Table 2.

Relationship of plasma lipids (PGE2 and 15-HETE) and PBL inflammatory gene transcriptome biomarkers with JSN at 24 months in patients with symptomatic knee osteoarthritis (SKOA).

	Area under ROC (AUC)			Logistic Regression
	AUC	P value	P value adj.	(no adjust.)	Age, Gender, BMI	Age, Gender, BMI, KL
JSN>0mm(N=72 vs. 39)
15-HETE	0.412	0.180	0.325	0.293	0.348	0.311
PGE-2	0.490	0.435	0.461	0.774	0.618	0.909
PBL IL-1β expression	0.622	0.016	0.073	0.144	0.142	0.186
PBL TNF-α expression	0.579	0.096	0.212	0.436	0.396	0.338
PBL COX-2 expression	0.699	0.0001	0.0015	0.0005	0.0003	0.0002
Age	0.487	0.395	0.461
Gender	0.480	0.358	0.461
BMI	0.527	0.285	0.405
KL	0.585	0.066	0.161
JSN>0.2mm (N=64 vs. 39)
15-HETE	0.410	0.177	0.325	0.311	0.339	0.306
PGE-2	0.494	0.461	0.461	0.609	0.499	0.878
PBL IL-1β expression	0.609	0.032	0.123	0.135	0.124	0.181
PBL TNF-α expression	0.579	0.102	0.212	0.392	0.345	0.257
PBL COX-2 expression	0.681	0.0004	0.0035	0.0011	0.0007	0.0003
Age	0.509	0.429	0.461
Gender	0.481	0.368	0.461
BMI	0.534	0.243	0.364
KL	0.623	0.015	0.073
JSN>0.5mm (N=44 vs. 39)
15-HETE	0.334	0.052	0.141	0.068	0.060	0.065
PGE-2	0.492	0.453	0.461	0.535	0.436	0.684
PBL IL-1β expression	0.672	0.003	0.017	0.077	0.072	0.118
PBL TNF-α expression	0.611	0.045	0.141	0.449	0.378	0.284
PBL COX-2 expression	0.712	0.0001	0.0015	0.0005	0.0004	0.0003
Age	0.539	0.240	0.364
Gender	0.486	0.416	0.461
BMI	0.539	0.228	0.364
KL	0.604	0.050	0.141

For measuring a biomarker’s predictivity for radiographic progression in medial knee OA, we used the area under the receiver operating characteristic (ROC) curve. Generalized Linear Model (GLM) -based (logistic regression) p-values are provided after adjusting individually and in combination for three covariates: age, gender and body mass index (BMI) and KL score.

^*15-HETE data were available only for N=43 SKOA subjects. JSN=joint space narrowing; AUC=area under ROC curve.