Clustering of gout-related comorbidities and their relationship with gout flares: a data-driven cluster analysis of eight comorbidities.

OBJECTIVES

To study the aggregation of multiple comorbidities in people with gout and explore differences in prognosis of gout flares among different subgroups.

METHODS

Hierarchical clustering was performed to identify homogeneous subgroups among 2639 people with gout using eight comorbidities. A one-year follow-up of acute gout flares in 463 of these people was conducted; the incidence and the timing of gout flares in each cluster were assessed to explore prognosis of gout flares. Binary logistic regression was applied to assess factors associated with gout flares.

RESULTS

In baseline study, we identified five subgroups (C1-C5). C1 (n = 671, 25%) was characterized by isolated gout with few comorbidities. C2 (n = 258, 10%) were all obese. Almost all people in C3 (n = 335, 13%) had diabetes (99.7%). All people in C4 (n = 938, 36%) had dyslipidemia. C5 (n = 437, 17%) had the highest proportion of cardiovascular disease (CVD, 53%), chronic kidney disease (CKD, 56%), and cancer (7%). In follow-up study, C5 had the highest incidence (71.9%) and earliest onset (median 3 months) of gout flares. C2 had the lowest incidence (52.1%) and the latest onset (median 10 months) of gout flares. The highest relative risk for gout recurrent was seen for C5 (OR = 2.09). Other factors associated with the risk of gout flares were age at diagnosis of gout, duration of gout, presence of tophi, and smoking ≥ 20 cigarettes/day.

CONCLUSIONS

We clustered people with gout into five groups with varying comorbidities. People with CVD, CKD, and cancer had the highest risk of gout flares and should receive comprehensive care.