Performance Assessment of a Novel Multianalyte Methodology for Celiac Disease Biomarker Detection and Evaluation of the Serology-Alone Criteria for Biopsy-Free Diagnosis

A total of 703 adult and pediatric patients whose serum samples were submitted to our laboratory for CD serology were included in the study. Some serum samples were tested for more than one CD antibody. Briefly, 407 specimens were tested for anti–tTG IgA, 316 specimens were tested for anti–tTG IgG, 306 specimens were tested for anti–DGP IgA, and 306 specimens were tested for anti–DGP IgG test on QUANTA Lite (Inova Diagnostics, San Diego, California) enzyme-linked immunosorbent assay (ELISA). Patient samples were also tested using Aptiva IgA and IgG reagents (Inova Diagnostics) for all four analytes (anti–tTG IgA, anti–tTG IgG, anti–DGP IgA, and anti–DGP IgG). Of the 703 patients, we had access to electronic patient information of 142 patients. A retrospective chart review was conducted to explore clinical history and review clinicians’ notes. Fifteen patients were excluded based on the following criteria: 9 patients were diagnosed with CD without biopsy confirmation, and 6 patients formerly diagnosed with CD were on gluten-free diet at the time of serology testing. Of the 127 patients included, 58 were biopsy-proven CD patients, and in 69 patients the presence of CD was excluded per the clinicians’ notes (Figure 1). Chart review to evaluate the reasons for ordering CD antibody tests on the negative patients revealed symptoms as listed in Supplemental Table 1 (see the supplemental digital content containing 2 tables and 2 figures at https://meridian.allenpress.com/aplm in the December 2023 table of contents). All samples were collected, handled, and deidentified in accordance with the approved protocol from University of Utah Institutional Review Board.

Two serum aliquots from each patient were stored at −80°C until use, and then tested for anti–tTG IgA, anti–tTG IgG, anti–DGP IgA, and anti–DGP IgG using the QUANTA Lite ELISA and Aptiva particle-based multianalyte technology.

In the particle-based multianalyte technology, analytes coupled to paramagnetic microparticles carry a unique signature, which allows for their classification by an optical module. Particles are incubated with diluted patient samples, undergo a wash cycle, and then incubated with anti–human IgA or anti–human IgG conjugated to phycoerythrin. After another washing cycle is complete, particles are analyzed through digital imaging technology installed in the platform.²⁰  QUANTA Lite assays were performed according to manufacturer’s package insert. Briefly, antigen is immobilized to the surface of the microwell plate. Diluted patient samples and controls are added to specific wells and incubation allows the patients’ antibodies to bind to the antigen. A wash step removes unbound antibodies, and an enzyme-labeled secondary antibody is added to each well. A second incubation allows the enzyme-linked antibodies to bind to the patient’s antibodies. After another wash cycle, appropriate substrate is added to the wells. Antibody detection is accomplished by measuring the enzyme activity, which develops measurable color intensity, using a spectrophotometer. Antibody presence is evaluated by comparing the optical density developed in the patient wells with control wells.

For particle-based multianalyte technology and ELISA testing, cutoff values as per manufacturer recommendations were adopted, namely ≥5 fluorescent light units (FLU) for all Aptiva assays and ≥4 U/mL for QUANTA Lite tTG IgA, ≥6 U/mL for QUANTA Lite tTG IgG, and, lastly, ≥20 U/mL for both QUANTA Lite DGP IgG and IgA.

In addition to the analyte-specific microparticles, Aptiva system contains an internal control microparticle that is coated with polyclonal goat anti–human IgA antibodies. IgA in the sample binds to the beads and mean fluorescence intensity (MFI) values proportional to the antibody concentration are generated. In the event of low or undetectable IgA levels (<2000 MFI), the system reports a flag and does not produce anti–DGP IgA or anti–tTG IgA results. This serves as a built-in internal control to capture IgA-deficient patient sera, a common occurrence in CD patients.²¹  The IgA detection using internal control microparticles on the Aptiva system was compared to our current IgA measurement using Optilite (The Binding Site, Birmingham, United Kingdom), a turbidimetric analyzer. In our laboratory, we use a reflex approach for CD testing. If the patient sample is IgA-deficient (<2 mg/dL), then the sample is reflexed to anti–tTG IgG and anti–DGP IgG testing. But if the sample has low IgA values according to age-specific reference ranges, but above the detection limit, then the sample is tested for anti–tTG IgA, anti–tTG IgG, anti–DGP IgA, and anti–DGP IgG analytes. To verify the ability of Aptiva to flag only IgA-deficient samples and not samples with low and detectable levels of IgA, residual sera from 100 samples submitted for CD testing were tested for total IgA in house using Optilite and compared to MFI results obtained from the Aptiva platform.

Clinical performance for each analyte (anti–tTG IgA, anti–tTG IgG, anti–DGP IgA, and anti–DGP IgG) was assessed using information gathered from medical chart review. Fifty-eight CD–confirmed positive samples were used to evaluate clinical sensitivity and 69 disease control samples were used to evaluate clinical specificity. Positive predictive value (PPV), negative predictive value (NPV), positive likelihood ratio (LR), negative LR, and odds ratio were also calculated from these clinical data sets. A receiver operating characteristic (ROC) curve was generated using Analyse-it Software, version 5.92 (Leeds, United Kingdom). This was further used to assess area under the curve (AUC) and optimal cutoffs for each CD biomarker. Youden J statistics were used to select optimal cutoffs with the best balance of sensitivity and specificity. To compare which diagnostic tests improved the outcome to a statistically significant degree between the 2 methods, the McNemar test for significant proportion changes was applied. Percent-agreement analysis and 95% CIs were calculated using Microsoft Excel version 5.66 (Microsoft Corp, Redmond, Washington); Spearman correlation coefficients (ρ and P values) were generated using GraphPad Prism V.7 (GraphPad Software, San Diego, California); the McNemar test was applied using the function mcnemar.test in the base package of R²² ; UpSet plots and hierarchical clustering dendrogram were generated using the following: Python, version 3.8; Numpy, version 1.19.1; Pandas, version 1.2.5; Matplotlib, version 3.3.2; Seaborn, version 0.11.1 (Werfen, San Diego, California).

Total percent agreement (TPA), positive percent agreement (PPA), and negative percent agreement (NPA) between ELISA QUANTA Lite and Aptiva methods were determined. Anti–tTG IgA showed a TPA of 95% (PPA: 92%, NPA: 99%), anti–tTG IgG displayed a TPA of 71% (PPA: 41%, NPA: 99%), anti–DGP IgA demonstrated a TPA of 81% (PPA: 62%, NPA: 98%), and anti–DGP IgG displayed TPA of 91% (PPA: 85%, NPA: 97%) between the 2 methods (Table 1) with Spearman ρ values ranging from 0.60 to 0.84 (P < .001). Supplemental Figure 1 shows correlations displayed in hierarchical clusters along with the corresponding correlation coefficients from Spearman analysis based on comparisons between analytes tested on both methods.

All 96 samples with sufficient IgA levels assessed using Optilite generated high MFI values (range, 5460–10 532) on Aptiva IgA control bead (Supplemental Table 2). Two samples that were confirmed IgA-deficient on Optilite also resulted in lower MFI values (<2000) on Aptiva and were flagged. Another 2 samples were IgA-low when compared with the age-specific normal range but above the limit of detection on Optilite. One of these samples was flagged by Aptiva and did not produce anti–tTG IgA and anti–DGP IgA results. The second sample resulted in an MFI value above the 2000 cutoff on Aptiva and produced anti–tTG IgA and anti–DGP IgA results as expected. Of note, samples with the lowest IgA concentrations (1 mg/dL, 1 mg/dL, 5 mg/dL, and 48 mg/dL) on Optilite displayed the lowest 4 MFI values on Aptiva (32 MFI, 48 MFI, 4134 MFI, and 1914 MFI, respectively).

To assess the overall diagnostic accuracy of both methodologies, ROC curve analysis was performed using the presence or absence of CD as binary classifier. Overall high accuracy was observed for all 4 analytes on QUANTA Lite and Aptiva, with an AUC (95% CI) that equaled to 0.897 (0.837–0.956) for anti–DGP IgA, Aptiva, and the highest acquired that equaled to 0.990 (0.975–1.004) for anti-DGP IgG, QUANTA Lite (Figure 2, A through D). Aptiva anti–tTG IgA showed a higher AUC (95% CI) of 0.960 (0.924–0.996) compared with QUANTA Lite anti–tTG IgA, which showed an AUC of 0.931 (0.867–0.995). Also, Aptiva anti–tTG IgG showed a higher AUC of 0.919 (0.851–0.986) on Aptiva compared with QUANTA Lite, which had an AUC of 0.906 (0.854–0.958).

At the manufacturer-suggested cutoffs, clinical sensitivity, specificity, predictive values, and LRs were calculated for all 4 antibodies using clinically defined samples (n = 127) (Table 2). Anti–tTG IgA showed equal clinical sensitivity (91%; 95% CI: 86%–96%) and specificity (99%; 95% CI: 96%–100%) on QUANTA Lite and Aptiva. For anti–tTG IgG, sensitivities were comparable: 72% (95% CI: 65%–80%) and 69% (95% CI: 61%–77%) on QUANTA Lite and Aptiva, respectively, and higher specificity (95% CI) was observed on Aptiva (100%, 95% CI: 98%–100%) compared with QUANTA Lite (94%, 95% CI: 90%–98%). Anti–DGP IgA displayed lower sensitivity on Aptiva (64%, 95% CI: 55%–72%) compared with QUANTA Lite (83%, 95% CI: 76%–89%); similar specificities of 97% (95% CI: 94%–100%) and 99% (95% CI: 96%–100%) were observed on QUANTA Lite and Aptiva, respectively. Anti–DGP IgG showed greater sensitivity (90%; 95% CI: 84%–95%) on Aptiva and higher specificity (99%; 95% CI 96%–100%) on QUANTA Lite. The PPVs and NPVs are summarized in Table 2. Except for anti–DGP IgG, Aptiva displayed a generally higher specificity versus QUANTA Lite. PPVs on both methods ranged from 91% to 98%, NPVs were higher for anti–tTG IgA (93%, QUANTA Lite and Aptiva) and anti–DGP IgG (86%, QUANTA Lite; 92%, Aptiva), and lower for anti–tTG IgG (80%, QUANTA Lite; 79%, Aptiva). Anti–DGP IgA demonstrated 87% and 79% NPVs for QUANTA-Lite and Aptiva, respectively. Using the Aptiva system, the anti–tTG IgA assay displayed a high positive LR (95% CI) of 63.1 (52.2–73.9) and a negative LR (95% CI) of 0.1 (0.0–0.1), anti–DGP IgG showed a positive LR (95% CI) of 15.5 (12.9–18.1) and negative LR (95% CI) of 0.1 (0.1–0.2), anti–DGP IgA showed a positive LR (95% CI) of 44.0 (36.4–51.6) and negative LR (95% CI) of 0.4 (0.3–0.5), and anti–tTG IgG showed an infinite positive LR (could not be calculated due to a specificity of 100%) and negative LR (95% CI) of 0.3 (0.2–0.4). Except for anti–tTG IgG, QUANTA Lite assays also demonstrated similar likelihood and odds ratios, as listed in Table 2. Applying the McNemar test for significant proportion, the outcome of the diagnostic tests between the 2 methods showed a statistically significant difference for anti–DGP IgA (P = .01) and anti–DGP IgG (P = .04); however, anti–tTG IgG (P = .11) and anti-tTG IgG (P > .99) remained comparable.

Table 3 shows the sensitivity, specificity, and predictive values at the manufacturer-suggested and optimal cutoffs. The optimal cutoffs were points with the best balance of sensitivity and specificity derived from ROC analysis. Both the manufacturer-suggested (≥5 FLU) and the optimal calculated cutoffs (≥3 FLU) resulted in comparable sensitivity and specificity for the anti–tTG IgA. Applying the optimal cutoff (≥1 FLU) to anti–tTG IgG showed an increase in sensitivity (from 69% to 90%) and NPV (from 79% to 92%), while specificity (from 100% to 96%) and PPV (from 100% to 95%) showed a slight decrease. For anti–DGP IgA, there was a 19% increase in sensitivity (from 64% to 83%) and a 10% increase in NPV (from 76% to 86%) when using the optimal cutoff (≥2.6 FLU). However, the change in cutoff greatly compromised the specificity (from 99% to 88%) and PPV (from 97% to 86%). Using the optimal cutoff (≥4.5 FLU) for anti–DGP IgG resulted in a similar overall performance compared to the manufacturer-suggested cutoff (≥5 FLU), due to the nearly equal cutoffs (Table 3).

Using CD diagnosis as reference, we evaluated the ESPGHAN-recommended use of 10 times greater than or equal to ULN anti–tTG IgA titers at the manufacturer-suggested and optimal cutoffs. The performance characteristics of QUANTA Lite and Aptiva anti–tTG IgA assays at this high titer cutoff were evaluated (Table 4). At ≥10× the ULN cutoff, specificity, and PPV was 100%. The ≥10× of the optimal cutoff resulted in higher sensitivity (from 62% to 69%) and greater NPV (from 76% to 79%) compared to the ≥10× of the manufacturer-suggested cutoff on QUANTA Lite, but similar sensitivity (72%) and NPV (81%) were observed on Aptiva at both cutoffs.

Serology tests serve as an obligatory tool to guide biopsy decisions, as duodenal biopsy remains the gold standard for CD diagnosis. According to the updated 2020 ESPGHAN guidelines, children with high titers of anti–tTG IgA antibodies may be diagnosed with CD without biopsy, provided the anti-EMA antibody is positive in a second sample. With increasing requests for CD serology tests, automated high-throughput assays are an emerging trend, especially in high-volume laboratories. In this study, we evaluated the performance characteristics of a novel multianalyte particle-based fluorescence detection system, Aptiva, which received Food and Drug Administration clearance recently. This platform offers simultaneous and semiquantitative assessment of patient sera for antibodies to anti-tTG and anti-DGP, IgA, and IgG isotypes. The automated Aptiva assays were compared to the manual QUANTA Lite ELISA assays, which only allows for individual detection of these antibodies. We also assessed the clinical performance of the Aptiva system’s anti–tTG IgA assay in the diagnosis of CD using the ESPGHAN-recommended high-titer cutoff.

Using 703 total patient samples, our comparison demonstrated excellent qualitative agreement between Aptiva and QUANTA Lite anti–tTG IgA assays. However, the anti–DGP IgG assay only showed moderate agreement and anti–tTG IgG and anti–DGP IgA assays displayed an even lower PPA between the 2 methods. This prompted comparison of the Aptiva and QUANTA Lite assays with the clinical diagnosis. In total, 127 patients were available with clinical diagnosis as established by the clinician. As expected, clinical performance analysis demonstrated high sensitivity and specificity for anti–tTG IgA, the most efficient serology test for CD diagnosis, by both Aptiva and QUANTA Lite assays. Among the 58 CD patients, noticeably only 1 patient had false-positive results, which were borderline values on both platforms for anti–tTG IgA. This was a patient diagnosed with Crohn disease, and although positive for anti–tTG IgA, the patient’s duodenal biopsy showed no evidence for CD according to the clinician’s report. A negative biopsy does not always rule out CD due to the patchy nature of the disease and inadequate sampling. It is also possible that the patient has an early immune reaction before the onset of mucosal damage that could be evidenced in a biopsy. A large metanalysis study has reported the increased probability of CD in patients with inflammatory bowel disease and increased risk of inflammatory bowel disease in patients with CD.²³  However, currently there are no specific recommendations for screening inflammatory bowel disease patients for CD.

IgA isotypes are generally regarded as having higher accuracy for CD diagnosis in comparison to the IgG tests, except anti–DGP IgG, which can be useful for screening IgA-deficient patients.^21,24,25  IgA deficiency affects 2% to 3% of CD patients; in these cases anti–DGP IgG has demonstrated better accuracy compared with anti–tTG IgG. However, limited evidence advocates for the use of anti–DGP IgG in children younger than 2 years old due to its overall lower accuracy compared with anti–tTG IgA, unless confirmed IgA-deficient.^26–28  In our evaluation, the anti–DGP IgG assay showed a better sensitivity on Aptiva than QUANTA Lite, which can be beneficial in the screening and diagnosis of IgA-deficient individuals. Anti–tTG IgG and anti–DGP IgA assays may benefit from a higher specificity due to only a secondary use of these assays to verify anti–tTG IgA screening results in selected situations. Both these assays demonstrated better specificity on Aptiva compared to QUANTA Lite. Furthermore, predictive values determined in this study were calculated on a CD frequency of 46%. Since our study cohort included patients submitted to CD testing, there is potential for preselection bias. Because predictive values are dependent on the prevalence of the disease, we also calculated the likelihood and odds ratios, which are independent of the pretest probability of the disease. For anti–tTG IgA (P > .99) and anti–tTG IgG (P = .11) tests, the performance characteristics were largely comparable between the 2 platforms (McNemar test for significant proportion changes; Table 2). A high odds ratio was observed for anti–tTG IgA on both platforms and for anti–tTG IgG on Aptiva, indicating a strong association between test positivity and CD diagnosis. Anti–DGP IgA, with a sensitivity of 64.0% versus 83.0% and a specificity of 99.0% versus 97.0%, displayed a higher positive LR of 44.0 versus 28.6 on Aptiva versus QUANTA Lite, respectively (P = .01, McNemar test for significant proportion changes), which makes a positive test result on Aptiva more informative for CD diagnosis compared with QUANTA Lite. Conversely, anti–DGP IgG, with a sensitivity of 90% versus 81% and a specificity of 94.0% versus 99.0%, displayed a lower positive LR of 15.5 versus 55.9 on Aptiva versus QUANTA Lite, respectively (P = .04, McNemar test for significant proportion changes), making a negative test result on Aptiva more beneficial to exclude CD diagnosis compared to QUANTA Lite. Despite superior performance of CD serology assays in these platforms, recommendations based on larger validation studies do not favor combining several tests in lieu of anti–tTG IgA and total IgA screening tests. Previous studies have favored the use of anti–DGP IgA in patients with borderline anti–tTG IgA results. Recent data show only limited value for anti–DGP IgA in these settings due to its high false-positivity rate and suggest cautionary use.^12,29  Indeed, our analyses using UpSet plots demonstrated that only fewer than half (46.6%) of CD patients were positive for all 4 antibodies in our cohort (Supplemental Figure 2). Aptiva system’s anti–tTG IgA displayed positivity in most CD patients separately and/or in combination with other CD markers, demonstrating its superior characteristics for predicting CD. Noticeably, anti–DGP IgA was positive only in 37% of CD patients and anti–DGP IgG was falsely positive in 5.8% of non-CD patients. In a study performed in pediatric patients with moderately increased anti–tTG IgA titers, anti–DGP IgA only showed a moderate improvement in performance when screening for CD in patients without type 1 diabetes but showed a high false-positivity when screening type 1 diabetes patients for CD.³⁰  These data further emphasize the importance of appropriate utilization of these CD markers. To maneuver appropriate test utilization, the Aptiva system has capabilities to link with a laboratory information system and report results specific to reflex testing panels designed by the laboratories.

To evaluate whether optimal cutoffs derived from ROC analysis showed improved performance characteristics compared to the ones suggested by the manufacturer, we compared the clinical performance characteristics at 2 different sets of cutoffs on the Aptiva system. At the optimal threshold, sensitivity increased significantly for Aptiva system’s anti–tTG IgG (from 69% to 90%) and anti–DGP IgA (from 64% to 83%) assays and were maintained for anti–tTG IgA (91%) and anti–DGP IgG (from 90% to 91%) assays (Table 3). However, specificity was compromised for the anti–tTG IgG assay (from 100% using manufacturer-suggested, to 96% using optimal cutoff) and considerably decreased for anti–DGP IgA assay (from 99% using manufacturer-suggested, to 88% using optimal cutoff) using the Aptiva system. At the manufacturer-suggested cutoff (>5 FLU), there were no false-positive results using the anti–tTG IgG assay and only 1 false-positive result (borderline result: 5.3 FLU) was observed for the anti–DGP IgA assay using the Aptiva system. Anti–DGP IgA and anti–tTG IgG assays are not recommended as first-line screening tests and higher specificity may be beneficial to correctly identify patients free of disease after the initial screening. Therefore, in our analysis, the manufacturer-suggested cutoffs met the expected clinical performance characteristics for all 4 biomarkers using the Aptiva system.

Current NASPGHAN and ESPGHAN guidelines suggests testing for total IgA in addition to anti–tTG IgA to ensure the presence of sufficient IgA antibodies, which could impact the results of CD tests involving the IgA isotype. The combination of anti–tTG IgA and total IgA has been regarded as the most reliable screening strategy for CD diagnosis.^12,29  Total IgA is detected using nephelometric or turbidimetric analyzers and shuttling the samples between different locations can have significant impact in workflow efficiencies, especially in high-volume testing laboratories. The Aptiva system has built-in internal control microparticles to detect IgA and the system flags the result in the event of very low or undetectable IgA levels. The IgA detection using internal control microparticles on the Aptiva system was compared with our current IgA measurement using the Optilite system’s turbidimetric method. Only 1 patient sample was discrepant between the 2 methods (Supplemental Table 2). This was a patient sample with IgA values lower than the reference interval but above the detection limit of the turbidimetric measurement for IgA. All other patient samples displayed excellent agreement between the 2 methods. These results demonstrate the acceptability of the internal control microparticles for IgA detection built into the Aptiva system and essentially eliminates the need to utilize a separate analyzer for the actual IgA quantitation by nephelometric and turbidimetric methods during CD screening.

Serologic diagnosis of CD is extremely desirable because it precludes the need, costs, and risks of an invasive procedure. Currently, the American College of Gastroenterology suggests a positive anti–tTG IgA result, or IgG-based test in cases of IgA deficiency, plus a duodenal biopsy as the criteria for CD diagnosis. The 2020 ESPGHAN guidelines advocated for a biopsy-free CD diagnosis in pediatric patients with ≥10× ULN anti–tTG IgA levels in combination with another serum sample positive for anti–EMA IgA antibodies.¹²  Several studies have further investigated whether serology-alone diagnosis could be predictive of CD accurately in adults and children.^13–19  Despite convincing data published in the literature, gastroenterologists have approached the use of this strategy with caution. Experts have raised concerns about the lack of harmonization and variability involved in the currently available anti–tTG IgA assays when using them for a biopsy-free CD diagnosis. Werkstetter et al³¹  found a specificity of 93.5% and PPV of 99.1% from a study that employed assays from different manufacturers.³¹  However, another meta-analysis study in which methodology was not collected showed a specificity of only 48.8% for a serology-alone diagnosis at high titers of anti–tTG IgA antibodies.³²  A strategy proposed to navigate this issue was to validate the high-titers cutoffs method specifically in individual labs.³³  In our study we evaluated the diagnostic characteristics of the QUANTA Lite and Aptiva anti–tTG IgA assays and, using ≥10× ULN anti–tTG-IgA titers, both platforms provided excellent specificity and PPV. Our study did not involve anti–EMA IgA testing; however, even in the absence of anti–EMA IgA results, anti–tTG IgA assays on both platforms demonstrated 100% specificity and PPV at this high-titer cutoff. Similar observations were made by Gidrewicz et al³⁴ ; using the EUROIMMUN (EUROIMMUN US, Mountain Lakes, New Jersey) assay they reported specificity of 97.3% on patients with anti–tTG IgA titers ≥10× ULN, positive anti–EMA IgA, and symptoms. Notably, specificity (95.9%) did not change very much in the absence of EMA positivity.³⁴  Our data demonstrated higher positive predictivity compared to Gidrewicz et al³⁴  and Wolf et al,³⁵  who found 97.3% and >98% positive predictivity, respectively, using EUROIMMUN anti–tTG IgA immunoassay.^34,35  Pacheco et al³⁶  reported a specificity and PPV of 99.5% and 95.4%, respectively, using the Bioplex 2200 anti–tTG assay, which was comparable to the characteristics observed in our study. Thus, our study supports the ESPGHAN recommendations for diagnosis of CD in patients with ≥10× ULN anti–tTG IgA titers without a biopsy using the Aptiva system. Although NASPGHAN guidelines have not released any updates about this approach, both Aptiva and QUANTA Lite systems at ≥10× ULN anti–tTG IgA could be considered for a biopsy-free diagnosis when clinical scenario precludes biopsy on patients. Such a diagnosis should be accompanied with close monitoring to corroborate the reversal of serology and reconstitution of villi upon initiation of a gluten-free diet. Additionally, CD diagnosis without a biopsy should be avoided in patients with type 1 diabetes. Both ESPGHAN recommendations and other studies have cautioned the use of a biopsy-free approach on these patients due to the high false-positivity observed, possibly due to the higher CD screening frequencies of this population.^12,31,35,36 

Our study is limited due to the inclusion of only a small number of patients with clinically defined CD. Despite a limited sample number, clinical performance characteristics of Aptiva assay compared well with the published claims in the manufacturer’s package insert for CD diagnosis. Future studies using well-characterized patient cohorts for diagnosis of dermatitis herpetiformis and for monitoring patients on a gluten-free diet will be beneficial to understand the performance characteristics of Aptiva assays to evaluate these conditions.

In conclusion, the Aptiva system’s multianalyte assays demonstrated excellent performance characteristics for CD detection in our study design in addition to exhibiting capabilities for automation, high-throughput, improved workflow, and faster turnaround times. Our data also support the use of serology-alone criteria for CD diagnosis at ≥10× ULN anti–tTG IgA titers using the Aptiva system. Larger studies are warranted for a widespread adoption of Aptiva’s multianalyte system for CD serology.

The study team would like to thank Werfen (Inova Diagnostics, San Diego, California) for testing blinded ARUP (Salt Lake City, Utah) samples using Aptiva Celiac Disease IgA and IgG Reagents kits, which allowed us to complete this work. We would like to thank specifically Andrea Seaman, BS, from Werfen, who helped coordinate this testing. We would also like to thank Dipanwita Banerjee, MAS, and Hailey Wells at the ARUP Laboratories for their help with collecting and aliquoting samples, and we would like to acknowledge Lauren M. Zuromski, MAS, at the ARUP Laboratories for the statistical analysis support provided.