Table 3

Results of the GEE model for longitudinal multinomial regression on the risks of sepsis (n=137) and death (n=53) after SI (n=1017)

OR95% CIOR95% CI
Age (by 10 years)1.411.15 to 1.742.471.61 to 3.79
Sex (male vs female)0.990.63 to 1.551.450.74 to 2.83
FFbH (by 10% improvement)0.920.84 to 1.000.860.76 to 0.98
GC (<5 mg/day=reference)Ref..Ref..
 GC (5 to <10 mg/day vs ref.)1.260.82 to 1.930.930.47 to 1.83
 GC (≥10 mg/day vs ref.)1.660.96 to 2.882.401.04 to 5.55
 TNFi0.640.42 to 0.970.480.24 to 0.95
 Other bDMARD0.450.25 to 0.800.160.05 to 0.54
Heart failure (yes vs no)1.380.74 to 2.563.561.73 to 7.33
Chronic renal disease (yes vs no)1.931.19 to 3.141.510.72 to 3.17
  • The adjusted ORs specify the increase or decrease in the risk of developing the outcome (sepsis or death).

  • bDMARD, biologic DMARD; csDMARD, conventional synthetic DMARD; DMARD, disease-modifying antirheumatic drug; GC, glucocorticoids; GEE, generalised estimating equation; SI, serious infection; TNFi, tumor necrosis factor-α inhibitor (adalimumab, etanercept, infliximab, golimumab, certolizumab), other bDMARD (tocilizumab, rituximab, abatacept). Physical function (FFbH) and doses of GC refer to measurements of most recent study visit, DMARD exposure is the current exposure at SI. The median time between the last study visit and the SI was 3.7 months (first quartile: 1.9 months, third quartile: 6.2 months).