A rapid ADAMTS13 activity assay using a portable readout device Authors: Theodora J. Steeghs, Nerissa M. M. Coolen, Elena Krivosheeva, Steven P. L. Kuijpers, Zixuan Li, Jona Merx, Thomas J. Boltje, Nicole M. A. Blijlevens, Saskia E. M. Schols, Cornelis van 't Veer, Waander L. van Heerde Published: 2026 ABSTRACT Key Points•A microfluidic platform simultaneously quantifies ADAMTS13 activity and detects ADAMTS13 antibodies in patients with TTP.•The novel ADAMTS13 activity assay uses a portable readout device that is applicable for near-patient testing. Introduction Thrombotic thrombocytopenic purpura (TTP) is a rare type of thrombotic microangiopathy with a high mortality rate within 24 hours of onset if left untreated.1,2 TTP is caused by severe deficiency of the plasma protein ADAMTS13 (a disintegrin and metalloprotease with thrombospondin type 1, member 13).3 Most commonly, TTP is acquired because of autoantibodies directed to ADAMTS13, known as immune-mediated TTP (iTTP), but low ADAMTS13 levels can also originate from variants in the ADAMTS13 gene, leading to the even rarer congenital-type TTP (cTTP).4 The annual incidence of clinically diagnosed iTTP was estimated to be 2.17 cases per million (95% confidence interval, 2.00-2.34), whereas the prevalence of cTTP has been postulated to account for <5% of all acute cases of TTP.1,5 ADAMTS13 plays a critical role in controlling von Willebrand factor (VWF) multimer size by cleaving VWF multimers in the VWF A2 domain. Normally, the VWF A2 domain is not accessible but is exposed when VWF unrolls, a shear stress-related process. ADAMTS13 binds to the VWF A2 domain, subsequently cleaving at the Y1605-M1606 sequence. Deficiency of ADAMTS13 leads to high molecular weight VWF multimers that promote platelet aggregation, resulting in platelet- and VWF-rich microthrombi, thrombocytopenia, and ischemic organ damage.6 Distinguishing TTP from other thrombotic microangiopathy syndromes, such as severe thrombocytopenia or ischemic disorders, is crucial because patients with TTP are likely to respond to daily plasma exchange (PEX) therapy, whereas those without ADAMTS13 deficiency often require other treatments.5 Although plasma ADAMTS13 activity levels <0.1 IU/mL confirm the diagnosis of TTP, assay results are frequently unavailable within a clinically relevant time frame in centers without specialized laboratory facilities, with turnaround ranging from 24 hours to several days.7,8 Given these barriers and the severity of TTP, clinicians initiate treatment with PEX and immunosuppressive therapy based on clinical characteristics using the PLASMIC score rather than awaiting definitive results.7,9 Although this approach is often lifesaving, it has significant drawbacks, such as unnecessary or suboptimal treatment. New TTP-specific therapies have entered the clinical arena that can only be started when definitive results are available, such as recombinant ADAMTS13 (rADAMTS13), to restore ADAMTS13 activity levels in cTTP, or caplacizumab, a monoclonal antibody against the A1 domain of VWF that blocks platelet binding.7,10, 11, 12 Besides the effect on treatment strategy and patient burden, studies have demonstrated that rapid ADAMTS13 activity testing significantly reduces the economic burden of managing patients with suspected TTP by limiting unnecessary PEX.13,14 The described challenges highlight the need for a rapid, accurate ADAMTS13 near-patient test to improve management of patients with TTP.15,16 To bridge the current gap in clinical need and laboratory testing possibilities, we describe the design, proof of principle, and applicability of a novel luminescence-based ADAMTS13 assay (ADAMTS13lum) in a point-of-care setup to facilitate quick diagnosis and monitoring of patients with suspected TTP. Materials and methods Reagents and plasma samples Glycerol and EDTA were obtained from VWR (Radnor, PA). HEPES [4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid], TCEP [tris(2-carboxyethyl)phosphine], sodium chloride (NaCl), calcium chloride (CaCl2), disodium hydrogen phosphate (Na2HPO4), potassium dihydrogen phosphate (KH2PO4), potassium chloride (KCl), bovine serum albumin (BSA), sucrose, heparin, and absolute ethanol were purchased from Merck (Darmstadt, Germany). Phosphate-buffered saline (10mM Na2HPO4, 1.8mM KH2PO4, 137mM NaCl, and 2.7mM KCl; pH 7.4) was prepared in-house. Furimazine was obtained from Aobious (Gloucester, MA). Coomassie stain and 4% to 20% polyacrylamide gels were obtained from Bio-Rad (Hercules, CA). rADAMTS13 (and TAK755) was supplied by Takeda (Lexington, MA). A patient-derived monoclonal inhibitory ADAMTS13 antibody (immunoglobulin G II-1) was generously supplied by Sanquin (Amsterdam, The Netherlands).17 Citrated normal pooled plasma (NPP) was received from Radboud University Medical Center (Nijmegen, The Netherlands). ADAMTS13-deficient plasma (<0.01 IU/mL ADAMTS13) was obtained from Technoclone (Vienna, Austria). Design, expression, purification, and cleavage of an NLuc-VWFA2 chimera by ADAMTS13 A novel luciferase–VWF A2 domain chimeric protein was designed with a free cysteine at the C-terminus to immobilize the protein onto the microfluidic surface. After cleavage of the VWF A2 part of the chimera by ADAMTS13, the luciferase part is released and can be measured when the supernatant is transported to a luciferase substrate. In detail, the expression vector generated encoded a chimeric protein construct consisting of NanoLuc luciferase (residues 1-170), a spacer followed by the VWF A2 domain of human VWF (residues 1574-1668), a long spacer to extend the A2 domain away from the surface, followed by the polyhistidine sequence for purification, and finally, a free cysteine at the C-terminus for immobilization. Figure 1 visualizes the chimeric protein Nter-Nluc-VWFA2-Cys-Cter using the AlphaFold 3 model.18 A free cysteine in the NanoLuc sequence was mutated (C166F) to prevent disulfide bonds with the C-terminal cysteine and immobilization of the chimera to the surface by the NanoLuc part. Synthesized coding sequences of the different parts were obtained from GenScript and inserted by restriction digestion and ligation into the pET28a(+) vector and subcloned and amplified using Escherichia coli TOP10. The generated plasmid was sequenced to check the appropriate DNA inserts. The correct target vector was transformed into E coli BL21 (DE3), and the chimeric protein was expressed by IPTG (isopropyl β-d-1-thiogalactopyranoside) induction. The expressed Nter-Nluc-VWFA2-Cys-Cter protein was purified by Ni-NTA and size exclusion chromatography and stored in 20mM HEPES, 300mM NaCl, and 2mM TCEP (pH 7.5) at −80°C. Cleavage of the chimera by ADAMTS13 was analyzed by adding 0.1, 1.0, and 10.0 IU/mL rADAMTS13 to the purified Nter-Nluc-VWFA2-Cys-Cter in 25mM HEPES and 125mM NaCl (pH 7.4), and after incubations of 10, 60, and 120 minutes at 37°C, reactions were stopped by adding sodium dodecyl sulfate (SDS) sample buffer. Subsequently, samples were subjected to SDS–polyacrylamide gel electrophoresis (SDS-PAGE) on a 4% to 20% polyacrylamide gel that was Coomassie stained.Figure 1.Design of the luminescent ADAMTS13-specific chimera. For the microfluidic ADAMTS13lum assay, a novel ADAMTS13-specific chimera was designed. The chimera contains the NanoLuc luciferase (yellow) and part of the A2 domain of VWF (green). The VWF A2 domain contains the cleavage site of ADAMTS13 at Y214-M215 (highlighted in red in the AlphaFold images). The C-terminus contains a free cysteine that allows immobilization via thiol linkage. Design of the luminescent ADAMTS13-specific chimera. For the microfluidic ADAMTS13lum assay, a novel ADAMTS13-specific chimera was designed. The chimera contains the NanoLuc luciferase (yellow) and part of the A2 domain of VWF (green). The VWF A2 domain contains the cleavage site of ADAMTS13 at Y214-M215 (highlighted in red in the AlphaFold images). The C-terminus contains a free cysteine that allows immobilization via thiol linkage. Microfluidic assay design and chimera coupling A cyclic olefin copolymer (COC) injection-molded microfluidic cartridge was designed and manufactured by Axxicon. To reduce crossover by light scattering of adjacent chambers, 2.5% carbon was added to the COC. The microfluidic cartridge has 2 designated sections for reagent deposition: the reagent chambers and the detection chambers (Figure 2A). The reagent chambers have an elliptical shape of 1.3 μL in total. Each detection chamber (800 × 800 × 180 μm) contains 110 nL. The COC cartridges were first sprayed with a proprietary antifouling coating with a water droplet static contact angle of ∼55° and an advancing contact angle of ∼70°. Next, maleimide coating was applied in reagent chamber 2 for the coupling of the chimera. Reagent dispensing was done with a SciFlexArrayer S100 from Scienion. The Nter-Nluc-VWFA2-Cys-Cter chimera was diluted in coupling buffer (50mM Na2HPO4, 75mM NaCl, 5mM EDTA, and 1.4% glycerol; pH 7.2) before printing to achieve a stock concentration of 70μM. The SciFlexArrayer dispensed 2 droplets of 300 pL in 147 locations (88.2 nL total volume) in the reagent chamber with maleimide coating, and subsequently, the cartridges were washed with phosphate-buffered saline and Milli-Q water. After drying, 300 pL of 10% sucrose, 0.5% BSA, 25mM HEPES, and 125mM NaCl (pH 7.4) was placed on the spots with immobilized Nter-Nluc-VWFA2-Cys-Cter to protect the chimera during storage. Finally, furimazine, dissolved in 4.5% ethanol, 10% sucrose, 0.5% BSA, 25mM HEPES, and 125mM NaCl (pH 7.4), was printed with the SciFlexArrayer at 1 location in the detection chambers to reach a final furimazine concentration of 60 μg/mL when the 110 nL detection chambers were filled. The cartridges were dried for 2 hours (30°C; <5% relative humidity), and subsequently, the cartridges were laminated with a custom-made 25-μm thick clear polyethylene adhesive foil (Adhesive Research) to create the microfluidic channels and chambers. The final product was placed in a pouch with a silica bag before storage at 4 to 8°C.Figure 2.Microfluidic ADAMTS13lum assay principle. (A) The design of the microfluidic cartridge with 4 sample lines. (B) A schematic overview demonstrates the Nter-Nluc-VWFA2-Cys-Cter chimera bound in the bottom spotting chambers and the furimazine substrate placed in the detection chambers. In the presence of ADAMTS13, Nter-Nluc-VWFA2-Cys-Cter is cleaved, and NanoLuc luciferase is released. The migration of the fluids through the microfluidics with capillary force results in the transfer of the sample, including the released luciferase, to the detection chambers with the furimazine substrate, resulting in the production of photons. Figure created with biorender.com. Steeghs D. (2026) https://biorender.com/v85o0n6. Microfluidic ADAMTS13lum assay principle. (A) The design of the microfluidic cartridge with 4 sample lines. (B) A schematic overview demonstrates the Nter-Nluc-VWFA2-Cys-Cter chimera bound in the bottom spotting chambers and the furimazine substrate placed in the detection chambers. In the presence of ADAMTS13, Nter-Nluc-VWFA2-Cys-Cter is cleaved, and NanoLuc luciferase is released. The migration of the fluids through the microfluidics with capillary force results in the transfer of the sample, including the released luciferase, to the detection chambers with the furimazine substrate, resulting in the production of photons. Figure created with biorender.com. Steeghs D. (2026) https://biorender.com/v85o0n6. The microfluidic measurements were performed using a prototype readout device. The readout device has been described in detail previously.19 Briefly, plasma samples were diluted fivefold with buffer containing 17 IU/mL heparin, 5mM calcium, 0.5% BSA, 25mM HEPES, and 125mM NaCl (pH 7.4) and applied to the microfluidic sample inlet to fill the reagent chambers. After reaching the stop valve located at the end of reagent chamber 2 that contained the immobilized Nter-Nluc-VWFA2-Cys-Cter, an incubation step of 20 minutes was introduced, during which the ADAMTS13 in the sample was allowed to cleave the chimera to release the NLuc luciferase. Then, by using the trigger line, the stop valve was broken, and the released luciferase was transported by flow to the final detection chambers, which contain the furimazine substrate. The generated photons were counted using the prototype readout device for 2 minutes to quantify the amount of released luciferase (Figure 2B). The data were transmitted via USB to a computer with the interface software to visualize the signals in real time. The data were also transferred to a CSV file. Preparation of plasma samples Specificity and sensitivity of the designed ADAMTS13lum assay were tested in duplicate with ADAMTS13-deficient plasma spiked with different ADAMTS13 concentrations in the range of 0 to 0.88 IU/mL ADAMTS13 activity. The ADAMTS13 activity levels in the spiked plasma samples were determined using the reference method Technozym ADAMTS-13 Activity kit (Technoclone), according to the manufacturer’s protocol. To determine sensitivity to inhibitors, NPP was spiked with an inhibitory ADAMTS13 antibody (0.03nM, 0.3nM, 3nM, 30nM, and 300nM) and tested in triplicate with the microfluidic ADAMTS13lum. Cartridge setup for cTTP or iTTP discrimination using the ADAMTS13lum assay Microfluidic cartridges were prepared to detect inhibitory antibodies against ADAMTS13, in parallel with quantifying the ADAMTS13 activity levels in plasma samples. Besides the standard preparation of the microfluidic cartridge, rADAMTS13 was printed in the first reagent chamber (Figure 2). The rADAMTS13 was diluted to 30 IU/mL in printing buffer containing 10% sucrose, 0.5% BSA, 25mM HEPES, and 125mM NaCl (pH 7.4). The SciFlexArrayer dispensed 14 droplets of rADAMTS13 in the first reagent chamber of 1 channel (4.2 nL) to obtain a final concentration of 0.1 IU/mL rADAMTS13 after filling the cartridge. These cartridges were tested in duplicate with ADAMTS13-deficient plasma, and NPP spiked with 30nM anti-ADAMTS13 antibody mimicking cTTP and iTTP, respectively. Comparison of the ADAMTS13lum assay with commercially available diagnostic ADAMTS13 assays To determine the performance of the microfluidic ADAMTS13lum, it was compared with 2 commercially available assays commonly used for TTP diagnosis: Technozym ADAMTS-13 Activity kit (hereafter, Technozym ELISA) read on a SpectraMax M3 (Molecular Devices), and the automated Technofluor ADAMTS-13 Activity kit (hereafter, Ceveron FRET) run on Ceveron S100 (Technoclone). The reference methods were performed according to the manufacturers’ package insert protocols. Citrated plasma samples from patients with TTP (n = 39) from the UK TTP registry biobank (approval number 08/H0810/54) were received from University College London Hospital (London, United Kingdom). These plasma samples originated from patients with either confirmed non-TTP thrombotic microangiopathy or diagnosed iTTP or cTTP on treatment or in remission. The microfluidic ADAMTS13lum was calibrated by measuring 2 internal controls in parallel in the same cartridge (see formula 1). These controls were ADAMTS13-deficient plasma spiked without (RLU0) or with (RLU1) 1 IU/mL rADAMTS13.Formula1:ADAMTS13lumactivity(IU/mL)=RLUsample−RLU0RLU1−RLU0 The plasma samples from patients with TTP were tested in duplicate under blinded conditions, meaning no ADAMTS13 activity levels were known before the method comparison. For each sample, all 3 assays were performed on the same day in the same laboratory. Statistical analysis For each microfluidic sample line, the average of the photon counts per second in the 4 detection chambers was plotted against time. Detection chambers were excluded in case of an observable error in filling behavior. The maximal signal output in relative light units (RLU) per second (maximum RLU/s) was the main assay outcome. The calibration curve was prepared using simple linear regression. The quantitative level of agreement between the Technozym ELISA, Ceveron FRET, and ADAMTS13lum was calculated using the Spearman correlation coefficient. The qualitative level of agreement between the methods was calculated using linear weighted Cohen kappa. Quantitative statistical analysis was performed using GraphPad Prism version 10.1.0; qualitative statistical analysis and raw data output were analyzed using Excel. Results Design of an immobilized luciferase-VWFA2 chimera for a luminescent ADAMTS13 assay To develop the microfluidic ADAMTS13lum assay, a novel chimeric protein was expressed in E coli and purified. This chimeric protein, called Nter-Nluc-VWFA2-Cys-Cter, contains a modified NanoLuc protein (Figure 1, yellow), a part of the A2 domain of VWF (Figure 1, green), and 2 linkers. Figure 3 shows the purified Nter-Nluc-VWFA2-Cys-Cter at the calculated molecular weight of 40 kDa. Incubation of the chimera with rADAMTS13 resulted in the expected bands at 24 kDa, corresponding to NLuc, and 16 kDa, corresponding to the cleaved VWF A2, in a time- and concentration-dependent manner (Figure 3). These results reveal that ADAMTS13 successfully recognized and cleaved the A2 domain present in the chimera, demonstrating the applicability of the chimera for detecting ADAMTS13 activity.Figure 3.Cleavage of the NLuc-VWFA2 chimera by ADAMTS13. SDS-PAGE shows the intact Nter-Nluc-VWFA2-Cys-Cter substrate at 40 kDa. Adding increasing levels of ADAMTS13 results in the formation of the cleaved products at 24 kDa (NLuc) and 16 kDa (cleaved VWF A2). The cleavage by ADAMTS13 is dependent on ADAMTS13 levels and incubation time. Cleavage of the NLuc-VWFA2 chimera by ADAMTS13. SDS-PAGE shows the intact Nter-Nluc-VWFA2-Cys-Cter substrate at 40 kDa. Adding increasing levels of ADAMTS13 results in the formation of the cleaved products at 24 kDa (NLuc) and 16 kDa (cleaved VWF A2). The cleavage by ADAMTS13 is dependent on ADAMTS13 levels and incubation time. ADAMTS13lum assay sensitivity and specificity in a microfluidic cartridge The Nter-Nluc-VWFA2-Cys-Cter chimera, immobilized in the microfluidic cartridge, was incubated with different concentrations of rADAMTS13 spiked in ADAMTS13-deficient plasma. After cleavage by ADAMTS13, the luciferase was released in the supernatant and transferred to the detection chambers with furimazine, resulting in luminescent signals. Figure 4A shows that increasing the ADAMTS13 concentration corresponds to increased levels of luciferase in the supernatant, resulting in higher luminescent signals in RLU/s. ADAMTS13-deficient plasma samples showed a max signal of 2663 RLU/s, which was set as the background signal. Spiking deficient plasma with 0.05 IU/mL rADAMTS13 resulted in an increase of the max signal to 3391 RLU/s. Further increasing the ADAMTS13 levels up to 0.88 IU/mL resulted in a maximum signal of 15 586 RLU/s and demonstrated a linear correlation between the ADAMTS13 activity levels and the maximum RLU/s with an R2 value of 0.98 (Figure 4B).Figure 4.ADAMTS13lum assay sensitivity in a microfluidic cartridge. Six samples with different ADAMTS13 levels up to 0.88 IU/mL were tested in the microfluidic cartridge in duplicate. (A) Real-time photon count per second is shown. (B) The correlation between the ADAMTS13 activity levels and maximal signal output in relative light units per second signals is demonstrated with simple linear regression analysis. ADAMTS13lum assay sensitivity in a microfluidic cartridge. Six samples with different ADAMTS13 levels up to 0.88 IU/mL were tested in the microfluidic cartridge in duplicate. (A) Real-time photon count per second is shown. (B) The correlation between the ADAMTS13 activity levels and maximal signal output in relative light units per second signals is demonstrated with simple linear regression analysis. Detection of inhibitory ADAMTS13 antibody by printed ADAMTS13 in the ADAMTS13lum assay To verify that our system is also applicable for iTTP, NPP was spiked with an inhibitory ADAMTS13 antibody. Figure 5A shows clear inhibition of signals with increasing levels of anti-ADAMTS13. The 50% inhibitory concentration was determined to be 3.7nM, and ∼20% residual ADAMTS13 activity was observed at 30nM, both consistent with a previous study and the ADAMTS13 plasma concentration of 3.5nM to 7.0nM.17,20 The multichannel microfluidic cartridge design allows parallel testing of samples. By printing rADAMTS13 in a reagent chamber of 1 channel of the cartridge, an inhibitor can be detected. As shown in Figure 5B, channels either without or with rADAMTS13 printed in the microfluidic chambers (reagent chamber 1) were compared. The ADAMTS13-deficient plasma sample simulated a sample from a patient with cTTP, whereas an NPP sample spiked with an anti-ADAMTS13 antibody simulated a sample from a patient with iTTP. The additional rADAMTS13 was detected only in ADAMTS13-deficient plasma but not in NPP with inhibitory ADAMTS13 antibodies (Figure 5B), indicating the sensitivity of the assay for inhibitors against ADAMTS13.Figure 5.Detecting inhibitory ADAMTS13 antibodies using the ADAMTS13lum assay. (A) The most prevalent form of TTP is immune-mediated ADAMTS13 deficiency. By spiking increasing levels of an inhibitory ADAMTS13 antibody to NPP, the sensitivity of the microfluidic ADAMTS13lum assay was tested. The residual activity was established by the inhibition of the RLU/s signal of the samples spiked with inhibitors compared to the NPP without inhibitors. A characteristic sigmoidal inhibition curve is shown. (B) The parallel microfluidic lines allow for simultaneous detection of inhibitory ADAMTS13 antibodies. To detect the presence of inhibitors in the plasma samples, the first reagent chamber was printed without or with rADAMTS13. When adding ADAMTS13-deficient plasma, the addition of rADAMTS13 in the first reagent chamber resulted in increased residual ADAMTS13 activity compared to the sample lines without printed rADAMTS13. When NPP was spiked with 30nM inhibitory ADAMTS13 antibodies, the printed rADAMTS13 did not result in increased residual ADAMTS13 activity because of the inhibition. Detecting inhibitory ADAMTS13 antibodies using the ADAMTS13lum assay. (A) The most prevalent form of TTP is immune-mediated ADAMTS13 deficiency. By spiking increasing levels of an inhibitory ADAMTS13 antibody to NPP, the sensitivity of the microfluidic ADAMTS13lum assay was tested. The residual activity was established by the inhibition of the RLU/s signal of the samples spiked with inhibitors compared to the NPP without inhibitors. A characteristic sigmoidal inhibition curve is shown. (B) The parallel microfluidic lines allow for simultaneous detection of inhibitory ADAMTS13 antibodies. To detect the presence of inhibitors in the plasma samples, the first reagent chamber was printed without or with rADAMTS13. When adding ADAMTS13-deficient plasma, the addition of rADAMTS13 in the first reagent chamber resulted in increased residual ADAMTS13 activity compared to the sample lines without printed rADAMTS13. When NPP was spiked with 30nM inhibitory ADAMTS13 antibodies, the printed rADAMTS13 did not result in increased residual ADAMTS13 activity because of the inhibition. Microfluidic ADAMTS13lum assay compared to commercially available diagnostic ADAMTS13 activity assays To evaluate the performance of the microfluidic ADAMTS13lum against reference assays, plasma samples from patients (n = 39) with confirmed non-TTP thrombotic microangiopathy or from patients diagnosed with iTTP or cTTP on treatment or in remission were tested in duplicate under blinded conditions in the ADAMTS13lum assay and the reference assays Technozym ELISA and Ceveron FRET. Figure 6A-C shows the correlation between the 3 assays. The microfluidic ADAMTS13lum showed an excellent correlation with the Ceveron FRET (r = 0.90) and the Technozym ELISA (r = 0.88). According to the International Society on Thrombosis and Haemostasis (ISTH) guidelines,21 ADAMTS13 levels can be subdivided into 3 categories: <0.1 IU/mL, ADAMTS13 activity is positive for TTP; 0.1 to 0.2 IU/mL, ADAMTS13 activity is borderline, and >0.2 IU/mL is negative for TTP. To assess agreement between the ADAMTS13lum and the reference methods concerning categorizing the samples, a linear weighted Cohen kappa was calculated. The calculated Cohen kappa value was similar for the reference methods (0.762) and for the ADAMTS13lum compared with the Ceveron FRET (0.703) or the Technozym ELISA (0.657), demonstrating similar qualitative performance. Figure 6D-F visualizes the agreement between the methods per category. In 3 cases, there was a major disagreement between the 3 methods. Two of these samples originated from patients with iTTP receiving PEX. One was reported as <0.1 IU/mL by the Technozym ELISA (0.09 IU/mL), whereas the Ceveron FRET reported ADAMTS13 activity between 0.1 and 0.2 IU/mL (0.18 IU/mL), and the ADAMTS13lum reported >0.2 IU/mL (0.23 IU/mL). The other iTTP sample was reported to be <0.1 IU/mL by the ADAMTS13lum (0.06 IU/mL), whereas it was determined to be >0.2 IU/mL by the Ceveron FRET (0.32 IU/mL) and the Technozym ELISA (0.24 IU/mL). The third sample originated from a treated patient with cTTP, where the Technozym ELISA reported <0.1 IU/mL ADAMTS13 activity (0.09 IU/mL) compared with the >0.2 IU/mL reported by the ADAMTS13lum (0.21 IU/mL) and the Ceveron FRET (0.24 IU/mL). Although these differences in ADAMTS13 activity resulted in variable categorization, the absolute difference between the reported ADAMTS13 activity levels was close to the borders of the respective category. Furthermore, disagreement between the 3 methods in nonborderline cases was rare. With this patient data set, it was demonstrated that the microfluidic ADAMTS13lum level shows results comparable to those of reference methods in plasma samples.Figure 6.ADAMTS13lum assay method comparison with plasma samples from patients with TTP. The method comparison with reference ADAMTS13 assays was done with 39 plasma samples from patients with TTP from University College London Hospital. The plasma samples were measured in duplicate and in parallel with the microfluidic ADAMTS13lum and 2 reference assays. The correlation between the 3 systems, namely, ADAMTS13lum vs Ceveron FRET (A), ADAMTS13lum vs Technozym ELISA (B), and Technozym ELISA vs Ceveron FRET (C), showed high levels of agreement for the quantitative data. Based on the calculated ADAMTS13 levels, the samples can be divided into TTP categories. Therefore, the level of agreement for qualitative data is established for ADAMTS13lum vs Ceveron FRET (D), ADAMTS13lum vs Technozym ELISA (E), and Technozym ELISA vs Ceveron FRET (F). The classification is indicated on the x-axis for the reference methods and compared with the indicated test with a color-code legend. ADAMTS13lum assay method comparison with plasma samples from patients with TTP. The method comparison with reference ADAMTS13 assays was done with 39 plasma samples from patients with TTP from University College London Hospital. The plasma samples were measured in duplicate and in parallel with the microfluidic ADAMTS13lum and 2 reference assays. The correlation between the 3 systems, namely, ADAMTS13lum vs Ceveron FRET (A), ADAMTS13lum vs Technozym ELISA (B), and Technozym ELISA vs Ceveron FRET (C), showed high levels of agreement for the quantitative data. Based on the calculated ADAMTS13 levels, the samples can be divided into TTP categories. Therefore, the level of agreement for qualitative data is established for ADAMTS13lum vs Ceveron FRET (D), ADAMTS13lum vs Technozym ELISA (E), and Technozym ELISA vs Ceveron FRET (F). The classification is indicated on the x-axis for the reference methods and compared with the indicated test with a color-code legend. Discussion This study presents proof of principle of a new microfluidic assay to quantify ADAMTS13 activity levels in plasma samples using a previously described detection platform based on luminescence. The assay uses a novel luminescent chimera Nter-Nluc-VWFA2-Cys-Cter that is recognized and cleaved by ADAMTS13 when immobilized on the cartridge surface, even in a plasma environment. The released luciferase detected in the system is directly related to the ADAMTS13 activity level in the plasma. To distinguish iTTP from cTTP, we furthermore showed the sensitivity of the ADAMTS13lum assay for detecting inhibitory antibodies against ADAMTS13. Although the ADAMTS13 activity level is the most relevant marker to diagnose and initiate appropriate treatment in TTP,5 distinguishing between iTTP and cTTP in early stages has important treatment advantages. For patients with TTP, early treatment is crucial because approximately half of all deaths occur within 24 hours of onset.1 Because of the severity of the disease, suspected TTP is usually followed up with immediate intervention while awaiting definitive diagnostic results. Standard treatment consists of daily PEX to replenish ADAMTS13 and the removal of large VWF multimers, in combination with immunosuppressive therapy (eg, rituximab and corticosteroids) to suppress the level of inhibitory autoantibodies in patients with iTTP.22 Several days of PEX can be unnecessarily given to patients who are eventually diagnosed as negative for TTP.23,24 A method comparison study was conducted to compare our microfluidic ADAMTS13lum assay with current diagnostic ADAMTS13 activity assays. There are several standard assays for measuring ADAMTS13 activity; most of them involve a recombinant VWF peptide substrate to measure the activity of ADAMTS13.25,26 In 2004, Kokame et al27 demonstrated that the minimal region of the VWF A2 domain required for an ADAMTS13 substrate should consist of the 73 residues D1596 to R1668 of the VWF amino acid sequence, known as the VWF73 peptide. Subsequently, the same group developed an ADAMTS13 activity assay based on the cleavage of the VWF73 peptide monitored by FRET, which remains the gold standard for ADAMTS13 activity measurement.28 Other ADAMTS13 activity assays have been developed using a similar approach or by employing an antibody that recognizes the neoepitope generated after cleavage of the Y1605-M1606 bond in the peptide.29,30 We compared the ADAMTS13lum assay with 2 standard ADAMTS13 activity assays, the Ceveron FRET and the Technozym ELISA. Previous studies already reported a high level of agreement between these 2 standard ADAMTS13 activity assays and the gold standard, indicating that these assays are suitable as references for method comparison studies.26,31 In our study, the correlation between the 2 commercial methods was determined as r = 0.97, which is the same as previously reported by Singh et al26 (r = 0.97). The microfluidic ADAMTS13lum assay also displayed high correlation with both reference methods, demonstrating that the platform is able to accurately quantify ADAMTS13 activity levels in a broad spectrum of ADAMTS13 concentrations. In addition, the qualitative comparison of the diagnostic categories showed a substantial level of agreement (0.6-0.8) according to the Cohen kappa values. The kappa value between the reference methods was lower than described in the recently published review,32 which might be because of the high number of borderline samples included in our study. In 3 cases, there were inconsistencies with the categorization of the sample between the 3 methods. Although different categories were assigned based on the ADAMTS13 activity level, the absolute differences were close to the borders of the respective categories. This was also visualized using Bland-Altman analysis, demonstrating the relative position of these 3 samples to the category borders (supplemental Figure 1). Most importantly, the agreement of the microfluidic ADAMTS13lum assay was comparable to the agreement between the reference tests for the sample classification. Of all TTP cases, >95% are caused by the presence of inhibitory antibodies against ADAMTS13, and the remaining 5% account for cTTP.33 The detection of inhibitory antibodies in the ADAMTS13lum can be established by printing rADAMTS13 in the microfluidic reagent chambers. Suppressed luminescent signals from the channels that contain rADAMTS13 are indicative of the presence of an inhibitor. During this study, proof of concept was demonstrated, although no analytical performance, such as limit of detection for the antibodies, was established yet. In addition, it would be interesting to investigate how other types of antibodies, including low and noninhibitory antibodies, affect the ADAMTS13lum assay. At this stage, the ADAMTS13lum can be incorporated into a laboratory workflow but cannot be used for near-patient testing. To make this platform suitable for such an application, the system should be free from manual pipetting steps. Incorporation of automatic trigger lines and application of whole blood and plasma separation by plasma filters are currently under investigation.34 Such automation of the microfluidic ADAMTS13lum assay will allow near-patient testing, which will reduce the turnaround time, contributing to reduced patient burden as well as reduced economic burden. Rapid ADAMTS13 testing in the first sample of suspicion is reported to be cost-effective because of reduced therapeutic PEX.13,14,35,36 Connel et al37 and McGinnis et al38 even demonstrated a significant reduction in plasma use at tertiary care centers after introducing rapid ADAMTS13 testing. Besides the financial aspect, delayed ADAMTS13 testing may cause a suboptimal treatment strategy in patients without TTP or a delay in optimal treatment in patients with confirmed TTP. Advancements in TTP-specific therapies, such as rADAMTS13 (Adzyma; Takeda) or caplacizumab,10, 11, 12 can only be started once definitive results are available.39 According to the ISTH guidelines, experienced clinicians can initiate caplacizumab based on clinical prediction scores, such as the PLASMIC or French score, in combination with a high clinical suspicion of TTP, without waiting for the results of ADAMTS13 activity testing.7 In addition, it would be interesting to investigate whether capillary blood is compatible with this platform to further simplify the test and allow rapid diagnosis that can be completely performed by laypersons. The analytical turnaround time of the microfluidic ADAMTS13 assay is ∼30 minutes, which is similar to that of the Ceveron FRET (25 minutes).32 Moreover, our microfluidic setup approach allows the detection of ADAMTS13 activity, anti-ADAMTS13 antibodies, and internal controls in a single cartridge. In conclusion, the novel portable setup establishes the capability of a rapid miniaturized ADAMTS13lum assay for diagnosis and follow-up of TTP with a high level of agreement with commercially available ADAMTS13 activity assays. The system even allows distinction between cTTP and iTTP. Future research will focus on optimizing the setup for clinical applications performed by health care workers and even laypersons. Conflict-of-interest disclosure: The authors declare the following financial interests/personal relationships that may be considered as potential competing interests. W.L.v.H. reports financial support from Takeda Development Center Americas, Inc. T.J.S., N.M.M.C., E.K., S.P.L.K., and C.v.V. report employment with Enzyre. W.L.v.H. reports board membership and employment with Enzyre. T.J.S., W.L.v.H., and C.v.V. are inventors on patents WO2024153750A1 and WO2023187005A1 issued to Enzyre. Z.L., J.M., and T.J.B. are inventors on patent WO2024153750A1. The remaining authors declare no competing financial interests.