Table 4

Alternate multiple linear regression model performed in the discovery cohort and including both angiogenic factors independently associated with the late-NVC pattern and SSc-related disease characteristics associated univariately with P ≤ 0.10 with the late-NVC pattern

NVC pattern

P value

Multiple

linear

regression

(r)

P value


Early/Active

(n = 36)

Late

(n = 24)


Modified Rodnan skin score > 14, n (%)

6 (17)

13 (54)

0.005*

0.50

< 0.0001

History of digital ulcers, n (%)

13 (36)

16 (67)

0.03*

0.25

0.04

Positive antitopoisomerase antibodies, n (%)

12 (33)

16 (67)

0.02*

FVC < 75, n (%)

5 (14)

8 (33)

0.1*

DLCO/VA < 75%, n (%)

9 (25)

14 (58)

0.02*

Treatment with angiotensin-converting enzyme inhibitors, n (%)

8 (22)

11 (46)

0.09*

Treatment with endothelin-receptor antagonists and/or PDE5 inhibitors, n (%)

4 (11)

8 (33)

0.08*

EPCs (106 Lin-mononuclear cells), median (range)

64 (21-573)

31 (5-300)

< 0.0001*

-0.45

0.005

VEGF (pg/ml), median (range)

470 (156-1,146)

814 (189-1,564)

0.01*

0.37

0.01


DLCO/AV, diffusing capacity for carbon monoxide/alveolar volume; FVC, forced vital capacity; NVC, nailfold videocapillaroscopy.

    * Variables included in multiple regression analysis

Avouac et al. Arthritis Research & Therapy 2013 15:R55   doi:10.1186/ar4217

Open Data