Background Gouty arthritis (GA) is a progressive inflammatory disease of the joints resulting from the deposition of monosodium urate crystals due to hyperuricemia. A subset of patients with comorbidities cannot be adequately treated with standard anti-inflammatory therapies. Inadequately treated GA patients develop tophaceous disease and acute flares. Building treatment prediction models is an important step for optimizing clinical practice and assisting with the design and analysis of clinical studies, particularly to account for potential bias and confounding factors.
Objectives To identify the most significant patient characteristics that predict new flares in GA patients treated with a single dose of canakinumab 150 mg s.c. vs. triamcinolone acetonide 40 mg i.m. in two 12-week studies (β-RELIEVED, N=230; β-RELIEVED II, N=226) followed by 12-week extensions.
Methods We conducted a retrospective analyses to explore the impact of demographic and baseline characteristics on incidence of new flare. The effect of each baseline variable on reflare was assessed using the logistic regression model. For each variable, the model included factors for treatment and the individual variable. An initial multivariate model including the effects of treatment and all variables that were individually significant was also examined. The final multivariate model excluded variables that were not significant (p<0.05) in the initial multivariate model.
Results The triplet of baseline parameters as number of tophi locations, number of joints affected by acute GA and number of flares in the last year are the only statistically significant covariates in the initial screening models and remain significant in the multivariate model (Table). Parameters excluded from the model were: age (p=0.3607), sex (p=0.5738), race (p=0.1450), BMI (p=0.4368), urate level (p=0.2104), LogCRP (p=0.2052), duration of GA (p=0.2669). Number of tophi locations is a quantitative measure of GA severity indicating the degree to which tophi formation extends over the whole body. Number of joints affected by acute flares dichotomized on the optimal cut-off point >1 vs 1 is a spatial criterion indicating the extent of acute conditions. The frequency of reflaring – number of flares in the last year – illustrates the temporal dimension (Poisson process) of this model of GA severity.
Conclusions Based on this spatio-temporal model the number of tophi locations, number of joints affected by acute GA, along with the number of flares is predictive of subsequent reflaring in this population of GA patients. Thus, the phenomenon of repeated acute GA is impacted mainly by the extent of advanced GA, i.e. the degree of systemic disease and frequency of flares, and not by factors immediately preceding an inflammation (CRP, urate level, etc.) thought to be contributing to the reflaring.
Disclosure of Interest T. Bardin Grant/Research support from: Menarini, Consultant for: Novartis, Ipsen, Menarini, Ardea, Biocryst, Speakers Bureau: Novartis, R. Alten Grant/Research support from: Novarttis, Consultant for: Novartis, N. Schlesinger Grant/Research support from: Novartis, Consultant for: Novartis, URL Pharma, Savient, Takeda, Rx Ensyme, Speakers Bureau: Novarttis, A. Shpilsky Shareholder of: Novartis, Employee of: Novartis, T. Kiechle Shareholder of: Novartis, Employee of: Novartis, A. So Grant/Research support from: Novartis, Consultant for: Novarttis, Ardea, Speakers Bureau: Novartis, Ardea, Menarini
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.