Table 2

Test characteristics of different cut-offs for the number of erosive joints with initiation of MTX or any DMARD during the first year as well as arthritis persistency over 5 years as outcome, in all early arthritis patients studied

		Leiden-EAC (n=902)*								ESPOIR (n=700)†
Outcome	N erosive joints	N patients (%)‡	Sens	Spec	PPV	NPV	LR+	LR-	AUC	N patients (%)*	Sens	Spec	PPV	NPV	LR+	LR-	AUC
MTX	None	395			311
	≥1 joint	507 (56)	0.77	0.57	0.53	0.79	1.78	0.41	0.67	389 (56)	0.60	0.51	0.67	0.43	1.23	0.78	0.56
	≥2 joints	387 (43)	0.64	0.71	0.58	0.76	2.19	0.51	0.68	261 (37)	0.41	0.69	0.69	0.41	1.35	0.84	0.55
	≥3 joints	307 (34)	0.51	0.77	0.58	0.71	2.19	0.63	0.64	177 (25)	0.29	0.80	0.71	0.40	1.48	0.88	0.55
	≥4 joints	227 (25)	0.40	0.84	0.62	0.69	2.51	0.72	0.62	133 (19)	0.21	0.85	0.71	0.39	1.44	0.92	0.53
	≥5 joints	179 (20)	0.32	0.88	0.64	0.67	2.74	0.77	0.60	96 (14)	0.16	0.90	0.73	0.39	1.61	0.93	0.53
DMARD	None	395						311
	≥1 joint	507 (56)	0.72	0.75	0.85	0.57	2.90	0.37	0.74	389 (56)	0.58	0.57	0.87	0.21	1.34	0.74	0.57
	≥2 joints	387 (43)	0.57	0.86	0.89	0.50	4.03	0.50	0.72	261 (37)	0.40	0.75	0.88	0.20	1.56	0.81	0.57
	≥3 joints	307 (34)	0.46	0.89	0.89	0.45	4.09	0.61	0.67	177 (25)	0.28	0.86	0.91	0.19	2.04	0.84	0.57
	≥4 joints	227 (25)	0.35	0.93	0.91	0.42	5.21	0.70	0.64	133 (19)	0.21	0.90	0.91	0.19	2.04	0.88	0.55
	≥5 joints	179 (20)	0.28	0.97	0.94	0.40	8.51	0.74	0.62	96 (14)	0.15	0.95	0.94	0.18	3.04	0.89	0.55
PERSISTENCY	None	395				234
	≥1 joint	473 (54)	0.60	0.64	0.85	0.31	1.67	0.62	0.62	304 (56)	0.59	0.56	0.85	0.25	1.34	0.73	0.58
	≥2 joints	358 (41)	0.47	0.80	0.89	0.30	2.41	0.65	0.64	202 (37)	0.40	0.74	0.87	0.23	1.55	0.81	0.57
	≥3 joints	283 (33)	0.38	0.85	0.90	0.28	2.63	0.72	0.62	135 (25)	0.27	0.85	0.88	0.22	1.78	0.86	0.56
	≥4 joints	209 (24)	0.29	0.92	0.92	0.27	3.45	0.78	0.60	98 (18)	0.20	0.90	0.90	0.21	2.10	0.88	0.55
	≥5 joints	169 (19)	0.24	0.95	0.94	0.26	4.55	0.81	0.59	70 (13)	0.15	0.96	0.94	0.21	3.95	0.88	0.56

This table describes the test characteristics when the cut-off of erosiveness was set at different number of erosive joints. Patients with equal or more number of erosive joints as specified were compared with patients with less erosive joints as specified. The predictive ability of each cut-off was tested against the initiation of MTX, any DMARD during the first year or disease persistency over 5 years.
*868 patients had sufficient follow-up to define persistency in the Leiden-EAC.
†538 patients had sufficient follow-up to define persistency in ESPOIR.
‡Number and percentage of patients positive in this analysis.
AUC, area under the receiver operating characteristic curve; DMARD, disease modifying antirheumatic drug; ESPOIR, Étude et Suivi des POlyarthrites Indifférenciées Récentes; Leiden-EAC, Dutch Leiden-Early Arthritis Clinic; MTX, methotrexate; NPV, negative predictive value; PPV, positive predictive value.