Article Text
PDF
Abstract
Objective Construction of a simple nailfold videocapillaroscopic (NVC) scoring modality as a prognostic index for digital trophic lesions for day-to-day clinical use.
Methods An association with a single simple (semi)-quantitatively scored NVC parameter, mean score of capillary loss, was explored in 71 consecutive patients with systemic sclerosis (SSc), and reliable reduction in the number of investigated fields (F32–F16–F8–F4). The cut-off value of the prognostic index (mean score of capillary loss calculated over a reduced number of fields) for present/future digital trophic lesions was selected by receiver operating curve (ROC) analysis.
Results Reduction in the number of fields for mean score of capillary loss was reliable from F32 to F8 (intraclass correlation coefficient of F16/F32: 0.97; F8/F32: 0.90). Based on ROC analysis, a prognostic index (mean score of capillary loss as calculated over F8) with a cut-off value of 1.67 is proposed. This value has a sensitivity of 72.22/70.00, specificity of 70.59/69.77, positive likelihood ratio of 2.46/2.32 and a negative likelihood ratio of 0.39/0.43 for present/future digital trophic lesions.
Conclusions A simple prognostic index for digital trophic lesions for daily use in SSc clinics is proposed, limited to the mean score of capillary loss as calculated over eight fields (8 fingers, 1 field per finger).
Statistics from Altmetric.com
Introduction
Occlusive vasculopathy is the hallmark lesion of all forms of systemic sclerosis (SSc).1 2 This is clinically mirrored not only by Raynaud's phenomenon, but also by digital trophic lesions whose severity is reflected in the peripheral vascular category of the Disease Severity Scale (DSS) by Medsger et al.3
Digital microvascular disease in SSc has recently been studied by (semi)-quantitative nailfold videocapillaroscopy (NVC) as a complete and elaborate research tool.4,–,6 As a research tool, capillaroscopic (semi)-quantitative scoring has an indisputable value, but it has not yet found its way into day-to-day clinical practice. One of the reasons for this is the complexity of capillaroscopic scoring; either several hallmark parameters of the scleroderma pattern are scored (several of the following: capillary loss, giants, ramifications) or several fields are scored (eg, 32 fields (F32; 8 fingers, 4 fields per finger)).
Our aim was to investigate if there is a simple prognostic index for digital trophic lesions limited to one hallmark parameter (rather than several) and calculated over a smaller number of fields than 32. We first explored the association between a single (and most reliable) hallmark parameter—namely, (semi)-quantitatively scored capillary loss over eight fingers (F32)—and digital trophic lesions.7 We then investigated whether we could reliably reduce the number of investigated fields to (semi)-quantitatively score capillary loss. Lastly, we explored the cut-off value at which the prognostic index (mean score of capillary loss as calculated over a reduced number of fields) was of value to the clinician in being informative about digital trophic lesions.
Methods
Patients
Seventy-one consecutive patients with SSc (mean±SD age 52±14 years) visiting the Scleroderma Clinic of the Ghent University Hospital were enrolled (see file 1 in online supplement).
Nailfold videocapillaroscopy
Collection and blinding of the NVC images
Details are given in file 2 in the online supplement.
Mean score of capillary loss based on (semi)-quantitative scoring over 32 fields
(Semi)-quantitative scoring was performed as described previously.5 7 Capillary loss was counted over F32 (8 fingers, 4 fields per finger, fingers 2–5 of each hand) according to the gold standard (figure 1A). The described (semi)-quantitative rating scale (0, no changes; 1, <33% capillary reduction vs normal (9/mm); 2, 33–66% capillary reduction; 3, >66% capillary reduction) was adopted to calculate the mean score of capillary loss, as recently described elsewhere.7 8 This is a continuous variable ranging from 0 to 3.5 The mean scores of capillary loss were classified into mean score classes 1, 2 and 3 (1, <33% changes; 2, 33–66% changes; and 3, >66% changes) when categorical variables were needed.
Number of fields studied: gold standard (F32) and successive simplifications (F16–F8–F4). (A) F32: eight fingers (arrows), four fields of 1 mm per finger, giving a total of 32 fields. (B) F16: eight fingers (arrows), two fields of 1 mm per finger, giving a total of 16 fields. (C) F8: eight fingers (arrows), one field of 1 mm per finger, giving a total of eight fields. (D) F4: one finger (arrow), four fields of 1 mm in that finger, giving a total of four fields.
Evaluation of digital trophic lesions
The digital vascular disease was evaluated according to the peripheral vascular category of the DSS by Medsger (see file 3 in online supplement) at baseline and at a follow-up visit 6–12 months later.9 Digital trophic lesions were defined as stage 2 (pitting scars), stage 3 (digital tip ulcerations) or stage 4 (digital gangrene) of the peripheral vascular category of the DSS described by Medsger.
Statistical methods
The association between digital trophic lesions and capillary loss was investigated by the Fisher exact test. The reliable reduction in the number of fields for mean score of capillary loss (32 fields (F32)–16 fields (F16)–8 fields (F8)–4 fields (F4)) was assessed by absolute agreement intraclass correlation coefficient (ICC) based on two-way mixed model analyses (figure 2). Receiver operating curve (ROC) analysis was performed with the variable mean score of capillary loss versus present/future digital trophic lesions. The cut-off value of mean score of capillary loss as a clinical prognostic index for present/future digital trophic lesions was based on a trade-off between accuracy and clinical usefulness. Statistical analyses were performed using SPSS Version 17.0 (SPSS, Chicago, Illinois, USA) and MedCalc 10.1.8.0 (MedCalc Software, Mariakerke, Belgium).
Reliability of simplification of the numbers of semi-quantitatively scored fields. Scatterplots investigating the relationship between mean score of capillary loss as calculated over F32/F16, F32/F8, F32/F4.
Results
Characteristics of the patients
A description of the characteristics of the patients is given in file 4 in the online supplement.
Association between mean score of capillary loss and digital trophic lesions
A statistically significant and clinically meaningful association was found between the mean score class of capillary loss, as calculated over F32, and the presence of digital trophic lesions (p=0.004). Moreover, patients with digital trophic skin lesions were only found in mean score classes 2 and 3 (table 1). Nineteen of 55 patients (35%) with mean score classes 2 and 3 had digital trophic lesions compared with no patients (0%) with mean score class 1.
Association between progressive stages of capillary loss (mean score classes 2 and 3) and the presence of digital trophic lesions in the patients with systemic sclerosis
Reliable reduction in the number of fields investigated to obtain the mean score capillary loss
The mean score of capillary loss was calculated over F32 (8 fingers, 4 fields per finger), F16 (8 fingers, 2 fields per finger), F8 (8 fingers, 1 field per finger) and finally F4 (1 finger, 4 fields) (figure 1A–D). Reliability analysis was performed to assess whether the simplified approaches were reliable gauges for the score, as evaluated over 32 fields. The scatterplots of the gold standard (F32) and the simplifications can be seen in figure 2. The ICC of F16/F32 was 0.97 (95% CI 0.95 to 0.98), of F8/F32 was 0.90 (95% CI 0.84 to 0.93) and of F4/F32 was 0.76 (95% CI 0.63 to 0.85).10 Calculation of the mean score of capillary loss could be reliably reduced from eight fingers (32 fields, 4 fields per finger) to eight fingers (8 fields, 1 field per finger), as shown by the ICC.10
Cut-off of mean score of capillary loss as a clinical prognostic index for present/future digital trophic lesions
The performance of the mean score of capillary loss to detect present/future digital trophic lesions can be seen in the ROC analysis in figure 3 in file 5 in the online supplement. ROC analyses were calculated for mean score of capillary loss by investigating F32–F16–F8–F4. The area under the curve, 95% CI (see table SI in file 5 in the online supplement), sensitivity and specificity corroborate the reduction in the number of investigated fields of mean score value of capillary loss to F8. A mean score value of capillary loss of 1.67 (see table SII in file 5 in the online supplement) was chosen as the cut-off value for the clinical prognostic index and had a sensitivity of 72.22/70.00, specificity of 70.59/69.77, positive likelihood ratio of 2.46/2.32 and negative likelihood ratio of 0.39/0.43 for present/future digital trophic lesions.
Discussion
This is the first study to show that a simple NVC scoring modality—namely, the calculation of mean score of capillary loss—investigated over eight fields (8 fingers, 1 field per finger) can be constructed and used as a simple prognostic index for digital trophic lesions in SSc-related microvascular disease, which makes it possible to use it as a day-to-day in-office clinical tool.
This simple NVC scoring modality is now constructed with one single parameter (rather than several parameters which are used in NVC as a research tool) and the scoring of a reduced number of fields (rather than 32). The advantage of capillary loss as a single parameter is twofold. On the one hand it allows interested clinical rheumatologists to use this tool as the mere counting of capillaries does not require much training. On the other hand, it is a parameter that can be easily applied with different techniques that are currently being used to perform capillary microscopy, such as automated capillary counting.11 Reliability analysis (figure 2) shows that the number of investigated fields (for this single parameter) can be restricted from 32 fields (F32) (8 fingers, 4 fields per finger) over 16 fields (F16) (8 fingers, 2 fields per finger) down to 8 fields (F8) (8 fingers, 1 field per finger) (figure 1A–C).
The clinical value of this study lies in the fact that an association with future digital trophic lesions has been established. To date, no simple tool has been available to help the clinical rheumatologist in decision making concerning digital trophic lesions. The mean score of capillary loss as calculated over F8 may be used as a clinical prognostic index. Based on ROC analysis, the cut-off value of 1.67 has a sensitivity and specificity of about 70% in informing the clinician of pending or future digital trophic lesions. This may well help the clinician to make a therapeutic decision.
However, this index still needs to be validated. We also need to draw attention to the fact that it was not the purpose of this study to substitute or compete with already existing research tools which have better performance, sensitivity and specificity. For clinical research and therapeutic trials, the numbers of investigated fields and parameters must be optimal and multicentre efforts are being made to obtain such an index.4 5 12
In conclusion, we have constructed, for the first time, a simple NVC scoring modality limited to the mean score of capillary loss calculated over eight fields that can be used as a day-to-day in-office tool as a clinical prognostic index for present and future digital trophic lesions.
Acknowledgments
The authors thank Melissa De Decker for her unconditional daily support, their colleagues for sending their patients to the Scleroderma Clinic and their senior professors for making this collaboration possible.
References
Supplementary materials
Web Only Data
Files in this Data Supplement:
Footnotes
VS, FDeK and CP contributed equally to this work.
Funding This capillaroscopic study has been made possible by a grant from the ‘Fonds voor Wetenschappelijk Reuma Onderzoek/Fonds de la Recherche Scientifique en Rhumatologie’. JTVanP is supported by a research grant from the Fund for Scientific Research-Flanders. SD is supported by a grant from the Rotary – National Association to support Disabled People.
Competing interests None.
Ethics approval This study was conducted with the approval of Ghent University Hospital.
Provenance and peer review Not commissioned; externally peer reviewed.