assessed by repeated mea- disialotransferrin values (N20%) [53], and may be used for preliminary surements of pooled serum samples, according to the CLSI EP05-A2 proce- screening of CDG into Type-I or Type-II. It should be noted that CDG test- dure [55]. Two pools (low and high CDT) were prepared and analyzed in ing and diagnosis is usually done in newborns, because many CDG pa- duplicate in two separate runs over 20 days. No outlier was detected dur- tients suffer from mental and psychomotor retardation that are ing this experiment. As indicated in Table 2, the within laboratory impre- evident already from an early age. cision was b5%, which is in accordance with previous results [19]. This result was confirmed in another experiment by repeated measurement 6. Validation of the HPLC candidate RMP (n = 34) on non-consecutive days of two sets of pooled sera (a low CDT pool with 1.20%, and a high CDT pool with 3.07% disialotransferrin), yield- 6.1. Detection limit ing coefficients of variation (CV) of 7.0% and 2.9%, respectively. The inter-laboratory imprecision has been quantified in a laboratory Low end-precision performance characteristics at the detection limit network running the HPLC cRMP, both with frozen and lyophilized sam- are very important if the result may have clinical or forensic signifi- ples. Six ring trials have been conducted from 2004 to 2013 and, as ex- cance, which does not apply in the case of very low CDT results. Medical pected, the imprecision decreased when the disialotransferrin or forensic CDT requests are typically used in case of suspected chronic concentration in the sample increased. The inter-laboratory CV ranged heavy alcohol use, or for routine company health care testing [54], and from 5.6% for low CDT samples to 3.7% for high CDT (elevated) values are especially meaningful when the CDT level exceeds the upper level [25,27]. These data were compared to the results from an EQAS in of the reference interval of 0.6–1.7% (see Section 8). 2013–2015 (Equalis, Sweden). Based on 28 test results from 15 to 20 The limit of detection (LOD) and the lower limit of quantification laboratories using single measurement of fresh serum samples, the (LLOQ) of the method should be well below the upper level, and prefer- mean inter-laboratory CV was 5.8% at a mean CDT level of 2.3%. ably also below the lower level, of the reference interval, to ensure reli- ability of elevated results. According to the EP17-A2, the LOD = mean of 6.4. Measurement uncertainty blanks + 1.645 SDblank + 1.645 SDlow sample leading to a calculated LOD of 0.16% disialotransferrin, which is close to the empirically [19] esti- Both the ISO 15189 and ISO 15193 require determination of mea- mated value of 0.10%. For the LLOQ, it has been proposed to use 10 surement uncertainty, when reporting an analytical result. The mea- SDlow sample which means an LLOQ of 0.8% disialotransferrin, but the fac- surement uncertainty concept is defined in the “Guide to the tor 10 is arbitrary. We instead prefer to use 5 SDlow sample which means expression of uncertainty measurement” (GUM) [56] and is briefly an LLOQ of 0.4% disialotransferrin. Another approach is a calculation of summarized below. In this model, the composite uncertainty (u) can the LLOQ based on imprecision. From the results presented in Section be calculated [57], using IQC, EQC, and intra-individual biological varia- 6.3, it is estimated that the LLOQ at a CV of 10% is well below the level tion data by the following equation: of 1.0% disialotransferrin.
It should be noted that these analytical thresholds depend on the ucomposite ¼ √ u2 IQC þ u2 EQC þ u2 intra‐individual transferrin concentration in the sample and the performance character- istics of the UV–vis detector. As very few disialotransferrin values below where u2(IQC) is the variance of internal quality control, u2(EQC) the var- 0.5% are observed when running human sera, the analytical sensitivity iance related to accuracy and dispersion of external quality control  Author's Personal Copy F. Schellenberg et al. / Clinica Chimica Acta 465 (2017) 91–100 97 Fig. 5. Measurement of CDT (% disialotransferrin) using the HPLC cRMP in serial dilutions of high CDT samples with low CDT samples (determined; y axis), compared to the calculated values (target; x axis) (Pearson's correlation coefficient r = 0.999). values, and u2(intra-individual) the intra-individual variance. The unit of u is calculated, using a k-value of 2 (95% CI), resulting in a U value of the same as for the measured analyte [57], hence % in case of CDT mea- 0.312% at a CDT concentration around 1.7%. surement. The u2(IQC) value was calculated from the repeated CDT pool Another component of uncertainty is the accuracy of the measure- measurements and found to be 0.0076% (considered as the intra-labora- ment, defined by a bias (E) to the “true” value. In a given laboratory tory variance). The u2(EQC) value was calculated from the 28 EQAS re- using a given procedure, the true value is expressed as follows: sults mentioned above in Section 6.3 (CDT from 1.96% to 3.75%, mean CDT(true) = CDT(measured) + E ± U at 95% CI. The HPLC cRMP has been 2.17%) and was 0.0097% (considered as the inter-laboratory variance). developed as a standardization procedure and thus no bias factor E For calculating the u2(intra-individual) variance, 17 subjects with repeat- has to be applied. Therefore, a CDT result (in %) obtained can be ed CDT measurements (n = 2–4) and at least one month between two expressed as follows: CDT(true) = CDT(measured) ± 0.31 at 95% CI (see consecutive measurements were used. Patients were in “alcohol steady also Section 8). state” conditions during the study period and had CDT values in the ~1– 2% range (overall mean value 1.72%). The CV of each set of two consec- 6.5. Analytical specificity utive measurements for each patient was calculated, showing a mean CV of 6.5%. This result was considered as the intra-individual variation After incubating serum samples with anti-transferrin antiserum (i.e. and allowed the calculation of u2(intra-individual) found at 0.0070%. The immunosubtraction), no peaks were observed in the HPLC chromato- resulting value of the composite uncertainty u(composite) was then calcu- gram at 470 nm (see Fig. 1A, insert) [19]. This demonstrated the speci- lated and found to be 0.156%. ficity of the absorbance measurement of transferrin glycoforms by the According to the NIST guideline 1297 [58], “what is often required is HPLC cRMP in serum samples (see Section 4.1). a measure of uncertainty that defines an interval about the measure- ment result y within which the value of the measurand Y is confidently 7. Method comparison and standardization believed to lie”. The measure of uncertainty intended to meet this re- quirement is termed expanded uncertainty (U). The expanded uncer- 7.1. Capillary electrophoresis methods tainty is obtained by multiplying the composite uncertainty with a coverage factor k [58]. A 95% confidence interval (CI) corresponds to a Three capillary electrophoresis (CE) procedures for CDT have been k-value of 1.96, and a 99% CI to a k-value of 2.58. compared with the HPLC cRMP, following the manufacturer's instruc- The ISO 15193 advocates reporting clinical results with a measure- tions for use. The results produced by the CEofix™ kit (Analis, Suarlee, ment uncertainty. For forensic purposes, a CI of either 95% or 99% is re- Belgium) [61] (CE1) run on a mono-capillary electrophoresis system quired [59,60], corresponding to a coverage factor of 2 or 3. The were highly correlated (r2 = 0.972) with those obtained by the HPLC expanded uncertainty for CDT measurement with the HPLC cRMP was cRMP. The Capillarys (Sebia, Evry, France) multi-capillary method Table 2 Repeatability, day-to-day, and within-laboratory imprecision of CDT (% disialotransferrin) measurement with the HPLC cRMP at two different levels, according to the CLSI EP05-A2 protocol. Sample Mean CDT value (% disialotransferrin) Repeatability Between day Within laboratory SD Imprecision SD Imprecision SD Imprecision Low pool 1,18% 0,046 3,90% 0,026 2,20% 0,053 4,50% High pool 3,05% 0,045 1,50% 0,028 0,90% 0,053 1,80%  Author's Personal Copy 98 F. Schellenberg et al. / Clinica Chimica Acta 465 (2017) 91–100 (CE2) results were also highly correlated with the HPLC cRMP [56,57], 7.4. Standardization procedure although the CE2 method produces consistently lower CDT results (r2 = 0.996; CE2 = 1.074 × HPLC – 0.46 [62], and r = 0.986; CE2 = The wide range of upper reference limit values for “CDT” given by 0.968 × HPLC − 0.248 [63]). Also the recently introduced V8 CDT (Hel- the manufacturers of commercial assays (from 1.3% for Sebia [63] to ena Biosciences, Tyne and Wear, UK) procedure (CE3) produced consis- 2.35% for Siemens [20]), and the variable regression equations given tently lower CDT results than the HPLC cRMP (r2 = 0.979; CE3 = above, clearly demonstrated the urgent need for standardization of 0.854 × HPLC − 0.03). Similar results (r2 = 0.941; CE3 = CDT measurement. The HPLC cRMP was shown to yield homogeneous 0.84 × HPLC + 0.03) were obtained in another study [64]. and stable results over time in an international network of six reference laboratories, using frozen serum samples as control materials [25–27]. CDT was measured by the network laboratories in a five-level set of cal- 7.2. HPLC methods ibrators produced at the MCA Laboratory (Queen Beatrix Hospital, Winterwijk, The Netherlands) to assess target values. The manufac- Two commercial HPLC assays for CDT have been evaluated and com- turers of the above mentioned commercial CDT procedures then mea- pared with the HPLC cRMP. The %CDT-by-HPLC method (Bio-Rad, Her- sured CDT using their standard procedure. The inter-method CV was cules, CA) [65] run on an Agilent system showed a high correlation initially 8.8% but decreased to 3.4% after mathematical re-calibration. with the HPLC cRMP but produced slightly lower results (r2 = 0.998; A set of 66 serum samples (CDT range 1.10–6.10%) was analyzed by %CDT-by-HPLC = 0.987 × cRMP − 0.205). seven commercial CDT procedures and the HPLC cRMP, before and after A study carried out to evaluate the correlation between the standardization. The correlation and the concordance between different ClinRep® CDT (Recipe, Munich, Germany) and the cRMP HPLC proce- procedures were tested, using the Pearson's test. The concordance cor- dures, using 77 serum samples covering a wide range of CDT concentra- relation coefficient ρc evaluates the degree to which pairs of observa- tions (0.5–16.1%), showed good agreement between the methods (r2 = tions fall on the 45° line through the origin [67]. It comprises a 0.993; ClinRep CDT = 1.017 × cRMP [95% CI 0.997–1.037] − 0.007 [95% measurement of the coefficient of correlation ρ, which reflects the pre- CI −0.196–0.064]) (Fig. 6). The Cusum test also indicated a linear agree- cision, and the calculation of a bias correction factor Cb, which is a mea- ment, and a paired t-test confirmed the absence of a significant differ- sure of accuracy. Finally, the concordance correlation coefficient is ence between the results obtained by both HPLC methods, thereby calculated using the formula: ρc = ρ × Cb. confirming the results of a previous study [66]. A scale of interpretation of the concordance correlation coefficient values ρc was used, to estimate agreements between the HPLC cRMP and each field method. The strength of agreement was classified accord- 7.3. Immunochemical method ing to McBride [68] into four categories: “almost perfect” (ρc N 0.99), “substantial” (0.95 N ρc N 0.99), “moderate” (0.90 N ρc N 0.95) and The CDT results of a homogeneous latex-enhanced immuno inhibi- “poor” (ρc b 0.90). In head-to-head comparison of the seven commercial tion nephelometric assay, based on a monoclonal antibody directed procedures, only nine of 21 pairs were considered as “substantial” and against CDT (N Latex CDT; Siemens, Marburg Germany) [20], were com- one as “almost perfect”, whereas, after standardization, the number of pared with those of the HPLC cRMP. Although the immunoassay is based “substantial” pairs increased to 19 (Table 3). When comparing the com- on a different analytical principle than HPLC, the comparison showed mercial methods with the HPLC cRMP, the number of “moderate” agree- good correlation but a systematic bias and slope (r = 0.989; N Latex ments decreased from four to zero and the number of “almost perfect” CDT = 0.838 × cRMP CDT + 0.354). The N-Latex CDT immunoassay ap- increased from zero to three, whereas four pairs had a “substantial” pears not to be influenced by the common transferrin variants [20]. agreement both before and after standardization. These data clearly Fig. 6. Measurement of CDT (% disialotransferrin) values by the Recipe® Clin Rep HPLC method and the HPLC cRMP in 77 patient samples.  Author's Personal Copy F. Schellenberg et al. / Clinica Chimica Acta 465 (2017) 91–100 99 Table 3 Strength of agreement, according to McBride [68], in head-to-head comparison of seven commercial procedures themselves, and between these procedures and the HPLC cRMP, before and after standardization using the secondary calibrators. Comparisons Agreement categories Strenght of agreement Before standardization After standardization Between commercial procedures Poor (b0.90) 5 0 Moderate (0.90–0.95) 6 1 Substantial (0.95–0.99) 9 19 Almost perfect (N0.99) 1 1 Between commercial procedures and the RMP Poor (b0.90) 0 0 Moderate (0.90–0.95) 3 0 Substantial (0.95–0.99) 4 4 Almost perfect (N0.99) 0 3 demonstrated that standard materials calibrated by the HPLC cRMP can Conflict of interest be used for standardization of the commercial CDT procedure results [27]. The authors declare that they have no competing financial interests. 8. Reference interval Acknowledgments The 2.5–97.5 percentile reference interval for “social drinkers” was We thank Prof Dr. Torsten Arndt (former WG-CDT member) for initially estimated to be 0.67–1.67% [19]. Later, a set of 245 samples valuable contributions, Carla Siebelder for skilful assistance in preparing from a collaborative WHO-ISBRA study on different populations [9], and handling of controls and calibrators for the CDT network laborato- and 97 control subjects from the Karolinska University Hospital (Stock- ries, and Dr. André Naus for performing the Bhattacharya calculations. holm, Sweden), were used to calculate a reference interval according to CLSI C28-A3. In the 342 samples, the mean CDT value was 1.12% and the References calculated central 95% reference interval plus CI was 0.71% disialotransferrin (90% CI 0.68–0.75) to 1.60% (90% CI 1.51–1.70%). No [1] G. de Jong, J.P. van Dijk, H.G. van Eijk, The biology of transferrin, Clin. Chim. Acta 190 (1990) 1–46. clinically significant sex difference in CDT was noted for non-drinkers [2] J. Caslavska, J. Joneli, U. Wanzenried, et al., Determination of genetic transferrin var- [9,46]. However, pregnancy was demonstrated to increase CDT mea- iants in human serum by high-resolution capillary zone electrophoresis, J. Sep. Sci. sured with the HPLC cRMP [69], yielding results in the high normal or 37 (2014) 1663–1670. slightly elevated range at the end of pregnancy (i.e. in the third trimes- [3] W.L. van Noort, G. de Jong, H.G. van Eijk, Purification of isotransferrins by concanav- alin A sepharose chromatography and preparative isoelectric focusing, Eur. J. Clin. ter), and returning to normal within 2–3 months post-partum. Chem. Clin. Biochem. 32 (1994) 885–892. An independent way to obtain a reference interval is the use of [4] H.G. van Eijk, W.L. van Noort, G. de Jong, J.F. Koster, Human serum sialotransferrins Bhattacharya calculation [70]. This procedure allows the determination in diseases, Clin. Chim. Acta 165 (1987) 141–145. [5] H.G. van Eijk, W.L. van Noort, The analysis of human serum transferrins with the of reference values from unselected populations with an underlying PhastSystem: quantitation of microheterogeneity, Electrophoresis 13 (1992) Gaussian or Gamma distribution of “normal” values. In a set of 5000 pa- 354–358. tient CDT results obtained in Malmö (Sweden) with the HPLC cRMP, the [6] O. Mårtensson, A. Härlin, R. Brandt, K. Seppä, P. Sillanaukee, Transferrin isoform dis- tribution: gender and alcohol consumption, Alcohol. Clin. Exp. Res. 21 (1997) reference interval was found to be 0.77–1.72% (central 95%) using a 1710–1715. Gamma distribution, or 0.68–1.65% using a Gaussian distribution [7] H. Stibler, C. Allgulander, S. Borg, K.G. Kjellin, Abnormal microheterogeneity of (Wielders and Naus, 2009, unpublished). Combining these data, it was transferrin in serum and cerebrospinal fluid in alcoholism, Acta Med. Scand. 204 (1978) 49–56. concluded that the upper reference limit (URL) for the HPLC cRMP [8] H. Stibler, Carbohydrate-deficient transferrin in serum: a new marker of potentially will be 1.7%, which agrees with that originally reported [19]. harmful alcohol consumption reviewed, Clin. Chem. 37 (1991) 2029–2037. For forensic applications, the use of measurement uncertainty is rec- [9] J. Bergström, A. Helander, Clinical characteristics of carbohydrate-deficient transfer- rin (%disialotransferrin) measured by HPLC: sensitivity, specificity, gender effects, ommended [59] and fulfils the recommendations of the ISO 15189. and relationship with other alcohol biomarkers, Alcohol Alcohol. 43 (2008) Using the expanded uncertainty calculated above at 95% CI, a true CDT 436–441. value of 1.7% could be, at most, measured as 2.0% in a single [10] T. Arndt, Carbohydrate-deficient transferrin as a marker of chronic alcohol abuse: a measurement. critical review of preanalysis, analysis and interpretation, Clin. Chem. 47 (2001) 13–27. This 2.0% cut-off value guaranties a high level of specificity and could [11] F. Bortolotti, G. De Paoli, F. Tagliaro, Carbohydrate deficient transferrin (CDT) as a be considered in forensic medicine as the highest numerical value to be marker of alcohol abuse: a critical review of the literature 2001-2005, J. Chromatogr. expected for CDT, using the HPLC cRMP in occasional or non-drinkers B 841 (2006) 96–109. [12] J.R. Delanghe, M.L. De Buyzere, Carbohydrate deficient transferrin and forensic med- considering all sources of imprecision. icine, Clin. Chim. Acta 406 (2009) 1–7. [13] F. Bortolotti, M.T. Trevisan, R. Micciolo, et al., Re-assessment of the cut-off levels of 9. Conclusion and approval by the IFCC CDT for automated immunoassay and multi-capillary electrophoresis for application in a forensic context, Clin. Chim. Acta 413 (2013) 1–4. [14] T.M. Maenhout, G. Baten, M.L. De Buyzere, J. Delanghe, Carbohydrate deficient trans- The IFCC initiative for standardization of the alcohol biomarker CDT ferrin in a driver's license regranting program, Alcohol Alcohol. 47 (2012) 253–260. [71] is completed with the description and validation of a candidate ref- [15] G. Gough, L. Heathers, D. Puckett, et al., The utility of commonly used laboratory tests to screen for excessive alcohol use in clinical practice, Alcohol. Clin. Exp. Res. erence measurement procedure, as presented in this work. The proce- 39 (2015) 1493–1500. dure and its use are described in detail, along the general guidelines of [16] F. Schellenberg, J. Weill, Serum desialotransferrin in the detection of alcohol abuse. ISO 15189 and ISO 15193, fulfilling the demands for a RMP and allowing Definition of a tf index, Drug Alcohol Depend. 19 (1987) 181–191. [17] R.F. Anton, C. Dominick, M. Bigelow, C. Westby, Comparison of Bio-Rad %CDT TIA for standardization of the commercial CDT assays. As such, this mea- and CDTect as laboratory markers of heavy alcohol use and their relationships surement procedure is approved by the IFCC Working Group on CDT, with gamma-glutamyltransferase, Clin. Chem. 47 (2001) 1769–1775. by the Executive Committee of the Scientific Division of the IFCC and ac- [18] J.-O. Jeppsson, H. Kristensson, C. Fimiani, Carbohydrate-deficient transferrin quanti- cepted by the IFCC National Societies. From now on it will be known as fied by HPLC to determine heavy consumption of alcohol, Clin. Chem. 39 (1993) 2115–2119. the IFCC-RMP for CDT, while CDT results standardized according to this [19] A. Helander, A. Husa, J.-O. Jeppsson, Improved HPLC method for carbohydrate-defi- RMP should be indicated as CDTIFCC. cient transferrin in serum, Clin. Chem. 49 (2003) 1881–1890.  Author's Personal Copy 100 F. Schellenberg et al. / Clinica Chimica Acta 465 (2017) 91–100 [20] J.R. Delanghe, A. Helander, J.P.M. Wielders, et al., Development and multicentre eval- [46] B. Song, J. Zhu, J. Wu, et al., Determination of carbohydrate-deficient transferrin in a uation of the N Latex CDT direct immunonephelometric assay for serum carbohy- Han Chinese population, BMC Biochem. 15 (2014) 5. drate-deficient transferrin, Clin. Chem. 53 (2007) 115–121. [47] H.K. de Wolf, K. Huijben, M. van Wijnen, M. de Metz, J.P.M. Wielders, A novel C2 [21] F. Bortolotti, G. De Paoli, J.P. Pascali, F. Tagliaro, Fully automated analysis of carbohy- transferrin variant interfering with the analysis of carbohydrate-deficient transfer- drate-deficient transferrin (CDT) by using a multicapillary electrophoresis system, rin, Clin. Chim. Acta 412 (2011) 1683–1685. Clin. Chim. Acta 380 (2007) 4–7. [48] A. Grahn, P. Bengtson, E. Eklund, J. Asin-Cayuela, A novel mutation on the transferrin [22] B. Wuyts, J.R. Delanghe, The analysis of carbohydrate-deficient transferrin, marker of gene abolishes one N-glycosylation site and alters the pattern of transferrin iso- chronic alcoholism, using capillary electrophoresis, Clin. Chem. Lab. Med. 41 (2003) forms, mimicking that observed after alcohol consumption, Clin. Biochem. 49 739–746. (2016) 511–513. [23] J.-O. Jeppsson, T. Arndt, F. Schellenberg, et al., Toward standardization of carbohy- [49] M. Guillard, Y. Wada, A. Hansikova, et al., Transferrin mutations at the glycosylation drate-deficient transferrin (CDT) measurement: I. Analyte definition and proposal site complicate diagnosis of congenital disorders of glycosylation type I, J. Inherit. of a candidate reference method, Clin. Chem. Lab. Med. 45 (2007) 558–562. Metab. Dis. 34 (2011) 901–906. [24] W. Oberrauch, A.C. Bergman, A. Helander, HPLC and mass spectrometric character- [50] A. Zühldorf, M. Said, C. Seger, et al., It is not always alcohol abuse – a transferrin var- ization of a candidate reference material for the alcohol biomarker carbohydrate-de- iant impairing the CDT test, Alcohol Alcohol. 51 (2016) 148–153. ficient transferrin (CDT), Clin. Chim. Acta 395 (2008) 142–145. [51] J.P. Bergström, A. Helander, Influence of alcohol use, ethnicity, age, gender, BMI and [25] A. Helander, J.P.M. Wielders, J.-O. Jeppsson, et al., Toward standardization of carbo- smoking on the serum transferrin glycoform pattern: implications for use of carbo- hydrate-deficient transferrin (CDT) measurements: II. Performance of a laboratory hydrate-deficient transferrin (CDT) as alcohol biomarker, Clin. Chim. Acta 388 network running the HPLC candidate reference measurement procedure and evalu- (2007) 59–67. ation of a candidate reference material, Clin. Chem. Lab. Med. 48 (2010) 1585–1592. [52] E. Landberg, E. Åström, B. Kådegal, P. Påhlsson, Disialo-trisialo bridging of transferrin [26] C. Weykamp, J.P. Wielders, A. Helander, et al., Toward standardization of carbohy- is due to increased branching and fucosylation of the carbohydrate moiety, Clin. drate-deficient transferrin (CDT) measurements: III. Performance of serum native Chim. Acta 414 (2012) 58–64. and serum spiked with disialotransferrin proves that harmonization of CDT assays [53] A. Helander, J. Bergström, H.H. Freeze, Testing for congenital disorders of glycosyla- is possible, Clin. Chem. Lab. Med. 51 (2013) 991–996. tion by HPLC measurement of serum transferrin glycoforms, Clin. Chem. 50 (2004) [27] C. Weykamp, J.P.M. Wielders, A. Helander, et al., Harmonization of measurement re- 954–958. sults of the alcohol biomarker carbohydrate-deficient transferrin by use of the tool- [54] U. Hermansson, A. Helander, L. Brandt, A. Huss, S. Rönnberg, Screening and brief in- box of technical procedures of the international consortium for harmonization of tervention for risky alcohol consumption in the workplace: results of a 1-year ran- clinical laboratory results, Clin. Chem. 60 (2014) 945–953. domized controlled study, Alcohol Alcohol. 45 (2010) 252–257. [28] W. Greg Miller, G.L. Myers, M. Lou Ganzer, et al., Roadmap for harmonization of clin- [55] CLSI, EP05-A2, Evaluation of Precision Performance of Quantitative Measurement ical laboratory measurement procedures, Clin. Chem. 57 (2011) 1108–1117. Methods, CLSI, Wayne, PA, USA, 2004. [29] ISO 15193, In vitro diagnostic medical devices – measurement of quantities in sam- [56] Joint Commitee for Guides in Metrology. Evaluation of measurement data - Guide to ples of biological origin, Requirement for Content and Presentation of Reference the expression of uncertainty in measurement (GUM). JCGM 100:2008. http:// Measurement Procedures, 2009. www.bipm.org/utils/common/documents/jcgm/JCGM_100_2008_E.pdf [30] A. Helander, G. Eriksson, H. Stibler, J.-O. Jeppsson, Interference of transferrin isoform [57] European Diagnostic Manufacturers Association, EDMA Position Paper – Estimation types with carbohydrate-deficient transferrin quantification in the identification of of Uncertainty of Measurement in Medical Laboratories, www.edma-ivd.be2006/09. alcohol abuse, Clin. Chem. 47 (2001) 1225–1233. [58] NIST Technical Note 1297, Guidelines for Evaluating and Expressing the Uncertainty [31] J.P. Bergström, A. Helander, HPLC evaluation of clinical and pharmacological factors of NIST Mseasurement Results, http://www.nist.gov/pml/pubs/tn1297/upload/ reported to cause false-positive carbohydrate-deficient transferrin (CDT) levels, tn1297s.pdf. Clin. Chim. Acta 389 (2008) 164–166. [59] R.G. Gullberg, Estimating the measurement uncertainty in forensic blood alcohol [32] T. Arndt, B.B. van der Meijden, J.P.M. Wielders, Atypical serum transferrin isoform analysis, J. Anal. Toxicol. 36 (2012) 153–161. distribution in liver cirrhosis studied by HPLC, capillary electrophoresis and trans- [60] T. Vosk, A. Emer, Forensic Metrology, a Primer for Lawyers and Judges, CRC/Taylor ferrin genotyping, Clin. Chim. Acta 394 (2008) 42–46. Francis Group, second ed., 2014. [33] S.H. Stewart, S. Comte-Walters, E. Boewen, R.F. Anton, Liver disease and HPLC quan- [61] A. Helander, J.P.M. Wielders, R. te Stroet, J.P. Bergström, Comparison of HPLC and tification of disialotransferrin for heavy alcohol use: a case series, Alcohol. Clin. Exp. capillary electrophoresis for confirmatory testing of the alcohol misuse marker car- Res. 34 (2010) 1956–1960. bohydrate-deficient transferrin, Clin. Chem. 51 (2005) 1528–1531. [34] International Vocabulary of Metrology – Basic and General Concepts and Associated [62] N. Kenan, S. Husand, A. Helander, Importance of HPLC confirmation of problematic Terms (VIM), ISO, third ed., 2006 (Geneva). carbohydrate-deficient transferrin (CDT) results from a multicapillary electrophore- [35] J.-O. Jeppsson, Isolation and partial characterization of three human transferrin var- sis routine method, Clin. Chim. Acta 411 (2010) 1945–1950. iants, Biochim. Biophys. Acta 140 (1967) 468–476. [63] F. Schellenberg, J.P.M. Wielders, Evaluation of capillary electrophoresis assay for CDT [36] A. Barroso, E. Gimenez, F. Benavente, J. Barbosa, V. Sanz-Nebot, Analysis of human on SEBIA's Capillarys system: intra and inter laboratory precision, reference interval transferrin by capillary electrophoresis and capillary liquid chromatography-mass and cut-off, Clin. Chim. Acta 411 (2010) 1888–1893. spectrometry. Application to diagnosis of alcohol dependence, Anal. Chim. Acta [64] M. Marinova, C. Artusi, S. Baggio, M. Zaninotto, M. Plebani, First evaluation of a 804 (2013) 167–175. multi-capillary electrophoresis CDT assay on Helena Biosciences' V8 analyser, Clin. [37] I. Zegers, H. Schimmel, To harmonize and standardize: making measurement results Biochem. 47 (2014) 228–232. comparable, Clin. Chem. 60 (2014) 911–913. [65] A. Helander, J. Bergström, Determination of carbohydrate-deficient transferrin in [38] B.M.R. Appenzeller, R. Wennig, Altered distribution of transferrin isoforms according human serum using the Bio-Rad %CDT by HPLC test, Clin. Chim. Acta 371 (2006) to serum storage conditions, Clin. Chem. 51 (2005) 2159–2162. 187–190. [39] A. Helander, G. Nordin, Insufficient standardization of a direct CDT immunoassay, [66] J.P.M. Wielders, R. te Stroet, N. Dirsch, Excellent correlation of recipe's CDT method Clin. Chem. 54 (2008) 1090–1091. with the proposed IFCC reference HPLC method, Ned. Tijdschr. Klin. Chem. [40] ISO 15189, Medical Laboratories – Requirement for Quality and Competence, 2012. Labgeneesk. 34 (2009) 70. [41] E. Pronicka, M. Adamowicz, A. Kowalik, et al., Elevated carbohydrate-deficient trans- [67] L. Lin, A concordance correlation coefficient to evaluate reproducibility, Biometrics ferrin (CDT) and its normalization on dietary treatment as a useful biochemical test 45 (1989) 255–268. for hereditary fructose intolerance and galactosemia, Pediatr. Res. 62 (2007) [68] G.B. McBride, A proposal of strength-of-agreement criteria for Lin's concordance co- 101–105. efficient, NINA client report, HAM (2005) 2005–2062. [42] A. Helander, N. Kenan Modén, Effect of transferrin glycation on the use of carbohy- [69] N. Kenan, A. Larsson, O. Axelsson, A. Helander, Changes in transferrin glycosylation drate-deficient transferrin (CDT) as an alcohol biomarker, Alcohol Alcohol. 48 during pregnancy may lead to false-positive carbohydrate-deficient transferrin (2013) 478–482. (CDT) results in testing for riskful alcohol consumption, Clin. Chim. Acta 412 [43] W. Okyere-Boakye, Studies on Genetic Variants of Human Plasma Transferrin(Thesis) (2011) 129–133. London Univ, 1997. [70] A.J. Naus, A. Borst, P. Kuppens, The use of patient data for the calculation of reference [44] M.I. Kamboh, R.E. Ferrell, Human transferrin polymorphism, Hum. Hered. 37 (1987) values, J. Clin. Chem. Clin. Biochem. 18 (1980) 621–625. 65–81. [71] A. Helander, J. Wielders, R. Anton, et al., Standardisation and use of the alcohol bio- [45] K. Kishi, Y. Ikehara, T. Yasuda, K. Mizuta, W. Sato, Transferrin polymorphism detect- marker carbohydrate-deficient transferrin (CDT), Clin. Chim. Acta 459 (2016) ed in human urine using isoelectric focusing followed by immunoblotting, Forensic 19–24. Sci. Int. 45 (1990) 225–230. Saatja: MKM info </O=JM/OU=EXCHANGE ADMINISTRATIVE GROUP (FYDIBOHF23SPDLT)/CN=RECIPIENTS/CN=EE984254B57E459B9B4B36E7ED2B33E3-MKM INFO> Saaja: SM Info, [email protected] Teema: FW: Kroonilise alkoholi liigtarbimise tuvastamiseks mõeldud markeri implementeerimine Eestis Tere Edastan vastavalt kuuluvusele. Dokumendihaldustalitus From: Krause, Karit <[email protected]> Sent: Monday, June 26, 2023 5:08 PM Subject: Kroonilise alkoholi liigtarbimise tuvastamiseks mõeldud markeri implementeerimine Eestis Tere! Umbes 2 aastat tagasi pöördusin Tervisearengu Instituudi poole seoses kroonilise alkoholi liigtarbimise skriininguga. Sealt suunati edasi Justiitsministeeriumi ja Sotsiaalministeeriumi kontaktide poole. Kahjuks ei ole selle teemaga minule teadaolevalt siiski edasi mindud. Nimelt on olemas võimalus kasutada vereseerumi markerit, mis on mujal maailmas laialt levinud nii kohtupraktikas, arsti abiliseks kroonilise alkoholi liigtarbimisega (ja/või muude haigustega) patsientide ravis kui ühiskonda teenindavate ametikohtade skriinimisel (juhilubade/relvalubade taotlemisel, piloodid, kaptenid, bussijuhid jne). Väga paljudes riikides on eelnevalt nimetatud juhtudel CDT-IFCC määramine ja skriining tavaprotseduur. Ka Eestis võiks sarnane määramine toimuda, et meie ühiskorda tervemaks ja turvalisemaks muuta. Eestis on olemas hetkel 3 laborit, kus määratakse standardiseeritud CDT-d, kuid, et puudub ühtne süsteem ja reeglistik on selle määramise hulk hetkel väga marginaalne. CDT-IFCC on Rahvusvahelise Kliinilise Keemia ja Laboratoorse Meditsiini Föderatsiooni poolt väljatöötatud kalibreeritud CDT väärtus, mis näitab alkoholi liigtarbimist möödunud 2 nädala kohta (sarnaselt glükeeritud hemoglobiiniga diabeediravis). Marker nimega CDT-IFCC ehk süsivesik-defitsiitne transferriin. Kalibreeritud süsteemidel, näiteks kapillaarelektroforeesil, on CDT-IFCC tulemused tõesed ja omavahel võrreldavad üle maailma. CDT on ainus IFCC poolt standardiseeritud ning kõige spetsiiflisem kroonilise alkoholi liigtarbimisega seotud marker, mis on kasutusel juba 2018ndast aastast. Lisasin manusesse mõne artikli CDT-IFCC kohta ning OECD hinnangu alkoholi liigtarbimisega seotud ennetustöö võimaluse kohta Eestis. Kas oskate soovitada, kelle poole peaksime pöörduma, et aidata määrustega ka Eesti ühikonda tervislikkuse ja ohutuse suunas sammuke edasi astuda? Ette tänades ja tagasisidet ootama jäädes! Parimatega / Best regards, Karit Krause Müügijuht / Sales Manager Laboritarbed ja -instrumendid / Laboratory Business T +372 6515 168 | GSM +372 56237 168 | [email protected] <mailto:[email protected]> | www.mediq.ee <http://www.mediq.ee/> | Mediq Eesti OÜ | Väljaotsa 2, 76505 Saue, Estonia | Please consider the environment before printing this email _________________________________________________________________________________________________ Disclaimer This e-mail is meant exclusively for the intended recipient. If you received this e-mail in error, any use, copying or distribution of this e-mail is strictly prohibited. Please notify the sender of the error by return e-mail and destroy/delete this e-mail, as well as any attachment. Mediq B.V. is registered in the chamber of commerce under number 56038062