The study was carried out in the cohort of patients with SSc that was used in the validation study of NEMO score. This cohort was initially composed of 102 patients with SSc who were referred to the Scleroderma Clinics of the Rheumatic Disease Unit of the Gaetano Pini Institute of Milan. Since four of these patients dropped out from the follow-up, the present study was completed in 98 patients.
All enrolled patients met the ACR/EULAR classification criteria for SSc (4), and they were also sub-classified as having limited cutaneous SSc (lcSSc) or diffuse cutaneous SSc (dcSSc) according to the LeRoy et al. criteria (12). At the time of the enrolment, it was preliminarily established to include around half the patients with inactive disease (EScSG score < 3) and a similar proportion of
patients with active disease (EScSG score ≥ 3) (7).
Exclusion criteria were pre-existing conditions that may induce additional microvascular changes, such as diabetes, smoking and onychophagic habitus, presence of anti-phospholipid antibodies, and pregnancy (13-16). Another exclusion criterion was current treatment with beta-blockers which may exacerbate Raynaud’s phenomenon (RP) (17).
At the time of study enrolment, all of the patients were receiving low-dose acetylsalicylic acid and calcium channel blockers (CCBs), and in addition around one third of the patients were on treatment with other vasoactive agents (31 with monthly infusion of iloprost, 5 with weekly infusion of alprostadil, 7 with oral bosentan and 4 with sildenafil). None of the patients was taking anti-coagulant therapy. Three out of these latter patients had pulmonary hypertension. For ethical reasons these vasoactive therapies were maintained throughout the study.
NVC was performed in all the patients at enrolment time (T0) by using a videocapillaroscope with a x 200 magnification lens. All fingers of both hands, excluding thumbs, of each patient were examined by positioning each digit in such a way that the capillaroscopic light was 90 degrees incident on the centre of the nailfold. Four adjoining 1-mm fields—two on the right and two on the left side, starting from the middle of the nailfold and for a total extension of 4 mm—were examined (6,7). The derived digital images were stored using dedicated software (VideoCap; Scalar Co. Ltd., Tokyo, Japan). One experienced investigator (FP) was responsible for reviewing and scoring the stored NVC images of all the study patients, according to the NEMO definition (8).
Assessment of digital ulcers
All the enrolled patients were carefully observed every three months for the following two years, with particular attention given to the new appearance of at least one IDU in the distal fingers. A IDU was defined as a painful area, of at least 6 mm in diameter at its longest point, with visible depth and loss of dermis, located at the volar surface of the digit, distal to the proximal interphalangeal digital crease.
Statistical analysis was performed using MedCalc software package, 2014 version (MedCalc® Inc., Ostend, Belgium).
The Mann Whitney test was applied to compare T0-NEMO score values recorded in patients who subsequently developed IDUs to those assessed in patients who did not. This non-parametric method was adopted because the NEMO score variable did not have a normal distribution (Shapiro-Wilk test, p > 0.05).
Other variables as disease duration, history of previous ulcers and types of vasoactive therapies, were also tested for a possible association with the subsequent development of IDUs, using chi-square cross tabs for categorical variables and Mann-Whitney test for discrete variables.
Receiver operating characteristic (ROC) curves were constructed by plotting the sensitivity and 1-specificity values of the T0-NEMO scores in identifying patients who developed IDUs during the following two-year observation time. The area under the curve (AUC) was calculated together with the related 95% confidence intervals (CI) by applying the Hanley-McNeil test.
We also identified the sensitivity, specificity, positive (P) and negative (N) predictive values (PV) of the different T0-NEMO scores with the best performance in capturing patients who developed IDUs in the subsequent follow-up observation.
The Kaplan-Meier curve analysis and log rank test were used to evaluate the occurrence and the time of the appearance of IDUs during the follow-up in patients having different levels of T0- NEMO score.
Finally, a logistic regression model was tested in which the appearance of new digital ulcers in the follow-up represented the independent variable, while dependent variables were those that showed to be separately associated with the development of IDUs.