The TEST 1 Automated System A New Method for Measuring the Erythrocyte Sedimentation Rate

Transcription

1 The TEST 1 Automated System A New Method for Measuring the Erythrocyte Sedimentation Rate Mario Plebani, MD, 1,2 Stefania De Toni, MD,1 Maria Colomba Sanzari, MD,' Daniela Bernardi, MD,1 and Elisabeth Stockreiter, MD1 Key Words: Erythrocyte sedimentation rate; Westergren ESR method; ICSH reference method; Automated analyzer We evaluated performance of the TEST 1 (SIRE Analytical Systems, Udine, Italy), a fully automated analyzer for the measurement of the erythrocyte sedimentation rate (ESR). Intra-assay reproducibility was satisfactory for a wide range of ESR values, whereas there was a significant decrease in ESR data when the samples were stored at 4 Cfor up to 24 hours. We compared TEST I with the Westergren ESR method approved by the International Council for Standardization in Haematology (ICSH) and with the Diesse Ves-Matic analyzer. Linear regression analysis comparing the TEST 1 and the ICSH reference method yielded satisfactory correlations for K3 EDTA- and sodium citrate-anticoagulated samples. Bland-Altman analysis showed no evidence of a systematic bias between the TEST I and the reference method. A close correlation was found between the TEST I system and the Diesse Ves-Matic analyzer despite a significant positive systematic bias. Reference values for men and women were analyzed according to nonparametric statistics. The TEST I was easy to use, had a satisfactory operative praticability, required minimal maintenance, and reduced contact with potential biohazards. This system enables the determination of ESR with any common standard-sized tube; the use of samples anticoagulated with K, EDTA can widely reduce the workload in clinical laboratories. 334 AmJCIinPathol 1998;110: The erythrocyte sedimentation rate (ESR) test has long been used in clinical laboratories because it is straightforward, speedy, and inexpensive.1 Although ESR is a diagnostically nonspecific and insensitive laboratory test,2 it is still considered a reliable indirect indicator of the acute phase inflammatory response,3 and it is usually increased in the presence of infectious diseases and acute or chronic inflammatory processes.4,5 ESR is widely ascertained to monitor the course of many rheumatic and collagen-vascular diseases and their response to therapy,6 and it is commonly used as a diagnostic or classification criterion for temporal arteritis and polymyalgia rheumatica and as a prognostic index for monitoring disease activity and for establishing remission in patients with rheumatoid arthritis.7 The ESR may be elevated in many other diseases, including cancer and Hodgkin disease.4,5 The reference method for measuring the ESR proposed by the International Committee for Standardization in Haematology (ICSH)8 and based on the technique described by Westergren during the 1920s9 uses blood diluted with sodium citrate in open-ended vertically oriented glass tubes 300-mm long and 2.5 mm in diameter. The ESR test has been performed in practically the same way for more than 70 years, thus allowing comparable reference values to be maintained within the same laboratory and between different laboratories over time. At present the ICSH reference method is essentially the same as it always has been, with a minor modification: the use of undiluted K3 EDTA-anticoagulated blood.10 The Westergren modified method, still considered the "gold standard" for measuring ESR, is laborious and cumbersome and, above all, involves a high biohazard risk to laboratory operators.11,12 It is not recommended for routine ESR determination in clinical laboratories, which often do not have sedimentation tubes with an inner diameter large enough to analyze clinical samples with a hematocrit of 0.35 L/L or Abstract

2 Materials and Methods System Descriptions TEST1 The TEST 1, a closed automated analyzer, determines the ESR in any standard-sized primary tube with a perforating stopper. The tubes are placed in appropriate racks, and their contents are rotated slowly for about 2 minutes. The sample loader simultaneously accepts 4 racks containing 15 tubes each. By using a closed aspiration needle, the blood is directly drawn from the collection tube, distributed in a capillary, and centrifuged at about 20g. The sensing area temperature is maintained at 37 C. The system uses an infrared ray microphotometer with a light wavelength of 950 nm. The electrical impulses, collected using a photodiode detector, are directly correlated to the concentration of RBCs present at that capillary level. The impulses measured per unit of time are then used to delineate a sedimentation curve for each sample. By using a linear regression model, the mean decrease in the signal per unit of time, called "medium signal," and the square root of the "integral signal" are transformed to comparable Westergren values. The system operates at an analytic rate of 110 specimens per hour in continuous loading, providing a finding every 32 seconds; 200 LtL of whole blood are required for each sample. Data on patients can be directly transmitted by connecting the instrument to the laboratory information system by means of a standard interface (RS232). In our study, samples were collected by means of venipuncture and injected into 2 different types of tube: Vacutainer 13 x 75 mm containing 3 ml of blood anticoagulated with K3 EDTA (7.5%, ml) (Becton Dickinson, Meylan, France), and Vacu-Tec (Diesse Diagnostica Senese) containing sodium citrate (the same solution as that used by the Diesse Ves-Matic). Results were expressed as ESR = x x mm/h, with "x" representing the unknown. ICSH Reference Method Following the ICSH recommendations for the reference method,10 the Westergren ESR was performed within 4 hours of collection on undiluted blood anticoagulated with K3 EDTA. All samples had an original hematocrit of 0.35 L/L or less. Glass pipettes (300-mm long) were used with a 200-mm sedimentation scale marked on the pipette and a constant bore no lower than 2.55 ±0.15 mm. Immediately before setting up the test, each specimen was mixed by 8 complete manual inversions. During sedimentation, the pipettes were mounted vertically on appropriate supporting racks and protected from direct light, extreme heat, and vibrations. The stands were kept at room temperature, which never exceeded 25 C. Results were expressed as ESR (undiluted) = x x mm/h. Diesse Ves-Matic The Diesse Ves-Matic System has been programmed for the simultaneous determination of ESR in 60 blood samples. This automated system uses rectangular plastic vacuum Vacu-Tec tubes (90-mm long) containing sodium citrate (0.36 ml of a mol/L concentration) as the anticoagulant. The collection tubes, containing 1 ml of sample, are placed in individual cuvette holders and rotated gently for 2 minutes. Sixty photoelectronic elements (photodiode and phototransistor) then record the blood level at which light transmission occurs, before and after a 20-minute interval. During this stage, the tubes are maintained at an angle of 18 to the vertical. Results are expressed as ESR = x x mm/h. Specimen and Sample Population Samples obtained from 297 outpatients (174 men and 123 women; age range, years) who had given informed consent were used to establish the local set of reference values. None of the patients had been ill during the 3 months before the study, and none were taking drugs. A second batch of samples (n = 448) was obtained from a large population of patients from different departments of Padua University-Hospital (Italy). Two hundred seventy-five K3 EDTA-anticoagulated specimens with a packed cell volume (PCV) of 0.35 or less were used to compare results from the reference ICSH method with those from the TEST 1 following recommendations of the National Committee for Clinical Laboratory Standards (NCCLS).5 All blood samples were collected in K3 EDTA Vacutainers and in sodium citrate Vacu-Tec tubes and assayed within 4 hours of venipuncture, following the aforementioned guidelines for each test method. The PCV or hematocrit, Am J Clin Pathol 1998;110: more. Many systems using closed blood collection tubes and sometimes automatic readers (eg, the Sediplast ESR system [Polymedco, Cortland Manor, NY]) therefore recently have been developed and introduced into clinical practice for measuring ESR. In particular, the Diesse VesMatic System (Diesse Diagnostica Senese, Milano, Italy), a closed analyzer using rectangular sealed vacuum tubes containing sodium citrate, seems the most widely used method in Europe and North America.4 We evaluated a new closed automated analyzer, the TEST 1 (SIRE Analytical Systems, Udine, Italy), which uses quantitative capillary photometry technology for measuring the ESR. With this instrument any common standard-size collection tube with a perforating stopper may be used. Our study evaluated the performance of the TEST 1 ESR System and compared its performance with that of the ICSH reference method and with the routinely used technique.

3 Plebani et al / EVALUATION OF THE TEST 1 SYSTEM FOR MEASURING ERYTHROCYTE SEDIMENTATION RATE ITable II Intra-Assay Reproducibility for TEST 1* in K3 EDTA Specimens Sample ESR Mean ± SD (mm/h) 10 ± ± ± ± ± ± ±3.32 distributions were created, and the 97.5 percentile was determined for each method used.15 Coefficient of Variation (%) obtained by the hematologic instrument Technicon H*2 (Bayer, Tarrytown, NY), also was determined for each sample. Evaluation Protocol The intra-assay precision of the TEST 1 was determined over a wide range of concentrations by performing 10 replicate measurements of 7 samples anticoagulated with K3 EDTA, with an ESR value in the range of 10 to 117 mm/h. The mean, SD, and coefficient of variation were also calculated. A quality control program was conducted by making a daily selection of an EDTA blood sample with a PCV of 0.35 L/L or less that was known to have high ESR value and by performing the TEST 1 and the reference method, following the NCCLS recommendations.5 The stability of ESR samples in time was evaluated by storing 53 specimens for 24 hours at 4 C. The samples were returned to room temperature for 45 minutes and remixed before the second measurement. The inaccuracy was evaluated by comparing the ESR data of samples anticoagulated with K3 EDTA from the TEST 1 with those obtained by using the ICSH reference method. The ESR determinations in specimens anticoagulated with sodium citrate were compared with undiluted ESR values for the ICSH reference method and corrected for lack of dilution by using the following formula: Undiluted Reference ESR x '13 To further study agreement, the ESR determinations performed on the TEST 1 with sodium citrate-anticoagulated blood were compared with data obtained with the system used routinely in our laboratory, the Diesse VesMatic. Moreover, the data obtained were used to compare the Diesse Ves-Matic with the reference ICSH method for sodium citrate-anticoagulated specimens. To calculate the reference ranges, in view of the rise in ESR in relation to age,14 values were collected separately for each decade of adult life in men and women. Frequency 336 Am J Clin Pathol 1998;110: To estimate within-run precision, we computed means, SDs, and coefficients of variation. Differences between means for stability were evaluated by using the Student's 2-way t test. The data from method comparison studies were analyzed by simple least squares linear regression to obtain the y-intercept (i), slope (s), SD of regression line (Sy/x), and the Pearson correlation coefficient (r). Moreover, data also were compared by using Altman and Bland analysis,16 allowing the assessment of bias, agreement limits, and accuracy. Presence of accuracy can be established by using a 95% confidence interval (CI) for the mean difference between the methods. There is therefore no evidence of systematic bias when the 95% CI includes zero. Estimates for the reference ranges and for comparison between them were based on a nonparametric statistical method using percentiles. Differences between percentiles were considered significant at P<.05 by the Mann-Whitney test and differences between age-related groups by 1-way analysis of variance and the Scheffe F-test.1718 Results Reproducibility The intra-assay precision of the TEST 1 was evaluated in 7 K3 EDTA-anticoagulated specimens with ESR values ranging from 10 to 117 mm/h. ITable II gives the results expressed as mean ± SD and coefficient of variation for 10 replicates. The ESR quality control chart is shown in IFigure II. Over time, no significant variations were observed in the agreement between the method evaluated and the reference method; the SD of the bias (mean 0.3 ± 4.6) always was within acceptable limits. Stability Study Storage of blood samples at 4 C for up to 24 hours caused a significant decrease in ESR values obtained with the TEST 1; the mean difference was 8.6 mm/h (95% CI, ) for K, EDTA-anticoagulated samples (P , Student's paired t test), and 3.8 mm/h (95% CI, ) for specimens anticoagulated with sodium citrate (P =.0001, Student's paired t test), reflecting decreases of 20.9% and 13.5%, respectively. Accuracy K3 EDTA-anticoagulated samples from 275 patients with a PCV of <0.35 (median, 0.33) were studied to compare ESR = erythrocyte sedimentation rale. *SIRE Analytical Systems, Udine, Italy. Statistical Analysis

4 ** y * ^ *f\#* &: S 60 t #* >*Jr? * * -.3J 40%2^r* * Days Figure I I Erythrocyte sedimentation rate quality control chart obtained by plotting day to day the differences between values from the reference method and those from the TEST 1 (SIRE Analytical Systems, Udine, Italy) method. the TEST 1 with the ICSH reference method. Linear regression analysis showed a satisfactory correlation between the 2 methods IFigure 21. A close correlation (r = 0.86; P =.0001; slope = 0.99; intercept = 2.38; and Sy/x = 16.11) was also obtained between the TEST 1 sodium citrate-anticoagulated ESR determinations and the ICSH reference data corrected for lack of dilution on 239 specimens, with hematocrit values of 0.35 or less (median, 0.329). A comparative study was made between the TEST 1 and the reference method by plotting the mean values for ESR with both methods against the difference between the 2 methods following Bland-Altman analysis. The mean bias was mm/h (limits of agreement, to 37.94) for samples anticoagulated with K3 EDTA. We found no systematic bias between the 2 methods, as demonstrated by the 95% CIs, which ranged from to Similar results were obtained when ESRs were assayed on sodium citrate-anticoagulated samples, demonstrating that the TEST 1 system is highly accurate (95% CI, ICSH (mm/h) Figure 21 Comparison between the reference method (International Council for Standardization in Haematology [ICSH]) and TEST 1 (SIRE Analytical Systems, Udine, Italy). Samples were collected in K3 EDTA. y = 0.99x ; r = 0.85; P =.0001 ;Sy/x = to 0.01). The mean bias was mm/h (limits of agreement, to 29.48) ITable 21. Moreover, a satisfactory correlation (r = 0.83; P =.0001) was found when the ESR Diesse Ves-Matic values (y) obtained in the same samples were compared with the ICSH corrected reference data (x), with a linear regression equation of y = 1.10 x (Sy/x = 20.26). Bland-Altman analysis, on the other hand, showed a negative mean bias equal to mm/h (limits of agreement, to 32.63), thus demonstrating that the routine system tends to overestimate ESR values (95% CI, to 4.80). A close agreement (r = 0.91; P =.0001) also was found between the values obtained on TEST 1 with sodium citrate-anticoagulated samples and Diesse Ves-Matic data from 745 patients, on the basis of the regression equation of y = 0.85 x There was, however, a positive systematic bias (mean bias, 3.84 mm/h, with an interval of agreement of to 32.7) between the 2 methods, indicating that the 20 ITable 21 Bland-Altman Analysis (ICSH vs TEST 1*) Bias Agreement limits 95% confidence interval TEST 1, EDTA TEST 1, Sodium Citrate to to to to 0.01 ICSH = International Council for Standardization in Haematology. *SIRE Analytical Systems. Udine. Italy. Am J Clin Pathol 1998; 110:

5 Plebani et al / EVALUATION OF THE TEST 1 SYSTEM FOR MEASURING ERYTHROCYTE SEDIMENTATION RATE ITable 31 Reference Values Men (n = 174) Method/Age Group (y)* Upper L i m i t (97.5%) (mm/h) 97.5 Percentile (95% CI) ESR Mean (mm/h) Upper L i m i t (97.5%) (mm/h) 97.5 Percentile (95% CI) ESR Mean (mm/h) ESR = erythrocyte sedimentation rate; CI = confidence interval; ICSH = International Council for Standardization in Haematology. *The numbers in each age group were as follows: men. years, 49; years, 52; years, 29; years, 44; women, years, 34; 31^10 years, 33; years, 26; years, 30. T TEST 1 from SIRE Analytical Systems, Udine, Italy; Diesse Ves-Matic, Diesse Diagnostica Senese, Milano, Italy. TEST 1 tended to underestimate values compared with the routine Ves-Matic system (95% CI, 2.78 to 4.90). present. ITable 41 gives the 97.5 percentiles and 95% CIs for men and women distinguishing between 2 age groups (younger than 50 years and older than 50 years). Reference Range The reference intervals for ESR assayed in each method are shown in ITable 31. With the 1-way analysis of variance and the Scheffe F-test, no statistically significant differences were found in the first 3 age-decade groups for men or women. When, however, data were assembled, a significant difference (P<.05) was found between previously reported groups (younger than 50 years) and the last group (older than 50 years). Moreover, the statistical differences for the 97.5 percentile were examined between the TEST 1 and the routine technique and between the TEST 1 and the ICSH reference method for men and women by using the Mann-Whitney test. On comparing the ICSH reference EDTA and the TEST 1 EDTA methods, no significant difference was found between data for subjects older than 50 years (men and women), whereas a significant difference (P<.05) was found for subjects younger than 50 years. When the Diesse VES-MATIC system was compared with the reference method, a significant difference between all age-related groups of reference subjects always was 338 Am J Clin Pathol 1998; 110: Discussion It is important to establish whether this new technique for ESR measurement is of value and, if so, whether it is necessary. The TEST 1 system, designed to overcome some drawbacks of the Westergren ESR method, incorporates important improvements: (1) It obviates contact between operator and blood, thus reducing the biohazard risk and is less time consuming. (2) It can be used with any closed blood collection tube in a routine setting and, in particular, with undiluted K3 EDTA-anticoagulated blood, which is the recommended type of sample. (3) Clinical laboratories that wish to continue with the consolidated sodium citrate-type of diluted specimen can use this system. A satisfactory intra-assay reproducibility was obtained over a wide range of ESR values, and the reduction in precision at low rates of sedimentation does not affect the clinical reliability of measurements. The lack of control ICSH ICSH diluted TEST 1, f EDTA TEST 1, sodium citrate Diesse Ves-Matic* Women (n = 123)

6 Table 41 Reference Values for Men and Women by Age Groups* Women years ICSH, EDTA ICSH, diluted TEST 1, f EDTA TEST 1, sodium citrate Diesse Ves-Matict 36.0 ( ) 18.9( ) 31.7( ) 16.6( ) 16.0( ) Men > 50 years 41.9( ) 25.8( ) 37.1 ( ) 24.7( ) 24.0( ) years 26 ( ) 10.4( ) 14.92( ) 8.1 ( ) 16.0( ) > 50 years 36.8 ( ) 19.82( ) 34.8( ) 17.3( ) 16.0( ) materials precluded traditional studies on inter-assay precision. However, following the suggestions made by the NCCLS,5 at various frequencies we checked the agreement between the TEST 1 and the reference method and found no significant variations in agreement over time. Moreover, the ESR procedure is inherently quite stable and does not require calibration. Storage of samples at 4 C for up to 24 hours caused a significant decrease in ESR values in K3 EDTA- and sodium citrate-anticoagulated specimens. This finding contradicts the finding in 1965 by Gambino,19 but is in agreement with the NCCLS recommendation to set up the test within 4 hours if blood samples are left at room temperature or within 12 hours if they are stored at 4 C. Regarding the comparison between TEST 1 values and those obtained with the reference method and/or routine techniques, it should be noted that we used 2 types of statistical analysis to evaluate agreement between the technique considered and the reference method. It is in fact widely accepted that the Bland-Altman approach is more accurate than regression analysis. In fact, the value of the correlation coefficient in the latter depends on the range of data sampled, and extreme data may lead to misleadingly inflated correlation coefficients. This may explain why a significant bias was found in studies comparing the reference method and the Diesse Ves-Matic technique despite a satisfactory correlation coefficient (r = 0.83; P =.0001). On the other hand, we found a close correlation between the reference method and the TEST 1 in K3 EDTA- and sodium citrate-anticoagulated specimens without significant bias. Finally, the reference interval data were analyzed by using nonparametric statistics (percentiles) because ESR values did not conform to a true gaussian distribution. On comparing the 97.5 percentiles (upper limit of the reference range), no significant differences were found between the ICSH reference method and the TEST 1 in men and in women older than 50 years, while a statistically significant difference was found between the reference method and the Diesse Ves-Matic technique. Conclusion The new fully automated system for ESR measurement is safe, precise, accurate, and easy to use in a clinical setting. As this technique allows the use of K3 EDTA-anticoagulated specimens and any perforating stopper collection tubes and has a high productivity (110 samples per hour), it can easily be integrated in modern hematologic analyzers. This might enable the creation of a work station for K3 EDTA-anticoagulated specimens that would allow us to obtain, in a random-access way, CBC counts, including the leukocyte differential count and, when required, reticulocyte and/or ESR determination, thus contributing to optimizing the work flow in clinical laboratories. From the 'Department of Laboratory Medicine, University Hospital of Padua, Italy, and 2Centre of Biochemical Research, Castelfranco Veneto, Italy. Address reprint requests to Dr Plebani: Laboratorio Centrale, Azienda Ospedaliera di Padova, Via Giustiniani, 2, Padova, Italy. References 1. Hertzman A, Evans TI, Sanders KM, et al. Effects of blood storage on the erythrocyte sedimentation rate. ] Rheumatol. 1993;20: Kallner A. On the temporal development of erythrocyte sedimentation rate using sealed vacuum tubes. Am ] Hematol. 1991;37: Van Leeuwen MA, van Rijswijk MH. Acute phase proteins in the monitoring of inflammatory disorders. BailUeres Clin Rheumatol. 1994;8: Zlonis M. The mystique of the erythrocyte sedimentation rate: a reappraisal of one of the oldest laboratory tests still in use. CUnlabMed. 1993;13: Am J Clin Pathol 1998; 110: ICSH = International Council for Standardization in Haematology. *Data are given as 97.5 percentile (95% confidence interval). t TEST 1 from SIRE Analytical Systems, Udine. Italy; Diesse Ves-Matic, Diesse Diagnostica Senese S.r.l., Milano. Italy.

7 Plebani e t al / EVALUATION OF THE TEST 1 SYSTEM FOR MEASURING ERYTHROCYTE SEDIMENTATION RATE 5. National Committee for Clinical Laboratory Standards. Methods for the Erythrocyte Sedimentation Rate (ESR) Test: Approved Standard. 3rd ed. Villanova, Pa: National Committee for Clinical Laboratory Standards; NCCLS document H2-A3. 6. Weinstein A, Del Giudice J. The erythrocyte sedimentation rate: time honored and tradition bound. J Rheumatol. 1994;21: Wolfe F, Michaud K. The clinical and research significance of the erythrocyte sedimentation fate. ] Rheumatol. 1994;21: Westergren A. Studies of the suspension stability of the blood in pulmonary tuberculosis. Acta Med Scand. 1921;54: International Council for Standardization in Haematology. ICSH recommendations for measurement of erythrocyte sedimentation rate. J Clin Pathol. 1993;46: Caswell M, Stuatt ]. Assessment of Diesse Ves-matic automated system for measuring erythrocyte sedimentation rate. J Clin Pathol. 1991;44: Bain BJ. Some influences on the ESR and the fibrinogen level in healthy subjects. Clin Lab Haematol. 1983;5: National Committee for Clinical Laboratory Standards. How to Define, Determine, and Utilize Reference Intervals in the Clinical Laboratory: Proposed Guideline. Villanova, Pa: National Committee for Clinical Laboratory Standards; NCCLS document C28-P. 16. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1: DDV Software, Astute. Statistics add-in for Microsoft Excel. DDV Software - University of Leeds, Old Medical School, Leeds, UK, Gardner H], Gardner SB, Winter PD. Confidence interval analysis microcomputer program. BMJ. 1989; Gambino SR, DiRe JJ, Monteleone M. The Westergren sedimentation rate using K, EDTA. Tech Bull Registry Med Technologists. 1965;35: Patton W N, Meyer PJ, Stuart J. Evaluation of sealed vacuum extraction method (Seditainer) for measurement of erythrocyte sedimentation rate. J C!m Pathol. 1989;42: Am J Clin Pathol : International Committee for Standardization in Hematology. Reference method for the erythrocyte sedimentation rate (ESR) test on human blood. Br J Haematol. 1973;24: International Committee for Standardization in Haematology. Guidelines on selection of laboratory tests for monitoring the acute phase response. J Clin PathoL 1988;41: