A multicenter evaluation of the Technoscreen ADAMTS13 activity semi‐quantitative screening test for thrombotic thrombocytopenic purpura diagnosis and exclusion

Thrombotic thrombocytopenic purpura (TTP) is a rare but potentially fatal microangiopathy, with an untreated mortality rate of around 90%. TTP is caused by severe deficiency in ADAMTS13, which results in accumulation of ultra large von Willebrand factor multimers, triggering a consumptive thrombocytopenia, microangiopathic hemolytic anemia and end‐organ dysfunction and damage. Demonstration of severe ADAMTS13 deficiency is diagnostic for TTP, but long turnaround times for quantitative activity testing often necessitates empirical plasma exchange and/or caplacizumab treatment.


| INTRODUCTION
Thrombotic thrombocytopenic purpura (TTP) is a rare but potentially fatal disease, with 1-6 episodes per million people per year reported in several databases. [1][2][3][4] TTP represents a hematological emergency, with a near 90% mortality rate if not treated promptly. 5 Up to 50% of deaths occur within 24 h of presentation, underscoring the importance of swift commencement of treatment with therapeutic plasma exchange (TPE) and/or caplacizumab. 6 Diagnosis of TTP can be challenging due to overlapping clinical features with many other thrombotic microangiopathies (TMA), including hemolytic uremic syndromes, catastrophic antiphospholipid syndrome, malignant hypertension, disseminated intravascular coagulation, and TMAs associated with pregnancy, transplant, malignancy or pharmacological agents. [7][8][9] Despite shared clinical features, pathophysiology and treatment modalities vary substantially. TTP is caused by severe deficiency of ADAMTS13 (A Disintegrin And Metalloproteinase with Thrombospondin-1-like motifs, member 13), 5,10,11 which reduces ultra large von Willebrand factor (ULVWF) multimers into smaller units.
ADAMTS13 deficiency results in accumulation of ULVWF multimers, which can, in certain situations, bind to and activate platelets, provoking systemic platelet aggregation and leading to development of clinical TMA. 10,11 TTP is diagnosed when ADAMTS13 activity levels are <10% of normal, although ADAMTS13 activity of 10%-20% might be compatible with TTP. [10][11][12][13] ADAMTS13 deficiency in TTP most frequently results from auto-antibodies against ADAMTS13 (immune-mediated TTP); though congenital TTP also exists. [10][11][12][13][14] Quantification of ADAMTS13 activity is generally restricted to specialized laboratories. Moreover, the most popular method is ELISA (enzyme linked immunosorbent assay), most economically performed as a batch assay, 15 which can delay test results by up to 7 days from presentation. 12,[14][15][16] This delay is so ubiquitous that consensus guidelines, recommending target time to result of <72 h, acknowledge results within 7 days are 'acceptable'. 10,11 Accordingly, access to fast, easily accessible, widely available, methods would be useful. The Technoscreen assay is a semi-quantitative flow through assay for determination of ADAMTS13 activity in plasma, manufactured by Technoclone (Vienna, Austria). This bench-top assay does not require use of analytic instruments, and is intended as a screening assay, with ADAMTS13 deficiency needing to be subsequently confirmed by quantitative methods. 17 Clinical risk assessment tools (e.g., French TMA and PLASMIC scores), have been developed to augment clinical judgement. [18][19][20][21] This contemporary approach has proven successful, and since the introduction of TPE to treatment algorithms, mortality has fallen to 5-20%. 3,5,22,23 Although delays to definitive diagnosis of TTP may be compensated by empirical therapy, this is sub-optimal for patient care.
Inexperienced centers may be reluctant to commence therapy, with one report indicating TPE delay ≥48 h from presentation in 15% of cases. 1 Diagnostic confirmation also allows use of adjunctive therapies, including multi-modal immunosuppression or targeted therapy (e.g., caplacizumab if available). 10,11,24,25 Similarly, where ADAMTS13 results indicate alternative diagnoses, initiation of more effective targeted therapy (e.g., eculizumab for atypical hemolytic uremic syndrome) could be delayed. 26 29 ) and also the number of TPE procedures (by 62%) 29 for TMA patients. Thus, optimizing speed and accuracy of diagnosis facilitates appropriate therapy, and minimizes resource intensive potentially harmful treatments. We therefore evaluated the Technoscreen assay for ADAMTS13, with the above considerations in mind.

| Study settings and objectives
This report comprises an academic study, independent of the manufacturer of the Technoscreen assay, comprising five different sites within NSW Health Pathology, all associated with tertiary metropolitan hospitals (Figure 1). Reagents were purchased by study centers from the local supplier, Helena Laboratories (Melbourne, Australia), who delivered reagents according to their standard logistical pathway.
The main objective was to compare the Technoscreen assay to current standard practice, as a potential screening test for ADAMTS13 activity. For site ('A'), standard practice was referral of quantitative testing to one of two possible central sites (i.e., either site C or site E, depending on courier arrangements) ( Figure 1). Site 'B' also referred samples to site E. In two sites (C, D), quantitative testing was available onsite; here, the assessment aimed to validate the method for potential use for earlier diagnosis/exclusion of TTP at partner sites without onsite quantitative testing. Early experience with the assay at site 'A' suggested inter-observer error may affect results; thus, a secondary site-specific objective was to assess inter-observer reliability in reading assay results.

| ETHICS STATEMENT
This study was designed to assess accuracy of a screening test compared to current standard of practice, thus representing a quality assessment study for which formal ethics approval was not deemed necessary. Samples were plasma in excess to requirement for quantitative ('diagnostic') ADAMTS13 testing (i.e., no additional samples collected). Local institutional guidelines addressing collection of data for quality audits were adhered to. Most samples were frozen immediately to À80 C, although some were assessed prospectively as fresh (unfrozen) samples. Clinical data were limited to basic patient demographics.

| The Technoscreen assay
Each assay is performed using a test device containing a test and control well. Patient plasma is incubated with a lyophilized VWF fragment for Site B, Lots SA03C00.01 and SA21C00.01, for Sites C and D, lots SA85B00.01 and SA12C00.01 were respectively used. Figure 2A shows a sample test result using this method.

| Quantitative ADAMTS13 activity testing
For Sites A and B, referral laboratories performed quantitative testing, typically in batches, using either Technoclone Technozym ELISA assay 15 or AcuStar (Werfen, Bedford, USA) Chemiluminescence assay. 15,30 For sites C and D, quantitative testing was available onsite, respectively using ELISA or AcuStar, the latter being an automated test using magnetic beads coated with GST labelled VWF peptide 73 fragments. 30 These fragments are cleaved by ADAMTS13 and bind to isoluminol-labelled anti-VWF, which emits light signals proportional to the degree of ADAMTS13 activity. The Technozym ADAMTS13 activity ELISA also uses a GST-VWF73 peptide fragment, which is immobilized to plastic wells, and when cleaved by ADAMTS13, captures a horse radish peroxidase conjugated monoclonal antibody to generate a colorimetric signal using a substrate.
Sensitivity, specificity, PPV, NPV for detection and exclusion of severe ADAMTS13 deficiency is not reported by either of the manufacturers in their currently provided information for use (IFU) documentation. The detection limit for the AcuStar is reported as 0.1 IU/dL, and the coefficient of variation for a low level (28.7 IU/dL ADAMTS13) control is given as 4.3% for between run testing. For the ELISA, the manufacturer reports a detection limit of 0.2 IU/dL, and the inter-assay coefficient of variation for a low level (8.2 IU/dL ADAMTS13) control is given as 8.0%.
F I G U R E 1 A flow chart of the study as performed and described in this report.

Results of quantitative assays were tabulated alongside
Technoscreen test results and simple patient demographic data using Excel Worksheets. were tested post freezing. Frozen samples were thawed in a 37 C water bath for 10 min and recentrifuged at 1200 g for 15 min prior to testing.

| Sample testing and assessment
All assays were performed in accordance with manufacturer instructions.
All samples processed were visually confirmed as acceptable and not affected by any interfering substances (icterus, hemolysis, lipemia).
Results were interpreted by the primary operator, using the lot specific reference color chart, between 2 and 5 min after the final reagent step (product information indicates to leave 1-2 min after addition of the last reagent, and to read within 10 min). For site A, a single operator was involved; however, results were photographed alongside the reference color chart and also assessed by a further four assessors (two clinical scientists, two hematology registrars), for a total of five assessments. All assessors were blinded to quantitative results. The final result was established by a majority assessment. An example is shown in Figure 2B. For other sites, the final result was taken as that of the primary assessor at that site for that test. For site C, four different operators were involved in the study; for sites B and D, a single operator was involved. The tests were run on the same patient plasma sample at the same time from the same lot (SA03C00.01, quantitative result of 150 IU/dL), with left (10 IU/dL) from a previously used kit and right (40 IU/dL) from a new kit. (D) Quantitative ADAMTS13 activity levels plotted against corresponding Technoscreen ADAMTS13 activity semi-quantitative levels for Sites A and B data using the problematic lot SA03C00.01. (E) Combined quantitative data from all sites using data from three lots (SA85B00.01, SA12C00.01 and SA21C00.01) that appeared to be fit for purpose.  Based on contingency assessment (Table 1), performance data were calculated for identification of ADAMTS13 deficiency: (with 95% confidence interval [CI] reported in parentheses) sensitivity 100% (57%-100%); specificity 67% (50%-80%); PPV 31% (14%-55%); NPV 100% (50%-100%) ( Table 3).

| Statistical analysis
Six results (13%) were amended from initial readings from primary assessor following assessment by four additional assessors, based on majority assessment. Two tests initially considered with an acceptable control were invalidated by majority result. Two tests reading 0 IU/dL were amended to 10 IU/dL, changing positive results to negative.
Two tests reading 10 IU/dL were amended to 0 IU/dL, changing negative results to positive. None of these involved true positive results (as defined by quantitative result). There was good inter-observer reliability for Technoscreen ADAMTS13 activity levels (ICC = 0.85; 95% CI 0.81-0.89). Table 2 Figure 2C shows examples). There were too few duplicated samples to assess statistically; however, this inconsistency remains an important observation.

| Sites B-D
These sites used similar approaches, with similar findings generally observed; thus, composite data is shown in Table 1 and Figure 2D,E.
Data appeared more reliably aligned between screen/quantitative values compared for certain lots, suggesting possible batch lot issues for the Technoscreen assay, particularly for lot SA03C00.01 at Sites A and B.
In total, valid results were available for 70 samples, with test failures identified in a further 12 samples, for a total of 82 tests.
Quantitative ADAMTS13 activity levels ranged from <1 to 120 IU/dL, with median of 52 IU/dL. Eighteen samples had an ADAMTS13 activity level consistent with TTP (<10 IU/dL), and all of these were recorded as 0 IU/dL (true positive) on the Technoscreen assay. An additional three samples were recorded as 0 IU/dL on the Technoscreen assay, with quantitative results ≥10 IU/dL using the quantitative assay (thus, 3 false positives). Forty-nine results with a quantitative ADAMTS13 activity of ≥10 IU/dL were recorded as ≥10 IU/dL on the Technoscreen assay (i.e., true negatives). No false negatives were identified. These results are shown in Table 1 and Figure 2E.

| DISCUSSION
This study's objective was to compare performance of the Technoscreen to current standard of care quantitative testing of ADAMTS13 activity. Potential benefits are rapid, semi-quantitative assessment of ADAMTS13 activity to guide initial management of TTP, thus aiding quick clinical decisions on most appropriate treatment (e.g., TPE/ caplacizumab if TTP) and otherwise avoid this treatment for non-TTP diagnoses. 2,22,27 Our study is important as it adds to the very small pool of literature available on the Technoscreen assay. 17

| Quality issues
This study identified potential quality issues with the Technoscreen assay. Site A had an unexpectedly high number of low reading test results (0 and 10 IU/dL), which were discordant with quantitative results. Also, test devices from two different lots at Site A yielded different results, an issue that was also observed between test devices from different kits from the same lot with reagents of different ages.
A representative from the local supplier, Helena Laboratories, was informed of these issues and attended Site A to observe the applied assay technique and confirmed that this was consistent with manufacturer recommendations. The storage refrigerator storing the kit was confirmed to be functioning correctly, as was the freezer storing test samples. There was no observable difference between fresh and frozen plasma specimens. This issue remained unresolved by the conclusion of data collection. Although further investigation of this potential issue is indicated, we feel this is consistent with a problematic lot of reagent (SA03C00.01) that somehow passed manufacturer quality control checks and was released for use, but was found to be problematic on laboratory use. This lot was also independently observed to be problematic at site B.

| Accuracy and reliability
The study samples reflected a broad range of ADAMTS13 activity levels, <1 to 150 IU/dL, with a high number of overall samples (n = 23) consistent with TTP (i.e., <10 IU/dL ADAMTS13). Several samples were also derived from patients with moderate reductions in ADAMTS13 levels (Figure 2), reflective of the population from which specimens were drawn, being hospitalized sick patients, with inflammation known to modestly reduce plasma ADAMTS13 levels. 4,5,33 Table 3 summarizes performance characteristics of the Technoscreen assay from this and previous studies. Our study found a high sensitivity (100%) and NPV (100%), which is consistent with previous findings. 17,31,32 These appear to be the most robust attributes of the Technoscreen assay, and is supportive of its proposed role as a screening assay. Notably, specificity and PPV from our study depended on study site, or more accurately on kit lot number used, which was substantially lower in Site A than published studies, but higher in Sites B to D. Close assessment of Study A results ( Figure 2D) 10 Two clinicopathological assessment scores are currently recommended: the French TMA score, with an optimal specificity of 94% if both criteria are met; 10 and the PLASMIC score, with a high score (6,7) suggesting an intermediate likelihood of TTP of 62%-82%. 10 We found Technoscreen has low specificity and PPV for severe ADAMTS13 deficiency when compared to quantitative testing (

| CONCLUSION
There is a well-established need for rapid turn-around ADAMTS13 activity assays to optimize diagnosis or exclusion of TTP in newly presenting TMAs. Where rapid quantitative assays such as the AcuStar assay (<40 min to test result) are available, these should be performed. 10,11,15,30 Our study assessed whether Technoscreen could be alternatively employed in laboratories without access to quantitative assays. An advantage of the screening method is that it does not require any special laboratory equipment, and as a method is much cheaper than setting up an instrument or ELISA method. Conversely, quantitative assays permit quantitation of ADAMTS13 inhibitor titers (useful for management of autoimmune TTP and distinguishing these from congenital TTP), simply by testing a dilution series and using the Bethesda method, 30 and the screening assay cannot provide this information. We also highlight unexpected findings of discrepant results between kits, suggesting underlying quality issues (lot batch differences or potential reagent degradation). Thus, the manufacturer needs to improve internal quality control processes, or to assess for potential conditions that cause reagent degradation, or laboratories need to validate each lot as fit for purpose before using for patient testing. Otherwise, these results support previously published findings that Technoscreen has high sensitivity and NPV, with strong interobserver reliability for positive results (i.e., Technoscreen = 0 IU/dL group), as compared to severe ADAMTS13 deficiency quantified by standard of care testing. Using the Technoscreen assay to rapidly and reliably exclude a TTP diagnosis in patients with clinically intermediate to high likelihood of TTP therefore has potential to reduce use of empiric TPE/caplacizumab and associated risks of iatrogenic harm and resource wastage. As indicated by the manufacturer, all test results suggestive of TTP (i.e., values of 0 IU/dL using Technoscreen), must be verified by quantitative assay. Finally, given test limitations and costs of additional testing, we would suggest use of the Technoclone only in laboratories that do not have easy or timely access to quantitative testing.

AUTHOR CONTRIBUTIONS
All authors contributed to either the study execution, study design, or data analysis. Jared Stephenson and Emmanuel J. Favaloro wrote the original draft of the manuscript, which was then revised according to input from other contributors; all authors approved the manuscript for submission and publication.