Article Text
PDF
Abstract
Background The treatment of low grade Spetzler-Martin (SM) brain arteriovenous malformations (AVMs) has been debated in unruptured cases. Nevertheless, in clinical practice there are cases where treatment is preferred; in these cases a very low complication rate is mandatory. In ruptured cases, early and complete obliteration of the nidus is the preferred strategy.
Objective To achieve curative embolization, ideally in a single session, by dual microcatheterization techniques with arterial and/or venous access, according to the angioarchitecture.
Materials and methods This is a prospective, single-center study carried out between January 2008 and January 2016. Patients with ruptured and unruptured brain AVMs, with SM grades I and II, treated by endovascular means, were included. Demographics, clinical presentation, angioarchitecture, and procedure-related complications were analyzed. Angiographic and clinical results were reported.
Results Seventy-three patients, aged 40.5±17.8 years, were included. More than 60% of the patients presented with ruptured AVMs. Initial SM grades were I for 22% and II for 78% of the patients. Preprocedural modified Rankin Scale (mRS) score was 0–2 for 53 (72.6%), 3 for 12 (16.4%), 4 for 5 (6.8%) and 5 for 3 (4.1%) patients. Procedure-related morbidity was 2.7% and procedure-related mortality was 0%. Ninety percent (90.5%) of the patients were independent in their everyday lives (mRS score 0–2) at 6 months. In all but one case (95%) the embolization was curative.
Conclusion Stand-alone endovascular treatment for SM grade I and II brain AVMs seems safe and effective, allowing for complete obliteration of the nidus, with low complication rates. A study of larger cohorts is needed.
- arteriovenous malformation
- catheter
- embolic
- liquid embolic material
Statistics from Altmetric.com
Introduction
The management of Spetzler-Martin (SM) grade I and II brain arteriovenous malformations (bAVMs) depends on the presence or absence of rupture at clinical presentation. In the ARUBA study on unruptured bAVMs, the risk of treatment was higher than the risk of the natural history of the disease, especially for low-grade cases.1 This conclusion has since been challenged2; at present there is insufficient pragmatic evidence to provide clear guidelines.3 Moreover, when treatment has been started, the natural history of the bAVM is modified and an increased risk of hemorrhage is present, as compared with intact, unruptured bAVMs.
A clear short and mid-term benefit4–6 is evident for patients with ruptured lesions, in order to prevent re-rupture, which appears to be the most important factor leading to disability and mortality6 for these patients. A fast eradication of the nidus should be the goal of treatment and long latency periods for cure seem to be less desirable,7 because the patient is not protected against re-rupture8 during this time.
The role of radiosurgery is clearly a valid option for treatment of unruptured cases,9 10 especially in low SM bAVMS. Both ruptured and unruptured cases seem amenable to microsurgery, with rapid eradication and seemingly acceptable complication rates.2 11 The role of endovascular treatment in low-grade bAVMs remains in many cases unclear in the published literature. Targeted embolization in ruptured cases is a well-established practice,12–14 but there is a diversity of embolic agents, practices, goals, techniques, and evidently, outcomes published on endovascular eradication of bAVMs12–14 of all grades.
Nevertheless, the evolution of endovascular approaches has been important in the past 10 years, with new non-adhesive embolic agents allowing for better control of the procedure and higher percentages of nidus eradication.15 16 Even though data are limited to case series, there is increasing evidence of rapid and effective eradication of AVM nidi, with acceptable complication rates. The safety and effectiveness of exclusive curative embolization based on prospective data needs further assessment in well-designed randomized trials and registries.3 17 We feel that patients and practitioners will find useful the results of this prospective cohort of carefully selected cases, treated by endovascular means with rapid anatomic cure as the goal of the therapeutic approach.
Methods
This is a prospective, single-center, observational cohort study extending over 8 years, from January 1, 2008 to January 1, 2016. All procedures were performed by the same team of experienced consultant interventional neuroradiologists and supporting personnel. Imaging and clinical evaluation, embolization technique, and follow-up were performed according to pre-established protocols and details were entered in a prospectively maintained database.
The scientific hypothesis of the study was that stand-alone endovascular treatment could be a safe and effective treatment for SM grade I and II brain AVMs. Primary end points were safety, expressed as clinical outcome according to the modified Rankin Scale (mRS) score of 0–2 at 6 months, and effectiveness, expressed as total exclusion of the nidus immediately after the procedure and at 6 months. Secondary end points included a comparison of two groups—ruptured versus unruptured bAVMs—for safety, effectiveness, and clinical outcomes of the proposed therapeutic strategy, in order to gain further insight for the design of larger prospective studies. The study received approval from the institution’s ethics committee.
Patient selection
The treatment strategy followed our pre-established protocols for AVM treatment. The study presented was observational, thus did not influence the therapeutic decision-making. In our center a multidisciplinary discussion (team with neuroradiologists, interventional neuroradiologists, neurosurgeons, and anesthetists) is employed in the therapeutic decision-making; for the past 10 years the center has had an interventional prioritization for patients with neurovascular pathologies.
Inclusion criteria were patients with both ruptured and unruptured brain AVMs that were classified as SM grade I or II and that were amenable to exclusively endovascular treatment. Ruptured AVMs were embolized in the first 3 weeks after hemorrhage. Since the ARUBA study, unruptured, low SM AVMs have been amenable to conservative management, apart from selected cases that have been considered for invasive treatment, as follows.
Non-ruptured AVMs for which endovascular embolization, radiosurgery, or surgery had already been attempted, without nidus eradication, were considered as having a higher risk of hemorrhage and were treated. Non-ruptured cases were treated when a patient actively demanded intervention and only when nidus eradication was judged feasible and safe in a single session and only after informed consent had been received from the patient,.
Exclusion criteria, apart from higher SM classification rates, included unruptured cases for which nidus eradication was not considered feasible and safe by exclusively endovascular means, other concomitant pathologies were present such as aneurysms not related to the AVM (not in the axis of the AVM or the axis of its feeders), carotid stenosis >40%, imaging evidence of other vascular malformations (eg, cavernous malformations, facial or body AVM), or microsurgical resection of AVM. All AVM cases were diagnosed by MRI or CT angiography, and further confirmed and analyzed by selective digital subtraction angiography (DSA).
Clinical evaluation
Every patient underwent full clinical and neurological evaluation by a senior interventional neuroradiologist and a senior anesthesiologist before treatment, at awakening, at discharge, and at the 6-month follow-up. Glasgow coma scale scores were assessed before and after treatment. mRS scores were assessed for the same periods. Results were cross-checked by an independent senior neurosurgeon. A mRS score of 0–2 at 6 months was considered to be a good clinical outcome and a mRS score ≥3 at 6 months was considered to be a poor outcome.
Preoperative and postoperative imaging evaluation
Angiographic results: Pre-embolization, immediate post-embolization, and follow-up angiographic results at 6 and 12–18 months after embolization were classified as total exclusion-eradication of the nidus, small residual nidus <3 mm in diameter, or residual nidus in a percentage of the total volume of AVM. Four-axis selective DSA scans were performed at 6 months after intervention and, if complete occlusion was confirmed, MRI was programmed at 24 months' follow-up.
MRI imaging: All patients systematically underwent a brain MRI imaging at 24–28 hours before and after treatment. This was performed in a 3T system (Achieva, Philips Medical System, Best, The Netherlands), with an imaging protocol of a T2-weighted spin-echo imaging with a fluid-attenuated inversion recovery (FLAIR), a T2* sequence, a 3DT1 sequence, and diffusion-weighted imaging (diffusion gradients in x, y, and z directions with two b values (0 and 1000 s/mm2)), in order to evaluate the potential presence of arterial or venous ischemia.
Endovascular embolization protocol
Informed consent was obtained from the patient in all unruptured cases; decision-making took into account the presence or absence of special angiographic characteristics, patients’ cardiovascular risk factors, and his/her personal wishes. Procedures were performed in a biplane flat panel angiographic suite (Allura Xper FD20; Philips Healthcare, The Netherlands). Treatment was conducted under general anesthesia with continuous pressure monitoring. Intravenous heparin (3000 IU) was administered before hyperselective intracranial vessel navigation. Double arterial, venous and double arterial-and-venous techniques in our institution have been described elsewhere18 and were chosen according to the angioarchitecture and location of the AVM.
Statistical analysis
Sample size estimation was performed using one-proportion, z, Χ2 test, setting the estimated population proportion with good clinical outcome at 90% and deriving from published data a 75% of good clinical outcome for all treatments used (neurosurgery, endovascular methods, and radiosurgery). Allowing for an α error of 0.05 and a power of 90%, the calculated sample size for this study was 68. To this number an additional 5% of patients not evaluable for whatever reason were added. Given previous recruitment of our department, 8 years was the estimated time for collecting the required number of patients. Descriptive statistical analysis was performed, using the Student t-test for quantitative data, after appropriate testing for normal distribution (d’Agostino-Pearson test) and a Χ2 test for qualitative data. Two groups were identified: with or without previous rupture of the bAVM. Group comparisons were made after verification of the comparability of the two groups by the main characteristics—namely, age, sex, size of nidus, and SM grade.
The Statistica software (StatSoft, GE) was used for the statistical analysis. The level of statistical significance was determined as p≤0.05.
Results
Of 187 patients with brain AVMs treated between January 2008 and January 2017, 73 cases with AVMs SM grade I and II were included in the study.
Patients’ baseline characteristics
Seventy–three patients (45 men, 28 women), aged 40.5±17.8 (mean ±SD), were included in the study. Clinical presentation and symptoms indicating rupture were seen in 60.3% (44/73), epileptic seizures in 8.2% (6/73), headaches in 21.9% (16/73), neurologic deficit in 2.7% (2/73) of cases; in 6.8% (5/73) discovery was an incidental finding. Baseline characteristics (age, sex, SM grade and size of the nidus) were comparable among the groups of ruptured versus unruptured bAVMs.
Location and angioarchitecture; AVM characteristics
Cases were classified initially as SM I in 21.9% (16/73), II in 76.7% (56/73), and SN III in one case (1.4%). This last case was a patient previously treated elsewhere and subsequently referred to our department; SM III refers to the initial situation; the patient was classified as SM II in our department. The mean±SD size of the nidus was 2.7±1.2 cm. In 35.6% (26/73) of the cases the AVM was frontal, 13.7% (10/73) were temporal, 2.7% (2/73) were pre-Rolandic, 1.4% (1/73) were Rolandic, 21.9% (16/73) were parietal, 15.1% (11/73) were occipital, and 9.6% (7/73) cerebellar.
In 19.2% (14/73) of cases an aneurysm was revealed on DSA; a Willis aneurysm was present in 2.7% (2/73), an afferent artery aneurysm in 11.0% (8/73), and an intranidal aneurysm in 5.5% (4/73) of the cases. Venous stenosis was found in 11.0% (8/73) and venous ectasia in 17.8% (13/73) of the cases. Venous drainage was exclusively cortical in 84.9% (62/73) of the AVMs and in 15.1% (11/73) the deep venous system was also involved.
Procedure/technique specifications
In 85% (62/73) of cases a single embolization session was enough for total exclusion of the nidus, while in 15% (11/73) of the remaining cases two embolization sessions were required. Among the latter, the first embolization session was dedicated to associated aneurysm occlusion in three cases (3/11, 27%) and the second to the nidus exclusion. Mean nidus embolization time, calculated as the interval between dimethyl sulfoxide injection and retrieval of the microcatheter was 29.5±17.5 min.
Adjunctive treatment before endovascular embolization was not deemed necessary in 85% (62/73) of the patients (figure 1). Hematoma evacuation before embolization was performed in 2.7% (2/73) cases, radiosurgery in 1.4% (1/73); ventriculostomy in 2.7% (2/73), and aneurysm coiling in 8.2% (6/73). An embolization session in another center, which referred the patients afterwards, had taken place in 15.1% (11/73) of the cases.
Patient presenting with convulsions, arriving at a regional hospital with post-critical syndrome; a ruptured left frontal brain arteriovenous malformation (bAVM) was discovered in the diagnostic investigation; the patient’s condition deteriorated clinically to Glasgow Coma Scale score 6. The patient was intubated, ventilated, and transported rapidly to the center of reference for further care. Absence of focal neurologic deficit; on the DSA scan, lateral (A) view, left frontal AVM of ~3 mm diameter, arterial feeders from the frontal anteromedial branch (arrow) and the orbitofrontal branch (thin arrow) of the left anterior cerebral artery (ACA), as well as the frontal trunk of the middle cerebral artery (MCA) (dotted arrow), exclusively cortical venous drainage, with multiple venous pouches ((B) three-dimensional rotational angiography from the left internal carotid artery (ICA), the asterisk showing an ectasia retained as the hemorrhage site); endovascular treatment was performed (double femoral access), with initially, single pedicle arterial microcatheterization from the orbitofrontal branch (with 3 mm detachable microcatheter, Apollo, Medtronic, California, USA), followed by double simultaneous arterial injection of Onyx through two 1.5 mm detachable microcatheters from the ACA and the frontal trunk of the MCA; (C), single shot showing the cast of Onyx from the first microcatheterization, as well as the dual microcatheterization by the ACA (arrow) and the MCA (dotted arrow); (D), single shot showing the cast of Onyx at the control and the end of the procedure; control at 6 months showing DSA from the left ICA, lateral (E) view, showing complete exclusion of the AVM; absence of thrombotic or hemorrhagic procedure-related complication.
The technical approach included single arterial microcatheterization and Onyx injection in 23.3% (17/73), single venous approach in 21.9% (16/73), multiple arterial in 41.1% (30/73), and arterial and venous microcatheter injection in 13.7% (10/73) of the cases (figure 2). The mean±SD volume of Onyx-18 was 1.9±1.4 mL. Complementary treatment after the embolization session included hematoma evacuation in 5.5% (4/73) and ventriculostomy in 8.2% (6/73) of the cases; in 86.3% (63/73) of the cases no other intervention was deemed necessary. Radiosurgery was performed after the embolization in one case.
Patient with non-pulsatile tinnitus, with treatment difficult. Digital subtraction angiography: (A) frontal and (B) lateral view, showing a left fronto-opercular arteriovenous malformation (AVM), with a nidus of maximal diameter 2.5 mm, with exclusively cortical venous drainage towards the superior sagittal sinus (SSS) and the left Labbe vein to the left lateral sinus (LS); (C) three-dimensional rotational angiography showing a single collector divided into two draining veins, both of them dilated at origin; dual femoral arterial and jugular vein access was used; (D) single shot during the procedure showing arterial microcatheterization with a 3 mm Apollo (Medtronic, California, USA) microcatheter (dotted arrow) by the left fronto-opercular artery to the nidus and dual venous microcatheterization with a microcatheter (Marathon, Medtronic, California, USA) (white arrow) into the single venous collector and a second one for coiling (Echelon-10, Medtronic, California, USA) (black arrow), left in the junction of the two venous collectors. An inverse ‘pressure cooker technique’ was performed, with coiling of the collectors initially, followed by controlled injection of Onyx-18 retrograde in the nidus, by the Marathon microcatheter and finishing with Onyx injection through the arterial detachable microcatheter. The Marathon microcatheter was cut at the level of the jugular vein, in order to avoid traction to the venous system, as has already been described19; (E) the cast of Onyx at the end of the procedure, Xper-CT, 3D reconstructions; (F) and (G), DSA on frontal and lateral views, respectively, showing complete obliteration of the nidus at 6 months’ control; absence of periprocedural or postprocedural complications during hospitalization and follow-up.
Per-procedural and technical complications did not differ significantly among the different therapeutic strategies (single arterial, multiple arterial, single venous, arterial and venous; Χ2=10.333, df 6, significance level p=0.1113), nor did they differ according to the presence or absence of specific angiographic characteristics of the AVMs (Χ2=2.882, df 10, significance level p=0.9841), and not even according to the venous drainage type (Χ2=0.555, df 2, significance level p=0.7576).
Treatment strategy varied significantly according to the nidus location and the Spetzler-Martin classification (p=0.010 and p=0.020—tables 1 and 2, respectively). The mean time for the Onyx injection in single microcatheter techniques was 20±21 min and in dual microcatheter techniques 36±9 min; a Welch test revealed statistically significant difference among them (test statistic t(d): 3.316, df: 28.0, two-tailed probability p=0.003).
Endovascular treatment strategy according to nidus location
Endovascular treatment strategy according to the Spetzler-Martin classification
Procedure-related complications
During the procedures two AVM ruptures (2.3%) occurred. In the hospitalization period three cases (3.5%) of HLH occurred, one of which resolved with steroid treatment during hospitalization. Procedure-related mortality for the series was 0%. Permanent, procedure-related morbidity in the series was 2.7%. The group of unruptured AVMs had no permanent, procedure-related morbidity and no deaths.
Clinical and angiographic outcomes
Clinical outcome during hospitalization included a stable or improved clinical situation in 93.2% (68/73) of the cases and deterioration in 6.8% (5/73). The mRS scores for the different end points are summarized in table 3.
Modified Rankin scale (mRS) score modifications
Clinical outcome, as measured by the mRS score at discharge and at 6 months after the intervention, correlated significantly with the initial clinical presentation (respectively, Χ2=42.084, df 16, significance level p=0.0004 and Χ2=60.399, df 20, significance level p<0.0001, respectively). When the sample was divided into two groups, according to the presence or absence of rupture of the AVM at presentation, perprocedural and postprocedural complications were significantly higher in the ruptured group (p=0.040 and p=0.050, respectively, with Χ2 for trend p<0.010). The mRS score at discharge was also significantly higher in the ruptured group (p<0.0001). The size of the nidus and the number of embolization sessions did not differ between the two groups.
Immediate angiographic control at the end of the procedure showed absence of nidus opacification in all but one cases (1.4%); for the latter case a remnant of <10% with late filling was present. In another two cases a remnant was present, but this was embolized with total occlusion of the nidus 4 days later. The last follow-up DSA, performed at a median of 13 months (95% CI 12 to 18 months), showed a total occlusion of the nidus in all but one case. Adjunctive radiosurgery after embolization was not needed in this series.
Discussion
While the decision to treat a ruptured AVM is well established, even though sometimes technically challenging, the management of unruptured brain AVMs remains controversial. Moreover, several therapeutic modalities have been proposed, with different approaches, latency periods, and clinical and angiographic outcomes. To date, in the absence of clear-cut guidelines for unruptured cases, treatment should focus on low procedure-related complication rates and high and rapid total cure rates. Since the annual risk of hemorrhage from unruptured, low-grade bAVMs has been estimated at 2.2–3%,19 complication rates should remain, at most, within these percentages.
Nidus eradication with one or at most two interventions seems to represent the ideal situation, especially for SM grade I and II malformations. For this reason microsurgery for small AVMs has been considered better than radiosurgery or interventional neuroradiology in classic case series.11 In a recent study by Moon et al,2 of 85 patients with grade I and II SM unruptured bAVMs, all patients were cured; 23.5% experienced temporary postoperative neurologic deficits and 3.5% had new, permanent, clinical impairment.
Significant disparity exists in the published literature about the outcomes of embolization as a curative treatment for bAVMs. However, in our series, of a similar sample size (73 patients) to that of Moon et al, we report comparable results, with a high cure rate (95%) and low permanent morbidity (2.7%). To the best of our knowledge, this is the first case series to report such results for curative endovascular embolization in low-grade bAVMs.
The use of non-adhesive embolic agents has resulted in the publication of higher nidus obliteration rates of between 51% and 90%,15 20 but there is still a lot of confusion about techniques, material and outcomes, when endovascular approaches are used.17 In the series presented here, the use of more than one microcatheter injection (41% double arterial and 14% arterial and venous) simultaneously resulted in significant reduction of the procedure time and in higher total nidus occlusion rates in a single session.
Multiple microcatheter injections did not yield higher perprocedural or technical complications in the series presented here; on the contrary, the procedures were less time-consuming, which favored better clinical outcomes. This is in accordance with the findings of Renieri et al,21 who reported higher nidus occlusion rates for small AVMs and reduction of the number of procedures, as well as improved safety, with the use of double microcatheter injections. Similar observations were reported with transvenous embolization techniques.18 22
In many ways, the techniques proposed here are highly personalized and depend on the angioarchitecture of the AVM, as does the choice of single or dual microcatheterization and arterial and/or venous access. A consistent standardized protocol for treatment in a highly skilled center of reference seems to affect clinical outcome. In the approach presented here, mastering the techniques and, most importantly, the decision about the preferred embolization strategy are, in our opinion, vital determinants of good outcome.
Even though low SM malformations were considered as ‘benign’ in the ARUBA study,23 the cumulative risk of hemorrhage and sometimes the psychologic impact for young patients demanding to be cured cannot always be neglected. In several centers of reference the publication of ARUBA did not change the management of unruptured bAVMs, including low-grade SM lesions24; in some centers, even higher rates of SM grade I and II malformations are being treated after ARUBA, than previously before the study was complete. Several recent studies challenge the results of ARUBA, showing very good clinical outcomes and high occlusion rates with both microsurgery and embolization.25 26
The study we present here supports these results, with very high immediate total occlusion rates and low complication rates. The series comprises mostly young patients of around 40 years of age. Around 60% of them had ruptures at presentation; for the remaining 40% the reasons for treatment varied from initiation of treatment (embolization or radiosurgery) in another center, which itself modifies the natural history of the lesion unfavorably, to patients’ demand or the presence of specific characteristics that implied a worse natural history.
Even though the majority of the cases in our study were ruptured or had a modified natural history due to previous therapeutic attempts, some patients simply demanded eradication of the nidus. In patients with a 45-year life expectancy at the time of consultation, the cumulative risk of hemorrhage is difficult to estimate. The risk of the intervention remains, in our opinion, justified in selected patients; because with a single-session treatment we totally obliterated the nidus is an advantage to consider, both for hemorrhage risk reduction and cumulative procedure-related complication rates.
Potts et al,27 in a prospective study of 232 ruptured and unruptured low-grade bAVMs, treated by microsurgery, reported very high nidus eradication rates (94% angiographically confirmed) and 3.5% procedure-related complications. In another retrospective study with ARUBA-eligible patients treated with microsurgery or multimodality techniques,28 a permanent complication rate of 7% was reported. These results are in line with the clinical outcomes presented here. As in the series of Potts et al, we found that the presence of rupture at presentation was a strong predictor of the clinical outcome. They concluded that the ’gold standard' should be surgery with adjunctive preoperative embolization.
In our study two embolization sessions were deemed necessary in only 15% of the patients, as opposed to 43% of adjunctive preoperative embolization used in the Potts series. In another recent series on 85 unruptured low SM bAVMs, Moon et al 2 used preoperative embolization in 52% of cases and obtained similar angiographic and clinical outcomes. In our opinion, the strategy we present here represents a valid alternative, with, in most cases (85%), single-session curative embolization and similar clinical and angiographic outcomes.
Previously, microsurgery was considered superior to radiosurgery and endovascular treatment, mainly owing to the faster, one-session, higher total obliteration rates.11 Even though microsurgery remains the gold standard of treatment for these lesions, the endovascular strategy presented in this study, when mastered appropriately, seems to demonstrate equally high nidus obliteration rates, with procedure-related morbidity that remains among the lowest published.
Until further data come to light, the decision to treat an unruptured bAVM should be taken only when the expected procedure-related complication rate is lower than the natural history of rupture of the AVM. Meanwhile, for the treatment of patients with ruptured, low-grade bAVMs, the curative endovascular strategy presented seems a valid alternative to classic approaches, with good clinical and angiographic outcomes.
Limitations
Even though this study is prospective, the number of subjects remains limited; as a consequence, statistical comparisons, even when statistically significant, lack power. Larger studies are needed confirm the findings for the group comparisons. Long-term follow-up is also needed to confirm the results; this is underway for the presented series. Most unruptured cases in this cohort had been previously partially treated in other center, without achieving eradication of the nidus, thus the choice of treatment was justified. Nevertheless this represents a bias, for the group comparisons. The design of the study did not allow for comparisons with intact, non-ruptured bAVMs. Further studies are needed to explore further the usefulness of the technique in intact, non-ruptured bAVMs.
Conclusions
Stand-alone, single-session endovascular treatment of SM grade I and II brain arteriovenous malformations seems to be a valid strategy in ruptured and selected unruptured cases, offering high nidus eradication rates and good clinical outcomes. The appropriate choice of embolization strategy is important and multiple microcatheter techniques should be mastered in order to obtain optimal results. More studies are needed to confirm these results.
References
Footnotes
Contributors All authors made substantial contributions to the conception or design of the work; or the acquisition, analysis, or interpretation of data for the work; agreed to the final draft of the work and gave final approval of the version to be published and agree to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.
Competing interests The authors declare no competing interests for the study presented.
Ethics approval Dupuytren University Hospital-Haute Vienne ethics commitee.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Supplementary data are available upon request from the corresponding author.
Patient consent for publication Obtained.