Table 2.

Unadjusted and adjusted liner regression models for assessing the correlations among different clinical parameters and T cell signal transducer and activator of transcription-1 (STAT)-1 expression levels in 23 patients with systemic lupus erythematosus (SLE)^*

	Univariate fold change	P-value	Multivariate† fold change	P-value
Age (per 10 years)	0·92 (0·66–1·28)	0·605	0·95 (0·69–1·34)	0·757
Sex (male/female)	0·60 (0·24–2·49)	0·253	0·66 (0·27–1·59)	0·330
SLEDAI score (per 1 point)	1·06 (0·99–1·13)	0·055	1·03 (0·96–1·11)	0·348
Anti-dsDNA (per 10 IU/ml)	1·02 (0·97–1·07)	0·412	–	–
C3 (per 10 mg/dl)	0·94 (0·83–1·07)	0·345	–	–
Lupus nephritis (yes/no)	2·27 (1·20–4·29)	0·014	2·04 (1·04–3·99)	0·038
Steroid dosage equivalence to methylprednisolone (mg/day)	1·09 (0·95–1·25)	0·189
Potent immunosuppressants‡ (yes/no)	1·45 (0·70–3·00)	0·300	–	–

^*SLE patients from both the screening and validation group were included, but three patients in the screening group were excluded due to lack of enough specimens for analysing the expression levels of mRNA.

^†After analysis with the multiple linear regression model adjusted for age and gender, only lupus nephritis, but not SLE disease activity index (DAI) score, was correlated significantly with increased expression of STAT-1.

^‡SLE patients received azathioprine, mycophenolic acid or cyclosporin in addition to steroid and hydroxychloroquine.