Presentation: Utility and Reliability of Different Capillaroscopic Measures for the Identification of Secondary Raynaud’s Phenomenon (2007)

12 Utility and Reliability of Different Capillaroscopic Measures for the Identification of Secondary Raynaud’s Phenomenon

Objective. Nailfold videocapillaroscopy (NVC) is a widely accepted test used in the differential diagnosis of Raynaud’s phenomenon. Because the overall level of agreement regarding which and how many parameters have to be studied during the examination is surprisingly thin, the goal of our study has been to analyze the inter- and intra-observer agreement and the predictive value of main NVC measurements.
Methods. Two hundred and seventeen consecutive subjects were studied. The subjects were classified as patients with or without a specific diagnosis before NVC. The second group of subjects were followed up for 1 to 8 years after the capillaroscopy was done to ascertain their health status. In all cases the following parameters were evaluated: number of capillaries, intercapillary distance, avascular areas, capillary disorganization, length of the capillary, total width, percentage of minor abnormalities (tortuousity, crossed, enlarged), percentage of major abnormalities (megacapillaries, bushy, meandering, branching), haemorrhages, skin transparency, and subpapillary plexus visibility. All images taken were coded and analysed blindly by two observers.
Results. A total of 877 nailfold images were analyzed. The Receiver Operating Characteristic (ROC) curve analysis could not detect any significant difference in diagnostic performance between the number of capillaries/mm (area under the curve 0.645; 95% C.I. 0.542-0.740), the interpeak distance (area under the curve 0.539; 95% C.I. 0.435-0.641) or the avascular areas (area under the curve 0.637; 95% C.I. 0.533-0.732). The inter and intra-observer agreement, sensibility, specificity, positive and negative likelihood ratio and predictive value of all the other parameters are shown in the table below.
Conclusions. This study has shown that an analytical and objective approach of NVC can be useful for quantitative and reproducible recording of different parameters. We suggest that a scoring method in which several parameters are combined by a semiquantitative approach could be the most reliable method to analyse the nailfold images. This combined analysis should help the reproducibility and widespread applicability of this procedure in clinical practice.
Statistical analysis of various capillaroscopic parameters
Statistical analysis of various capillaroscopic parameters
ReproducibilityRepeatabilitySensibilitySpecificity+ LR- LR+PV- PV
Number of capillaries/mm0.80.688.220.61.10.537.576.5
Intercapillary distance0.80.282.425.41.10.637.372.7
Avascularity0.80.852.985.73.70.566.777.1
Total width0.080.0855.950.81.10.838.068.1
% of minor abnormalities0.50.376.534.91.10.638.873.3
% of major abnormalities0.90.826.581.01.390.942.967.1
Capillary disorganization0.80.741.269.81.30.842.468.7
Haemorrhages0.90.711.892.11.40.944.465.9
Skin transparency0.60.385.325.41.10.538.276.2
Subpapillary plexus visibility0.90.870.633.31.00.836.467.7
LR: likelihood ratio; PV: predictive value

 F. Ingegnoli, None.