Total Testing Process, Including Standardisation, Preanaltical Process

Clinica Chimica Acta 493 (2019) S673–S710

Contents lists available at SciVerse ScienceDirect

Clinica Chimica Acta

journal homepage: www.elsevier.com/locate/clinchim

Total testing process, including standardisation, preanaltical process

M372 clinician for the laboratory services and ultimately increase laboratory test utilization. Successful development of real-time display of the ongoing phases of laboratory tests and expert comments by laboratory doi:10.1016/j.cca.2019.03.1493 physician on the computerized order communication system

S. Kimb, T.H. Umb, C.R. Chob, D.W. Shina M373 aEmergency Medicine, Ilsan Paik Hospital, Inje University, Goyang, Republic of Korea Evaluation of BD Vacutainer® urinalysis tubes for serous ﬂuids bLaboratory Medicine, Ilsan Paik Hospital, Inje University, Goyang, cell counts Republic of Korea

a a a a b Background-aim R. Ramos , N. Del Amo , I. PeÑa , M.J. Ruiz , M.F. Calafell , M.B. Alvareza, F. Cavaa, R. Guillena aHospital Infanta Sofía-Brsalud, Spain Turnaround time (TAT) is a retrospective quality indicator for bHospital Tajo-Brsalud, Spain laboratory tests. The real-time display of the ongoing phases of laboratory tests would be helpful for clinicians to estimate the Background-aim duration required to obtain result reports. Additionally, the expert comments by laboratory physicians about the test results through In inflammatory and infectious diseases, the cellular components the Order Communication System (OCS) would contribute to the in serous fluids are increased rendering essential diagnostic infor- increase in laboratory test utilization. mation. Sample testing should occur within one-hour post collection to avoid cellular deterioration or lysis. The aim of this study was to Methods evaluate the usefulness of BD Vacutainer® Urianalysis tube (BDUt) for serous fluids cell counts. We upgraded our OCS program to enable the real-time display of the ongoing phases of laboratory tests process, which are categorized Methods as follows: 1) test request, 2) label printing, 3) sampling, 4) receipt by laboratory, 5) test carry out (manual or automated), 6) result White blood cells (WBC) and Red blood cells (RBC) counts from 9 verification, 7) interpretation by laboratory physician and 8) result fluid samples were processed and aliquoted into BDUt for further final report. Additionally, the feature enabling the easy adding and analysis at 4 and 8 hours after the initial analysis. Neubauer and browsing of expert comments with regard to test results in OCS was Fuchs-Rosenthal were used as a counting chamber. BDU contains installed. ethyl paraben, sodium propionate and chlorhexidine preservative. Results Results

This upgraded OCS system has enabled clinicians to know what Although t-test showed a nonsigniﬁcant reduction of WBC and phases the requested laboratory tests are in at present in real-time. RBC neither in the ﬁrst 4 hours nor 8 hours the decreased of Additionally, this system has allowed laboratory physicians to put concentration in both groups were up to 80% in some cases. interpretative expert comments about test results, even for not specialized routine tests (e.g., HbA1c, CBC) that does not normally Conclusions add interpretive comments.

Conclusions Despite of several papers have published that BDUt maintains the sample integrity for urine cells, our results does not recommend use for serous ﬂuid. We have successfully developed the OCS function for the real- time display of test phases and expert comments by laboratory doi:10.1016/j.cca.2019.03.1494 physicians. This function would contribute to the satisfaction of the

0009-8981/$ – see front matter Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

M374 parameter. 22 parameters were verified: IgG, IgM, IgA, Albumin in serum and CSF and Transferrin, Haptoglobin, soluble Transferrin Should we use a tube sorting device or not? Receptor, Ceruloplasmin, IgE, C3, C4, Alpha-1-antitrypsin, IgG subclasses, Haptoglobin, CRP and Beta-2-microglobulin. Verification E. Calci, Z. Kangal, S.E. Yilmaz results were exported from laboratory information system to Usak Public Health Laboratory, Turkey modified and verified Excel templates which were downloaded from The Association for Clinical Biochemistry and Laboratory Medicine’s Background-aim website. Verification analysis for each analyte was performed for 5 days with 5 replicates per day for two different concentration levels. Laboratory processes are defined in 3 stages. (pre-analytic, Accuracy and precision of repeatability, intermediate precision and analytic, post-analytic). The most common error rate of these reproducibility were calculated. processes is seen in the pre-analytic stage. In our study, we aimed to evaluate the tube sorting device which we used in the pre- Results analytical process. All verified parameters corresponded acceptance criteria from the Methods manufacturer.

The samples sent to the Uşak Public Health Laboratory from Conclusions different centers on the same day were included in the study. Acceptance and separation of the samples were done by using The protocol used proved to be a fast and reliable tool for carrying personnel and tube sorting device. The elapsed time was recorded. out the verification of a number of parameters. Verification provides The difference between recorded times was evaluated. a deeper insight into the quality of the material used in the medical laboratory, thereby increasing confidence in analytical results. In our Results case, all the verified parameters were consistent with the manufacturer’s data, so we could introduce the BN II analyzer within a short When different numbers of blood samples were examined at period of time into a clinical practice. different times during the day, it has been observed that the tube sorting device achieved the blood seperation and acceptance doi:10.1016/j.cca.2019.03.1496 procedure nearly half time shorter when compared to personnel.

Conclusions M376

The study showed that the blood acceptance and separation Turnaround time in emergencies process in the pre-analytical stage can be done with the tube sorting device with less error, less labor force and shorter time and we can A. Vílchez Rodríguez, S. Esteve Poblador, C. Valldecabres Ortiz, J. ﬁ evaluate the personnel more ef ciently in different departments. González Cantó, M. Ortuño Alonso Hospital Universitario La Ribera, Alzira, Spain doi:10.1016/j.cca.2019.03.1495 Background-aim

M375 Turnaround time (RT) is the interval between the arrival time of a sample to the laboratory and the time of clinical validation of results. Veriﬁcation of BN II analyzer according to ISO 15189 It is an indicator of laboratory quality, which has an impact on diagnosis and treatment. M. Krsnik, M. More, U. Čegovnik Primožič Our goal was to evaluate RT of samples received from the ﬂ University Medical Centre Ljubljana, Institute of Clinical Chemistry and Emergency Department, in a 5 days period and study the work ow Biochemistry, Ljubljana, Slovenia in preanalytical, analytical, postanalytical and total phases.

Background-aim Methods

Each examination procedure must be verified in order to Data of 385 stat requests were evaluate in an ADVIA 1800 guarantee that performance characteristics comply to the intended Chemistry System® (Siemens Healthineers). We calculated RT for scope of the test. There are limited guidelines available on how to each phase and for each of the four time periods studied: 1 (8-11h), perform analytical part of verification for ISO 15189 at the moment. 2 (11-15h), 3 (15-20h), and 4 (20-8h). RT was measured in minutes. Upon the start of the verification process, one is confronted with Data were processed with Excel 2013. many undefined steps, not mentioned in ISO 15189 or any other standard. That is why we prepared our own verification protocol Results which will be explained below. Analyser BN II was verified by our protocol. 49, 80, 125 and 131 samples were analyzed for periods 1, 2, 3 and 4 respectively. Average number of determinations for sample were Methods six, with the most frequent determinations being glucose, creatinine, urea, sodium, potassium and chloride in serum samples. Each step of verification must be planned in detail, well RT average was 24, 12, 12 and 48 minutes, for preanalytical, documented, reviewed and written in the verification plan for each analytical, postanalytical and total phases, respectively. For periods 1, Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

2, 3 and 4, RT average total was 50, 48, 50 and 40 minutes, − Androstenedione: Y=0.34X + 1.14 respectively. Constant “a” differed from 0, we can state that both methods Conclusions present constant differences. The plot slope did not contain the value 1, there are proportional differences. The mean difference leads to a During periods when our emergency samples were mixed with negative bias. The functional sensitivity concluded that the method inpatients and Primary Health samples (8-20h), RT was higher, was sensitive up to concentrations of 0.197 ng/mL mostly because of time used in preanalytical phase. During period of − 17-OHP: Y=0.49X + 1.02 20-8h, in which only stat samples were analyzed and there was less work load, RT decrease. It is necessary to analyze which factors fall “ ” on preanalytical phase and prioritize stat samples to adapt to The constant a differed from 0, we can state that both methods healthcare needs, although RT for stat samples are always less than present constant differences. There are proportional differences 60 minutes. between both methods. The mean difference leads to a positive bias. Value 0.158 ng/mL is proposed for functional sensitivity. doi:10.1016/j.cca.2019.03.1497 The MAGLUMI1000® method showed good repeatability and reproducibility.

Conclusions M377 The results obtained with both methods are not interchangeable, Measurement of free testosterone, androstenedione, and 17- the reference values cannot be adopted. In addition, proportional and hydroxyprogesterone: Comparison of current laboratory methods constant differences were observed. Therefore, new population reference values should be set using MAGLUMI1000®. J. Gorrín Ramos, N. López Lazareno, C. González-Mohíno Vázquez De Ágredos, A. Mata Fernández, M.M. Herranz Puebla doi:10.1016/j.cca.2019.03.1498 Hospital General Universitario Gregorio Marañón, Madrid, Spain

Background-aim M378 Testosterone is the main androgen. Its function is to develop Analysis of reasons for biological specimens rejection at a secondary sex features of male. Clinically, it is useful for the diagnosis university hospital clinical laboratory and follow-up of hypogonadism, impotence and prostate cancer. Androstenedione and 17-hydroxyprogesterone (17-OHP) are intermediate metabolites in the synthesis of androgens and estrogens. S. Kim Their clinical use is in the diagnosis and follow-up of patients with Laboratory Medicine, Ilsan Paik Hospital, Inje University, Republic of congenital adrenal hyperplasia and premature adrenarche. We Korea compared if a new method of measurement, (MAGLUMI1000®, SNIBE Co©., chemiluminescence immune assay (CLIA)) is equivalent Background-aim to an stablished method (PERSONAL LAB JUNIOR®, DRG international©, enzyme-linked immunosorbent assay (ELISA)) and adapt- Preanalytical errors cause decreasing the accuracy of clinical ability of the reference values for 3 study hormones. Repeatability, laboratory results. The purpose of this study was to analysis the reproducibility, and functional sensitivity were also studied. reasons of biological specimens rejections, regarding their rates in certain test groups, department in our hospital. Methods Methods Serum samples were aliquoted and frozen at –20°C until processing. Repeatability, reproducibility, and functional sensitivity The disqualified sample type and reasons of errors in the were studied using CLIA to determinate the concentration of the 3 preanalytical phase were investigated in our laboratory over a 1- study parameters in a quality control (BIO RAD©, Lyphocheck™ year period from January 2018 to December 2018. Data immunoassay plus control). were obtained using an specimen rejection documents and from the hospital information system. Type of inappropriateness − f-T: 311 serum samples were analyzed (66.6% in men, aged 2-88 were evaluated as follows: improper request, incorrect labeling, years). improper collection, inappropriate transport, inappropriate specimen − Androstenedione: 131 serum samples were analyzed (71.8% in quality. women, aged ranging 4-85 years. − 17-OHP: 150 serum samples were analyzed (82.7% in women, Results aged 1 month to 86 years). There were 1,124,081 biological specimens were submitted to Results clinical laboratory for 1-year. Among them 164 (0.01%) specimens were cancelled by the doctor for any reasons, 5,880 (0.52%) − f-T: Y=0.48X + 0.37 specimens were rejected and taken again due to disqualified specimen, and remaining 1,118,037 (99.46%) specimens were Constant “a” contained 0, we can state that there are no constant qualified. Internal medicine, emergency medicine and neurosurgery differences. The plot slope did not contain the value 1, there are were the department having the highest rejection rate. General proportional differences. The mean difference leads to a negative ward, emergency room and MICU were the ward having highest bias. A value of 0.7 pg/mL is proposed for functional sensitivity Abstracts / Clinica Chimica Acta 493 (2019) S673–S710 rejection rate. Hemolysis and clotting were the most common reasons for specimen inappropriateness. M380 Conclusions Demand management of procalcitonin requests in the biochem- We analyzed of the disquailified biological specimens rates and istry emergency laboratory the reasons. Quality improvement activity based on these results will reduce the disqualified sample submission and helps to reduce risk R. Rubio Sánchez, E. Lepe Balsalobre, M.D.M. Viloria Peñas, A. Moro of patients outcome and improve the quality of clinical laboratory. Ortiz Virgen de Valme University Hospital, Seville, Spain doi:10.1016/j.cca.2019.03.1499 Background-aim

Procalcitonin (PCT) is produced mainly in the parafollicular cells M379 of the thyroid and it is one of the most reliable diagnostic markers for bacterial infection; its level are related to the severity and An audit of timeliness of urinalysis in a Singapore public hospital mortality of the infectious disease. PCT results greater than 0.5 mg/L are related to the occurrence of sepsis. The implementation of the L. Lam, P.H. Pwint, H.F. Lee, A.P. De Quiroz, J. Lee PCT test supposes a high cost because each determination costs 8.71 Department of Laboratory Medicine, Ng Teng Fong General Hospital, €. Singapore The objective was to analyze the requests with joint demand of CRP and PCT, and to implement a protocol with adequacy of the PCT Background-aim demand for a better use of the resources of the Emergency Laboratory. Proper collection of urine specimens is important to avoid contamination, or deterioration of constituents. An increased time Methods lag between sampling and analysis, and a lack of temperature control for which urinalysis cannot be performed within 2 hours of Descriptive and retrospective study of all PCT requests received in collection, will lower the quality of urinary test results. Our objective the Biochemistry Emergency Laboratory during 2017; a total of 5016 is to explore the preanalytical handling of urinalysis specimens in Ng samples were analyzed. CRP and PCT measurements were performed Teng Fong General Hospital. on a Roche Diagnostics COBAS 8000 analyzer by turbidimetry and immunoassay, respectively. For the statistical analysis, the MedCalc Methods program were used. The ROC curve were used to analyze the diagnostic efficiency, as well as the cut-off point. Retrospective data of 21560 urinalysis performed over 1 year from November 2016 to October 2016 were retrieved from the Results Laboratory Information System. The urine collection time, laboratory receiving time, whether the urine specimens were refrigerated The ROC curve analysis shows, for a cut-off value of 20 mg/L of before reaching the laboratory and ordering locations were CRP, a sensitivity of 96.80%. The diagnosis of sepsis was confirmed in evaluated. 29.43% of cases due to elevated CRP and PCT results, and was ruled out in 21.05% of cases because both determinations were in the Results normal range. Only in the 0.48% of the cases studied the CRP results were normal while the PCT results were pathological. We found that 91.5% of the specimens were received in the laboratory within 2 hours of collection and not refrigerated Conclusions previously. Another 3.8% of the specimens were delivered from a nearby healthcare facility. Although the specimens were received Based on the published scientific evidence, we conclude that it is beyond 2 hours of collection time, they were refrigerated prior to necessary to implement a protocol to adapt the demand to avoid transportation and therefore met the preanalytical quality require- unnecessary tests and to make rational use of the laboratory ments. The remaining 4.7% of the specimens were received beyond resources. Because of this, the PCT test should not be performed if the 2 hours of collection time and were never refrigerated. Of the there is no CRP value higher than 20 mg/L, except in cases where 1006 urine specimens that exceeded the quality requirements for there is clinical suspicion of sepsis. It would also be necessary to handling, 82% and 18% were from inpatients and outpatients, perform the PCT determination, without regard the CRP results, in respectively. neonates and in the ICU samples. During the time studied, 1080 requests of PCT (21.53%) should Conclusions have been canceled. After the implementation of this protocol, the savings would have been 9406.8€. In addition, the response time of When compared to the CAP Q-Probes 1997 study, our audit the emergency laboratory would be reduced because the PCT showed excellent compliance to recommended standards and determination takes longer to be made, delaying the delivery of the quality guidelines. We will continue to monitor and emphasize to results. our clinical colleagues the importance of preanalytical handling of urinalysis specimens. doi:10.1016/j.cca.2019.03.1501 doi:10.1016/j.cca.2019.03.1500 Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

M381 Methods

Improving troponin T turnaround time by changing barcode type We made two pools of plasma: one with patients which INR was less than 1,5 and another group with INR more tan 1.5. In both pools R. Nar, S. Demir, E. Avcı were performed hemolysis in vitro adding growing concentration of Faculty of Medicine, Department of Biochemistry, Pamukkale University, Hemoglobin (Hb) from 25 to 2000 mg/dl (25, 50, 100, 200,400, Denizli, Turkey 500,1000,2000). All samples generated were processed in CA-500 (Sysmex®) to obtain basic coagulation tests (PT and APTT). Background-aim Differences pre and post hemolysis were evaluated according to Spanish Minimum Consensus Performance Speciﬁcations (SMCPS). Laboratory turnaround time (TAT) is one of the most important indicators of the quality and effectiveness of laboratory performance. Results In this study, we aimed to evaluate the effects of new barcode type usage for emergency department (ED) on TAT analysis of Troponin T Both in the INR group b 1.5 and in groupN 1.5, percentage of test T in the emergency department (ED) of university hospital. change (PC), for TP and APTT, in the different grades of hemolysis, was lower than that published by SMPC (31% and 24% respectively). Methods For the TP the maximum change was 6% with a concentration of Hb of 1 mg/dl whereas for the APTT was 9% with 0.5 mg/dl Hb. No This study was carried out in the Clinical Biochemistry Laboratory linear relationship was observed between increased hemolysis and of Pamukkale University in January-December 2018, for Troponin T PC. test. In July, we started to use new barcode type which has red mark on the upper side. TAT was calculated in January-June and July- Conclusions December 2018. TAT was calculated from the time when blood samples arrived at the laboratory and the time of completion and Our results have shown that TP and APTT, do not undergo reporting of the test.The agreed TAT for Troponin T in our laboratory clinically signiﬁcant changes after hemolysis and question the policy is 120 minutes. of rejection of samples published by CLSI guidelines. Our future goal is investigate effect of hemolysis in patients with Results anticoagulant treatment.

Between January-June 2018, total number of analyzed Troponin T doi:10.1016/j.cca.2019.03.1503 were 14413 and the TAT averages were 114 minutes. During July- December 2018, after new barcode the number of analyzed test was 14087 and the TAT averages were 92 minutes. M383

Conclusions Comparison of commercially available 25OHD and 1,25(OH)2D assays: Experience of pediatric hospital laboratory participating Everyday hundreds to thousands of samples (both routine and in DEQAS proﬁciency testing emergency) analyse in the Central Biochemistry Laboratory. During this laboratory workload technicians may be delayed to detect and E. Czekuc-Kryskiewicz, M. Wojcik, B. Paraﬁniuk, E. Skorupa, P. analyze emergency samples. TAT time was shortened by increasing Pludowski the awareness of technicians with new barcode application. Department of Biochemistry, Radioimmunology and Experimental Medicine, The Children’s Memorial Health Institute, Warsaw, Poland doi:10.1016/j.cca.2019.03.1502 Background-aim

M382 Serum 25OHD is a reliable biomarker of vitamin D status.

Hormonal active vitamin D metabolite - 1,25(OH)2D - is less often Influence of hemolysis in basic coagulation parameters used in clinical practice. Both accuracy and precision are important factors for proper diagnosis of vitamin D deficiency and activity. M.F. Calafell Masa, R. Ramos Corralb, M.J. Mallol Poyatoa, S. Marin The aim of the study was to comprise of methods for serum Yepesa, M. Martin Villara, M. Prada Ortiza, R. Guillen Santosb, F. Cava determination of 25OHD (automatic CLIA on IDS-iSYS and LIAISON b Valenciano analyzers) and 1,25(OH)2D (semi-automatic CLIA on IDS-iSYS aHospital del Tajo, BR Salud, Madrid, Spain analyzer and manual RIA) in pediatric and DEQAS samples. bSan Sebastian de los Reyes, Hospital Infanta Sofia, BR SALUD, Madrid, Spain Methods

Background-aim The intra- and inter-variability (CVintra and CVinter) of 25OHD and 1,25(OH)2D measurements on the IDS-iSYS platform in Hemolysis is preanalytical incidence with a prevalence around pediatric samples were calculated. The comparison of the IDS-iSYS 3,3%. It is the most frecuent cause of rejected samples in the CLIA methods of 25OHD and 1,25(OH)2D determinations with the laboratory. However, hemolysis effect in basic coagulation parame- LIAISON CLIA and manual RIA methods, respectively, was per- ters (prothrombin time (PT) and activated partial thromboplastin formed. The accuracy of the CLIA method of 25OHD quantiﬁcation time (APTT)) has been less studied. The aim of our study is evaluate on the IDS-iSYS was evaluated using DEQAS HPLC and LC-MS/MS the inﬂuence of hemolysis in these analytes. data. Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

Results bilirubin and turbidity and for the veriﬁcation of the rates of hemolysis, jaundice and lipemia”.

The CLIA methods of 25OHD and 1,25(OH)2D determinations on To assess the existence of interferences, the limit used by the the IDS-iSYS platform were characterized by high repeatability manufacturer (10%) was taken as the maximum admissible error

(CVintra 2.6% and 6.4%, respectively) and reproducibility (CVinter criterion. The interference was obtained for the following compounds: 9.1% and 9.6%, respectively). Measurement of 25OHD concentration amylase, aspartate aminotransferase (AST), direct bilirubin (DBil), in pediatric and DEQAS samples showed, respectively, 25.7% and creatinine kinase (CK), alkaline phosphatase (ALP), iron, lactate 49.4% bias between methods. The comparison of methods of 1,25 dehydrogenase (LDH), lipase, magnesium, phosphorus, potassium,

(OH)2D determination showed their good compatibility in pediatric triglycerides, thyroid antiperoxidase antibodies (ATPO), Anti-receptor samples (bias of 11.0%). antibodies of TSH (ATSHR), folate, ultrasensitive troponin T (hs-cTnT), insulin, parathyroid hormone (PTH) and progesterone. Conclusions Conclusions The study indicate high precision of serum 25OHD and 1,25(OH)

2D determinations using CLIA methods on the IDS-iSYS platform. This Working with new analyzers, it is important to study the made the possibility of their use in clinical practice of pediatric dependence of the results of the interference of the hemolysis on hospital with maintenance of the highest diagnostic standards. The the used differentanalytical techniques, as it has an impact on the disagreement noted between methods of 25OHD determination may medical decision making. be due to differences in the cross-reactivity with the vitamin D In clinical practice, it is recommended to determine the hemolytic metabolites. index, to detect and quantify the interference by hemolysis, assessing the possible rejection of samples and providing reliable reports doi:10.1016/j.cca.2019.03.1504 helping clinicians to interpret the results.

doi:10.1016/j.cca.2019.03.1505

M384

Study of interference produced by hemolysis in 73 analytical tests M385

Y. Douhal, V. Benito Zamorano, L.D.M. Rivas Chacón, E. Cuadrado Comparison study between two enzyme immunoassay methods Galván, G. Sánchez Helguera, V. Cámara Hernández, R. Torrado for the determination of mycophenolate acid Carrión, T. Pascual Durán Hospital Universitario de Getafe, Spain Á. Cabrera Argany, M. Kassih Ibrahim, G. Garcia Aguilar, H. Cabrera Valido, T. Dorta Ramos, A.M. SÁnchez De Abajo, J. Paco Ferreira, T. Background-aim HernÁndez Lemes Las Palmas de Gran Canaria, Complejo Hospitalario Universitario Insular One of the main objectives of a clinical laboratory is getting Materno Infantil, Spain precise and accurate values. For instance, to quantify the hemolysis, the study of serum index such as hemolytic one is a widespread Background-aim practice. Hemolysis can take place both in vivo and in vitro. The latter is being the most frequent one. Breaking the red cell can Mycophenolic acid (MPA) is an immunosuppressant that inhibits produce interferences through several mechanisms, altering the the de novo pathway of guanosine nucleotide synthesis acting as a absorption spectrum, by chemical interferences, by release intracel- reversible and uncompetitive inhibitor of inosine monophosphate lular components and by dilution of the components of the sample. dehydrogenase, reducing the intracellular reserves of guanosine Due to a change of analyzers, the present hemolysis interference nucleotides and blocking the lymphocyte proliferation study is aiming to quantify the possible change in 73 tests performed It is administered as mycophenolate, a prodrug that is metabo- by different techniques in such equipment: spectrophotometry, lized at the liver to the active substance, MPA. Renal, hepatic or indirect potentiometric and immunological techniques. cardiac transplant patients usually receive MPA along with another immunosuppressant, however it is also used as monotherapy for the Methods treatment of autoimmune diseases. the low frequency of adverse effects makes the MPA very useful, especially in the pediatric The tests were performed in a Roche Diagnostics Cobas 8000 population. equipment following the protocol of the Spanish Society of The aim of our study is to evaluate the interchangeability of MAP Laboratory Medicine (SEQCml): “Procedure for the study of interfer- levels, measured by enzyme immunoassay technique in two different ence by haemolysis, bilirubin and turbidity and for the veriﬁcation of analyzers: AU-680 (Beckman Coulter®) and Cobas C 501 (Roche®). the rates of hemolysis, jaundice and lipemia”. Methods Results 82 serum samples, mainly from pediatric nephrology service The results were compared with those reported in the inserts patients, were recollected. serum MPA levels were measured in both provided by Roche Diagnostics of the different techniques. The tests analyzers. First, from every serum sample an aliquot was frozen and were performed in a Roche Diagnostics Cobas 8000 equipment another sent to an external laboratory and analyzed in the same following the protocol of the Spanish Society of Laboratory Medicine week of its recollection in the Cobas C 501 (Roche®). Then all the (SEQCml): “Procedure for the study of interference by haemolysis, frozen aliquots were thawed and analyzed all together in our AU-680 Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

(Beckman Coulter®) analyzer. Bothe analyzers had a working range ordinate in the origin -134 (CI95%: -33, -234). The CI95% of the of 0.3-10 ug/mL. values in the ordinate in the origin did not contain the zero value. Following the 2011 recommendations of The Spanish Society of The CI95% of the values of the slope did not contain the value 1. Laboratory Medicine (SEQC) for methods comparison studies, We observed a mixed SE in the evaluated procedure (constant Passing Bablok regression and Bland Altman plot were obtained and proportional). We estimated the SE starting from the slope value using MedCalc statistical software. (b) being 3.17%.

Results Conclusions

We obtained the following results: Bland Altman plot with According to the differences analysis there is a constant and mean= 0 ± 0.81 ug/mL and Passing Bablok regression equation Y proportional SE. On the other hand the CI95% of the values of the = -0.32 ± 1.17X where Y= AU-680 and X= Cobas C 501 slope and ordinate in the origin indicate that the evaluated procedure provides significantly different values to the comparison Conclusions procedure, having also constant and proportional systematic differences. Although according to the statistical results, there is a propor- The SE calculated is lower than that recommended in the tional and constant differences between the two analyzers, we specifications according to the SEQCML for NT-proBNP (4, 72%), we consider them clinically insignificants, and we can conclude that can consider it acceptable for the evaluated method. both analyzers are interchangeable. The constant SE cannot be corrected, but the proportional SE can Thanks to those results, our laboratory can carry out the be minimized changing the calibrators or the process of calibration. measurement of all the immunosuppressants used in renal transplanted patients, becoming the reference kidney transplanta- doi:10.1016/j.cca.2019.03.1507 tion hospital in our province. doi:10.1016/j.cca.2019.03.1506 M387

Use of haemolysis as a quality indicator of pre-analytical M386 processes

Study of the bias of a measurement procedure D.H. Heredero Jungb, S. Elena Pérezb, F. Marqués Garcíab, C. Agulló Rocab, A. Alonso Pradab, M.J. Aldegunde Rodríguezb, L. Rollán Mansoa a S. Esteve Poblador, A. Vílchez Rodríguez, J. González Cantó, M. Ortuño Clinical Laboratory Service, Álvaro Cunqueiro Hospital, Vigo, Spain b Alonso, F. Javier Lirón Hernández Clinical Laboratory Service, University Hospital of Salamanca, Sala- Hospital Universitario La Ribera, Alzira, Spain manca, Spain

Background-aim Background-aim

N-terminal natriuretic peptide (NT-proBNP) assay is useful in Quality indicators concerning haemolysis in pre-analytical pro- diagnosis, prognosis and monitoring patients with heart failure. cesses are contemplated in consensus documents nowadays. Several Our goal was to verify the specifications of systematic error (SE) of these documents propose a maximum level of 0.5 g/L as a provided by the manufacturer, for a new procedure used in serum landmark to establish the quality indicator. This cut-off point may be NT-proBNP test values. too high to be able to evaluate the haemolysis. Our hospital receives samples from all over the province, which Methods means there are several collection points which need to be evaluated. We try to obtain a way to find a suitable indicator for each laboratory. The analizer used have been the IMMULITE® 2000 (Siemens Our aim is to establish cut-off points in the level of haemolysis to fi Healthineers), for comparison procedure; and the ADVIA Centaur serve as internal quality indicators, taking the clinical signi cance of XP® (Siemens), for evaluated procedure. ananalytes into account. The NT-proBNP test was done by chemiluminescent immunometric assay. Methods For the study of veracity 50 samples of patients were processed along various series, following the protocol by the Sociedad Española As a starting point, haemolysis in our laboratory was assayed to de Medicina de Laboratorio (SEQCML). determine the cut-off points of each analyte, according to the recent documents about harmonization. We studied both the analytical and Results biological variabilities, establishing three intervals for each parameter (1. below the analytical cut-off, 2. between the analytical and Study of veracity was made using 2 methods: biological cut-off and 3. over the biological cut-off). Differences analysis: we obtained an average value the relative To perform the study, we built a database of 28475 samples of differences of -34 pg/mL (CI95%: -98, -287). The 95% confidence patients from the 40 collection points in our province (labeled interval (CI95%) of the average of the absolute and relative alphabetically from A to AN). These laboratory requests include differences did not include the zero value. potassium (K), lactate dehydrogenase (LDH), folic acid (FOL), Lineal regression analysis: we obtained a correlation coefficient aspartate aminotransferase (AST) and alanine aminotransferase (r) of 0.992, making the selected value intervals, being greater than (ALT). We calculated the percentages of samples cointained in each fi 0.975. The obtained slope was 0.968 (CI95%: 0.944, 0.993) and the of the three haemolysis intervals for the ve parameters. Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

Results the ordering locations, we found the hemolysis rate to be highest from Emergency (12 – 17%), followed by ICU (9 – 15%), Inpatient For LDH, we found out that, being the average result in the Wards (10 – 13%) and Outpatient Clinics (6 – 10%). In terms of second interval around 10-11% for most of the collection points, the absolute number, the most hemolyzed samples come from Inpatient percentage of some of them exceeded this value (AA: 27.38%, AB: Wards (56%), followed by Outpatient Clinics (34%), Emergency (6%) 28.31%, AI: 35.08%, J: 35.8%, S: 41.46%). The results for K are quite the and ICU (4%). same, as the intervals are similar to those of LDH. The cut-off points of FOL, AST and ALT are higher, so they provided no successful way Conclusions to ﬁlter the samples. Our ﬁndings were consistent with other published data showing Conclusions Emergency department hemolysis to be much higher than the ASCP 2% benchmark. However, we were concerned to discover similarly Haemolysis levels for LDH and K in the second interval are good unsatisfactory hemolysis rate across all locations. Further analysis indicators, as they are affected soon by increased levels of showed with the exception of the staff clinic, all other inpatient haemoglobin. This kind of study can lead to improvements in the wards and outpatient clinics were unable to meet the desired performance of those collection points which need it, applying the benchmark. Although we were somewhat comforted that grossly corrective measures considered. hemolyzed specimens, which posed the greatest challenge to interpretation of patients’ results, was at a much lower rate of about doi:10.1016/j.cca.2019.03.1508 0.2%, we acknowledged there is abundant room for quality and process improvement.

doi:10.1016/j.cca.2019.03.1509 M388

An audit of sample hemolysis rate in a Singapore public hospital M389 L. Lam, S.N. Devi, N.W.T. Ong, E.X. Foo, J. Lee Department of Laboratory Medicine, Ng Teng Fong General Hospital, Comparative study of two methods for the determination of Singapore plasma proteins

Background-aim I. Peral Camacho, M.D.M. Viloria Peñas, E. Lepe Balsalobre, A. Moro Ortiz Haemolysed specimens delay clinical laboratory results, prolifer- University Hospital Virgen de Valme, Seville, Spain ate unnecessary testing, complicate physician decisions, injure patients indirectly, and increase healthcare costs. In vitro hemolysis Background-aim is a more frequent phenomenon compared to in vivo and occurs from poor sample collection. This causes unnecessary sample Inflammatory processes, neoplastic alterations or immunological rejection and thus poses a major challenge in hospitals. According disorders can be diagnosed and followed by quantification of the to American Society for Clinical Pathology (ASCP), a hemolysis rate of corresponding plasma proteins. These proteins are called acute phase 2% or less is considered the benchmark of best practice. proteins and among them are ceruloplasmin, la1-antitrypsin or Our objective was to perform the first audit of sample hemolysis haptoglobin as positive acute phase proteins or prealbumin as a rate in our 3-year old general hospital. Locations with higher than negative acute phase protein. acceptable haemolysis rate will be highlighted and communicated to The objective was to compare the results obtained from the respective unit leaders to generate awareness and identify ceruloplasmin, 〈1-antitrypsin, haptoglobin and prealbumin in two process improvement opportunities. teams in order to substitute one method for another in our hospital.

Methods Methods

All specimens analysed on our lab automation system over 9 We analyzed, consecutively and in parallel, 100 serum samples months (1 Sep 2017 – 31 May 2018) was extracted from Abbott from hospitalized patients, outpatients and primary care. The Middleware – Data Innovations Instrument Manager. The data is samples were processed in parallel and to minimize the preanalytical then sorted according to the hemolysis index into non-hemolyzed, error, the analysis was carried out in the two teams following a slightly hemolyzed, hemolyzed and severely hemolyzed categories sequential order: first they were processed in the BNII System using Microsoft Excel. The data is further categorized according to analyzer (Siemens) (which uses nephelometry as a measurement the ordering locations such as Emergency, Inpatient Wards, Outpa- method) and subsequently in the Cobas 8000 analyzer (Roche) tient Clinics and Intensive Care. (immunoturbidometric method) The correlation between the two methods was established with Results the Pearson coefficient (r). Linear regression was also obtained by Passing-Bablok, and differences were assessed with a Bland-Altman fi A total of 211,891 sample data were collected and analysed. We scatter plot, as well as their respective 95% con dence intervals (CI). noted 23,272 samples were hemolyzed giving an overall hemolysis rate of 11%. The overall hemolysis rate ranged from 10.4 – 11.8% Results throughout the study period. Our slightly hemolyzed rate ranged from 8.2 – 9.1%, whereas hemolyzed rate ranged from 2.1 – 2.6%. The The comparison included the folowing results: the correlation grossly hemolyzed rate ranged from 0.15 – 0.21%. With regards to coefficient was 0,9476 (0,9221 a 0,9650) for ceruloplasmin, 0,9535 Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

(0,9308 a 0,9688) for prealbumin, 0,3262 (0,1334 a 0,4953) for 〈1- performed on the blood samples and compared between the antitrypsin and 0,9800 (0,9700 a 0,9868) for haptoglobin. different methods of transport. Finally, the phlebotomy technique was observed at the ED (9 blood extractions). Conclusions Results Based on the obtained values we can conclude that there is a good correlation between both instruments for all parameters The blood samples that were transported manually had a lower evaluated except in the case of 〈1-antitrypsin, so it would be haemolysis frequency than those transported by pneumatic tube necessary to apply a change in the current reference values to (13% less at the ED and 6% at the UCI). Some aspects of the minimize the systematic error. phlebotomy technique could be improved (sample homogeneization, patient identiﬁcation etc). doi:10.1016/j.cca.2019.03.1510 Conclusions

We observed a higher haemolysis frequency when the samples M390 are transported by pneumatic tube compared to those transported manually. To reduce the impact of the transportation via pneumatic Interference by haemolysis, who is to blame? tube system, padding for the containers has been implemented. Also, nurse training in phlebotomy planned. A. Madurga Hernándezg, P. Salas Gómez-Pablose, X. Martinez f i b d Asensio , A. Valdés Castiello , S. Reyes Medina , A. Arizmendi Demay , doi:10.1016/j.cca.2019.03.1511 M. Ballesterosh, J. Tost Vallsc, S. Martinez Martineze, V. Camacho Guilléna aCoordinació de Processos de Suport, Comunicació i Recursos Humans, Catlab, Terrassa, Spain M391 bCoordinacio Infermeria i Seguretat del Pacient, Consorci Sanitari de Terrassa, Terrassa, Spain Accuracy enhancement of HDL-cholesterol tests applying the cCoordinacio Seguretat del Pacient, Consorci Sanitari de Terrassa, reassigned calibrator value from NIST SRM 1951C Terrassa, Spain d Director d’Ambit d’Atenció al Malalt Crític i Urgent, Consorci Sanitari de D. An, D.H. Lee, Y. Hwang Terrassa, Terrassa, Spain Department of Laboratory Medicine, Seegene Medical Foundation, Seoul, e Extra Analitica, Catlab, Terrassa, Spain Republic of Korea fInformatica, Catlab, Terrassa, Spain gLaboratori d’Urgències CST, Catlab, Terrassa, Spain Background-aim hQualitat, Consorci Sanitari de Terrassa, Terrassa, Spain i Urgencias, Consorci Sanitari de Terrassa, Terrassa, Spain Because the HDL cholesterol (HDL-C) test is used to predict the risk of cardiovascular disease, its accuracy is very important. We Background-aim participated in the accuracy-based proﬁciency tests (ABPT) program of CDC LSP, CAP ABL, Korea Institute of Clinical Quality Assurance, for Haemolysis is an important interference factor that has to be the evaluation of metrological traceability during geological surveys. considered as a security patient problem. It might lead to false laboratory The accuracy criteria for all the programs were fulﬁlled, but a results, plus might need to repeat phlebotomy and cause a delay in positive bias of 3 - 5% was observed for HDL cholesterol, and there completing the laboratory analysis. At our laboratory, we measure the was a bias difference in each calibrator lot. Hence, we tried to haemolysis index (H-index) via an automated system (Cobas6000, reassign the value of the calibrator using the NIST SRM (Standardized Roche Diagnostics). Measuring the percentage of samples that present Reference Material), and evaluated whether the HDL-C test bias (%) haemolysis is one of our quality control indicators. The H-index is was improved. assessed with bychromatic measurements at 570/600 nm and is reported in arbitrary units. At the Consorci Sanitari de Terrassa, the prevalence of Methods samples with haemolysis is higher than the other two hospitals that we work with (on average 11% during 2018). Causes for haemolysis are HDL-C tests were performed using the Hitachi 7600-210 (Hitachi, diverse and can occur at many points during sample processing, Japan) analyzer using Cholestest N HDL (Sekisui, Japan) reagents and therefore a multidisciplinary approach has been used to review them. calibrated with Cholestest N Calibrators. The analyzer was calibrated using two-level NIST SRM 1951c materials, whose concentrations Methods were 41.0 ± 0.9 mg/dL and 64.9 ± 1.7 mg/dL. The original calibrator (concentration: 58.2 mg/dL) was measured 10 times to calculate the An intern and extern serum index quality control was measured reassigned value. Fourteen standard materials provided by the on a weekly and monthly basis respectively. At the laboratory, the Korean CDC (KCDC) were analyzed in the system calibrated using centrifuges have been re-calibrated and centrifugation conditions the original and reassigned calibrator values. Additionally, the have been reviewed (Heraeus Megafuge1.0). In collaboration with concentrations of LSP materials from US-CDC were measured and the Emergency Department (ED) and the Intensive Care Unit compared with the reference values of LSP materials. Department (ICU), the effect of the pneumatic tube system transport has been studied. Blood samples were transported from the ED and Results ICU to the hospital laboratory manually by hospital staff (60 patients for ED and 90 patients for ICU) or with a pneumatic tube system (98 The mean reassigned values of original calibrator using NIST SRM patients for ED and 180 patients for ICU). The H-index studies were level I, level II, and both level I / II were 56.6, 53.7, and 54.7 mg/dL. Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

When the standard materials provided by KCDC were analyzed, the Conclusions mean bias (%) of the results using 3 reassigned calibrator values and those using original calibrator value were 1.8%, -4.8%, 0.2%, and 6.8%. Preanalytical stability of ACTH is effected by observational period The mean bias (%) of the results using reassigned calibrator values after centrifugation and separation of aliquots within K3 – EDTA decreased from 4.1% to -1.0%, as compared to those using original plastic tubes post centrifugation. calibrator value, when the three LSP materials from US-CDC were analyzed. The trends of HDL accuracy-based proﬁciency tests shifted doi:10.1016/j.cca.2019.03.1513 from a positive bias to negative bias, as the reassigned calibrator values were applied.

Conclusions M393

Recent progress in the production of health-related certiﬁed The bias (%) from the target value in the HDL-C test was reference materials by the joint research centre improved using the reassigned calibrator values obtained using NIST SRM. However, it is necessary to reset the calibrator value if the calibrator lot has changed. L. Deprez, S. Boulo, E. Monogioudi, G. Auclair, S. Mazoua, H. Schimmel, I. Zegers, S. Trapmann doi:10.1016/j.cca.2019.03.1512 European Commission, Joint Research Centre (JRC), Geel, Belgium

Background-aim

M392 The European Commission’s Joint Research Centre (JRC) produces reference materials which are tailored to meet the needs of European The impact of observational timing on the preanalytical stability policy in various ﬁelds including health applications. Certiﬁed of adrenocorticotropic hormone (ACTH) Reference Materials (CRMs) for In-Vitro Diagnostics (IVD) are essential for the development of reference systems for the stand- N. Bekaia ardisation of routine IVD measurements and their use is required by Research Institute of Clinical Medicine, USA the EU IVD Regulation (EU) 2017/746.

Introduction Methods

Adrenocorticotropic hormone (ACTH) is sensitive to temperature The production of CRMs for protein IVD is a challenging process and observation timing. The pre-analytical stability of ACTH was which requires investigations on several parameters such as the carried out in two groups of patients; one cohort of patients availability of raw materials, commutability, stability, the level of the comprised of healthy volunteers, whilst another cohort included analyte of interest, and the strategy for assigning a certified value. patients with renal failure and Adison disease. The impact of The JRC develops several CRMs in close collaboration with the IFCC observational timing and the significance of plasma separation from working groups. formed elements after centrifugation was assessed. Results Methods To support measurement standardisation of biomarkers for the During the three month a total of 24 patients were analyzed using early detection of Alzheimer’s disease, the JRC released a panel of Electrochemiluminescent immune assay (ECL) , One cohort of three pooled liquid frozen cerebrospinal fluid materials with patients included 10 volunteers , whilst another the total of 14 different levels of amyloid-ß 1-42 (Aß1-42). The CRMs were patients: 9 patients with renal failure and 5 patients with Adison characterised with reference methods based on isotope dilution disease. Two blood samples were obtained from both cohort of mass spectrometry and the certified Aß1-42 mass concentrations patients. Specimen were immediately centrifuged after blood with the associated expanded uncertainties were 0.45 ± 0.07 ⎧g/L collection. The pre-cooled sampling tubes were used. The samples in ERM-DA480/IFCC, 0.72 ± 0.11 ⎧g/L in ERM-DA481/IFCC and 1.22 were analyzed at two points: immediately and after 2 hours of ± 0.18 ⎧g/L in ERM-DA482/IFCC. storage at 22C temperature. Centrifuged blood samples (12 sample) In the field of autoimmune disorders the CRM ERM-DA483/IFCC was kept in K3 - EDTA sampling tubes , whilst some plasma (12 was produced from a plasmapheresis sample of a patient diagnosed sample) was moved into clean tubes .The content of ACTH was with vasculitis. This material was certified for the mass concentration measured in both type of tubes - immediately and after two hours. of the immunoglobin G proteinase 3 anti-neutrophil cytoplasmic autoantibodies (IgG PR3 ANCA) and the certified value with the Results associated expanded uncertainty was 270 ± 29 mg/L. Building on a long tradition in the development of CRMs for the The concentration level of ACTH kept in K3 - ETDA plastic tube catalytic activity concentration of enzymes, a new CRM for has decreased between initial measurement and measurement taken pancreatic amylase will be produced. As a first step in the process a 2h post centrifugation by 2.5%.On the other hand, neither time nor commutability study was performed on five candidate materials. The temperature had any effect on hormone concentration levels kept in outcome of this study showed which matrix composition is the most clean tubes. The difference in change between these two groups was suitable for the new CRM to ensure its commutability for several of 15 %. the routine IVD methods. Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

Conclusions Conclusions

The JRC continues the development of new CRMs for biomarkers We conclude that successful Total Lab Automation demands a in various clinical ﬁelds including Alzheimer’s disease, autoimmune holistic approach, including LIS and HIS order request redesign. disorders and cardiovascular diseases. doi:10.1016/j.cca.2019.03.1515 doi:10.1016/j.cca.2019.03.1514

M395 M394 Is sample type associated with high hemolysis index? Redesigning the laboratory information system and the sample request process are prerequisites for total lab automation in N. Nikolac Gabaj, M. Miler, J. Culej, L. Milevoj Kopcinovic, A. Unic, A. hospital settings Vrtaric, A. Topic Department of Clinical Chemistry, Sestre milosrdnice University Hospital J.M. Gillisb, A. Albersenb, I. Zwiersa, C.M. Cobbaertb Center, Zagreb, Croatia aCHIZ interim management & coaching, Oldemarkt, Netherlands bDepartment of Clinical Chemistry and Laboratory Medicine, Leiden Background-aim University Medical Center, Leiden, Netherlands Haemolysis is the most common preanalytical error and leading Background-aim cause of sample rejection. In order to decrease proportion of haemolysed samples, all potential contributors to haemolysis have Until recently about 10 diagnostic laboratories in our hospital to be controlled. In this study we aimed to investigate the possible used the same Laboratory Information System (LIS; GLIMS8). association between sample type and haemolysis index. Because each laboratory had set up its own LIS, there was no uniformity and no exchange of specimens across the local lab Methods departments was possible. Creating a common LIS with uniform architecture that enables sample & data logistics across the lab Information on sample type and haemolysis index (H) was departments and guarantees complete sample track & trace for all collected retrospectively from laboratory information system for all labs was necessary to support the recently implemented Total Lab samples admitted at the biochemistry department in 2018. Sample Automation in the complex clinical lab environment of an academic type was classified as follows: VS (venous serum with clot activator), hospital. OP (FX sodium fluoride/potassium oxalate plasma), HP (lithium heparin plasma), EP (K2EDTA plasma) and CS (capillary serum with Methods clot activator); all sampled in VACUETTE®, Greiner Bio-One, Kremsmuenster, Austria. H index was measured on Architect c8000 It was decided to configure a whole new LIS, i.e. GLIMS9, which (Abbott Laboratories, Abbott Park, Illinois, USA) according to supports it all. manufacturer’s declarations. Samples were considered haemolysed if H index was N 0.5 (free Hb N 0.5 g/L). Difference in proportion of Results haemolysed samples between sample types was tested using chi- square test. P level b0.05 was considered statistically significant. Regarding the architecture of the LIS, both specific and general requirements were formulated. Major starting points were: mid- Results dleware is not allowed unless there is added value for patient diagnostics; establishment of the Master-Slave concept, with the LIS Out of 196,122 samples analysed in biochemistry department in as master and automated analyzers and track system as slaves; full 2018, 7.1% of hemolysed samples were identified with constant track & trace on samples along its entire hospital trajectory; a central monthly percentage ranging from 6.1% to 7.3%. Statistically signifi- reception for all samples; a LIS configuration which reflects the cant difference was identified according to sample type (Pb0.001). actual routing of samples and processes in the total test process. An The highest percentage of haemolysed samples was obtained for CS entirely new aspect in the setup of the LIS as Master was the samples (4482/8752; 51.2%), followed by OP (2175/13772; 15.8%), management of a flexible track system connected to multiple pre- HP (65/790; 8.2%), VS (7099/171511; 4.1%) and EP (44/1297; 3.4%). and post-analytical modules, and to several 24/7 analyzers. Impor- tant in the configuration of the middleware of the track was to keep Conclusions it manageable. In the LIS over 10,000 tests were configured for all laboratories, with most of the samples being routed via the track. By The highest proportion of haemolysed samples was identified for using ‘group codes’ for tests where the routing on track is identical capillary serum and oxalate plasma samples. Since oxalate plasma is the configuration was simplified, resulting in about 1,000 tests used exclusively for glucose measurement and glucose is not configured in the middleware. Beyond the GLIMS redesign, the entire sensitive to haemolysis interference, this finding doesn’t require specimen request process, from digital order management in the corrective actions. Problem of poor quality of capillary samples on Hospital Information System (HIS) till reporting of the test results in paediatric and neonatology ward, on the other hand, requires urgent the HIS, was reconfigured and simplified. Other improvements were targeted efforts to reduce haemolysis. synchronization of the phlebotomy sampling time to allow accurate calculation of Total Turn-Around-Times and registration of non- doi:10.1016/j.cca.2019.03.1516 conformities near the patient. Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

M396 doi:10.1016/j.cca.2019.03.1517

ONE-month stability of B-type natriuretic peptide (BNP) in frozen samples using the lumipulse assay M397

V. Ramos-Arenasb, V. Campos-Rodríguezc, P. Pérez-Cañadasb,C. Veriﬁcation of the interchangeability of glycosylated hemoglobin Rodríguez-Rojasb, R. Cárdenas-Gámezb, L. Consuegra-Sáncheza,M. results (% HBA1C) by high performance liquid chromatography González-Moralesb, E. García-Martínb, M.D. Albaladejo-Otónb,L. (HPLC) García De Guadiana-Romualdob aDepartment of Cardiology, Hospital General Universitario Santa Lucia (Cartagena), Spain X. Tejedor Ganduxé, J. Minchinela Girona, M. Granada Ibern, J. b Barallat Martínez De Osaba, A. Leis Sestao, A. Martínez Iribarren, C. Department of Clinical Biochemistry, Hospital General Universitario Santa Lucia (Cartagena), Spain Morales Indiano, M. Llopis Díaz c Laboratori Clínic Metropolitana Nord-Hospital Universitari Germans Department of Internal Medicine, Hospital General Universitario Santa Lucia (Cartagena), Spain Trias i Pujol (Institut Català de la Salut), Spain

Background-aim Background-aim

The main objective of the clinical laboratory is to provide results Natriuretic peptides [B-type natriuretic peptide (BNP) and N- terminal pro-BNP (NT-proBNP)] are useful biomarkers for diagnosis of clinical utility, so it is essential to demonstrate the comparability of the results whenever an analyte can be measured in more than and prognosis of heart failure (HF). Stability of BNP assays must be examined, because there appears to be assay dependence, according one analytical system as in many laboratories, which use different to IFCC recommendations. Due to results reported about the BNP instruments in series for the determination of the same parameter. stability in frozen samples are controversial, the purpose of this They must ensure at all times the interchangeability of the results study was to evaluate its stability in frozen plasma EDTA samples (- over appropriate intervals, to ensure correct interpretation. 20 °C) after a period of 30 days, without the addition of protease Verify the comparability of results between 3 HPLC analyzers, for ﬁ inhibitors, measured with a Lumipulse G BNP assay. direct quanti cation of HbA1c.

Methods Methods

Blood samples were drawn into blood tubes containing ethylene- Based on the EP31-A-IR protocol of the Clinical and Laboratory diaminetetraacetic acid (EDTA), in two groups: group A, including 6 Standard Institute (CLSI), the test was performed on 3 HA8180 apparently healthy individuals, and group B, including 5 patients analyzers (Menarini, Akray) with 2 analytical series in which patient admitted to the Internal Medicine Department with HF. After samples were used at representative values (sample A and sample B) centrifuging immediately, baseline plasma BNP levels were assayed for the evaluation of HbA1c. The 2 levels of internal control (QCI) within 30 minutes from blood collection and one aliquot was frozen supplied by the manufacturer were used as control material, at -20 °C until tested 1 month later. BNP levels were measured ensuring its commutability at all times. on a LUMIPULSE G 600II analyser (Fujirebio), based on a chemiluminiscence immunoassay. Results A change was considered significant when percent deviation (PD %) from baseline BNP level, defined as 1.65*CVA (analytical variation The results for samples 1 and 2 analyzed in the 3 analyzers shows coefficient), was higher than 4.8% and 4.1% in groups A and B, that the number of replicates required based on the calculated respectively. Analysis of results was performed according to the parameter was 15. The acceptability criteria was based on biological criteria proposed by the SEQC-ML. Statistical analysis were per- variability (CVwithin-subject) of the HbA1c in healthy individuals. fi formed using SPSS v. 21.0. The analytical quality speci cations for the allowable difference between analytical systems in our laboratory organization should be Results set at desirable levels of 2/3 CVwithin-subject (CVwithin-subject*2/ 3=1.85*2/3=1.23%). Baseline BNP levels ranged from 4.5 to 13.6 pg/mL [mean The sample A showed an average of the results for each (standard deviation (SD): 10.1 pg/mL (3.5)] in group A and from instrument of 5.78% and the sample B showed an average of 65.3 to 1598.6 pg/mL [mean (SD): 606 pg/mL (546.5)] in group B. 10.83%. And the calculations corresponding to range and range At -20°C, BNP levels remained stable within 30 days from rejection limit were 0.00764 and 0.061 respectively for the sample A collection until measurement in both groups. PD for samples at and 0.02124 (range) and 0.1144 (range rejection limit) for the sample B. -20°C was -4.7% in samples of patients with physiological BNP levels (group A) and 3.4% in samples with pathological BNP levels (group Conclusions B).

Conclusions In this parameter studied, the calculated range has been lower than the range rejection limit, so that the 3 instruments provide BNP levels in frozen plasma samples, without protease inhibitors, results comparable to the evaluated concentration and the compa- remain stable at physiological and pathological levels for up to 1 rability of the methods is considered acceptable, given the known month from collection, as manufacturer recommends. Our study underlying bias between the analyzers. support that freezing of plasma samples is an alternative solution doi:10.1016/j.cca.2019.03.1518 when BNP cannot be measured in the ﬁrst hours from blood collection. Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

M398 Unlike hard copy transmission of VL/EID results via motorcycles, using the electronic results enabled the clinicians and clients to Reducing the viral load and early infant diagnosis results’ immediately have access to the results soon after the specimen is turnaround time in northwestern province, Zambia processed and authorized – reducing the turnaround time to the shortest possible time. A. Temboe, J. Bandaa, S. Musondae, I. Kabongua, H. Kassimd, M. Zuluc, H. Lumanob Conclusions aLaboratory Department, Solwezi General Hospital, Zambia bTechnical Unit, JSI-SAFE HQ Office, Lusaka, Zambia Electronic transmission of VL/EID results system greatly reduced cTechnical Unit, JSI-SAFE, Ndola Office, Copperbelt, Zambia the turnaround time and allowed prompt informed decisions to be dTechnical Unit, JSI-SAFE Solwezi Office, Zambia made by clinicians. It is recommended however, that both methods eTechnical Unit, PCI-HLAB Project, Lusaka, Zambia of results transmission be used concurrently as they supplement each other very well. Background-aim doi:10.1016/j.cca.2019.03.1519 Viral load (VL) is recommended as the preferred monitoring approach to determine the performance of Combined Antiretroviral Therapy (cART) in HIV-infected individuals. According to the Zambia M399 Consolidated Guidelines for Treatment and Prevention of HIV Infection (ZCG) 2018, the first VL is done at 6-months post-initiation, and if VL is less than 1000 copies/ml, 12-months post-initiation and Impact of analytical interference on serum melatonin levels every 12 months, if it remains below 1000 copies/ml. VL monitoring a a b for HIV-infected pregnant and breastfeeding women is done at D. Terzieva , D. Arabadzhiyska , D. Terziev a baseline for known HIV positives. Henceforth, VL to be done every 6 Department of Clinical laboratory, Faculty of Pharmacology, Medical months during pregnancy and breastfeeding period. However, if VL is University, Plovdiv b greater than 1000 copies, a repeat VL is done at 3 months following Second Department of Internal Medicine, Gastroenterology Section, enhanced adherence counseling, until suppression is achieved or the Faculty of Medicine, Medical University, Plovdiv cART regimen is switched. Early Infant Diagnosis (EID) of HIV is done in children below 24 Background-aim months of age born from HIV-positive mothers. In Northwestern province, like Viral Load, EID is done at the Polymerase Chain One of the leading source of errors which impact clinical Reaction (PCR) Laboratory at Solwezi General Hospital through the laboratory results encounter at the pre-analytical phase of testing. use of Dry Blood Spot (DBS) specimen. According to ZCG 2018, Factors as sample collection, improper handling, drugs, hemolysis, Nucleic Acid Test (NAT) is a preferred test for EID because NAT is a lipemia may influence the test result. Melatonin is a principal Point of Care Test (POC) and can be accessible at the point of service secretory product of the pineal gland. The hormone secretion is delivery and offer same-day results. Nevertheless, due to the stimulated by darkness and inhibited by light so that is called “the unavailability of POC NAT, DBS specimen are sent to the PCR hormone of darkness”. The data about analytical interference of Laboratory for processing. DBS specimen are collected from HIV hemolysis, bilirubin and triglycerides on serum melatonin concen- Exposed Infants (HEI) at birth, 6 weeks, 6 months, 9 months, 12 trations are scarce. The aim of our study was to evaluate the effect of months, 18 months, 24 months and 6 weeks after complete this factors on serum melatonin concentrations. breastfeeding cessation. All these DBS and VL specimen are sent to the PCR Laboratory for Methods processing. The turnaround time for these tests results is crucial in management of clients, more especially the HEIs. This is because Venous blood samples were collected from five clinically healthy clinicians’ decisions are solely dependent on these results. The subjects. Two tubes venous blood were taken from each patient. The objective of this study was to assess the impact of electronic VL/EID first tube was centrifuged and serum was obtained. Serum was results transmission in reducing the turnaround time in Northwest- spiked with increased concentrations of bilirubin (Bilirubin standard, ern province of Zambia. cat. N°S-106, Cormay) and triglycerides (Triglycerides standard, cat N° S-113, Cormay). For hemolysis making, the second tube was Methods mixed for 30 min. (Schüttelfrequnz, Germany). Free hemoglobin was measured on spectrophotometer (Specol 11, ref. range: up to 40 mg Direct inward system access (DISA) service at the PCR Laboratory %). Measurement of serum melatonin was performed using ELISA kit was used to populated VL/EID results recently processed. The (IBL, Hamburg, Germany) and Sirio S microplate reader (SEAC, Italy). password encrypted Excel Workbooks for both VL and EID results Results were compared to baseline values and bias (%) was raised from the DISA service were shared with responsible Clinicians calculated. and Data Entry Clerks (DECs) via email and WhatsApp group platform. Unsuppressed VL and DBS Detected spreadsheets are Results separated and colored red for quick intervention. Serum with hemolysis showed decreased melatonin concentra- Results tions in all tested samples (bias: -59.5 - -85.4%). The bilirubin interference is: at spiked concentration with bilirubin 8.267 ⎧mol/l From an average turnaround time of 30 days when using the – bias + 2.95% and with bilirubin 15.782 ⎧mol/l – bias + 17.71%. motorcycles, the turnaround time of electronic results reduced to Triglycerides also allowed positive interference on the melatonin less than 5 days – to allow sample preparation and processing. concentration: bias + 2.58 % (0.238 mmol/l) and + 16.61% (0.454 Otherwise, electronic results transmission can take less than 1 hour. mmol/l). Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

Conclusions sample rejections, quicker receipt and performance of routine and urgent tests. The successful implementation of this process encour- Our results show that hemolysis, triglycerides and bilirubin ages us to extend it to additional departments, including emergency affected serum melatonin concentrations. Hemolysis in specimen is rooms and intensive care units. a negative interferent, and bilirubin and triglycerides are positive interferents. Evaluation of the interference of this factors will help for doi:10.1016/j.cca.2019.03.1521 better clinical interpretation of serum melatonin results. doi:10.1016/j.cca.2019.03.1520 M401

Reliable detection of kEDTA sample contamination requires a M400 routine EDTA assay

Digital laboratory project - A revolution in laboratory services S. Whitehead, C. Ford, R. Gama Dept. of Blood Sciences, New Cross Hospital, Wolverhampton, UK D. Weinstein, R. Geler, M. Prienty, G. Rashid Meir Medical Center, Clalit Health Services, Kfar Saba, Israel Background-aim

Background-aim Potassium-ethylenediaminetetraacetic acid (kEDTA) contamination of serum samples is identified as unexplained hyperkalaemia The laboratory division at the Meir Medical Center provides a and hypocalcaemia and if unrecognised may adversely affect patient wide range of laboratory services to all medical center departments, care. We assessed the value of an EDTA assay in identifying kEDTA while continuously striving to improve its performance. In 2018, we contamination and studied the effect in vitro EDTA contamination on initiated and carried out the Digital Laboratory project in the Internal surrogate serum biomarkers of EDTA contamination. Medicine departments. The main objective was to replace the existing process of filling out test order forms manually by Methods implementing a controlled, electronic system to decrease errors, missing information and processing time in the labs. Planning and Serum spiked with k2EDTA plasma, to mimic sample contamina- initiating the new process involved verity of hospital units, including tion, was used to: 1) study its effect on potassium, calcium, zinc, information management, procurement, logistics, nursing and labo- magnesium and alkaline phosphatase and; 2) derive local cut-offs for ratory. Several stages were required: acquiring equipment, revising classifying EDTA contamination as part of approach to identifying it lab software settings and building new work processes in the labs in routine patient samples. Our current laboratory protocol for and in inpatient departments. identification of kEDTA contamination based on measurement of The aim of the study was to evaluate the success of the project by serum calcium was then compared to that of EDTA measurement. measuring improvements in lab services to the departments included, based on key parameters of sample rejection, turnaround Results and processing time. An EDTA concentration of N0.19 mmol/L was associated with a Methods significant change in serum potassium (increase of 0.54 mmol/L [11.9%]) and zinc (decrease of 71%); no significant reduction in We have based the analysis on data collected from our laboratory measured calcium, ALP and magnesium was observed. The serum information system, regarding these key parameters. EDTA assay detected contamination (ε0.2 mmol/L) in 31/106 patient samples with hyperkalaemia (potassium ε 6.0 mmol/L), 20 of which Results were undetected by the current laboratory protocol.

We measured a 5% decrease in internal medicine departments Conclusions rejected samples, compared with the corresponding period in 2017, before the project was implemented. Turnaround time of chemistry Measurement of EDTA is necessary to reliably identify significant and hematology tubes, showed 40% of tubes were handled by the lab sample contamination from small amounts of k2EDTA and prevent earlier in the morning, compared to corresponding period in 2017. In mislabelling genuine hyperkalaemia as kEDTA contamination. addition, more efficient performance time was measured for several Clinically significant sample contamination occurs from small routine laboratory tests, and successful performance time of 65 amounts of kEDTA which cannot be reliably identified using minutes for urgent marked tests in the system, which accordingly surrogate serum biomarkers. Unexpected hyperkalamia and hypo- require faster execution in the busy morning hours. calcaemia is commonly used to detect k2EDTA contamination however, our spiking data indicate that this will only identify gross Conclusions contamination. Serum zinc has the potential for detecting low levels of EDTA contamination although this is prone to an acute The Digital Laboratory project considers to be a revolution in Meir phase effect. Medical Center lab services and provides efficient and advanced service to the hospitalization departments. Our results indicate doi:10.1016/j.cca.2019.03.1522 progress and improvement in the service, such as reduction in Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

M402 plasma K3EDTA tube and place it on dry ice immediately. A previous study suggests that aprotinin, a serine protease inhibitor, Lithium heparin tubes should not be used for the determination permits to enhance the stability of ACTH. Our laboratory uses of gamma-glutamyl transferase activity K3EDTA + Aprotinin tubes for ACTH determinations in adults, and K3EDTA for paediatric determinations. Our main objective was M. Zendjabila, I. Rikaib, G. Benhamedb to evaluate the ACTH behaviour in different storage conditions and aDepartment of Biochemistry, Oran University Hospital, Oran, Algeria the effect of the addition of this proteolytic enzyme inhibitor bLaboratory of Front de Mer, Oran, Algeria (aprotinin).

Background-aim Methods

Gamma-glutamyl transferase (GGT) is routinely used as a highly We collected venous blood specimens from 10 healthy volunteers sensitive marker of hepatobiliary disease, alcohol consumption and at 8.00 a.m. From each subject, we collected 4 tubes: 2 K3EDTA, and more recently as a predictive biomarker of cardiovascular events. In 2 K3EDTA + Aprotinin. Specimens were immediately centrifuged current practice, most laboratories measure the GGT activity on tube (10 min; 4°C: 2000rpm; 22°C: 3000rpm). Each primary tube was containing lithium heparin. Except that there is no evidence that GGT then aliquoted in 5 tubes, having a total of 20 aliquots per patient. activity can be reliably determined with this anticoagulant. The Each aliquot was left at 2°C to 8°C or at room temperature, during purpose of this study is to evaluate the impact of the anticoagulant different periods of time (0, 2, 4, 8 and 24 hours), and then frozen at on the determination of GGT activity in blood samples. -20°C until analysis (Immulite 2000 XPi, Siemens). Samples were considered stable when they met the formula: Analytical CV% x 1.64 Methods b 13.42. Student’s t-test was used to compare all ACTH concentrations by tube type, at each time point. Blood samples of 30 patients were collected in three types of tubes: dry tube, tube containing lithium heparinate and another Results containing dipotassium ethylene diamine tetraacetate (EDTA). The statistical analysis includes correlation studies, Bland Altman’s A total of 200 aliquots were analyzed after thawing them at room analysis, as well as the comparison of biases to the analytical change temperature. Samples that remained at room temperature main- limit (ACL). tained the ACTH stability during 4 hours. Samples stored at 2°C to 8°C kept the ACTH stability during 8 hours. There was a loss of Results stability after the following hours at those conditions. Results were independent of aprotinin. Results obtained on heparinized plasma are globally over- estimated comparatively to the serum (p=0.001). Whereas, the Conclusions results obtained on EDTA plasma are more comparable to those of serum (p=0.356). Taking the dry tube as a reference, the calculated Storage conditions play an essential role to maintain the ACTH bias for heparinized plasma (+49.55%) samples is larger than the stability. There were differences between storage conditions, prefer- ACL (5.24%), while the bias is about -13.40% for EDTA plasma ring to keep the tubes refrigerated for up to 8 hours until analysis. As samples. no differences were seen among tube types, aprotinin seems to be not necessary, especially at 2° to 8°C storage and for cost- Conclusions effectiveness.

Our results show that the determination of GGT activity must be doi:10.1016/j.cca.2019.03.1524 done on serum samples. If ever plasma should be used, we recommend collecting specimens on EDTA anticoagulant. M404 doi:10.1016/j.cca.2019.03.1523 Evaluation of the sample interference indices (HIL) on the Alinity c system M403 M. Berman, A. Gruszynski Assessment of ACTH preanalytical conditions Abbott Diagnostics Division, Abbott Laboratories, Chicago, USA

M.B. Badal Cogul, D. Morell-Garcia, A. Rubio Alaejos, L. Valiña Amado, Background-aim J.M. Muñiz Fuentes, J.M. Bauçà, A. Barceló Benassar Department of Laboratory Medicine, Hospital Universitari Son Espases, The Alinity c system Sample Interference Indices method provides Palma de Mallorca, Spain a more accurate and consistent method for interpretation of interferents than time consuming visual interpretation. The Alinity Background-aim c HIL method uses speciﬁc wavelength pairs and an algorithm to provide a Sample Interference Index (HIL) that can correlate with Adrenocorticotrophin (ACTH) is a very useful marker for the sample interference due to turbidity, hemolysis and bilirubin present diagnosis of endocrine disorders. As a protein, ACTH seems to be in serum/plasma samples. These values in combination with degraded by the action of proteases. In order to maintain the interference studies can be used to determine the potential for HIL preanalytical stability conditions, it is accepted to collect samples in a interference for Alinity c clinical chemistry assays. Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

Methods were processed in our laboratory accredited by UNE-EN ISO 15189:2013. Calcium, phosphorus and magnesium were measured. The HIL methodology will be described and discussed. The study Aliquots were stored at 2-8°C until an analysis after four days and followed CLSI protocol EP7-A2. Interferences were studied up to seven days.Calcium and phosphorus were measured on cobas 8000® concentrations of 1000 mg/dL for hemoglobin, 1000 mg/dL for c701 (Roche®) using NM-BAPTA assay and molybdate assay, triglycerides, and 30 mg/dL for Bilirubin (unconjugated) in serum. respectively. Magnesium was measured on cobas 8000® c702 Serial dilutions of the sample pools were analyzed in replicates of 4 (Roche®) using xylidyl blue assay. Urinary pH was measured with on the Alinity c system. A cumulative summary was compiled of the pH strips Multistix® 10 SG Siemens and pH-indicator Hemolysis, Lipemia and Icterus indices for about 80 Alinity c Clinical MColorpHast™. We used MedCalc® for Bland-Altman and Passing- Chemistry assays. Bablok between acid and no-acid urines and stability after four and seven days from collection. Results Results Using known concentrations of hemoglobin, bilirubin and Intralipid, the Abbott Semi-Quantitative Index (concentration in mg/dL) and Calcium showed no systematic differences. Proportional differ- Qualitative Index (Blank, 1+, 2+, 3+ 4+) were conﬁrmed on the ences were found between acid and no-acid, after four days in no- Alinity c system. Correlation studies show a linear relationship (r = 1.0) acid, and after seven days for both urines.Phosphorus showed of the indices with increasing concentration of analyte. Using the HIL systematic differences in no-acid urines after four days and in both Qualitative Index scores combined with the speciﬁc assays interference urines after seven days. Proportional differences were found results provided a guide to potential interferents for the Alinity c clinical between acid and no-acid urines, after four days and seven days in chemistry assays. Approximately 50 of the Alinity clinical chemistry both urines. Magnesium showed systematic differences always. assay were not compromised using samples with elevated H, L or I index Proportional differences were found after four days in no-acid urines values (4+ estimates). However, accuracy of about 30 assays (example and after seven days in both urines. albumin) could be over-estimated in the presence of hemoglobin, Bilirubin and/or Lipemia denoted by H indices N/= 4+. Conclusions

Conclusions The differences found between acid and no-acid urines are so slight that we can work with no-acid urines. Samples can be The Alinity c system provides a simple automated procedure for processed until four days post-collection without relevant effects determining the sample indices (HIL) for patient specimens on the Alinity on the stability c system. These HIL values in combination with samples interference studies can be used to determine the potential for HIL interference in the doi:10.1016/j.cca.2019.03.1526 Alinity c clinical chemistry assays to avoid misdiagnosis. doi:10.1016/j.cca.2019.03.1525 M406

Patient safety and pseudothrombocytopenia: Case report M405 C. Collazo Acid and no-acid urines for the analysis of calcium, phosphorus Hospital Álvaro Cunqueiro, Vigo, Spain and magnesium Background-aim Á. Beteta-Vicente, E. Guillén, M. Forrellat, C. Vidal, D. Visiedo, A. Garijo, A. Fernández, M. Nicolás, P. Cañero EDTA-dependent pseudothrombocytopenia is a rare phenomenon CATLAB of in vitro platelet agglutination due to the presence of antiplatelet antibodies. In these cases, the peripheral smear must be observed in Background-aim order to confirm the presence of platelet aggregattes. In addition, blood sample anticoagulated with sodium citrate is the most suitable Calcium, phosphorus and magnesium in urine can be required for the sample for the platelet count. diagnosis of renal pathology. Hydrochloric acid (HCL) is a preservative The laboratory has the responsability of detecting these false avoiding the precipitation of calcium and keeping phosphorus in thromcytobopenias, and reflecting this information in the laboratory dissolution. Inserts indicate acid urines to analyze these ions, but it is report. Likewise, the clinician should consider this condition in order tedious in clinical laboratories. Acid is a corrosive that need to be handled to avoid innecessary diagnostic tests and procedures with the aim of specifically. Most measurements in urine are required in no-acid urines. preserving preserve patient safety. Acidifying urine means more containers, analysis time and storage space. Many laboratories use no-acid urines to analyze these ions. We Methods study pH influence on the accuracy of analysis for these analytes. We evaluated stability after four and seven days from collection. A 8-year old boy was referred to paediatric onco-hematology for presenting a two-year evolution thrombocytopenia with a platelet Methods count below 20 . 10^3/μL. He did not have a significant past medical history, besides his familial hypercholesterolemia, and his clinical Fifty five urine samples were chosen without preservative and examination was unremarkable. centrifuged at room temperature 10min 1760g. We prepared two The inicial working diagnosis was of a PTI and an aspiration aliquots with 1 mL, one acidified with 10 ⎧L HCL 20%. Both aliquots biopsy of the spine was made. There were no significative anomalies Abstracts / Clinica Chimica Acta 493 (2019) S673–S710 in the sample obtained. In that moment the clinician realized that in Results all the previous laboratory reports, there was a commentary on the platelet count remarking the presence of plentiful platelet aggre- After the implementation of this filter, half of determinations are gates. The hemogram and the blood smear was repeated, and a made. The majority of Intensive, Resuscitation, Internal Medicine and platelet count was made in a sample with sodium citrate, obtaining a Geriatrics, Pediatrics, Urgencies and Hematology, although there are normal count (160 . 10^3/μL). of all areas. The most requested request motive was “septic patient control”, followed by “suspicion of sepsis” Results Conclusions According to the WHO, patient safety is the absence of preventable harm to a patient during the process of health care In recent years, the need to assume has arisen the responsibility and reduction of risk of unnecessary harm associated with health of reviewing and evaluating analytical determinations by laboratory care to an aceptable minimum. professionals. A more unnecessary requests, higher risk and more The clinic laboratory works an importan role over patient safety, chances of error. not only in the analytical and preanalytical phase, but also in the The best strategies for the control of the demand of testing are -sometimes forgotten- post-analytical phase. Likewise, we should the act when the requesting doctor makes the request. make sure our laboratory reports are understood by clinicians and After detecting the high demand for procalcitonin, the need arises consider offering complementary tests if necessary. for its modulation without impairing the attention to patients and clinicians. A modulation of the preanalytic demand is decided, which Conclusions allows the application with clinical justification, since it is handled by most of the units. From the laboratory, we should enhance our relationship with clinicians trying to avoid misunderstandings as the reflected in the doi:10.1016/j.cca.2019.03.1528 case report. It should be remainded them that in those cases with EDTA-dependent pseudothrombocytopenia, a blood smear is mandatory to confirm platelet aggregates and that the evolution of the M408 platelet count requires a sample with sodium citrate. doi:10.1016/j.cca.2019.03.1527 Hemolysis index interfered by high bilirubin: A patient safety problem

N. Del Amo, R. Ramos, M.B. Alvarez, E. Marquez, M.J. Ruiz, R. Guillen, M407 F. Cava Laboratorio Central de la Comunidad de Madrid- BRSalud, Spain Demand modulation`s of procalcitonin from electronic request Background-aim S. De Las Heras Florez, M. Carretero Perez, C.T. Sanz Diaz, E. Mateos Rodriguez Nowadays serum indices (HIL), hemoglobin (H-index), icterus (I- Servicio de Análisis Clínicos, Hospital Universitario Nuestra Señora de index) and lipemia (L-index), have been automatized in many Candelaria, Santa Cruz de Tenerife, Spain laboratories for evaluating the sample quality. Data of HIL indices measures condition the rejection or acceptance of laboratory test Background-aim results. However, HIL indices are susceptible to interference and the laboratory must ensure the results accuracy so that patient safety During the last years the assistance activity and the demand for won’t be compromised. laboratory tests have increased, without being associated with an increase in economic resources. That makes it necessary to use Methods management tools, especially in expensive tests such as procalcitonin, a useful biomarker in the diagnosis of sepsis and A specific protocol was computerized in the laboratory informa- infections in other locations. tion system according to the results of an internal interference study In our laboratory, procalcitonin was introduced in April-2013, performed following CLSI recommendations in a Dimension EXL restricted to Paediatrics, Intensive and Resuscitation. It was possible system (Siemens Healthcare SLU). For the indices estimation, the to apply for any clinician, increasing their demand and generalizing analyzer adds water to 20 ⎧L of sample and the indices measure- in practically all areas, with 8,793 requests in 2017, more than a half ment wavelengths are 405 nm (hemoglobin), 452 (bilirubin) and from services not agreed at the beginning. After this increase, a 700 (lipemia/turbidity). preanalytic demand modulation protocol was established. Results Methods After protocol implementation, direct observation detected icteric A request filter was made. In the computer system procalcitonin samples (I index N or=3, equivalent to a bilirubin concentration is an open test. Upon request, a pop-up window opens: “The from 5 to 20 mg/dl) leaded to a hemolytic index N or = 3 (equivalent determination of procalcitonin is useful for the diagnosis and to a hemoglobin concentration from 50 to 200 mg/dl) without monitoring of the septic patient. Do you want to continue the hemolysis. Lactate dehydrogenase, aspartate aminotransferase, po- application? “, Once accepted, a drop-down will appear to choose a tassium, sodium and creatinine were incorrectly rejected or flagged reason for request:” suspicion of sepsis, treatment control in septic with a possible hemolysis interference comment. The algorithm was patients, fever and leukocytosis without focus, others “ modified and a flag is now displayed with a recommendation for Abstracts / Clinica Chimica Acta 493 (2019) S673–S710 hemolysis visual inspection when H index is suspected to be falsely was mostly affected during storage, regardless the condition, while in increased due to high bilirubin. PPT sample miRNA expression levels remained stable. Also for the exosome-associated miRNA fraction, samples collected in 2-K2E Conclusions displayed the lowest detectability and stability but, in this case, the storage conditions had comparably no effects on miRNAs stability in Although modern analyzers are able to identify HIL interference PPT and 1-K2E samples, although it resulted slightly improved in 1- much more accurately than the traditional visual approach, indices K2E samples. overlap may exist and visual inspection may still be required. HIL indices interferences are manufacturer and analyzer depen- Conclusions dent, so every laboratory must be aware of the interference and the mechanism in order to establish the suitable strategy for it detection Taken together these results indicate that blood collection in PPT, and correction. i.e., higher amount of spray-coated EDTA, guarantee a greater stability for free-circulating miRNAs (i.e., the biomarker-like frac- doi:10.1016/j.cca.2019.03.1529 tion), compared to K2E samples, and also for microvesicle-associated miRNAs (i.e., the hormone-like fraction) the performance of this tube are good. On the contrary, a stepwise centrifugation strongly affected miRNA detectability and stability of miRNAs and, therefore, it may M409 not be applied in miRNA testing.

Effects of common pre-analytical variables on detectability and doi:10.1016/j.cca.2019.03.1530 stability of microvesicle-associated and free circulating miRNAs

M. Faraldib, V. Sansonib, S. Peregob, M. Gomarascab, J. Kortasa,E. Ziemanna, G. Banfid, G. Lombardic M410 aGdańsk University of Physical Education and Sport, Gdańsk, Poland bLaboratory of Experimental Biochemistry and Molecular Biology, IRCCS How different normalization strategies affect circulating miRNA Istituto Ortopedico Galeazzi, Milano, Italy quantification? AN explicative study on sedentary and highly- cLaboratory of Experimental Biochemistry and Molecular Biology, IRCCS trained subjects Istituto Ortopedico Galeazzi, Milano & Gdańsk University of Physical Education and Sport, Gdańsk, Poland M. Faraldia, M. Gomarascaa, V. Sansonia, S. Peregoa, G. Banfic,G. dLaboratory of Experimental Biochemistry and Molecular Biology, IRCCS Lombardib Istituto Ortopedico Galeazzi, Milano & Vita-Salute San Raffale Univer- aLaboratory of Experimental Biochemistry and Molecular Biology, IRCCS sity, Milano, Italy Istituto Ortopedico Galeazzi, Milano, Italy bLaboratory of Experimental Biochemistry and Molecular Biology, IRCCS Background-aim Istituto Ortopedico Galeazzi, Milano & Gdańsk University of Physical Education and Sport, Gdańsk, Poland Pre-analytical phase standardization is a main requirement for cLaboratory of Experimental Biochemistry and Molecular Biology, IRCCS the clinical implementation of biomarkers and, specifically, of Istituto Ortopedico Galeazzi, Milano & Vita-Salute San Raffale Univer- circulating miRNAs. miRNAs are present in blood as free (e.g., sity, Milano, Italy associated with argonaut protein, low-density lipoproteins) and microvesicle-associated forms. Based on their release, free-circulat- Background-aim ing miRNAs are considered biomarkers while microvesicle-associated miRNAs mostly carry out the role of hormone-like compounds. The physical activity status represents an important pre-analyt- Aim of this study was to evaluate the effects of common pre- ical variable that might be considered in a clinical setting especially analytical variables on detectability and stability of a panel of in the case of the measurement of those biomarkers whose microvesicle-associated and free circulating miRNAs. circulating levels are strongly modified by chronic training. Hence, sedentary and highly trained subjects are thought to display really Methods different circulating miRNA profiles. However, the use of circulating miRNAs as biomarkers is still limited by the lack of standardized Venous blood from 10 male volunteers was collected into post-analytical data processing. This work aimed at comparing the K2EDTA tubes and plasma preparation tubes (PPT), containing a effect of different normalization approaches on RT-qPCR data. In double amount of EDTA. Plasma was obtained by centrifugation order to identify the most appropriate strategy to obtain the most according to the manufacturer’s indications (1-K2E, PPT), while an reliable results, normalization methods based on endogenous and aliquot of K2EDTA plasma was further centrifuged (2500g, 15min, exogenous miRNAs were assayed in order to evidence the differences room temperature) to deplete the platelets (2-K2E). Samples were in the expression level of free-circulating miRNAs between sedentary immediately frozen or stored for 24h at either room temperature or and highly-trained subjects. 4°C. The microvesicle-associated and free circulating fractions of 179 miRNAs was assayed by RT-qPCR. Methods

Results Circulating miRNAs were extracted from plasma of 10 sedentary and 14 trained males. A panel of 179 miRNAs (i.e., the most Detectability of free miRNAs was greater in PPT samples than in 1- represented in blood) was assayed by RT-qPCR. The relative expres- and 2-K2E samples; particularly, 2-K2E samples displayed the lowest sion of each miRNA was calculated by the 2-⊗⊗CT method using, as detectability over all the conditions. Detectability in 1-K2E samples normalizer, the average value of all expressed microRNAs, the averaged expression values of the most stable endogenous miRNAs, Results Abstracts / Clinica Chimica Acta 493 (2019) S673–S710 BT reduced bubble formation. For HT, poor stability was observed the single most stable miRNAs, and exogenous oligonucleotides. The for glucose determination while a better stability was found for BT. fi identi cation of the most stable miRNAs were performed using the This difference in stability might be accounted for by a better NormFinder algorithm, ranking the miRNA based on their expression separation between plasma and cells thus reducing glycolysis by stability. leukocytes. Interestingly, a significant decrease in hemolysis index was noticed for BT. CAT reduced hyperleukocytosis interference Results compared to HT. We hypothesized that the clot protects from release of cell contents and that the separator in BT might perform the same Different normalization methods lead to different, and sometimes function. In patients with hyperleukocytosis, there may be a contrasting, results. Among the endogenous-based methods, nor- reduction in interference with BT tubes compared to HT tubes but malization on the single miRNA hsa-miR-320d resulted the most this needs to be confirmed by further study. appropriate strategy giving the lowest technical variability among the replicates and the best method to highlight the existing Conclusions differences in the miRNA profile between the two sample popula- tions. On the contrary, exogenous oligonucleotides resulted the The new BT offers advantages in terms of a reduction in the time worst normalization methods. spent on laboratory management, a reduction in pre-analytical issues, higher speed/lower time centrifugation and better stability. Conclusions BT could be an alternative for laboratories that would like to gain the benefits of separator tubes but do not want to use gel-separator This study highlights the importance of an accurate choice of the tubes. appropriate normalization strategy for each experimental setting in order to avoid misleading interpretation of RT-qPCR data. doi:10.1016/j.cca.2019.03.1532 doi:10.1016/j.cca.2019.03.1531

M412

M411 Preanalytical interventions when using automated biochemistry tracks BD-BARRICOR® tubes in biochemistry emergency: A new window opening on blood sample processing and analysis C. Ison, R. Flatman, D. Kanowski, A. Jennings, G. Ward, L. Price Biochemistry Department, Sullivan Nicolaides Pathology, Bowen Hills, G. Grzychd, C. Herbauxc, G. Smitha, C. Sosnowskib, C. Tiernyb,T. Brisbane, QLD, Australia Brousseaub, P. Maboudoub aDepartment of Emergency and SAMU, CHU Lille, Lille F-59000, France Background-aim bDepartment of Biochemistry, CHU Lille, Lille F-59000, France c Department of Clinical Hematology, CHU Lille, Lille F-59000, France Sullivan Nicolaides Pathology receives and collects approximately d CHU, Laboratory of Endocrinology, Metabolism-Nutrition, Oncology, 20,000 pathology samples per day from hospitals, GP clinics and University of Lille, Lille, France community blood collection centres over an area in North Eastern Australia the size of Western Europe. In general only one serum Background-aim (SST) and one EDTA tube are collected for each patient episode. The pre-analytical phase of the pathology testing cycle is a significant If a PTS (Pneumatic Tube System) is used to send Heparin tube source of laboratory error. Pilot studies in 2016 demonstrated that (HT) samples to laboratory then pre-analytical complications, such significant resources were required for pre-analytical troubleshoot- as debris, fibrin or foaming and bubbles may occur. These may ing including sample recollection and erroneous results. The interfere with analyses and require extra manipulation and re- introduction of a GLP track for distribution of samples for automated analysis, causing an increase in turn-around time and contamination. biochemistry has further highlighted the need to identify samples BD-Barricor® tubes (BT) are a heparin plasma tube based on using a which are unsuitable for immediate loading onto a track. Identifica- mechanical barrier to separate the cellular material from the plasma. tion of the major sources of unsuitable samples will allow BT was compared to non-gel HT and non-gel serum Clot Activator determination of strategies to minimize unsuitable samples which Tube (CAT) in an emergency biochemistry laboratory department, in turn will lead to the reduction of pre-analytical laboratory errors. including the effect of tube type on hyperleukocytosis interferences. This study provides data on unsuitable samples from a private The aim is to provide an alternative separation method for community reference laboratory. laboratories that do not wish to use tubes with gels. Methods Methods Further studies identified that 5.4 % of samples referred for Blood was collected from 197 patients in four departments of Lille automated biochemistry could not be immediately loaded onto a University Hospital and transported by PTS. Samples were collected track and required human intervention. Centrifugation for biochem- into 3 tube types: BT, HT without gel and CAT and were analyzed istry samples is usually performed at the point of sample collection. simultaneously. Pre-analytical issues were evaluated. A stability Statistics on the number of samples requiring human intervention study was performed after 12h storage at RT without sealing. were collected for the 12 month period February 2018 - January Hyperleukocytosis interferences were assessed in a patient cohort 2019. (N50x109/L). Results Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

Overall 3.8-5.5 % of all samples required human intervention. The 5. EDTA tubes labeling: 400 samples classified by Hematology area main causes (monthly range; lowest –highest) over 12 months were and new immunology label design. AP as follows: insufficient sample (2.3-3.1%); unspun tube (0.7-1.0%); 6. Sample processing systems (Beckman Coulter): tube re-spin (0.3-0.9%); and fibrin clot (0.3-0.8%). a. Improvement of entry/exit samples flow and increasing tubes ‘rack capacity. AP Conclusions b. Automation of secondary tube plugging, new treatment in the tubes where a fibrin clot is found and priority samples rack These results have identified areas for further education of implementation. NIY collectors to improve sample quality. Graphical KPI reports have c. Automation of preferential treatment for priority samples. R been implemented for each phlebotomist throughout the practice. 7. IPS management: New workflow defined for priority samples. NIY 8. Centrifuge: Acquisition of a new one with high capacity to improve doi:10.1016/j.cca.2019.03.1533 workflows. AP 9. Preanalytical area space design: reorganization of spaces in order to improve workflows. NIY

M413 Results

Preanalytical area: Kaizen methodology implementation Defined KPIs showed that all 3 objectives had been achieved. Planned actions for CCP: AP:47%. NIY:40%. R:13%. L. Jimenez AÑon, C. Gomez Gomez, J. Barallat, M.A. Huertas Contreras, C. Rodriguez Vaca, E. Arias Puig, N. Calderon Cervantes, Conclusions M.A. Llopis Diaz Clinical Core Laboratory, Laboratori Clinic Metropolitana Nord, Bada- Kaizen methodology is a useful tool to improve workflows and lona, Spain utilization of resources and to increase efficiency.

Background-aim doi:10.1016/j.cca.2019.03.1534

Kaizen methodology is a quality strategy based in continuous improvement by reducing waste (muda), eliminating unnecessary M414 hard work and humanizing workplaces (gemba). The aim of this study is to implement Kaizen methodology in a ﬁ preanalytical area (9000 samples/day). Preanalytical aspects of investigations of prostate-speci c antigen in distillated and deionized water extracts of semen in forensic Methods samples

a b The methodology was implemented by performing the following V. Sidorov , L. Khorovskaya a steps: Forensic biological division, Saint-Petersburg State Healthcare Instituition Bureau of Forensic Medical Examination, Saint-Petersburg, 1. Objectives and key performance indicators (KPI) definition Russia 2. Process map bLaboratory Medicine Department, North-Western State Medical Uni- 3. Critical control points (CCP) detection versity named after I.I. Mechnikov, Saint-Petersburg, Russia 4. CCP categorization 5. CCP prioritization Background-aim 6. Action plan (APL) development 7. . APL launch Sperm presence establishment in forensic casework demands 8. KPI evaluation operative preanalytical stage with maintaining stability of Prostate- Specific Antigen (PSA). The aim of the study was to investigate Objectives and KPI were defined as: stability and activity of total PSA in dry sperm spots, extracted with deionized water in comparison with distillated water. Deionized 1. In patient samples (IPS) receiving time b 2 hours (KPI-1: Receiving water helps to minimize microbial contamination that could improve time) stability of the samples. 2. Decrease in samples arrival time to different laboratory analytical areas (KPI-2: % of samples into the analyzers) Methods 3. 90% samples received before 2:00 pm (KPI-3: % of samples received) Semen of volunteers diluted in distillated and deionized water in proportions of 1+1 (high concentration) and 1+25 (low concentra- Action plan: CCP (high-low impact): tion) was put on sterile gauze (0.5 sm x 0.5 sm) in volume of 1 μL 1. To avoid errors in specimen’s delivery by hospital orderlies: and dried in temperature 18-20° C for 3-4 hours. Samples were Sample delivery racks according to sample type. Action performed extracted with 100 μL of distillated and deionized water in +4 °C for (AP) 18 hours. The extracts were aliquoted to five portions per each level 2. Specimen’s delivery to different laboratory analytical areas: of concentration and stored at +18-20° C during 6 hours. PSA Schedules and responsible persons were established. AP amount was measured by an enzyme-linked immunosorbent assay 3. Removal of specimens from transport containers: New model method (ELISA) in the two obtained concentrations (26.3 ng/ml and design. Not implemented yet (NIY). Awaiting public tender. 5.4 ng/ml) five times during one day. The average, within- and 4. Transport routes analysis: Route evaluation to advance samples between series imprecision and percentage of lost PSA activity were arrival time. Rejected (R) calculated. Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

Results 307 adults (median (range) age: 47.7 (18-87.7) years) and 80 children (13.6 years (6 months-17.9 years)). A subset of this The within series CV% was bigger for low concentrations PSA population also had assay data for C1 inhibitor protein (96 adults (pb0.05) for samples extracted by distillated water compared with (31.3%) and 28 children (35.0%). As previously reported in the deionized (7.8% and 4.0% respectively). Between-series variation was literature, no need for age partitioning was evidenced for CP50, C3c bigger for both levels of concentrations in samples extracted by and C4. The RIs [90% confidence intervals] calculated from the pooled distillated water (pb0.05): 28.3% and 13.2% for low and 15.2% and data are: 35.4 [33.1; 37.2] to 76.3 [73.7; 83.6] U/mL for CP50 activity, 6.7% for high concentrations respectively. Activity of total PSA were 803 [745; 869] to 1641 [1587; 1715] mg/L for C3c and 119 [103; 135] significantly decreased (pb0.05) in samples extracted by distillated to 380 [364; 397] mg/L for C4. water compared with deionized for low concentrations of PSA: 64.9% Conversely, our results highlighted a positive association between and 81.2% after 3 hours and 51.6% and 71.6% after 6 hours age and C1 inhibitor protein levels. We therefore derived three age respectively. partitions and the related RIs: from 6 months to 30 years (220 [201; 241] to 388 [363; 411] mg/L), from 30 to 50 years (220 [201; 241] to Conclusions 388 [363; 411] mg/L), and after 50 years (246 [217; 272] to 412 [400; 431] mg/L). Samples of semen on dry spots for testing of total PSA extracted with deionized water at room temperature (+18-20° C) are more Conclusions stable compared with distillated water and could be stored up to 4 hour 30 minutes during working day while PSA extracts with We provide RIs for complement component assays in EDTA distillated water could be stored no more than 1 hour 30 minutes. plasma samples. C1 inhibitor protein levels were linked to age, and Thus implementation of deionized water for PSA extracts from so the use of age-specific RIs is mandatory. These RIs will be of value semen on dry spots improves the quality of preanalytical phase. for the diagnosis of complement-related diseases and for the accreditation of laboratories. doi:10.1016/j.cca.2019.03.1535 doi:10.1016/j.cca.2019.03.1536

M415 M416 Classical pathway activity, C3C, C4 and C1-inhibitor protein reference intervals using Optilite® reagents in EDTA plasma Strategy to avoid applying “In Vitro” hemolysis algorithms in “in vivo” hemolysis B. Lopeza, N. Bertiera, E. Ledoultb, R. Joudinauda, M. Maanaouic,E. Moitrota, A. Deleplancquea, S. Rogeaua, D. Launayb, G. Lefèvrea,M. M.E. López-Guío, M.J. Sacedo Calvo, A. Hernandez Jiménez, S. Buendía Labalettea, S. Dubucquoia Martínez, R. Gonzalez Cervera, J. Asensio Antón aDepartment of Immunology, CHU Lille, Lille F-59000, France Hospital Niño Jesús, Spain bDepartment of Internal Medicine and Clinical Immunology, CHU Lille, Lille F-59000, France Background-aim cDepartment of Nephrology, CHU Lille, Lille F-59000, France The aim of this study is to avoid eliminating results in patients Background-aim with “in vivo” hemolysis (IVVH) and altered hemolytic index (HI) due to the algorithms applied currently to “in vitro” hemolysis Although complement assays are now part of the routine (IVTH), using hemoglobin value. immunological assessment in many fields of medicine, the results Hemolysis is a common interference in the medical lab, and it can are very sensitive to pre-analytical handling. EDTA-containing produce spurious results due to several mechanisms. collection tubes are known to stabilize the analytes and to make In order to systematize and standardize the actions related to the the results more reliable. However, reference intervals (RIs) for EDTA IVTH, automatic algorithms based on the HI has been developed in our plasma samples have not previously been published. Thus, the lab. In this regard, results are informed with a comment if the interference objective of the present study was to determine RIs for classical is analytically significant and not informed if is clinically significant. pathway (CP50) activity and C3c, C4 and C1 inhibitor protein levels Because the idiosyncrasy of our hospital a strategy to inform in EDTA plasma samples, using Optilite® reagents from The Binding results in IVVH is mandatory. Site Group Ltd. HI is not always altered in patients with IVVH. However, for those with HI above 0, the identification of the clinical situation is essential. Methods Methods We retrospectively evaluated a large cohort of patients attending our university hospital and known to be free of complement- 507 request with IVVH diagnosis confirmed were selected from associated diseases. The need for age partitioning was assessed and January 2009 to December 2018. This requests belonged to 154 RIs were calculated for each parameter according to the Clinical and patients. Requests without HI, without hemogram, diagnosis not Laboratory Standards Institute’s C28-A3c protocol. clear or not confirmed and other interferences (as a lipemic serum) were discarded. Results From those, requests with a HI above 0 were selected (187 request from 78 patients) A total of 387 samples with assay data for CP50 activity and C3c and C4 protein levels were used to derive the RIs. These came from Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

As a IVTH group, 906 request (807 patients) from the same period presented as 95% intervals (2.5th-97.5th percentiles), age is pre- with a diagnosis non related with IVVH, apparently healthy and with sented as medians (25th-75th percentiles). a HI of 1 or above were selected. Statistical analysis was performed by Stata. HI was performed by Results AU680 from Beckamn coulter. RIs were calculated in four age groups, month: “b12”–4(4-5), Results 12-60 – 31(26-36), “60-120”–98(78-108) and “N120”–138(131- 158). No age and gender differences had RIs of MONO (0.37-1.26) Hemoglobin in the IVVH group was 9.11 g/dL (8.8-9.42; 95% IC) x109/L, BASO (0.00-0.05)x109/L, MCHC (320-353) g/L, RDW-CV (12- and in the IVTH group 13.78 g/dL (13.69-13.88; 95 % IC). Differences 14.5)%, RET (17.8-64.9)x109/L, LFR (91.7-99.3)%, IRF (0.7-8.3)%. RIs between groups were significant (pb0.05). in the age group “b12” and “12-60” months had no differences in Based on the promedio of hemoglobin in IVVH group, it was WBC (6.0-13.6)x109/L, EO (0.07-0.88)x109/L, PLT (267-580)x109/L, established 9.5 g/dl as a cut off to set up an alert during validation HGB (111-133)g/L, RBC (4.02-5.30)x1012/L, HCT (32-40)%, MCV process. The aim of this alert is to identify IVVH and do not treat (71.5-86.3)fL, MCH (23.9-29.3)pg. RIs in the age group “60-120” and this patients as an IVTH, and therefore do not apply hemolysis “N120” months had no differences in WBC (3.9-11.5)x109/L, EO algorithms. (0.02-0.65)x109/L, PLT (175-436)x109/L, RBC (4.09-5.33)x1012/L, A simulation was performed with all the requests from 2017 with MCH (24.9-30.4)pg. In the age group “b12” months RIs for NEUT was an altered HI. And the cut off of 9.5 g/dL evidenced a sensibility: 88.5 (0.6-3.6)x109/L, in children older than 12 months RIs NEUT was % and a specificity: 64%) (1.1-5.8)x109/L. Absolute count of LYMPH was different in all the age groups: “b12” (3.3-9.0)x109/L, “12-60” (1.6-7.1)x109/L, “60-120” Conclusions (0.9-5.0)x109/L, “N120” (1.1-3.8)x109/L. Red blood cells parameters differed in two groups: “60-120”–HGB (114-147)g/L, MCV (76.6- The strategy has certain limitations, patient with transfusions has 89.7)fL, HCT (35-43)% and “N120”–HGB (118-155)g/L, MCV (78.6- not been identified and we are not able to identify the 100% of the 92.9)fL, HCT (37-46)%. IVVH. Patient with both IVVH and IVTH can not be identified in an CONCLUSIONS isolated drawn, so if suspected in cases with HI above 1, new specimen need to be requested if possible. These data will help to implement hemogram parameters in the It is a good strategy for patient safety in order to treat IVVH everyday clinical practice needs. separately, with this strategy we would have detected 88% of HVVI with HIN=1. doi:10.1016/j.cca.2019.03.1538 doi:10.1016/j.cca.2019.03.1537

M418

M417 On the lookout for the best candidate material to develop a certiﬁed reference material for PR3 ANCA IGG antibodies: A Pediatric reference intervals for hemogram parameters commutability story

O. Melnichuk, N. Mayanskiy E. Monogioudib, J. Sheldonc, P.L. Meronia, I. Zegersb Laboratory Department, Scientific Center for Children’s Health/Moscow, aDepartment of Clinical Sciences and Community Health, Istituto Russia Auxologico Italiano, Milan, Italy bEuropean Commission, Joint Research Centre, Geel, Belgium Background-aim cProtein Reference Unit and Immunopathology Department, St Georges’ Hospital, London, UK Hemogram analysis is frequently used in practice by clinicians of all specialities. In order to interpret the results of hemogram Background-aim parameters correctly, it is necessary to know reference intervals (RIs). Autoantibody measurement is key for the diagnosis of autoimmune diseases. It is essential for these measurements to be reliable Methods and on the longer run standardised. Standardisation can be assisted through the use of an appropriate reference material (RM). A useful Venous blood samples were taken from 286 healthy children RM must fulfil a number of criteria. It must be homogeneous, stable (119 female, 167 male) at prophylactic medical examination. All at certain temperatures and preferably over longer periods of time. It samples were measured by a Sysmex XT-2000i analyzer. RI were must be fit for purpose and commutable, i.e. to behave as any calculated according to CLSI C28-A3 standard for the following patient sample. The Joint Research Centre of the European parameters: leukocytes (WBC), absolute number of neutrophils Commission has released a RM certified for PR3 ANCA IgG (NEUT), lymphocytes (LYMPH), monocytes (MONO), eosinophils autoantibodies (ERM-DA483/IFCC). Antibodies that are involved in (EO), basophils (BASO), platelets (PLT), hemoglobin (HGB), erythro- cases of small vessel associated vasculitis. Thus, prior to anything cytes (RBC), hematocrit (HCT), mean corpuscular volume (MCV), else, commutability of different possible formats for this material mean corpuscular hemoglobin (MCH), mean corpuscular hemoglo- were assessed. bin concentration (MCHC), red blood cell distribution width (RDW- CV), absolute reticulocyte number (RET), low fluorescence reticulo- Methods cyte fractions (LFR), immature reticulocyte fraction (IRF). RIs are Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

Commutability was assessed in 2 separate studies. During the first can serve as an effective quality indicator. In this study, we have one, 18 formats of the candidate RM were tested with eight compared and analyzed data collected from two tertiary care commercial immunoassays alongside 30 clinical samples. During hospitals with a primary focus on ruling out reasons for rejection the second study, 3 of the most promising candidate RMs and and identifying specific volatile test groups. dilutions thereof were analysed with another set of 30 clinical The aim of this study is to establish a cohort relation between the samples with three immunoassays so as to select the format that rate of sample rejection and quality indicators with an evidence- would be best for preparing the CRM. The data derived from these based explanation for rejected samples. studies were compared in a pairwise manner and Pearson’s Primary objective – Primary objective of this study is to develop a correlation coefficient for the samples of all pairs was calculated. Standard Operating Protocol (SOP) to answer and check all the specific Additionally, the intermediate precision and repeatability were causes of sample rejection & improve Total Test Process (TTP). evaluated. Secondary objective – 1. Check for novel factors for improving TTP, depending upon the Results comparative variables & compounding from the two tertiary care hospitals. Commutability can be analysed with different statistical ap- 2. To validate the SOP and minimizing unintelligible errors and proaches. Through the studies, the formats that led to results not improving the quality of sample/specimen collection. being within the 95% prediction interval of a pairwise analysis were rejected. The inter- and intra-plate variations were below 7% for all Methods assays. Eventually one format was selected for the candidate CRM, it was a plasmapheresis material with high concentration of PR3 ANCA This study is based upon retrospective data analysis of all the IgG that was converted into serum and freeze-dried. biochemistry samples received and the total number of samples rejected in 1 year from clinical biochemistry laboratories of two Conclusions tertiary care hospitals. Both hospitals are from Southern Karnataka region, one being 1800 bed & the other is 850-bed strength. The fi The results con rmed that there is a large variability in PR3 ANCA clinical biochemistry laboratory of the respective hospitals are IgG testing despite the overall good performance of individual following NABL (National Accreditation Board for Testing and immunoassays. However, through properly designed and performed Calibration Laboratories) guidelines for sample collection & rejection. commutability studies a suitable format for a CRM could be selected. This descriptive study is based on the analysis of different rejection Together with the detailed instruction for use, the CRM produced rate, types of unintelligible approach and level of inappropriateness. (ERM-DA483/IFCC) will allow control of the variation in PR3 ANCA Assays and blood samples of clinical chemistry, hormonal assay, IgG measurements. hematology, HbA1c, urine chemistry, arterial blood gas (ABG) etc. have been evaluated. We have used predefined criteria for sample doi:10.1016/j.cca.2019.03.1539 rejection to assess the level of appropriateness, which are hemolysis, insufficient volume, clotted, wrong vacutainer, mis-match, venous blood (for ABG) and test raised by mistake. M420 After calculating the mean rejection rate we compared the data with other national and international laboratories. Sample rejection rate as a major quality indicator: Comparative findings from clinical biochemistry laboratory of two tertiary Results care hospitals from Southern India The total number of clinical biochemistry laboratory samples from P. Doddamania,b, V.A. Ojhaa,b, A. Prashanta,b, P.A. Manjrekara,b,A. test group A received from 1st January 2016 to 31st December 2016 is Ranjana,b 1,62,079 of which 1617 (per year) samples were rejected with annual aDepartment of Biochemistry, JSS Medical College, JSS Academy of sample rejection rate is 1% (0.997%). In this study, we have followed Higher Education & Research, Mysuru, Karnataka, India the standardized sample rejection criteria, among which hemolysis bDepartment of Biochemistry, Kasturba Medical College, Manipal (50%) is the most common reason and mismatch (0.61%) is the least Academy of Higher Education & Research, Mangalore, Karnataka, India common reason. Among the inpatient departments, medicine has the highest percentage (31%) and pediatrics has the lowest percentage Background-aim (0.86%). The total number of clinical biochemistry laboratory samples from test group B received from 1st January 2016 to 31st December In a clinical biochemistry laboratory, there are many determi- 2016 is 3,65,000 of which 3,600 (per year) samples were rejected with nants responsible for sample rejection, the majority of them falling an annual sample rejection rate of 1% ( 0.9863%). into the pre-analytical phase of errors. With advances in Laboratory After comparing grouped criteria of sample rejection with other fi & Clinical Medicine, clinical diagnosis is largely based upon laboratories we found that hemolysis, insuf cient volume, and biochemical investigations. Collecting and analyzing data consis- clotted sample are common. Beside common criteria, some other tently are necessary tasks for assessing quality, monitoring stan- causes like the venous sample for arterial blood gas analysis with a dardized key processes, improving performance and patient safety in rejection rate of 12% are the 2nd most common cause. clinical laboratories. With a few comparative studies available, it has been found that Conclusions the rate of sample rejection is strongly related to quality control and Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

We need a median rejection rate to compare rejection rates from This work was made at the CPMC biochemistry laboratory, from various clinical biochemistry laboratories. Variation in rejection rates January 26 to March 20, 2017. is dependent upon different criteria adopted by different laborato- − To verify the analytical performance of the automated V8 system, ries. One significant retrospective finding of this study is that the we followed the recommendations of the SFBC protocol for quality of the collected sample is crucial. If the quality of the Verification / validation of the performance of an analytical collected sample or specimen is not clinically good then it initiates a method. vicious chain reaction, as when the quality of a blood specimen is − epeatability and reproducibility were assessed using control sera poor, it cannot be processed by the laboratory, this leads to a re- of two levels of concentrations 58 and 77g /L. request for blood specimen and therefore to an increased turn- − comparison was made with another capillary electrophoresis around time (TT) for the laboratory, which is positively correlated system, CAPILLARYS 2® (Sebia) from the biochemistry laboratory with the delay in diagnosis and undue increment in the expenses of CHU MOHAMED NEDIR of Tizi-Ouzou, Algeria. both for the patient and the hospital. − 123 sera were analyzed first on Capillarys2 ®, then sent to the In the era of evidence-based medicine, clinical biochemistry CPMC Biochemistry laboratory and kept at + 4°C. The analysis on laboratory lies at the core of diagnosis and treatment. Clinical V8® (Helena) was performed the next morning. diagnosis pertaining to the services rendered by biochemistry − The total proteins of these sera were assayed by the Biuret method laboratories cannot afford to increase the turn-around time (TT) on the INDIKO® (ThermoScientific) analyzer . and compromise on the total test process (TTP). − The evaluation of the agreement between the two analysers was Further national & international comparative studies involving carried out according to the recommendations of the Cofrac by the multiple laboratories should be done to design a manifesto to detect Bland-Altman diagram and the Passing-Bablok regression line. and correct varied determinants of sample rejection. − The statistical study was carried out by XL-Stat 2014 software. doi:10.1016/j.cca.2019.03.1541 Results

The coefficients of variation found for the study of the repeat- M421 ability and reproducibility of the different fractions (albumin, 〈1, 〈2, ®, ©) were all less than 4%. Evaluation of the capillary electrophoresis of serum proteins on The comparative study showed that there is a good agreement the automated Helena V8® system between the results of the two systems, the differences found by the Bland and Altman diagram were b2.1% for all the protein fractions, N. Ould Bessib, S. Bouchtebb, S. Bouchachib, A. Chikouchea the widest difference was that obtained for the Beta fraction (2.1%). aBiochemistry Laboratory, Department of Medicine, Algiers-1 University, The equations found for the Passing Bablok regression of the Pierre and Marie Curie Center, Algiers; Algeria values of the different fractions (expressed in g /L) obtained on V8 / bBiochemistry Laboratory, Department of Pharmacy, Algiers-1 Univer- Capilarys2 are the following: Albumin: y = 1.09x - 2.79; Alpha1: y = sity, Pierre and Marie Curie Center, Algiers, Algeria 1.1x + 0.31; Alpha2: y = 1.07x + 0.69; Beta: y = 1.03 x - 1.61; Gamma: y = 1.13x - 1.55. Background-aim CONCLUSIONS Proteins are the most abundant macromolecules of human serum, they provide essential functions to the life. Their dosage is This study has demonstrated the performance of the Helena V8, a daily need in human medicine. However, the fractionation of attesting that it could advantageously be an appropriate alternative serum proteins is essential for the interpretation of any for serum protein separation methods in medical analysis measured anomaly. This fractionation is possible thanks to laboratories. electrophoresis. Electrophoresis of serum proteins is a particularly useful analysis doi:10.1016/j.cca.2019.03.1542 in many pathological situations: to guide a diagnosis, to specify the severity of a disease or to follow the effectiveness of a therapy. It also provides a lot of information on the inflammatory, nutritional, M422 infectious state and in particular allows the screening and monitoring of monoclonal gammopathies. Extending lab automation to the ED by the use of TEMPUS600 Several electrophoretic techniques are currently available. The transportation: No time to clot! Evaluation of alternative tube most efficient technique currently used, with an excellent resolving types for routine chemistry analysis power, is capillary electrophoresis, which tends to replace, more and more, the other traditional methods. The CPMC (Pierre and Marie Curie Center) Biochemistry Labora- F. Verbeke, M. Oyaert, A. Vandevelde, S. Lambrecht, V. Stove tory has acquired an automated capillary electrophoresis system: Dept. of Laboratory Medicine, Ghent University Hospital; Dept. Diag- V8® Helena. International recommendations (ISO 15189) require the nostic Sciences, Ghent University, Corneel Heymanslaan 10, Ghent, validation of any new technique before it can be used routinely. In Belgium this context, our aim is to validate capillary electrophoresis on this E-mail address: [email protected] (F. Verbeke) system.

Methods Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

Background-aim Background-aim

The new core lab automation at Ghent University Hospital Determination of IgG subclasses concentration is often used when includes an innovative cartridgeless pneumatic tube sample trans- there is suspicion of antibody deficiencies or IgG4 related disease. portation system (Tempus600™) from the emergency department Validation of the individual results is often done by comparing the (ED). Samples cover 172 m in 20-25 s and enter the (pre)-analytical sum of the 4 IgG subclasses with the result obtained from Total IgG automation directly via a bulk loader, significantly reducing pre- assays (IgGt). However, often, discrepancies arise and additional laboratory TAT. SSTII, a low-vacuum serum tube (6.0 mL) with 30 testing becomes necessary. Presently, there is no international min clotting time, was previously the preferred tube, however now reference material for IgG calibration, with each manufacturer conflicted with reduced TAT. This study examined suitable succes- adopting a different calibration standard. The objective of this work sors to SSTII, albeit SSTII quality and user friendliness need to be was to investigate the potential impact of different calibration upheld. materials in the correlation between IgG subclasses and IgGt assays.

Methods Methods

The 5.0 mL BD Rapid Serum Tube (RST) contains thrombin Thirty-six serum samples were analyzed with IgG subclasses and resulting in 5 min clotting time. An exhaustive literature review total IgG assays from The Binding Site (TBS, Birmingham, UK) in the showed RST to maintain standard-of-care. Consecutively, RST OPTILITE turbidimeter (TBS), both calibrated against the standard replaced SSTII at the ED from 13-30/11/17 (n = 818). Evaluation of ERM-DA470k, and with total IgG from Siemens Healthcare (SH, RST was based on median haemolysis index (MHI) and haemolytic Erlangen, Germany) in an AU5800 series (Beckman Coulter, Califor- samples incidence (HSI) with samples being haemolytic from a nia, US), calibrated against a different standard. Statistical analysis haemolysis index (HI) ε 88. Retrospective comparison was made done using Excel and Analyse-it. with SSTII samples 1 year prior (i.e. 13-30/11/16). Following, 2 Li-heparin plasma tubes (PSTII BD Vacutainer 4.5 mL Results and Barricor 3.5 mL) were subjected to a prospective, randomized study at the ED (n = 91). First SSTII was drawn from patients admitted to ED The correlation coefficient (r) between the sum of IgG subclasses who gave informed consent, followed by PSTII and Barricor randomized. and IgGt determination was 0.974 for TBS assays and 0.939 for SH Transportation occurred randomized, via either Aerocom™ (cartridge assay. The Bland-Altman analysis showed a mean relative difference pneumatic tube system) or Tempus600™. LD, K, Na, Cl, AST, ALT, of 0.4% (95%CI: -2.0% to 2.8%) between the sum of IgG subclasses and creatinine, glucose, urea, HI and CK were determined. IgGt from TBS and of 24.3% (95%CI: 20.9% to 27.6%) when comparing with the IgGt assay from SH. Results Conclusions RST showed an increased MHI (p b 0.001; SSTII = 7; RST = 11) and increased HSI (p b 0.0001; SSTII = 3.7%; RST = 8.0%) compared The analysis shows that, although the correlation between IgG to SSTII. Next, Bland-Altman and Passing-Bablok analysis of Li- subclasses from TBS and the IgGt assays from both providers is very heparin tubes showed ample clinically significant differences com- good (RN0.93), the agreement is clearly superior when the assays are pared to SSTII, except K as expected. However an unexpected from the same provider, suggesting that the discrepancies frequently difference for LD was observed for both and PSTII increased HSI in observed between both results are likely to be due, mainly, to the ED-specific setting. different reference materials to which different providers calibrate their assays. Conclusions doi:10.1016/j.cca.2019.03.1544 Increased HSI and MHI rendered RST unsuitable. Based on haemolysis incidence, the low vacuum Barricor is the most suitable tube. However, various ED nurses expressed the Barricor slow filling M424 as undesirable. Tempus600™ did not alter chemistry results. doi:10.1016/j.cca.2019.03.1543 Stability of B-type natriuretic peptide (BNP) stored under different conditions when measured with the lumipulse assay

V. Ramos-Arenasb, V. Campos-Rodríguezc, P. Pérez-Cañadasb,C. M423 Rodríguez-Rojasb, R. Cárdenas-Gámezb, L. Consuegra-Sáncheza,M. González-Moralesb, E. Martín-Garcíab, M.D. Albaladejo-Otónb,L. IGG subclasses and total IGG assays: The importance of using García De Guadiana-Romualdob methods calibrated against the same certiﬁed reference material aDepartment of Cardiology, Santa Lucia (Cartagena) Cardiology, Hospital General Universitario, Spain J. Perurena Prieto, E. García Guantes, M. Hernández González, R. Dieli bDepartment of Clinical Biochemistry, Hospital General Universitario, Crimi Cartagena, Spain Immunology Unit, Hospital Universitario Vall d’Hebron, Spain cDepartment of Internal Medicine, Hospital General Universitario, Cartagena, Spain Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

Background-aim Background-aim

Natriuretic peptides [B-type natriuretic peptide (BNP) and N- Hyponatremia (HP) is the biochemical finding in blood, plasma or terminal pro-BNP (NT-proBNP)] are useful biomarkers for diagnosis serum sodium concentration (Na) b135 mmol/L. It is one of the most and prognosis of heart failure (HF). Stability of BNP assays must be common disorders of body fluids but it is also one of the most examined, because there appears to be assay dependence, according underdiagnosed. Early detection and treatment are essential to avoid to IFCC recommendations. The aim of this study was to evaluate BNP fatal consequences, especially in patients (PA) with moderate and stability in plasma EDTA samples in different handling conditions, severe HP (Na b130 mmol/L) (sHP). Therefore, the Medicine assayed with a Lumipulse G BNP assay. Laboratory implemented a diagnostic algorithm (DA) applied to clinical laboratory (LAB) and POCT analyzer (POCT) results to help Methods clinicians improve the detection and management of HP. This DA included a correction of Na due to hyperglycaemia. If HP Blood samples were drawn into blood tubes containing ethylenedi- was not corrected, an automatic comment (AC) in the POCT report aminetetraacetic acid (EDTA), in two groups: group A, including 6 was incorporated recommending to send a serum sample to the LAB healthy individuals, and group B, including 5 patients admitted to the to measure osmolality (OS). If HP was evidenced in LAB, OS was Internal Medicine Department with HF. After centrifuging immediately, automatically measured. When serum OS was b275 mOsm/Kg, an AC plasma BNP levels were assayed at baseline and then at 4, 12 and 24 in the report asked for a urine sample to measure electrolytes and hours after collection, stored under room temperature and refrigerated osmolality. If OS was N275 mOsm/Kg, the AC recommended to rule (4 °C). BNP levels were measured on a LUMIPULSE G 600II analyser out the presence of osmotically active substances. (Fujirebio), based on a chemiluminiscence immunoassay. The aim of this study is to evaluate the outcome of the DA on the A change was considered significant when percent deviation (PD elapsed time between the initial detection of a sHP and the next test %) from baseline BNP level, defined as 1.65*CVA (analytical variation request with Na ε130 mmol/L. coefficient) was higher than 4.8% and 4.1% in groups A and B, respectively. Analysis of results was performed according to the Methods criteria proposed by the SEQC-ML. This is an observational and retrospective study that included all Results PA (adults and children) from the Emergency and Nephrology Departments with Na b130 mmol/L. Na was measured by Indirect Baseline BNP levels ranged from 6.1 to 26.3 pg/mL [mean (SD)]: potentiometry (P) and in POCT by direct P. An episode of HP was 12.8 pg/mL (7.7)] in group A and from 191.0 to 824.7 pg/mL [mean considered as the period of time that began when a PA showed a Na (SD): 440.6 pg/mL (289.9) ] in group B. b130 mmol/L and finished when Na ε130 mmol/L. We obtained the At room temperature, BNP levels decreased progressively in both results from the laboratory information system during six months, groups, remaining stable within 12 hours from collection only in before (665 PA) and after (710 PA) the implementation and were group A. PD for samples at room temperature were -1.1% and -4.0% homogeneous. The database was analyzed with SPSS program by the at 4 hours, -4.5% and -13.8% at 12 hours and -8.0% and -14.4% at 24 Mann-Whitney U test. hours, in groups A and B, respectively. Similarly, at 4 °C, BNP levels also declined progressively, Results remaining stable for up to 12 hours at physiological levels, but their concentrations at pathological levels only were stable within 4 hours The time for the episode of HP (average ± standard deviation) from collection. For samples at 4°C, PD were -0.012% and 1.3% at 4 was: hours, -3.5% and -5.6% at 12 hours and -7.6% and -6.5% at 24 hours, in Before: 22.3 ±22.7hours. After: 19.0 ±23.1hours. groups A and B, respectively. This difference was statistically significant (p b0.001).

Conclusions CONCLUSIONS

Plasma BNP is stable for at least 4 hours from blood collection The incorporation of this DA by laboratory in clinical practice until its measurement, as recommended by the manufacturer. contributed to decrease the time between the initial HP and the next test request with Na ε130 mmol/L. These actions are relevant to doi:10.1016/j.cca.2019.03.1545 improve the detection, classiﬁcation and management of PA with HP.

doi:10.1016/j.cca.2019.03.1546

M425

Outcomes of a laboratory diagnostic algorithm implemented in M426 patients with hyponatremia Comparative analysis of glycosilated hemoglobin results obtained B. Montero San Martín, J.J. Sánchez-Pascuala Callau, P. Oliver Saez, P. from two analyzers: Variant II Turbo and Tosoh G8 Fernández Calle, M.J. Alcaide Martín, B. Fernández Puntero, M. Duque Alcorta, N. Rodríguez Roca, C. Pizarro Sánchez, A.L. Qasem Moreno, I. R. Rubio Sánchez, E. Lepe Balsalobre, M.D.M. Viloria Peñas, A. Moro Casares Guerrero Ortiz Laboratory Medicine, La Paz University Hospital, Madrid, Spain Virgen de Valme University Hospital, Seville, Spain Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

Background-AIM aHospital Universitario Sant Joan-Alacant, Spain bUniversity of Missouri, USA The glycosylated hemoglobin (HbA1c) is the reference test for the monitoring of glycemic control of diabetic patients in the long term. Background-aim There are different methods for the quantification of HbA1c. The method of determination of the IFCC (International Federation of Hypomagnesemia is a condition necessary to be detected in Clinical Chemistry and Laboratory Medicine) is suitable as a Emergency Department (ED) patients, as it often leads to hypocal- reference method, although it is impracticable for the routine cemia that does not resolve until the magnesium deficiency has been determination of HbA1c. Therefore, the method designated by the corrected. The objective was to show and monitor an intervention to DCCT (Diabetes Control and Complications Trial) as a identify ED patients with hypomagnesemia. comparator method is ion exchange HPLC, being widely used by the Clinical Biochemistry Laboratories for the determination of Methods HbA1c routinely. The aim of this study was to evaluate the concordance of the The stat laboratory is an independent laboratory located in the HbA1c measurement obtained in the Variant II Turbo (BIO-RAD) and central University Hospital Laboratory in a health district of Alicante, Tosoh G8 (HORIBA) analyzers. Spain. The laboratory serves a population of 234,403 and processes requests for in-patients and ED patients. A Cross-sectional study was Methods performed on ED patients who had hypocalcemia between July 2016 and December 2018. In a meeting between the laboratory and ED We analyzed 110 samples of whole blood-EDTA-K3 in a range of physicians we devised a strategy that consisted that the laboratory values between 3.8% and 13.9%. The samples were processed in information system would automate register serum magnesium (s- parallel and, to minimize the preanalytical error, the analysis was magnesium) when patient presented hypocalcemia (Albumin-ad- performed in the two teams following a sequential order: first in the justed total calcium N 7,5 mg/dL). We counted the number of Variant II Turbo analyzer (reference analyzer) and later in the Tosoh detected patients with hypomagnesemia (s-magnesium b1,8 mg/dL) G8 analyzer (analyzer in process of assessment). and calculated the cost in reagent of each identified case. The statistical analysis for the comparison of methods was carried out by Passing-Bablok regression and the analysis of the differences Results using Bland-Altman with the statistical program Method Validator®. The cut-point analysis was performed using ROC Curves with the S-magnesium was automatically added to 323 samples that statistical package SPSS 19.0. presented low albumin-adjusted total calcium values, with 161 hypomagnesemia results. 119 identified patients with hypomagne- Results semia in a 2,5 year period. Each case represented a cost of 0,84 euros.

The Passing-Bablok and Bland-Altman regressions showed that Conclusions systematic differences of constant and proportional type were not found between the two methods. The equation of the line obtained The intervention to identify patients with hypomagnesemia when comparing the results between the two analyzers was: Tosoh seems cost-effective. Moreover, patients have an excellent prognosis G8 = 0.000 + 0.956 Variant II Turbo, with a conﬁdence interval of since once the deﬁciency is detected and corrected the symptoms are 95%. The results showed a high correlation between both teams (r = reversible with treatment. 0.956). doi:10.1016/j.cca.2019.03.1548 Conclusions

The results obtained in the Tosoh G8 analyzer were traceable and M428 comparable to those of the Variant II Turbo analyzer, so it is possible to ensure the transferability of the results between both teams. In Total Cost of Ownership (tco) measurement is a new evidence- addition, the Tosoh G8 analyzer improves the workﬂow and based tool to know the real cost of laboratory tests productivity of the laboratory since the processing time of each sample is less (108 seconds). L. Stankevich, P. Gontard, B. Gorodetsky doi:10.1016/j.cca.2019.03.1547 Gontard & Cie Group, Switzerland

Background-aim

M427 Year after year, manufacturers offer us new analyzers postulating that the new model is better than the previous one. The laboratory, Serum magnesium, the forgotten test: Intervention from the in turn, needs to make the right choice between the technics and laboratory to identify emergency department patients with analyzers or have objective arguments in favor a new generation. hypomagnesemia Frequently, a focus is on reducing the cost of analytics, but this should not be considered in isolation. Pre-analytical processes in M. Salinasa, C.M. Puchea, M. Lopez-Garrigosa, E. Floresa, C. Leiva- production – tube philosophy and probe sorting - also can affect Salinasb costs. Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

Methods 2-8°C. Seven analytes were measured on each sample, i.e. albumin, haptoglobin, orosomucoid, alpha-1-antitrypsin, immunoglobulin G Methodology of the Total Cost of Ownership (TCO) as a practical (IgG), immunoglobulin A (IgA) and immunoglobulin M (IgM). Prior study was carried out in the core laboratories and enabled us to to aliquotation and analysis SSTs were allowed to clot for 30 min at a make an assumption of the real cost of pre-analytics, real throughput room temperature of 18-25°C and subsequently centrifuged at 2000g of analyzers and to structure direct and hidden costs. for 10 min. To evaluate the results, laboratory calculated coefficient In this study, four independent laboratories in France ran 1000 of variation (CV) was compared to CV obtained from quantification requests of clinical chemistry (CC) and immunoassay (IC) test panels during storage for each sample’s SST and aliquots tube. of 32 analytes (among them 400 requests CC, 350 - IC, 200 - mixed (CC+IC) and 50 urgent requests) using standardized protocols on Results different analytic platforms from Abbott (ARCHITECT c8000, ARCHI- TECT i4000, ARCHITECT i1000), Beckman Coulter (AU2700, AU5800, All but one analyte remained stable during storage in both SSTs DxI 800), Roche (COBAS c701, COBAS c502, COBAS e602) working in and aliquots tubes. Four of the seven samples had a satisfactory different combinations with different pre-analytical platforms (MPL correlation for all analytes when comparing laboratory calculated track, Inpeco track, standalone AutoMate 2500). Key operational CVs with CVs obtained from quantification during storage. That is, efficiency areas that contribute to TCO were measured: sorting CVs obtained from stored tubes was lower than expected demon- speed, maintenance time (separate human and machine time), strating that plasma proteins are stable in both SSTs and aliquots processing time (total (TAT), per tube, per test) and utility tubes. In the remaining three samples a satisfactory correlation for all consumption. analytes with the exception of IgM was found. IgM presented a higher CV than expected after storage for both SSTs and aliquots Results tubes with no preference for either of the two tube types. This may be due to increase in IgM concentration during storage and Obtained results demonstrate that not only analytic platforms but corresponding trend was also observed for the other analytes but pre-analytical platforms also affect the processing speed and the did not affect CV to the same extent. same analyzers demonstrate diverse efficiency when they are built into the track compared to working with standalone sorters. Conclusions Different kinds of tube use and priority of primary tube sequence (CC, IC) also impact TAT and costs and should be determined for each This preliminary study shows that there is no difference in laboratory individually. plasma protein stability comparing SSTs and aliquots tubes for up to 10 days of storage. The result thus suggests that storage of samples in Conclusions SSTs is comparable to that in aliquots tubes. However, to validate these findings a more comprehensive study with larger sample size Methodology of the TCO enables a detailed assessment of costs to would be necessary. be carried out and a new technological solution to meet the needs of laboratories can be chosen. This makes it possible to use evidence- doi:10.1016/j.cca.2019.03.1550 based real-life management tools for calculating and controlling the costs while still ensuring high-quality test results. M430 doi:10.1016/j.cca.2019.03.1549

kEDTA sample contamination: A reappraisal

M429 S. Whitehead, K. Chadwick, C. Ford, R. Gama Royal Wolverhampton NHS Trust Will stability of plasma proteins in serum gel separator tubes be equal to that in aliquots tubes? Background-aim

N. Vickius, M. Cianchetta-Sívori Potassium ethylenediaminetetraacetic acid (kEDTA) is a widely Department of Clinical Chemistry, Blekinge Hospital, Karlskrona, used blood anticoagulant in laboratory sample tubes. EDTA contam- Blekinge, Sweden ination of clinical chemistry serum samples may result in spurious hyperkalaemia and hypocalcaemia. If unrecognised, this could Background-aim adversely affect patient care. Strategies for identifying serum kEDTA contamination are typically based upon a combination of raised For optimization of laboratory work ﬂow it is necessary to reduce the potassium and low alkaline phosphatase and calcium results,but this time-consuming manual task of aliquotation. To achieve this, the stability approach is subjective. We performed a validation of an assay for of plasma proteins when stored in serum gel separator tubes (SSTs) EDTA on an Abbott Architect c16000 analyser with the aim of needs to be evaluated and compared to aliquots tubes as currently used introducing it into routine use. in the laboratory. The aim of this preliminary study was to determine plasma protein stability in SSTs compared to aliquots tubes. Methods

Methods Intra-/Inter- batch imprecision, linearity, recovery, interference and carryover were assessed. Serum was spiked kEDTA plasma to Nephelometric quantiﬁcation of plasma proteins in seven con- mimic sample contamination and the effect on EDTA, potassium, secutive samples was performed on BN ProSpec System simulta- calcium, magnesium, ALP and zinc concentration was established. neously in SSTs and aliquots tubes after 3, 7 and 10 days of storage at Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

Results pathology, only the relationship between the presence of spermatozoa in urine and the absence of proteinuria was studied. The assay displayed acceptable imprecision, recovery and linearity. Statistical evaluation was performed using the IBM® SPSS® No significant carryover was detected but the assay was subject to Statistics V24 program. positive interference from haemolysis. EDTA contamination was detectable when serum was contaminated with 1% (v:v) kEDTA plasma. Results A change in serum potassium of 0.54 mmol/L (11.9% was observed) at a measured EDTA concentration of 0.19 mmol/L; equivalent to a Of the 50 samples studied, 43 samples were male urine (mean contamination of 3.2% (v:v). At the same level of contamination, age 52 ± 17 years) and 7 female (mean age 42 ± 9 years). reductions in measured levels of calcium (4.7%), zinc (-22.2%) and ALP No proteinuria was detected in 74% of the samples. This was (4.7%) were observed; no change was observed for magnesium. not observed in 72.1% of the samples of men nor in 85.7% of women. Conclusions Conclusions The EDTA assay displayed acceptable performance. Even small levels of kEDTA contamination were detectable. Routine measure- It can be concluded that the simple presence of semen in urine is ment of serum EDTA concentrations has the potential to identify low not a determining factor for the identification of proteins in the levels of contamination which may not be detected using more urine, with less influence in women than in men. subjective measures. Additional studies should be carried out to determine the concentration at which the presence of semen determines the doi:10.1016/j.cca.2019.03.1551 erroneous identification of proteinuria.

doi:10.1016/j.cca.2019.03.1552

M431

Presence of semen in urine, (dis)connection with proteinuria M432

M.P. Barros, G. GaiÃo Marques The impact of preanalytical automation on risk assessed with Hospital de Santa Maria - Centro Hospitalar Universitário Lisboa Norte, FMECA EPE, Lisbon, Portugal C. Bellini, F. Cinci, R. Guerranti, C. Scapellato Background-aim Clinical Pathology, University Hospital of Siena, Italy

Urinalysis (UA) is the third most commonly used laboratory Background-aim diagnostic test, with 70% of UA errors occurring in the preanalytic phase. Since up to 70% of errors in laboratory testing occur in the Diagnosis made from the UA should be based on standardized preanalytical phase, to improve patient safety it is necessary to procedures for collection, transport and analysis. carefully manage the related risk, as required by accreditation: the The presence of proteinuria may be a sign of renal damage. assessment discloses criticalities and helps in decisions. We aim to According to the European Urinalysis Standards published in 2000, provide methodological elements for an effective management, by under the auspices of the European Federation of Laboratory objectively measuring the risk connected to the phases handled by Medicine, urine samples should be collected free of internal and man compared to those automated. Our clinical laboratory belongs to external contaminants, so sexual intercourses should be avoided at the University Hospital of Siena and offers diagnostics both to least one day because the presence of semen interferes positively inpatients and to outpatients. Since the management of these two with proteinuria. Several published articles corroborate this ﬂows has different degree of automation, it allows to evaluate the principle. risk associated with different kind of preanalytics. To this aim we applied the proactive methodology FMECA (Failure Mode Effects and Objectives. Determine if the presence of semen in urine lead to Criticality Analysis). identiﬁcation of proteinuria in urine test strips (TS) used in the urinalysis. Methods

Methods Our multidisciplinary team divided the 3 phases (pre- preanalytical, preanalytical outside and within the lab) in 11 Through the observation of urinary sediments in an automated subphases (formulation of clinical question, test selection, prescrip- equipment, a total of 50 samples were selected due to the presence tion; acceptance, preparation, identification, labelling, collection, of spermatozoa. conservation, transportation; preparation for analysis) and defined At the same time, the proteinuria values of each sample, obtained 18 main activities, recognising the ones already automated from through the use of TS were evaluated. Once it could not be excluded those still human handled. For each we identified failure modes and that the presence of proteinuria could be due to the underlying effects on clinical outcome. According to literature data and our quality indicators, we assigned scores to severity, probability and Abstracts / Clinica Chimica Acta 493 (2019) S673–S710 detectability using ten-point scales. We calculated the risk indexes Results (RI) that varied from 32 to 243. During a year 1 900 680 requests and 543 828 venous and capillary Results blood samples were considered. After implementation of harmonized QIs as a part of the patient-centered evaluation of pre-analytical errors The sum of RI obtained from human activities resulted much their number decreased from 2472 (2017) to 2212 (2018) with the higher than the one produced by automation. The most critical steps highest rates for Pre-HemV (0,47%), Pre-Clot (0,27%), Pre-NotRec (RIN150) were: manual acceptance of test orders, patient identifica- (0,13%) and Pre-MisS (0,03%) with 4.1, 4.3, 4.6 and 5.0 Six Sigma values tion, tube labelling and sample collection. Our results suggest to respectively. Reevaluated RPNs for hemolized (96 to 72), clotted (72 to introduce automated phlebotomy tray preparation systems, already 54) and unreceived (16 to 12) samples decreased by 13%, but for available for outpatients, in wards. misidentification errors did not changed (81 to 81).

Conclusions Conclusions

Although automation has a fundamental role, each organization is Phlebotomy educational activities based on the ISO 15189 different about workloads and competencies, so the most suitable requirements and the National recommendations as well as imple- management must be tailor-made. Our methodology represents a mentation of harmonized QIs decreased the number of pre-analytical useful tool to predict the risk related to scenarios with more or less errors and improved the laboratory pre-analytical process, but the automation and to choose the best balance according to the needs. need for risk management based on improving the misidentiﬁcation The cyclic repetition of this analysis allows to measure the errors still exsists. effectiveness of the action adopted. doi:10.1016/j.cca.2019.03.1554 doi:10.1016/j.cca.2019.03.1553

M434 M433 Comparative evaluation of 4 different BD Vacutainer® blood Role of laboratory in improving preanalytical quality and collection tubes to study the stability of 15 steroid hormones patient safety: Impact of educational activities on sample determined by LC-MS/MS collection based on the ISO 15189 requirements and the national recommendations C. Le Goff, C. Nix, E. Cavalier Clinical Chemistry Department, University Hospital of Liege, Belgium M.M. Kardum Paro, S. Perkov, Z. Šiftar, A. Radeljak Department of Medical Biochemistry and Laboratory Medicine, Merkur Background-aim University Hospital, USA For some steroid hormones, we could observe interferences in the Background-aim peaks measured by liquid chromatography tandem mass-spectrometry (LC-MS/MS). Therefore, the inﬂuence of gel or non-gel The patient-centered evaluation of pre-analytical errors through separators should be evaluated to assess the stability of steroid risk management based on the implementation of harmonized hormones. The aim of this study was to evaluated the performance of quality indicators (QIs) is an important tool used for pre-analytical the BD Vacutainer® Serum (BD Serum) tube, the BD Vacutainer® quality improvement and patient safety. The aim of this study was to SST™ II Advance (BD SST II Advance™) tube with gel separator, the assess the impact of phlebotomy educational activities based on the BD Vacutainer® Heparin (BD LiHep) tube and the BD Vacutainer® ISO 15189 requirements and the National recommendations for Barricor™ (BD Barricor™) tube, a lithium-heparin plasma tube with venous and capillary blood sampling using the Failure Mode and mechanical non-gel separator, for the stability of 15 steroid Effects Analysis (FMEA). hormones over different time intervals.

Methods Methods

From January to December 2018 in the Department of Medical After informed consent, 10 healthy individuals were included in Biochemistry and Laboratory Medicine an retrospective observation the study. From each subject, 4 blood sample tubes were drawn: BD analysis of QIs as part of the model of quality indicators (MQIs) was Serum, BD SST II Advance™, BD LiHep and BD Barricor™. Concen- applied. MQI launched by the International Federation of Clinical tration measurements of 15 steroids were performed by LC-MS/MS Chemistry and Laboratory Medicine (IFCC) Working Group on using the Steroids MassChrom kit with QTRAP6500. Primary samples „Laboratory Errorrs and Patient Safety” (WG-LEPS) with highest risk (T0) were centrifuged and aliquots were taken and frozen for further priority numbers (RPNs) included followed QIs: Percentage of analysis. The primary tubes were stored at 2-8°C. After 4 days (T4), 7 Number of misidentiﬁed samples/Total number of samples (Pre- days (T7) and 15 days (T15), aliquots were taken from the primary MisS), Precentage of Number of samples with free haemoglobin (Hb tubes and frozen for further analysis. The data were analyzed to b0,5 g/L) detected by visual inspection/Total number of checked compare the results between the tube at T0 and to compare the samples for haemolysis (Pre-HemV), Precentage of Number of results from the same tube at different time points to evaluate samples not received/Total number of samples (Pre-NotRec). RPNs within-tube steroid stability, using an Analysis of Variance (ANOVA), before and after QIs implementation of phlebotomy educational the Acceptable Change limit (ACL), that takes the analytical activities were calculated and compared to baseline. The goal was to coefﬁcient of variation (CV) in consideration and the Total Change decrease the RPNs at least by 10% from those last year. Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

Limit (TCL), that takes both the analytical and the biological CV in hospital ordering system December 1st 2015. The observation consideration. periods were 2,5 years before and after blocking, respectively.

Results Results

At T0, statistical differences were observed between the tubes for In hospital settings, LI test requests were 10,138 and 7,331 8 steroids. However, from the ACL and TCL point of view, the before and after intervention, respectively. Repeats decreased from between-tube comparison did not show any difference at initial 416 (4.1 %) before blocking to 33 (0.5%) after blocking. The time. The within-tube stability showed statistical differences for 7 number of LI tests requested by GPs was 14,887 before and steroids after 15 days, 2 steroids after 7 days and 3 steroids after 4 18,281 after the intervention, while the number of tests repeated days. was 579 (3.9%) and 643 (3.5%) before and after intervention, respectively. Conclusions Conclusions Although some statistical differences were observed, we found no absorption in the gel neither interference from the gel itself. The Blocking of LI tests is an effective way of reducing inappropriate biggest bias was observed with the serum tube after 15 days, so it is retesting in hospitals. But the decrease in LI tests in the hospitals preferable to not store the tube longer than 7 days. cannot only be explained by the blocking. The number of repeated LI tests ordered by GPs was comparable doi:10.1016/j.cca.2019.03.1555 in the two periods evaluated. Furthermore, the number of tests requested by GPs increased during the 2,5 years. We hypothesize that the considerable increase in the number of LI tests requested by GPs indirectly caused the decreasing number of M435 requests in hospitals as, once the patient´s LI status has been tested, hospital wards are not allowed to repeat the test. Furthermore, the Blocking of lactose intolerance (MCM6-GENE) repeated tests in blocking of requests from GPs also would expect to result in an hospitals approximately 3.5% reduction of requests, which corresponded to an annual reduction of ca. 250 LI requests. S. Lillob, T.R. Larsenb, L. Pennerupb, M. Nyboc, M.V. Bord, C. Lohman e a b Brasen , M. Fruekilde , S. Antonsen doi:10.1016/j.cca.2019.03.1556 aDepartment of biochemistry and immunology, Sygehus Sønderjylland, Aabenraa, Denmark bDepartment of Biochemistry, Odense Universitetshospital og Svendborg Sygehus, Svendborg, Denmark M436 cDepartment of Clinical Biochemistry and Pharmacology, Odense Universitetshospital og Svendborg Sygehus, Odense, Denmark The pitfall of hemolyzed samples: How to manage in vitro dDepartment of Clinical Biochemistry, Hospital of South West Denmark, hemolysis? Esbjerg, Denmark eDepartment of Immunology and Biochemistry, Sygehus Lillebælt, M. Muñoz-Calero, E. Salas Herrero, C. Montilla López Kolding, Denmark Hospital San Juan de Dios del Aljarafe Bormujos, Sevilla, Spain

Background-aim Background-aim

The ability to tolerate lactose is determined by intestinal activity Recently EFLM published recommendations for management of of the lactase enzyme. There are several DNA variants of the lactase hemolyzed samples (HS) in an attempt to harmonization. There is a gene (MCM6) rending the person tolerant to lactose. We perform a strong recommendation for implementation of automated algo- test for the single-nucleotide polymorphism (SNP) C/T 13910 DNA rithm-based decision rules with the measurement of serum index. variant, called Lactose intolerance (LI) test, which is the most Nevertheless, the document did not mention the difﬁculty of common SNP in the Northern European countries. Any repetition of distinguish between in vivo hemolysis (IVH) and spurious hemolysis. this test is considered inappropriate. A differentiation has to be carried out in every HS. Despite IVH is We investigated the effect of blocking repetition of this test on infrequent, the annulation of interfered parameters could cause the number of requests from hospitals and compared with delays in the patient diagnostic. unblocked requests from general practitioners (GPs) in the Region In IVH, levels of bilirubin increase rapidly. This particularity has of Southern Denmark. been used to implement our decision algorithm in HS. Rarely IVH courses with normal levels of bilirubin (TB). Only in these cases, Methods interfered parameters are automatically annulated.

Blocking, in which repeated requests for the LI test on the same Methods patient were not allowed for “life-time”, was introduced in the Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

When the sample is not hemolyzed, the results are released. Methods When the sample is mildly hemolyzed (hemoglobin from 10 to 25 mg/dL) a comment is added warning about the possibly of interfered Peak throughput was assessed on the VITROS XT 7600 using the test. When hemoglobin exceeds 25 mg/dL, the algorithm also takes following four XT MicroSlide assays: UREA-CREA Slides; ALB-TP Slides; into account the icteric index (II). If II is normal (II=1) the interfered TRIG-CHOL Slides and TBIL-ALKP Slides versus eight corresponding parameters are suppressed and a specific comment is added. When single slide test assays: UREA Slides; CREA Slides; ALB Slides; TP Slides; the II is abnormal (IIε2) the Total Bilirubin (TB) is implemented. If TRIG Slides; CHOL Slides; ALKP Slides; and TBIL Slides. The XT the TB is normal, the interfered parameters are suppressed. But MicroSlides with dual test capability were chosen through their high when the TB is altered (TBε1 mg/dL) the interfered parameters are correlation to the sample workflow, modelled using archived run data, shown with a comment to be valued together with clinical data of and obtained through Ortho’s secure, real time communications the patient by the clinician. eConnectivity database. The Vitros XT 7600 executed a workflow test design with protocols containing MicroSlides, MicroWells, and Results MicroTips combined and with MicroSlides only.

During the year 2018, 196.095 samples were analyzed, 12.004 Results (6%) presented hemolysis. 11.336 (95%) had normal II therefore interfered parameters were suppressed. Only 593 (5%) of the HS had The peak throughput achieved, was greatest when four of the XT altered TB and the interfered results were shown. MicroSlides products were studied; resulting in maximum throughputs of 720 TPH with four XT MicroSlides products and 606 TPH with Conclusions single slide test examining the throughput of MicroSlides only. The maximum throughputs of 731 TPH with four XT MicroSlides Most of HS received in laboratory are due to spurious hemolysis. products and 617 TPH with single slide test examining the Only in a few percentage, the interfered parameters will be released throughput of MicroSlides, MicroWells, and MicroTips combined. A to be assessed together with patient’s clinic because laboratory is not series of protocols tested the maximum throughput using four, three, able to distinguish the hemolysis type. two, and one XT MicroSlides products versus single slide test Despite IVH is an infrequent disease, the cancellation of interfered products. *For presentation and demonstration purposes only. parameters could cause delays in the adequate patient management. The icteric status can be easily incorporated in the algorithm decision Conclusions rules of hemolyzed samples improving patient safety. In addition, the impact of XT MicroSlides to workﬂow is reduction doi:10.1016/j.cca.2019.03.1557 in FTE time to manage the following daily, weekly, monthly tasks: incoming consumables inspection/unpacking/storage, hands-on time loading/unloading cartridges, and generation/disposal of solid waste. VITROS XT Platforms are currently under development and are M437 not available to the public or for sale until approved under the requirements of the DIRECTIVE 98/79/EC OF THE EUROPEAN Evaluation of the new multi-test Vitros® chemistry product PARLIAMENT AND OF THE COUNCIL of 27 October 1998 on in vitro Slides* on the Vitros XT 7600 integrated system diagnostic medical devices and the respective regulatory requirements of the target market K. Culkin, D. Minihane Blood Sciences Department, Pathology Laboratory, General Hospital,, St. doi:10.1016/j.cca.2019.03.1558 Helier, Jersey, JE1 3QS, United Kingdom

Background-aim M438 The VITROS® XT 7600 Integrated System (VITROS XT 7600) from Ortho Clinical Diagnostics can perform general chemistry and Integrative analysis of parameters affected by hemolysis in immunoassays by combining its MicroSlide, MicroTip, and MicroWell different matrices under defined preanalytical conditions (with enhanced chemiluminescence) technologies into one instrument platform. We conducted an evaluation to access the impact on N. Neuwinger, D. Meyer Zum Büschenfelde, R. Tauber, K. Kappert productivity at our institution, of the VITROS XT 7600 Integrated Labor Berlin – Charite Vivantes GmbH, and Charite – System for use with the new VITROS XT Chemistry Products Slides. A Universitaetsmedizin Berlin, Clinical Chemistry and Pathobiochemistry, series of new VITROS XT Chemistry Products Slides (XT MicroSlides) Institute of Laboratory Medicine, Berlin, Germany with dual test capability have been developed. They are intended to reduce sample size and enhance operational efficiency while Background-aim maintaining analytical performance versus the conventional single slide test. The six new XT MicroSlides products are the following: In laboratory medicine, preanalytics are less standardized than UREA-CREA Slides, ALTV-AST Slides, TRIG-CHOL Slides, ALB-TP Slides, analytics, but the validity of generated data depends on both phases. GLU-Ca Slides, and TBIL-ALKP Slides. These new slide products are Variations in blood withdrawal conditions are frequently evident in unique in that they allow testing for two different analytes to be clinical settings and have been suggested to impact measurements of performed simultaneously on a single slide in the clinical chemistry erythrocyte lysis parameters, such as potassium, free hemoglobin, laboratory on the VITROS XT 7600. This study examined both the and lactate dehydrogenase (LDH). LDH activity is routinely deter- workflow improvements when adding XT MicroSlides to current mined, e.g. for monitoring malignant diseases. Here we performed an workflow and the peak throughput impact to the addition of integrative analysis of parameters affected by hemolysis in different incremental XT MicroSlides to the workflow. matrices under defined preanalytical conditions, including different Abstracts / Clinica Chimica Acta 493 (2019) S673–S710 vacuum blood collection tubes (serum, SE; heparin plasma, HP) and other drinks and alcohol ingestion, last meal before fasting and their filling volumes. exercise. Also if availability of computerized order entry (CPOE) to automatically print the recommendations in PCCs and if customized, Methods according to the requested tests. An 11th question presented some recommendations, requesting their agreement. From n=30 healthy volunteers venous blood was withdrawn. Blood tubes were either filled (90-100%) or underfilled (40-50%). Results Each tube was subjected to analyses of electrolytes (sodium, potassium, chloride), enzymes (LDH, creatine kinase (CK)) as well 68 laboratories fulfilled the survey. In 46 (68%) HDs fasting was as quality indices (hemolysis, icterus, lipemia). Further, the blood always recommended, 15 just when tests requiring fasting, and in 2 filling velocity was quantified. Parameters of hemolysis and plasma always recommended except when only CBC or CBC plus coagulation yield were determined and additionally analyzed retrospectively in tests. In 41 of the 46 laboratories always recommending fasting, was nN74.000 in-house patients. to simplify/unify criteria; In 5 because no CPOE availability. 24 (35%) always recommended a 12-hours fasting; 14, 10-hours, Results and 18 12 or 10-hours depending weather lipid tests were requested or not. 12 did not specified the duration of fasting, or less than 10 Concerning different matrices (SE vs. HP), both hemolysis index hours. (-11.7%) and potassium concentrations (-7.3%) were expectedly In 48 (70%) water intake was allowed without restrictions during reduced in anticoagulated HP. LDH activity was higher in HP the fasting period and in 9 to a maximum of ½ liter. In 2, water compared to SE (+10.3%), contradicting hemolysis index and intake was not allowed and in 9 no recommendations. In 61 (90%), potassium. Analyzing the impact of underfilling, the analytes sodium, other liquids were not recommended. In 41 was alcohol ingestion chloride, and CK remained largely unchanged. However, underfilled specifically not allowed and in 15 no recommendations. tubes were characterized by higher LDH (SE: +21.6%; HP: +29.5%), In 45 (66%) a light meal was recommended before fasting. 41 hemolysis index (SE: +216.1%; HP: +191.7%) and potassium (SE: (60%) HDs had specific guidelines regarding non- strenuous exercise +4.3%; HP: +5.5%). As one potential underlying mechanism, we in the 3 days prior to analysis, and in 16 were not mentioned. determined tube filling velocity, which was ~3-fold higher in the first In 39 (57%), CPOE offered the possibility of printing the half compared to the second half in both HP and SE. Importantly, in recommendations automatically in the PCC that were automatically volunteers plasma yield inversely correlated with LDH activity, customized, according to the requested tests. 49 (72%) laboratories which was confirmed in our retrospective patient cohort agreed with the proposed recommendations.

Conclusions Conclusions

It is relevant to distinguish between SE and HP collection tubes There was a high variability in Primary Care Patient recommen- when using LDH as a risk stratiﬁcation parameter, since HP produces dations before the laboratory tests in Spain, and some do not follow higher levels. Underﬁlling of blood tubes leads to elevations in LDH international guidelines. There is a need to reach an agreement. activities, which should be considered in clinically implausible elevated LDH levels. doi:10.1016/j.cca.2019.03.1560 doi:10.1016/j.cca.2019.03.1559 M440

M439 Discordant findings and error types observed in urine drug screens: An external quality assessment program perspective Recommendations for the patient preparation for laboratory tests in primary care in Spain: A redconlab study P. Yipb, J. Stempd, E. Dunnc, H. Vandenberghea, D. Konfortee aConsultant, Mississauga, Canada b M. Salinasa, E. Floresa, C.M. Puchea, M. Lopez-Garrigosa, C. Leiva- Department of Laboratory Medicine, Sunnybrook Hospital and Univer- Salinasb sity of Toronto, Canada c aHospital Universitario Sant Joan-Alacant, Spain Dynacare, Brampton, Canada d bUniversity of Missouri, USA Institute for Quality Management in Healthcare, Toronto, Canada eLifeLabs Medical Laboratories, Toronto, Canada Background-Aim Background-aim To know the recommendations for the patient preparation for laboratory tests in Primary Care. Drugs are screened in urine for suspected overdose or compliance with treatment programs. Single-use tests based on immunoassay Methods methods are often employed due to their ease of use, quick availability of results, and multi-panel design for common drugs. Spain is divided in Health Departments (HDs) composed of The Institute for Quality Management in Healthcare (IQMH) provides fi several Primary Care Centers (PCCs) and a laboratory attending every ISO 17043:2010 accredited Pro ciency Testing (PT) programs for HD inhabitant. A cross-sectional REDCONLAB study was designed urine drug screens. The limitations of these drug screens are fl and a call for data posted by Email. Spanish laboratories were invited re ected in discordant PT results, which were reviewed for their to fill out a ten questions questionnaire about fasting recommenda- error types and root causes. tions, duration, if customized according to the requested tests, water, Abstracts / Clinica Chimica Acta 493 (2019) S673–S710

Methods Methods

Eighteen surveys distributed between May 2010 and October UCRE were measured by Dimension Vista Jaffe Creatinine method 2018 were included. PT samples consisted of drug-free, single-donor (CRE2 Siemens Healthcare Diagnostics). Linearity performance and urine supplemented with specific drugs. A total of 54 challenges (6 comparison of UCRE measured by urine and plasma mode (reference per year) were included in the study period. Participants’ results concentration) were performed. were assessed against the reference laboratory, and discordant findings (DF) were investigated by the participant laboratory using Results a standardized report. The test was not linear below 2mM. The bias was inversely Results related to UCRE (-4.2± 3.4, -9.5± 2.3, 19.6± 4.9 and -25.9± 8.6 %, respectively for UCRE 2-4, 1.5-2, 1-1.5 and b1mM). The linear The PT program has expanded to 17 drugs or drug classes for relation between UCREb2mM and reference concentration was assessment, and enrolled 177 participant laboratories within Canada. further used to correct UCRE (corrected UCRE =0.855UCRE+0.435, The number of positive drugs per challenge ranged from 1-4, and a R2 = 0,9764). Our laboratory alerted Siemens and the ANSM total of 111 intended positive drug results were distributed during (Agence nationale de sécurité du medicament et des produits de the study period. The error types for 637 discordant findings santé) providing individual patient reports to support the clinical included analytical method (46.6%), clerical (27.9%), technical relevance. The bias of CrCl compared to mGFR was studied in 8000 (7.7%), equipment (0.9%), and the remainder were unexplained samples between 2015-2018. For UCREb2 mM, there was a linear (16.8%). The mean number of DF per positive drug result was 5.9 DF/ relation between UCRE and the bias, some CrCl values being lower Pos. The most frequently misreported results even after normalizing than mGFR. The correction of UCRE restored the physiologically for the number of positive challenges were methamphetamine (17.8 expected overestimation of mGFR by CrCl. Siemens acknowledged DF/Pos), methadone (7.9 DF/Pos), amphetamine (6.7 DF/Pos), and the negative bias at low concentrations of the analytical domain in barbiturates (5.9 DF/Pos). Among the most frequent root causes, urine (29 June 2016 Security letter) and that the limit of method sensitivity and specificity were cited in 14.8% and 13.7% of quantification (0.442 mM) could not be reached. However, the cause investigations, respectively. For clerical and interpretation errors, the was not identified and no solution was provided. most frequent were noted for methamphetamine (19.8%), methadone metabolite (EDDP, 14.2%), and amphetamine (12.2%). Conclusions

Conclusions The reported bias potentially leads to a misdiagnosis, since urine creatinine concentration is commonly used to normalize biological The errors identiﬁed within PT surveys occurred throughout the parameters for urine concentration. This would for instance impact testing process with potential impact on patient reports. Although Albumin-to-Creatinine Ratio (ACR), recommended for risk classiﬁca- immunoassay drug screen results are considered preliminary or tion of patients with chronic kidney disease, metanephrines for the presumptive, their limitations are not necessarily apparent to end diagnosis of pheochromocytoma, bone turnover markers, and many users. other biomarkers. doi:10.1016/j.cca.2019.03.1561 doi:10.1016/j.cca.2019.03.1562

M441 M442

Clinical impact of bias of urine creatinine methodology Managing post-analytical phase of procalcitonin testing in intensive care unit improves the request appropriateness A. Bouttena, A. Barniera, T. Roberta, C. Bouchet-Seraphina, N. Setaa,E. Vidal-Petiotb, M. Flamantb E. Aloisiob, A. Dolcia, M. Panteghinib aBiochemistry Department, Assistance Publique - Hôpitaux de Paris, aClinical Pathology Unit, ASST Fatebenefratelli-Sacco, Milan, Italy Hôpital Bichat, Paris, France bSpecialization School in Clinical Pathology and Clinical Biochemistry, bPhysiology Department, Département Hospitalo-Universitaire FIRE, University of Milan, & Clinical Pathology Unit, ASST Fatebenefratelli- Assistance Publique - Hôpitaux de Paris, Hôpital Bichat, Paris Diderot Sacco, Milan, Italy University, Sorbonne, INSERM U 1149, Center of Research on Inﬂam- mation, Paris, France Background-aim

Background-aim The available scientific evidence supports the use of procalcitonin (PCT) measurement in serum for antibiotic stewardship in septic In 2015, our laboratory was alerted by the Physiology department patients hospitalized in intensive care unit (ICU). PCT should be about abnormalities in creatinine clearance values (CrCl), some of tested daily until its concentrations decrease by ε80% from the peak which lower than measured GFR ( mGFR, obtained from renal value or to absolute values δ0.5 ⎧g/L, at which time both antibiotic clearance of 51Cr EDTA) which is physiologically impossible. Due to therapy and PCT monitoring should be stopped. In our hospital, PCT tubular secretion of creatinine, CrCl overestimates mGFR by approx- can be ordered without restrictions by ICU clinicians, who should use imately 20-30%, whereas in the absence of tubular reabsorption or the test according to the algorithms described above. However, we metabolism, CrCl cannot be lower than mGFR. An underestimation of noticed a tendency to continue PCT monitoring for most patients urinary creatinine concentration (UCRE) was therefore suspected, well past the ‘80% decrease from peak value’ threshold. To avoid this especially in the lower range. drawback, we decided to introduce a comment in the PCT report to Abstracts / Clinica Chimica Acta 493 (2019) S673–S710 alert clinicians when the 80% decrease was reached, with the hope of analysis completion, etc.). Sample blood clotting errors came mainly improving request appropriateness. from intensive care (20.9%), coronary unit (13.3%), and emergency department (9.2%). Median TAT for potassium on samples with blood Methods clotting errors were 72 minutes, compared to 32 minutes for non- problematic samples. Since October 2017, a comment indicating “a decrease ε80% from peak value” was added, whenever relevant, to PCT reports of ICU Methods patients as a clinical validation rule. The number of PCT requests was audited from January to September 2018 (P2) and compared with To improve the quality of the samples in our lab, and to lower the same period of 2017 (P1), before the rule implementation. turnaround times, we decided to use lithium heparin plasma samples collected in BD Barricor™ tubes. These tubes have a mechanical Results barrier rather than gel to separate plasma from cells. However, before implementing these new tubes, we had to validate an 1103 PCT determinations for 181 ICU patients (mean, 6.1 tests per alternative centrifugation time on our Beckman-Coulter centrifuges patient; range 1-60) and 991 PCT determinations for 185 patients (5 min at 1912 x g vs the recommended 3 min at 4000 x g). Since (mean, 5.4 tests per patient; range 1-38) were ordered during P1 and plasma was not validated in our Beckman assays, we also compared P2, respectively. During P2, 258 comments were added to PCT results obtained for 65 routine chemistry and immunochemistry reports of 47 patients (mean, 5.5 comments per patient; range 1-25). analyses in Barricor™ tubes in comparison to serum separator tubes This resulted in a 10% decrease in total PCT requests and an average for samples from 119 healthy volunteers. Data were analyzed using reduction of 0.7 in requests per patient. This corresponded to a weighed Deming regression analysis and comparison acceptability yearly saving of ~6,000 €. was based in mean bias (clinical acceptance limits) and correlation coefficient (r ε 0.975). Conclusions Results The post-analytical phase is the most overlooked phase of the total examination process. However, its proper management is The use of Barricor™ tubes showed clinically acceptable equiva- crucial for patient safety and clinical governance of laboratory tasks, lence to serum tubes for the studied analytes, and the alternate including appropriate test utilization. We showed that adding a centrifugation speed did not significantly affect the results. standard comment to the PCT report for highlighting the clinical information given by the biomarker behaviour in ICU patients Conclusions undergoing antibiotics can increase the appropriateness for this critical, expensive and often misused test. After implementing the new sample tubes, blood clotting errors decreased to 372/month. For example, clotting errors from intensive doi:10.1016/j.cca.2019.03.1563 care samples went from 181/month to 8/month.

doi:10.1016/j.cca.2019.03.1564

M443

Improved sample quality and decreased turnaround time by M444 implementing BD Barricor™ plasma tubes on Beckman-Coulter AU5800 and DXI Number: Standardised post analysis in Het Netherlands – It’s prime time M. Belanger, P. Hétu Centre hospitalier de l’université de Montréal (CHUM), Centre de N. Brouwerc, W. Elzen Dend, M. Thelena,b, C. Cobbaertd recherche du CHUM, Canada aAmphia Hospital, Laboratory for Clinical Chemistry and Haematology, Breda, The Netherlands Background-aim bStichting Kwaliteitsbewaking Medische Laboratoriumdiagnostiek, Nij- megen, The Netherlands Serum is the sample type mostly validated by in-vitro cDiagnost-IQ, Dijklander Hospital, Purmerend, the Netherlands diagnostic companies. However, many patients have conditions d, Department Clinical Chemistry and Laboratory Medicine, Leiden that can affect the coagulation process: thrombocytosis, lympho- University Medical Centre, Leiden, the Netherlands cytosis, anticoagulant therapy, etc. Samples also come from clinics outside the hospital and sometimes the time needed for blood Background-aim coagulation in serum tubes is not respected, leading to microfibrin problems. Uniformity of language, necessary for unambiguous information In our efforts to LEAN the total laboratory automation process, we exchange between different specialists, is a hot topic in medical found that the Beckman-Coulter AU5800 analyzers flagged 867 healthcare. This is particularly important for the exchange of sample blood clotting errors in 1 month, representing an average of laboratory results to assure proper interpretation and treatment of 28 errors/day. Resolving each sample blood clot error takes more patients, which is only possible if the total testing process is than 15 min of manual handling by the technologists (recuperating standardized. The introduction of an accuracy-based external quality the problematic sample, re-run in front loading mode, verification of control program (SKML) has already led to standardized analysis of Abstracts / Clinica Chimica Acta 493 (2019) S673–S710 clinical chemistry tests. Laboratory specialists should now take specimens, particularly long term stability. This study performs the responsibility for realizing standardized post-analysis. evaluation of the short term and long term stability of common clinical chemistry tests for biobanking specimen. Methods Methods Embedded in the Calibration 2.000 program, which initially focused on analytical test standardization, the NUMBER (Dutch Blood specimens are drawn from healthy volunteers. For short UniforM Decision Limits and Reference intervals) project was term evaluations, specimens are stored at four different tempera- initiated in the Netherlands for standardization of the post-analytical tures (20, 4, -20, and -70 degree of Celsius), and analytes are phase. The NUMBER steering group, together with 16 laboratories, measured at the time of basal, 2, 6, 24, 48, 72 hours and 1 week for used a big data approach of existing primary care data to deduce 28 kinds of analytes including calcium, phosphorus, glucose, urea reference intervals (RI). The quality of the data used in the big data nitrogen, uric acid, cholesterol, total protein, albumin, total bilirubin, approach was judged in the SKML MUSE scoring system. Five alkaline phosphatase, AST, ALT, gamma GT, creatinine, sodium, workshops were organized to reach consensus. potassium, chloride, total carbonates, iron, total iron binding capacity, triglycerides, HDL, LDL, Immunoglobulin G, A, M, Comple- Results ment 3 and 4. For long term evaluations, specimens are stored at 2 different temperatures (-20 and -70 degree of Celsius), and analytes During three workshops agreement was obtained for the RI of 18 are measured at the time of 1, 3, 6 months, 1 and 2 years intervals for standardized clinical chemistry parameters. In the fourth workshop, same analytes as short term evaluations. Percent differences from essential factors were appointed for a successful implementation of basal level for each analyte are evaluated. these RI. Setting up a website with background information and education was an important outcome. In the fifth workshop an Results action plan was established and infographics for different users of laboratory results (patient, healthcare specialist, laboratory special- For short term evaluation, percent changes of total bilirubin, total ist) were developed. carbonates, Complement 3 and 4 show greater than 10 % from basal Besides implementation of the standardized RI, embedding the levels mainly at 20 degree of Celsius. However, for long term reference calculations in the Dutch EQAS SKML is essential to evaluations, percent changes of total bilirubin, alkaline phosphatase, accomplish a sustainable infrastructure and database for future data AST, ALT, total carbonates, HDL, LDL, Complement 3 and 4 show acquisition, both for expansion with new well standardized analytes greater than 10 % starting from 6 months storage mainly at -20 and for periodic retesting of the established RI in accordance with degree of Celsius. ISO15189 requirements. Conclusions Conclusions From this study, we can predict how biobanking specimens can The establishment and implementation of standardized RI based be used in clinical researches. Furthermore, alkaline phosphatase, on a big data approach is enabled in the Netherlands, due to a type 1 AST, and ALT can be suggested as easily available biomarkers for accuracy-based external quality assessment scheme for general evaluation of the stability of biobanking specimen. chemistry and fruitful collaborations with quality-minded laboratory specialists. doi:10.1016/j.cca.2019.03.1566 doi:10.1016/j.cca.2019.03.1565

M446

M445 Recommendations on measurement units- Why and how

Evaluation of the long-term stability of clinical chemistry Y.B.L. Hansen analytes for human biobanking specimen C-NPU, IFCC

M. Hana, N.K. Leec, H.J. Leec, S.S. Parkb Background-aim aDepartment of Laboratory Medicine, Kangdong Sacred Heart Hospital, Seoul, Republic of Korea Globally, clinical laboratories are producing, communicating, and bDepartment of Laboratory Medicine, Seoul National University Hospi- exchanging millions of laboratory examination results to multiple tal, Seoul, Republic of Korea parties every day. For quantative values, measurement units are cHuman Biobank, Biomedical Research Institute, Seoul National Univer- required to make numerical values comparable and meaningful. sity Hospital, Seoul, Republic of Korea However, a unit may have several expressions, e.g. “mmol/L=μmol/ mL=nmol/μL=pmol/nL”. A non-systematic use of several expres- Background-aim sions of measurement units may create errors in the vast communication between multiple health care providers and become a risk The necessity of biobanking is growing more and more nowadays to patient safety. The Committee of Nomenclature for Properties and for clinical research. Since most of the researches with biobanking Units (C-NPU), a collaboration between International Federation of specimen are for later use, the stability of specimen is the most Clinical Chemistry and Laboratory Medicine (IFCC) and International important factor for maintaining quality of research and biobanking. Union for Pure and Applied Chemistry (IUPAC), recommends using However, there are few researches on the stability of biobanking Abstracts / Clinica Chimica Acta 493 (2019) S673–S710 an unambiguous terminology of measurement units, for daily patient Methods care and scientific publications. Venous blood samples were collected from 43 hospital patients Methods by venipuncture in RST tubes and additionally in Barricor LH (lithium-heparin) tubes. Directly after venipuncture, tubes were In corporation with international scientific organisations, C-NPU carefuly inverted and centrifuged according to manufacturer instruc- has discussed measurement units and other metrological issues in tions. Thyroid hormones, free T3, free T4, and TSH were measured by laboratory terminology for more than 50 years. chemiluminescent microparticle immunoassay by Alinity analyzer (Abbott Laboratories, Lake Forest, USA). Venous blood samples Results collected in the RST tubes were centrifuged within thirty minutes of sampling, for ten minutes at 1300 g. This was followed by the C-NPU developed, published, and recommended principles and serum separation from the cells, which was used to measure the rules on measurement units. concentration of selected analytes. Samples collected by venipunc- Examples of some of the principles for SI unit: ture in Barricor LH tubes were also centrifuged within thirty minutes of sampling, for three minutes at 4000 g.The resulting plasma was fi 1. Base SI units have unambiguous international de nitions and also used to measure the concentration of selected analytes. First, we should be used compared results of all three analytes from RST tubes with Barricor™ 2. Unit of a given magnitude should have only one expression. tubes using Passing-Bablok and Bland-Altman analyses followed by fi 3. One SI pre x per unit calculating the difference of analytes concentration beetween fi 4. SI pre x only belongs to the numerator baseline and a 24h time interval in tubes stored at +4°C.

Consequently, “millimoles per litre” shall only be expressed as Results “mmol/L”. fi Examples of de nitions of some of non-SI units: The results of the test tube comparison in the initial (baseline) time 1. WHO International Unit’ (IU) comprises a heterogeneous group of (0) showed a statistically significant difference for a pair of values units, each defined by internationally certified reference material when determining TSH. Results for TSH obtained using a RST and (CRM), e.g. a WHO International Standard. Barricor tubes were 1,250 (0,580 - 2,534) mIU/L and 1,310 (0,605 - 2. Procedure defined unit (p.d.u.) comprises a heterogeneous group 2,606) mIU/L (p = 0,002), respectively. Passing-Bablok regression of units that are not traceable to an international unit or an analysis of measured values showed that for TSH and free T3 there is a international CRM. Each unit must be defined by the local proportional difference in serum and plasma while Bland-Altman laboratory. analysis showed an acceptable concordance for all measured analytes. Serum and plasma samples were tested for the analytes stability after a Conclusions 24 hour tubes storage at 2-8 °C and although plasma and serum free T3 and free T4 levels showed statistically significant difference between C-NPU has developed and recommended an unambiguous termi- those two measurements all mean differences were within the range nology of measurement units, in order to reduce risk of errors in of clinically acceptable deviations for all tested analytes. communication of laboratory results between multiple health care parties. A considerable harmonization task for non-SI units is required. Conclusions doi:10.1016/j.cca.2019.03.1567 Using Barricor tubes improves laboratory workflow and efficiency and also provides a prompt centrifugation and a high-quality plasma sample. Additionaliy, all analyses can be performed within 24h of specimen collection. M447 doi:10.1016/j.cca.2019.03.1568 Comparison of BD Barricor™ vs. BD Rapid Serum Tube (RST) for thyroid hormones

I. Lukicb, S. Mandicb, V. Horvatb, M. Lukica, V. Sericb M448 aDepartment of Clinical Laboratory Diagnostics, Osijek University Hospital, Josipa Huttlera 4, Osijek HR-31000, Croatia Impact of managing altered results in hemolyzed samples (HS) in bDepartment of Medical Chemistry, Biochemistry and Clinical Chemistry, an infant-maternity hospital using an unconventional approach Faculty of Medicine, University of Osijek, Josipa Huttlera 4, Osijek HR- b a a a b 31000, Croatia C. Robbiano , S. Birindelli , R. Dolcini , C. Defendenti , M. Panteghini aClinical Pathology Unit, ASST Fatebenefratelli-Sacco, Milan, Italy Background-aim bClinical Pathology Unit, Specialization School in Clinical Pathology and Clinical Biochemistry, ASST Fatebenefratelli-Sacco, University of Milan, In recent years, BD Barricor™ lithium-heparin plasma tube with a Milan, Italy mechanical separator was introduced to improve laboratory workﬂow and to provide a fast and high-quality plasma sample. Background-aim We aimed to compare results and stability of thyroid hormones, free T3, free T4, and TSH in the BD Barricor™ lithium-heparin plasma How to manage altered results in HS is a matter of debate and it is tube, versus BD Rapid Serum Tube, routinely used in the Osijek even more problematic in infants. In our infant-maternity setting, we Clinical Hospital Center laboratory for thyroid hormones analysis. apply an approach for reporting interfered test results from plasma HS that provides the result itself, the degree of hemolysis expressed Abstracts / Clinica Chimica Acta 493 (2019) S673–S710 in free hemoglobin (fHb) concentration and a warning rec- Working Group Reﬂective Testing, Dutch Society for Clinical Chemistry ommending sample recollection, if possible. We investigated the and Laboratory Medicine (NVKC), the Netherlands impact of this approach on phlebotomy quality and physicians’ decision-making process. Background-aim

Methods Reflective testing is the process of interpreting, commenting and, if necessary, adding tests by the clinical chemistry laboratory, in We measured fHb on Beckman AU680 scoring results into 5 order to complete the diagnosis sensibly and effectively. In the intervals of hemolytic index (HI), where HI1 is fHb 0.5-1 g/L and HI5 Netherlands, this aspect of consultation is mainly applied to is fHb N5 g/L. We estimated the total number of HS, the most laboratory research that is requested by general practitioners. interfered analytes, the most frequently repeated tests, the clinical Previous research has shown that reflective testing is experienced wards (CW) more affected by HS and those requiring more test as useful or meaningful, in 95% or more of the cases. The aim of the repetitions. We compared data from Apr-Nov 2017 (just after present study is to make an inventory of the need for reflective implementing the approach) vs Apr-Nov 2018. testing in an outpatient or hospital setting.

Results Methods

We detected 1447 HS (10.8%) out of 13,349 blood samples in 2017 Questionnaires were developed for the medical specialists in vs 1376 HS (9.7%) out of 14,137 samples in 2018 (P=0.003), with HI2 gynaecology, paediatrics, cardiology and geriatrics. Each question- (fHb 1-2 g/L) as the most frequent category (48% vs 39%). In the two naire contained of several cases – applicable to the medical specialty periods, the most interfered analytes were conjugated bilirubin (CB) – in which reflective testing can be used meaningfully. In each case, (37.9% vs 29.2%), ammonium (22.1% vs 13.7%), LDH (16.6% vs 15.0%), respondents were asked to choose between the following options: 1) Mg (16.1% vs 11.0%) and K (12.2% vs 8.6%). We noted a significant I would like to receive only the results of the requested tests; 2) I decrease in repetition for Mg (5.0% vs 3.1%, P=0.0018) and K (3.1% vs would like to receive the results with comments; 3) I would 2.3%, P=0.0015) and a significant increase for LDH (2.4% vs 3.2%, appreciate the appliance of reflective testing; 4) Otherwise. The P=0.018). CW more affected by HS were Neonatology, Pediatric and questionnaires were distributed to medical specialists and residents Neonatal ICUs and ER, decreasing from 31% to 27%, 29% to 17%, 22% to of 6 general hospitals in the Netherlands. 19%, and 23% to 22%, respectively, in the two considered periods. Maternal ward showed hemolysis in only 0.9% of samples. CW mostly Results repeating tests after HS were both ICUs, ER and Pediatric Surgery. Interestingly, Neonatal ICU mainly required repetition of CB. The questionnaire was completed by 87 respondents in total (gynaecology n=26, paediatrics n=32, cardiology n=16, geriatrics Conclusions n=13). One third of the respondents were residents. The appliance of reflective testing was preferred in 45% of the cases; adding an Our outline of laboratory report led to a general HS decrease. We interpretive comment to laboratory research was wanted in 36% of argue that the alert comment could be a driving force able to improve the cases. Ten percent of the respondents would only like to receive the quality of phlebotomy. Beside the objective constraint of age on the the results of the requested tests; this mainly concerned specialists possibility to repeat phlebotomy, the reduction of some test repetition in geriatrics and cardiology. The option “Otherwise” (9%) was and the increase of others seem to indicate an improved physicians’ relatively more chosen by paediatricians and gynaecologists. Further skill in evaluating the clinical impact of hemolysis on different tests. analysis of this option revealed that the specialists were generally positive about reflective testing, but that they would like to have doi:10.1016/j.cca.2019.03.1569 mutual consultation about this first.

Conclusions

M449 Medical specialists and residents reported to have a clear need for reﬂective testing, in case the laboratory research is indicative for this. Reﬂective testing in laboratory medicine: Is it helpful in an The addition of tests and/or an interpretive comment can be seen as outpatient or hospital setting? part of the consulting role of specialists in laboratory medicine.

W.P. Verboeket- Van De Venne, D.L. Bakkeren, R.L. Herpers, R.M. doi:10.1016/j.cca.2019.03.1570 Hoedemakers, W.P. Oosterhuis, H. De Waard, A.P. Van Rossum