Laboratory Management, Accreditation, Quality Assurance

Clinica Chimica Acta 493 (2019) S497–S532

Contents lists available at SciVerse ScienceDirect Clinica Chimica Acta

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

Laboratory management, accreditation, quality assurance

M294 M295

International external quality evaluation on new sweat test Evaluation and comparison of HbA1c determination methods analyzer: ISEsweat II O.M. Diz Mellado, M. Calero Ruiz, M. Rico Rodríguez R. Camero UGC Laboratorios, Hospital Universitario Puerta del Mar, Cádiz, Spain Tecnicas Cientiﬁcas Para Laboratorio, Spain Background-aim Background-aim Hemoglobin A1c is formed by an uncatalyzed reaction between Introduction. The concentration of chloride in sweat remains the gold the blood glucose and some Hemoglobin A amino acids. This reaction standard for conﬁrming the diagnosis of Cystic Fibrosis. The sweat test is a is proportional to the concentration of blood glucose. HbA1c is tedious laboratory test that traditionally requires 3 correlative steps: formed in two steps by the nonenzymatic reaction of glucose with stimulation, collection and analysis. A new sweat chloride analyzer that the N-terminal amino group of the ® chain of normal adult allows the direct determination of chloride in sweat using disposable hemoglobin (HbA). Glycated hemoglobin is a parameter that cards could reduce the handling of samples and facilitate the sweat test. estimates the average of blood glucose measurements in the last 2- 3 months. Objectives. To compare the analytical performance of the new Compare two automated analytical systems to measure HbA1c in ISEsweat II sweat chloride analyzer with traditional laboratory order to evaluate if they are interchangeable. methods through participation in an international program of external quality assessment. Methods

Methods 100 samples are obtained from the hospital and from the primary care area of our center. For the HbA1c analysis are used two Methods and materials. The chloride of 24 blind samples with methods:Turbidimetric Inhibition Immunoassay (C obas 6000, c501, disposable sensor cards on the ISEsweat II sweat chloride analyzer Roche Diagnostics®) and High Performance Liquid Chromatography (RIQAS sweat testing program from Randox Laboratories, Ltd) was (HPLC) (ADAMS HA8180V A. Menarini Diagnostics® ). Everyday the evaluated monthly during 2 years. analyzers are tested with two control levels (low and high). The EDTA anticoagulant tube is used for HbA1c measurement. Results Imprecision, linearity, detection limit, functional sensitivity and correlation studies are analyzed. The results of the ISEsweat II analyzer were compared with 41 laboratories, using 5 different chloride determination methods (Coulom- Results etry, Colorimetry, Titrimetry, Direct ISE Potentiometry, Indirect ISE Potentiometry) from 15 countries belonging to 3 continents. The cycle Linearity study: y= 0,998× + 0,143 (correlation coefficient: 0,999). average absolute SDI for cycle1 was 1.02, and for cycle2 was 0.46. Detection limit: 4.7% Functional sensitivity: A blood sample is used with a known Conclusions result of HbA1c (6,5%). This level is the one used to considered if diabetes is or not well controlled. The variation coefficient at this The analytical performance of ISEsweat is comparable to that of concentration is b0.01. conventional laboratory analytical methods for analyzing chloride in Correlation study: y=1,015× + 0,121 (correlation coefficient: sweat. On RIQA’s EQA an “acceptable performance” it’s a value for cycle 0,998). average absolute SDI below 2, on our cases it was 1.02 and 0.46. The results obtained confirm the validity of the new ISEsweat II sweat Conclusions chloride analyzer as an alternative to traditional laboratory methods approved in the current guidelines for the diagnosis of cystic fibrosis. A good correlation was observed between HPLC and inmunoturbidimetric methods, therefore are interchangeable and doi:10.1016/j.cca.2019.03.1047 acceptable for the diabetes control.

0009-8981/$ – see front matter Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

The limit detection indicated that values l^ess or equal than 4.7 M297 can not be reported accurately. It is possible to conclude that the HbA1c determination, by HPLC Management of demand of AST in outpatient patients fulﬁlls the required requirements for the diabetes control from the technical point of view, and therefore, its use and implementation in D. González Benito, S. García Castañón, C. Sopeña Sánchez, V. García the routine of the clinical laboratory is possible. Moreira, F.J. Cepeda Piorno, C. Alberdi García Del Castillo, A. Llorente One limitation of the study is that patients with variants or anemia Torres, E. Fernández Rodríguez have been studied, so no conclusions can be drawn in these groups. Clinical Analysis Department, University Hospital Of Cabueñes, Gijón, Spain doi:10.1016/j.cca.2019.03.1048 Background-aim

ALT, AST and GGT determinations are useful in the diagnosis and ﬁ M296 follow-up of hepatic diseases. Its use has signi cantly increased during last years. Its essential is appropriate use, avoiding the Study of laboratory tests requests to an emergency laboratory unnecessary parameter determination, and its application will be ﬁ from a general hospital emergency department adequate to the scienti c evidence. The aim of this study is to determine the effect that would be obtained after protocol implementation of AST refusement in B. Garcia-San Vicente, D. Martinez, A. Calderon, B. Fernandez, M. patients with average values of ALT and GGT, in a sample of Gonzalez, P. Muñoz, I. Palacios, L. Pastor outpatients of our Health Area. Hospital Universitario de Alava, Vitoria-Gasteiz, Spain

Methods Background-aim

An observational retrospective study was performed for all the Introduction. Urgent attention is a main and valuable issue for received requests from Primary and Specialized Care during the last health services. As it is a complex assistance, adecuated test term of 2017 in patients over the age of 18 years and with ALT, AST and laboratory request is prioritary to ensure an appropriated and GGT results. Hemolized samples were excluded, and also patients with a affordable patient management. diagnosis of liver disease or those who attend the Digestive Department Objectives: The aim of this work is to evaluate the appropriate- were excluded. Blood tests with average values of ALT and GGT were ness of the urgent Laboratory tests requested from the Emergency selected: ALT: 4-41 U/L, GGT men: 2-30 U/L and GGT women 1-24 U/L. Department to the Emergency Laboratory of the Hospital. The concentration measurement of ALT, AST and GGT was performed in blood samples in the ADVIA 2400 Clinical Chemistry Methods System. Normal values were considered as follows: AST: 4-41 U/L for fi men and 4-35 U/L for women. Methods^. We quanti ed the number of petitions and tests requested to the Emergency Laboratory from the Emergency Results Department within last year, 2018, and calculated AST/ALT, Urea/ Creatinine and CK/Troponine ratios to assess the proper demand. From the 11445 tests with results for AST, ALT and GGT, 7036 (61.48%) had average values for ALT and GGT. From those, 63 (0.90%) Results had AST values higher than normal. Gender distribution: 29 (1.09%) from 2658 tests in men, and 34 (0.78%) from 4378 sample in women. 18 We processed 36767 petitions, 23175 venous samples, 9257 urine samples had nontypical values of AST, from those, 14 belong to patients samples, 2397 arterial samples, 1897 respiratory samples, 27 cerebro- with hepatic diseases as hepatitis or liver cancer but that had not been spinal fluid samples and 20 other biological fluids samples. The most indicated in their diagnosis, and only four had no explanation. requested tests were : ions (Na,K), 21157; Creatinine,21100; Glucose, 2108; Complete blood count (CBC)20966; Protrombine ratio, 18516; C Conclusions Reactive Protein (CRP), 18487; Urea, 17829; Activated Partial Tromboplastine ratio,13885; ALT,133102; AST,11917; Urinalysis,7987; 61.48% of the AST requests would not have been performed if the Amilase, 4575; LDH,4240; Troponine,3890; Blood gas, 3940; BT,3421; protocol had implemented. The 99.10% AST requests with normal CK, 3101; BNP,2327;Procalcitonine 1647 and Calcium, 1484. values of ALT and GGT do not yield any benefit so the test performation AST/ALT ratio was 0.91 (desirable b0.25) , urea/creatinine ratio is not justified in outpatients without previous hepatic disease. was 0.85 (desirable b0.1) and CK/Troponine 0.80(desirableb1) This protocol implementation would help to use more efficiently the AST test with a cost decrease and a work overloaded reduction. Conclusions

doi:10.1016/j.cca.2019.03.1050 Most petitions covered Ions, Creatinine, Glucose and CBC AST/ ALT and Urea/Creatinine ratios were too far from desirable values CK/^Troponine ratio was whithin desirable limits M298 We should implement actions to improve a proper demand and agree suitable protocols with the Emergency Department Veriﬁcation of common biochemical reference intervals on Alinity ci in Karachi, Pakistan doi:10.1016/j.cca.2019.03.1049 F. Kanani, M. Zubairi, U. Ata, A.H. Kazmi The Indus Hospital, Pakistan Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

Background-aim Background-aim

Alinity ci was recently launched internationally by Abbott The term six sigma means 99.9997% accuracy or 3.4 defect rates Diagnostics. We were amongst the ﬁrst to install these analyzers on per million transactions. In laboratory studies, sometimes false Accelerator a3600 in Pakistan. It is imperative for the labs to verify negative, false positive or outlier values can be seen as a result of reference ranges before adopting any assay. This is especially our errors. Six sigma is a method used to identify and then minimize important in our setting as the reference intervals are generally these errors. established on the Western population, and local studies on Alinity Six-sigma metrics has been successfully performed in clinical are absent. laboratories for nearly 20 years. In our study, we evaluated We aimed to verify the reference intervals of routine clinical the analytical performances of ﬁfteen tests performed in our chemistry and immunoassay parameters on Alinity ci in the local laboratory with six sigma metrics and aimed to achieve high quality population according to CLSI EP 28 A3. targets.

Methods Methods

After informed consent, healthy, voluntary blood donors were This study was conducted in Haydarpaşa Numune Education screened on the basis of a standard questionnaire and physical and Research Hospital Emergency Biochemistry Laboratory. Coeffi- examination during regular blood drives. Additional 5 cc blood was cient and variation (CV %) and bias % were calculated using the taken for the verification of reference ranges. Each analyte was tested internal and external quality control data from January 2016 to in 20 samples on two analyzers each. If more than two of twenty December 2016 (12 months). The Sigma metric was estimated for values were outside the suggested reference intervals, the study was 15 clinical chemistry tests [albumin, alanine aminotransferase repeated on another twenty samples. The analytes included were (ALT), aspartate aminotransferase (AST), total bilirubin, direct urea, creatinine, sodium, potassium, chloride, bicarbonate, total and bilirubin, urea nitrojen, creatinine, calcium, creatine kinase (CK), direct bilirubin, alanine transaminase, aspartate transaminase, alka- sodium, potassium, chlorine, glucose, lactate dehydrogenase (LDH), line phosphatase, gamma glutamyl transferase, uric acid, magne- Magnesium ] on Abbott ARCHITECT ci8200 autoanalyzer. Sigma sium, phosphorous, calcium, amylase, CK, LDH, iron, total protein, metrics were calculated using bias %, CV % and total allowable albumin, C Reactive Protein, FreeT3, Free T4, TSH, prostate surface error (TEa) ratios of CLIA’88 and Ricos. Sigma metric equation is antigen, ferritin, Vitamin B12, Vitamin D, folate, AFP, beta hCG, FSH, given below: LH and prolactin. Sigma Metric ¼ ðÞTEa%−Bias% =CV% Two Sigma metrics were calculated for each parameter using 2 Results different levels of controls (Low and high, normal and high QC etc.) close to clinical decision threshold. The mean (SD) age of the study participants was 37 (8.4) years. Mean weight was 70.2 (14.4) kg and haemoglobin concentration was Results 14.9 (1.27) gm/dl. Manufacturer defined reference ranges were verified for 23 Sigma values were variable based on the CLIA and Ricos TEa clinical chemistry and 11 immunoassay parameters. Vitamin B 12 ^ targets. Magnessium had world performance with CLIA TEa limits, ranges fell below the criteria when determined on initial 40 samples. but it could not showed acceptable performans in Ricos TEa limits. The study was repeated on another 20 samples. Out of a total of 60 On the other hand, according to both CLIA and Ricos targets, CK had healthy blood donors, 22 had vitamin B12 levels less than 187-883 the highest sigma values at both control levels (⌠1 CLIA: 14.9, ⌠ 2 ng/ml. This was understandable as vitamin B12 deficiency is very CLIA: 9.3); (⌠1Ricos: 15.1, ⌠2Ricos: 9.4) and sodium had the lowest common in South East Asia, even among apparently healthy subjects sigma values (⌠1 CLIA: 1.93, ⌠ 2 CLIA: 1.54); (⌠1Ricos: −0.07, ⌠2Ricos: as evidenced by literature. As expected, vitamin D levels (6.3 to 39.1 −0.06). For direct bilirubin there was no specification from CLIA so ng/ml) were also at sub optimal levels in our population. we calculated sigma value of direct bilirubin from Ricos data and it had world-class performance. The other tests had sigma performance Conclusions ranged between 2 and 6.

Manufacturer defined reference ranges were verified on Alinity ci. Conclusions doi:10.1016/j.cca.2019.03.1051 The six sigma methodology is an important quality indicator in the evaluation of analytical performance. It can be a self-assessment tool regarding the performance of clinical laboratory. We’ve seen M299 some tests that we have insufficient performance therefore, we first reviewed our reproducibility and accuracy for these tests. At the Analytical performance evaluation using six sigma methodology same time, we think that the use of six sigma is very useful for the for clinical chemistry tests clinicians in the differential diagnosis and follow-up of the treatment. We also think that the standardization of total allowable error A. Kütükcüb, F. Özçelika limits will be more understandable and feasible for laboratory aDepartment of Medical Biochemistry, Health Science University, Sultan professionals. Abdülhamid Han Education and Research Hospital, İstanbul, Turkey bDepartment of Medical Biochemistry, Health Science University, doi:10.1016/j.cca.2019.03.1052 Haydarpaşa Numune Education and Research Hospital, İstanbul, Turkey Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

M300 M301

The increasing quality of the services provided by Romanian Veriﬁcation of result interchangeability in the clinical laboratory medical laboratories demonstrated by the participation in external quality control J.A. Delgado RodrÍguez, M.B. Badal Cogul, A. Rubio Alaejos, A.R. Pons Mas, C. GÓmez Cobo, J. Robles BauzÁ C. Popa, G. Sorescu, C.G. Dinulescu Department of Laboratory Medicine, Hospital Universitari Son Espases, CALILAB, Romania Palma de Mallorca, Spain

Background-aim Background-aim

Introduction. External quality control is for medical laboratories According to ISO 15189:2012, clinical laboratories need to that use it as an important tool in improving processes by providing compare procedures, analyzers and methods and assess the inter- them with opportunities for improvement (OFIs), according to the changeability and equivalence of test results from identic procedures World Health Organization. or analyzers. This verification process makes it possible to establish The aim of this paper is to show the increase in quality of the preventive corrective actions towards an improvement in patient services provided by medical laboratories in Romania by objectively safety. analyzing the coefficients of variation of the results provided by the Our objective was to apply the verification protocol for test result participants to the external quality control organized by a national interchangeability of the accredited biochemical tests in our provided (CALILAB). laboratory, in accordance with document EP-31-A-IR (CLSI). The second aim of this project was to compare the results of this Methods interchangeability verification process with the previous one, carried out 3 years ago using the same technology. The coefficient of variation used to assess the performance of medical laboratories has been calculated and analyzed for a variable Methods number of 340 laboratories according to the external quality control schemes taken in the study: Serum Biochemistry and Hematology. Serum, urine, K3EDTA-whole blood and lithium heparin-whole The value of the variation coefficients of the results reported by blood patient samples and certificated control materials were used the medical laboratories participating in the external quality control for the study of 52 accredited biochemistry tests in our laboratory. organized by CALILAB in Romania for the analyzes / parameters of Approximate analyte concentration was estimated on the basis of the studied schemes was compared with the value of the coefficients clinical decision levels and known imprecision; the latter obtained of variation presented in the literature by other international from the intraserial standard deviation (RS, n = 20) and the total providers of external quality control. standard deviation (TS, n = number of controls assayed in a ε6- month period). Results Magnitudes were defined as interchangeable when the difference between the average results of each autoanalyzer (DbM)bLimit of The comparison results highlighted: acceptability (LoA), which was calculated by following the document EP-31-A-IR (CLSI). - ^the increasing in quality of the services provided by medical laboratories in Romania by the decrease the values of the Results coefficients of variation for the parameters analyzed in the external quality control organized by CALILAB in the period Interchangeability of all analyzers was verified, except for TSH 2008-2018; and GGT. As a corrective action, were removed the conflictive tests - the values of the coefficients of variance for the parameters ^ from the uninterchangeable autoanalyzers, although in response to analyzed in the external quality control organized by CALILAB in their demand and loading by equipment, TSH was implanted in a 2018 are lower than the values of the coefficients of variation of newly acquired autoanalyzer. the same parameters established by the consensus of the The results highlight the importance of this kind of studies, given providers of external quality control at international level. that in the previous one, these tests were interchangeable. This shows the variability over time of the results, as a consequence of Conclusions equipment wear, changes in the reference material, new laboratory staff… The quality of the services provided by medical laboratories in Romania has been highlighted by the decrease of values of the Conclusions coefficients of variation of the results by the laboratories participating in the external quality control organized by CALILAB, coefficients Clinical laboratories must periodically verify the interchangeabil- of variation lower than the values of the established by the ity of the test results and establish the pertinent corrective actions, as consensus of the providers of external quality control at international well as to evaluate the effectiveness of the adopted measures, in level. order to improve result quality and to ensure patient safety. doi:10.1016/j.cca.2019.03.1053 doi:10.1016/j.cca.2019.03.1054 Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

M302 do the test based on the Laboratory Developed Test (LDT), and also do the QC based on the Standard Operation Procedure (SOP) that is Diagnosis of seasonal influenza: Workload in an emergency managed by each FCM Lab. The conventional FCM software doesn’t laboratory trough the last two influenza seasons have the QC management function for LDT, so original QC chart is created and the QC results are managed in the each Labs. It is B. Garcia-San Vicente, D. Martinez, A.B. Bravo, B. Fernandez, E. Gallo, complicated for the FCM testing workflow. Therefore we developed M. Gonzalez, A.I. Martinez, L. Pastor the “Panel Setup/QC” function, it includes the management and the Hospital Universitario de Alava, Vitoria-Gasteiz, Spain report function of QC for LDT, and decrease complicated steps and improve the FCM testing Workflow. Background-aim Methods Introduction. The overall health impact of Influenza varies widely from season to season, it has an effect on the Emergency Laboratory 1. Adjust the Gain/Photomultiplier voltage to the optimized value workload. for analysis by using the QC material based on SOP in Lab. 2. Create the compensation table. Aim. The aim is to test diagnosis of influenza impact on 3. Save the measurement conditions as the panel information. Emergency Laboratory workload, trough the last two influenza 4. Create the QC file for LDT and set the baseline and the upper/lower seasons. limits for each channel. 5. Measure the QC material in daily. Methods. It is an observational retrospective study. An estimation 6. Report the QC chart. of diagnosis of Influenza workload for the Emergency Laboratory was done for the 2016-2017 ( November2016-April 2017) and for the Results 2017-2018 (November 2017-April 2018) influenza seasons. There was obtained the number of tests for detection of Influenza virus About QC, Adjustment of Gain/PMTV and Compensation, Running performed , from Laboratory Information System (Omega-Roche sample, we evaluated the difference in the case of between the Diagnostics) and also was used the Catálogo de Pruebas de manual operations and using the “Panel Setup/QC” function. The BioquímicaClínica y Biología Molecular. Consellería de Sanitat. usage materials that we assume the TBNK 6 color of LDT panel are Generalitat Valenciana to calculate the relative unit cost (CRU) followings. produced each period. - Sample: CD Chex (Streck) - reagents: Kombitest 6 color cocktail (Exbio) Results. The resultant Emergency Laboratory workload due to - Beads for compensation: CompTrol beads with CD8 antibody with diagnosis of Influenza was 17741 CRU, in the first season 2016-2017, the different flurochromes and 27866 CRU, in the season 2017-2018 (+57 % higher)

Conclusions The results are followings. I. The steps for FCM testing Workflow are decreased in 56.5% by Influenza diagnosis implies a considerable workload for the using the “Panel Setup/QC” function. Emergency Laboratory. (there are 23 steps for the manual operations, but only 10 steps by The workload has varied widely trough the two seasons. using it.) Compared with the previous season, 2017-2018 Influenza season, one of the most severe in this century, has supposed a serious Conclusions increase in the Emergency Laboratory workload. The Catálogo de Pruebas de BioquímicaClínica y Biología Molec- We can realize the followings for LDT by developing the “Panel ular. Consellería de Sanitat. Generalitat Valenciana, agreed by Setup/QC” function. professionals, enables workload evolution study. - Management of QC results by using Levey-Jennings chart doi:10.1016/j.cca.2019.03.1055 - Creating and output reports

As a result, we decrease the complicated steps for FCM testing Workﬂow. M303 doi:10.1016/j.cca.2019.03.1056 ”Panel Setup/QC” function for laboratory developed test—Im- provement of the Flow Cytometry testing workﬂow

M. Kinishib, T. Tsujia, Y. Tsuruokab, S. Yonedab, H. Tatsutanib M304 aCell Technology Engineering1, Sysmex Corporation bSoftware Technology Engineering 2, Sysmex Corporation Relevant quality indicators from pre-analytical phase to continuously improve the accredited medical laboratory performance in Background-aim an emergency clinical hospital

a b b So far, the IVD applications like TBNK / CD34 are provided by D. Popa , C.D. Neculoiu , S.N. Moga a Flow Cytometry (FCM) manufactures. Then, they also provide the Emergency Clinical County Hospital, Brasov, Romania b Quality Control (QC) material and the software functions for QC Emergency Clinical County Hospital Brasov, Faculty of Medicine, management and Report. However, almost FCM laboratories (Labs) Transylvania University Brasov, Romania Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

Background-aim improper transport. A risk assessment analysis of the pre-analytical phase was performed in the Clinical Laboratory of a six-ﬁfty bedded Considering that on average 70% of clinical laboratory errors tertiary care hospital in India. Aims of the study included improve- occur in the pre-analitycal phase of total testing process, the aim of ment of in-time sampling, reduction in sample rejection and this study is to analyze the speciﬁc quality indicators-QIs and to increased patient satisfaction. identify the relevants ones to inhance patient safety by continous improvement of our clinical laboratory activity. Methods

Methods FMEA (Failure Mode effect Analysis) was used as a tool for Risk Assessment of pre-analytical processes. The probable areas of risk in Prospective study, carried out for 18 months in a clinical laboratory the pre analytical phase were first identified and a RPI (Risk Priority for in-patients, by analyzing collected data on e-requests and types of Index= Severity × Detectability × Probability) was assigned to each biological samples received for clinical chemistry and hematology source of error. In areas where the RPI was above 50 (out of a total of compartments. The calculated values obtained for the 12 selected Qis 125), appropriate actions were proposed to reduce the estimated were expressed as %, defects per million -DPM and on six sigma scale. risk. The effectiveness of the actions taken shall be reviewed every six months. Results Results During the follow-up time we had received 29454 request forms and 36746 biological samples from the clinical departments of the More than 90% of the samples requiring re-dos were from the in- stationary. The data analysis of selected Qis values showed the patients department. The primary causes of rejection included highest % rate for the number of inpatients requests with erroneous clotted and haemolysed samples, quantity not sufficient, wrong data entry(test name)-Pre InpMT and the number of misidentified sample, wrong container and doubtful integrity of samples. They requests-PreMisR(0.89 and 0.94%),respectively the lower one for the were attributed to incorrect sample collection practices (wrong unidentified sample-PreUnIns (0.065%).The 3.9 Sigma score value container, wrong barcode), incorrect phlebotomy practices for serum associated to the the critical errors corresponding to the Pre InpMT samples (collection with needle and syringe, opening the cap of the and to PreMisR showed an immediate need to training for all vacutainer, pushing the piston of the syringe, inadequate filling authorized staff as mandatory corrective action in the pre- resulting in altered plasma-anticoagulant ratio). This reflects inade- preanalytical phase.We have notice a good acceptability –4.4 sigma quate training of the nursing personnel involved in sample collection value- for the hemolysed primary samples evaluated by visual in the IPD. inspection. Reporting errors Sigma score 4.4 associated to the biological samples for hematology compartment was over the Sigma Conclusions value 4.2 obtained for the clinical biochemistry specimens. Our lab proved a good performance showed by a Sigma score between 4.1 to Extensive training sessions on “Best Practices in Phlebotomy” 4.4 for 8 monitored Qis, but the accuracy improvement of entering were organized for nursing staff every month. This was to ensure data process in e-request form it’ s a must. training of all new staff and re-training of the existing personnel. In addition, regular trainings were also imparted on appropriate sample Conclusions collection practices.

Study results were used as entry data for management analysis to ensure risk mitigation especially in the extraanalytical phase by doi:10.1016/j.cca.2019.03.1058 improving communication and training of medical staff in order to increase lab performance doi:10.1016/j.cca.2019.03.1057 M306

Outsourcing laboratory tests: Making a breakthrough control real

M305 D. Weinstein, B. Simkovitz, I. Heler, O. Cohen, G. Rashid Meir Medical Center, Clalit Health Services, Kfar Saba, Israel Pre-analytical errors in the clinical laboratory: A risk assessment analysis Background-aim

S. Senguptaa, A. Handoob Meir Medical Center clinical laboratories catalog contains over aClinical Chemistry & Immunoassays, Laboratory Services, BLK Super 450 tests, in a wide range of fields. However, there are tests that Speciality Hospital, Delhi, India require a dedicated instrument, a special method of performance or bLaboratory Services, BLK Super Speciality Hospital, Delhi, India performed only at national centers. For this purpose, we have established a process of tests delivery for outsourcing. The process Background-aim involves several factors within the hospital and outside it, including hospital departments, lab reception office, transportation unit and Pre analytical variables account for 45-60% of the total errors in a the laboratories operating at the outsourcing sites. Accordingly, to clinical laboratory. They are encountered principally in the following ensure its success, process control is required to provide efficient and areas: improper test request, misidentification of the patient, continuous service. Our aim was to developed a process map in order labeling errors, sample collected in unsuitable vacutainer/additive, to locate control points and to establish a tight control on the whole inadequate volume, compromised sample integrity and delayed/ process. Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

Methods biochemistry and haematology were integrated and microbiology separated. The 60.9% of stat laboratories were integrated in part with Several stages were defined, including communication improve- laboratory core. Requests were done electronically by 75.6%. The ments with all outsourcing labs, a wide electronic lab catalog majority of laboratories (48.8%) had been 100-300 request/day with updating and full documentation of all incoming outsourcing lab 1000-3000 tests/day. The 46.3% of laboratories were using Interna- result reports. We also preformed a trial period to test improvements tional and conventional units in clinical reports. All laboratories feasibility. participated in internal and external quality assessment programs. The vast majority of laboratories reported critical results (97.6%). Results These values are established by consensus with clinicians (59%). Still telephone was the main communication system (69%). Clinicians or fi As of June 2016, the control has entered into a defined monthly nurse were the responsible for receiving the critical value noti ca- framework. The monthly control data for all 2017 was summarized tion (80%). The majority of laboratories validation was by laboratory and analyzed. In terms of overall output, a 7% increase in number of medical specialist and laboratory technician (31.7%) or only by samples sent for outsourcing compared to 2016, 2509 samples in technician (24.4%). With regard to human resources, the vast 2017 versus 2341 samples in 2016. In measuring the control process majority of stat laboratories have been one medical specialist performance, 99.1% results were obtained in 2017, 2435 results out (biochemistry or analysis) (75%) and in some laboratories also there of 2456 of samples shipped for outsourcing. This is in comparison to were a microbiology or haematology specialist. In 75.6% there were a 42% at the beginning of process map designing. In addition, by these specialist on call 24 hours/day by physically present or telephoni- controlled processes we performed a full measurement of samples cally. Laboratory staff was consisted by technicians (100%) and turnaround time, from sampling transport date until the result nurses (29.3%). received from all outsourcing sites. Conclusions Conclusions This survey was helpful in order to know the state of the art in The outsourcing process is a central part of the laboratory services stat tests laboratories in Catalonia. for the medical staff, by enables comprehensive laboratory service and continuity of patient’s treatment. Results shows we have doi:10.1016/j.cca.2019.03.1060 succeeded by tight control process to significantly improve the tractability of outsources lab results. In this way we have effectively expanded the control limits of our clinical laboratory division, which strives to continuously service improvements. M308 doi:10.1016/j.cca.2019.03.1059 Sigma metrics for assessing the analytical quality of the new multi-test VITROS® XT chemistry products slides11*For presentation and demonstration purposes only. VITROS XT Slides are currently under development and are not available to the public M307 or for sale until approved under the requirements of the DIRECTIVE 98/79/EC OF THE EUROPEAN PARLIAMENT AND OF Survey on stat tests in Catalonia clinical laboratories THE COUNCIL of 27 October 1998 on in vitro diagnostic medical devices and the respective regulatory requirements of the target A. Arbiol-Roca, D. Dot-Bach market. Laboratori Clínic Territorial Metropolitana Sud, L’Hospitalet de Llobregat, Barcelona, Spain M. Barbero, T. Dimagno, C. Graby, T. Huynh Ortho Clinical Diagnostics Background-aim Background-Aim In order to know the state of the art and new trends on Stat Tests, the Catalan Association of Clinical Laboratory Sciences (ACCLC) did a The new VITROS® XT Chemistry Products Slides (UREA-CREA, survey on the vast majority of Clinical Laboratories in Catalonia. ALTV-AST*, TRIG-CHOL*, ALB-TP*, GLU-Ca*, and TBIL-ALKP*) have been developed with dual test capability for use on the VITROS XT Methods 7600 Integrated System. The analytical performance of these new XT assays has been evaluated for analytic quality using the sigma During 2017, the ACCLC distributed an online survey to 69 metrics methodology. hospital laboratories. A web-survey with 12 point questionnaire was sent to investigate: laboratory model, degree of computerisation, Methods human resources, quality control, validation of results and critical results, among others. The collected information was analyzed. For precision determinations, the total within-lab precision was calculated with quality control and patient sample using two replicates Results per day, twice per day over 20 days (total n = 80) following CLSI EP5 guidelines. For bias determinations, two different methods were used. Survey participation was (41/69) 59%. There were 8 tertiary For the first method over one hundred patient samples spanning the hospitals (19.5%) (ε400 patients/day) and 21 secondary hospitals measuring range were analyzed in singleton on a XT 7600 System (51.2%) (ε170 patients/day). There were 28 teaching hospitals against a reference method. The data was analyzed following CLSI EP9 (68.3%) and 26 hospitals with Intensive Unit Care (63.4%). The guidelines using a Passing-Bablok regression, and the percent bias was 53.7% were public hospitals. The 68.3% of stat laboratories, calculated from the regression line. The second method used the Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

College of American Pathologists (CAP) 2018 Chemistry survey fluids of Catalonian Health Institute (Spain) consensus after ten years of to calculate the percent bias. Two different allowable total error (TEa) experience. values (Ricos and CLIA) were used to calculate the sigma metrics using the coefficient of variation (CV) and bias as determined above: sigma Methods metrics = (TEa − %Bias)/CV. Yearly average was recorded for each IQ and laboratory, and the Results average of the yearly interlaboratory median was calculated to assess intergroup evolution. Using the regression equation from the patient samples to For samples, IQ formulas were modified, and the activity of the determine Six Sigma quality, six of the XT assays (UREA, TRIG*, most requested test in each type of specimen was chosen as GLU*, TBIL*, ALTV* and AST*) had greater than six sigma perfor- denominator: creatinine, complete blood cell, prothrombin time or mance versus the Ricos TEa. Another three XT assays (CREA, CHOL* diuresis for serum, EDTA, plasma-citrate-coagulation (Plasma-CC), and ALKP*) had greater than five sigma performance. The remaining and 24h urine sample, respectively. For serum and first morning three XT assays (Ca*, ALB* and TP*) had less than three sigma urine sample, IQs were calculated respect to the total number of performance due to the very small Ricos TEa (2.55%, 4.07%, and 3.63% requests to evaluate its evolution respect to the previous QS. respectively). When evaluating the assay quality versus the CLIA TEa Hemolyzed serum sample QI was calculated differently with the requirement, all assays except ALB* had greater than six sigma use of automated hemolysis index detection in all laboratories. performance and ALB* had greater than five sigma performance. The results were substantially equivalent when the sigma metric were Results determined using the CAP survey fluids for the bias estimate instead of the regression equation. QS for total errors in requests, requests with data missing and total errors in samples improved (0.96% vs 1.3%), (0.82% vs 1.31%) Conclusions and (3% vs 5%), respectively. QS for total errors in serum with respect to the total requests worsened (0.6% vs δ0.50%). QS for total errors in The data presented here demonstrate that the new XT Micro- first morning urine improved (1.22% vs. 1.25%). Incorrect patient Slides on the XT Systems show excellent analytical performance for data also improved (0.03%-0.01%), but could not reach desirable 0% precision and accuracy when judged using the Sigma Metrics QS as sentinel QI. methodology against the CLIA TEa requirement. New QS for total errors in different types of samples based on the activity were: Serum (1.16%), EDTA (0.67%), Plasma-CC (2.29%) and doi:10.1016/j.cca.2019.03.1061 24h urine (8.81%), respectively. Hemolyzed serum QI for primary care (2.07%-1.01%) and hospitals (3.91%-1.71%) centers improved in the last year but the QS has not yet been established due to the variability between laboratories. M309 Conclusions Continuous monitoring and readjustment are required in quality indicators and their specifications due to organizational changes The results of the QI will be used as new QS. Organizational changes such as the fusion of hospital and primary laboratories may M.A. Llopis Diazf, C. Biosca Adzetf, N. Serrat Orúsb, J. Minchinela have an impact on the results of the QI (increase in the number of Gironaf, R. Ruiz Morere, M.I. Llovet Lombarteg, G. Busquets Soriah,M. samples, changes in the sample containers, analyzers, etc.). A more Montesinos Costah, A. Blanco Fontc, M. Simón Palmadaa, C. Perich reliable denominator in different types of samples avoid that QI of Alsinai, M. Ibarz Escuerd less frequent samples can be underestimated. The monitoring of QI aConsorci del Laboratori Intercomarcal de l’Alt Penedès, Vilafranca del allows take actions to avoid occurrence of errors in order to increase Penedes, Barcelona, Spain patient safety. bLaboratori Clínic (ICS) Camp de Tarragona, Hospital Universitari Joan XIII, Tarragona, Spain doi:10.1016/j.cca.2019.03.1062 cLaboratori Clínic Hospital, ICS-Metropolitana Sud, Universitari de Bellvitge, Bellvitge, Spain dLaboratori Clínic ICS Lleida, Hospital Universitari Arnau de Vilanova, LLeida, Spain M310 eLaboratori Clínic L’Hospitalet, ICS-Metropolitana Sud, L’Hospitalet de LLobregat, Barcelona, Spain Communication of critical values in the clinical laboratory fLaboratori Clínic Metropolitana Nord, (ICS) Hospital Universitari Germans Trias i Pujol, Badalona, Barcelona, Spain I. Olmos Sanchez, L. Criado Gomez, S. Villanueva Curto, B. Perez g Laboratori Clínic Terres del Ebre (ICS), Tortosa, Tarragona, Spain Sebastian, N. Seco Moro h Laboratori Clínic Territorial de Girona (ICS), Spain HOSPITAL iLaboratoris Clínics Hospital Vall d’Hebron (ICS), Barcelona, Spain Background-aim Background-aim Introduction. A critical value is the one which results of a The aim was to evaluate the evolution of preanalytical quality diagnostic test which expresses a medical situation that endangers indicators (QI) during 2014-2017 in order to evaluate quality the patient´s life if there isn’t a proper intervention, hence the specifications (QS) previously established by the clinical laboratories importance of informing the clinician about it immediately Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

Aim. To review and evaluate the reported critical values by means phase were sample rejection rate (SRR) which included number of of the existing protocol at the laboratory. clotted specimen, haemolysed sample, insufficient sample, wrong tube/label and diluted sample. For analytical phase the indicators Methods were concordant proficiency testing (CPT), equipment uptime and repeat testing. Turnaround time (TAT) compliance and critical alert The critical values are analyzed retrospectively in a period of five callout were included for the post-analytical phase. The trend was months. observed for all QI for the two years. Data was collected into an excel file for analysis and % performance compared over both the years. Results SRR and TAT were evaluated with the sigma scale.

We obtained a total of 467 critical values, from which a Results percentage of 44% were notified (No. 207). The rest of them weren’t notified because they had similar previous values. Regarding the A total of 1,414,244 (6, 90,751 in 2015 and 7, 23,493 in 2016) action taken by the clinician, 23% (no=47) repeated the analytics, specimens were received for testing. In 2015, 6564 (0.95%) 20% (no=40) adjusted the medication, 12% (no=24) contacted the specimens were rejected giving a sigma level for SRR of 3.90. Sample patient and 9 %( no=19) sent the patient to the emergency service. dilution was the most common pre-analytical error with a sigma 30% (no=63) no actions were taken by the clinician or they are value of 4.2. Whereas in 2016, 4716 (0.65%) samples were rejected unknown because they are not shown in the clinical history. giving a sigma level of 4.00 with haemolysed sample being the common error giving a sigma value of 4.4. There was an improve- Conclusions ment in proficiency testing from 97.7 % in 2015 to 99 % in 2016. Equipment uptime also showed a significant improvement from 91.1 In view of our results, we consider that our protocol is useful % in 2015 to 96.8% in 2016. No. of samples for repeat testing however because, thanks to it and to the rapport between the laboratory and increased from 0.30 % in 2015 to 0.40 % in 2016. A total of 80,554 the clinical services, we have been able to resolve situations that (0.95%) samples were not meeting TAT in 2015 with a sigma level of endanger the patients’ lives. 2.70. Whereas 42,158 samples (0.65%) were reported to not meet the We are surprised by the fact that in sixty-three patients no action TAT in 2016 giving a sigma level of 3.10. For the critical values was taken or it was not shown in their clinical histories. We objectify communicated to inpatients within one hour an improvement from that approximately half of them were admitted patients with glucose 97.8 % in 2015 to 98.9 % in 2016 was observed. and/or potassium alterations, on whom the action was probably limited to clinical observation. Conclusions From the laboratory we plan to update our protocol by conducting consensus surveys with the clinicians, with the aim of The results of our study indicate a general improvement for continuing to improve the quality of care. most of the QIs. Most significant improvement was found in the pre- and post-analytical phase due to sensitisation of staff which doi:10.1016/j.cca.2019.03.1063 included training events and issue of informative documents to decrease error of request data input. The use of quality indicators to assess and monitor the quality system of the laboratory is an important tool to assure the improving process and guarantee the M311 patient safety.

Evaluation of quality indicators by six sigma principle: Experience doi:10.1016/j.cca.2019.03.1064 at a tertiary care hospital

S. Sharma, K. Taneja Chacha Nehru Bal Chikitsalya, India M312

Background-aim Impact of total allowable error according to different recommendations on the risk management index of clinical chemistry tests Quality in laboratory has a big impact on diagnosis and patient management as 80-90 % of all diagnosis is made on the basis of S. Szakony laboratory tests. The identification of reliable quality indicators is a Department of Laboratory Medicine, St Imre Teaching Hospital, crucial step enabling users to quantify the quality of laboratory Budapest, Hungary services. The ISO15189 mandates the implementation of quality indicators. In this study we evaluated the performance of the Background-aim laboratory through some set of quality indicators. Sigma value depicts the performance of laboratory and its quality measures. Contemporary quality control (QC) design focuses on minimizing Hence in the present study six sigma principle was applied to the the risk of patient harm. Following new expectations we reviewed quality indicators to evaluate the clinical biochemistry laboratory the QC strategy for our 23 clinical chemistry tests. performance. Methods Methods We have been using independent controls (Bio-Rad) for years. QC The present study was conducted in the Department of Biochem- results are evaluated using the Unity software (Bio-Rad). The istry at Medanta-The Medicity Hospital from January 2015 till Westgard Advisor of Unity was used to select the QC rules for each December 2016. The quality indicators (QI) for the pre-analytical test. Since there is no national total allowable error (TEa) Abstracts / Clinica Chimica Acta 493 (2019) S497–S532 recommendation in Hungary, therefore the biological variability (BV) Results data was used for TEa. BV minimum bias and imprecision was chosen as default settings. Since September 2018, we have been able to test The findings of the laboratory system assessment show that the Bio-Rad’s new Mission:Control (MC) software version 2, which strongest areas of the country’s laboratory system at the policy and makes recommendations for the frequency of QC and calculates the regulatory level are “Coordination and management” and “Labora- patient’s risk assessment. tory information system”. The laboratory-related coordination at the Ministry of Health is well established and functional, and the Results national laboratory data collection and analysis activity is centralized and is implemented by the Center of Health Development of MOH. In the case of tests with Sigma N4 (16 tests), the risk management Gaps are found in “Infrastructure”, “Regulations” and “Human index (RMI) was b1, which means a controlled risk. The initial settings resources”. The poor result of “Infrastructure” is due to financial did not need to be changed in terms of QC rule and QC frequency shortcomings. The main problems detected in the area of “Human according to MC. But for tests with Sigma b4 (7 tests), RMI was N1. For resources” include insufficient financial and organizational support instance in the case of total protein the default setting resulted in RMI for the continuous education of laboratory workers, a shortage of 4.03 (TEa: 5.45%, Sigma: 2.48, 1 daily QC, repeat 1:2s rule, 50 samples trained personnel and incomplete national registration system of per day, severity of harm: minor, probability of harm: 50%). As a results laboratory professionals. of changes suggested by MC (3 QC per day, 1:2s rule), RMI decreased to 3.8. Leaving all other settings unchanged and increasing TEa to 6% Conclusions (Rilibak recommendation) Sigma was 2.82 and RMI was 1.59. Based on MC suggestion (3 QC per day, 1:2s rule), RMI was reduced to 1.48. 1. A national regulatory body needs to be established for the Increasing TEa to 8% (SEKK recommendation), Sigma and RMI became registration of all laboratories and laboratory professionals. 4.06 and 0.09 respectively. Following the CLIA recommendation (TEa: 2. A formal continuous education system for laboratory professionals 10%), Sigma and RMI changed to 5.3 and b0.001 respectively. In the should be set up. latter two cases, there was no need for further action. doi:10.1016/j.cca.2019.03.1066 Conclusions

There are 2 to 11-fold differences between the loosest and the M314 tightest TEa values in the various recommendations. In 6 out of 7 tests, the default TEa (BV minimum bias and imprecision) was the Are Mongolian laboratories ready for accreditation? lowest of the optional values. Consequently the well-chosen TEa had a greater impact on RMI than QC frequency or QC rules. T. Enkhjargala, M. Koguchib, D. Khishigbuyana a doi:10.1016/j.cca.2019.03.1065 Mongolian Association of Health Laboratorians, Mongolia bSysmex Corporation, Japan

Background-aim M313 Accurate test results enable health professionals to make the right Assessment of laboratory system in Mongolia diagnostic and therapeutic decisions. In order to demonstrate the quality and reliability of their services, medical laboratories seek T. Enkhjargal, J. Dulamjav accreditation to ISO 15189. There are more than 300 medical Mongolian Association of Health Laboratorians, Mongolia laboratories in Mongolia, and only seven are accredited so far. We have initiated a project to assist medical laboratories in their efforts Background-aim to obtain the accreditation. As a first step of the project activity, we carried out a gap analysis of laboratories. Medical laboratory is integral to many clinical decisions on prevention, diagnosis, treatment and management of patient disease. Methods Therefore, the quality of laboratory services is paramount. One of important factors that influence a laboratory’s overall activity is Six laboratories representing private and public, urban and rural support of policy makers. An assessment of the national laboratory medical laboratories are selected for participation in the project. The system in Mongolia will help to identify its strengths and gaps. gap analysis of the participant laboratories is carried out using an Excel program that incorporates all ISO 15189 requirements. Methods Results The questionnaire-based assessment of strategic organization and support at the national level (e.g. defining policies and regulatory The findings reveal that the participant laboratories are strongest framework) was performed with participation of members of the in the following categories: Organization and management, Quality Laboratory professional council of the Ministry of Health of Mongolia of examination results, Personnel and facility management and (MOH), MOH officers in charge of diagnosis and reference services, Laboratory information management. The majority of the laborato- and officers in charge of monitoring and assessment of health ries are hospital based, and their organization and management are servicesь and their answers were substantiated with related well established and functional mostly due to centralized adminis- documents. The assessment results were generated and analyzed trative guidance. The concept of quality control is effectively adapted using an Excel questionnaire program, and scores below the 75% in medical laboratories, and data management is usually in line with limit were considered insufficient. the requirements. Weaker areas include Evaluation and audits, and Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

Document control. Even though the laboratories do conduct M316 evaluations and control, they do not do it regularly and, most importantly, do not keep records routinely, which cause the higher ”Safety and hygiene” risk management in the quality approach— gap rate. Biochemistry Laboratory Experience of CHU Ibn Rochd- Casablanca Conclusions S. Aatfaoui, F. Boulhen, N. Kamal Policies to meet ISO 15189 requirements are in place in the Laboratory of Biochemistry of CHU Ibn Rochd of Casablanca, Morocco participant laboratories, but their documentation and records keeping are insufﬁcient. Background-aim doi:10.1016/j.cca.2019.03.1067 Safety and hygiene are among the priorities of any medical analysis laboratory, they are an integral part of any quality process, including the ISO 15189 qualiﬁcation, it is an obligation in the framework of the Moroccan regulation. The quality of performance M315 of a medical analysis laboratory is based on the control of its organization and its processes. the neglect of security may cost a lot Design and implementation of quality control plans that inte- to the laboratory : its reputation, the health of the staff and the grate moving average and internal quality control: Incorporating patients, the quality of the results, the biomedical material invested, the best of both worlds as well as the hygiene of the environment.

H. Van Rossum, D. Van Den Broek Methods The Netherlands Cancer Institute, The Netherlands Initially, we began by structuring the risk management approach Background-aim using the 5M method, from which we extracted the main risk factors for adverse effects. We identiﬁed them and calculated the criticality New moving average quality control (MA QC) optimization of each by FMEA method, then we prioritized them by a farmer methods have been developed and are available for laboratories. diagram and we ﬁnalized our work with a plan of corrective and Having these methods will require a strategy to integrate MA QC and preventive actions based on a collective brainstorming and experi- routine internal QC. ence other biomedical laboratories in managing these risks.

Methods Results

MA QC was considered only when the performance of internal The main risks identified were biological risk, chemical risk, QC was limited. A flowchart was applied to determine, per test, physical risk, fire risk and environmental risk. Prevention measures whether MA QC should be considered. Next, MA QC was examined for some of these risks were almost non-existent, while for others using MA Generator (www.huvaros.com), and optimized MA QC the risk was unavoidable. Our action plan was based on this procedures and corresponding MA validation charts were obtained. prospective identification and we started by highlighting the critical When a relevant systematic error was detectable within an average risks first, by initiating a prevention procedure based on the single daily run, the MA QC was added to the QC plan. For further document and the moroccan regulations in the field of security. implementation of MA QC for continuous QC, MA QC management software was configured based on earlier proposed requirements. Conclusions Also, protocols for MA QC alarm work-up were designed to allow detection of temporary assay failure based on previously described The laboratory remains a sector with a wide variety of experiences. occupational exposures and the assessment of occupational risks in the face of these hazards is a regulatory obligation that everyone Results must invest in.

Based on the ﬂowchart, 10 chemistry, 2 immunochemistry and 6 doi:10.1016/j.cca.2019.03.1069 hematological tests were considered for MA QC. After obtaining optimal MA QC settings and the corresponding MA validation charts, the MA QC of albumin, bicarbonate, calcium, chloride, creatinine, M317 glucose, magnesium, potassium, sodium, total protein, hematocrit, hemoglobin, MCH, MCHC, MCV and platelets were added to the QC Evaluation of the effect of more frequent reporting in analytical plans. performance through the use of an international external quality assessment scheme Conclusions

S. Doherty, G. Kansanaho, M. Rodríguez, L. Adams, S. Fitzgerald The presented method allows the design and implementation of Randox Laboratories Ltd., Crumlin, United Kingdom QC plans integrating MA QC for continuous QC when internal QC has limited performance. Background-aim doi:10.1016/j.cca.2019.03.1068 To ensure reliable reporting of patient test results, the periodical assessment of the analytical performance of the laboratories through Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

External Quality Assessment (EQA) schemes is relevant. An EQA turnaround time (BC-TAT) can improve adequacy of antibiotic scheme allows comparison of results of participant laboratories by treatment and prevent unnecessary and superfluous empiric their methodology, instrument and reagent, which facilitates the treatment identification of potential failures and inter-laboratory agreement to To achieve BC-TAT reduction, improve patient outcomes and ensure the test system accuracy needed to meet regulatory reduce unnecessary antibiotic consumption. requirements. Frequent reporting allows early identification of system errors and implementation of any necessary corrective Methods actions with minimum disruption to the laboratory. In this study the EQA scheme, Randox International Quality Assessment Scheme The study was conducted at Rambam Health Care Campus, Israel. (RIQAS), was used to evaluate how the frequency in the reporting The microbiology laboratory (micro-lab) operates between 07:00- affects performance of clinical chemistry parameters 20:00. On July 2016, an intervention was implemented to allow BC processing 24/7, through collaboration between the microbiological Methods and STAT laboratories, computer and infectious disease divisions and the hospital administration. Outside the micro-lab working hours, BC General Clinical Chemistry Programme samples (52 parameters) are transferred to the STAT laboratory and are incubated in a satellite were assessed by participants either every 2 weeks or every month incubator (Bactec FX-40, BD, USA), connected to the microbiological depending on their chosen frequency option. The performance laboratory information system (LIS). At the beginning of each indicators, %Deviation from the mean for comparison (%Dev) and working day, the bottles are transferred to the main incubator the Target Deviation for Performance Assessment (TDPA) were (Bactec FX, BD, USA) at the micro-lab for continued processing. A studied. TDPA is a target %deviation which represents the lowest hospital-wide education campaign was launched to encourage BC average 10% of the most poorly performing participants, for any delivery to the relevant lab 24/7 immediately after BC drawing. given parameter, averaged over 1 year. The differences between TDPA on bi-weekly and monthly frequencies were compared. The Results parameter lipase was studied in detail to establish reasons for the differences. The time from BC drawing to its arrival to the laboratory during nighttime decreased by 82% from 6.8 hours to 1.2 hours. During the Results 2-year period of the intervention, 642/15540 cultures, representing 4.1% of total cultures handling in the STAT laboratory, were signaled For the majority of parameters (50 out of 52), the TDPAs for 2017 as positive in the STAT laboratory, allowing immediate processing at for samples assessed bi-weekly were smaller than for samples the micro-lab in the morning. BC-TAT, calculated as the time from assessed monthly. For lipase, the biweekly TDPA was 10.2%, while arrival to the micro-lab or STAT lab until results were reported to the the monthly TDPA was 21.3%. The average % deviation of lipase for physician, were improved by 12% for positive and 4.5% for negative samples assessed biweekly was 1.56% (n = 12463) while the % BCs. deviation for samples assessed monthly was 4.94% (n = 17791). Other factors such as age of the samples, length of time of Conclusions participation in the programme, location of participants, method and instrument did not have significant impact in the reported trend Our results demonstrate that collaboration between different clinical laboratories can improve processes. Our next objective is to Conclusions examine the effects of the intervention on clinical outcomes and antibiotic consumption. The results from this report indicate a trend in the improvement of the analytical performance by increasing the frequency of doi:10.1016/j.cca.2019.03.1071 reporting in the EQA RIQAS. doi:10.1016/j.cca.2019.03.1070 M319

System performance evaluation of STA satellite max, new bench M318 top analyzer for the routine hemostasis lab

Blood culture incubation 24 hour a day: Interdisciplinary J.R. Palomo Martínez, M.á. Somoza López, R. Fernández Palomares, F. collaboration in a large tertiary Israeli hospital García Gómez Laboratorio Análisis Clínicos, Clínica CEMTRO, Madrid, Spain J. Attiasb, M. Paula, M. Kaplanb, Y. Geffenb aInfections Diseases Department, Rambam Health Care Campus, Haifa, Background-aim Israel bLaboratory Department, Rambam Health Care Campus, Haifah, Israel The STA Satellite Max is a new bench top analyzer designed to perform in vitro tests for the diagnosis and monitoring of disorders Background-aim related to hemostasis. It is meant for routine labs with small hemostasis activity or can be used as back-up instrument in larger Sepsis is, the most severe manifestation of acute infection, can organizations. It performs clotting, colorimetric and turbidimetric cause multi-organ failure and ends in death in 30–50% of cases. Blood assays. cultures (BC) are considered the most sensitive method for detecting Our laboratory was the ﬁrst European center to evaluate this new bacteremia and are commonly obtained in patients with fever, instrument and assess its ergonomics along with reliability and some chills, leukocytosis, focal infections and sepsis. Short blood culture analytical performances, under a routine-like setting of tests. Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

Methods excluding the requests without diagnose, and we stratiﬁed them by FC levels: b50 μg/g, 50-200 μg/g and N200 μg/g. Over 3 weeks, our lab had to test at least 30 patients’ samples on STA Satellite Max each day, along with intra-assay reproducibility Results and cross contamination evaluations, using citrated plasma samples from patients. We registered 10003 FC requests from January of 2013 until Reproducibility was assessed over 10 replicate measurements on September of 2018 (2013: 29, 2014: 483, 2015: 1425, 2016: 2341, prothrombin time (PT), activated partial thrombin time (APTT), 2017: 3179 and from January until September of 2018: 2549). ﬁbrinogen (Fib), antithrombin (AT) and d-dimer (DDi) levels, using The sample of patients that were suspected or suffer from IBD STA-NeoPTimal 10, STA-Cephascreen 4, STA-Liquid Fib, STA- was formed by 3434 patients: 27% (911) FC b 50 μg/g, 25% (871) FC Stachrom ATIII 3 and STA-Liatest DDi respectively. 50-200 μg/g and 48% (1652) FC N 200 μg/g. On the other hand, 2872 Cross contamination was evaluated by comparison of mean patients were not suspected to have IBD: 50% (1427) FC b 50 μg/g, results of 10 APTT tests measured in a raw versus 10 APTT tests 28% (800) FC 50-200 μg/g and 22% (645) FC N 200 μg/g. measured in combination with PT tests. Conclusions Results FC request has increased exponentially because it is not an No major issue appeared when using the analyzer, during the invasive method, its cheap to perform and it has a high sensitivity. routine testing of patients’ samples. Moreover, a 48% of the patients that were suspected or suffer from Intra-assay CVs obtained were of 1.49%, 0.25%, 2.57%, 1.96% and IBD have and active organic disease and a 25% are in the grey zone. 2.79% for PT (sec), APTT (sec), Fib (sec), AT (%) and DDi (μg/mL) Otherwise, just a 22% of patients that were not suspected to have IBD respectively. have an active organic disease and 28% are in the grey zone. In order to Relative deviation of mean APTT during cross contamination that we should consider applying a demand management to FC request. assessment was of 0.83%. All those results easily meet acceptance criteria. doi:10.1016/j.cca.2019.03.1073

Conclusions

M321 STA Satellite Max shows good reproducibility and cross contamination results. Easy-going for work and reliable on the tests evaluated, HS troponin I performance characteristics on Alinity I in Karachi, the instrument can be used in the hemostasis clinical laboratory. Pakistan doi:10.1016/j.cca.2019.03.1072 F. Kanani, A.H. Kazmi The Indus Hospital, Pakistan

M320 Background-aim

Faecal Calprotectin as an evidence-based medicine biomarker of We recently shifted from conventional troponin I (c trop I) assay inflammatory bowel disease to high sensitive troponin I (hs trop I) assay on Alinity i by Abbott Diagnostics. Alinity i system has recently been launched internation- M.A. Caro Miró, S. Gómez Vera, L. Herranz Arriero, J.J. Morales Lomas ally, and we are among the first ones in Pakistan to adopt the assay Complejo Hospitalario Universitario de Badajoz, Spain on this analyzer. Our objective was to to verify the performance characteristics of hs trop I on two Alinity i analyzers. Background-aim Methods Inflammatory bowel diseases (IBDs), such as Crohn’s disease (CD) and Ulcerative colitis (UC), are chronic diseases that result from the Precision: This was verified by running commercial controls at inflammation of the lower gastrointestinal tract. These pathologies three different levels in five replicates for five days. are usually related to high risk of surgery and an increased risk of Accuracy: External proficiency sample were run and results colorectal cancer. compared to Architect group mean. Calprotectin is a calcium and zinc binding protein implicated in Sensitivity: Limit of Blank, Limit of Detection and Limit of the regulation of the inflammatory process that represents about 60% Quantification were determined and compared to vendor claims. of soluble proteins of granulocytes cytoplasm. Method comparison: 40 samples were analyzed for c trop I on Faecal Calprotectin (FC) is an evidence-based medicine biomarker Vitros ECi and hs trop I on both Alinity I analyzers across entire to diagnose IBD with a high sensitivity that can reduce unnecessary analytical measuring range (AMR). invasive colonoscopies due to its ability to differentiate organic from Verification of 99th percentile URL: functional IBD. It is recommended to apply a grey zone between 50 and Hs trop I was performed on 40 healthy male voluntary blood donors 200 μg/g and levels N200 μg/g are indicative of active organic IBD. (20 on the each Alinity i). HbA1c and estimated creatinine clearance of the donors were also determined to further define normality. Methods Results We analysed the number of FC requests in our Hospital during the last 5 years. Then we selected the patients that were suspected or Precision study showed a precision of 5.6/5.7, 1.4/2.7 and 1.1/4.1 suffered from IBD and the ones that were not suspected to have IBD, for Alinity i 1 and 2 at low, medium and high levels. The samples Abstracts / Clinica Chimica Acta 493 (2019) S497–S532 were within total allowable error in the CAP proficiency testing Results survey when compared to Architect series, though with a negative bias. Limit of blank was 0.93 and 0.27 ng/L and limit of detection was In all 27 clinical chemistry, and 13 immunoassay analytes were 1.7 and 0.7 ng/L on Alinity I 1 and 2 respectively. Limit of tested. All clinical chemistry and immunoassay parameters met quantification was verified to be less than 5.1 ng/L as claimed (3.6 manufacturer’s claims for precision, accuracy, linearity and report- and 4.4 ng/L).The assay demonstrated a CV of less than 6% at 99th able range on both Alinity c and i systems. In all 87.5 % of assays were percentile URL on both Alinity i. Passing Bablock between c trop I on at greater than 5 and 6 sigma levels. CO2, Chloride and Free T4 were VItros ECi and hs trop i on the two Alinity i instruments showed a between 4 and 5 sigma, while sodium and Free T3 were between 3 slope of 0.988 (r = 0.99)and 0.921 (r = 0.96), (95% CI included 1.0) and 4 sigma. with a negative intercept of 3.98 and 3.35, and no significant difference in mean concentrations. Hs trop I levels were below the Conclusions gender specific cut offs in all healthy donors, being undetectable in 35 out of 40 tested so far. It needs to be seen if 99th percentile URL Abbott’s claims for precision and accuracy were verified. In all levels are lower in our ethnic group. 87.5 % of the assays evaluated were at more than 5 and 6 sigma. Limitations of the study: Conclusions • Total allowable error margins have been set too wide by accrediting The performance characteristics of hs troponin I assay are verified bodies • on Alinity i. Extended analysis is required to give an accurate picture of the instrumental and assay performance. • doi:10.1016/j.cca.2019.03.1074 Accuracy or true value should be judged against a reference method.

doi:10.1016/j.cca.2019.03.1075 M322

Method performance veriﬁcation of Alinity c and i systems in M323 Karachi, Pakistan

Continuous improvement of laboratory quality systems through F. Kanani, M. Zubairi, A.H. Kazmi upper management support: Lodwar county referral hospital The Indus Hospital, Pakistan (LCRH), Kenya experience Background-aim J. Maragia Marcomic Kenya -Ministry of Health and Sanitation, Turkana County, Kenya Alinity series was recently launched internationally with great fanfare. We are among the first in Pakistan to install Accelerator Background-aim a3600 and Alinity ci series as part of Total Laboratory Automation. Initial studies published on Alinity ci have been carried out in Abbott laboratories. Independent data is needed to verify manufacturer’s Incognizant of Resolution AFR/RC58/R2 (2008) and Maputo claims. We aimed to verify performance specifications (precision, Declaration that emphasizes the strengthening of laboratory accuracy, linearity and reportable range) of common quantitative system, the top management in LCRH resolved to implement and clinical chemistry and immunoassay parameters on two Alinity c and bolster this noble course by enrolling LCRH laboratory in Strength- i systems as per CLSI guidelines. Method verification studies are the ening Laboratory Toward Accreditation (SLMTA) process. Since fi exceptions rather than the rule in our part of the world. 2015, the top-level management through the of ce of the director zealously committed itself to successfully compete at the interna- Methods tional level of performance. It knew without management commitment, laboratory involvement and practice, the effort would be stymied and abortive. This study is aimed at demystifying the The method performance was verified on two Alinity i analyzers. massive support the management has accorded the laboratory for • Simple and Complex Precision: These were determined by continual and sustainable improvement. running commercial controls at three different levels in five replicates for five days. Mean, standard deviations and coefficient of Methods variation were determined, and compared with the company claims. • Accuracy: 3-7 proficiency testing samples of various levels were Before the SLMTA process commenced in LCRH, the management tested in 3 replicates and results compared with Architect group was brought on board and sensitized about the entire process by the mean. experts with an aim of fostering commitment and buy-in. This was • Linearity/ Reportable Range: 7 levels spanning analytical measuring fundamental in enlightening the management on how the process range of parameters not having multi point calibrators were run in will provide a controlled and efficient high level of technical 3 replicates. Mean values at each level were compared to target competence and quality service to its customers. All the departmen- values to see if they were within total allowable error. Recovery and tal heads that were directly or indirectly linked to the laboratory linearity were calculated. were involved which included but not limited to Nursing head, • Sigma metrics: Sigma metrics for the assays were calculated from clinician, laboratory staff, and maintenance staff—towards the accuracy and precision (TAE-bias/CV). mission of sustainable quality practices. Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

Results = 120; SD 63 mins) and 164 minutes (n = 11; SD 90 mins), respectively. During three periods of prolonged CSF TAT during the night, The Laboratory moved from zero to five stars in a period of three there were 17, 52 and 61 cases of ABID in the BTS versus 1, 2 and 3 years. Currently earmarked for accreditation. There was an overhaul cases of CSF at the same time. This correlated with a CSF TAT of 262, renovation, extension, and reorganization of laboratory floor plan for 300 and 246 minutes, respectively. optimal workflow. Equipment was put on the service contract and a reduced equipment downtime from 30% to 2% due to controlled Conclusions temperature and periodic preventive maintenance. The management engaged in resource mobilization and advocacy which increased staff Our analysis showed a relationship between BTS workload, long level from 8 to 15, participation in external quality assurance CSF TAT and manpower. During periods where a single technologist schemes, a budget for commodities, training and mentorship. A was on duty and the BTS workload was high, the TAT for CSF was strong Lab-Clinical Interphase as evidenced by joint CMEs, improved prolonged. Our root cause analysis and value stream maps showed inter-professional communication, improved quality of diagnosis, that workload numbers provide superficial information but a deep notification of critical lab results, joint MDTs, timely results dive into test complexity showed involvement to prolonged TAT. submission, joint ward rounds, Mentorship of HCWs on proper These tools can also be used to imply a future state that can meet the documentation in lab request forms intended service level by process improvements and optimisation that will result in a faster diagnosis and improved patient outcome. Conclusions doi:10.1016/j.cca.2019.03.1077 Management commitment and laboratory staff teamwork is an impetus in continual and sustainable laboratory quality management system. M325 doi:10.1016/j.cca.2019.03.1076 Assessment of the performance of two quality control materials, in-kit and a new third party control, for anti-Müllerian hormone in cobas e411 analyzer M324 M. Simón Velasco, N. Rodríguez Roca, M.J. González Villalba, P. Using quality improvement tools in the investigation of pro- Fernández-Calle, A.L. Qasem Moreno, M. Gómez López, V. Escribano longed turn around time and improving the efficiency in the core Hernández laboratory Laboratory Medicine, La Paz University Hospital, Spain

E. Lee, A. Omar, M. Luceri, M. Hey, W.W. Chey, M.S. Wong Background-aim Department of Laboratory Medicine, Khoo Teck Puat Hospital, Singapore The anti-Müllerian hormone (AMH) is an established biomarker Background-aim for assessing ovarian reserve; consequently, it requires accurate AMH measurements. The strategy of internal quality control to assure Khoo Teck Puat Hospital is a 626 bed Acute Care hospital that analytical performance can be assessed by in-kit (IK) controls, employs multi-skilled technologists to run the core laboratory during the supplied by the manufacturer or by third party (TP) controls, night shift (2100h to 0900h the next day). During this shift, one medical unafﬁliated to the reagents and calibrators used in the measurement. technologist manages both the Blood Transfusion Services (BTS) and Therefore, TP controls have distinct advantages; they can detect a Cerebrospinal Fluid (CSF) analysis. Upon monthly review of CSF Turn shift due to degradation of the calibration materials and avoid Around Times (TAT), there was a consistent and distinct prolonging of releasing erroneous patient results. TAT during this shift. The two hour Key Performance Indicator for CSF The aim of this study is to compare the ability of an IK to a TP TAT was prolonged to as much as 300 minutes for 12.4% of the total daily control in order to detect possible changes in instrument workload compared to 63.6% of the night workload. performance.

Methods Methods

Data on the total number of samples received and the TAT for the PreciControl AMH Plus (IK) (Roche) is a lyophilized equine serum tests performed by technologists in a 24-hour cycle was collected from matrix in two concentrations ranges. The new Liquichek AMH the Laboratory Information System for a period of one month. We Control (TP) (BioRad) is a liquid serum of human origin, in three used Lean methodology and Quality Improvement tools to identify the concentrations ranges. Both controls were stored in aliquots of 150 various contributors and potential root causes. A relationship map was μLat− 20°C. Materials were processed over a period of 2 months, plotted over 24 hours incorporating Blood Transfusion and CSF two days a week by duplicate (n = 22), using one reagent lot and workloads, their respective TATs and manpower allocation. two different calibrations, on the cobas e411 analyzer (electrochemiluminescence ECLIA). Statistical analysis was per- Results formed using Microsoft Excel.

There were 129 requests for CSF and 1019 requests for Blood Results Typing and Antibody screen during this period. Of the 1019 antibody screens performed, 981 subsequently required Antibody identiﬁcation The assigned values of the two levels of control materials for IK (ABID). The average daily TAT for CSF was 130 minutes (n = 129, SD and TP were 0.76 ng/mL vs 0.95 ng/mL and 4.64 ng/mL vs 4.56 ng/mL 76 mins). The CSF TATs for day and night shifts were 111 minutes (n respectively. Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

The imprecision (coefficients of variation) obtained were 7,58% and period were resolved, providing reassurance in the service provided. 3,28% for the IK, while TP controls showed lower results 2,35% and 3,06%. Sebia had confirmed the loan instrument would be the analyser The IK control presented a bias of −5.12% at low level and 2.84% permanently used, meaning installation and validation requirements at high level. A bias of −2.65% and −3.86% were obtained in both had been met. levels of TP control respectively. Procurement proved to be problematic and time consuming, delaying the implementation of CE by several months. Conclusions The final roadblock was that of IT interfacing and validation, which required a good deal of planning and organisation to ensure “go-live” Both control materials presented adequate assigned values was successful. This resulted in a further delay in switching over to concentrations, which monitor assay performance at different reporting via CE, however the transition occurred in May 2018. medical decision levels. TP appears to be more stable than the IK, probably due to its liquid stable presentation that minimize pipetting Conclusions error. Both materials showed a negative bias in the concentration near to the clinical decision level used for assessing the ovarian The transition between technologies was smooth, with no reserve, being less noticed in the TP control, maybe due to its human significant issues encountered. origin. The new Liquichek AMH Control (TP) of BioRad can be an adequate alternative to monitor the measurement of AMH. doi:10.1016/j.cca.2019.03.1079 doi:10.1016/j.cca.2019.03.1078

M327

M326 Moving average quality control—Practical experience

Making the switch from high performance liquid chromatogra- M. Vershininaa, N. Steriopolob, V. Ibragimovab phy (HPLC) to capillary electrophoresis (CE)—The Edinburgh aFSAI “National Medical Research Center for Children’s Health” of the experience Ministry of Healthcare of the Russian Federation, Moscow, Russia bFSBI “Central Clinical Hospital with Polyclinic” Department for P. Ryan Presidential Affairs (of the Russian Federation), Moscow, Russia NHS Lothian, Livingston, United Kingdom Background-aim Background-aim The use of traditional quality control in hematology testing with The haemoglobinopathy laboratory at the Royal Infirmary of use of the commercial control blood has several drawbacks: Edinburgh provides a screening service for the South East of expenses, problems with appropriate storage, stored material may Scotland, testing approximately 4000 patient samples annually. slowly deteriorate even if maintained in an appropriate manner, the Historically the service has been provided using high performance short shelf life, the inter-vial variability. liquid chromatography (HPLC), however the department has re- In addition, the control blood is the material that consist of cently changed to capillary electrophoresis (CE). stabilized red blood cells, latex particles instead of white blood cells, The change to CE was driven by several factors: the current contract animal platelets and other cells and particles. So it is very far from with the HPLC supplier was due to expire, an increased awareness of the real patient’s blood. Recently, there has been renewed an interest the need for a contingency due to ISO15189 requirements and the in quality control algorithms using patient samples: moving average need to future-proof against possible workload rises. (MA) algorithm. There were several barriers to overcome in changing methods. In our laboratory, the MA algorithm is being implemented in The primary barriers related to resistance to change within technical, addition to the use of the commercial control blood. scientific and medical staff groups, however there were also logistical barriers e.g. installation/maintenance, procurement/financial issues Methods and IT integration. Samples (sp) of patients with EDTA received from the hospital Methods departments. Exclusion criteria: age b 12 and N 86 years, patients with hematology disorders and under chemotherapy. A CAPILLARYS 2 FLEX PIERCING instrument was provided by The study of blood (EDTA) of patients was carried out on the Sebia on loan, allowing laboratory staff to become familiar with use hematological analyzer CELL-DYN Ruby (Abbott Diagnostics). For of the equipment and interpretation of the results. A period of daily quality control used control blood CBC-3K (R&D Systems) and a evaluation was carried out over 4 months where all samples received standard set of rules Westgard. for HPLC screening were also tested by CE. Over this period, over Measurements of leukocyte count (WBC), erythrocyte (Er), 1500 samples were tested, providing results covering common platelet (Plt) and hemoglobin concentration (Hgb) were chosen as variant haemoglobin traits and several variations of disease states. a model.

Results Results

The CE results agreed with the HPLC results, providing conﬁdence Based on the available capabilities for the calculation of MA, an in the method. approach was chosen with direct calculation of the sample average Logistical barriers were also addressed through this loan instru- and truncation limits, corresponding to the reference intervals of ment, as technical difﬁculties encountered during the evaluation WBC, Er, Plt, Hgb. Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

At the ﬁrst stage, we retrospectively estimated the average value imprecision and its corresponding 50% percentile average bias of hematological tests for our patient population (WBC – 13 059 sp, (pB50%). Plt – 15 347 sp, Er – 14 031 sp, Hgb – 11 801 sp), for HGB and Er additionally with sex stratiﬁcation (male - 4 473 sp and 6 091 sp; Results female – 7 328 sp and 7 940 sp respectively). Further, on the basis of the obtained averages, operational control Mean group imprecision decreased 41.9% except for P4 that charts (MA-chart) were built. The accuracy requirements were the increased 36.4%. Mean group bias decreased 45.9%, except for Ca19-9 same as for the control blood charts (LJ-chart). Daily mean values of that increased 114%, P4 41%, AD 28% and SDHEA 12%. Major patients with appropriate inclusion criteria and results within difference was observed in insulin, pCV50% and pB50% decreased reference intervals were calculated. Daily average number of patients 72% and 73% respectively. is: WBC, Plt - 80-100; Er, HGB male – 30 – 40, Er, Hgb female – 50 – 60. Conclusions MA-chart analyzed together LJ-chart for decision-making about the correctness of the obtained results of hematological analysis. Measurement quality improved both in imprecision and harmonization; this can be attributed to advances in technology and Conclusions platforms, clinical guidelines and laboratory quality awareness. Clinical chemistry analytes, which have traceability to international The following preliminary conclusions were drawn: reference standards, demonstrate greater harmonization than im- • Long-term CV can be calculated from MA-chart data. munoassays in general. P4 performance may be due to introduction • MA-chart allows to detect lot to lot variability of reagents, which is of new assay design by a major method. Increase in pB50% for Ca 19- impossible for the control blood due to the short shelf life. 9 and SDHEA may be attributed to one major method, not • MA-chart is sensitive to changes in the surveyed population harmonized with the rest. For AD, an important difference between • MA-chart can be an indicator of the unsuitability of the control automated and manual methods increased pB50%. Laboratory blood for further use (especially for Plt). performance evolution helped to achieve more reliable and repro- ducible results having a positive impact on patient safety. doi:10.1016/j.cca.2019.03.1080 doi:10.1016/j.cca.2019.03.1081

M328 M329 Clinical laboratories´ imprecision and bias evolution 2003–2018, fl from an external quality assessment scheme overview Quality management system in ow cytometry unit from a clinical laboratory. Implementation of international organization for standardization (ISO) 15189 C. Raspagliesia, K. Isaacka, S. Quirogac, L. Del Vecchiob, C. Fenilib,M. Torresb aCarrera de Especialización en Bioquímica Clínica, Departamento de B. Álvarez Flores, R. Guillén Santos, M. Medrano Élez, B. Soto Del Análisis Clínicos, Hospital Universitario CEMIC Centro de Educación Pecho, J. García De La Fe, L. Conejo Sánchez, F. Cava-Valenciano Médica e Investigaciones Clínicas, Buenos Aires, Argentina Laboratorio Central de la Comunidad de Madrid, BRSalud, Spain bCEMIC Programa Buenos Aires, Instituto Universitario CEMIC, Argentina Background-aim cDepartamento de Análisis Clínicos, Hospital Universitario CEMIC Centro de Educación Médica e Investigaciones Clínicas, Buenos Aires, Argentina The Laboratory accreditation verifies laboratories have an appro- Background-aim priate Quality Management System (QMS). International Standard (ISO) by ISO 15189 has been applied in clinical laboratories in many ISO 17043 accredited Buenos Aires Program (ProgBA) is an areas like biochemistry, immunology, pathology and microbiology. In fl External Quality Assessment Scheme (EQA) founded in 1979, with ow cytometry, recently there are more laboratories in the need of fl more than 1000 Latin American laboratories. ProgBA evaluates implementing these quality systems but is not easy because ow fl fi quality of results through cumulative indicators for assessing cytometry is a very exible technique in which it is not easy to nd performance. The aim is to compare the measurement´s state of the standardization in the use of the reagents, protocols, instruments art of some clinical chemistry and immunoassays analytes included and even criteria of acquisition of the events of each sample. in ProgBA between years 2003 and 2018. Our objective is explain our experience in the implementation of standard 15189 in the study of T lymphocyte subpopulations and in Methods the study of HLA-B27 and the challenges we had to overcome to obtain and renew the quality system. Lyophilized human sera samples, prepared from repeated pools Methods were sent to participating laboratories for monthly process. 46 analytes were included in the group: Cholesterol, Glucose, Creati- nine, Bilirubin, Uric acid, Amylase, CPK, AST, ALT, Calcium, Magne- For the quality control of the cytometers we use CS&T beads (BD) sium, Iron, LDH, ALP, Proteins, Sodium, Potassium, Triglycerides, GGT, and for the cytometer calibration 7 Color Setup Beads (BD). For the Urea, Albumin, HDL, TSH, FT4, T3, T4, FSH, LH, Estradiol, Cortisol, acquisition and the analysis, we use BD FACSCanto Software and the Testosterone, Progesterone (P4), Insulin, Prolactin, Ferritin, PSA, CEA, reagents Multitest CD3/CD8/CD45/CD4 and HLA-B27 Kit (BD). We IgE, HCG, AFP, Ca125, CA15-3 Ca19-9, 17-OHprogesterone, SDHEA, had to participate in external quality controls (EQC) and to use Androstenedione (AD). For performance comparison through time, internal controls (Multicheck CD4 and CD4 Low multicheck) using fl we considered yearly 50% percentile (pCV50%) average laboratory Two 8 Color ow cytometers (BD FACSCanto II). We needed to Abstracts / Clinica Chimica Acta 493 (2019) S497–S532 perform the validation test of the method, as well as records of Methods equipment maintenance, incidents, improvement actions, personnel training, Standard Operating Procedures, equipment and program A retrospective descriptive study of the determinations made manuals, technical data sheet, and more information such as external during six consecutive months has been carried out, the data have documentation, degrees of compliance and verification of internal been obtained from Modulab (Werfen), our laboratory information controls. system (LIS). The determination of CRP was performed in an Architect c16000 system (Abbott) by immunoturbidimetry and the Results determination of PCT in a Cobas E411 system (Roche) by chemilu- minescent immunoassay (ECLIA). We obtained accreditation UNE-EN ISO15189 by the National Accreditation Entity with the following scope: Results Name Description CD4 % Percentage of T CD4 T lymphocytes In the period of the study, 11,080 requests with CRP and PCT CD4 lymphocytes Absolute counts of T CD4 cells (cells/mm3) were requested and jointly determined. 14.8% (1,640) of these, had a CD8% Percentage of CD8 T lymphocytes CRP result b5 mg/L, 21.8% (2.418) CRP b10 mg/L, 26.8% (2,972) CRP CD8 lymphocytes Absolute counts of T CD8 cells (cells/mm3) b15, 31,3% (3,466) CRP b20 mg/L and 34.7% (3,842) CRP b25 mg/L. CD3 % Percentage of CD3 T lymphocytes 89.9% of the patients with CRP b5 had a negative value of PCT CD3 lymphocytes Absolute counts of T CD3 cells (cells/mm3) (b0.5 ng/mL), 90.3% with CRP b10, 89.6% with CRP b15, 88.8% with CD4/CD8 Ratio CD4/CD8 CRP b20 and 87.5% with CRP b25. HLA-B27 HLA-B27 PCT was positive in patients with CRP b5 in 10.1% of cases, with CRP b10,9.7%, CRP b15, 10.7%, CRP b20, 11.3%, and CRP b 25 12.5%. Conclusions If a filter is established in order not to perform the PCT determination (unless there is a high clinical suspicion) depending We obtained the ISO 15189 accreditation for the proposed on the level of CRP, it is estimated that it could mean an annual achievements but we had to deal with some obstacles like the saving of 26240 euros (CRP b5), 38688 euros with CRP b10), 42624 absent of internal control for the study of HLA-B27. The implemen- euros with CRP b15), 55456 euros with CRP b20) and 61472 euros tation of ISO 15189 has been useful to work correctly and a way with CRP b25). according to the established standards, but also for the control of risks. The next challenge is to accredit the ISO 15189 in hematolog- Conclusions ical pathologies, for which we will need greater standardization in the procedures (preanalytical, analytical and reporting) and the With CRP cut-off points b5, b10 and b15 mg / L, similar results are obtaining of EQC that allow us to perform an adequate inter- obtained (in %), being the CRP point b10 with which we obtain the comparing to the concrete pathology that we want to accredit. For best results (90.3% of negative PCT). this purpose, Euroflow system could be useful since it has standard When the CRP value increases (b15, b20 and b25), the data gets procedures for the processing of samples, calibration and analysis worst, which, in our opinion, we do not recommend to base the filter described in detail. on CRP values higher than 10 mg/L. We believe that the lowest impact for patients (9.7% of PCTN 0.5 doi:10.1016/j.cca.2019.03.1082 ng / mL) is achieved using a CRP value b10 with a substantial annual saving (direct cost 38.688 euros), since the number of determinations that are saved is much higher than the cut-off point CRP b5 and that difference is not maintained with the CRP cut b15. M330 We must remark that this protocol would be a base to start working and to which other parameters such as the number of Economic impact of a protocol on adequacy of procalcitonin leukocytes could be added and always recommending the determi- demand nation when there is a high clinical suspicion.

b a a R. Coca Zuñiga , A. González Raya , M.L. Hortas Nieto , E. Martín doi:10.1016/j.cca.2019.03.1083 Sálidoa, G. Callejón Martína, A. Lendinez Ramireza, M. Cantero Sancheza M331 aHospital Costa del Sol, Marbella, Spain bHospital Universitario Virgen de las Nieves, Granada, Spain Performance evaluating of Abbott 25-OH-vitamin D assay: comparison with HPLC and LC-MS/MS systems Background-aim

E. Avcia, R. Nara, S. Demira, H. Senolb Procalcitonin (PCT) is a very demanded test in the emergency aMedical Biochemistry Department, Faculty of Medicine, Pamukkale units with a high economic cost (8 euros/determination). The University, Denizli, Turkey synthesis of PCT can be induced directly by bacterial endotoxins or bMedical Biostatistics Department, Faculty of Medicine, Pamukkale indirectly by proinflammatory cytokines, so it constitutes an early University, Denizli, Turkey marker of systemic bacterial infections. According to several studies, PCT begins its elevation only 4 hours before the C-Reactive Protein Background-aim (CRP) does. The objective of this work is to evaluate the impact of the Vitamin D deficiency is a worldwide health problem caused implementation of a protocol of action in the control of the demand mainly by insufficient exposure to sunlight and dietary intake. of requests of PCT in the emergency laboratory in function of several Vitamin D deficiency represents several clinical findings such as CRP values cut-off points (5, 10, 15, 20 and 25 mg/L). muscle weakness, orthostatic hypotension, eczema etc. Liquid Abstracts / Clinica Chimica Acta 493 (2019) S497–S532 chromatography-tandem mass spectrometry (LC-MS/MS) as the best aCellule de Bio-Informatique (CBI), CHU de Bordeaux, Bordeaux, France method for quantifying vitamin D metabolites due to improved bDirection des Services Informatiques (DSI), CHU de Bordeaux, Bor- sensitivity, accuracy, and reproducibility. High-Pressure Liquid deaux, France Chromatography (HPLC) can effectively separate 25-hydroxyvitamin cLaboratoire de Biochimie & Cellule de Biologie Délocalisée (CBD), RMSB D3 [25(OH) D3], D2 [25(OH) D2], and other Vitamin D metabolites. UMR5536 CNRS-Université Bordeaux, CHU de Bordeaux, France Abbott diagnostic claimed, their 25- OH Vitamin D assay could dLaboratoire de Biochimie & Cellule de Biologie Délocalisée (CBD), CHU determine 25(OH) Vitamin D metabolites, with excellent accuracy de Bordeaux, Bordeaux, France and sensitivity. In the present study, we aim to investigate the eService du Biomédical, CHU de Bordeaux, Bordeaux, France performance of Abbott 25-OH-Vitamin D assay in contrast to HPLC fTechnical Specialist Department, Werfen, Paris, France and LC-MS/MS. Background-aim Methods CHU de Bordeaux is a large teaching hospital in France with a We randomly have chosen 80 serum specimens from the patients’ robust, decentralized POC testing (POCT) program. Management of samples pool during four days period. Serum specimen aliquoted decentralized blood gas testing across a network of acute care into three parts and analyzed with immunoassay [Abbott Architect i- settings and a vast number of operators (more than 800) requires a 2000 (Abbott Park, IL, USA)], HPLC and LC-MS/MS systems [Zivak rigorous quality framework, close partnership between the lab and HPLC and Zivak Tandem Gold Triple quadrupole (Istanbul, Turkey)]. clinical areas and reliable technology to ensure success. Continuous variables were expressed as mean ± standard deviation This analysis describes the management of 12 GEM® Premier™ (SD), median (minimum-maximum values) and categorical variables 4000 and implementation of 7 new GEM Premier 5000 blood gas as number and percent. Shapiro–Wilk tests were used for testing analyzers networked to GEMweb® Plus 500 Custom Connectivity normality. We used kappa analysis to evaluate agreement between (Instrumentation Laboratory) in acute care settings across two gold standard and HPLC and IA measurements. Sensitivity, Specific- hospital sites using ISO 22870 framework. ity, Negative and Positive predictive values were used f to analyze the performance of HPLC and IA measurements. Wilcoxon signed Methods rank test was used for determining the difference between gold standard values and other technics. Venn diagrams were used to Three dedicated lab technicians supported the roll-out of the new examine consistency between 3 methods. All statistical analyses blood gas analyzers using ISO 22870 framework to meet the clinical analysed by SPSS, 24.0 and p-value less than 0.05 was considered and quality requirements of each area throughout the pre-installa- statistically significant. tion, validation and go-live phases. The process required a close and constant partnership between the laboratory, clinical care services, Results biomedical staff, material management, suppliers and many other stakeholders. Pre-planning included: definition of requirements, We accepted deficiency/insufficiency/sufficiency/toxic levels re- compliance, connectivity, patient records, method validations, doc- spectively 0-10& 10-20& 20-70&N70 (ng/ml) those were defined by ument management, competency training, and consumables, among World Health Organization (WHO). Patients’ age means value was others. All phases of the roll-out examined key quality indicators to the 50.2±17.69 year. D vitamin mean values 21.2±14.49 nmol/L, measure success. HPLC was 22.72±14.83 nmol/L and IA 19.45±14.73 nmol/L. There was strong accordance among three-assay method. As a gold Results standard method LC-MS/MS, sensitivity, specificity, positive and negative predictive values for HPLC and IA were respectively 88.9- The 7 new GEM Premier 5000 analyzers were added to the 93.3%, 94.3-82.9%, 95.2-87.5%, and 86.8-90.6%. network in less than 6 months, with a carefully executed quality plan. Over 495 care staff in 10 sectors were trained as a part of the Conclusions program. Management and implementation of the new systems were facilitated by the built-in risk management features of the In deficiency clinic, IA more compatible then HPLC with the gold GEM Premier 5000 with iQM2 and GEMweb Plus, specifically as standard. HPLC was successful in insufficiency than the IA method. In it pertained to quality management (including COFRAC reports), the present study Abbott 25-OH-Vitamin D assay is appropriate for error detection, device management and operator competency determining Vitamin D status. training. doi:10.1016/j.cca.2019.03.1084 Conclusions

Implementation and management of a broad blood gas program requires rigorous quality standards, processes and technology. Active M332 collaboration of the main clinical and laboratory stakeholders in a decentralized roll-out is fundamental. GEM Premier analyzers tied Point-of-care (POC) deployment and management of blood gas into GEMweb Plus can be effective tools to facilitate accreditation analyzers following an international organization for standardi- compliance and decentralized testing management, particularly as it zation (ISO) 22870 quality framework pertains to risk-management requirements.

d d d d b G. Castaing , C. Terral , N. Berthon , R. Viaud , S. Senrens ,M. doi:10.1016/j.cca.2019.03.1085 Galvezb, J. Domingorenab, D. Rouxb, J. Hauraye, J. Redine, J. Catele,Q. Vua, S. Brandf, L. Dusseaua, M. Beauvieuxc Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

M333 M334

Evaluation of automated quality control (QC) feature of Alinity ci Implementation of a system of management of quality pre- System analytical

N. Kumproaa, S. Promnucha, P. Bumrungpola, S. Punprasita, C. Chenb, L. Criado Gomez, N. Seco Moro, S. Villanueva Curto, M.B. PÉrez S. Vanavanana Sebastian, J. Reig Del Moral aDivision of Clinical Chemistry, Department of Pathology, Faculty of Laboratorio Analisis Clinicos, Hospital Universitario de Móstoles, Spain Medicine Ramathibodi Hospital, Mahidol University bScientiﬁc Services, Abbott Laboratories Background-aim

Background-aim The pre-analytical phase is the most critical in the clinical laboratory. Around 70% of errors occur, affecting significantly to To evaluate the automated QC feature of Alinity ci system (Abbott the safety of the patient. The aim of our study was to establish a Laboratories) and its benefits in laboratory operations. system of management of the pre-analytics incidents and check if we comply with the quality specifications given by the SEQC in Methods 2017.

A Sigma metric study was carried out on 4 immunoassays and 7 Methods clinical chemistry assays to compare automated QC mode with manual QC using Technopath multiconstituent control materials. Computer tests representing the following incidences were Sigma metric values were calculated based on 5-day precision and created: erroneous identification, not received serum samples, not the bias of the mean values against target values to assess the received urine samples, inadequate serum samples, and hemolyzed quality of QC results. In the automated QC mode, controls were sample. pipetted and tested directly from the onboard control vials. The The specifications were established: control lot number, target values and range were read by the 1. Number total of rejections/number total of analytical requests instrument from the barcode of the vials. In the manual QC mode, (NT) (%) (optimal 1.39, desirable 2.21, minimum 3.16) controls were manually pipetted into sample cups prior to testing 2. Number of discrepancies in identification/NT (%) (O 0.001, D 0.01, in each run. A manual input of the control lot number, target values M 0.032) and range was also required. The performances of precision, sigma 3. Number of serum samples rejected/number of creatinine determi- metrics and turnaround time were evaluated on both QC modes. nations (NTC) (%)(O 0.448, D 1.16, M 2.09) The turnaround time was determined as the cumulative time of the 4. Number of not received serum sample / NTC (%) (O 0.09, D 0.18, M QC procedure from the QC material preparation through QC test 0.33) completion. 5. Number of hemolyzed sample / NTC (%) (O 0.218, D 0.73, M 1.67) 6. Number of inadequate serum sample / NTC (%) (O 0.005, D 0.035, Results M 0.098) 7. Number of not received urine sample / number of determinations The precision and sigma metrics resulting from automated QC are with urine (urianalysis and/or biochemistry( (%) (O 1.848, D 3.106, comparable with those from manual QC mode. The within-laboratory M 5.01) precision for the 7 assays ranged from 0.52% to 3.95% CV for 8. Number of not received urine sample / NT (%) (O 0.342, D 0.825, M automated QC and from 0.41% to 3.31% CV for manual QC. Ninety- 1.351) one percent of the assays tested on the automated QC mode (or 10 out of 11) and 82% on the manual QC mode (or 9 out of 11) were He was carried out in the full year 2018, and quarterly way of found to be operating at 5 Sigma or above. Time savings of 22.3% seven primary care centers attends to our laboratory. Was Compared were achieved by using the automated QC mode as compared with with the specifications given by the SEQC in 2017. manual QC. Results Conclusions

The results were satisfactory. Speciﬁcations 1, 2, 3 and 5, came to While the precision and Sigma-metrics performances were desirable in all primary care centres during the four quarters of the comparable between the automated and manual QC modes, the year. Speciﬁcations 4 and 6 remained within levels minimum and automated QC feature of Alinity ci System enables the saving of time several centers in 1 or 2 quarters were above. Instead those related in daily QC procedure and increase of walkaway time, leading to a to urine (7 and 8) had in all health centres and in several quarters a reduced labor requirement, improved turnaround time and en- level above the minimum. It is a primary care centre with the best hanced operational productivity. indicators, all of them at desirable levels and optimal except in the urine in two quarters. doi:10.1016/j.cca.2019.03.1086 Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

Conclusions M336

We think that a better tool is very useful for extraction points, for Cooled storage space and solid infectious waste production: this reason the clinical laboratory implement an annual report to results of a comparative study across six immunochemistry each centre with its results, high lighting the strengths and analysers suggesting actions of continuous improvement as the management of samples of urine in this case, and we will continue measuring P. Findeisena, I. Zahna, D. Krempela, C. Garcia Rabanedac, T. Haro these indicators as management of pre-analytical phase. Romeroc, T. De Haro-Muñozc, M. Barral Juezc, S. Engelmannb aLaboratory Dr. Limbach and Colleagues, Heidelberg, Germany doi:10.1016/j.cca.2019.03.1087 bRoche Diagnostics GmbH, Mannheim, Germany cUGC Análisis Clínicos intercentros, Hospital Universitario Campus de la Salud, Granada, Spain

M335 Background-aim

Strategy for the anemia parameters demand management in During this study, we analyzed two aspects – reagent- and waste primary care handling – that are of high relevance for everyday organization in diagnostic laboratories. a a b a N. Del Amo , E. Marquez , R. Ramos , S. Garcia-Valdecasas , M.J. We compared six commercially available immunochemistry a a a a Ruiz , M.B. Alvarez , R. Guillen , F. Cava analysers, with respect to cooled space required to store reagents a Laboratorio Central de la Comunidad de Madrid, BRSalud, Spain and their production of solid infectious waste. bSpanish Society of Laboratory Medicine (SEQC), Patient Safety Commission, Spain Methods

Background-aim The comparative evaluation was performed at two laboratories in Germany and Spain on analyzers from five different manufacturers: Laboratory tests adequacy is a key element towards quality of two cobas e 801 analytical modules (Roche Diagnostics), ARCHITECT fi medical laboratory outcomes, in terms of patient safety, ef ciency i2000SR (Abbott), UniCel DxI 800 (Beckman Coulter), Liaison® XL and effectiveness. Several studies revealed high test request (DiaSorin), ADVIA Centaur XPT and IMMULITE 2000 XPi (Siemens fi variability and a signi cant test over request. The aim of the present Healthineers). work is to assess the impact of a demand management strategy for Demands for cooled storage space were determined for typical anemia parameters in the context of primary care. clinical laboratory assays on basis of the respective manufacturer’s reagent packaging. Our analysis included a total of 18 assays Methods (Ferritin, PTH, Estradiol, FSH, beta-HCG, LH, Progesterone, Prolactin, Testosterone, AFP, CA 125, CA 15-3, CA 19-9, CEA, fPSA, tPSA, free T4, A multidisciplinary group was created for the evidence medicine TSH) covering reagents from five indication areas. protocols revision and algorithm elaboration. The algorithm was The production of potentially infectious solid waste was evalu- introduced thanks to the laboratory information system middleware. ated by processing standardized hospital /commercial laboratory like The experience was piloted in three primary care centers prior to workloads on each analyser and determining the weight of the solid total implantation. waste output. In addition, the solid waste production per determination including that contributed by the empty reagent packs was Results calculated per system.

The algorithm designed condition biochemical anemia parameters Results to hemoglobin and medium corpuscular volume (MCV) results. If no anemia is evidenced, no further investigation is made. If anemia is The cooled storage space requirement for the test panel differed evidenced, according to MVC result different test are completed: iron signiﬁcantly among the included analysers. Total storage volumes and ferritin (b80 fL), vitamin B12 and folic acid (N98 fL) or iron, ferritin, (L) ranged from 79L (cobas e 801 system) up to 925L (Immulite lactate dehydrogenase and reticulocyte count (80-98 fL). The results of 2000 XPi) for an identical number of determinations including the strategy application were the decrease of the anemia parameters calibrators. as follows: iron 43% (CI: 34-51), ferritin 42% (CI: 33-51), transferrin The amount of potentially infectious solid waste ranged between 0.6g 92% (CI 91-94), vitamin B12 66% (CI 53-79) and folic acid 70% (59-82%) (cobas e 801 system) per determination up to 2.8g (ADVIA Centaur XPT). Considering the solid waste contributed by empty reagent packs Conclusions per determination and identical workloads representing both commercial and hospital like workloads, a waste reduction of up to Mayor decrease was obtained for transferrin, vitamin B12 and 77.5% was observed on the cobas e 801 system in comparison to folic acid. Transferrin was inadequate used for iron deﬁciency ADVIA Centaur XPT. anemia. Vitamin B12 and folic acid were wrongly demanded for no anemic and/or microcytic anemias. Conclusions The impact of test over request is minimized due to the applied algorithm, patient safety is not compromised and an improvement of Our study provides laboratory managers with comparative cooled resources was achieved. storage space and solid waste data over various commercially available immunochemistry systems. doi:10.1016/j.cca.2019.03.1088 doi:10.1016/j.cca.2019.03.1089 Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

M337 M338 Estimation of measurement uncertainty in some haematological quantities in Sysmex XN analyzers Indirect estimating T3 and T4 reference intervals based on outpatient population in a regional hospital in Taiwan L. García Tejada, L. Sánchez Navarro, L. Rapún Mas, R. Rigo Bonnin Bellvitge University Hospital, Spain C. Kao, L. Hsu Department of Laboratory, St. Martin De Porres Hospital, Chiayi City, Background-aim Taiwan

Background-aim The latest update of the ISO 15189 accreditation standard requires that laboratories know and estimate the uncertainty of the results provided by their measurement systems. Thyroid disease is common in most people. It is important to ’ This metrological concept is taking a huge interest in clinical compare the patient s thyroid function test results with reference biochemistry, but it is still quite unknown in the haematology field. intervals(RIs) derived from a matched population. Ideally, each The aim of this study is to estimate the measurement uncertainty laboratory should determine RIs based on representative studies of of seven haematological quantities results measured in Sysmex XN the target population. However, to establish RIs using traditional fi analyzers (Roche Diagnostics®). method is dif cult to perform, time-consuming, costly, and often inaccurate. In light of these difficulties, most laboratories elect not to ’ Methods establish their own RIs, but rather choose to verify manufacturer s RIs which may be established in different countries and do not necessarily apply to the local population. Some alternative methods The quantities evaluated were: number concentration of eryth- have been suggested and include the use of outpatient population rocytes (RBC), reticulocytes (RET), leukocytes (WBC) and platelets instead of healthy volunteers. The aim of this study was to estimate (PLT-I for impedance and PLT-F for fluorescence method), mass RIs for T3 and T4 using matched clinical population. concentration of haemoglobin (HB) and volume fraction of erythro- cytes (HTO). Methods Expanded measurement uncertainty (U) was estimated for nine Sysmex XN analyzers. For each quantity and analyzer, the following formula was applied: We employed a posteriori study, so-called indirect method, where results from specimens were collected for routine clinical 2√(ucal^2+up^2+uδ^2), being: care purposes and used to determine the RIs. Thyroid function test 1. ucal: uncertainty associated with the values assigned to the (TSH, free T4, T3, and T4) results were measured using two Roche calibration material (XN CAL PF for PLT-F and XN CAL for the rest Cobas e601 analyzers and over one and a half years period (Jan. of quantities). 2017~Jun. 2018) were retrieved from our 531-bed regional hospital fi 2. up:uncertainty associated with imprecision expressed as coef - outpatient electronic medical record system to constitute the original cient of variation (CV), obtained by measuring XN CHECK Level L2 database for this study. The division of metabolism and endocrinol- controls during a period of 4 months. ogy outpatients were excluded. For each T3 and T4 test, the presence 3. uδ:uncertainty related to the bias of the measured system. The of both TSH and free T4 tests without any abnormality were included √ ™ ™ formula applied was ( ^2+ux^2+uμ^2 ) , being the bias, ux to assure normal thyroid status. The RIs were estimated the central uncertainty associated to the mean value of the control results, and 95% range of the population by a nonparametric method using EP uμ uncertainty corresponding to the value assigned to the control Evaluator software. material. All data were obtained from the external quality scheme assessment SNCS-IQAS Online. Results

Since control lot changes every two months, weighted data from Original T3 and T4 manufacturer’s RIs were 84.6~201.8 (ng/dL) two different lots were employed to estimate up and uδ. and 5.1~14.1 (ug/dL). The newly implemented T3 and T4 RIs Finally, in order to obtain a single uncertainly value for each we established from our own hospital outpatients were 50.9~142.1 quantity, a weighted U for the nine analyzers was calculated. (n = 1772) and 4.6~9.7 (n = 499) respectively and successfully veriﬁed by the conventional approach with 20 healthy Results volunteers.

The weighted U obtained for each quantity was: 3.53% for RBC, Conclusions 2.80% for HB, 6.43% for HTO, 8.07% for WBC, 12.33% for PLT-F and 19.45% for RET. This study showed the “indirect method” for RIs estimate is suitable, faster, cheaper, and can provide large patient numbers for Conclusions more robust assessment for laboratories use the manufacturer’sRI derived from unmatched population there is a need to establish a Uncertainty allows quantifying the quality of the result and new RI. evaluating its reliability, providing a correct clinical interpretation of it. This study could help and motivate clinical laboratories to perform doi:10.1016/j.cca.2019.03.1091 uncertainty studies in the haematology ﬁeld. doi:10.1016/j.cca.2019.03.1090 Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

M339 aBIONET S.A., Santiago, Chile bClinical Laboratory Bioanalise, Teresina, Piaui, Brazil Three-years of experience of accuracy-based lipid proficiency cClinical Laboratory Network from the state of Jalisco, Public Health testing service in Korea State Laboratory (LESP), Comisión para la Protección contra Riesgos Sanitarios del Estado de Jalisco (COPRISJAL), Mexico d J. Kim Facultad de Medicina, Universidad Católica de Cuenca, Cuenca, Yonsei University College of Medicine, South Korea Ecuador eInternational Laboratories Services, INTERLAB S.A., Guayaquil, Ecuador Background-aim fLaboratory of Thrombosis and Hemostasis, Department of Hematology- Oncology, School of Medicine, Pontificia Universidad Católica de Chile, Accuracy-Based Lipids (ABL) Proficiency Testing (PT) program Santiago, Chile g started from 2016 by Korean External Quality Assessment Service LAC, Montevideo, Uruguay h (KEQAS) to minimize matrix effect. We present the three-years of Section of Clinical Biochemistry, Department of Neurosciences, Bio- experience of ABL PT. medicine and Movement Sciences, University of Verona, Verona, Italy iServicio de Acreditación Ecuatoriano (SAE), Quito, Ecuador j Methods Universidad de Guayaquil, Guayaquil, Ecuador kUniversidad Peruana Cayetano Heredia, Lima, Perú We made six kinds of commutable frozen sera according to the Background-aim CLSI 37A guideline and distribute in two rounds every year from 2016 to 2018. We got the reference values of total cholesterol (TC), HDL cholesterol (HDLC), LDL cholesterols (LDLC) of each fresh frozen The Working Group for Preanalytical Phase in Latin American pool from the reference measurement laboratories, CEQAL, the one of (WG-PRE-LATAM) of the Latin America Confederation of Clinical the cholesterol reference measurement laboratory network labora- Biochemistry (COLABIOCLI) was established in 2017, and its main tory and total glycerides and free glycerol reference values by isotope purpose is: to study preanalytical variability and to establish dilution gas chromatography mass spectrometry from National guidelines for preanalytical procedures in order to be used by clinical Medical Reference Laboratory (NMRL) in Korea Center for Disease laboratories and healthcare professionals in Latin America. This study Control and Chemical Metrology Laboratory, Health Sciences Au- on behalf of COLABIOCLI WG-PRE-LATAM aiming at evaluating thority (HAS) in Singapore or ReCCS in Japan, respectively. We whether an Ecuadorian breakfast can interfere with thyroid function evaluate the average %bias of participating laboratories according to assays. National Cholesterol Education Program (NCEP) bias limit. Methods Results We studied 20 healthy volunteers who consumed an Ecuadorian The number of participating laboratories of TC, HDLC, LDLC, Total breakfast containing a standardized amount of carbohydrates, glycerides, and Triglycerides were increased from 164 to 223, 163 to proteins, and lipids. We collected blood specimens for thyroid 223, 158 to 214, 98 to 139, and 61 to 82, respectively. The average %bias stimulating hormone (TSH), and free thyroxine (fT4) before fi (consensus vs. reference values) of all participating laboratories for TC, the breakfast and 1, 2, and 4 hours thereafter. Signi cant HDLC, LDLC, Total glycerides, and Triglycerides were +0.14%, −0.54%, differences between samples were assessed by the Wilcoxon +2.9%, −1.08%, and −1.32%, respectively. Although the average %bias ranked-pairs test. or absolute %bias showed within the bias limit of NCEP, the exceeding cases of bias limit sometimes occurred. The average % bias of LDLC Results showed exceeded NCEP bias limit most frequently (8 out of 18 pools). Instrument-specific bias estimation report seemed to stimulate each The Ecuadorian breakfast impressively decrease thyroid stimu- manufacturer to keeping traceability. lating hormone (TSH) 1 hour after breakfast vs. baseline specimen 1.33 mIU/mL (0.90 – 1.83) vs. 1.82 mIU/mL (1.03 – 2.21), respec- Conclusions tively; P b 0.001. Free thyroxine (fT4) decreased 7.1% four hours after Ecuadorian breakfast 12.7 pmol/L (11.6 – 14.4) vs. 13.6 pmol/L (12.1 – Although the average %bias of participating laboratories for TC, 15.7), respectively; P = 0.014. HDLC, LDLC, Total glycerides, and Triglycerides showed within the bias limit of NCEP, the exceeding cases of bias limit sometimes occurred Conclusions especially in LDLC during recent 3 years in Korea and ABL PT can be useful to keeping traceability not worrying about matrix effect. Findings of this study reveal that an Ecuadorian breakfast can influence the thyroid function assays and might expose patient doi:10.1016/j.cca.2019.03.1092 safety to some risks. Therefore, the COLABIOCLI WG-PRE-LATAM calls attention and highlights that the fasting time needs to be carefully considered when performing blood testing in order to prevent spurious results and thus, reduce laboratory errors. M340 doi:10.1016/j.cca.2019.03.1093 Breakfast jeopardize thyroid function assays: an evaluation on behalf of COLABIOCLI WG-PRE-LATAM

G. Lima-Oliveirah, W. Bajañae, E. Arandaf, M.E. Arredondoa, L.M. Brennan-Bourdonc, M.D. Campelob, E. Espinozai, S. Floresk, P. Ochoad, V. Vegaj, B. Varelag Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

M341 M342

Management 3.0 applied in clinical laboratories Automatic immunofixation on the Interlab G26 EasyFix system: Analysis of workflows and benefits V. Luis Consultoria y Capacitacion, Santiago, Chile G. Marchesini, A. Paternoster, G. Marzotto, L. Bedin, T. Guidolin, D. Giavarina Background-aim Laboratory Medicine, St. Bortolo Hospital, Vicenza, Italy

Clinical laboratories have been working on quality for more than Background-aim 50 years. In all this time, the technologies have improved and the analytical quality has reached a level of excellence in most clinical The aim of this study was to assess the impact on our laboratory laboratories, the non-analytical quality focused on the processes and workflows of the introduction of a completely automated analyser workflows (Lean) has begun to work no longer more than 10 years A (Interlab G26 EasyFix) for performing serum and urine immuno- new methodology that was born in the area of software develop- fixation (s-IFE and u-IFE). ment and that is very useful for the empowerment, leadership and commitment of the work teams (Lean Agile) has been working for Methods less than 5 years in the area of Health. This methodology is based on a series of activities that seek to improve the value delivered to the This assessment required the prior identification of some Key client focused on a human factor. There are a series of tools that can Point Indicators (KPIs) and their comparison against the preceding be used in agile methodologies, among them is SCRUM, Kanban, adopted solution (the semi-automated Hydrasys analyser in Extreme programing. For this study, Management 3.0 created by conjunction with the Assist sampler). The following KPIs have been Jurgen Appelo was used. This is a series of concrete practices that identified: Turn Around Time (TAT), Full Time Equivalent (FTE), help inspire managers and team members with the idea of Average Walk Away time, number of interventions by the operator, generating positive change in the organization. The objective of this involved personnel, number of installed instruments. The study of study is to determine the contribution of Management 3.0 in the the workflows has been performed by using Gantt diagrams. The management applied to Clinical Laboratories. related results refer to a typical 8 hours working day during which an average of 4 s-IFEs and 16 u-IFEs are performed. Methods Results A series of meetings was organized with the directors of 5 different clinical laboratories in Santiago of Chili to determine the The following results have been obtained respectively on the type of Management 3.0 activity to be carried out with the laboratory Interlab and on the Sebia systems: TAT (6h 8’ Vs 6h 16’), FTE (0,03 Vs staff, teams of 4 to 7 people were created with which they worked 0,07), Average Walk Away Time (1h 23’ Vs 6’), number of operator once a week in sessions of no more than two hours. According to the interventions (4 Vs 28), involved personnel (1 technician + 1 objectives set, we worked on Mental Map activities, Moving validator Vs 2 technicians + 1 validator) – number of installed Motivators, Delegation Poker and Skills Matrix. The results of each instruments (1 Vs 2). The Interlab analyser allowed to significantly of these activities were placed in the office of the director of the improve the FTE (by 57%), the Average walk Time (by 13 times) and clinical laboratory in order to visualize what was working. the number of operator interventions (by 600%). It shall be noted that the number of installed instruments (1 Vs 2) has been reduced, Results as well as the operator dedicated time.

A week after the first activities planned with the work teams, Conclusions important changes were observed in the work environment, improved communication flow between people and allowed to At equal analytical and diagnostic performances, the INTERLAB advance in a better way with the following planned activities. Once G26 EasyFix instrument, by representing an all-in-one solution, has these activities that took about a month were completed, a series of proved its capacity to optimize the workflows in the clinical actions were implemented that improved the flow of information laboratories while guaranteeing the full automation and traceability among the staff, who participated in the activities were more of samples, antisera and gels. Furthermore, the significant improve- committed and motivated to do new Management 3.0 activities ment of technical FTE and of the involved personnel allowed to and generated an interest of the rest of the staff to know and estimate a reduction of costs, relieving human resources that can be participate in any of the activities carried out allocated to other laboratory sectors. The Interlab G26 system results, therefore, to be a valid alternative to the SEBIA system for Conclusions performing second level tests aimed at immunologically characterising the serum and urine monoclonal components (s-IFEs Lean Agile helps make changes in the organization in a short and u-IFEs). time. The activities carried out in Management 3.0 allow adding value to the product generating a change in the work environment doi:10.1016/j.cca.2019.03.1095 that allows to do the activities in a better way, in a shorter time at a low cost. Workers who can make decisions and trust them are more committed and empowered workers in their daily activities. doi:10.1016/j.cca.2019.03.1094 Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

M343 M344

Risk assessment for planning the reagent water quality control in Urine test strip ascorbic acid interference: A risk analysis the clinical laboratory N. Nikolac Gabaj, M. Miler, A. Vrtaric, J. Culej, M. Bozovic, A. Topic, L. M.E. Mendes, A. Nascimento De Carvalho, R.K. Gonçalves De Almeida, Milevoj Kopcinovic N.M. Sumita Department of Clinical Chemistry, Sestre milosrdnice University Hospital Central Laboratory Division, Hospital das Clinicas of University São Centre, Zagreb, Croatia Paulo, Brazil Background-aim Background-aim Ascorbic acid (AA) interference causes falsely negative glucose The aim was to present the Failure Modes and Effects Analysis (U-Glc), blood (U-Hb), nitrite (U-Nit) and bilirubin (U-Bil) results in (FMEA) risk analysis tool for planning the reagent water(RW) quality urine test strip analysis. Our aim was to determine the frequency of control in the clinical lab to improve cost reduction and patient AA positive urine samples submitted for analysis to our department safety.This is a critical process and an error–prone,based on the and to evaluate their effect on patient safety. severity of possible harmful events may cause a potential dangerous impact to the patient. Methods

Methods Results from N = 27,856 samples submitted to our laboratory for urinalysis in one year were retrospectively retrieved from the The FMEA in this process involved:a multidisciplinary team; laboratory information system (LIS). Samples were collected using collecting and organizing information on the process;risk analysis; VACUETTE® TUBEs Z Urine No Additive (Greiner Bio-One, identifying failures modes for each step;determining the potential Kremsmuenster, Austria) and analysed on the Iris IQ200 (Iris effect of each failure mode; ranking the severity of failure mode Diagnostics, Chatsworth, USA). iChem Velocity test strips (Iris effects(S); ranking the probability of occurrences(O) and detection Diagnostics, Chatsworth, USA) were used for U-Glc, U-Hb, U-Nit , capacity(D) of each failure mode. The risk priority score(RPS) was U-Bil and AA determination. Risk analysis combined the impact of calculated by multiplying S*D*O considering:RPSδ40 acceptable and/ erroneous results (due to AA interference) on patient safety or residual risk and RPSN40 means high risk. Identifying and (severity-S) with their frequency (occurrence-O). Risk was identified prioritizing the critical failures modes. Implementing actions through as high, intermediate and low. the Quality Control Plan (QCP), covering all uses of RW in the laboratory.New evaluation was made after the action plans were Results completed. Negative AA results were detected in 25,012 (89.8%) urine Results samples, while 1199 (4.3%) and 1646 (5.9%) were mildly (20 mg/ dL, 1+) and highly positive (40 mg/dL, 2+), respectively. We The RW process was mapped and the risks were analyzed for identified 4 possible errors: false negative U-Glc, U-Hb, U-Nit and each step. The RPS (pre and post) was calculated. The QCP generated U-Bil. Five S classes were identified ranging from the lowest harm S1 was defined with the following topics: training for staff; daily to the highest S5 (i.e. S1 for U-Bil, S2 for U-Glc, S4 for U-Nit and S5 resistivity measurement; weekly microbiological control of water for U-Hb). Based on LIS results, O was categorized as O1b3%, O2=3- (heterotrophic bacteria); monthly control (endotoxins and total 10%, O3=10-25%, O4=25-50%, O5N50%. Accordingly, errors were organic compounds);control of particulate matter filtration/the type classified as O2 for U-Glc, and O3 for U-Bil, U-Nit and U-Hb. The risk of water; preventive maintenance of the purification system (PS); analysis 5×5 matrix revealed that false negative U-Hb was associated calibration of manometer, resistivimeter and thermometer; control with high risk, false negative U-Nit with intermediate risk, while of pre-filters of the PS; traceability of the filters; cleaning of the false negative U-Glc and U-Bil were associated with low risk on water tank (1/year); redundancy of PS; contingency plan; criteria of patient safety. acceptance/ type of reagent water. Conclusions Conclusions The frequency of AA positive urine samples submitted for analysis The risk evaluation for planning and control RW helps to reduce was quite high in our study. Furthermore, AA interference with U-Hb the occurrence of adverse events in the clinical laboratory. The and U-Nit analysis causing false negative results, is associated with reliability of RW in the lab depends on the risks identified and higher risk of harmful effects on patient safety. AA test strip analysis evaluated so that the QCP is appropriate to the needs, prevents errors should be routinely included in urinalysis protocols and AA results by the use of the RW and can avoid waste. FMEA and QCP brought should be taken into account when interpreting urinalysis test results. more quality to RW, reducing errors and costs, increasing the effectiveness of the process and patient safety. doi:10.1016/j.cca.2019.03.1097 doi:10.1016/j.cca.2019.03.1096 Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

M345 M346

PRECISION OF THE NEW MULTI-TEST VITROS® CHEMISTRY Homocysteine: Validation and comparison of two methods using PRODUCT SLIDES* ON THE VITROS XT 7600 INTEGRATED SYSTEM samples from neurology patients

J. Miller, T. Dimagno W. Masrib, E. Plouvierb, T. Wtulichb, H. Broutierb, C. Petitb,M. Ortho Clinical Diagnostics Bendaouda, H.A. Cungb, Y. Costaa aLaboratory, GHEF Marne La Vallee Site, Marne La Vallee, Est – Ile de Background-aim France, France bLaboratory, GHEF Meaux Site, Meaux, Est – Ile de France, France Ortho Clinical Diagnostics (Ortho) has developed the new VITROS® XT 7600 Integrated System (VITROS XT 7600) which Background-aim utilizes digital chemistry technology to support the new VITROS XT Chemistry Products Slides (XT MicroSlides). The XT MicroSlides with High homocysteine (Hcy) levels have been observed in throm- dual test capability are intended to reduce sample size and enhance boembolic diseases, psychiatric (schizophrenia, depression) and operational efficiency while maintaining analytical performance neurodegenerative (Parkinson’s disease, Alzheimer’s disease) pa- versus the conventional single slide test. Digital chemistry technol- thologies, showing a potential role of Alzheimer’s in the pathogen- ogy enables the use of imaging algorithms to improve chemistry esis of these disorders, hence the importance of developing Hcy results, such as within-lab precision. This study will examine within- assay techniques in our laboratory. The objective of this study was to lab precision, comparing the VITROS XT 7600 with digital chemistry validate the method of enzymatic plasma quantification Hcy on capability versus the VITROS 5600 Integrated System (VITROS 5600) Cobas 6000 c502® Roche Diagnostics™ and to compare the results with traditional spectrophotometry. obtained by our method on neurology patients with those obtained by liquid chromatography tandem mass spectrometry (LC-MSMS). Methods Methods The precision of six XT MicroSlides: UREA-CREA Slides, ALTV-AST Slides*, TRIG-CHOL Slides*, ALB-TP Slides*, GLU-Ca Slides*, and TBIL- The Hcy Enzymatic Assay (Hcy-EA) is based on an enzyme cycling ALKP Slides* for two serum concentrations was evaluated using assay principle that assesses the co-substrate conversion product. quality control materials on a VITROS XT 7600. Total within-lab The validation parameters such as repeatability, intermediaire precision (reported as percent coefficient of variance) for a single precision, accuracy, quantification limits as well as linearity and calibration was evaluated with two runs per day with two replicates measurement uncertainty were evaluated using internal and exter- per run over 20 days, for a total of 80 replicates following CLSI EP05 nal quality controls. The results obtained by our method and those guidelines. by the LC-MSMS technique on 30 patients were compared. The In addition, customer quality control precision data from single statistical data are processed by the SHGTA-04 (SPSS software) with test slides from VITROS 5600s were extracted and evaluated. The Ricos objectives. customer quality control data points were screened for statistical outliers and outliers were removed. The %CV was calculated for each Results lot/fluid/system combination and compared to the within-lab precision for the XT MicroSlides on the VITROS XT 7600. The repeatability and reproducibility study on Hcy concentrations at 12 and 40 ⎧mol / L showed coefficients of variation respectively Results b1.5% and b2%. Accuracy and precision were consistent with the objectives. The linearity covered a range of 50 ⎧mol/L and The within-lab precision for VITROS XT MicroSlides on the quantification limit was 5.5 ⎧mol/L. A result of Hcy at 12 is made VITROS XT 7600 was compared to the corresponding single test +/- 0.5 ⎧mol/L. The statistical analysis of the patient results allowed slides within lab precision on the VITROS 5600. The XT MicroSlides to show a correlation (r: 0.988, p b0.0001) and a strong agreement lots reported within-lab precision as improved or equal for 19 out of between the two assay techniques. The results obtained by the Cobas 24 examined assay/fluid combinations. The exceptions being CHOL* are overestimated by 20% compared to those obtained by LC-MSMS, at PV1 (1.9% vs. 1.5%); CREA at PV2 (1.6% vs. 1.4%); TBIL* at PV1 highlighting the need to change the threshold of normality and (3.8% vs. 3.1%); ALTV* at PV1 (2.4% vs. 1.9%); and ALTV* at PV2 (1.7% clinical decision-making currently defined for LC-MSMS from 10 vs. 1.4%). ⎧mol/L to 12 ⎧mol/L for Hcy-EA.

Conclusions Conclusions

The data presented here demonstrate that the new XT Micro- The validated Roche Diagnostics™ enzymatic plasma Cobas 6000 Slides using new imaging algorithms provide comparable or c502® Hcy-EA method demonstrated good performance in quanti- improved precision relative to single test slides from internal and fying plasma Hcy levels. The concordant results obtained in this external precision studies. study allow us to realise Hcy-EA at Meaux and a resumption of an outsourcing activity for a better follow-up of neurology patients. doi:10.1016/j.cca.2019.03.1098 doi:10.1016/j.cca.2019.03.1099 Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

M347 Background-aim

Monitoring of quality indicators—First results Identification and quantification of monoclonal component (MC) in serum and urine (Bence Jones protein, BJP) by electrophoresis and D.A.T.K. Polak Erceg, KlasićBezek, Novački immunofixation, are included in clinical protocols for patients with Special Hospital for Medical Rehabilitation Krapinske Toplice, Krapinske monoclonal gammopathies (MG); however, the sensitivity of these Toplice, Croatia techniques may not be sufficient when MC levels are low. Since 2001, main clinical practice guidelines recommend quanti- Background-aim fication of serum free light chains (FLCs) due to its higher analytical sensitivity as well as diagnostic and prognostic value, while Quality Indicators are measurable, objective, numerical indicators maintaining the recommendations regarding BJP studies. of the effectiveness of key segments of a system. In this paper we have presented the first results of monitoring the turnaround time (TAT), Aims the number of hemolyzed samples, incorrect laboratory reports and unacceptable performances in EQA schemes per year which are good To assess the concordance between results derived from serum examples of key process Quality Indicators in the laboratory. FLCs quantification and urine BJP studies.

Methods Methods

The data available from the Laboratory Information System were Two years-long retrospective study in which 499 patients with used for the period from 01.01.2018. until 31.12.2018. Blood Cell clinical suspicion of MG were included. For the entire cohort, urine Count, Prothrombin time (PT), Potassium, Glucose and hsTnI are BJP studies and serum FLCs values [Kappa, Lambda and Kappa/ selected emergency tests for monitoring TAT. TAT is expressed as the Lambda ratio (rFLC)] were collected. number of urgent analyzes performed over 60 min. in relation to the Urine proteinograms and immunofixations were analyzed by total number of urgent analyzes, then converted into the number of capillary electrophoresis (Capillarys, Sebia), and agarose gel electro- defects per million occasions and value Six Sigma using statistical phoresis (Hidrasys, Sebia), respectively. FLCs were quantified by calculators as well as the number of all samples with visible hemolysis nephelometry (Freelite®, Binding Site). and incorrect laboratory reports compared to the total number of Statistical analysis was carried out with MedCalc software (v. samples and reports. Pre-defined Eligibility Criteria were Sigma N 3.0. 13.0). Chi-squared test was applied to evaluate the concordance between BJP and FLCs results, both rFLC alone and the combined Results evaluation (CE) of serum FLCs (Kappa, Lambda and rFLC), and p b 0,05 was considered as statistically significant. The number of urgent analyzes performed over 60 min. in relation to the total number of urgent analyzes and the value of the Results Sigma were: for Blood Cell Count 29/2444, Sigma 3,8; for PT 328/ 7364, Sigma 3,3; for Potassium 421/8726, Sigma 3,2; for Glucose 194/ Comparision of BJP results with rFLC showed that rFLC was 5098, Sigma 3,3; for hsTnI 60/2735, Sigma 3,6. The number of abnormal in 24,2% patients, while no monoclonal band (MB) was hemolyzed samples and incorrect laboratory reports in relation to identified in urine, probably due to the lack of sensitivity described the total number of samples and reports and the value of the Sigma for BJP. On the other hand, for 7,2% patients a urine MB was were: 367/60730, Sigma 4.1 and 308/47588, Sigma 4.0. The number identified in the presence of a normal rFLC, likely because of the of unacceptable performances in EQA schemes per year/Total polyclonal nature of FLCs. When comparison was made with the CE, number of performances in EQA schemes per year was 5/250, Sigma similar results were observed: 36,4% of patients showed some CE 3,6 (2,0% unacceptable performances). disturbance in the absence of a MB in urine. These differences were statistically significant. Conversely, none of the patients with normal Conclusions CE showed a urine MB.

Data analysis has determined that the Sigma value for selecting Conclusions Quality Indicators pre-analytical, analytical and post-analytical processes meets the predeﬁned criteria. Monitoring of Quality In those cases in which serum FLCs are within the established Indicators is used for self-assessment of a Medical-Biochemistry reference values, BJP analysis in urine does not provide additional Laboratory. information to the diagnosis or follow-up of MG. doi:10.1016/j.cca.2019.03.1100 doi:10.1016/j.cca.2019.03.1101

M348 M349

Contribution of Bence-Jones proteinuria to the monoclonal Unnecessary tests in an emergency department: Analysis of not- gammopathy diagnosis based on serum free light chains seen results

D. Queimaliños Pérez, M. Rodríguez Mata, O. Ortiz Pastor, C. Pérez V. Escribano, I. Casares, N. Rodríguez, M.J. Alcaide, B. Fernández, A. Portugués, B. Gutiérrez Cecchini, Z. Corte Arbolaya, R. Venta Obaya Buño Clinical Biochemistry, University Hospital San Agustín, Avilés, Spain Laboratory Medicine, La Paz University Hospital, Madrid, Spain Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

Background-aim Background-aim

Laboratories have to face to an increasing demand of workload BD Barricor tubes are plasma blood collection tubes with mechan- combined with a reduction of resources. ical separator. We aimed to test if these tubes are superior to BD Implementing strategies to reduce the use of unnecessary and Li-heparin gel tubes and can be used intermittently in the laboratory. redundant tests is an important tool to manage increasing health care costs. Methods We observed that there were laboratory reports that were not visualized by clinicians at the Emergency Department (ED) of a Verification was conducted in Clinical Hospital Centre Rijeka from tertiary level hospital. February till April 2017. The samples were taken from out-patients Aim: evaluate the number of laboratory requests at the ED that who signed informed consent. Tubes were randomized by the order of were not visualized by any clinician. draw. Blood was taken in BD Li-heparin gel tube and BD Barricor tube. Analyses were done 2-4 h after blood collection. Verification protocol Methods was adopted from CLSI GP34-A Validation and verification of tubes for venous and capillary blood specimen collection. All samples were Retrospective observational unicentric study, between May 4 and visually inspected and tested on Roche Cobas 6000 (Roche, Mannheim, June 21, 2018. We use the Computerized Laboratory Information Germany) for: glucose (GLU), urea (BUN), creatinine (CRE), choles- Management System to get the information of total number of terol (CHOL), total proteins (TP), AST, LDH, CRP, sodium (Na), analytics, and a report of non-visualized episode´s number. potassium (K), chloride (Cl) and calcium (Ca). We accessed precision Data was processed by a Microsoft Excel database and statistically (CV), trueness (bias), and comparison of measurements by Passing- treated in SPSS programme. Bablok analysis using MedCalc (MedCalc, Ostend, Belgium). Precision and trueness (bias) were judged according to the biological criteria. Results Results There were 46,287 analytical requests made to emergency laboratory, 10,434 from the ED (Children´s and General Hospital), CV (%) for BD Barricor and BD Li-heparin tubes were: GLU 1.10, 175 not visualized by clinicians (1.7%). 1.16; BUN 2.11, 1.34; CRE 3.39, 2.01; CHOL 1.03, 0.63; TP 0.87, 0.81; AST 52 (29.7%) were registered in the morning shift, 80 (45.7%) in the 3.11, 2.41; LDH 1.48, 1.20; CRP 11.69, 16.39; Na 0.91, 0.46; K 1.35, 0.75; afternoon and 43 (24.6%) in the night. Cl 0.90, 0.63; Ca 1.96, 1.25. Bias (%) BD Barricor from BD Li-heparin The age distribution was b1 year 0.6%, 1-15 years 10.8%, 16-65 tubes was: GLU 0.94, BUN 0.34, CRE 1.43, CHOL 0.68, TP 0.26, AST 0.81, 65.2%, N65 23.4%; and the analytical priority normal 94.9%, urgent LDH 2.25, CRP 0.7, Na 0.14, K 0.23, Cl 0.02, Ca 0.74. Correlation 1.1% and vital 4%. coefficients for all tests were N0.870 except for Na 0.617 which is 58.3% of requests had hemogram, 56.0% biochemical tests, 57.1% unsatisfactory. Passing-Bablok analysis revealed constant difference coagulation tests, 41.7% urinalysis, 4.6% troponin I and 1.7% D dimer. for GLU (Y=-0.05(-0.05-(-0.05))+1.0(1.0-1.0)X), CRE (Y=-1.14(- Patient´s destination after discharge from emergency service was 2.25-(-1.0))+1.0 (1.0-1.01)X) and Ca (Y=-0.11(-0.32-(-0.02))+1.04 home in 67.4%, voluntary discharge 9.2%, transfer to another hospital (1.0-1.13)X). Other methods are comparable in both tested tubes. 1.7%, hospital admission 15.4 %, and death 0.6 %. The potential explanation for non-visualizing the reports were: Conclusions duplicated requests (more than 1 request for the same patient) 14.9%, referral to another specialist doctor 22.9%, voluntary discharge All tubes showed good quality based on visual inspection. CV and 9.2%, visualization of wrong report 4.6%, results not used for patient Bias of BD Barricor tubes are satisfactory. We used broader criteria diagnosis 47.5%, death 0.6%. for CRE, Na, Cl and Ca. Electrolyte comparison mismatch occurred because of the limitation of the analytical performance. BD Barricor Conclusions tubes are not superior when compared to BD Li-heparin gel tubes for tested parameters. The not visualized reports mean performing an invasive technique that it will not contribute to benefit the patient nor will be doi:10.1016/j.cca.2019.03.1103 used in a clinical decision, that may impact on patient care. It is necessary to monitor this performance indicator and communicate/ discuss these results with the ED to optimize laboratory tests. M351 doi:10.1016/j.cca.2019.03.1102 Use of within- and between-subject biological variation data for analytical performance specification in external quality assurance scheme: A retrospective study M350 A. Terreni, G. Avveduto, P. Pezzati, Q. Massimo Verification of BD Vacutainer® Barricor™ plasma blood collection SOD Sicurezza e Qualità, CRRVEQ, Dipartimento dei Servizi, AOU tube Careggi, Florence, Italy

V. Supak Smolcica,b, L. Ongaroa, L. Bilic-Zullea,b Background-aim aClinical Department for Laboratory Diagnostics, Rijeka Clinical Hospital Center, Rijeka, Croatia External Quality Assurance (EQA) schemes providers evaluate bDepartment for Medical Informatics, Rijeka University School of laboratory performances based on analytical performance specifica- Medicine, Rijeka, Croatia tion (APS). Presently, Providers establish and applies APS using Abstracts / Clinica Chimica Acta 493 (2019) S497–S532 different criteria. The 1st EFLM Strategic Conference held in Milan in Methods 2014, defined three models to be used to derive APS: based on the effect of analytical performance on the clinical outcome (M1); based Home-made materials (12 lyophilized human sera prepared out on components of biological variation (BV) of the measurand (M2); from different pools) were sent to participant laboratories. Target based on the state of art of the measurement (M3). Plasmatic Na, K, values were assigned as consensus mean for each peer group and Cl, and Total Protein were assigned to M2. between-laboratory CV% was calculated; acceptance limits were stated as +3SD. Performance statistics for individual laboratories Methods were calculated through annual cumulative results: performance index PI was computed as meanCV% and meanBIAS%; PIs were Since for those measurands, rigorously determined BV data have ranked and calculation of percentiles p25, p50, p70 and p90 defined been established by the EFLM Biological Variation Study (EuBIVAS), bands from very good to unacceptable (A B C D E). Labs that we calculated Total Analytical Error % (TAE), as following: TAE completed the survey for PI calculation were: FT3 35; FT4 184; T3 optimal 1.65 (0.25CVi)+0.125 (CVi2+CVg2)1/2; TAE desirable 1.65 173; T4 175; TSH 191; FSH 159; LH 156; PRL 162; hCG 159; E2 148; (0.50CVi)+0.25 (CVi2+CVg2)1/2; TAE minimum 1.65(0.75CVi) P4 128; To 122; Csol 135, Insulin 155. +0.375 (CVi2+CVg2)1/2 Results Results We calculated % labs in different performance groups (A+B We obtained, respectively: Na 0.38, 0.77, 0.15 ; K 2.32, 4.65, 6.97; acceptable, C regular, D+E bad) in each method. Acceptable Cl 0.61,1.22,1.84 and Total Protein: 1.73, 3.47,5.20. The use of Ricos et performance for major methods was: COBAS ROCHE: TSH 62% n = al. database on BV, produced the following TAE: Na 0.36, 0.73, 1.09; K 85, FT4 66% n = 84, LH 65% n = 76, FSH 66% n = 75, PRL 46% n = 2.80,5.61,8.41; Cl 0.74, 1.47,2.21; Total Protein 1.82,3.63,5.45. When 75, hCG 63% n = 72, E2 74% n = 76, To 75% n = 66, P4 41% n = 62, TEA derived from EuBIVAS were used to evaluate the 2018 Centro di CSOL 61% n = 61; ABBOTT ARCHITECT: TSH 41% n = 36, FT4 21% n Riferimento Regionale Verifica Esterna Qualità (CRRVEQ ) Clinical = 34, LH 21% n = 28, FSH 86% n = 28, PRL 87% n = 31, hCG 77% n Chemistry EQA scheme results, we showed the following percent- = 31, E2 43% n = 30, To 38% n = 21, P4 96% n = 25, CSOL 46% n = ages of laboratory meeting respectively optimal, desirable and 24; ACCESS BECKMAN: TSH 54% n = 11, FT4 45% n = 11, LH 29 % n minimum goal: Na 20%, 38%, 50%; K 70%,92%, 95%; Cl 22%, 45%, = 7, FSH 17% n = 6, PRL 50% n = 6, hCG 14% n = 7, E2 0 % n = 9, To 60%; plasma protein 50%, 80%, 91%. When TEA derived from Ricos et 13% n = 8, P4 0% n = 9, CSOL 72% n = 7; CENTAUR SIEMENS: TSH al. were used, we showed : Na 18%, 33%, 47%; K 78%, 95%, 100%; Cl 50% n=35, FT4 50% n=20, LH 42% n=17, FSH 17% n=6, PRL 81% 26%, 53%, 68%; Total protein 54%, 80%, 92%. The CRRVEQ, similarly to n=16, hCG 0% n=12, E2 38% n = 8, To 0% n = 8, P4 71% n = 7, CSOL several international providers, adopts ASP based on a combination 25% n = 12; IMMULITE SIEMENS TSH 29% n = 35, FT4 45% n = 33, of BV and state of the art, and, as expected, the percentage LH 40% n = 25, FSH 0% n = 28, PRL 6% n = 30, hCG 22% n = 32, E2 of laboratories evaluated as having positive performances was 19% n = 19, To 0% n = 11, P4 43% n = 19, CSOL 35% n = 15. higher: Na 88% ; K 92% and 95% ( K b3.0 mmol/L and N3.0 mmol/L respectively); Cl 92% and 85% ( b90 mmol/L and N90 mmol/L Conclusions respectively); Total Protein 92% and 89% (b5.0 gr/dL and N5.0 gr/dL respectively). Differences in acceptable performance and mean CV% were found for all hormones, reflecting different laboratory implementation and Conclusions robustness of platforms. EQAS is a useful tool for follow up of method performance, as results reflect the state of the art in hormone assays. We can affirm that, mainly for Na and Cl, the APS based on BV Traceability as claimed by manufacturers did not lead to acceptable may represent a difficult goal to achieve for the majority of harmonization in some of the hormones, as reflected in differences in laboratories. mean BIAS%. doi:10.1016/j.cca.2019.03.1104 doi:10.1016/j.cca.2019.03.1105

M352 M353

Performance of hormone assays in EQAS “Buenos Aires” ProgBa – Determination of sigma score based on biological variation for Cemic haemostasis assays: Fit-for-purpose for daily practice?

C.A. Fenili, L. Del Vecchio, M. Porta, S. Quiroga, M. Torres M.J. Van Essen-Hollestellec, J. Ruinemans-Koertsa, R.N. Idemad,P. Programa Internacional de Aseguramiento Externo de la Calidad Meijerc, M.P. De Maatb (ProgBA), Hospital Universitario CEMIC, Buenos Aires, Argentina aDepartment of Clinical Chemistry and Hematology, Rijnstate Hospital, Arnhem, the Netherlands Background-aim bDepartment of Hematology, Erasmus Medical Center, Rotterdam, the Netherlands EQAS “Buenos Aires” ProgBA – CEMIC was established in 1979 cECAT Foundation (External quality Control for Assays and Tests), and got its accreditation under ISO/IEC 17043:2010 in 2011. We Voorschoten, the Netherlands present results from 2017-2018 to evaluate method performance of dLaboratory for Clinical Chemistry and Hematology, Amphia Hospital, several immunoassays of routine hormones. Breda, Noord Brabant, the Netherlands Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

Background-aim Even relatively small analytical variations may indeed influence the proportion of patients who could be identified as suitable for Internal quality control (QC) rules for laboratory tests can be discharge. To monitor baseline drifts and calibration accuracy at derived from analytical performance specifications (APS) using the very low hsTn concentrations, in March 2017 we introduced a serum six-sigma method. We tested the applicability of this paradigm to pool with a concentration between LOB and LOD as an additional routine haemostasis measurements. internal quality control material (IQC3). Here we show the impact of this additional quality tool on EQAS performance. Methods Methods Three laboratories using different instruments and reagents calculated sigma scores for their prothrombin time (PT), activated In our laboratory, we measure hsTnT on two interchangeable partial thromboplastin time (APTT), fibrinogen and antithrombin Roche Cobas e411 platforms (LOB and LOD: 3 and 5 ng/L, (AT) measurements. Sigma scores were calculated using biological respectively) and participate to the UK-NEQAS, which includes a variation (BV) data from the literature in combination with internal low concentration sample (LCS) in each monthly exercise. We and external QC data. Internal data were derived from one quality evaluate EQAS results according to an allowable total error (TEa) of sample in the normal range collected in three consecutive months ±22.5% (biological variability derived) between our result and the and external data were based on 6 external quality surveys with 2-3 mean of participants using the same measuring system. The IQC3 is data points per survey. prepared from fresh leftover human sera with hsTnT concentrations between 3 and 5 ng/L and stored at –20 °C in 250-μL aliquots. IQC3 Results target value and acceptability range are preliminarily determined by calculating mean ± 30% of 10 measurements performed in optimal Wide ranges in sigma scores for the PT (0.1-6.8), APTT (0.0-4.3), conditions. The IQC3 is then assayed twice daily and after every new “ ” fibrinogen (1.5-8.3) and AT (0.1-2.4) were observed when QC data calibration. If results are out of control , immediate corrective was combined with the minimum, median and maximum value of actions are implemented before reports related to the samples BV data, due in particular to a large variation in within-subject and analysed in the affected run are issued. between-subjects coefficients of variation. When the median BV values were applied, most sigma scores were below 3.0, for internal Results QC data; 75% and for external QC data; 92%. Before the IQC3 introduction, we measured hsTnT on LCS from 26 Conclusions EQAS exercises, with 11 results (42.3%) not meeting TEa. After the IQC3 introduction, only one out of 21 exercises (4.8%) did not meet Our findings demonstrate that: 1) The sigma scores for common TEa (P = 0.009 between the two periods). Results for the failed haemostasis parameters are relatively low and 2) The application of exercise were 9.1 vs. 7.2 ng/L (TE +26.4%). the six-sigma method to BV-derived APS is hampered by the large variation in published BV data. An updated database is needed, in Conclusions which only BV studies are included which fulfil standardised criteria. Since the six-sigma concept is based on requirements for monitoring, Implementing an IQC at hsTn concentrations between LOB and and many haemostasis tests are only designed for diagnostic LOD is vital for assuring the suitable accuracy at such low, but purposes, a fit-for-purpose APS is needed to achieve clinically clinically relevant concentrations. relevant quality goals. doi:10.1016/j.cca.2019.03.1107 doi:10.1016/j.cca.2019.03.1106

M355 M354 Developing pre- and post-analytical error monitoring in labora- Daily monitoring of a control material with a concentration tory medicine between LOB and LOD improves the accuracy of highly sensitive troponin assay B. De La Salleb, R. Marringtona, F. Mackenziea aBirmingham Quality (UK NEQAS), United Kingdom b E. Aloisiob, S. Pasqualettia, A. Dolcia, M. Panteghinib West Herts Hospitals NHS Trust Operating UK NEQAS Haematology and aClinical Pathology Unit, ASST Fatebenefratelli-Sacco, Milan, Italy Transfusion, United Kingdom bSpecialization School in Clinical Pathology and Clinical Biochemistry, University of Milan, Clinical Pathology Unit, ASST Fatebenefratelli-Sacco, Background-aim Milan, Italy There have been many initiatives to collect and collate Background-aim benchmarking data on error rates in the Pre and Post-Analytical aspects of Laboratory Medicine. We report here on the ﬁndings from Rapid and safe rule out of acute myocardial infarction in patients the UK where there has been an established pre and post-analytical admitted to emergency department (negative predictive value N99%) quality monitoring service (PREPQ) offered by UKNEQAS since 2017, can be achieved by setting limit of blank (LOB) or limit of detection following the Pilot phase from 2014 to 2016. The service is (LOD) of highly sensitive cardiac troponin assays (hsTn) as decision directed by a multi-disciplinary steering group of advisors that thresholds. Accurate calibration of hsTn in the low concentration includes international experts in the ﬁeld of pre and post-analytical range is therefore of the upmost importance for this application. variables. Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

Methods Methods

Participating Laboratories submit their error rates for a range of A set of six control materials, concentrations (56,4 to 278,2 U/L), parameters either by Discipline, (e.g. Haematology) or by the whole prepared by Stichting Kwaliteitsbewaking Medische Laboratorium service (e.g. Blood Sciences). The time periods are Calendar months. Diagnostiek (SKML, Holland) were distributed in a single shipment To avoid different counting practices Participants may classify their to 214 Spanish laboratories who stored it at -20°C until analysis. error rates by Request or Specimen. During 6 consecutive days, a control vial was measured by duplicate Participants are regularly offered clinical scenarios presenting in a single analytical run. pre- or post-analytical errors, which they interpret based on their local practice. An analysis of the responses is shared as a Results commentary to improve knowledge and drive harmonisation. 3254 results of ALP were grouped according to the combination Results of measurement procedure-traceability-instrument. Bias of each group mean (percentage deviation to the reference value) was The data is normalised into the currency of Sigma Scores, in compared with the desirable specification derived from biological addition to the raw data, since the size of Laboratories varies by over variation (6.7%). Intra-laboratory coefficients of variation were several orders of magnitude. Trend data is presented in a standard calculated and compared with the desirable specification for UKNEQAS style familiar to participants. Data has been returned from imprecision derived. a maximum of 32 Laboratories for eleven quality indicators. In 190 out of 214 laboratories used the IFCC recommended method November 2018, for the key indicators of Sample time/temperature (4-Nitrophenyl phosphate substrate with AMP buffer), traceable to critical failures, the median Sigma Score is 3.85 but the spread is IFCC method; 17 laboratories used same substrate with DEA buffer, 3 from 2.22 to 4.63. For Patient ID failures the data is a healthier 4.83, laboratories used dry chemistry and 4 did not inform about with a range of 3.3 to 5.64. Some Laboratories can be at opposite traceability. From the major group, Beckman AU, Bio-Systems BA ends of the Sigma Score spectrum for different indicators. and Siemens Dimension/Vista gave standardize results. Abbott Architect and Siemens Advia obtained results slightly below the Conclusions acceptable limit (deviations of -7.3% and -7.8%, respectively). Roche- Cobas produced the lowest results (-11%). In opposition, 16 Despite having detailed guidance, there remains confusion as to laboratories using Siemens Advia with the DEA buffer method, what data should be collected and how it could be collected in a showed highly negative deviant results (-108%). standardised fashion. The challenge going forward is to review the quality indicators in use against the published literature, to work Conclusions towards harmonisation of indicators within Europe and to have these built into the specifications for laboratory information The 4-Nitrophenyl phosphate substrate with DEA buffer was management systems. endorsed to be abandoned. Further combined work between labs and providers is encour- doi:10.1016/j.cca.2019.03.1108 aged, to reduce the discrepancies evidenced in this study for some methods groups.

doi:10.1016/j.cca.2019.03.1109 M356

Standardization assessment of alkaline phosphatase measurements in a category 1 external quality assurance program M357

E. González Lao, Z. Corte Arboleya, J. Díaz Garzón, F. Marqués García, Analytical performance specifications based on the state-of-the- P. Fernández Fernández, P. Fernández Calle, C. Ricós Aguilá, B. Boned art for the magnitudes included in the Spanish newborn Juliani, M. Simón Palmada, J.V. García Lario, C. Perich Alsina, J. screening program Minchinela Girona, X. Tejedor Ganduxe Spanish Society of Laboratory Medicine (SEQCML), Analytical Quality L. Guiñonc, A. Solerc, J.L. Marina, A. Molinaa, R.M. Lopeza, J. Garciaa,W. Commission, Spain Jimeneza, A. Mirab, L. Alvarezc aBiochemistry and Molecular Genetics, Biomedical Diagnostic Center, -Background-aim Hospital Clinic of Barcelona, Spain bDirection, Biomedical Diagnostic Center, Hospital Clinic of Barcelona, Introduction. Since 2015, the Analytical Quality Commission of Spain c Spanish Society of Laboratory Medicine (SEQCML) introduced a Quality Department, Biomedical Diagnostic Center, Hospital Clinic of category 1 external quality assurance program (“SCR”), running once Barcelona, Spain per year, that used: Background-aim - Commutable materials (fresh frozen human serum). - Reference method target values covering a wide measurable When setting analytical performance specifications, models based range. on the effect on the clinical outcome or on biological variation are - Replicated analysis to assess the analytical imprecision. preferred. However, for some magnitudes as the ones included in the newborn screening, there is no data available based on these two Objective. To assess whether the different method groups for models. In these cases, quality specifications based on the state-of- serum alkaline phosphatase (ALP), participating in the SCR 2015 to the-art are useful to keep the analytical error under control. 2017 programs, perform in a standardized way. Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

The objective of this work is to obtain quality speciﬁcations for measured on a single instrument. Therefore, evidence and guidance the total error (TE) for the magnitudes included in the External is required for routine operation and workﬂow in a centralised Quality Assessment Program (EQA) of newborn screening organized laboratory, where serial troponin samples from the same patient by the Spanish Association of Newborn Screening (AECNE). might be measured on two different instruments of the same type.

Methods Methods

A total of 23,778 results from 19 laboratories were collected from the In this sub-analysis of the multicentre TRAPID-AMI study, we Spanish EQA in newborn screening during the period between May evaluated the performance of the accelerated 0/1-h AMI algorithm 2015 and September 2018. Magnitudes included were thyroid-stimu- when measuring serial troponin samples from the same patient with lating hormone (TSH), phenylalanine (Phe), tyrosine (Tyr), immunore- symptoms suggestive of AMI on two parallel instruments. Patients active trypsinogen (IRT), free carnitine (C0), acetylcarnitine (C2), eligible for inclusion provided informed consent for remeasurements propionylcarnitine (C3), butyrylcarnitine (C4), isovalerylcarnitine (C5), and 707 samples were available. The 0-h and 1-h samples were glutarylcarnitine (C5DC), hexanoylcarnitine (C6), octanoylcarnitine (C8), measured with the Elecsys® Troponin T-high sensitive (cTnT-hs) decanoylcarnitine (C10), myristoylcarnitine (C14), palmitoylcarnitine assay on two different cobas 8000 analysers (Roche Diagnostics). (C16) and stearoylcarnitine (C18). AMI diagnosis was determined by 1-h criteria for rule-out (cTnT-hs For each result, TE in percentage was calculated by comparing the b12 ng/L and change b3 ng/L at 1 h) and 0/1-h criteria for rule-in value reported by the participant with the target value (robust (cTnT-hs N52 ng/L or change N5 ng/L at 1 h); remaining individuals mean). Quality speciﬁcations were calculated as the 90th percentile, were classiﬁed to the observation zone. Outcomes were analysed considering only the 75% of each laboratory best results of TE. for different data combinations, e.g. all samples measured, or 0-h and 1-h samples randomly assigned, on the two different Results instruments.

The magnitudes that had the best analytical performance were Results TSH, IRT, C16 and C18, with a TE lower than 20%. For a greater number of magnitudes such as Tyr, C0, C2, C3, C4, C5, C6, C8 and C14, When running all samples from the same patient on two different the TE committed were between 20 and 35%. Phe, C5DC and C10 instruments of the same type, results were in agreement for 691/707 showed the worst analytical performance, with a TE higher than 35%. (97.7%) samples analysed: 354 rule-out, 88 rule-in and 249 observation zone. Results differed (instrument 1/instrument 2) for Conclusions 16 (2.3%) samples: 1 observation/rule-out; 1 rule-in/observation and 14 rule-out/observation. No reclassifications from rule-in to rule-out TE specifications based on the state-of-the-art for the magnitudes were observed and only 1 reclassification from rule-in to observation of the Spanish EQA in newborn screening have been established. (0.1%). The potential variation introduced by the parallel instrument These data can help laboratories to establish quality specifications for setup is thus small compared with variations introduced by an these magnitudes and control their analytical performance. instrument switch when using traditional diagnostic protocols, e.g. a single cutoff. doi:10.1016/j.cca.2019.03.1110 Conclusions

The 0/1-h algorithm appears to be safe and effective for triaging M358 patients with suspected AMI when measured on two different instruments of the same type. Robustness of the troponin 0/1-h algorithm for early diagnosis of acute myocardial infarction when measured on two different doi:10.1016/j.cca.2019.03.1111 instruments of the same type

A. Haushofera, J. Seiera, M. Stüblera, G. Lirkf, A. Zieglere, E. Giannitsisc, B. Lindahld, C. Muellerb M359 aCentral Laboratory, Klinikum Wels-Grieskirchen, Wels, Austria bDepartment of Cardiology, Cardiovascular Research Institute Basel, Accuracy evaluation of ﬁve analytical systems for high- and low- University Hospital Basel, Basel, Switzerland density lipoprotein cholesterol assays in Korea cDepartment of Internal Medicine III, Cardiology, Angiology & Pulmonology, Heidelberg University Hospital, Heidelberg, Germany S. Kimc, H. Kimd, Y. Yund, W. Mina, J. Kimf, G.C. Kwonb, J. Songe,C. dDepartment of Medical Sciences, Uppsala Clinical Research Center, Chog Uppsala University, Uppsala, Sweden aDepartment of Laboratory Medicine, Asan Medical Center, Seoul, South eRoche Diagnostics International Ltd, Rotkreuz, Switzerland Korea fUniversity of Applied Sciences Upper Austria, Hagenberg im Mühlkreis, bDepartment of Laboratory Medicine, Chungnam National University Austria Hospital, Seoul, South Korea cDepartment of Laboratory Medicine, Green Cross Laboratories, Background-aim Gyeonggi-do, South Korea dDepartment of Laboratory Medicine, Konkuk University School of Accelerated protocols for acute myocardial infarction (AMI) Medicine, Seoul, South Korea diagnosis, based on relatively low troponin concentrations and small eDepartment of Laboratory Medicine, Seoul National University acute changes of 3–5 ng/L over 1 h in blood, have been validated in Bundang Hospital, Seoul, South Korea several research studies. However, blood samples were typically fDepartment of Laboratory Medicine, Yonsei University College of Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

Medicine, Seoul, South Korea bDepartment of Immunology, CHU Brest, Brest F-29200, France gKorea Centers for Disease Control & Prevention, Chungcheonbuk-do, cDepartment of Immunology, CHU Lille, Lille F-59000, France South Korea dDepartment of Toxicology, CHU Lille, Lille F-59000, France eDepartment of Immunology, Hospital Cochin, Assistance Publique- Background-aim Hôpitaux de Paris (AP-HP), Paris, France

Accurate and precise measurement of blood cholesterol, including Background-aim high-density lipoprotein cholesterol (HDL-C) and low-density lipoprotein cholesterol (LDL-C), is essential for generating the correct Although the use of EDTA-containing collection tubes is known to burden and trends of dyslipidemia. We evaluated the performance of stabilize the complement analytes and to make the results more five HDL-C/LDL-C cholesterol assays currently used at clinical reliable, no external quality assessment (EQA) scheme based on laboratory in Korea to assess the traceability of current HDL-C/LDL- EDTA plasma samples is available to date in France. Consequently, a C in vitro diagnostic products to commutable frozen serum (CFS) number of clinical laboratories currently participate to EQA pro- reference materials. gramme on samples whose matrix is different from their routine practice. The aim of this work was to offer a new external quality Methods assessment scheme, as an inter-laboratory exchange (ILE). The different steps of the validation procedure are presented, together The HDL-C/LDL-C assays were categorized as five groups with the one-year ILE implementation experience. according to the combination of instrument and reagent; Toshiba- Kyowa, Hitachi-Sekisui, Siemens ADVIA, Roche Cobas, and Beckman Methods Coulter AU. The 12 levels of CFS pools were prepared according to CLSI 37-A and sent at frozen state. The design based on two days, We used the complement Optilite® assays for classical pathway with one run per day and three replicates per run. Target reference activity, C3c, C4 and C1 inhibitor protein (The Binding Site Group Ltd, values were measured using reference ultracentrifugation method Birmingham, UK). Stability studies were performed using the for HDL-C and ®-quantification for LDL-C at CEQAL. Target acceptable change limit (ACL), defined as the analytical coefficient performance goal based on the analytical performance criteria for of variation (CVa, for each analyte) * 2,77 as the main criteria. The tests used to assess cardiovascular disease risk of the Cholesterol effect of 2 commercial diluents, used to adjust the samples’ Reference Method Laboratory Network as follows: maximal allow- concentration was studied through the percent of recovery (%R), able bias δ ±5% for HDL-C and δ ±4% for LDL-C; maximal allowable defined as the proportion of corrected to nondiluted analyte value imprecision δ 4.0% at ε 42 mg/dL or δ 1.7 at b 42 mg/dL for HDL-C (acceptable variation: 80-120%). and δ ±4% for LDL-C, In addition we observed total error according to the National Cholesterol Education Program; δ 13% for HDL-C Results and δ 12% for LDL-C. Validation procedure: (i) whole blood EDTA samples were stable Results at 4°C up to 72h; (ii) when separated within the 12h of sampling, EDTA plasmas were stable until the 72e hour at 4°C and -20°C but The target values of five materials were 37.1 – 57.2 mg/dL for not at room temperature; (iii) no matrix effect was evidenced for the HDL-C and 76.9 – 130.9 mg/dL for LDL-C. Bias ranged from -0.9% to 2 tested commercial diluents; (iv) the diluted samples were stable at 6.2% for HDL-C and from −3.3% to 8.5% for LDL-C. Imprecision -20°C until the 4th week. The ILE program was started on January showed total CVs b 2.2% for HDL-C and b 1.3% for LDL-C. Total error 2018. A total of 6 sendings of 2 samples was made during the year ranged from 3.4% to 6.7% for HDL-C and from 6.2% to 9.2% for LDL-C. 2018. From an initial number of 3 participants, we moved to 4 at the Except Toshiba-Kyowa, the other HDL-C assays met target perfor- 5th sending. Each participating laboratory received a personalized mance goal. Whereas, all five LDL-C assays did not meet target report after each response, together with an annual report of performance goal. performance. Due to the limited number of participants, both classic and robust z-score were included in the reports. The shipping was Conclusions systematically controlled using additional samples, and no CV exceeded -10 to +10% when measured back to the laboratory in Even routine assays have advanced considerably over recent charge. years, manufacturers’ still have to strive for accurate HDL-C/LDL-C measurements for patients with standardization and calibration Conclusions verification. The newly implemented ILE will be useful for the accreditation of doi:10.1016/j.cca.2019.03.1112 the complement activity of French laboratories using EDTA plasma samples.

doi:10.1016/j.cca.2019.03.1113 M360

Preanalytical considerations, stability study and external quality assessment scheme implementation for complement components M361 dosage on EDTA plasma UK NEQAS for serum indices: Three years on …. B. Lopezc, S. Rogeauc, A. Deleplancquec, E. Moitrotc, E. Berthea,C. Goulvestree, S. Hillionb, E. Vinnerd, M. Labalettec, S. Dubucquoic R. Marrington, J. French, A. Robins, F. Mackenzie aDepartment of Immunology, CHU Amiens, Amiens F-80000, France Birmingham Quality (UK NEQAS), United Kingdom Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

Background-aim Background-aim

The presence of Haemoglobin (Haemolysis), Bilirubin (Icterus) We sought to define the performance of automated chemistry and Lipids (Lipaemia) in serum can affect the measurement and platforms at two large academic medical centers by calculating and reporting of many Clinical Chemistry laboratory test results. Quality comparing sigma metrics for 28 analytes. Assurance of Serum Indices is not as rigorous as for other Clinical Chemistry assays. Three years ago Birmingham Quality established Methods an EQA scheme for Indices in Serum. This poster explores how the scheme has developed and where we are now. Performance characteristics of chemistry assays on two Roche Cobas analyzers (University of Florida Health Jacksonville) and four Methods Abbott Architect analyzers (Vanderbilt University Medical Center) were estimated using 12 months of Bio-Rad quality control (QC) data The UK NEQAS for Serum Indices (HIL) EQA scheme that not only at two concentrations. Method imprecision was calculated as the looks at Haemolysis (H), Icterus (I) and Lipaemia (L) as individual cumulative QC coefficient of variation (CV) and percent bias was analytes but also looks at the impact these serum indices have on calculated by comparison of analyzer mean to peer group means. particular specific analytes, which change from month to month, Sigma values were calculated for each method as [(TEa – Bias%)/CV%] called ‘Analyte X’. Three specimens are distributed monthly and using allowable total error (TEa) from two sources: the CLIA laboratories return their H, I, L and ‘Analyte X’ results, and whether evaluation limits and desirable biological variation (Ricos C et al.). they would report the measured analyte based on the serum indices. Average sigma values were generated for each site and graded as optimal N6 sigma; good 5–6 sigma; marginal 3–5 sigma; or poor b3 Results sigma. Analysis of NIST SRM1950 standards for a subset of analytes allowed an estimation of absolute bias. To accommodate the results reported, Birmingham Quality has two data presentations — standard report format for numerical data, Results and pie charts for category data. The results returned so far have shown that there are significant differences in practice for the Sigma metrics were highly comparable across both study sites. measurement and interpretation of serum indices both within-, and Considering CLIA TEa, just over half (UF 57%; VUMC 54%) of the 28 between- manufacturers. For example, an unspiked sample was analytes met the six-sigma standard of performance. Electrolytes (Na, distributed which had endogenous elevated triglycerides. Beckman K, Cl, Mg) and metabolites (total bilirubin, BUN, CO2) failed to meet Synchron and Beckman AU Olympus identified a significant amount six-sigma. Notably, there were dramatic differences in sigma values of icterus present, whereas other major methods hadn’t. Analytically calculated using CLIA and Ricos TEa criteria. Almost 40% of the analytes this is because of secondary interference due to overlapped had at least one QC that performed poorly using Ricos. Only 4 of the 28 absorbance not being corrected. Clinically this means that there is assays (CK, GGT, Lipase and triglycerides) demonstrated optimal the potential that results would not be reported because of an performance at both study sites using Ricos and CLIA criteria. Analysis incorrect icterus results being reported on a lipaemic sample, when of NIST SRM1950 at both study sites gave comparable sigmas for all analytically they may be valid for icterus. analytes except total bilirubin, cholesterol, Mg and total protein.

Conclusions Conclusions

The UK NEQAS for Serum Indices (HIL) has shown that there are Neither Abbott nor Roche analyzers met six-sigma quality significant variations in practice between participants for the use of standards for all analytes tested. CLIA TEa and RICOS TEa are serum indices which consequently affects patient care both in terms significantly different, with wider acceptability criteria for CLIA. of repeat testing and validity of test results. Unfortunately, EQA alone Variations between individual analyzers and manufacturers has not been enough to drive change in laboratory practices as and limitations in automation would make tailored QC rules based observed by repeat distributions over time showing similar percent- on sigma metrics difficult to implement in a high-volume ages of results that would or would not be reported for Analyte X laboratory. based on HIL value. doi:10.1016/j.cca.2019.03.1115 doi:10.1016/j.cca.2019.03.1114

M363 M362 Key factors to achieving International Organization for Stand- Sigma metrics of 28 common chemistry tests for two ardisation (ISO) 22870 accreditation with a broad point-of-care- manufacturers testing (POCT) program

J.H. Nicholsb, M. Browna, J.M. Colbyb, M. Feldhammera M. Millera, D. Patela, R. Hobmanb aCollege of Medicine, Department of Pathology and Laboratory Medi- aBiochemistry Department, Manchester University NHS Foundation cine, University of Florida, Jacksonville, FL, USA Trust, Wythenshawe, United Kingdom bDepartment of Pathology, Microbiology and Immunology, Vanderbilt bTechnical Specialist Department, Werfen UK, Warrington, United University Medical Center, Nashville, TN, USA Kingdom Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

Background-aim aCardiovascular Research Institute Basel (CRIB), Department of Cardi- ology, University Hospital Basel, University of Basel, Basel, Switzerland b When the Biochemistry department on the Wythenshawe site of Roche Diagnostics GmbH, Penzberg, Germany c Manchester University NHS Foundation Trust first achieved ISO Roche Diagnostics International Ltd, Rotkreuz, Switzerland d 15189 laboratory accreditation in 2016, this was a big step towards UC Davis Medical Center, Sacramento, CA, USA e extending the standards to our POCT program. Wythenshawe UT Southwestern Medical Center, Dallas, TX, USA achieved 22870 accreditation in September 2018, making the department only the second NHS trust to be awarded POCT Background-aim accreditation in the UK. Based on interest from peer institutions, we propose a roadmap for achieving ISO 22870, using our blood gas The biotin-streptavidin-based Elecsys® Troponin T-high sensitive testing program as an example. (cTnT-hs) assay (Roche Diagnostics) has a high negative predictive value for ruling out acute myocardial infarction (AMI), but biotin N20 Methods μg/L can reduce recovery by N10%. We assessed the risk of patient misclassification due to biotin interference. Wythenshawe is a 900-bed teaching hospital providing acute care services to adult and paediatric patients. POCT service includes more Methods than 300 devices—29 blood gas analysers, 180 hand-held glucose, ketone, tHb and chemistry devices— equating over 2million individ- Biotin was measured in two cohorts using an Elecsys® biotin ual patient tests each year. assay. The acute coronary syndrome (ACS) cohort comprised 797 Achieving ISO 22870 involved the efforts of the POCT team, initial (0-hr) and 646 3-hr blood samples from 850 patients with clinical teams, learning and development, the main laboratory and suspected AMI in the US. The US laboratory cohort comprised 2023 our supplier partners to establish: e-learning, a Quality Management random samples from a US laboratory network; biotin concentra- System (QMS), define Key Performance Indicators (KPI), and audit tions were extrapolated for higher values using pharmacokinetic for improvement opportunities. data to simulate future use of high-dose biotin for multiple sclerosis. The blood gas testing service at Wythenshawe includes 29 GEM® Prevalence of biotin N20 μg/L and 99th percentile biotin were Premier™ 5000 with iQM2® (Instrumentation Laboratory) analysers calculated, and the impact of elevated biotin on cTnT-hs was interfaced into GEMweb® Plus which is a key element to the POCT modelled in both cohorts. In the US laboratory cohort, the program. misclassification risk was determined using global (excluding US) 14 ng/L and US 19 ng/L cTnT-hs cutoffs, and for a biotin washout Results time of 3 hrs based on pharmacokinetic data.

The efforts performed for the blood gas testing can be used as a Results model for other institutions to achieve ISO 22870: ACS cohort: one (0.13%; 30.23 μg/L) initial and one (0.15%; 24.48 1. Build e-learning program – standardised analyser platform with μg/L) 3-hr sample had biotin N20 μg/L; 99th percentile biotin was Operator Competency modules in GEMweb Plus 2.62 μg/L (initial) and 2.38 μg/L (3-hr), N7 times lower than the assay 2. Establish standardised documentation within the QMS system – interference threshold. US laboratory cohort: 15 (0.74%) samples had Maintenance-free analysers simplify staff-time, documentation biotin N20 μg/L; 99th percentile biotin was 16.62 μg/L. Using and elevate quality conservative assumptions in the ACS cohort (including tripling the 3. Identify KPIs for POCT dashboard – built-in KPIs and iQM2 risk- highest observed biotin concentration per CLSI EP07 guidelines), management features facilitate monitoring for continuous biotin interference could lead to a falsely low value for an initial improvement cTnT-hs result between 19 and 45.24 ng/L; the likelihood of false- 4. Set up monitoring process – sample handling reports in GEMweb negative AMI prediction was 0.026% at 0 hrs. Using extrapolated Plus enable monitoring for operator training biotin data from the US laboratory cohort, the misclassiﬁcation risk due to biotin interference using 14 ng/L and 19 ng/L cTnT-hs cutoffs, Conclusions respectively, was: 0.025% and 0.026% at 0 hrs; 0.00049% and 0.00048% at 3 hrs. Seeking accreditation aligns with the Wythenshawe objectives to grow our clinical expertise and expand our research programs. Being Conclusions awarded ISO 22870 validated the quality of our comprehensive POCT service. The built-in features of the GEM Premier 5000 with iQM2 In our study, biotin interference had a minimal impact on cTnT-hs and GEMweb Plus not only facilitated ISO quality requirements diagnostic performance and the risk of false-negative AMI prediction automatically, but also helped free up POCT staff to focus on the due to biotin was low. broader framework of the overall accreditation program. doi:10.1016/j.cca.2019.03.1117 doi:10.1016/j.cca.2019.03.1116

M365 M364 Real-time monitoring of drug-laboratory test interactions with Evaluating the clinical risk of biotin interference with the an automated decision support application Elecsys® Troponin T-high sensitive assay

J. Van Balvereng,h, W. Verboeket- Van De Venned, L. Erdem-Eraslana, N. Trand, D. Dierckse, R. Twerenbolda, A. Zieglerc,A. A. De Graafc, R. Mussoni, W. Oosterhuisd, I. Van Der Sijse, R. Verheulb, Schuetzenmeisterb, D. Kasapicc, B. Mummad Abstracts / Clinica Chimica Acta 493 (2019) S497–S532

R. Kustersg,h, R. Hoedemakersf application (Gaston, Medecs B.V.). The DLTIs were described in a aDepartment of Clinical Chemistry, Erasmus University Medical Centre, validated database from the Dutch Society for Clinical Chemistry. The Rotterdam, the Netherlands application is able to generate a DLTI-based advisory text based on bDepartment of Clinical Chemistry, LabWest/HMC Westeinde, The predeﬁned aberrant laboratory test results and medication data from Hague, the Netherlands individual patients and present this alert text to the laboratory cDepartment of Clinical Chemistry, Medical Spectrum Twente, Enschede, specialist in the laboratory information system. The software the Netherlands application was successfully connected and installed in one hospital dDepartment of Clinical Chemistry, Zuyderland Medical Centre, Heerlen, laboratory in 2018 with two other hospitals to follow in 2019. the Netherlands Generated real-time DLTI alerts were collected and monitored during eDepartment of Hospital Pharmacy, Erasmus Medical Centre, University 4 weeks. Medical Centre Rotterdam, the Netherlands fLaboratory for Clinical Chemistry and Haematology, Jeroen Bosch Results Hospital, ’s-Hertogenbosch, the Netherlands g Laboratory for Clinical Chemistry and Haematology, Jeroen Bosch A mean of 45 DLTI alerts were generated per day. Twenty-one out Hospital,’s-Hertogenbosch, the Netherlands of 34 clinical rules were generated at least once in this period. The h Department of Health Technology and Services Research, Technical most frequently reported interactions were magnesium - proton Medical Centre, University of Twente, the Netherlands pump inhibitors (14%), creatine kinase – statins (13%) and potassium i Laboratory for Clinical Chemistry and Haematology, University Medical - ACE-inhibitors (13%). Most DLTI alerts were from the internal Centre Utrecht, The Netherlands medicine department (43%), cardiology department (22%) and the emergency department (10%). Background-aim Conclusions The lack of knowledge of the presence of Drug-Laboratory Test Interactions (DLTIs) can cause misinterpretation of laboratory test In this study, an automated decision support application was results and delayed or erroneous diagnosis with extra healthcare implemented to facilitate signalling the presence of drug laboratory costs and even harm to patients. There are over 50.000 physiolog- test interactions. A mean of 45 DLTI alerts per day were generated in ical and/or analytical drug-test interactions described. In this pilot this study. The clinical relevance of the alerts for laboratory study, an automated decision support application was used to specialists and physicians will be examined. detect drug laboratory test interactions in real-time. doi:10.1016/j.cca.2019.03.1118 Methods

In this multicentre study, 34 clinical rules about DLTI were programmed and validated in an automated decision support