Table 3.

Independent predictors of elevated preoperative CRP (as a binary variable).

Predictor	Adjusted odds ratio	95% confidence interval	P value
Fracture occurred ≤8 wk preoperatively	5.95	(2.35–15.1)	<.001
Active periprosthetic joint infection	4.88	(2.19–10.9)	<.001
Age, per 5-yr increase	0.83	(0.70–0.995)	.03
Whole model area under receiver operating characteristic curve: 0.71			<.001

Independent predictors of elevated preoperative CRP determined by multivariate analysis, with CRP treated as a binary outcome variable. An elevated CRP level was defined as a laboratory measurement of serum CRP greater than 10 mg/L. Variables assessed and ultimately excluded from the multivariate model due to lack of significance were surgical indication (other than fracture or periprosthetic infection), sex, history of prior shoulder surgery, history of prior shoulder infection, tobacco use status, history of rheumatoid arthritis or autoimmune disorder, and history of diabetes mellitus. CRP C-reactive protein.

Bold indicates statistically significant values.