Between January 2000 and April 2007, 122 patients with posttraumatic elbow stiffness were treated with open surgical release by a single surgeon. Of these patients, seventy-six had <100° of maximum flexion. They had skeletally mature elbows that had been essentially normal prior to injury, and all had a stable joint and radiographic evidence of fracture union after the injury. Patients with an intra-articular injury were included when it was determined that fracture-healing did not affect ulnohumeral motion on the basis of radiographic findings of a congruent articular surface and an intact joint space. We excluded twenty-five patients who had (1) incongruent joint surfaces or substantial osteophytes that might affect motion, (2) skin or muscle contractures as the main features of stiffness, (3) associated central nervous system injury, or (4) severe elbow ankylosis with <30° of total motion or radioulnar synostosis associated with heterotopic ossification. Five patients who had a mechanical block mainly caused by abundant anterior callus following an intercondylar fracture were also excluded. One patient died and three patients were lost to follow-up. Therefore, forty-two elbows in forty-two patients were included in this study. Our institutional review board approved a protocol for the retrospective review of their medical records and for inviting patients with inadequate follow-up to return for examination and radiographs.
There were twenty male and twenty-two female patients with a median age of thirty-seven years (range, fifteen to sixty-nine years). The dominant side was involved in twenty-two patients. The median time from the injury to the surgical release was ten months (range, four to eighty-four months). The indication for surgical release was a persistent loss of elbow flexion that caused a limitation in the activities of daily living for at least six months after the injury despite nonoperative treatments. In one patient, surgical release was performed at four months after the injury because of progressive ulnar neuropathy. When radiographs and symptoms suggested the presence of heterotopic ossification, surgery was performed after symptoms such as pain, swelling, and tenderness had subsided and the heterotopic bone showed maturation radiographically, which was indicated by a smooth, well-demarcated cortical margin and definite trabecular markings.
The initial injuries consisted of an intra-articular distal humeral fracture in nine patients, dislocation associated with fracture in nine, olecranon fracture in eight, radial head fracture in four, isolated elbow dislocation in three, humeral shaft fracture in three, ulnar shaft fracture in two, local injuries without fracture in two, radial shaft fracture in one, and lateral epicondyle fracture in one. Initial treatment included open reduction and internal fixation of the fracture in twenty-four patients and cast or splint immobilization in eighteen. The duration of immobilization before elbow motion was allowed varied, ranging from five days to twenty weeks. Before we performed contracture release, six patients had had a total of seven surgical procedures subsequent to the initial treatment. These included osteosynthesis for nonunion in three patients, release due to elbow stiffness with or without excision of heterotopic bone in three patients, and reconstruction of the lateral ulnar collateral ligament in one patient. No patient had a history of infection.
Preoperative Assessment
The flexion and extension arc was measured with a use of a handheld goniometer positioned along the lateral aspect of the arm and forearm and centered on the lateral humeral condyle. Preoperatively, flexion averaged 89° (range, 70° to 95°) and the mean flexion contracture was 34° (range, 10° to 55°). Preoperative pronation averaged 66° (range, 10° to 80°), and supination averaged 77° (range, 10° to 90°).
Evidence of irritation of the ulnar nerve was assessed, and the status of ulnar neuropathy was graded according to the McGowan classification15. Nine of the forty-two patients had Grade-1 neuropathy, consisting of tingling and paresthesias without objective sensory or motor changes. Five patients had Grade-2 neuropathy, which is sensory disturbance and motor weakness without clawing. No patient showed Grade-3 neuropathy. We did not routinely perform electrophysiological studies. They were performed only when dysfunction of the ulnar nerve was suspected.
Biplanar radiographs were obtained for preoperative evaluation. We particularly checked whether heterotopic bone was visible on the radiographs. Computed tomography was not used to assess the heterotopic ossification. Heterotopic bone was clearly visible, with a smooth well-demarcated cortical margin and definite trabecular markings, in twenty-three patients. It was observed in the posterior aspect of the elbow around the olecranon in nineteen patients, in both the anterior and the posterior aspect in three patients, and in the anterior aspect in one patient. Heterotopic ossification was barely visible in eleven patients and not clearly seen in eight patients. All suspicious lesions were observed in the area just distal to the medial epicondyle on the anteroposterior radiograph (Fig. 1).
Operative Technique
Our goal was to achieve a maximum arc of motion with use of extra-articular procedures, which included release of contracted capsuloligamentous structures, excision of heterotopic bone, and lengthening of the triceps depending on the cause of the stiffness. Since no patient had intrinsic lesions limiting the full arc of motion, we considered the procedure successful only when >130° of flexion and <10° of extension loss could be demonstrated by passive motion with the force applied by the surgeon's two fingers while the patient was under anesthesia.
All patients underwent surgery in the supine position and with the use of a tourniquet. The skin incision was determined by several factors including the location of the previous incisions, the site of the heterotopic bone, and whether there was an ulnar neuropathy. A posterior skin incision was used in twenty patients; a single medial incision, in fourteen; both medial and lateral incisions, in seven; and both posterior and anterior incisions, in one.
Regardless of the skin incision used, the approach was started with the dissection of the triceps muscle from the posterior aspect of the humerus. The ulnar nerve was always identified proximally and traced distally. The posterior aspect of the capsule was separated from the triceps and excised while the ulnar nerve was protected. Heterotopic bone in the posterior aspect of the distal part of the humerus around the olecranon was excised along with the posterior aspect of the capsule. Passive flexion was attempted at this stage, and invariably there was a failure to regain full flexion. Therefore, the release was extended to the posteromedial aspect of the capsule between the medial side of the olecranon and the trochlea, which coincided with the posterior band of the medial collateral ligament.
The ulnar nerve was completely released if it was entrapped within fibrous tissue or heterotopic bone within the cubital tunnel. The released nerve was transposed anteriorly into subcutaneous tissues at the final stage of the operation. Complete release and anterior transposition of the ulnar nerve was also indicated in patients who had preoperative symptoms suggesting ulnar neuropathy around the elbow. Another indication for transposition was a roughened ulnar nerve bed after resection of the posterior band of the medial collateral ligament in order to avoid postoperative ulnar neuropathy due to adhesions. On the basis of these guidelines, we transposed the ulnar nerve in forty patients, including three who had anterior transposition of the ulnar nerve during a previous operation.
Resection of the posterior band of the medial collateral ligament was started posteriorly and carefully advanced anteriorly while the flexion arc was frequently checked until >130° of flexion could be achieved. The anterior band of the medial collateral ligament, which is the main valgus stabilizer of the elbow, was not released. If heterotopic bone was embedded in the posterior band, it was excised together with the fibrous capsule. When the heterotopic bone was located in the anterior band, no attempt was made to remove it if the arc of elbow motion was not affected. After posterior release and resection of the posterior band, >130° of flexion could be achieved in all patients except four who had heterotopic bone anteriorly that blocked maximum flexion.
Release of adhesions from between the radial head and the surrounding anular ligament was required in six patients who had limitation of forearm rotation secondary to a radial head fracture. The radial head was approached by making a separate lateral incision or elevating the lateral skin flap of the posterior incision. The Kocher interval between the extensor carpi ulnaris and the anconeus was developed, and the anular ligament was divided while the lateral ulnar collateral ligament was preserved.
After the posterior release was completed, an anterior approach was performed if there was a persistent flexion contracture of >10° or any anterior heterotopic ossifications were restricting flexion. We usually approached the anterior aspect of the elbow from the medial side, elevating the brachialis and the superior portion of the common flexor muscles from the anterior surface of the humerus. The anterior aspect of the capsule was fully exposed and carefully resected in thirty-six patients to decrease the flexion contracture. In four patients, heterotopic bone that blocked maximum flexion was found anteriorly. It was resected with careful protection of the neurovascular structures. This required a separate incision on the anterior side in addition to the posterior incision in one patient, and both medial and lateral incisions were used in three patients. We eventually achieved our flexion-extension arc goal in all patients when these procedures were completed.
Twenty-four additional procedures were performed at the time of the surgical release in twenty-one patients. Implants for fracture fixation were removed in thirteen patients. The anterior band of the medial collateral ligament was repaired in three elbows in which the ligament had ruptured from the humeral origin when passive manipulation was used to gain maximum flexion. The lateral collateral ligament was repaired in two patients because of incomplete transection during excision of heterotopic bone. In four patients, contracture of the triceps secondary to long-standing stiffness limited full flexion despite sufficient release of the elbow having been accomplished. In these cases, we performed a V-Y-type aponeurotic lengthening to gain more flexion. Two patients who had severely restricted forearm rotation required resection of the deformed radial head to gain rotational motion. Before wound closure, elbow stability was assessed to ensure that all patients could initiate range-of-motion exercises immediately. The pneumatic tourniquet was released, and careful hemostasis was obtained. The wound was closed over a suction drain.
Postoperative Management
A long arm splint was applied following the surgery. For patients with marked loss of flexion, splinting in the flexed position was preferred as long as the circulation of the skin was not compromised. However, when impairment of circulation was a concern, particularly in patients with a long posterior incision, we positioned the elbow in <90° of flexion.
A single 700-cGy dose of radiation was given to twenty-three patients within forty-eight hours after the operation as prophylaxis against recurrence of heterotopic ossification. Radiation was indicated when heterotopic bone was found during the operative procedures. The selection of patients for radiation was arbitrary and not randomized.
Physical therapy was started after two to four days. It consisted of active-assisted and gentle passive flexion and extension exercises of the elbow. Continuous passive motion was routinely used for the initial three to five days so that the patients could obtain rapid confidence with the exercise. We did not use a postoperative brachial plexus block. Forceful passive manipulation by the therapist was not allowed. A night splint in the position of maximum tolerable flexion was used by selected patients who showed resistance to an increase in flexion while the extension was easily restored, but use of this splint was discontinued two or three weeks after the surgery. A hinged-type brace was applied for the five patients who had repair of the medial collateral or lateral collateral ligament during the operation. A turnbuckle brace was used by two patients during the early phase of this study. Physical therapy continued until the arc of flexion-extension was not basically affected by the range-of-motion exercises. When elbow motion reached a plateau, patients felt that active maximum motion was possible without additional effort and persisted without continuing physical therapy. The duration of physical therapy varied widely among patients, ranging from two to six months with a mean of 3.7 months.
Evaluation
The follow-up period was a minimum of twenty-four months and a mean of 38.5 months (range, twenty-four to eighty-nine months), excluding one patient who was evaluated at a second operation that was performed at eight months after the first operation. Twenty-eight patients returned to our clinic for the evaluation. Fourteen patients who refused an evaluation were contacted by telephone. All of them had been discharged from our clinic after their arc of elbow motion plateaued. They stated that the arc of motion had not essentially changed from the time that they had been discharged, which ranged from six to twenty months after the operation with a mean of eleven months. We defined the arc of motion measured at discharge as the final motion and completed the evaluation by means of an interview.
The Mayo Elbow Performance Index (MEPI), which includes the evaluation of pain (maximum score, 45 points), motion (20 points), stability (10 points), and daily function (25 points)16, was used to assess total elbow function before the operation and at the time of final follow-up. Complications including remaining symptoms of ulnar neuropathy were carefully assessed. We observed whether there was newly formed heterotopic bone by comparing the anteroposterior and lateral radiographs made immediately after the operation with those made at the time of discharge from our clinic or final follow-up. We determined that heterotopic ossification had recurred radiographically after the surgical release when newly formed bone was clearly seen on the radiographs at the time of follow-up. A patient with radiographic evidence of recurrence was considered to have clinical recurrence of heterotopic ossification only when he or she had failed to regain a functional arc of motion.
Statistical Analysis
The Wilcoxon signed-rank test was used to assess improvements in the range of motion after the operation. We used univariate analysis to examine the relationship between final outcomes, as represented by the final flexion, total arc of motion, and MEPI score, and clinical variables (age, sex, type of initial injury, initial articular involvement, period from injury to operation, duration of initial immobilization, preoperative flexion and extension arc, and use of prophylactic radiation). The simple correlation test was performed with use of the Spearman rank correlation coefficient to determine the association between continuous variables and the final data, and the Mann-Whitney U test or the Kruskal-Wallis test was used to assess categorical variables. Results were considered significant when the p value was <0.05.
Source of Funding
There was no external funding source for this study.
Intraoperative Findings
During the procedure, the presence of heterotopic bone was determined through gross observation. The mature heterotopic bone that was seen on the preoperative radiographs was easily observed during surgery. When the mature bone was embedded within the excised capsular structures, we also defined it as heterotopic bone although it was not clearly seen on the radiographs. We observed heterotopic bone in forty patients (95%). The heterotopic bone that appeared responsible for mechanical limitation of motion was posterior and extended to the posteromedial aspect of the capsule in thirteen patients, was limited to the posteromedial aspect of the capsule in twenty-three patients, was in both the anterior and the posterior aspect of the elbow in three patients, and was in the anterior aspect only in one patient. Additional heterotopic bone was separately located in the posterior radiocapitellar area in five patients and was excised. All patients in whom the heterotopic ossification was barely visible on radiographs and six of the eight patients in whom heterotopic bone was unclear radiographically were found to have heterotopic bone during surgery. The heterotopic bone was identified with visual inspection and was not confirmed histologically. In these patients, the heterotopic ossification was always found in the posteromedial aspect of the capsule.
Flexion-Extension Arc
The mean flexion at the final follow-up evaluation was 124° (range, 90° to 140°), which had improved significantly from 89° preoperatively (p < 0.05). Twenty-one (50%) of the forty-two patients regained flexion of =130°, and thirty-seven patients (88%) regained flexion of =120°. The mean flexion contracture decreased significantly from 34° preoperatively to 9° (range, 0° to 30°) postoperatively (p < 0.05). The mean total arc of motion increased from 55° preoperatively to 115° postoperatively, with a mean improvement of 60° (range, 10° to 85°) (p < 0.05) (Table I). A functional arc of motion from 30° to 130° was achieved in twenty-one patients (50%), and a total arc of =100° was obtained in thirty-seven patients (88%). Mean forearm supination improved from 76° to 80°, and mean pronation improved from 67° to 73° (p < 0.05).
Statistical analysis showed that an increased period from the injury to the operation was related to decreases in final flexion and the total arc of motion, although the association was relatively weak (r = -0.37 for flexion, r = -0.39 for the total arc). When we compared the results of the patients who had undergone the operation at twelve months or earlier (n = 24) with those of the patients who had had the operation at more than twelve months after the injury (n = 18), the differences in the mean flexion (128° versus 119°) and the total arc (121° versus 106°) were significant (both, p < 0.05), indicating that the delay of more than one year adversely affected the final range of motion. With respect to initial articular involvement, patients who had an extra-articular injury achieved more flexion and a greater total arc than did those who had an intra-articular injury (mean flexion, 128° versus 122°; mean total arc, 121° versus 110°), and the difference in the total arc between the two groups was significant (p < 0.05). With the numbers studied, no significant association was found between the final range-of-motion data and other clinical variables including age, sex, type of initial injury, duration of initial immobilization, preoperative arc of motion, and prophylactic radiation.
MEPI
The mean MEPI score increased significantly from 73 points preoperatively to 94 points (range, 72 to 100 points) postoperatively (p < 0.05), with the result rated as excellent in thirty-two patients, good in eight, and fair in two. All patients had either no pain or mild intermittent pain. The fair results were related to a lack of improvement in motion and associated functional limitations in one patient and to a mild degree of pain, instability, and functional limitations in the other. Among the clinical variables, only the preoperative degree of flexion had a significant relationship with the MEPI score: less preoperative flexion was related to increased scores (p < 0.05, r = -0.34).
Complications
Ten patients (24%) showed radiographic evidence of recurrence of heterotopic ossification, but clinical recurrence associated with a failure to gain substantial motion was observed in only two patients. Prophylactic radiation was given to two patients with radiographic evidence of a recurrence, and to one patient with a clinical recurrence. In one patient, repeat resection of heterotopic bone was performed eight months after the first operation. After the second release, the heterotopic ossification did not recur and the final flexion increased from 95° to 115°. Another patient accepted the result and refused a second operation.
One patient developed an infection immediately after the contracture release, and it required debridement and drainage followed by systemic administration of antibiotics. The infection was successfully eradicated and did not appear to affect the final range of motion. Ulnar nerve symptoms persisted in nine patients at the time of final follow-up. Eight patients had McGowan Grade-1 symptoms without sensory or motor disturbances, and one had persistence of a previous McGowan Grade-2 neuropathy in spite of submuscular repositioning at the time of the surgical release. No patient showed worsening of the ulnar neuropathy after the operation. Two patients had valgus instability, which was secondary to rupture of the medial collateral ligament in one and to radial head resection in the other. One patient who had a repair of the lateral collateral ligament showed varus instability. The instabilities measured <10° when compared with the normal side and did not appear to affect daily function.