Radiological features do not predict failure of two-stage arthroplasty for prosthetic joint infection: a retrospective case–control study

Table 2 Analysis for prediction of failure (binary logistic regression)

Condition	Number cases	Number controls	P value	Odds ratio(OR)	95% C.I. for OR
	N = 12	N = 29			Lower	Upper
Metal remaining	1 (8.3%)	0 (0%)	1.0	1 x 10¹⁰	0	∞
New metal	1 (8.3%)	0 (0%)	1.0	1 x 10¹⁰	0	∞
Cement	4 (33.3%)	9 (31%)	0.909	1.134	0.132	9.739
Restrictor	2 (16.7%)	4 (13.8%)	0.404	3.449	0.189	63.108
New fracture	3 (25%)	5 (17.2%)	0.654	0.481	0.020	11.807
Antimicrobial device	1 (8.3%)	3 (10.3%)	0.932	0.865	0.031	24.047
Drain	3 (25%)	4 (13.8%)	0.345	3.108	0.296	32.623

The table displays binary logistic regression analysis performed for seven variables with the outcome of prosthesis failure, with the model adjusted for age, gender and number of previous revisions.

ISSN: 1471-2474