Comparison of Six Automated Treponema-Specific Antibody Assays

Transcription

1 JCM Accepted Manuscript Posted Online 11 November 2015 J. Clin. Microbiol. doi: /jcm Copyright 2015, American Society for Microbiology. All Rights Reserved. 1 Comparison of Six Automated Treponema-Specific Antibody Assays 2 3 Borae G. Park 1, 4 borae.park@gmail.com 5 1 Department of Laboratory Medicine, 6 Severance Hospital, Yonsei University College of Medicine, Seoul, South Korea 7 8 Ji Houn Yoon 1, 9 Joonji422@yuhs.ac 10 1 Department of Laboratory Medicine, 11 Severance Hospital, Yonsei University College of Medicine, Seoul, South Korea John Hoon Rim 1, 14 Johnhoon1@yuhs.ac 15 1 Department of Laboratory Medicine, 16 Severance Hospital, Yonsei University College of Medicine, Seoul, South Korea Anna Lee 2, 1

2 Department of Laboratory Medicine, 21 Seoul Medical Science Institute, Yongin, South Korea Hyon-Suk Kim 1 24 kimhs54@yuhs.ac 25 1 Department of Laboratory Medicine, 26 Severance Hospital, Yonsei University College of Medicine, Seoul, South Korea Corresponding author 29 Hyon-Suk Kim, 30 Address: Department of Laboratory Medicine, Severance Hospital, Yonsei University 31 College of Medicine, 50-1, Yonsei-ro, Seodaemun-gu, Seoul, , Korea 32 address: kimhs54@yuhs.ac 33 Telephone number: Fax number: Keywords: Syphilis, Treponemal test, Treponema-specific antibody, Reverse screening 2

3 37 algorithm 38 3

4 1 Comparison of Six Automated Treponema-Specific Antibody Assays 2 3 Abstract 4 Six different Treponema (TP)-specific immunoassays were compared against the 5 fluorescent treponemal antibody absorption test (FTA-ABS). A total of 615 samples 6 were tested. The overall percent agreement, analytical sensitivity, and specificity of 7 each assay compared to FTA-ABS were as follows: Architect Syphilis TP, 99.2%, %, 100%; Cobas Syphilis, 99.8%, 99.4%, 100%; ADVIA Centaur Syphilis, 99.8%, %, 100%; HISCL Anti-TP Assay Kit, 99.7%, 98.7%, 100%; Immunoticles Auto3 10 TP, 99.0%, 97.5%, 99.6%; Mediace TPLA, 98.0%, 98.1%, 98.0%. All discrepant results 11 between TP-specific assays were associated with samples from non-clinical cases (11 12 immunoassay false positives and 7 from previous syphilis cases). Our study 13 demonstrated that TP-specific immunoassays generally showed high sensitivities, 14 specificities, and percent agreements when compared to FTA-ABS, with rare cases of 15 false positive or false negative results. Therefore, most TP-specific immunoassays are 16 acceptable for use in screening for syphilis. However, it is important to perform a 17 thorough review of a patient s clinical and treatment history when interpreting the 18 results of syphilis serology. 1

5 19 Introduction 20 Syphilis is commonly diagnosed based on the results of a combination of serological 21 tests to detect Treponema (TP) antibodies and non-tp antibodies (1). A traditional 22 screening algorithm for syphilis that began with a non-tp assay failed to detect 3% of 23 syphilis cases, in a previous study (2). Recently, reverse screening algorithm with an 24 automated TP-specific assay has been recommended by the European Centers for 25 Disease Control and Prevention (ECDC) (3). CDC continues to recommend the 26 traditional algorithm yet also recognizes the recent trends of the widely used reverse 27 algorithm and recommends extra TP tests to resolve discordant results (4). The reverse 28 algorithm has been found to show a superior diagnostic performance, with sensitivities 29 ranging from 99.38% to 99.85%, specificities from 99.98% to 100%, and accuracies 30 from 99.93% to 99.96% compared with a 24.2% missed-diagnosis rate and 75.81% 31 sensitivity of the traditional algorithm (4). 32 Various automated TP-specific immunoassays have been developed that use either 33 whole cells or antigens, such as 15TpN, 17TpN, and 47TpN, derived from the Nichols 34 strain of Treponema pallidum to detect IgG, IgM, or total immunoglobulins (1). Initially, 35 enzyme immunoassays (EIAs) were commonly used to detect TP-specific antibodies (6, 36 7). However, gradually, the use of chemiluminescence immunoassays (CLIAs) to detect 2

6 37 TP-specific IgG and IgM antibodies has been increasing (5, 7). Additionally, 38 quantitative TP-specific immunoassays using turbidimetry, based on a latex 39 agglutination method, have been widely used in Asia. However, comparative analyses 40 on the performances of these various methods are lacking. 41 The aim of this study was to evaluate the performance of 6 commonly used TP- 42 specific immunoassays, including CLIA and turbidimetry assays in comparison with the 43 fluorescent treponemal antibody absorption test (FTA-ABS) Material & Methods 46 Study design 47 A total of 615 samples were tested using the 6 kinds of automated TP-specific 48 immunoassays. These samples included 613 leftover serum samples that had been sent 49 for TP or non-tp assay and 2 international standards for syphilis (05/122 and 50/132, 50 NIBSC; National Institute for Biological Standards and Control, Hertfordshire, UK). 51 The samples included those from 105 medical check-ups of healthy individuals, preoperative evaluations of patients with variable underlying disease, and suspicious of current or previous syphilis. The median age of all patients was 48 (range ). All immunoassay results were compared with those of FTA-ABS (Zeus 3

7 55 Scientific, NJ, USA). In addition, the Venereal Disease Research Laboratory test 56 (VDRL) was performed and clinical history was reviewed for samples with discordant 57 TP-specific assay results. Samples were stored at 4 C until all testing was complete. 58 The study protocol was reviewed by the institutional review board at our hospital Treponema-specific immunoassays 61 All six kinds of TP-specific immunoassays were performed following the 62 manufacturer s instructions. Assays evaluated included Architect Syphilis TP (Abbott 63 Diagnostics, Tokyo, Japan), Cobas Syphilis (Roche Diagnostics, Mannheim, Germany), 64 ADVIA Centaur Syphilis (Siemens Healthcare Diagnostics, NY, USA), HISCL Anti-TP 65 Assay Kit (Sysmex Corporation, Kobe, Japan), Immunoticles Auto3 TP (A & T 66 Corporation, Kanagawa, Japan), and Mediace TPLA (Sekisui Medical Co, Tokyo, 67 Japan). Two of these assays are quantitative assays: Immunoticles Auto3 TP and 68 Mediace TPLA. Characteristics of the 6 different immunoassays are presented in Table FTA-ABS and VDRL tests 72 The FTA-ABS assay, which uses the nonviable Nichols strain of T. pallidum for 4

8 73 detection of TP-specific total antibodies, was performed according to the manufacturer s 74 instructions. Every batch of patient samples was tested with negative and positive 75 controls, and the results of positive samples were graded on a scale from 1+ to The quantitative VDRL test was performed using a BD VDRL Antigen kit (Becton, 77 Dickinson and Company, MD, USA) according to the manufacturer s instructions. 78 Serum samples were quantitated to an endpoint titer of 1: All FTA-ABS and VDRL tests were reviewed by 2 clinical pathologists in the 80 laboratory Neutralization assay using TP-specific antigen 83 Neutralization assay reagents were additionally provided for two quantitative 84 immunoturbidimetric assays: Immunoticles Auto3 TP and Mediace TPLA. Purified TP 85 antigens provided by each vendor were mixed with patient samples and incubated for minutes at room temperature. Both raw samples and neutralized samples were tested at 87 the same time. If the TP antibody titer was considerably lower than the value before 88 neutralization, the sample considered as a true positive for TP-specific antibodies Statistical analysis 5

9 91 Statistical analyses were performed using MedCalc Statistical Software version (MedCalc Software bvba, Ostend, Belgium). We evaluated 6 TP assays for analytical 93 sensitivity, specificity, and percent agreement by kappa (κ) coefficients. Linear 94 regression analysis was used to compare quantitative results, and Kruskal Wallis test 95 was used to compare variance among different groups. Mann Whitney U test was used 96 to evaluate differences between 2 groups. P-values of < 0.01 were considered 97 statistically significant Results 100 Results of automated immunoassays 101 A total of 157 samples of 155 patients (median age 56, range 19 93) and 2 standards 102 showed positive results. The overall percent agreement and corresponding κ values of 103 each assay s results compared with those of FTA-ABS were as follows: Architect 104 Syphilis TP, 99.2%, κ = 0.978; Cobas Syphilis, 99.8%, κ = 0.996; ADVIA Centaur 105 Syphilis, 99.8%, κ = 0.996; HISCL Anti-TP Assay Kit, 99.7%, κ = 0.991; Immunoticles 106 Auto3 TP, 99.0%, κ = 0.974; Mediace TPLA, 98.0%, κ = (Table 2). Analytical 107 sensitivities of assays were 96.8%, 99.4%, 99.4%, 98.7%, 97.5%, and 98.1%, and 108 specificities were 100%, 100%, 100%, 100%, 99.6%, and 98.0%, respectively. Eighteen 6

10 109 samples that showed discrepant results between the TP-specific assays were from non- 110 infectious cases. Seven of these results were from cases confirmed by clinical history to 111 be previous syphilis cases, and 11 of these results (2 of healthy individuals, 1 of a year-old allergic dermatitis patient without previous history of syphilis, and 8 patients 113 with variable malignancies who were tested to evaluate preoperative condition without 114 infectious symptoms) were considered to be false positive reactions of the immunoassay 115 (Table 3) Neutralization assay using TP-specific antigen in a quantitative turbidimetry assay 118 The 11 false positive, non-specific reactions were found only in the turbidimetry 119 immunoassays, Immunoticles Auto3 TP and Mediace TPLA assay (Table 3). Nine of 120 these results were from the Mediace TPLA. All 11 results were confirmed to be false 121 positives by neutralization assay. Ten true positive sera, diluted to various 122 concentrations, were tested in parallel using the neutralization assay as a control; none 123 of these samples was identified as false positives Analyses of the quantitative assays 126 The linear analytic measurable range (AMR) for the Immunoticles Auto3 TP assay was 7

11 U/mL and for the Mediace TPLA was U/mL, according to the 128 manufacturer s instructions. A total of 531 data points were within the linear AMR. The quantitative assays showed good correlation by linear regression analysis within the 130 linear AMR (y = x , R 2 = , p < ) (Fig. 1). All data points 131 over the AMR were obtained after dilution, and 9 outliers were excluded. Outliers were 132 defined as those points that were over 1000 units in Immunoticles Auto3 TP, but less 133 than 250 titer units in Mediace TPLA. The regression analysis of these 606 data points 134 showed better correlation (y = x , R 2 = , p < ) than the 135 analysis with the full range of data (Fig. 1). 136 In addition, quantitative results from immunoassays were associated with graded 137 results of FTA-ABS (Fig. 2). Quantitative assays showed false positive results, but these 138 assays could differentiate between high-grade positivity (+3 or +4) and low-grade 139 positivity (+1 or +2). However, there were no statistical differences between +3 and or between +1 and +2 in the Immunoticles Auto3 TP and Mediace TPLA assays (Fig. 2). 141 The units of the 2 biological standards were 0.3 IU/ampoule (05/122) and 3 IU/ampoule 142 (05/132). They measured at 455 and 3,400 TU by Mediace TPLA and 260 and 1, units by Immunoticles Auto3 TP assays

12 145 Discussion 146 Our findings indicated that commercial TP-specific immunoassays currently in use 147 show high sensitivities, specificities, and percent agreements when compared with FTA- 148 ABS. However, rare cases of false positive or false negative resulted. False positive 149 reactions in both non-tp and TP-specific assays are more likely to be seen with specific 150 conditions, such as viral or bacterial infection, autoimmune disease, pregnancy, post- 151 immunization, diabetes, and old age (1, 7). Therefore persistent or transient reactivity in 152 a non-tp or TP assay should be considered as a false-positive reaction without a known 153 history of syphilis or a positive anti-tp IgM test. In our study, false positive results were 154 only seen with turbidimetric immunoassays, especially Mediace TPLA. Previously, % false positives were reported for a latex-agglutinated assay used on samples from 156 non-syphilis patients (8). The neutralization assay was employed to rule out false 157 positive reactions, and it effectively distinguished between true and false positive 158 reactions. Therefore, neutralization assay confirmation should be considered, especially 159 with results of Mediace TPLA testing. 160 Additionally, false positive results can be ruled out by algorithm. If positive TP with 161 negative non-tp results are observed in a patient without a history of treatment, a 162 second TP-specific test could be performed to rule out early or late/latent disease (4, 9). 9

13 163 False negative non-tp tests can result from the prozone phenomenon in early, 164 secondary, or latent syphilis. This phenomenon has rarely been observed to affect TP- 165 specific assays (1, 7). In addition, insufficient antibody production, usually in the 2 to weeks after infection, might lead to false negative results in TP-specific immunoassays. 167 False positive results are rare in active syphilis cases because most immunoassays 168 currently in use can detect both IgM and IgG subtype of TP-specific antibodies. 169 According to our results, all 7 false negative results were from previous or treated 170 syphilis cases. 171 In this study, discordant results were observed between the various TP-specific 172 immunoassays. Regardless of which algorithm is used and because of the analytical 173 limitations of syphilis immunoassays, it is important that the clinician always take the 174 clinical history of a patient into consideration. 175 Several commercial, quantitative TP assays have been introduced, and our study 176 showed good correlation between the 2 quantitative, turbidimetric assays. In addition, 177 quantitative results correlated with low and high titer results from FTA-ABS testing. 178 However, the quantitative results from testing of the biological standards from NIBSC 179 (provided in 2014) did not match to the unit value of each material in our study. In our 180 study, Mediace TPLA and Immunoticles Auto3 TP measured NIBSC standard material 10

14 181 of 0.3 IU for Treponema-specific IgG and IgM, and 3 IU of syphilis antibody as 455 TU, Units and 3,400 TU, 1,610 Units, respectively. So, the unit values of the TPLA 183 results for standards should be re-evaluated. 184 Previous reports have shown that quantitative results of syphilis assays differed 185 according to the clinical cases and over 1,000 U of TP antibody values were helpful to 186 determine syphilis infection (8). However, the linear AMR of quantitative assays are 187 under 300 U. Therefore, sufficient dilutions may be necessary in quantitative assays to 188 differentiate infectious from non-infectious (previous) syphilis cases. 189 Although hands-on times did not differ between assays, the analytical assay times were 190 shorter in quantitative turbidimetric assays than in qualitative immunoassays (Table 1). 191 Additionally, the costs of turbidimetric assays ($1 2) were lower than those of 192 immunoassays ($3 5). However, the quantitative turbidimetric assays required a longer 193 analytical time for positive samples and were relatively expensive in areas with a high 194 prevalence of syphilis due to the need to include a dilution process of samples to show 195 which data points were over the AMR. Therefore, the prevalence of syphilis must be 196 considered in order to determine the kind of TP-specific assay to be used in laboratory. 197 Recently, automated assays to detect TP antibodies and reverse algorithm to detect 198 clinical syphilis are frequently used (10). However, overtreatment following false 11

15 199 positive test results is known to occur. On the other hand, the sensitive detection of TP- 200 specific antibody provides an advantage in areas of low syphilis prevalence (5). 201 In our study, various TP-specific immunoassays performed well in comparison with 202 FTA-ABS. Therefore, these TP-specific immunoassays are sufficient to screen for 203 syphilis. However, it is important to perform a thorough review of each patient s clinical 204 and treatment history when interpreting the results of syphilis serology because of their 205 analytical limitations Acknowledgement 208 This research received no specific grant from any funding agency in the public, 209 commercial, or not-for-profit sectors

16 211 References Morshed MG, Singh AE Recent trends in the serologic diagnosis of syphilis. Clin 213 Vaccine Immunol 22: Centers for Disease Control and Prevention Syphilis testing algorithms using 215 treponemal tests for initial screening--four laboratories, New York City, MMWR 216 Morbidity and mortality weekly report 57: French P, Gomberg M, Janier M, Schmidt B, van Voorst Vader P, Young H, Iust IUSTI: 2008 European Guidelines on the Management of Syphilis. Int J STD AIDS 20: Centers for Disease Control and Prevention Discordant results from reverse sequence 220 syphilis screening--five laboratories, United States, MMWR Morb Mortal Wkly Rep : Tong ML, Lin LR, Liu LL, Zhang HL, Huang SJ, Chen YY, Guo XJ, Xi Y, Liu L, Chen FY, 223 Zhang YF, Zhang Q, Yang TC Analysis of 3 algorithms for syphilis serodiagnosis and 224 implications for clinical management. Clin Infect Dis 58: Binnicker MJ, Jespersen DJ, Rollins LO Treponema-specific tests for serodiagnosis of 226 syphilis: comparative evaluation of seven assays. J Clin Microbiol 49: Cortez KJ, Greenwald MA Current trends in donor testing to detect syphilis infection. 228 Curr Infect Dis Rep 16:

17 Yukimasa N, Miura K, Miyagawa Y, Fukuchi K Evaluation of new automated syphilis 230 test reagents 'immunoticles auto3' series: performance, biochemical reactivity, and clinical 231 significance. J Infect Chemother 21: Binnicker MJ Which algorithm should be used to screen for syphilis? Curr Opin Infect 233 Dis 25: Lee K, Park H, Roh EY, Shin S, Park KU, Park MH, Song EY Characterization of 235 sera with discordant results from reverse sequence screening for syphilis. Biomed Res Int :

18 238 Figure Legend Fig. 1. Linear regression of two quantitative Treponema-specific immunoassays. (A) A 241 total of 531 data points within the linear analytical measurable range (AMR) of each 242 assay (300 unit for Immunoticles and 250 titer units for Mediace), and (B) a total of data points obtained after dilution of samples, showing the data over the AMR and 244 excluding 9 data points as outliers Fig. 2. Comparison between the quantitative results of Treponema-specific 247 immunoassays and grades of the FTA-ABS (fluorescent treponemal antibody 248 absorption) test. (A) Comparison between Mediace TPLA and FTA-ABS results, and 249 (B) between Immunoticles Auto3 TP and FTA-ABS results. All P-values were 250 determined by Mann Whitney U tests

19 Table 1. Characteristics of six different immunoassays to detect Treponema-specific antibody Reagent Instrument Principle Quantitative or qualitative Target antigen* (TpN15, TpN17, and TpN47) Abbott Roche Siemens Sysmex A & T Co. Sekisui Syphilis TP Cobas Syphilis Total antibodies to Architect 2000i (Abbott) Cobas e601 (Roche) Chemiluminescent microparticle immunoassay Electrochemiluminescence immunoassay Syphilis (SYPH) ADVIA Centaur (Siemens) Direct chemiluminometric immunoassay HISCL Anti-TP Assay Kit HISCL-2000i (Sysmex) Chemiluminescent enzyme immunoassay IMMUNOTICLES Auto3 TP CA-400 (Furuno, Hyogo, Japan) Immuno turbidimetry Mediace TPLA CA-400 Immuno turbidimetry Qualitative Qualitative Qualitative Qualitative Quantitative Quantitative Microparticles coated with recombinant Treponema-specific antigens Acridinium-labeled anti human IgG and IgM conjugate Ruthenium Labeled biotinylated Treponema specific antigen Streptavidin-coated microparticle Acridinium ester-labeled biotinylated T. pallidum recombinant antigens (TpN15 and TpN17) Streptavidin-coated magnetic latex particle Alkaline phosphataselabeled biotinylated recombinant Treponema-specific antigens Streptavidin-coated magnetic particles Nichols strain of Treponema pallidum Treponema pallidumsensitized polystyrene latex particle Treponema-specific antigen Treponema-specific antigen-sensitized latex No. of calibrators 1 point 2 point 2 point 1 point 6 point 5 point Controls 1 pos, 1 neg 1 pos, 1 neg 1 pos, 1 neg 1 pos, 1 neg 1 pos, 1 neg 1 pos, 1 neg Specimen Plasma, Serum Plasma, Serum Plasma, Serum Plasma, Serum Plasma, Serum Serum Sample volume 30 µl 10 µl 100 µl 20 µl 17 µl 16 µl Cut-off index (COI) S/CO 1.0 COI 1.0 Index 1.0 COI U 10 TU (1 TU = 2 miu) Total assay time 29 min 18 min 29 min 17 min 11 min 11 min Linear analytical measurable range U TU *All target antigens are composed of TpN15, TpN17, and TpN47, except Siemens Syphilis

20 Table 2. Evaluation of various Treponema assays in comparison with the FTA-ABS assay FTA-ABS result Sensitivity Specificity Agreement Kappa value Reactive Non-reactive Abbott Architect Reactive ( ) Non-reactive Standard error = Roche Cobas Reactive ( ) Non-reactive Standard error = ADVIA Centaur Reactive ( ) Non-reactive Standard error = Sysmex HISCL Reactive ( ) Non-reactive Standard error = A&T Immunoticles Reactive ( ) Non-reactive Standard error = Sekisui Mediace Reactive ( ) Non-reactive Standard error = *FTA-ABS: Fluorescent treponemal antibody absorption test

21 Table 3. Data on 18 discordant results between Treponema-specific immunoassays of 615 specimens False Positive (11) False Negative (7) No. of FTA- A&T VDRL Abbott Roche Siemens Sysmex Sekisui results ABS Co R R - 2 R (1:1) R 1 - R (+1) R (+3) R R R R R R (+2) - R R - - R 2 R (1:1) R (+2) - R R R R R 1 R (1:1) R (+2) R R R R - R 1 R (1:1) R (+1) - R R R - - *R: Reactive, VDRL: Venereal Disease Research Laboratory test, FTA-ABS: Fluorescent treponemal antibody absorption test

22

23

24

25