Multivariable models exploring determinants of joint damage in patients with gout separately for (a) erosions using zero-inflated negative binomial regressions and (b) joint space narrowing (JSN) using negative binomial regressions.
| (a) | Not Being Erosive* | Erosion Score (count)¥ | ||||||
|---|---|---|---|---|---|---|---|---|
| β | OR† | 95% CI (OR) | p | Β | Exp(β)‡ | 95% CI [Exp(β)] | p | |
| Age, yrs | −0.15 | 0.86 | 0.74–0.99 | 0.036 | 0.05 | 1.06 | 1.02–1.09 | 0.002 |
| Sex, female | 3.36 | 28.80 | 0.87–955.74 | 0.06 | 0.50 | 1.65 | 0.72–3.75 | 0.23 |
| Disease duration, yrs | 0.04 | 1.04 | 1.01–1.07 | 0.018 | ||||
| No. tophi | 0.07 | 1.07 | 1.03–1.12 | 0.001 | ||||
| sUA ≤ 0.36 mmol/l, yes/no | 4.39 | 80.53 | 1.25–5192.79 | 0.039 | ||||
| (b) | JSN score (count)¥ | |||||||
| Age, yrs | # | # | # | # | 0.01 | 1.01 | 0.99–1.02 | 0.63 |
| Sex, female | # | # | # | # | −0.13 | 0.88 | 0.55–1.39 | 0.58 |
| Tophaceous gout, yes/no | # | # | # | # | 0.57 | 1.76 | 1.23–2.53 | 0.002 |
Significant values are shown in bold face.
↵* Logistic model, predicting being nonerosive (erosion score being “certain zero”).
↵¥ Negative binomial model, predicting expected count.
↵† Factor change in odds for 1-unit increase in the independent variable.
↵‡ Factor change in expected count for 1-unit increase in the independent variable. # No estimates, because a negative binomial model has no “certain zeros.” sUA: serum uric acid.