Objective: Previous publications have reported that increased absolute monocyte counts are associated with treatment non-response in patients with rheumatoid arthritis (RA). This study investigated whether full blood count (FBC) components from routine clinical testing before treatment with a biological disease-modifying anti-rheumatic drug (bDMARD) were associated with treatment non-response after 6 months of treatment.
Method: From a UK-based prospective multicentre study of patients with RA starting a bDMARD, data from 246 patients attending five of the participating centres were retrieved. FBC components were analysed for their association with European Alliance of Associations for Rheumatology non-response after 6 months of treatment using backward stepwise logistic regression, adjusting for potential confounders. Final models underwent resampling with 200 repeats of out-of-bag bootstrapping to assess model performance using area under the receiver operating characteristics (AUROC) curves. Model fit was compared using the Akaike information criterion (AIC).
Results: After 6 months of treatment, the only FBC component predictive of non-response was pretreatment absolute monocyte count [adjusted odds ratio (ORadj) 9.56, 95% confidence intervals (CI) 1.61-59.86, p = 0.01, AUROC = 60.42%). The model including monocytes as a predictor demonstrated superior performance to the covariates-only model (AIC 184.36 vs 188.51, respectively).
Conclusion: In the largest study to date, increasing absolute monocyte counts were associated with bDMARD non-response after 6 months of treatment, replicating previous reports. Validation and mechanistic studies are required to inform future treatment selection.