Table 2

Univariate analysis of baseline variables as possible predictors of radiographic damage

Variable	OR (95% CI)	p Value†	Sensitivity	Specificity	PPV	NPV
Continuous variables are dichotomised by their median values and anti-CCP, RF, DRB1*04 as to whether they are present or absent. Radiological joint damage is defined as an end point Larsen score of ⩾10 or more (the median value).
†Upper limit of expected number of false significances = 0.26.
CCP, cyclic citrullinated peptides; CI, confidence interval; DAS28, 28 joint disease activity score; ESR, erythrocyte sedimentation rate; HAQ, health assessment questionnaire; NPV, negative predictive value; OR, odds ratio; PPV, positive predictive value; RF, rheumatoid factor; VAS, visual analogue scale.
Larsen score ⩾4	12.9 (7.6 to 22.0)	0.0005	0.79	0.77	0.79	0.78
Anti-CCP present	3.6 (2.3 to 5.7)	0.0005	0.71	0.65	0.65	0.66
RF present	2.7 (1.7 to 4.3)	0.0005	0.73	0.61	0.61	0.63
ESR ⩾29 mm/h	2.7 (1.7 to 4.2)	0.0005	0.63	0.63	0.63	0.61
C reactive protein ⩾19 mg/l	2.2 (1.4 to 3.5)	0.0005	0.60	0.62	0.62	0.58
Age ⩾55 years	2.1 (1.4 to 3.3)	0.001	0.59	0.61	0.61	0.58
Male sex	1.6 (1.0 to 2.5)	0.044	0.38	0.72	0.59	0.52
Previous or current smoker	1.6 (1.0 to 2.5)	0.040	0.66	0.66	0.66	0.55
Disease duration ⩾6 months	0.8 (0.5 to 1.2)	0.274	0.51	0.49	0.49	0.45
DRB1*04 present	1.3 (0.7 to 2.5)	0.371	0.58	0.48	0.48	0.59
DAS28 ⩾5.1	1.2 (0.8 to 1.9)	0.307	0.55	0.55	0.55	0.51
Pain VAS ⩾44 mm	1.0 (0.6 to 1.5)	0.951	0.53	0.52	0.52	0.47
HAQ score ⩾0.9	1.0 (0.6 to 1.6)	0.976	0.53	0.51	0.51	0.49