Biomarkers in Early-stage Spondyloarthritis

Currently, no laboratory test or biomarker is cable of differentiating between patients with early-stage axSpA and persisting low back pain of other causes. Human leukocyte antigen (HLA) B27 is not an acute biomarker, but a prognostic marker for the development of axSpA.

C-reactive protein (CRP) is commonly used as an acute biomarker in different inflammatory conditions. Previous studies have shown that only 40-60% of ankylosing spondylitis (AS) patients with clinical signs of active disease have elevated levels of CRP.

Earlier studies regarding the complement C3 split product C3d have shown a correlation between inflammatory diseases like AS and elevation of C3d. Those studies were based on a small number of cases, but a newer study from 2016 has shown a correlation between different biomarkers, including C3 and new fatty lesions at MRI in Golimumab treated AS patients.

Objective To investigate the predictive value of baseline levels of C3d and hsCRP identifying bone marrow oedema (BME) at MRI of the sacroiliac joints (SIJ) at baseline.

Design:

The study is a prospective cohort study including patients, aged 18-45 years, having unspecific low back pain registered in the Spines of Southern Denmark (SSD) cohort including patients referred to the Spine Centre of Southern Denmark.

At baseline, all patients filled in a standardised questionnaire and were interviewed identifying inflammatory symptoms followed by a clinical examination focusing on symptoms related to the SIJ. Blood samples were taken and analysed for HLA-B27 (prognostic biomarker). Furthermore, all patients underwent an MRI of the spine including the SIJ.

In total, 1037 patients were included, 696 had blood tests taken and gave consent for further use for research purposes.

For research purposes, the blood samples were stored in the Molecular Biology of Infectious Agents in the Early Diagnosis of Spondyloarthritis biobank (MICSA). The biobank is hosted by the research group at Graasten Rheumatological Hospital. Of the 696 patients who gave consent,188 patients were examined by a Rheumatologist at baseline. Of those 96 patients fulfilled the criteria of having axSpA according to The Assessment of SpondyloArthritis international Society (ASAS) criteria, 38 patients had inflammatory changes at MRI or had one feature consisting with spondyloarthritis, but did not fulfil the ASAS criteria. 82 randomly selected patients with low back pain were included as a control group.

In total, 216 patients were defined as the cohort included for further analysis in this study.

Data:

The Department of Clinical Biochemistry at Vejle Hospital has previously assisted in the collection and storage of MICSA samples. It is technically possible to use the stored frozen MICSA samples for analysing C3d.

In total, 184 (of 216) baseline samples were accessible for further analysis. The baseline samples will be analysed for C3d with the use of EDTA (ethylenediaminetetraacetic acid) plasma and treated according to the local protocol for such analyses.

MRI

All patients in the MICSA cohort have had an MRI of the spine including the SIJ at baseline. Two experienced radiologists have systematically described the MRI regarding signs of inflammatory changes including BME.

Statistics:

STATA (version 15.0) will be used for statistical analysis. Parametric data will be reported as the mean and standard derivation (SD) or One-way analysis of variance (ANOVA). In between-group comparisons for continuous and categorical demographic variables will be performed with the independent sample t-test and Pearson Chi-square test or Fisher´s exact test. The Kruskal-Wallis test and interquartile range will be used to describe nonparametric data. Logistic regression analysis will be used to test independent variables as predictive factors for BME detected by MRI, resulting in odds-ratios, sensitivity, specificity, and area under the curve (AUC). This will be followed by multivariate logistic regression to evaluate the potential effect of other variables. A p-value < 0.05 is considered statistically significant.

Sample size and power calculation:

A previous study with patients suspected of having spondyloarthritis has found a significant difference in mean levels of C3 between groups with 23 and 22 participants, respectively.

By the use of means and SD from the above study and alfa= 0.05, beta= 0.2, power 0.8 and enrollment ratio 1:2 the sample size in our study is calculated to n:10 and n:20 respectively.

It is therefore expected, that a significant difference in C3d levels between the three groups in the present study can be detected based on the size of each group in the MISCA cohort.