Figure 4.

Risk of hospitalization for flares in people with gout who initiated urate-lowering therapy (ULT) within 12 months of diagnosis, compared with people who did not initiate ULT. Outputs from Cox proportional hazards models are shown, highlighting the change in hazard ratio for hospitalizations in relation to time elapsed following initiation of ULT. Adjustment was performed for the following covariates: age, sex, calendar year of gout diagnosis, diuretic use and comorbidities at diagnosis (hypertension, CKD, IHD, heart failure, diabetes mellitus, prior CVA, obesity, smoking status, alcohol excess, history of urolithiasis). A logarithmic y-axis was used, to reflect the exponential distribution of hazard functions