RHEUMATOLOGY
Rheumatology 2012;51:2234–2238
doi:10.1093/rheumatology/kes214
Advance Access publication 25 August 2012
Concise report
Acro-osteolysis in systemic sclerosis is associated
with digital ischaemia and severe calcinosis
Emma M. Johnstone1, Charles E. Hutchinson2, Andy Vail1, Aurelie Chevance1
and Ariane L. Herrick1
Abstract
Objectives. Acro-osteolysis (bony resorption of the terminal digital tufts) is a well-recognized, but
under-researched, manifestation of SSc. Our aim was to investigate the hypothesis that acro-osteolysis
is associated with (i) the severity of digital ischaemia and (ii) the presence of calcinosis.
Methods. This was a retrospective study of 101 patients with SSc in whom hand radiographs taken
between 2001 and May 2008 were available for review. These radiographs were graded for severity of
acro-osteolysis on a 0–4-point scale for each finger (0 = normal bone structure, 4 = severe pencilling of the
terminal phalanges). From these scores, patients were subdivided into the following two groups: normal/
minimal acro-osteolysis and moderate/severe acro-osteolysis. The presence or absence of calcinosis
(mild, moderate or severe) was also documented.
CLINICAL
SCIENCE
Results. Of the 101 patients, 68 were grouped as normal/minimal acro-osteolysis and 33 as moderate/
severe acro-osteolysis. Forty-five had severe digital ischaemia: 25 (76%) of the patients with moderate/
severe acro-osteolysis compared with 20 (29%) of those with normal/minimal acro-osteolysis (multifactorial analysis: P < 0.001). Patients with moderate/severe acro-osteolysis were more likely to have severe
calcinosis (33% vs 13%), but this was not statistically significant after adjustment for potential
confounders.
Conclusion. Acro-osteolysis was strongly associated with severe digital ischaemia. The potential association with severe calcinosis merits further study. Prospective studies are required to investigate
acro-osteolysis as a marker of digital vascular disease progression and of treatment response.
Key words: SSc, acro-osteolysis, calcinosis, digital ischaemia, Raynaud’s phenomenon, hand radiographs.
Introduction
Digital problems are a major contributor to the morbidity,
disability and pain of SSc. They reflect the vascular
abnormalities and fibrosis that characterize the disease
process and result in digital ischaemia (which can progress to digital-tip ulceration and gangrene), and in sclerodactyly, often with contracture and ulceration over the
extensor aspects of the interphalangeal joints. For reasons that are not fully understood, a significant proportion
of patients also develop subcutaneous calcinosis, which
1
University of Manchester, Manchester Academic Health Science
Centre, Salford Royal Hospital, Salford and 2Department of Health
Sciences, University of Warwick, Coventry, UK.
Submitted 1 April 2012; revised version accepted 2 July 2012.
Correspondence to: Ariane L. Herrick, Clinical Sciences Building,
Salford Royal Hospital, Salford M6 8HD, UK.
E-mail: [email protected]
can ulcerate through the skin; when it is severe, calcinosis
in itself causes pain and loss of function.
Acro-osteolysis (resorption of the terminal tuft of the
digit) is a characteristic feature of SSc and has been estimated to occur in around 20–25% of patients [1, 2]. It is
often assumed that acro-osteolysis results from impairment of blood supply, although pressure from skin tightening may also play a role [3]. Acro-osteolysis has been
relatively little researched, although two recent studies of
hand radiographs have documented acro-osteolysis and
its associations (including with calcinosis) in patients with
SSc [1, 2], and an association with secondary hyperparathyroidism has recently been suggested [4]. Similarly, little
is known about the pathogenesis of SSc-related calcinosis. Increased understanding of the inter-relationships
between digital ischaemia, loss of digital tissue (including
bone resulting in acro-osteolysis) and calcinosis may provide further insights into the pathogenesis of the SSc
! The Author 2012. Published by Oxford University Press on behalf of the British Society for Rheumatology. All rights reserved. For Permissions, please email: [email protected]
Radiographic acro-osteolysis and calcinosis in SSc
disease process. Our specific aim was to investigate the
hypothesis that acro-osteolysis is associated with the severity of digital ischaemia (in which case it may serve as a
marker of disease severity) and the presence of calcinosis.
Patients and methods
minimal, moderate or severe acro-osteolysis)
ascribed to each patient on the following basis:
was
0 Normal/minimal acro-osteolysis: maximum score for an
individual finger = 1 and total score (sum of all fingers) 48.
FIG. 1 Radiographic scoring systems.
This was a retrospective study of hand radiographs from
patients with SSc attending Salford Royal Hospital.
Radiographs were selected on the basis that (i) the patient
had signed informed consent to his/her data being used
for research purposes (study approved by Salford and
Trafford Research Ethics Committee), (ii) the radiograph
was taken between 2001 and May 2008 and (iii) the radiograph was available for review.
A total of 164 hand radiographs were recorded from
2001 to 2008. Of these, 25 were available to view online
using the Picture Archive and Communication System,
and the remaining 139 films had been archived. Of these
139, 85 were available for review, giving a total of 110
radiographs. Nine radiographs were repeat films (films
from the same patients), meaning that radiographs from
101 patients were analysed. In the case of repeat films,
the most recent radiograph was included for analysis in
the study.
Patient details were extracted from an electronic database: gender, age and duration of RP at the time of radiograph, disease subtype (limited or diffuse cutaneous [5]),
ACA status, smoking status and presence or absence of
severe digital ischaemia [defined by a history of one or
more of the following: admission for intravenous (i.v.)
prostanoid therapy, surgical debridement or amputation].
Analysis of radiographs
Radiographs were examined for the following two
features: acro-osteolysis and calcinosis.
Acro-osteolysis
For the purpose of the study, a grading scale for acroosteolysis was developed. This scale, from 0 to 4, took
into account the presence and severity of acro-osteolysis
in each finger (including thumbs), with 0 being normal
bone structure and 4 being severe pencilling of the terminal phalanx (Fig. 1A). The scoring system was as
follows:
0 Normal terminal phalanges. No resorption.
1 Minimal acro-osteolysis, small amount of resorption at
the terminal tuft.
2 Resorption of most of the distal tip of the terminal tuft.
3 Resorption of most of the terminal tuft, leaving only
one side intact.
4 Complete resorption of the terminal tuft, with obvious
pencilling.
Each radiograph was viewed by two observers (one a
consultant musculoskeletal radiologist), and a consensus
was reached for the score for each finger. Because the
presence and severity of acro-osteolysis varied between
fingers within any one patient, an overall rating (normal/
www.rheumatology.oxfordjournals.org
(A) Acro-osteolysis scoring system for individual
fingertips and (B) examples of calcinosis scores (i) moderate and (ii) severe.
2235
Emma M. Johnstone et al.
1 Moderate acro-osteolysis: maximum score for an
individual finger 42 and total score from 9 to 16
inclusive.
2 Severe acro-osteolysis: maximum score for an individual finger 53 or total score 517.
Results
Of the 101 patients, 68 (67%) had an acro-osteolysis
rating of 0 (normal/minimal), 7 (7%) had a rating of 1 (moderate) and 26 (26%) had a rating of 2 (severe). Because
of the small number of patients with moderate acroosteolysis, those with moderate and severe acroosteolysis were considered together for the purposes of
statistical analysis.
Demographic and clinical features of patients are
shown in Table 1 alongside results of the multifactorial
analysis. Distributions of age, sex and disease subtype
were similar between those with and without moderate/
severe acro-osteolysis. Although patients with moderate/severe acro-osteolysis were more likely to smoke
than those without, this difference did not reach statistical significance. Patients with longer duration of RP
were more likely to have moderate/severe acroosteolysis. ACA-positive status was associated with
increased age and duration of RP, with more severe
calcinosis and severe digital ischaemia. The apparent
protective effect of ACA-positive status in the multifactorial model, despite the absence of a unifactorial
relationship, seems to be an artefact of these
associations.
Calcinosis
Presence or absence of radiographic calcinosis was
determined by a musculoskeletal radiologist. Calcinosis
was further subdivided into mild, moderate and severe
depending on the density of calcinosis and the number
of separate sites of calcinosis. A single site of low-density
calcinosis was grouped as mild, medium density calcinosis at one or more sites or a single site of high-density
calcinosis was grouped as moderate and more than one
site of high- or mixed-density calcinosis was grouped as
severe. Examples are shown in Fig. 1B.
Statistical analysis
Factors were grouped and described using standard summary statistics. Association with acro-osteolysis status
was assessed using a single multifactorial regression analysis. All factors for inclusion (listed in Table 1) were determined in advance and were retained in the model
regardless of statistical significance to prevent complications of multiplicity or conditionality in the selection of factors. Duration of RP was log-transformed to normalize the
distribution, and calcinosis was regrouped, combining the
none and mild categories, as no cases of moderate or
severe acro-osteolysis were observed in the few patients
classified as having mild calcinosis.
Acro-osteolysis and severity of digital ischaemia
Results are shown in Table 1. Forty-five patients had
severe digital ischaemia, of whom 39 had had i.v. prostanoids, 19 debridements and 7 amputations. Patients with
moderate/severe acro-osteolysis were more likely to have
TABLE 1 Demographic and clinical factors potentially associated with moderate and severe acro-osteolysis
Factor
Age, years
Sex
Subtype
Duration, years
ACA statusa
Smokedb
Calcinosis
Severe DI
Level
Median (range)
Male subjects
Female subjects
Limited
Diffuse
Median (range)
Negative
Positive
No
Yes
None
Mild
Moderate
Severe
No
Yes
i.v. prostanoids
Debridement
Amputation
No and
minimal
acro-osteolysis
(n = 68)
55
14
54
48
20
9
40
28
55
11
41
8
10
9
48
20
18
6
1
(29–82)
(21)
(79)
(71)
(29)
(1–61)
(59)
(41)
(83)
(17)
(60)
(12)
(15)
(13)
(71)
(29)
(26)
(9)
(1)
Moderate and
severe
acro-osteolysis
(n = 33)
54
6
27
22
11
17
20
12
20
10
19
0
3
11
8
25
21
13
6
(27–84)
(18)
(82)
(67)
(33)
(1–55)
(63)
(38)
(67)
(33)
(58)
(0)
(9)
(33)
(24)
(76)
(64)
(39)
(18)
Comparison
OR (95% CI)
P-value
per decade
1.3 (0.73, 2.2)
0.40
vs male subjects
0.94 (0.19, 4.8)
0.94
vs limited
per doubling
1.0 (0.24, 4.3)
2.2 (1.1, 4.4)
0.99
0.03
vs negative
0.17 (0.04, 0.82)
0.03
vs not
2.7 (0.70, 11)
0.15
vs none/mild
vs none/mild
1.2 (0.20, 7.1)
2.5 (0.57, 11)
0.84
0.23
vs not
10 (2.8, 38)
<0.001
All data are given as n (%) unless specified otherwise. DI: digital ischaemia; OR: odds ratio. aData from 100 patients. bData
from 96 patients.
2236
www.rheumatology.oxfordjournals.org
Radiographic acro-osteolysis and calcinosis in SSc
severe digital ischaemia than those without (P < 0.001).
This was also true for each of the components
defining severe digital ischaemia, despite the reduced
numbers of observations for each: i.v. prostanoids
(P = 0.008), surgical debridements (P = 0.002) and amputations (P = 0.04).
Acro-osteolysis and calcinosis
Forty-one patients (41%) had radiographic calcinosis
(Table 1). Although patients with acro-osteolysis were
more likely to have severe calcinosis, this association
was not statistically significant after adjustment for other
factors.
Discussion
We have demonstrated that acro-osteolysis is associated
with digital ischaemia, confirming the findings of Avouac
et al. [1] and Koutaissoff et al. [2]. This association is
strong (P < 0.001) even after adjustment for the potential
confounders listed in Table 1. The key new findings from
our study, discussed below, are that the association with
digital ischaemia is strengthened in an SSc population
with a high prevalence of digital problems (including
examining for associations with different degrees of severity of digital ischaemia), and that any association with
calcinosis may exist only when calcinosis is severe. The
major focus of our study was on acro-osteolysis rather
than a more comprehensive description of radiographic
abnormalities in patients with SSc. We developed an
acro-osteolysis scoring system to minimize subjectivity.
Although different patterns of acro-osteolysis have
been described, in this study we concentrated on distal
bone loss.
Although an association between severe digital ischaemia and acro-osteolysis does not indicate causation,
nonetheless a possible explanation for the association
between acro-osteolysis and severity of digital ischaemia
is that impaired blood flow may drive the terminal tuft
resorption, which may therefore be a result of ischaemic
atrophy. This seems intuitive, given that patients with SSc
and severe digital ischaemia develop fingertip ulcers and
digital pitting; it seems likely that this soft tissue loss is
paralleled by bone loss.
Avouac et al. [1] reported acro-osteolysis in 26 (22%) of
an unselected cohort of 120 patients with SSc. The definition of acro-osteolysis (e.g. whether its presence in one
finger alone was sufficient for the diagnosis) was not
given, but the findings were similar to ours in that patients
requiring prostacyclin were more likely to have
acro-osteolysis than those without it (10/20 vs 16/100,
P = 0.005) [1]. In our study, a higher proportion of patients
had required i.v. prostanoids than in the cohort of Avouac
et al. [1] (39% vs 17%). This higher proportion of patients
requiring prostanoids in our study (and the higher prevalence of acro-osteolysis in our study; 33% vs 22% in the
cohort of Avouac et al. [1]) most likely reflects its retrospective nature; those patients who had hand radiographs
would have been likely to have digital problems. Avouac
et al. [1] also reported that acro-osteolysis was associated
www.rheumatology.oxfordjournals.org
with digital ulceration, a finding confirmed by Koutaissoff
et al. [2], who scored acro-osteolysis in 167 patients with
SSc (25% had acro-osteolysis). Koutaisssoff et al. [2] also
reported (as we too found) that patients with longer disease duration are more likely to have acro-osteolysis, consistent with the clinical observation that clinical features of
digital vascular injury in SSc tend to progress over the
years.
These three studies (Avouac et al. [1], Koutaissoff et al.
[2] and our own) all point to acro-osteolysis being associated with clinical features of digital ischaemia. Although
numbers are small, it is of interest that in our cohort, moderate/severe acro-osteolysis was present in 86% of those
with amputations and in 68% of those requiring debridement; in other words, those in whom digital ischaemia is
most severe. This further supports the hypothesis that
acro-osteolysis is associated with severity of digital
ischaemia.
Although it seems likely that digital ischaemia plays a
key role in the pathogenesis of acro-osteolysis (emphasizing the need for effective vasodilator therapy), prospective studies are required to confirm this. Relevant to
this is that a recent longitudinal study by Avouac et al. [6]
reported that digital ulcers (and calcinosis) predicted progression of acro-osteolysis.
Our study also lends limited support for the association
between acro-osteolysis and calcinosis previously
reported by Avouac et al. [1, 6]. Our cohort had a higher
prevalence of calcinosis than that of Avouac et al. [1] (41%
vs 23%), again probably reflecting a bias in our population
toward patients with clinically significant digital problems.
In our study we graded the severity of calcinosis and
found an association between acro-osteolysis and
severe calcinosis, but not with more minor degrees. The
statistical significance of this association lessened with
multifactorial analysis. Koutaissoff et al. [2] reported that
terminal tuft calcinosis (present in 15% of their cohort)
was associated with digital ulceration and pitting scars.
Avouac et al.’s [6] study showed that digital ulceration
was an independent predictor of calcinosis and
acro-osteolysis. An association between calcinosis and
nailfold capillaroscopic changes has been reported [7],
although we did not confirm this [8]. Further studies are
indicated to elucidate the inter-relationships between digital blood flow (using modern imaging techniques), acroosteolysis and calcinosis.
In conclusion, there is now good evidence to suggest
that acro-osteolysis is another ischaemic manifestation of
SSc. Acro-osteolysis may also be associated with calcinosis. Although these findings provide increased insight
into pathogenesis, perhaps a more important conclusion
is that scoring systems for acro-osteolysis (building on
that used in this study) should be developed. The scoring
system used in this study was ad hoc. Although this
has good face validity and was rated independently
by two observers who reached consensus on each
digit, it remains a weakness of our study that we do not
have formal inter- or intra-rater reliability assessments
for this scale. A validated acro-osteolysis scoring
2237
Emma M. Johnstone et al.
system could become a marker in studies of digital vascular disease progression and of treatment response.
Rheumatology key messages
Acro-osteolysis
ischaemia.
. Acro-osteolysis
flow.
. Acro-osteolysis
sible marker of
.
is associated with severity of digital
may be driven by impaired blood
should be investigated as a posSSc-related digital ischaemia.
functional correlations. Semin Arthritis Rheum 2011;40:
455–60.
3 Benitha R, Modi M, Tikly M. Osteolysis of the cervial
spine and mandible in systemic sclerosis: a case
report with computed tomography and magnetic
resonance imaging findings. Rheumatology 2002;41:
1198–200.
4 Braun-Moscovici Y, Furst DE, Markovits D et al. Vitamin D,
parathyroid hormone, and acroosteolysis in systemic
sclerosis. J Rheumatol 2008;35:2201–5.
5 LeRoy EC, Black C, Fleischmajer R et al. Scleroderma
(systemic sclerosis): classification, subsets and pathogenesis. J Rheumatol 1988;15:202–5.
Disclosure statement: The authors have declared no
conflicts of interest.
6 Avouac J, Mogavero G, Guerina H et al. Predictive factors
of hand radiographic lesions in systemic sclerosis: a preospective study. Ann Rheum Dis 2011;70:630–3.
References
7 Vayssairat M, Hidouche D, Abdoucheli-Baudot N,
Gaitz JP. Clinical significance of subcutaneous calcinosis
in patients with systemic sclerosis. Does diltiazem induce
its regression? Ann Rheum Dis 1998;57:252–4.
1 Avouac J, Guerini H, Wipff J et al. Radiological hand involvement in systemic sclerosis. Ann Rheum Dis 2006;65:
1088–92.
2 Koutaissoff S, Vanthuyne M, Smith V et al. Hand
radiological damage in systemic sclerosis:
comparison with a control group and clinical and
2238
8 Herrick AL, Moore TL, Murray AK et al. Nail-fold capillary
abnormalities are associated with anti-centromere antibody and severity of digital ischaemia. Rheumatology
2010;49:1776–82.
www.rheumatology.oxfordjournals.org