Background/Purpose: Nailfold capillaroscopy (NFC) is essential in the evaluation and classification of systemic sclerosis (SSc). Semiquantitative capillaroscopy scoring is a promising tool for assessing disease activity, severity and change in SSc, however there is no consensus yet over which capillaroscopy abnormalities should be analyzed and how. We aimed to investigate the reliability of the qualitative and semiquantitative scoring of NFC assessment between two raters and test-retest for each rater in a SSc cohort.

Methods: This is a pilot study from one EUSTAR centre, where 2 raters assessed the NFC images of 48 consecutive patients with SSc. Data were analyzed in 3 ways: step 1. qualitatively by ‘normal’ / ‘abnormal’ category, step 2. ‘early’, ‘active’, ‘late’ scleroderma patterns and unclassifiable in any pattern, and step 3.semiquantitatively by calculating the mean score for capillary loss, disorganization of the microvasculat array, giant capillaries, microhaemorrhages and capillary ramifications; combinations of giant capillaries and microhaemorrhages (as a surrogate for vascular activity) and disorganization and ramifications (surrogate for vascular damage) were also assessed. Variables for all steps were calculated for all fingers and for each finger. Inter-rater/intrarater agreement was assessed by Cohen’s kappa coefficients for qualitative variables and by intraclass correlation coefficients (ICC) for mean score values of abnormalities. Differences in scores between patients with digital ulcers (DUs) history and Bosentan treatment were analyzed by logistic regression.

Results: Interrater reliability ranged from good to excellent agreement for mean score values of abnormalities in all fingers (ICC coefficients 0.745 to 0.897) and was excellent for activity (ICC coefficient of 0.923) and damage combinations (ICC coefficient of 0.918). Assessment of abnormalities in a qualitative manner (normal/abnormal or with capillaroscopy patterns) showed weaker interrater agreement than the semiquantitative assessment (k coefficient <0.7). When scores were assessed in each finger, interrater reliability was good to excellent for mean scores of abnormalities and activity and damage combinations (ICC coefficients 0.781 to 0.867 for mean abnormalities scores and 0.713 to 0.856 for combinations), whereas for qualitative assessments interrater reliability was much weaker (k coefficients <0.7). Intrarater variability was good to excellent for mean score values of abnormalities and activity and damage combinations in all fingers and separate fingers for both raters; for qualitative assessment only one of the raters had good test-retest reliability. There was no difference in scores between patients with/without DUs history or Bosentan treatment.

Conclusion: Reliability of NFC assessment is essential in SSc trials/clinical practice to ensure quality of data. This pilot study demonstrates very good reliability between raters of the semiquantitative NFC assessment in a SSc cohort. Combinations of capillaroscopy abnormalities had very good reliability and might be preferred because they are less time consuming. A longitudinal assessment and confirmation of our results in other cohorts is needed.

« Back to 2016 ACR/ARHP Annual Meeting

ACR Meeting Abstracts - https://acrabstracts.org/abstract/excellent-reliability-of-semiquantitative-nailfold-capillaroscopy-assessment-in-a-systemic-sclerosis-cohort-a-pilot-study/