Delayed Proximal Migration of the Radius Following Radial Head Resection for Management of a Symptomatic Radial Neck Nonunion Managed with Radial Head Replacement: A Case Report and Review of the Literature

Abstract

Background

Long-term outcomes of radial head resection for radial head fracture have shown mixed outcomes, depending on the integrity of the soft-tissue stabilizers of the elbow, forearm, and wrist.

Methods

We report a case of a symptomatic delayed proximal migration of the radius after radial head excision for radial neck nonunion which was managed with a staged radial head replacement. Informed consent was obtained from the individual in this case report.

Results

At 7 months after radial head replacement, the patient had 90 degrees of forearm supination and 85 degrees of pronation. Elbow range of motion was from 10 degrees short of full extension to 155 degrees of flexion. Her Disabilities of the Arm, Shoulder, and Hand score was 21.4 at 7 months and 6.48 at 38 months.

Conclusion

There is insufficient evidence to reliably predict which patients can be managed definitively with radial head excision without risk of later proximal migration of the radius. The authors suggest the use of acute radial head arthroplasty when the index injury is secondary to a traumatic mechanism, reserving radial head excision for patients with radio-capitellar arthritis. Further research of the pathology and healing of concomitant soft-tissue injuries seen in conjunction with radial head fractures is warranted to guide their treatment.

Introduction

Many patients have good functional outcomes after radial head resection; however, those with inadequate stability may develop sequelae including valgus instability, proximal migration of the radius, or distal radial ulnar joint (DRUJ) subluxation with pain and dynamic instability.¹-⁵ Though opinions differ as to the clinical significance of proximal migration,⁶ some authors have advocated for early prosthetic replacement of the excised radial head to prevent this potential complication.⁷ More recently, use of a radial head replacement as a temporary spacer to allow for soft tissue healing with later removal of symptomatic hardware has been advocated.⁸-¹⁰ However, late symptomatic proximal migration after radial head replacement removal has also been described.⁹,¹¹ The ideal management of these complicated patients is unclear. We provide a case report that describes management of a late, symptomatic proximal radial migration following prior radial head resection for symptomatic radial neck nonunion.

Case Report

A 60-year-old right-hand-dominant female sustained right radial head and left radial neck fractures after a fall. She was initially treated non-operatively with a sling and early range of motion for both injuries. The patient presented to our institution 8 weeks after the initial injury due to persistent left elbow pain. Elbow range of motion on the right was from -5° to 145° flexion with 90° of pronation and 90° of supination, as measured with a goniometer. Elbow range of motion on the left was from -5° to 150° flexion with 90° of pronation and 90° of supination. Additionally, the patient had pain to palpation over the radial neck on the left with palpable clicking during forearm pronation and supination. She had no wrist or forearm pain with palpation or range of motion and her DRUJ was stable when shucked. At this time she was diagnosed with an asymptomatic delayed union of a right Mason I radial head fracture and a symptomatic delayed union of a left radial neck fracture.

The patient returned 6 weeks later, now 14 weeks post-injury, with persistent left elbow pain with activities of daily living and aching at rest despite anti-inflammatories and physical therapy. She was diagnosed with a persistent symptomatic left radial neck delayed union (Figure 1). She was taken to the operating room 5 months from the date of her injury for a left radial head resection via a Kaplan approach. Intraoperative exam under fluoroscopy revealed a stable ulnohumeral joint in flexion and extension, and a pivot-shift test assessing the status of the lateral ulnar collateral ligament confirmed posterolateral rotatory stability.¹² The elbow and DRUJ were stable after radial head resection. The elbow was dressed in a soft, bulky dressing until 2 weeks. and the patient was allowed to move the elbow freely and started on a physical therapy program. There were no early postoperative complications.

Figure 1.

(A) Anteroposterior and (B) lateral radiographs taken of the left elbow 6 weeks after the initial visit (14 weeks after injury) show a radial neck nonunion. The patient continued to be symptomatic on this side.

The patient did well for 4 months, after which she developed wrist pain with daily activities. Clinical examination revealed elbow range of motion from -5° to 150° flexion with 90° of pronation and 90° of supination. Radiographs and magnetic resonance imaging (MRI) revealed proximal migration of the radius with 5 mm ulnar positive variance and increased signal on the coronal and sagittal T2 MRI sequences consistent with ulnar impaction syndrome (Figure 2). Clinical diagnosis was chronic longitudinal instability of the forearm with delayed proximal migration of the radius resulting in positive ulnar variance with subsequent ulnar impaction syndrome.¹³

Figure 2.

(A) Postero-anterior radiograph of the left wrist taken 1 month after radial head excision shows neutral ulnar variance. Postero-anterior radiograph of the (B) left and (C) right wrist taken 4 months after radial head excision shows +5 mm positive ulnar variance on the left and +3.5 mm positive ulnar variance on the right. (D) An MRI of the left wrist shows evidence of ulnar impaction syndrome.

Given the patient’s symptoms and the findings of ulnar impaction on MRI, revision surgery was performed. Surgery consisted of delayed radial head replacement, 9 months after the initial injury and four months after her initial radial head excision. The radiocapitellar joint was exposed via the previous Kaplan approach and a radial head arthroplasty was performed (Acumed, Hillsboro, OR). Prior to placement of the implant, intraoperative fluoroscopy confirmed a 4 mm ulnar positive variance with the forearm in supination, which was reduced to 1 mm after radial head replacement (Figure 3). The elbow and DRUJ were stable to pronation and supination after the procedure and the DRUJ was stable to shuck testing (Figure 4). The patient was immobilized in a long arm splint until 2-week follow-up, at which point her wrist pain had subjectively started to improve. She then began physical therapy to improve range of motion. Radiographs of her wrist taken at 6 weeks postoperatively revealed a 3.5 mm ulnar positive variance, but the patient continued to report improved wrist pain compared with preoperatively (Figure 5). Radiographs of her wrists taken at 7 months post-operatively revealed a +2 mm ulnar positive variance compared to a +3 mm ulnar positive variance on the asymptomatic right side. At 7 months post-operatively she had 90 degrees of forearm supination in 90 degrees of flexion and 85 degrees of pronation. Elbow range of motion was from 10 degrees short of full extension to 155 degrees of flexion. Her Disabilities of the Arm, Shoulder, and Hand (DASH) score at that time was 21.4. At 38 months post-operatively she had a DASH score of 6.48, consistent with continued improvement and less disability.

Figure 3.

Intraoperative fluoroscopic images of the left wrist (A) with +5 mm positive ulnar variance before radial head replacement and (B) with +1 mm positive ulnar variance after replacement.

Figure 4.

(A) Anteroposterior and (B) lateral radiographs of the elbow following radial head replacement

Figure 5.

(A) Postero-anterior radiograph of the left wrist 6 weeks following radial head replacement shows some loss of the improved positive ulnar variance post-replacement (+5 mm → +1 mm → +3.5 mm). (B) Anteroposterior radiograph of the asymptomatic right wrist with +3.5 mm ulnar positive variance.

Discussion

Patients with operative fractures of the radial head and neck can be treated with fragment excision, open reduction and internal fixation, radial head replacement, or complete radial head excision. In addition to patient factors such as age, handedness, and activity level, the decision on how best to manage an individual injury is informed both by the pattern of the fracture and the concomitant soft-tissue injuries. The case presented here demonstrates how unrecognized or sub-clinical soft-tissue injury may result in proximal migration of the radius after isolated radial head excision, but that excellent clinical results may still be achieved if it is recognized in a timely manner and treated with radial head replacement.

Multiple studies have shown that the majority of patients have good functional outcomes after radial head excision.²,⁵,⁶,¹⁴-¹⁸ Antuna et al14 followed 26 patients younger than forty years of age with radial head fractures treated with primary radial head excision for a minimum of 15 years. Overall patients did well with a mean Mayo Elbow Performance Score (MEPS) of 95, mean DASH score of 6, and 92% with good or excellent results. However, 3 patients (11%) had wrist pain and radiographic evidence of proximal migration and 1 had DRUJ instability. Similarly, Iftimie et al² reported on 27 patients who underwent resection arthroplasty after radial head fracture without elbow instability. At mean 17-year follow-up, they found overall very good functional results but noted symptomatic proximal migration (wrist pain) in 2 patients (7%) who were subsequently treated with an ulnar shortening osteotomy. The authors suggest these patients likely had unrecognized or sub-clinical Essex-Lopresti injuries, similar to our patient. Moreover, the timing of radial head excision does not seem to influence outcomes, as patients tend to do well after both early and late excision.¹⁵,¹⁶

However, not all patients have successful results following radial head excision. In 1931, Brockman reported 2 cases of radial head excision for fracture with subsequent wrist pain due to proximal migration of the radius.¹ Taylor in 1964 described radiographic DRUJ subluxation in 37 of 58 cases of radial head excision, with symptoms of pain or instability present in 29 patients (50%) 7 years after surgery.⁴ Similarly, in 1957 McDougall and White reported that 12 of 44 patients (27%) had wrist symptoms after excision of the radial head for fracture due to proximal migration of the radius.³ They hypothesized that the degree of proximal migration was related to the severity of the soft tissue damage at the time of the initial injury, with acute migration occurring at the time of injury if the DRUJ was disrupted and the radial head dislocated while late migration occurred as a results of stretching of the articular disc and interosseous membrane. Proximal migration was also thought to be related to the activity of the patient, as it was found to be more common in manual laborers.

It should be noted that not all patients who have radiographic evidence of proximal migration have clinical symptoms. Morrey et al⁶ reported 20-year follow-up data on 13 patients after radial head excision with an average proximal migration of 1.9mm (0-5mm) and no significant correlation between loss of motion (pronation/ supination) and the amount of proximal migration of the radius. Only 4 of their 13 patients (30%) with proximal migration experienced wrist pain. Similarly, Yalcinkaya presented 14 elbows after radial head excision, 8 of which had proximal migration and an increase in carrying angle, with overall good clinical outcomes.⁵ Goldberg et al¹⁷ noted asymptomatic proximal migration in 8 of 37 patients (22%) despite radiographic evidence of migration as well as osteoarthritic changes.

In patients with radial head fracture with soft-tissue injury, such as medial collateral ligament tear, interosseous membrane injury (Essex-Lopresti injury), or DRUJ instability, lateral mechanical support must be provided by either the native radial head or by a radial head replacement in order to give the best chance of functional recovery. As early as 1953, some authors were advocating early prosthetic replacement of the excised radial head to prevent proximal migration.⁷ More modern implants have been shown to have overall excellent clinical outcomes.⁸-¹⁰,¹⁹ Nonetheless, some patients with acute complex elbow injuries who are treated with radial head replacement develop symptoms related to the arthroplasty itself, such as pain or a mechanical block to motion, and require later removal of the prosthesis. Knight and Rymaszewski suggested using the radial head replacement as a temporary spacer in unstable elbows, allowing the soft-tissue injuries to heal, and removing the prosthesis later if it becomes symptomatic.⁹ Harrington et al⁸ and Wretenberg et al¹⁰ both supported this idea and reported several patients with unstable elbow injuries who had radial head arthroplasties that were later removed after soft-tissue healing, leaving the patients with asymptomatic, stable elbows.

Even with adequate time for soft tissue healing, there can be symptomatic proximal migration of the radius. Van Riet et al. reported one case of radial head arthroplasty in whom soft tissues did not heal over a prolonged period of time, leading to a poor functional result after removal of the radial head implant.²⁰ Many authors believe that the interosseous membrane never truly heals but that it may scar down in a relatively contracted position, similar to many other soft tissue injuries.²¹ For this reason, chronic cases with established proximal migration are difficult to treat due to the radioulnar length inequality. Ulnar shortening does not offer a predictable solution in these patients,²²,²³ and one must consider the potential complication of nonunion of the ulnar osteotomy,²⁴ particularly in a patient like ours who has already shown poor potential for union. Szabo et al.²¹ presented a technically difficult solution to this challenging problem involving guided growth of the radius using Ilizarov technique followed by allograft radial head implantation. An alternative salvage procedure is acute shortening of the ulna and creation of a one-bone forearm. This procedure is advocated for by some authors,²²,²³ but others have reported unpredictable results.²⁵

In our case of the radial neck nonunion, the radial head is analogous to the radial head replacement acting as spacer prior to its excision. Though the initial evaluation showed no evidence of interosseous membrane injury, she likely had a sub-clinical injury that did not heal despite the prolonged 5-month period after her original injury before the resection of her radial head. Alternatively, it is possible that the interosseous membrane and other soft-tissues were attenuated and could not stand up to the repetitive stresses they faced after radial head resection. She went on to develop symptomatic proximal migration of the radius after radial head excision, despite documented elbow stability without forearm or wrist pain at the time of radial head resection. It may be that all radial head and neck fractures have some element of an interosseous ligament injury, i.e. in order for these fractures to occur the radius must be loaded and forces are transmitted through the interosseous ligament resulting in some injury. The presence of this type of sub-clinical injury and potential reasons for poor healing or longterm interosseous ligament incompetence is a potential area for further research.

Fortunately, in our case, we had the option of performing a delayed radial head replacement, which resolved the patient’s symptoms from ulnar impaction syndrome, without the need for gradual pre-implantation lengthening. We believe this offered our patient a more reliable recovery compared with procedures directed at the ulna, such as an ulnar shortening osteotomy because by replacing her radial head, we treated the cause of her wrist symptoms, which was the proximal migration of her radius, rather than the symptom of ulnar impaction. This relieved her pain despite losing some of the gained ulnar positive variance during the post-operative period and she continues to be functioning very well 3 years after surgery. It should be noted that there is some variability in the measurement of ulnar variance depending on the position of the arm and the quality of the radiographic technique that may contribute to the measured differences in this patient.²⁶, ²⁷ The present case is novel as it demonstrates a sub-clinical Essex-Lopresti lesion that became apparent after radial head resection for a radial neck nonunion which was successfully treated with delayed radial head replacement.

At this point in time there is an insufficient body of clinical evidence to reliably predict which patients can be managed definitively with radial head excision that will not be compromised by delayed symptomatic proximal migration of the radius versus those that should be managed with acute radial head arthroplasty. Following experience gained from the present case presentation, the authors plan to adopt a clinical algorithm to proceed with acute radial head arthroplasty when the index injury was secondary to a traumatic mechanism. Radial head excision will be reserved for isolated presentations with radio-capitellar arthritis. Further research of the pathology and healing of concomitant soft-tissue injuries seen in conjunction with radial head fractures is warranted to guide their treatment.