CAMERA-II Clinical Study Procedures and summary of results
The design, intervention and main analyses of the CAMERA-II study are reported in detail elsewhere.[29] To summarize, CAMERA-II was a 2-year, prospective, randomized, placebo-controlled, double-blind multicenter tight control and treat-to-target (remission) strategy trial among patients with early RA (<1 year since diagnosis). Patients were 18 years or older and naïve to DMARD therapy, including glucocorticoids.
At study baseline, all patients initiated a monthly step-up strategy using oral MTX, at a starting dosage of 10 mg per week, and were randomized to also receive either oral prednisone, 10 mg per day (MTX+pred strategy), or placebo (MTX+plac strategy). Rheumatologists assessed each patient monthly and a computer program indicated whether the patient had achieved response (>20% improvement) compared with the previous visit. If response was not sufficient and remission had not been achieved, MTX dosage was increased by 5 mg per week until the patient had achieved remission (swollen joint count (SJC) =0 and 2 of the following criteria: tender joint count (TJC) ≤3, visual analogue scale (VAS) score ≤20 mm and ESR ≤20 mm/hr). At the maximum (30 mg per week) or maximum tolerable MTX dosage, if a step-up in treatment was indicated, MTX was administered at the same dosage subcutaneously. As the next step, cyclosporine was added to the regimen. However, shortly after start of the trial, cyclosporine was replaced with adalimumab.[29] All patients received folic acid, calcium carbonate with vitamin D and a bisphosphonate. The medical ethics committee of the University Medical Center Utrecht had approved the study. All patients had provided written informed consent before entering the study.
Onset of efficacy was more rapid in the MTX+pred strategy group, and at two years the MTX+pred strategy group had achieved a greater reduction in disease activity, as measured with the DAS28, and had less progression of erosive joint damage, fewer adverse effects and less frequent need for additional biological (b)DMARD treatment [18].
The MBDA score
The development and validation of the MBDA score are reported in detail elsewhere.[24,25] In short, 130 candidate biomarkers were tested in feasibility studies, of which 12 were selected for final algorithm development and validation. The biomarker selection and algorithm were optimized to maximize the strength of the association of the MBDA score with DAS28-CRP in a cohort of patients on diverse treatments.[25] Concentrations of these 12 MBDA protein biomarkers (CRP, epidermal growth factor, interleukin (IL) 6, leptin, matrix metalloproteinase 1 (MMP-1), matrix metalloproteinase 3 (MMP-3), resistin, serum amyloid A (SAA), tumour necrosis factor receptor type I (TNF-RI), vascular cell adhesion molecule 1 (VCAM-1), vascular endothelial growth factor A (VEGF-A) and cartilage glycoprotein 39 (YKL-40)) were measured by multiplex immunoassay using the Meso Scale Discovery MULTI-ARRAY® platform. Biomarker concentrations were combined in the validated MBDA algorithm to generate the MBDA score, a whole number from 1 to 100, for which the established categories of disease activity are: low (<30), moderate [30−44] and high (>44).[24] Biomarker measurement and MBDA score calculation were performed in the CLIA-certified laboratory of Crescendo Bioscience, Inc., South San Francisco, CA, USA using the same instrument, reagents and algorithm as for the VectraÒ DA test, which is commercially available in the United States.
Multiple biomarker-based disease activity assessment in CAMERA-II
MBDA biomarkers were evaluated in serum samples obtained at baseline and at 1, 2, 3, 4, 5, 6, 9 and 12 months. Numbers of samples available for the present study varied between time points, based on patient compliance and the volume of available sample. Of 104 patients in CAMERA-II for whom baseline sera were available for MBDA testing, MBDA scores and DAS28 were analyzed for the 92 who had at least one MBDA test result for months 1, 2, 3, 4, 5, 6, 9 or 12. For this 92-patient cohort, the average number of post-baseline tests per patient was 3.7.
Statistical Analyses
To evaluate changes from baseline for DAS28 and MBDA score and comparisons of change in DAS28 or MBDA score over time between patients treated with the MTX+pred or MTX+plac strategy, a t-test was performed for each time point evaluated. Association between change from baseline to 12 months for DAS28 and MBDA score was assessed using Spearman’s rank correlation.
Concentrations of individual biomarkers were analyzed for the subset of 51 patients who had an MBDA test at baseline and at least one time point from month 1 to 5, to focus on the initial biomarker responses to treatment and exclude possible effects from exposure to cyclosporine or adalimumab. The average number of post-baseline tests per patient was 3.3 in this subset. Biomarker concentrations were analyzed after base-10 logarithm (log10) transformation, to approximate a normal distribution. The changes from baseline in log10 biomarker concentrations were assessed for months 1−5 for each treatment arm by Wilcoxon signed rank test and compared between treatment arms by Mann-Whitney U-tests. The means of the changes were calculated as averages of individual changes in log10 values, and standard error (SE) values were determined accordingly. For presentation in graphs, each mean change (D) was back-transformed by raising 10 to the D power, thus reversing the log10 transformation to generate a fractional value, relative to baseline, on a linear scale. Thus, any time point demonstrating no change from baseline was represented on the graph with a value of 1.0 and, for example, a 20% reduction from baseline was represented with a value of 0.8. Response profiles are the courses of changes from baseline for the MTX+plac and MTX+pred strategy arms.
For the individual biomarkers, profile categories were defined, dependent on their response to MTX+plac, and their response to concomitant prednisone, i.e. the difference in response to MTX+plac and MTX+pred. This was based on visual inspection of curves representing change from baseline in biomarker concentration for each treatment strategy arm and on p-values for changes from baseline and for the difference between treatment strategy arms. The software package R 2.15.1 (www.r-project.org) was used for the analyses. No clinical or biomarker data were imputed. A p-value of <0.05 was considered statistically significant. No adjustments were made for multiple testing.