Rate of serious infections in spondyloarthropathy patients treated with anti-tumour necrosis factor drugs: a survey from the Italian registry GISEA.

To determine the incidence of serious infections (SIs) among the spondyloarthropathy (SpA) patients from the "Gruppo Italiano per lo Studio delle Early Arthritis" (GISEA) registry and treated with tumour necrosis factor (TNF) inhibitors (TNFIs), and to identify the factors associated with the development of the infections. This observational study on 3321 GISEA-registered SpA patients collected real-world demographic and clinical data relating to their biological drug treatments. The overall incidence of infections was analysed by type of SpA. A total of 3321 SpA patients (1731 males, 52.2%; mean age 47±13 years; median disease duration 3 years, interquartile range [IQR] 0-8) were eligible for inclusion in the analysis. Two hundred and fifty-nine patients experienced at least one of 391 microbiologically diagnosed SIs, 32% of which were recorded during the first 12 months of treatment. The overall incidence of SIs was 43.9/1000 patient-years of follow-up (95% confidence interval [CI] 39.6-48.4): 29.9/1000 (95% CI 23.1-38.1) among those treated with adalimumab (ADA); 36.1/1000 (95% CI 30.0-43.1) among those treated with etanercept (ETN); and 61.4/1000 (95% CI 53.3-70.5) among those treated with infliximab (INF). The highest incidence was observed among the patients with psoriatic arthritis (PsA), but the difference was statistically significant only in comparison with the patients with undifferentiated SpA (p=0.002), whose incidence of SIs was also lower than in the patients with ankylosing spondylitis (AS) (p=0.034). Multivariate models showed that the number of comorbidities (hazard ratio [HR] 1.29, 95%CI 1.2-1.4; p<0.001), age at the start of TNFi treatment (HR 0.99, 95%CI 0.97-0.99; p=0.030), steroid use (HR 1.40, 95%CI 1.1-1.8; p=0.012) and male sex (HR 0.72, 95%CI 0.5-0.9; p=0.012) were all statistically significant predictors of infection. The factors independently associated with a lower risk of SIs were the use of ETN (HR 0.52, 95%CI 0.4-0.7; p<0.001) or ADA (HR 0.59, 95%CI 0.4-0.8; p=0.002) rather than INF. The incidence of SIs was higher among patients with PsA or AS than among those with undifferentiated SpA, and among patients treated with INF than among those treated with ADA or ETN. Male sex, steroid use and the number of comorbidities were all factors predictive of SIs.