Research into best practice for capillaroscopy

Date: Wed 8th November 2017

Dr. Graham Dinsdale, Research Associate at the University of Manchester is at the American College of Rheumatology Conference in San Diego, California this week.

Dr. Dinsdale is presenting on his research study looking at how many fingers should be examined during a capillaroscopy (looking at the blood vessels in the nail bed) to identify scleroderma.

This research project was funded by SRUK as it is vital that the correct procedure be undertaken when conducting a capillaroscopy to ensure that people showing signs of scleroderma are not missed.

Background: Nailfold capillaroscopy plays an important role in diagnosing systemic sclerosis (SSc), with abnormal nailfold capillary appearance being included in the 2013 ACR/EULAR diagnostic criteria [1]. Common queries from clinicians who assess patients with suspected SSc are: which finger(s) should be imaged, and how many digits in total, given that assessing 8 fingers (excluding thumbs) takes time in a busy clinical practice? Our aim was to demonstrate the sensitivity of assessing different (combinations of) fingers for the presence of two markers of capillary abnormality: (1) presence of giant capillaries, and (2) overall image grade, compared to assessment of all 8 fingers.

Methods: Nailfold images (all fingers from each of 101 patients with SSc) and subsequent multiobserver assessments from a large study of quantitative capillaroscopy [2] were characterised by digit. Using custom software, observers counted giant vessels and graded the image overall (including normal/early/active/late). Patients were defined as “true case” for each of 2 parameters (giants, and image grade) if at least one of 8 fingers tested positive for the parameter (i.e. ≥ 1 giant vessels in one or more fingers, or one or more fingers given an 'abnormal' [early/active/late] grade). Seven single-finger, or finger combinations (derived from the middle and ring fingers), were then tested for sensitivity of achieving the correct result against the 8-finger “gold standard” true cases.

Results: For each of seven combinations of finger(s), sensitivity percentages for the two parameters are shown in Table 1. For the 8-finger “gold standard”, sensitivity against the diagnostic criteria was 53.0% (71 +ve cases from 134 assessments) and 73.1% (98 +ve cases from 134 assessments) for presence of giants and image grade, respectively. Pairs of fingers have higher sensitivity than single fingers in all cases, and the 4-finger combination shows a sensitivity of 85.9% and 91.8% for giants and image grade, respectively.

Conclusion: 1. Assessing only middle and ring fingers on both hands detects abnormality in 85-90% of cases of established SSc (halving imaging time). 2. Assessing only ring fingers (sensitivity 73-80%) brings a 75% reduction in imaging time. 3. Some cases of abnormality will be missed by not examining all fingers.

References 1. Van den Hoogen F, Khanna D, Fransen J, et al. Ann Rheum Dis 2013; 72: 1747-55. 2. Dinsdale G, Moore T, O'Leary N, et al. Microvascular Res 2017; 112: 1-6