Systemic sclerosis (SSc) is the result of a combination of autoimmune and vascular phenomena in association with fibrosis that triggers changes in the skin, soft tissue, joints and bones.1 It mostly affects women and has a variable clinical course. It is a rapidly progressive condition and, in some cases, has a high risk of mortality.

An early manifestation of SSc is the involvement of the hands,2 which occurs in between 46% and 97% of the cases.3 Despite the low sensitivity and specificity for the detection of early disease-related changes,4 plain radiography of the hands is useful for the follow-up of patients, with calcinosis and acro-osteolysis being the characteristic radiographic findings.5,6

The term acro-osteolysis refers to the resorption of the terminal digital tufts.7 We should point out that the distal phalangeal tuft does not have a universal radiographic definition. However, it can be considered the region of the distal phalanx in which the cortical bone changes and, rather than being completely smooth, acquires a tapered aspect8 (Fig. 1).

Fig. 1.

Radiographic image of the normal tuft. The orange-colored region delineates the area corresponding to the PTF. It indicates the typical tapering aspect of the cortical bone. The coloring appears only in the online version of the journal.

(0.04MB).

The radiographic changes in the phalangeal tuft of the fingers (PTF) in patients with SSc is not exclusive to that disorder. Acro-osteolysis has also been described in individuals with mixed connective tissue diseases (MCTD).6,9 On the other hand, in 1929, Edeiken10 reported 3 cases of SSc with sclerosis of the distal phalanx, a finding that was subsequently characterized in patients with rheumatoid arthritis (RA).11

As a outcome of the discrepancy between evidence concerning the radiological involvement of PTF in musculoskeletal diseases, the description of these structures is overlooked in the interpretation of a plain radiograph of the hands in clinical practice.

The objective of this review is to establish the abnormal findings in the PTF observed in plain radiographs in adults with SSc.

This review was carried out following the parameters of the PRISMA guidelines.12 The protocol was registered (registration no. CRD42015025185) in the PROSPERO database of systematic reviews of the University of York, which can be consulted in the following link: http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42015025185.

In the initial search, we identified 8 articles in MEDLINE, 10 in BIREME, 11 in EMBASE and 25 in Google Scholar. After exclusion of duplicates and screening, we obtained 41 articles eligible for full-text review, in addition to 6 relevant citations in the respective bibliographies. It was impossible to acquire 2 complete texts due to the publication date or to discrepancy in the literature citation provided by the database.

The qualitative synthesis included 3 case reports, 5 case–control studies, 11 cross-sectional studies and 3 cohort studies. The flow chart of the systematic review and the articles included in the analysis is shown in Fig. 2.

Fig. 2.

Flow chart corresponding to the systematic review. A summary of the process of selecting the articles included in the systematic literature review. SSc, systemic sclerosis; VHL, Virtual Health Library.

*Two complete articles were not available.

(0.36MB).

In relation to the risk of bias, only the studies of Baron et al.3 and La Montagna et al.15 were qualified for the 3 items as being “low risk”. The selection risk was more often rated as “high” because a large number of the individuals included had a history of traumatic injury to the fingers. More than half of the studies were designated as showing an “unclear” risk of detection bias since the authors did not specify whether the radiological evaluation was blinded or did not establish the interobserver variability. The results of the evaluation of risk of bias are summarized in Figs. 3 and 4. With respect to the Oxford levels of evidence, 8 articles fell into level 4, 11 into level 3 and this classification was not applied in the remainder as they were case reports.

Fig. 3.

Overall evaluation of risk of bias. Synthesis of the risk of selection, performance and detection bias of all the studies included in the systematic review. R, risk.

(0.07MB).

Fig. 4.

Independent evaluation of risk of bias. Risk of selection, performance and detection bias in each of the studies included in the systematic review. R, risk.

(0.21MB).

No study reported proliferation or erosions of the PTF as radiographic findings in patients with SSc and only the study by Bassett et al.16 described sclerosis of the PTF in 5 patients (9.09%).

All of the case reports described resorption of the PTF as the only clear finding in the radiographs of the hands of SSc patients. Taking into account the remaining articles, the prevalence of PTF resorption was 28.3% (95% CI: 0.256–0.312; p<.001) and the I2 value=80.4%, indicating a high heterogeneity (Fig. 5). In this analysis, we omitted the investigation by Allali et al.,17 as they did not discriminate the prevalence of tuft resorption in hands and feet and the study of Johnstone et al.7 as, in their results, a grade of acro-osteolysis of 0 involved individuals with and without tuft resorption (grade 0: individual score of each phalange ≤1 or total score ≤8). Fig. 5 corresponds to the forest plot of the prevalence of PTF resorption in SSc patients and the weight recorded for each study established utilizing a fixed-effects model. The data from the initial and final radiographic evaluation of the cohorts studied by Avouac et al.2 and by Blocka et al.6 were included separately.

Fig. 5.

Prevalence of PTF resorption in systemic sclerosis patients. Forest plot of the prevalence of PTF resorption in the studies included in the systematic review with the weight recorded in each and calculated overall prevalence.

CI, confidence interval; Fi, final cohort evaluation; In, initial cohort evaluation.

(0.43MB).

The studies that compared individuals with SSc versus (vs) MCTD-CREST (calcinosis, Raynaud's phenomenon, esophageal dysfunction, sclerodactyly and telangiectasia)11,16 found no statistically significant association related to PTF resorption. However, in these subgroups with calcinosis of the PTF, Yune et al.18 found for SSc vs MCTD an OR of 89.85 (95% CI: 5.36–1506.3; p=.002). Koutaissoff et al.1 reported a similar datum in their comparison of individuals with SSc vs “controls without disease”, observing tuft calcinosis with an OR of 60.31 (95% CI: 3.64–999.39; p=.004). Taking these 2 articles into account, the prevalence of tuft calcinosis was 15.6% (95% CI: 0.113–0.210; p<.001) and an I2 value=0%.

On the other hand, Avouac et al.,5 Freire et al.4 and Koutaissoff et al.1 compared individuals with SSc with “controls without disease”, finding for PTF resorption an OR=11.50 (95% CI: 0.67–196.45; p=.09), OR=18.5 (95% CI: 1.04–330; p=.05) and OR=54.34 (95% CI: 7.38–400.3; p<.01), respectively. Table 1 summarizes in detail the studies included in the synthesis, the populations evaluated and the associations reported.

The present study found that PTF resorption was the radiographic finding affecting the tuft associated with a higher rate of SSc in individuals and a prevalence of 28.3%, and a statistically significant difference when compared with controls without disease, with an OR between 11.5 and 54. Only the study of Avouac et al.5 did not observe a significant association, a fact that may be explained by the limited number of controls involved in the report (20 individuals with a history of traumatic injury to the fingers).

Many of the studies selected for the qualitative synthesis included control groups with a history of traumatic injury. This fact could lead to bias due to a confounding factor; the other review, a publication written by our group (the results of which are available in http://www.crd.york.ac.uk/PROSPERO/display_record.asp?ID=CRD42015025185), found that patients with a history of traumatic injury were more likely to develop changes in the PTF than individuals with no previous traumatic injury affecting their hands. Thus, it was decided to employ the concept of “controls without disease” and not utilize the term “healthy controls”.

In the comparison with radiographic findings in the PTF of patients with SSC and with other diseases, the evidence was limited. Three reports involved patients with MCTD-CREST6,9,18 and 1 included controls with RA.5 None of these studies found statistically significant associations in terms of tuft resorption, possibly as a result of the small number of individuals in the control groups.

The study of Yune et al.,18 which had a large control group (296 patients with MCTD), did not evaluate PTF resorption, but did report calcinosis of the PTF as a significant finding in individuals with SSc. This observation agrees with results documented by Koutaissoff et al.,1 who also graded the severity of the calcinosis in accordance with the extent and density of the calcifications. Other studies pointed out the high prevalence of soft tissue calcinosis, but did not differentiate between calcinosis of the tuft (or the pulp of the finger) and calcification of other sites of the hand, and, thus, were not included in the analysis.

It was not possible to perform a meta-analysis because of the high statistical and clinical heterogeneity (only 1 study compared SSc and RA and 4 compared SSc and controls without disease). On the other hand, cross-sectional and cohort studies did not provide data on disease duration or the sex or age of participants and, therefore, it was not possible to carry out a more in-depth analysis.

We mention 2 classifications for grading the severity of acro-osteolysis in SSc patients. Koutaissoff et al.1 classify acro-osteolysis ranging from absence of resorption to complete resorption of the tuft. They present images illustrating each grade and define as abnormal a score greater than 2 for each finger. On the other hand, Johnstone et al.7 propose a score that is specified in the comments on their study provided in Table 1. To date, the performance and reliability of these scores in unknown as the authors did not evaluate their inter- and intraobserver variability.

In accordance with the results of Johnstone et al.,7 individuals with moderate-to-severe acro-osteolysis were more likely to require debridement, amputation or the use of intravenous prostanoids than those with mild acro-osteolysis. This suggests, as was proposed by Avouac et al. in 2006,5 that the grade of acro-osteolysis is proportional to the severity of digital ischemia. Likewise, the 2011 prospective study of Avouac et al.2 found that the patients who developed acro-osteolysis and those in whom it got worse were more likely to ultimately meet severity criteria such as flexion contracture, severe vascular damage and interstitial lung involvement. Thus, this change in the PTF is a marker of the prognosis and the severity of SSc.

Chung et al.19 recently presented a radiographic scoring system to evaluate calcinosis in SSc patients that has an excellent inter- and intraobserver reliability (intraclass correlation coefficient 0.89 and 0.93, respectively). For this calculation, the authors divided the hand into 22 anatomical regions (10 corresponding to the PTF) and estimated the percentage of area covered with calcium and the severity of the calcinosis according to a simple scoring system previously proposed by Koutaissoff et al.1

An important limitation to this study lies in the differences between the diagnostic criteria employed by the authors of the primary studies, which include research from 1969 to 2013. This could explain the high clinical and statistical heterogeneity, such as that found in the prevalence of PTF resorption in this review.

Finally, current evidence suggests that acro-osteolysis is another of the ischemic manifestations of SSc and, thus, its radiographic characterization is useful in establishing the prognosis and severity of the disease. In this respect, the implementation of radiological scores standardizes criteria for the evaluation of these aspects, although further research is necessary to determine their reliability and validity.

In conclusion, resorption and calcinosis of the PTF are the characteristic radiographic findings in patients with SSc. However, studies with a greater methodological soundness are required to enable us to firmly establish the associations between these phenomena and the presence of other connective tissue diseases.

This project was performed without outside funding.

The authors declare they have no conflicts of interest.