Rationale of salvage procedures for failed resection arthroplasties of the distal radio-ulnar joint

Introduction When a distal radio-ulnar joint (DRUJ) causes refractory pain and there is extensive damage to the joint cartilage, the only available surgical option is to sacrifice the joint by complete resection of the distal ulna (popularized by Darrach), by partial resection with or without interposition (e.g. Bowers’ hemiresection or Watson’s ‘matched ulna’) or by fusion of the radius and distal ulna with proximal pseudoarthrosis [Sauvé-Kapandji (S-K) procedure]. These options, especially Darrach’s resection, have been widely used, and the satisfactory pain and mobility outcomes achieved (>80%) have long been documented1. However, it is increasingly acknowledged that these arthroplasties can fail, with patients reporting pain and weakness in grip and object lifting, and even clunking with pronosupination due to instability of the distal forearm2,3. This instability takes place in the anteroposterior plane, with X-rays showing radioulnar convergence and wear in the medial cortex of the radius at the ulnar stump end4 (Figure 1). This complication is reported as radioulnar impingement syndrome2 or convergent instability of the distal ulnar stump5, and various dynamic radiographic studies have used different methods to determine its frequency and the influence of distinct procedures or techniques on this loss of stability. Since the introduction of a novel radiographic view by Lees and Scheker in 19976, radiographic radio-ulnar convergence has been observed in 100% of cases and affects patients equally after Darrach’s resection, S-K procedure, or partial resection (Figure 2). One explanation that has received little attention is that DRUJ is a load joint and critical for lifting objects. When a weight is held in the hand with the elbow flexed and forearm in neutral rotation, the radius is centred on the ulnar head, producing a transverse load and exerting a compressive force7. Moreover, the contraction of the muscles connecting the ulna and radius favours convergence once these bones are free of contact with the ulnar head. García-Elias demonstrated the role of the anterior brachial muscle, pronator quadratus and long abductor muscle of the thumb in this phenomenon8 (Figure 3), whereas other authors highlighted the role of the short thumb extensor and the flexor digitorum profundus of the second and third fingers. The percentage of distal ulna resections with symptoms of radioulnar impingement ranges between 8% and 50%, depending on the series. Although no association has been established with any specific radiographic parameter, an excessively high ulnar osteotomy appears to have a negative effect9. It has been claimed that the ulnar stump meets an area of the radius that is unfavourable for contact8, but we believe that there may be more important biomechanical causes of this complication. In DRUJ surgery, the instability of the distal ulna can be reduced by preserving soft tissue stabilizers such as the triangular fibrocartilage complex (TFCC) and radio-ulnar ligaments, posterior ulnar tendon sheath, interosseous membrane and * Corresponding author Email: hdezcp@hotmail.com


Introduction
When a distal radio-ulnar joint (DRUJ) causes refractory pain and there is extensive damage to the joint cartilage, the only available surgical option is to sacrifice the joint by complete resection of the distal ulna (popularized by Darrach), by partial resection with or without interposition (e.g.Bowers' hemiresection or Watson's 'matched ulna') or by fusion of the radius and distal ulna with proximal pseudoarthrosis [Sauvé-Kapandji (S-K) procedure].
These options, especially Darrach's resection, have been widely used, and the satisfactory pain and mobility outcomes achieved (>80%) have long been documented 1 .
However, it is increasingly acknowledged that these arthroplasties can fail, with patients reporting pain and weakness in grip and object lifting, and even clunking with pronosupination due to instability of the distal forearm 2,3 .This instability takes place in the anteroposterior plane, with X-rays showing radio-ulnar convergence and wear in the medial cortex of the radius at the ulnar stump end 4 (Figure 1).This complication is reported as radio-ulnar impingement syndrome 2 or convergent instability of the distal ulnar stump 5 , and various dynamic radiographic studies have used different methods to determine its frequency and the influence of distinct procedures or techniques on this loss of stability.
Since the introduction of a novel radiographic view by Lees and Scheker in 1997 6 , radiographic radio-ulnar convergence has been observed in 100% of cases and affects patients equally after Darrach's resection, S-K procedure, or partial resection (Figure 2).One explanation that has received little attention is that DRUJ is a load joint and critical for lifting objects.When a weight is held in the hand with the elbow flexed and forearm in neutral rotation, the radius is centred on the ulnar head, producing a transverse load and exerting a compressive force 7 .
Moreover, the contraction of the muscles connecting the ulna and radius favours convergence once these bones are free of contact with the ulnar head.García-Elias demonstrated the role of the anterior brachial muscle, pronator quadratus and long abductor muscle of the thumb in this phenomenon 8 (Figure 3), whereas other authors highlighted the role of the short thumb extensor and the flexor digitorum profundus of the second and third fingers.
The percentage of distal ulna resections with symptoms of radioulnar impingement ranges between 8% and 50%, depending on the series.Although no association has been established with any specific radiographic parameter, an excessively high ulnar osteotomy appears to have a negative effect 9 .It has been claimed that the ulnar stump meets an area of the radius that is unfavourable for contact 8 , but we believe that there may be more important biomechanical causes of this complication.
In DRUJ surgery, the instability of the distal ulna can be reduced by preserving soft tissue stabilizers such as the triangular fibrocartilage complex (TFCC) and radio-ulnar ligaments, posterior ulnar tendon sheath, interosseous membrane and

Trauma & Orthopaedics
The objectives of this review were to describe reported procedures for revision surgery after failed resection arthroplasty of the distal radioulnar joint, to critically evaluate the indications for these procedures and to investigate their functional outcomes and complications.

Materials and methods
We performed an Internet search of the MEDLINE and EMBASE databases.

Results
There was a weak level of scientific evidence, with no randomised controlled trials comparing different operative managements and no prospective case series.This review is limited to data from retrospective studies and case reports.

Conclusion
Distal tenodesis is ineffective.Placement of a partial or total distal radioulnar joint prosthesis is the sole option for re-establishing the distal pivot point necessary for adequate tensioning of the interosseous membrane and for restoring the optimal transfer of loads from radius to ulna and vice versa.However, concerns have been raised about the utilization of these procedures in young and in high-demand patients.Allograft Licensee OA Publishing London 2013.Creative Commons Attribution License (CC-BY) Competing interests: none declared.Conflict of interests: none declared.
pronator quadratus muscle 10 .However, none of these structures offer protection against convergence (Figure 4).Radio-ulnar impingement appears to be inevitable, but the onset of symptoms is unpredictable and depends more on the physical demands of the patient than on variations in DRUJ resection techniques.
The objectives of this review were to outline published procedures for revision surgery after failed resection arthroplasty of the DRUJ, to critically evaluate the indications for these procedures and to investigate their functional outcomes and complications.

Materials and methods
An Internet search of MEDLINE and EMBASE databases was performed on 11 February 2013, using the search terms: 'Failed Darrach procedure', 'Failed Sauvé-Kapandji procedure', 'Failed Bower´s hemiresection procedure', 'Failed matched ulna arthroplasty', 'Failed distal ulna resection', 'Salvage of failed DRUJ resection', 'Radio-ulnar impingement syndrome', 'Treatment of radio-ulnar impingement', 'Stabilization of ulnar stump' and 'Salvage procedure for distal end of the ulna'.The reference lists in all relevant identified papers were also screened to identify other studies for inclusion in this review.Letters to the Editor, purely radiological studies and articles on DRUJ primary surgery or the non-surgical management of failed DRUJ arthroplasty were excluded from this review.

Results
Only a weak level of scientific evidence was found.There were no randomized controlled trials comparing different operative management options and no prospective case series.This review was therefore limited to retrospective studies and case reports.
The identifier 'failed Darrach procedure' traced eight studies published between 1986 and 2010, 'failed Sauvé-Kapandji procedure' traced 13 articles between 1991 and 2011, and 'failed Bower´s hemiresection procedure' one article in 1996, whereas there were no results for 'failed matched ulna arthroplasty' .The articles on revision surgery for distal ulna resection procedures can be classified as soft tissue procedures, bone procedures and arthroplasties with implants or other salvage techniques.Competing interests: none declared.Conflict of interests: none declared.

Soft tissue procedures
Soft tissue procedures can be divided between distally-based tenodeses of the ulnar stump and interposition arthroplasties.
Tenodesis has mainly been carried out using the hemi-tendon of flexor carpi ulnaris (FCU), extensor carpi ulnaris (ECU) or both (Figure 5).The original article to propose stabilization of a failed Darrach procedure with FCU tenodesis was published in 1984 by Tsai and Stilwell 11 .Kessler and Hecht 12 were the first to suggest using a strip of ECU to stabilize distal ulnar subluxations in rheumatoid patients, applying a similar method to that described for the FCU.There are no detailed reports on the application of ECU tenodesis in isolation to salvage a painful distal radio-ulnar impingement, but it has been used in primary ulnar head resection to prevent instability.Breen and Jupiter combined distally based FCU with proximally based ECU tenodesis to achieve both static and dynamic stabilization of the distal ulnar stump and obtained good outcomes in three patients 13 .
In 1992, Johnson 14 proposed pronator quadratus interposition and, in 2003, Shah and Klimisch 15 reported the interposition and theoretical tenosuspension of the radius by means of tenodesis with the tendon of the brachioradialis.Finally, authors such as Berger and Cooney 16 and Sotereanos 17 recommended interposing an allograft of Achilles' tendon between the radius and ulnar stump (Figure 6).Kleinman and Greenberg 18 applied a combination of soft tissue procedures, stabilizing the distal ulnar stump with a distally based ECU tenodesis and adding a dorsal transfer of the pronator quadratus muscle to interpose a soft tissue cushion between radius and ulna.

Bone procedures
Some procedures to palliate instability have focused on the bone segments of the forearm.Although shortening of the unstable ulnar stump is highly controversial, Cooney et al. 19 performed a large number of wide distal ulna resections due to tumours and reported a very low incidence of complications and symptomatic radio-ulnar impingement.Wide ulnar resection has subsequently been applied in patients with failed Darrach's surgery, among other conditions, with apparently good outcomes being reported by Wolfe et al. 20 .Competing interests: none declared.Conflict of interests: none declared.
Based on the assumption of a lower incidence of symptomatic radio-ulnar convergence with Bower's hemiresections or Watson's matched ulnar resections than with the Darrach and Sauvé-Kapandji techniques, Watson and Brown 21 performed joint levelling by lengthening the ulna to achieve a situation similar to the 'matched ulna', obtaining good results but acknowledging the technical difficulty of the procedure.Ulna lengthening can be conducted by step osteotomy or by tricortical bone grafting 22 .
Shortening the radius is conceptually equivalent to lengthening the ulna and is probably responsible for fewer complications, although the only published data is from a study of two patients by García-Elias in 2002, who reported good results 8 .González del Pino and Fernández 23 did not address the biomechanics of radio-ulnar impingement but proposed a revision technique to convert a painful hemiresection into a Sauvé-Kapandji arthroplasty (Figure 7).

Arthroplasties with implant
The third DRUJ salvage option is to implant a prosthesis (Figure 8).McMurtry et al. 24 reported on 40 Swanson silicone ulnar head arthroplasties in 37 patients, with a minimum of 12-months follow-up; eight of the procedures were for patients with failed Darrach surgery, but a satisfactory outcome was obtained in only three of these.
The largest sample of failed resection arthroplasty cases was studied by van Schoonhoven et al. in 2000 25 , who used Herbert prostheses (Herbert Ulnar Head Prosthesis, Gebrüder Martin GmbH & Co KG, Tuttlingen, Germany) and reported good pain outcomes, a mean pronosupination of 151° and a 43%-73% improvement in the grip strength of the contralateral hand.A follow-up of the same patients at 12 years 26 found that the good results had persisted and that the implants had not required revision.Competing interests: none declared.Conflict of interests: none declared.

Critical review
In 2006, Fernández et al. 27 proposed a modification of the Herbert prosthesis procedure, using a spherical ulnar head, for the revision of failed Sauvé-Kapandji.It maintains the arthrodesis of the distal ulna, which is given a concave shape to fit the prosthesis, and they even suggested osteotomy of the distal metaphysis of the radius to fit the implant without removing the fusioned ulnar head.Willis et al. 28 used an Avanta uHead prosthesis (Small Bone Innovations Inc, Morrisville, PA, USA) to treat 22 patients with failed distal ulna resection and reported excellent results in 18 of these at 2 years, although there were also 2 cases of aseptic loosening.
Watts et al. 29 proposed a revision algorithm for using different implants as a function of the soft tissue stabilizers of the DRUJ and the continuity with the styloid process (Figure 9).Total arthroplasty is mandatory when the soft tissues do not guarantee the stability of the distal ulna prosthesis, with the most widely used implant being Scheker's prosthesis 30 .

Other salvage techniques
Finally, the creation of a one-bone forearm is presented as the last resort for patients in whom previous techniques have failed.It consists of excision of the distal ulna, followed by osteotomy, transfer and fixing of the radius in slight pronation by means of a plate.This procedure should not be confused with radio-ulnar arthrodesis, although both result in the loss of pronosupination.In a study of 23 patients with tumours and two instabilities due to failed ulnar resection, Peterson et al. 31 reported that less favourable outcomes were associated with traumatic disease and a history of iterative surgical procedures.Allende and Allende 32 described one-bone forearm procedures in seven patients with post-traumatic forearm defects, all of whom evidenced a stable and pain-free forearm at 6 years post-surgery and had returned to their previous occupation.

Discussion
Although symptomatic radio-ulnar impingement is relatively frequent after DRUJ resection arthroplasty, no prospective or randomized clinical studies have compared different techniques for its treatment, and there has been only one experimental controlled study 33 .Proposed procedures have been highly varied and are only supported by retrospective  Competing interests: none declared.Conflict of interests: none declared.
studies with small sample sizes and sometimes heterogeneous DRUJ lesions, yielding only level IV evidence and a recommendation grade of C.
More articles have been published on failures with Sauvé-Kapandji and Darrach techniques than on failures with hemiresection and matched Ulna techniques, suggesting that problems may be more frequently associated with complete than the with partial distal ulna resections.
Distal tenodesis techniques have generally been discredited.García-Elias 8 demonstrated the anatomical foundations of the incompetence of these stabilizations as a function of wrist position.All tenodeses remain lax in ulnar deviation of the hand.In a biomechanical study of cadavers, Sauerbier et al. 33 compared the effec tiveness of ulna stabilization after distal resection by tenodesis with ECU and FCU hemitendon, pronator quadratus interposition, or placement of distal ulna prosthesis.They observed that radio-ulnar convergence was not prevented by tenodesis or interposition method after distal ulna resection and that instability may be exacerbated by a pronator quadratus interposition.
The DRUJ can be compared to a bell (Figure 5c).If the clapper moves and hits the bell, it clangs (the wrist hurts).We can stop the movement of the clapper by increasing the tension of the rope holding it (as in tenodesis), but we cannot prevent the clanging if it is the bell (radius) that moves, which is exactly occurs in the wrist.
Interposition of a tendon allograft between the ulna and radius was studied by Sotereanos et al. in 4 patients in 2002 17 , and the same author reported a good outcome of this approach in 16 out of 17 patients studied in 2008 34 .
Surgical shortening of the ulnar stump is acknowledged to be irrational and inappropriate as it always results in a new and more proximal impingement, as reported by Bell et al. 2 .Therefore, this procedure is currently in disuse, despite the possibility of obtaining good results from wide resections in tumour disease.Radio-ulnar joint levelling procedures, either by ulnar lengthening or radius shortening, have been described only in case reports or short case series 8,21,22 .
One way of contemplating these treatments is to think of a forearm in which a portion of distal ulna is removed as if it were a grade 3 lever (Figure 10).If d is the resected length of the ulna, the application of a force (f) produces a radio-ulnar convergence with contact at point p.The momentum of the pair of forces is the product of the force applied by distance d; therefore, the momentum increases with a wider ulnar resection.Likewise, in a simple support, the reaction to the application of a weight is always perpendicular to the surface, and the vertical component of the reaction must be equal to the weight borne.The angle of the two vectors increases with ulnar shortening, and the reaction at the point of contact is proportional to the secant of the angle, i.e. the reaction is greater.These mathematical arguments support joint levelling procedures but not iterative shortening of the ulnar stump, consistent with clinical observations.Nevertheless, wide ulnar resections appear to be better tolerated.One explanation may be that when the arm of the ulnar 'lever' is very short, the radio-ulnar convergence requires a very long downward movement of the radius, probably limited by the soft tissues of the elbow, and there can be no reaction force if there is no point of contact.Another possible salvage treatment for failed DRUJ resection arthroplasty is to implant an ulnar head prosthesis, which was found to restore the stability of the DRUJ effectively by simulating the geometry of the ulnar head 33 .However, except for a 12-year follow-up study conducted by van Schoonhoven et al. in 2012 26 , no data are available on the long-term survival of these implants, raising questions about their suitability for younger and higher-demand patients.
Finally, the one-bone forearm procedure is supported by studies of patients who had a different pro-file 31,32 from that of patients with dynamic radio-ulnar impingement usually encountered in clinical practice.This technique has been associated with a high rate of complications, including a lack of bony union, and the function is markedly reduced by the loss of pronosupination.It is therefore important that patients have realistic expectations of the outcome.

Conclusion
There is inadequate scientific evidence to recommend a specific treatment for failed DRUJ resection arthroplasty.Distal tenodesis is ineffective.The placement of a partial or total DRUJ prosthesis is the only option that re-establishes the distal pivot point necessary for adequate tensioning of the interosseous membrane, restoring optimal load transfer from the radius to the ulna and vice versa; however, concerns have been raised about its application in young and highdemand patients.Tendon allograft interposition is an alternative to prosthesis substitution.and c) Forearm with resection arthroplasty of the distal ulna.If d is the resected length of the ulna, the application of a force "f" determines a radio-ulnar convergence with contact at point p.The momentum of the pair of forces is the product of the force applied by distance "d"; therefore, the momentum increases with wider resection of the ulna.(d) Likewise, in a simple support, the reaction to the application of a weight "w" is always perpendicular to the surface, and the vertical component of this reaction must be equal to the weight borne.The angle of the two vectors "µ" increases with the shortening of the ulna, and the reaction at the point of contact is proportional to the secant of the angle, i.e. a greater reaction is produced.

Figure 2 :
Figure 2: Radio-ulnar impingement in Darrach's arthroplasty and partial ulna resection.(a) Darrach procedure in a rheumatic patient, showing exostosis on the ulnar stump and wear of the contact area on the radius.(b and c) Bowers' hemiresection and dynamic radiographic study evidencing radio-ulnar convergence in (c).

Figure 1 :
Figure 1: (a) Radio-ulnar impingement in Sauvé-Kapandji procedure.X-ray shows radio-ulnar convergence and wear in the medial cortex of the radius at the ulnar stump end.(b) Antero-posterior projection of wrist with Sauvé-Kapandji, with the elbow at 90° flexion and intermediate pronosupination.(c) Scheker's radiographic view: in the position reported in (b), the patient holds a weight of 2.2 kg, and radio-ulnar impingement is observed.

Figure 5 :
Figure 5: Anatomic diagram of ulnar stump stabilization by tenodesis.(a) Most techniques use hemitendon of extensor carpi ulnaris (ECU) or flexor carpi ulnaris (FCU).(b) Tenodesis by the Breen and Jupiter technique.(c) Comparison of the biomechanics of tenodesis with the functioning of a bell.The clapper of the bell represents the ulnar stump and the bell itself the radius.The rope represents the tendons that stabilize the ulnar stump in tenodesis.Tenodesis is ineffective when the bell (radius) moves against the fixed clapper (ulnar stump).

Figure 6 :
Figure 6: Radio-ulnar interposition arthoplasty with Achilles' tendon allograft.(a) Exposure of painful distal ulna prosthesis.(b) Implant removal.(c) Preparation of Achilles' tendon allograft as a cushion.(d) Placement of two suture anchors on the ulnar side of the radius.(e) Interposition of the allograft between radius and ulna; the allograft is stabilized by the suture anchors.(f) Part of the tendon allograft is used to wrap the ulnar stump.

Figure 7 :
Figure 7: Conversion of distal ulna hemiresection in Sauvé-Kapandji procedure as suggested by González del Pino.(a) Painful hemiresection arthroplasty.(b) Operative view: A is interposed between the radius and ulnar styloids to favour distal radio-ulnar arthrodesis.The bone defect created leads to proximal ulna pseudoarthrosis.(c) Radiographic view of the procedure.

Figure 10 :
Figure 10: Comparison of the forearm with a grade 3 lever.(a) Normal forearm.(band c) Forearm with resection arthroplasty of the distal ulna.If d is the resected length of the ulna, the application of a force "f" determines a radio-ulnar convergence with contact at point p.The momentum of the pair of forces is the product of the force applied by distance "d"; therefore, the momentum increases with wider resection of the ulna.(d) Likewise, in a simple support, the reaction to the application of a weight "w" is always perpendicular to the surface, and the vertical component of this reaction must be equal to the weight borne.The angle of the two vectors "µ" increases with the shortening of the ulna, and the reaction at the point of contact is proportional to the secant of the angle, i.e. a greater reaction is produced.