Nailfold capillaroscopy - how many fingers should be examined to detect abnormality?

Graham Dinsdale1, Chris Roberts2, Tonia Moore3, Joanne Manning3, Michael Berks4, John Allen5, Marina E Anderson6, Maurizio Cutolo7, Roger Hesselstrand8, Kevin Howell9, Carmen Pizzorni7, Vanessa Smith10, Alberto Sulli7, Marie Wildt8, Christopher Taylor4, Andrea Murray1, Ariane L Herrick1,11

1. Division of Musculoskeletal & Dermatological Sciences, University of Manchester, Salford Royal Hospital NHS Foundation Trust, Manchester Academic Health Science Centre, Manchester, UK

2. Centre for Biostatistics, Division of Population Health, Health Services Research & Primary Care, University of Manchester, Manchester, UK

3. Salford Royal Hospital NHS Foundation Trust, Salford, UK

4. Centre for Imaging Sciences, Division of Informatics, Imaging & Data Sciences, University of Manchester, Manchester, UK

5. Microvascular Diagnostics, Northern Medical Physics and Clinical Engineering, Freeman Hospital, Newcastle upon Tyne, UK

6. Institute of Ageing and Chronic Disease, University of Liverpool, Liverpool, UK

7. Research Laboratory and Academic Division of Clinical Rheumatology, Dept. Internal Medicine, University of Genova, Italy

8. Department of Clinical Sciences, Section of Rheumatology, Lund University, Lund, Sweden

9. Institute of Immunity and Transplantation, University College London, Royal Free Campus, London, UK

10. Department of Rheumatology, Ghent University Hospital, Faculty of Internal Medicine, Ghent University, Ghent, Belgium

11. NIHR Manchester Musculoskeletal Biomedical Research Centre, Central Manchester NHS Foundation Trust, Manchester Academic Health Science Centre, Manchester, UK.

Corresponding author: Dr Graham Dinsdale, Clinical Sciences Building, Salford Royal Hospital, Stott Lane, Salford, M6 8HD, UK. Email: graham.dinsdale@manchester.ac.uk. Telephone: (+44) (0)161 206 2935

Word count (excluding title page, abstract, references, figures and tables): 2,000

Competing interests: The authors declare no conflict of interests.

1

ABSTRACT

Objectives: Nailfold capillaroscopy is being increasingly used by rheumatologists in the diagnosis of systemic sclerosis (SSc). However, assessment of all nailfolds can be time-consuming in a busy outpatient clinic. Our aim was to answer the question as to how many (and which) fingers a clinician should routinely assess to capture accurately the 'true' state.

Methods: 2994 assessments (by an international panel of expert observers) of 1600 images from 173 participants (101 with SSc, 22 with primary Raynaud's phenomenon, and 50 healthy controls) were included in this analysis. Seven single-finger, or finger combinations (derived from the middle and ring fingers), were then tested for sensitivity for the presence of two markers of capillary abnormality (presence of giant capillaries and an SSc 'grade' [early, active or late]), compared to assessment of all 8 fingers.

Results: For the 8-finger “gold standard”, sensitivity against the diagnostic criteria was 74.6% (53.0% for presence of giants alone and 73.1% for image grade alone). Examining only one finger gave low sensitivity (ranging from right middle 31.7% to left ring 46.6%). Examining both ring fingers gave a sensitivity of 59.8%, whereas examining the 4-finger combination of both ring and both middle fingers gave a sensitivity of 66.7%.

Conclusion: During routine capillaroscopic examination all eight nailbeds (excluding thumbs) should ideally be examined, otherwise some abnormalities will be missed. Examining only four fingers reduces capillaroscopy sensitivity.

KEY WORDS: Capillaroscopy, diagnosis, Raynaud's phenomenon, systemic sclerosis.

KEY MESSAGES

1. All eight nailbeds (excluding thumbs) should be examined to diagnose systemic sclerosis

2. Examining fewer than eight nailbeds reduces the chance of detecting capillary abnormalities.

3. Examining capillaries from only four nailbeds reduces sensitivity from 75% to 67%

2

INTRODUCTION

Abnormal nailfold capillaries are one of the 2013 American College of Rheumatology (ACR)/European League Against Rheumatism (EULAR) classification criteria for systemic sclerosis (SSc)[1,2]. Typical nailfold capillary abnormalities include widened capillaries (including 'giant' capillaries, i.e. homogenously enlarged vessels with a diameter of at least 50µm, considered pathognomonic of SSc), areas of avascularity, distortion of the normal nailfold architecture, and haemorrhages[3-5]. In the patient presenting with Raynaud's phenomenon (RP), abnormal nailfold capillaries predict SSc[6,7], and are therefore crucial in early diagnosis[8-10].

The gold standard for nailfold capillaroscopy is to examine eight fingers, omitting thumbs, where it is more difficult to visualise/classify capillaries[11]. It is generally accepted that if abnormality exists in any finger/nailbed, then that patient has 'abnormal nailfold capillaries'. Degrees of abnormality have been graded according to the 'early', 'active' and 'late' patterns defined by Cutolo et al [5], with a patient given the most severe grade observed across all eight fingers. As the use of capillaroscopy becomes more widespread, a key question for busy clinicians is 'How many fingers do we need to examine?' If, for example, abnormality could be diagnosed by examining two or four (as opposed to eight) nailfolds, then this would save time. The problem is that nailfold morphology can vary markedly between nailfolds, as exemplified in Supplementary Figure 1 where abnormalities are only visible in some but not all fingers. The aim of this study was to answer the question as to how many (and which) fingers a clinician should routinely assess with nailfold capillaroscopy in order to be sure that the resulting overall picture accurately represents the 'true' state of the nailfold capillaries. Specific objectives were to:

1.Examine different fingers and combinations of fingers, comparing results to the 'gold standard' of all eight fingers.

2.Examine different fingers and combination of fingers, comparing results in relation to diagnostic groups (healthy controls, primary Raynaud's phenomenon [PRP], SSc).

PATIENTS AND METHODS

Patients

A total of 123 patients (101 with SSc, 22 with PRP), and 50 healthy controls (HC), were recruited at Salford Royal NHS Foundation Trust, a tertiary referral centre for SSc, and imaged at a baseline visit as part of a previously-reported study examining reliability of nailfold capillary assessment[12,13]. Imaging sessions were carried out after 20 minutes acclimatisation at 23°C in a temperature/humidity-controlled laboratory. Participants were asked to refrain from caffeine and smoking for 4 hours before imaging. All participants gave informed, written consent. The study was approved by the Greater Manchester East Research Ethics committee (reference: 11/NW/0444).

Image acquisition

3

At each study visit, panoramic nailfold videocapillaroscopy images were captured using a microscope with 300x magnification (KK Technology, Honiton, UK). Green LED illumination was used to ensure maximum capillary contrast. Images were recorded from all 10 digits, including the thumbs. Custom capture software allowed grayscale images to be recorded as panoramic mosaics across the entire nailfold (to avoid potential issues relating to any within-nailfold heterogeneity, as demonstrated in Supplementary Figure 1) by registering and then digitally stitching individual camera frames[14,15].

Image analysis and grading

Images were assessed in a random order by 10 capillaroscopy experts (hereafter “observers”) from 4 European countries, who were blinded to the image’s metadata (e.g. participant group, finger). A custom software tool, written specifically for this study, was used to assess the images, as previously described[12,13]. In this analysis we only consider 2 nailfold parameters: (1) the number of giant vessels (homogenously enlarged vessels with a diameter >50 µm) recorded in each nailfold, and (2) overall image grade (normal, early, active, late, non-specific or two different ungradeable categories). The early/active/late classification criteria used for overall image grade was as defined by Cutolo et al[5]. Abnormality was defined independently for the two parameters considered: (1) the presence of one or more giant capillaries in an image, and (2) the identification of any one of the three SSc patterns of 'early', 'active' or 'late'. An observer could choose to assess one or both parameters. There were 2994 assessments of 1600 images from 173 participants included in this analysis.

Finger combinations

Since it would be impractical to examine all possible combinations and numbers of nailfolds (for example, there are 45 'two-finger combinations' that can be selected from ten nailfolds) we chose to concentrate on the middle and ring fingers. This was because (a) a recent multicentre study[16] suggested the middle finger of the dominant hand gives the best capillaroscopic prediction of digital ulcer risk, and (b) early capillaroscopy work suggested that appearances in the ring fingers best discriminated between PRP and SSc[17].

Statistical analysis

We investigated factors that might influence the probability of observing either a giant vessel in an image and/or an abnormal image grade using a mixed effects logistic regression (MELR) model with a random effect for participant including fixed covariates for age, sex and diagnostic group of the participant, finger (index, middle, ring, little) and hand (left or right) of the image and rater. As the purpose of these analyses is to assess the probability of detecting apparent abnormality whether correctly or incorrectly, we classified images graded as non-specific as a “failure to detect” or “negative”, so they are included in the analysis. Marginal probabilities for the detection of giant vessels or abnormality in images, that is the effect averaged across other model covariates, were

4

estimated from the model[18]. All statistical analyses were carried out using Stata 14 (StataCorp. 2015. College Station, TX).

RESULTS

Table 1 gives frequencies for the presence of giant capillaries and for image grade, subdivided by digit. Table 1 also records the proportion of ‘unclassifiable’ images. An extended version of Table 1, with data further subdivided by clinical group, is available as supplementary data.

Unclassifiable images

For the assessment of giant vessels, 21.5% (644/2,994) of image assessments were unclassifiable. When the MELR model was fitted there was evidence that the unclassifiable proportion varied between fingers. The ring finger (p≤0.001) and the little finger (p<0.001) were less likely to be missing than the thumb, and observations from the left hand were more likely to be available than from the right (p=0.012).

For the assessment of image grade, 56.8% (1,700/2,994) of image assessments were unclassifiable. When a MELR model was fitted thumbs were more likely to be unclassified than other fingers (p<0.001) with the ring finger being least likely to be unclassified, followed by the little finger. There was no difference between hands.

Detection of abnormality using visualization of giant capillaries or an abnormal image grade

Dichotomous variables were constructed for: (1) the observation of giant vessels in an image, and (2) an overall assessment being abnormal (one of ‘early’, ‘active’, or ‘late’).

Presence of giant vessels. When a MELR model was fitted to the dichotomous variable for giant vessels, there was evidence (p<0.001) of a difference between fingers with the thumb having the lowest abnormality rate of 4.4% (95% CI: 2.5% to 6.2%). Estimates from the MELR model after adjustment for other covariates suggest that giant vessels were most likely present and detectable in the ring finger. The abnormality rates (95% CI) for other fingers were 9.2% (6.3% to 12.0%), 10.7% (7.6% to 13.7), 14.7% (11.2% to 18.3%), and 10.8% (7.7% to 13.9%) for the index, middle, ring and little fingers respectively. There was a substantial difference between clinical groups (p<0.001) with the marginal estimate of giant vessel abnormality rate (95% CI) being 15.5% (11.3% to 19.7%) for SSc, dropping to 3.5% (0.3% to 6.8%) for PRP, and 1.0% (0 to 2.0%) for healthy controls.

Abnormal image grade. Again there was evidence of a difference between fingers (p<0.001) with thumbs having the lowest abnormality rate (14.0%, 95% CI: 10.8% to 17.2%). The abnormality rates (95% CI) for other fingers were 18.9% (15.2% to 22.5%), 18.8% (15.2% to 22.5%), 19.7% (18.1% to

5

25.8%), 26.2% (22.1% to 30.3%), and 27.7% (23.5% to 31.9%) for the index, middle, ring and little fingers respectively. Again, there was a substantial difference between clinical groups with the marginal estimate of image grade abnormality rate (95% CI) being 32.8% (27.3% to 38.4%) for SSc, dropping to 9.0% (3.5% to 14.4%) for PRP, and 3.1% (1.0% to 5.3%) for healthy controls.

Selected finger combinations – sensitivity, specificity, and reliability

Table 2 gives the frequency and percentage of participants with: (a) one or more giant capillary, (b) an abnormal image grade, and (c) a composite measure of (a) and (b) for pre-specified fingers or combinations of fingers by diagnostic group. The rates of detection of the three measures (a, b and c above) in the SSc group can be thought of as the sensitivity, whereas the failure to detect (a, b and c) in the PRP and healthy control groups can be thought of as the specificity. For example: considering giant vessels, in the left ring finger there is a sensitivity of 26.7% and a specificity of 98.7% for healthy controls and 93.7% for PRPs. Considering any middle or ring finger as a diagnostic criteria, the corresponding sensitivity is 43.8% and specificities are respectively 95.5% and 85.7% for patients with PRP and healthy controls. And considering image grade, for any middle or ring finger, the sensitivity is 65.4% with specificities 90.9% and 78.6% for healthy controls and patients with PRP respectively. When giant vessels and image grade are combined (again for any middle or ring finger), the sensitivity is 66.7% with specificities 90.9% and 76.2% for healthy controls and patients with PRP respectively. In comparison, when giant vessels and image grade were combined for the 8-finger “gold standard”, sensitivity against the diagnostic criteria was 74.6% (53.0% for presence of giants alone and 73.1% for image grade alone), with specificities of 86.9% and 70.7% for healthy controls and patients with PRP respectively.

DISCUSSION

Our results suggest that examining fewer than eight nailfolds (symmetrically 2 or 4) reduces sensitivity to detect capillary abnormalities. However, if a clinician is pressed for time, then the best two-finger combination is both ring fingers with a sensitivity of 59.8% for detecting either giant vessels or an abnormal image grade. A four-finger combination of both middle and ring fingers increases sensitivity to 66.7%. These values compare to an eight finger gold standard correctly identifying abnormality (giants or abnormal grade) in 74.6% of SSc cases. If only one finger is examined, then examining the non-dominant ring gives only 46.6% sensitivity: therefore at least two fingers should be examined. It can be noted from the data in table 2 that the sensitivities are only slightly improved by adding the detection of giant capillaries to the abnormal image grade (i.e. going from section (b) to section (c) in table 2). Sensitivities for detection of giant capillaries alone are lower for all finger combinations than for abnormal image grade alone. Although in practice clinicians are more likely to look at image grade rather than at giant capillaries alone, nonetheless giant capillaries are useful in diagnosis, especially for those less experienced in capillaroscopy, and are very specific for SSc (almost never occurring in healthy controls) as demonstrated in table 2. The occurrence of giant capillaries in 22% of patients with PRP is surprising and may reflect how some patients with PRP, especially at a tertiary centre, may be in early transition to a SSc-spectrum

6

disorder. Additionally (although not specially addressed in this study), giant capillaries are the most easily visible abnormal feature when using lower magnification capillaroscopy systems.

The optimal number of nailfolds examined is likely to be dependent upon the question being asked. To predict digital ulcer risk using capillaroscopy, one study[19] suggested that 8 fingers (at least 1 field per finger) should be examined whereas another suggested that examining the middle finger of the dominant hand was sufficient[16]. Other work examining treatment response advocated imaging 8 fingers[20].

In conclusion, this analysis of a large number of 'whole nailfold' images demonstrates that to detect abnormality all eight nailbeds (excluding thumbs) should be examined, since there is a drop in sensitivity if only four fingers are examined (from 74.6% to 66.7%). The relevance of our findings will depend on local arrangements for performing capillaroscopy, but allows informed choices to be made.

FUNDING

This work was supported by Scleroderma and Raynaud's UK, grant number MU3.

7

REFERENCES

1. Van den Hoogen F, Khanna D, Fransen J, Johnson SR, Baron M, Tyndall A, et al. 2013 Classification

criteria for systemic sclerosis. Arthritis Rheum. 2013; 65: 2737-47.

2. Van den Hoogen F, Khanna D, Fransen J, Johnson SR, Baron M, Tyndall A, et al. 2013 Classification

criteria for systemic sclerosis. Ann Rheum Dis 2013; 72: 1747-55.

3. Maricq HR, LeRoy EC. Patterns of finger capillary abnormalities in connective tissue disease by

‘widefield’ microscopy. Arthritis Rheum 1973;16:619–28.

4. Herrick AL, Cutolo M. Clinical implications from capillaroscopic analysis in patients with Raynaud’s

phenomenon and systemic sclerosis. Arthritis Rheum 2010;62:2595–604.

5. Cutolo M, Sulli A, Smith V. How to perform and interpret capillaroscopy. Best Practice Res Clin

Rheumatol 2013; 27: 237-48.

6. Koenig M, Joyal F, Fritzler MJ, Roussin A, Abrahamowicz M, Boire G, et al. Autoantibodies and

microvascular damage are independent predictive factors for the progression of Raynaud's

phenomenon to systemic sclerosis: a twenty-year prospective study of 586 patients, with validation

of proposed criteria for early systemic sclerosis. Arthritis Rheum 2008;58:3902-3912.

7. Ingegnoli F, Boracchi P, Gualtierotti R, Biganzoli EM, Zeni S, Lubatti C, et al. Improving outcome

prediction of systemic sclerosis from isolated Raynaud’s phenomenon: role of autoantibodies and

nail-fold capillaroscopy. Rheumatology 2010; 49, 797–805.

8

8. LeRoy EC, Medsger TA. Criteria for the classification of early systemic sclerosis. J Rheumatol.

2001;28:1573–6.

9. Matucci-Cerinic M, Allanore Y, Czirják L, Tyndall A, Muller-Ladner U, Denton C, et al. The challenge

of early systemic sclerosis for the EULAR Scleroderma Trial and Research Group (EUSTAR)

community. It is time to cut the Gordian knot and develop a prevention or rescue strategy. Ann

Rheum Dis 2009;68:1377–80.

10. Avouac J, Fransen J, Walker UA, Riccieri V, Smith V, Muller C, et al: Preliminary criteria for the

very early diagnosis of systemic sclerosis: results of a Delphi Consensus Study from EULAR

Scleroderma Trials and Research Group. Ann Rheum Dis 2011 ;70:476-81.

11. Hughes M, Moore T, O’Leary N, Tracey A, Ennis H, Dinsdale G, et al. A study comparing

videocapillaroscopy and dermoscopy in the assessment of nailfold capillaries in patients with

systemic sclerosis-spectrum disorders. Rheumatology 2015; 54: 1435-42.

12. Dinsdale G, Moore T, O'Leary N, Tresadern P, Berks M, Roberts C, et al. Intra-and inter-observer

reliability of nailfold videocapillaroscopy — A possible outcome measure for systemic sclerosis-

related microangiopathy. Microvascular Res 2017; 112: 16.

13. Dinsdale G, Moore T, O’Leary N, Berks M, Roberts C, Manning J, et al Quantitative outcome

measures for systemic sclerosis-related microangiopathy – reliability of image acquisition in nailfold

capillaroscopy. Microvascular Res 2017;113:56-9.

9

14. Allen PD, Taylor CJ, Herrick AL, Moore T. Image analysis of nailfold capillary patterns from video

sequences. In International Conference on Medical Image Computing and Computer-Assisted

Intervention 1999 Sep 19 (pp. 698-705). Springer, Berlin, Heidelberg.

15. Anderson ME, Allen PD, Moore T, Hillier V, Taylor CJ, Herrick AL. Computerised nailfold video

capillaroscopy - a new tool for the assessment of Raynaud's phenomenon. J Rheumatol 2005; 32:

841-8.

16. Cutolo M, Herrick AL, Distler O, Becker MO, Beltran E, Carpentier P, et al. Nailfold

videocapillaroscopic features and other clinical risk factors for digital ulcers in systemic sclerosis: a

multicenter, prospective cohort study. Arthritis Rheumatol 2016; 68: 2527-39.

17. Houtman PM, Kallenberg CG, Fidler V, Wouda AA. Diagnostic significance of nailfold capillary

patterns in patients with Raynaud's phenomenon. An analysis of patterns discriminating patients

with and without connective tissue disease. J Rheumatol 1986; 13: 556-63.

18. Bartus T. Estimation of marginal effects using margeff. Stata J 2005;5(3):309-29.

19. Smith V, De Keyser F, Pizzorni C, Van Praet JT, Decuman S, Sulli A, et al. Nailfold capillaroscopy

for day-to-day clinical use: construction of a simple scoring modality as a clinical prognostic index for

digital trophic lesions. Ann Rheum Dis. 2011; 70(1):180-3.

20. Trombetta AC, Pizzorni C, Ruaro B, Paolino S, Sulli A, Smith V, et al. Effects of longterm treatment

with bosentan and iloprost on nailfold absolute capillary number, fingertip blood perfusion, and

clinical status in systemic sclerosis. J Rheumatol. 2016; 43(11):2033-2041.

10

Supplementary Figure 1. Nailfold images of 8 digits (excluding thumbs) from one patient with SSc. There are clear signs of abnormality in the right-index (enlarged vessels), left-ring (derangement/avascularity), and left-little (avascularity). Other images contain vessels that appear generally normal. Additionally, the right-index image shows why it is important to image the whole nailfold: examining only the left section (where capillaries are within normal limits) could lead to incorrect interpretation of the nailbed.

11

Table 1. Frequencies and percentages of the occurrence of two nailfold parameters (presence of giant vessels, and image grade) in all participants by digit. Values are expressed as frequency (%). An extended version of this table, with data subdivided by participant group (controls, patients with primary Raynaud’s, patients with SSc), is available as supplementary data.

Left hand [frequency (%)] Right hand [frequency (%)]

Thumb Index Middle Ring Little Thumb Index Middle Ring Little Total

Giants

No 213 (70.5) 194 (64.9) 200 (65.8) 208 (68.4) 217 (70.9) 211 (69.2) 192 (63.8) 185 (63.4) 208 (68.0) 195 (70.9) 2023 (67.6)

Yes 17 (5.6) 32 (10.7) 40 (13.2) 51 (16.8) 34 (11.1) 13 (4.3) 29 (9.6) 30 (10.3) 46 (15.0) 35 (12.7) 327 (10.9)

Unclassifiable 72 (23.8) 73 (24.4) 64 (21.1) 45 (14.8) 55 (18.0) 81 (26.6) 80 (26.6) 77 (26.4) 52 (17.0) 45 (16.4) 644 (21.5)

Image grade

Normal 53 (17.5) 48 (16.1) 65 (21.4) 65 (21.4) 61 (19.9) 53 (17.4) 53 (17.6) 59 (20.2) 78 (25.5) 56 (20.4) 591 (19.7)

Early 15 (5.0) 19 (6.4) 22 (7.2) 16 (5.3) 18 (5.9) 11 (3.6) 17 (5.6) 14 (4.8) 24 (7.8) 24 (8.7) 180 (6.0)

Active 18 (6.0) 28 (9.4) 30 (9.9) 51 (16.8) 31 (10.1) 9 (3.0) 17 (5.6) 28 (9.6) 32 (10.5) 29 (10.5) 273 (9.1)

Late 16 (5.3) 19 (6.4) 29 (9.5) 23 (7.6) 36 (11.8) 23 (7.5) 23 (7.6) 20 (6.8) 25 (8.2) 36 (13.1) 250 (8.4)

Unclassifiable 200 (66.2) 185 (61.9) 158 (52.0) 149 (49.0) 160 (52.3) 209 (68.5) 191 (63.5) 171 (58.6) 147 (48.0) 130 (47.3) 1700 (56.8)

Number of assessments

302(100.0)

299 (100.0) 304 (100.0) 304 (100.0) 306(100.0)

305 (100.0) 301(100.0)

292 (100.0) 306 (100.0) 275(100.0)

2994 (100.0)

Number of images 159 (9.9) 161 (10.1) 163 (10.2) 163 (10.2) 160 (10.0) 160 (10.0) 159 (9.9) 159 (9.9) 163 (10.2) 153 (9.6) 1600 (100.0)

12

Table 2. Frequencies and percentages of participants with: (a) one or more giant capillary, (b) an abnormal image grade, and (c) a composite measure of (a) and (b), for pre-specified fingers or combinations of fingers broken down by participant group. The ‘Any Finger’ combination refers to the 8-finger “gold standard”.

HC PRP SSc

Freq. (%)

Assessments

(n)

Participants(n) Freq. (%)

Assessments

(n)

Participants(n) Freq. (%)

Assessments

(n)

Participants

(n)

(a) Presence of Giant

Ring Left 1 (1.3) 80 47 3 (6.3) 48 22 47 (26.7) 176 94

Ring Right 3 (3.7) 81 49 1 (2.2) 45 22 42 (23.3) 180 92

Either Ring 3 (4.0) 75 47 3 (7.0) 43 22 61 (37.2) 164 88

Middle Left 1 (1.3) 77 47 2 (4.4) 45 22 37 (20.3) 182 94

Middle Right 1 (1.3) 80 47 3 (6.7) 45 22 26 (15.6) 167 90

Either Middle 1 (1.4) 73 45 4 (9.5) 42 22 46 (28.4) 162 88

Any Middle or Ring 3 (4.5) 66 44 6 (14.3) 42 22 67 (43.8

) 153 85

Any Finger 3 (4.9) 61 41 9 (22.0) 41 21 71 (53.0

) 134 75

(b) Abnormal Grade

Ring Left 3 (3.8) 80 47 6 (12.5) 48 22 81 (46.0

) 176 94

Ring Right 6 (7.4) 81 49 4 (8.9) 45 22 71 (39.4) 180 92

Either Ring 6 (8.0) 75 47 6 (14.0) 43 22 94 (57.3

) 164 88

Middle Left 4 (5.2) 77 47 5 (11.1) 45 22 72 (39.6

) 182 94

Middle Right 4 (5.0) 80 47 6 (13.3) 45 22 52 (31.1

) 167 90

Either Middle 5 (6.8) 73 45 7 (16.7) 42 22 87 (53.7

) 162 88

Any Middle or Ring 6 (9.1) 66 44 9 (21.4) 42 22 100 (65.4

) 153 85

Any Finger 8 (13.1) 61 41 11 (26.8

) 41 21 98 (73.1) 134 75

(c) Giant or Abnormal Grade

Ring Left 3 (3.8) 80 47 7 (14.6) 48 22 82 (46.6

) 176 94

Ring Right 6 (7.4) 81 49 4 (8.9) 45 22 77 (42.8) 180 92

Either Ring 6 (8.0) 75 47 7 (16.3) 43 22 98 (59.8

) 164 88

Middle Left 4 (5.2) 77 47 5 (11.1) 45 22 73 (40.1

) 182 94

Middle Right 4 (5.0) 80 47 6 (13.3) 45 22 53 (31.7

) 167 90

Either Middle 5 (6.8) 73 45 7 (16.7) 42 22 89 (54.9

) 162 88

Any Middle or Ring 6 (9.1) 66 44 10 (23.8) 42 22 102 (66.7

) 153 85

Any Finger 8 (13.1) 61 41 12 (29.3

) 41 21 100 (74.6) 134 75

13

14

Supplementary Table 1. Frequencies and percentages of the occurrence of two nailfold parameters (presence of giant vessels, and image grade) in each participant group, broken down by digit. Values are expressed as either ‘frequency (%)’ or ‘number’..

Left hand Right hand

Thumb Index Middle Ring Little Thumb Index Middle Ring Little TotalGiants

Healthy ControlNo 60 (77.9) 54 (74.0) 60 (77.9) 69 (86.3) 66 (83.5) 51 (68.9) 60 (71.4) 57 (71.3) 67 (82.7) 61 (82.4) 605 (77.7)Yes 0 (0.0) 1 (1.4) 1 (1.3) 1 (1.3) 1 (1.3) 0 (0.0) 1 (1.2) 1 (1.3) 3 (3.7) 0 (0.0) 9 (1.2)Unclassifiable 17 (22.1) 18 (24.7) 16 (20.8) 10 (12.5) 12 (15.2) 23 (31.1) 23 (27.4) 22 (27.5) 11 (13.6) 13 (17.6) 165 (21.2)Number of assessments 77 73 77 80 79 74 84 80 81 74 779Number of images 44 43 47 47 45 45 48 47 49 47 462

PRPNo 33 (71.7) 33 (75.0) 37 (82.2) 41 (85.4) 42 (87.5) 38 (80.9) 37 (78.7) 35 (77.8) 40 (88.9) 38 (86.4) 374 (81.5)Yes 2 (4.3) 1 (2.3) 2 (4.4) 3 (6.3) 1 (2.1) 0 (0.0) 2 (4.3) 3 (6.7) 1 (2.2) 2 (4.5) 17 (3.7)Unclassifiable 11 (23.9) 10 (22.7) 6 (13.3) 4 (8.3) 5 (10.4) 9 (19.1) 8 (17.0) 7 (15.6) 4 (8.9) 4 (9.1) 68 (14.8)Number of assessments 46 44 45 48 48 47 47 45 45 44 459Number of images 22 22 22 22 22 22 22 22 22 21 219

SScNo 120 (67.0) 107 (58.8) 103 (56.6) 98 (55.7) 109 (60.9) 122 (66.3) 95 (55.9) 93 (55.7) 101 (56.1) 96 (61.1) 1,044 (59.5)Yes 15 (8.4) 30 (16.5) 37 (20.3) 47 (26.7) 32 (17.9) 13 (7.1) 26 (15.3) 26 (15.6) 42 (23.3) 33 (21.0) 301 (17.1)Unclassifiable 44 (24.6) 45 (24.7) 42 (23.1) 31 (17.6) 38 (21.2) 49 (26.6) 49 (28.8) 48 (28.7) 37 (20.6) 28 (17.8) 411 (23.4)Number of assessments 179 182 182 176 179 184 170 167 180 157 1,756Number of images 93 96 94 94 93 93 89 90 92 85 919

All ImagesNo 213 (70.5) 194 (64.9) 200 (65.8) 208 (68.4) 217 (70.9) 211 (69.2) 192 (63.8) 185 (63.4) 208 (68.0) 195 (70.9) 2023 (67.6)Yes 17 (5.6) 32 (10.7) 40 (13.2) 51 (16.8) 34 (11.1) 13 (4.3) 29 (9.6) 30 (10.3) 46 (15.0) 35 (12.7) 327 (10.9)Unclassifiable 72 (23.8) 73 (24.4) 64 (21.1) 45 (14.8) 55 (18.0) 81 (26.6) 80 (26.6) 77 (26.4) 52 (17.0) 45 (16.4) 644 (21.5)Number of assessments 302 (100.0) 299 (100.0) 304 (100.0) 304 (100.0) 306 (100.0) 305 (100.0) 301 (100.0) 292 (100.0) 306 (100.0) 275 (100.0) 2994 (100.0)Number of images 159 161 163 163 160 160 159 159 163 153 1600Values followed by a number in parentheses are “frequency (%)”; those with no parentheses are simply “number”.

Left hand Right hand

Thumb Index Middle Ring Little Thumb Index Middle Ring Little TotalImage grade

Healthy ControlNormal 24 (31.2) 20 (27.4) 34 (44.2) 29 (36.3) 34 (43.0) 22 (29.7) 29 (34.5) 33 (41.3) 44 (54.3) 33 (44.6) 302 (38.8)Early 2 (2.6) 2 (2.7) 4 (5.2) 2 (2.5) 3 (3.8) 1 (1.4) 3 (3.6) 2 (2.5) 5 (6.2) 5 (6.8) 29 (3.7)Active 0 (0.0) 0 (0.0) 0 (0.0) 1 (1.3) 0 (0.0) 0 (0.0) 1 (1.2) 2 (2.5) 1 (1.2) 1 (1.4) 6 (0.8)Late 1 (1.3) 1 (1.4) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 2 (0.3)Unclassifiable 50 (64.9) 50 (68.5) 39 (50.6) 48 (60.0) 42 (53.2) 51 (68.9) 51 (60.7) 43 (53.8) 31 (38.3) 35 (47.3) 440 (56.5)Number of assessments 77 73 77 80 79 74 84 80 81 74 779Number of images 44 43 47 47 45 45 48 47 49 47 462

PRPNormal 13 (28.3) 13 (29.5) 14 (31.1) 15 (31.3) 16 (33.3) 18 (38.3) 18 (38.3) 10 (22.2) 20 (44.4) 15 (34.1) 152 (33.1)Early 4 (8.7) 3 (6.8) 4 (8.9) 4 (8.3) 4 (8.3) 2 (4.3) 4 (8.5) 4 (8.9) 4 (8.9) 5 (11.4) 38 (8.3)Active 1 (2.2) 0 (0.0) 1 (2.2) 2 (4.2) 0 (0.0) 1 (2.1) 0 (0.0) 2 (4.4) 0 (0.0) 0 (0.0) 7 (1.5)Late 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 0 (0.0) 1 (2.1) 0 (0.0) 0 (0.0) 0 (0.0) 2 (4.5) 3 (0.7)Unclassifiable 28 (60.9) 28 (63.6) 26 (57.8) 27 (56.3) 28 (58.3) 25 (53.2) 25 (53.2) 29 (64.4) 21 (46.7) 22 (50.0) 259 (56.4)Number of assessments 46 44 45 48 48 47 47 45 45 44 459Number of images 22 22 22 22 22 22 22 22 22 21 219

SScNormal 16 (8.9) 15 (8.2) 17 (9.3) 21 (11.9) 11 (6.1) 13 (7.1) 6 (3.5) 16 (9.6) 14 (7.8) 8 (5.1) 137 (7.8)Early 9 (5.0) 14 (7.7) 14 (7.7) 10 (5.7) 11 (6.1) 8 (4.3) 10 (5.9) 8 (4.8) 15 (8.3) 14 (8.9) 113 (6.4)Active 17 (9.5) 28 (15.4) 29 (15.9) 48 (27.3) 31 (17.3) 8 (4.3) 16 (9.4) 24 (14.4) 31 (17.2) 28 (17.8) 260 (14.8)Late 15 (8.4) 18 (9.9) 29 (15.9) 23 (13.1) 36 (20.1) 22 (12.0) 23 (13.5) 20 (12.0) 25 (13.9) 34 (21.7) 245 (14.0)Unclassifiable 122 (68.2) 107 (58.8) 93 (51.1) 74 (42.0) 90 (50.3) 133 (72.3) 115 (67.6) 99 (59.3) 95 (52.8) 73 (46.5) 1,001 (57.0)Number of assessments 179 182 182 176 179 184 170 167 180 157 1,756Number of images 93 96 94 94 93 93 89 90 92 85 919

All ImagesNormal 53 (17.5) 48 (16.1) 65 (21.4) 65 (21.4) 61 (19.9) 53 (17.4) 53 (17.6) 59 (20.2) 78 (25.5) 56 (20.4) 591 (19.7)Early 15 (5.0) 19 (6.4) 22 (7.2) 16 (5.3) 18 (5.9) 11 (3.6) 17 (5.6) 14 (4.8) 24 (7.8) 24 (8.7) 180 (6.0)Active 18 (6.0) 28 (9.4) 30 (9.9) 51 (16.8) 31 (10.1) 9 (3.0) 17 (5.6) 28 (9.6) 32 (10.5) 29 (10.5) 273 (9.1)Late 16 (5.3) 19 (6.4) 29 (9.5) 23 (7.6) 36 (11.8) 23 (7.5) 23 (7.6) 20 (6.8) 25 (8.2) 36 (13.1) 250 (8.4)Unclassifiable 200 (66.2) 185 (61.9) 158 (52.0) 149 (49.0) 160 (52.3) 209 (68.5) 191 (63.5) 171 (58.6) 147 (48.0) 130 (47.3) 1700 (56.8)Number of assessments 302 (100.0) 299 (100.0) 304 (100.0) 304 (100.0) 306 (100.0) 305 (100.0) 301 (100.0) 292 (100.0) 306 (100.0) 275 (100.0) 2994 (100.0)Number of images 159 161 163 163 160 160 159 159 163 153 1600Values followed by a number in parentheses are “frequency (%)”; those with no parentheses are simply “number”.

15