Gijsbert J. Hötte ^1* Maaike d. Bie ² Ronald O. De Keizer ^1* Pieter Martijn Kolijn ² Roosmarijn C. Drexhage ^3* Sharon Veenbergen ^2* Marjan A. Versnel ⁴ Petrus M. VAN HAGEN ⁵ Dion Paridaens ¹ Willem A. Dik ^2*

• ¹ Rotterdam Eye Hospital, Rotterdam, Netherlands
• ² Laboratory Medical Immunology, Department of Immunology, Erasmus MC University Medical Center Rotterdam, Rotterdam, Netherlands
• ³ Department of Internal Medicine, section Endocrinology, Erasmus MC University Medical Center Rotterdam, Rotterdam, Netherlands
• ⁴ Department of Immunology, Erasmus MC University Medical Center Rotterdam, Rotterdam, Netherlands
• ⁵ Department of Internal Medicine, Section Allergy and Clinical Immunology, Erasmus MC University Medical Center Rotterdam, Rotterdam, Netherlands

Background: For the selective detection of thyroid-stimulating hormone receptor antibodies with stimulating properties (thyroid-stimulating immunoglobulins; TSI), a novel and rapid bioassay (Turbo TSI) has been introduced. We evaluate the clinical performance of Turbo TSI in Graves’ orbitopathy (GO) patients and compare it to a bridge-based TSI binding immunoassay and third generation TSH-R-binding inhibitory immunoglobulins (TBII) assay. Also, we investigate the association of Turbo TSI and TBII measurements with GO activity and severity, as well as response to intravenous methylprednisolone (IVMP), and compare results to previous findings on the bridge-based TSI binding immunoassay. Methods: Turbo TSI, TBII and bridge-based TSI binding immunoassay measurements were performed in biobank serum from 111 GO patients and control cases (healthy controls [HC; n=47], primary Sjögren’s disease [SD; n=10], systemic sclerosis [SSc; n= 10], systemic lupus erythematosus [SLE; n=10]). Clinical characteristics and response to treatment were retrospectively retrieved from GO patient files. Results: Turbo TSI had the highest sensitivity (97.3%) and negative predictive value (96.1%), while bridge-based TSI binding immunoassay showed the highest specificity (100%) and positive predictive value (100%). Differentiating GO patients from control cases, receiver operating characteristic (ROC) analysis showed an area under the curve (AUC) of 98.5%, 95.7% and 99.8% for Turbo TSI, TBII and bridge-based TSI binding immunoassay, respectively. Turbo TSI (p<0.001) and TBII (p<0.01) levels were higher in patients with active compared to inactive GO. Correlation with CAS was stronger for Turbo TSI (r=0.42) than TBII (r=0.25). No statistically significant differences were observed in IVMP responders vs. non-responders for Turbo TSI (p=0.092) and TBII (p=0.21). For identifying active GO, an AUC of 75% with Turbo TSI and 67% with TBII was found. For IVMP response, AUC was 66.3% with Turbo TSI and 62.1% with TBII. In multivariate logistic regression analyses, both assays were independently associated with disease activity (p<0.01 for both assays) and IVMP response (p<0.01 for Turbo TSI; p<0.05 for TBII). Conclusions: The new Turbo TSI functional bioassay has good clinical performance. Although turbo TSI is a stronger marker of activity and IVMP response than TBII, results are comparable to our previously published findings on the bridge-based TSI binding immunoassay.