Good Functional Recovery of Complex Elbow Dislocations Treated With Hinged External Fixation: A Multicenter Prospective Study

Abstract

Background

After a complex dislocation, some elbows remain unstable after closed reduction or fracture treatment. Function after treatment with a hinged external fixator theoretically allows collateral ligaments to heal without surgical reconstruction. However, there is a lack of prospective studies that assess functional outcome, pain, and ROM.

Questions/purposes

We asked: (1) In complex elbow fracture-dislocations, does treatment with a hinged external fixator result in reduction of disability and pain, and in improvement in ROM, function, and quality of life? (2) Does delayed treatment (7 days or later) have a negative effect on ROM after 1 year? (3) What are the complications seen after external fixator treatment?

Methods

During a 2-year period, 11 centers recruited 27 patients 18 years or older who were included and evaluated at 2 and 6 weeks and at 3, 6, and 12 months after surgery as part of this prospective case series. During the study period, the participating centers agreed on general indications for use of the hinged external fixator, which included persistent instability after closed reduction alone or closed reduction combined with surgical treatment of associated fracture(s), when indicated. Functional outcome was evaluated using the Quick Disabilities of the Arm, Shoulder and Hand (QuickDASH; primary outcome) score, the Mayo Elbow Performance Index (MEPI), the Oxford Elbow Score, and the level of pain (VAS). ROM, adverse events, secondary interventions, and radiographs also were evaluated. A total of 26 of the 27 patients (96%) were available for followup at 1 year.

Results

All functional and pain scores improved. The median QuickDASH score decreased from 30 (25^th–75^th percentiles [P₂₅–P₇₅], 23–40) at 6 weeks to 7 (P₂₅–P₇₅, 2–12) at 1 year with a median difference of −25 (p < 0.001). The median MEPI score increased from 80 (P₂₅–P₇₅, 64–85) at 6 weeks to 100 (P₂₅–P₇₅, 85–100) at 1 year with a median difference of 15 (p < 0.001). The median Oxford Elbow Score increased from 60 (P₂₅–P₇₅, 44–68) at 6 weeks to 90 (P₂₅–P₇₅, 73–96) at 1 year with a median difference of 29 (p < 0.001). The median VAS decreased from 2.8 (P₂₅–P₇₅, 1.0–5.0) at 2 weeks to 0.5 (P₂₅–P₇₅, 0.0–1.9) at 1 year with a median difference of −2.1 (p = 0.001). ROM also improved. The median flexion-extension arc improved from 50° (P₂₅–P₇₅, 33°–80°) at 2 weeks to 118° (P₂₅–P₇₅, 105°–138°) at 1 year with a median difference of 63° (p < 0.001). Similarly, the median pronation-supination arc improved from 90° (P₂₅–P₇₅, 63°–124°) to 160° (P₂₅–P₇₅, 138°–170°) with a median difference of 75° (p < 0.001). At 1 year, the median residual deficit compared with the uninjured side was 30° (P₂₅–P₇₅, 5°–35°) for the flexion-extension arc, and 3° (P₂₅–P₇₅, 0°–25°) for the pronation-supination arc. Ten patients (37%) experienced a fixator-related complication, and seven patients required secondary surgery (26%). One patient reported recurrent instability.

Conclusions

A hinged external elbow fixator provides enough stability to start early mobilization after an acute complex elbow dislocation and residual instability. This was reflected in good functional outcome scores and only slight disability despite a relatively high complication rate.

Level of Evidence

Level IV, therapeutic study.

Introduction

Background

Complex elbow dislocations, with injuries to osseous and ligamentous structures, is an important cause of instability of the elbow [19]. The goal in management of complex elbow dislocations is to reconstruct a stable joint that tolerates a functional after-treatment [19, 28, 33, 38]. Elbows with residual instability frequently are treated with primary ligament repair with or without plaster immobilization. However, ligament repair has its disadvantages. Overtightening or malpositioning of the ligaments beyond the isometric point may contribute to stiffness and instability. Furthermore, ligament repair increases the risk of ulnar nerve injury and necessitates an extensive surgical approach [24, 26, 35]. Moreover, ligamentous repair may not be sufficient to stabilize the elbow in such a way that immediate active movement is tolerated [5, 28]. Plaster immobilization is unattractive, because earlier studies found that mobilization is essential during healing of injured ligaments because the functional load on the collagen fibers prevents contracture and the risk of stiffness [1, 11, 17, 20, 27, 29, 37]. Another alternative is the hinged external fixator, which stabilizes the elbow and protects the elbow against valgus and varus stress and allows flexion and extension. Theoretically, this will allow the ligaments to heal without additional reconstruction and without compromising a functional after-treatment.

Rationale

Previous reports on ROM and patient-reported outcome scores after the use of a hinged external fixator in these types of injuries show promising but varying results [16, 18, 31, 32, 34, 38, 39]. This is mostly because the majority of these studies were small retrospective case series. There is a lack of prospective studies regarding the use of a hinged external fixator in patients with instability after a complex elbow dislocation.

The aim of our study was to prospectively evaluate patients with acute complex elbow dislocations and residual instability who were treated with a hinged external elbow fixator and early mobilization in terms of (1) functional outcome; and (2) fixator-related adverse events.

Study Questions

We attempted to answer the following questions: (1) In complex elbow fracture-dislocations, does treatment with a hinged external fixator result in reduction of disability and pain, and in improvement in ROM, function, and quality of life? (2) Does delayed treatment (7 days or later) have a negative effect on ROM after 1 year? (3) What are the complications seen after external fixator treatment?

Patients and Methods

Study Design and Setting

This study was a prospective multicenter case series. Surgeons representing 15 hospitals participated. All surgeons were selected based on their clinical case experience with this type of injury and the hinged elbow fixator. We assessed patients for eligibility for this study between December 15, 2009, and December 13, 2011.

Participants/Study Subjects

During the study period, the participating centers agreed on general indications for use of the hinged external fixator, which included (1) residual elbow instability after open reduction and internal fixation of all associated fractures and/or radial head replacement, or (2) persisting postreduction elbow instability of dislocations that were accompanied by fractures that did not require fracture treatment. Inclusion criteria for the study were patients 18 years or older with a complex elbow dislocation who were treated with a hinged elbow fixator (Orthofix^® elbow fixator; Orthofix International, Bussolegno, Italy; FDA-approved since September 15, 1999) for instability after closed reduction alone or closed reduction combined with open treatment of associated fracture(s) when indicated. A complex elbow dislocation was defined as any type of elbow dislocation with fractures of the radial head, coronoid process, or proximal ulna (olecranon). Residual instability was defined as spontaneous redislocation of the joint, or as redislocation during flexion and extension or the pivot shift test. Valgus or varus laxity without (sub)dislocation was not defined as residual instability and was not considered an indication for fixator placement. These tests were performed in the operating room directly after surgery. For patients who had closed reduction alone, spontaneous redislocation was used as an indication for fixator placement [25]. Exclusion criteria were pathologic fractures, preexistent injuries of the affected arm, collateral ligament repair, a fracture of the ipsilateral distal humerus, and additional traumatic injuries to the affected arm (ie, ipsilateral distal radius fracture). Patients with insufficient understanding of the Dutch language or patients for whom problems in maintaining followup were expected also were excluded. All patients gave written informed consent to participate in this study, which was approved by the medical research ethics committees of all participating hospitals. The study protocol was published elsewhere [36].

During the study period, 42 patients experienced a complex elbow dislocation and were screened for eligibility. Fifteen patients were excluded: seven patients had a stable elbow after open reduction and internal fixation, four had additional injuries to the ipsilateral arm, two had a fracture of the proximal humerus, one had only a subluxation of the radial head, and one did not consent to participate (Fig. 1). Twenty-seven patients from 11 hospitals were included (Table 1). Those 27 patients were treated by 14 different surgeons. Eight surgeons treated only one patient, three surgeons treated two patients, one surgeon treated three patients, one surgeon treated four patients, and one surgeon treated six patients. The majority of the patients were female (52%) with a median age of 52 years (25^th–75^th percentiles [P₂₅–P₇₅], 38–59). All but one patient completed 1 year followup. This patient died of a nonsurgery-related accident and completed only 6 weeks of followup.

Fig. 1.

The study flowchart is shown. ORIF = open reduction and internal fixation.

Table 1.

Baseline characteristics

Characteristic	n = 27
Female†	14 (52%)
Age (years)*	52 (38–59)
BMI (kg/m2)*	26 (23–28)
ASA score†
1	19 (70%)
2	7 (26%)
3	1 (4%)
Tobacco use†	7 (26%)
Alcohol use†	19 (70%)
Injury to dominant arm†	13 (48%)
Type of dislocation†
Posterior	14 (52%)
Posterolateral	10 (37%)
Lateral	1 (4%)
Unknownǂ	2 (7%)
Associated fractures†
Radial head	9 (33%)
Radial head + coronoid process	9 (33%)
Coronoid process	6 (22%)
Radial head + coronoid process + olecranon	1 (4%)
Radial head + olecranon	1 (4%)
Coronoid process + olecranon	1 (4%)
Radial head fractures† [14]	20 (74%)
Mason I	2 (10%)
Mason II	5 (25%)
Mason III	13 (65%)
Coronoid process fractures† [30]	17 (63%)
Regan and Morrey I	11 (65%)
Regan and Morrey II	5 (29%)
Regan and Morrey III	1 (6%)
Olecranon fractures† [21]	3 (11%)
Mayo IIIA	1 (6%)
Mayo IIIB	1 (6%)
Monteggia fracture	1 (6%)
Operative fracture treatment†	19 (70%)

Data provided as *median with the first and third quartiles in brackets or as ^†patient numbers with the percentage in parentheses; ^ǂtype of dislocation unknown because of prehospital reduction or absence of prereduction radiographic images when reduction occurred at another hospital; ASA = American Society of Anesthesiologists.

Fracture Characteristics

Nine patients (33%) presented with a terrible triad injury defined as an elbow dislocation accompanied by fractures of the radial head and coronoid process (Table 2). Nine patients (33%) had an isolated fracture of the radial head. In six patients (22%), the dislocation was accompanied by an isolated fracture of the coronoid process. One patient (4%) had combined fractures of the coronoid process and olecranon, one patient (4%) had combined fractures of the radial head and olecranon, and one (4%) sustained fractures of the radial head, coronoid process, and olecranon. In 20 patients (74%) at least one of the fractures required open treatment, and seven patients underwent only closed reduction before hinged external fixation. Time to surgery was a median of 6 days (P₂₅–P₇₅, 1–10).

Table 2.

Characteristics of the 27 patients in order of inclusion

Patient number*	Sex	Age (years)	Trauma mechanism	Fracture (classification)	ORIF	Radial head prosthesis	Adverse events	Secondary surgery	1-year QuickDASH	1-year OES	1-year FE arc (degrees)	1-year PS arc (degrees)
1	M	39	LET	R (II)	R	–	Wound infection Limited ROM	None Arthrolysis	4.6	90	65	180
2	F	60	LET	R (III)	–	+	None	None	0.0	100	115	180
3	M	56	LET	C (II)	C	–	Ulnaropathy	Ulnar nerve release	6.8	92	115	180
4	F	50	HET	R (I), O (Monteggia)	O	–	None	None	11.4	60	120	180
5	M	35	LET	R (III), C (I)	R	–	None	None	2.3	98	130	180
6	F	65	LET	R (II), C (II)	–	–	Incongruent joint	Replacement HEF	29.6	73	120	165
7	M	41	HET	R (II), C (I)	R	–	Pin-tract infection Joint crepitus	None ROH (radius)	2.3	96	135	160
8	M	54	LET	R (III), C (II), O (IIIB)	C,O	+	Late infection	ROH (olecranon + radius)	9.1	88	115	160
9	F	53	LET	R (III), C (I)	R	–	Incongruent joint Pin-tract infection Limited ROM	Replacement HEF None Arthroplasty	22.7	67	80	160
10	F	64	LET	R (II)	–	+	None	None	0.0	98	150	170
11	M	30	HET	C (II)	–	–	HEF malfunction Persistent instability	Replacement HEF LCL reconstruction	20.5	58	115	170
12	F	57	LET	R (III), C (I)	–	+	None	None	6.8	92	125	160
13	F	66	HET	R (III)	R	–	Incongruent joint Pin-tract fracture ulna	Replacement HEF None	11.4	77	60	170
14	F	54	LET	C (II), O (IIIA)	O	–	None	None	0.0	88	120	170
15	M	31	LET	R (I), C (I)	C	–	Septic arthritis Pin-tract fracture humerus	Debridement Plate humerus	11.4	81	105	155
16	M	43	LET	C (III)	–	–	Pin-tract infection	None	4.6	81	125	165
17	F	64	LET	R (III), C (I)	R	–	None	None	NA	NA	NA	NA
18	F	80	LET	R (II), C (I)	R	–	Incongruent joint	Replacement HEF	6.8	92	105	145
19	M	57	LET	R (III)	R	–	None	None	13.6	90	145	170
20	M	26	HET	R (III)	–	–	Incongruent joint (3)	Replacement HEF (3)	18.2	71	110	170
21	F	47	HET	R (III)	R	–	None	None	2.3	96	110	170
22	F	48	LET	R (III), C (I)	R	–	Pain	ROH (radius)	9.1	98	145	165
23	F	29	LET	R (III)	R	–	None	None	0.0	94	145	160
24	M	37	HET	C (I)	–	–	Pin-tract infection	None	6.8	69	80	160
25	M	24	LET	R (III)	R	–	None	None	9.1	90	100	170
26	F	52	LET	C (I)	–	–	None	None	15.9	73	150	155
27	M	49	HET	C (I)	–	–	None	None	0.0	98	150	170

* Patients 6, 11, 16, 20, 24, 26, and 27 were treated conservatively; ORIF = open reduction internal fixation; OES = Oxford Elbow Score; FE = flexion-extension; PS = pronation-supination;; LET = low-energy trauma; HET = high-energy trauma; R = radial head (Mason classification [14]); C = coronoid process (Regan and Morrey classification [30]); O = olecranon (Mayo classification [21]); ROH = removal of hardware; HEF = hinged external fixator; LCL = lateral collateral ligament.

Surgical Procedure

If instability was present after fracture treatment, a hinged external fixator was mounted. With the elbow in 90° flexion, the central axis of rotation was located by overlapping the capitellum and trochlea in a lateral fluoroscopic image. Perfect overlap of these structures resulted in a circle with the center of this circle representing the axis of rotation. Along the axis of rotation, a 2-mm K-wire was inserted. Its position was confirmed on the AP and lateral planes (Fig. 2). The central connecting unit of the external fixator then was applied over the K-wire. The lateral aspect of the humerus was exposed by an approximately 4-cm incision just distal to the insertion of the deltoid muscle taking the radial nerve into account. The humeral screws were inserted and the clamp cover was tightened. Subsequently, the ulnar screws were drilled laterally through a 4-cm incision. After tightening this clamp, the image intensifier was used to check reduction and congruency of the joint and alignment of the fixator. Flexion and extension were required to go smoothly without compromising congruency during movement. A good indicator for perfect alignment was the K-wire, which had to have no resistance in the center of the connecting unit during motion of the elbow. Furthermore, no widening of the joint space was accepted during flexion and extension on AP and lateral view radiographs. Finally, the link-locking screws were tightened, the K-wire removed, and the wounds on the upper arm and forearm approximated.

Fig. 2A–B.

(A) Locating the center of rotation is done by overlapping the trochlea and capitellum of the humerus projecting them as a perfect circle. The center of this circle is considered the axis of rotation. (B) The depth of the Kirschner wire is checked in AP view. Care should be taken not to drill too deep to avoid harming the ulnar nerve.

Aftercare

A protocol of supervised active and passive extension, flexion, and pronation and supination exercises was started immediately after surgery if tolerated (Fig. 3) [36]. After 6 weeks, the external fixator was removed in the outpatient department without any form of anesthesia. All patients received 50 mg indomethacin twice daily for 6 weeks as heterotopic ossification prophylaxis, unless NSAIDs were contraindicated. A proton pump inhibitor also was administered.

Fig. 3A–B.

A study patient is shown with his arm in (A) full flexion and (B) in extension immediately after surgery.

Outcome Assessment and Data Collection

Followup data were collected at 2 weeks, 6 weeks, 3 months, 6 months, and 12 months after surgery. Standard radiographs of the elbow were made at the time of admission, within 48 hours after surgery, and at each followup.

Variables, Outcome Measures, Data Sources, and Bias

The primary outcome was the Quick-Disabilities of the Arm, Shoulder and Hand (QuickDASH) scores after 1 year, reflecting functional outcome and pain [2, 12]. Secondary outcome measures were level of pain measured with a VAS, the Mayo Elbow Performance Index (MEPI) [22], the injury-related quality of life measured with the Oxford Elbow Score [4, 7], and health-related quality of life measured with the SF-36 [40]. Scores for the SF-36 physical and mental component summaries were converted to a norm-based score and compared with the norms for the general population of the United States [40]. Permission for translation and use of the Oxford Elbow Score for this study was obtained from Oxford and Isis Outcomes, part of Isis Innovation Limited (http://www.isis-innovation.com/). In addition, ROM was measured using a goniometer. All physical examinations were performed by an investigator or research assistant from the principal site in the presence of the treating surgeon. Complications and secondary interventions were recorded. Radiographs were evaluated by two observers independently (GITI, DDH) for type of dislocation, type of fractures, joint congruency, fracture consolidation, and the presence of heterotopic ossifications. Radial head fractures were classified using the Mason classification [14]. Fractures of the coronoid process were classified according to the Regan and Morrey classification [30]. Fractures of the olecranon were classified according to the Mayo classification [21]. Fractures were considered healed if one of the following three criteria was met: (1) bridging of fracture by callus/bone trabeculae or osseous bone; (2) obliteration of fracture line/cortical continuity; or (3) bridging of fracture at three cortices. Heterotopic ossifications were classified as present if there were immature calcifications, small mature ossifications, large mature ossifications, or complete bone bridging/ankylosis [9]. Radiographic results showed that 18 (78%) patients showed radiographic healing and that 13 (57%) patients showed signs of heterotopic ossifications at 1 year (Table 3).

Table 3.

Radiographic results at 1 year

Characteristic	n = 23
Fracture consolidation
Yes	18 (78%)
No	3 (13%)
NA	2 (9%)
Heterotopic ossifications
Yes	13 (57%)
No	10 (43%)

Radiographs at 1 year were not available for four patients. None of the patients showed radiographic signs of instability at 1 year; NA = not applicable (these patients had a radial head fracture treated with radial head replacement).

Statistical Analysis

Data were analyzed using SPSS Version 20.0 (Chicago, IL, USA). Normality of continuous data was tested with the Shapiro-Wilk test and by inspecting frequency histograms (Q-Q plots). Descriptive analysis was performed to describe baseline characteristics (intrinsic, injury, and intervention-related variables) and outcome measures. Continuous data are reported as medians and percentiles (nonparametric data) or as means and SD (parametric data) and categorical data as numbers with percentages. A Wilcoxon signed rank test was used to compare functional outcome scores at 1 year with those at the first followup measurement (ie, 2 weeks for ROM and 6 weeks for the QuickDASH, Oxford Elbow Score, MEPI, and SF-36). A Mann-Whitney U test was performed to assess statistical significance of difference in ROM between patients who received early treatment (ie, within 7 days after initial injury) and those who received delayed treatment (ie, 7 days or later after initial treatment). A p value less than 0.05 was considered the level of statistical significance.

Results

Patient-reported Pain, Functional Outcome, and Quality of Life

All outcome measures except for the SF-36 Mental Component Summary improved after the initial assessment (Fig. 4). The median QuickDASH score decreased from 30 (P₂₅–P₇₅, 23–40) at 6 weeks to 7 (P₂₅–P₇₅, 2–12) at 1 year with a median difference of −25 (p < 0.001). The median level of pain (VAS) decreased from 2.8 (P₂₅–P₇₅, 1.0–5.0) at 2 weeks to 0.5 (P₂₅–P₇₅, 0.0–1.9) at 1 year with a median difference of −2.1 (p < 0.001). The median MEPI increased from 80 (P₂₅–P₇₅, 64–85) at 6 weeks to 100 (P₂₅–P₇₅, 85–100) at 1 year with a median difference of 15 (p < 0.001). The median Oxford Elbow Score increased from 60 (P₂₅–P₇₅, 44–68) at 6 weeks to 90 (P₂₅–P₇₅, 73–96) at 1 year with a median difference of 29 (p < 0.001). The median SF-36 Physical Component Summary increased from 40 (P₂₅–P₇₅, 36–42) at 6 weeks to 52 (P₂₅–P₇₅, 47–55) at 1 year, with a median difference of 14 (p < 0.001).The SF-36 Mental Component Summary, however, remained similar (6-week median 58 [P₂₅–P₇₅, 46–61], 1-year median, 56 [P₂₅–P₇₅, 51–60], median difference, −2; p = 0.784).

Fig. 4A–F.

Changes during followup in the (A) QuickDASH score, (B) pain, (C) Mayo Elbow Performance Index (MEPI), (D) Oxford Elbow Score (OES), (E) SF-36 Physical Component Summary (PCS) score, and (F) SF-36 Mental Component Summary (MCS) are shown. The dotted lines in the (E-F) SF-36 physical and mental component summaries represent the US population norm of 50 ± 10 (SD) points. All outcome scores except for the SF-36 Mental Component Summary show improvement with time.

ROM

ROM for flexion-extension and pronation-supination arcs improved during followup (Fig. 5). The median flexion-extension arc improved from 50° (P₂₅–P₇₅, 33°–80°) at 2 weeks to 118° (P₂₅–P₇₅, 105°–138°) at 1 year, with a median difference of 63° (p < 0.001). The median flexion improved from 100° (P₂₅–P₇₅, 90°–110°) to 140° (P₂₅–P₇₅, 129°–145°), with a median difference of 33° (p < 0.001) and the median extension improved from 40° (P₂₅–P₇₅, 30°–60°) to 20° (P₂₅–P₇₅, 0°–26°), with a median difference of −30° (p < 0.001). Similarly, the median pronation-supination arc improved from 90° (P₂₅–P₇₅, 63°–124°) to 160° (P₂₅–P₇₅, 138°–170°), with a median difference of 75° (p < 0.001). The median pronation improved from 55° (P₂₅–P₇₅, 33°–85°) to 83° (P₂₅–P₇₅, 75°–85°), with a median difference of 15° (p = 0.001) and the median supination improved from 30° (P₂₅–P₇₅, 20°–45°) to 80° (P₂₅–P₇₅, 68°–85°), with a median difference of 45° (p < 0.001). At 1 year, the residual deficits compared with the uninjured side were 30° (P₂₅–P₇₅, 5°–35°) for the flexion-extension arc and 3° (P₂₅–P₇₅, 0°–25°) for the pronation-supination arc.

Fig. 5A–B.

Changes in (A) arcs of flexion-extension and (B) pronation-supination during followup are shown The dotted lines represent functional elbow ROM on positional and functional tasks as reported by Morrey et al. [23]. ROM shows improvement with time.

The study population was divided into a group that was treated within 7 days after initial injury (early treatment, n = 14) and a group that was treated 7 days or later after initial injury (delayed treatment, n = 13). There was a 15° difference in the arc of flexion and extension favoring the early treatment group after 1 year: 128° (P₂₅–P₇₅, 114–145) versus 113° (P₂₅–P₇₅, 80–119), respectively (p = 0.02). This difference was attributable mainly to the greater extension deficit in the late treatment group: 8° (P₂₅–P₇₅, 0–25) for the early treatment group versus 25° (P₂₅–P₇₅, 13–30) for the late treatment group (p = 0.03).

Fixator-related Complications

Ten patients (37%) experienced 12 fixator-related complications, requiring secondary intervention in seven patients (26%) (Table 2). Five patients (19%) had elbow incongruency resulting from fixator malalignment. In all patients incongruency was recognized between 5 and 25 days after fixator placement. In these five patients, seven procedures for fixator replacement were required, all of which occurred on the same day or the first day after incongruency was recognized. One patient experienced a hardware defect which required fixator replacement. Four patients (15%) had a pin-tract infection, of whom two were treated with oral antibiotics alone. The other two patients required débridement in the outpatient clinic combined with antibiotic treatment. One patient had a pin-tract fracture of the ulna that was treated conservatively (leaving the fixator in situ) and one patient had a pin-tract fracture of the humerus 5 months after removal of the fixator, requiring plate fixation. No redislocations occurred after removal of the fixator; however, one patient had chronic posterolateral rotatory elbow instability and required a lateral collateral ligament reconstruction.

Discussion

Residual instability after a complex elbow dislocation is a serious condition with potentially life-changing sequelae and its treatment poses a challenge, even for experienced surgeons. The goal in management of complex elbow dislocations is to reconstruct a stable joint that tolerates functional after-treatment [19, 28, 33, 38]. A hinged fixator may be used to achieve this. Previous studies reported promising but variable results regarding ROM and patient-reported outcome scores [16, 18, 31, 33, 37, 38]. The variability in reported results may have been a function of the shortcomings of retrospective analysis. Because most of the studies on this topic have been small and retrospective, had inconsistent surgical indications, had substantial loss to followup, and used inconsistent approaches to measure of outcomes, those studies are difficult to evaluate. Therefore, we aimed to assess patients with complex elbow dislocations who were treated with a hinged external elbow fixator and early mobilization, prospectively and in a consistent fashion in terms of (1) functional outcome; (2) ROM; and (3) fixator-related adverse events.

This study had some limitations. First, the sample size was small in relation to the number of participating centers, but reasonable given that complex elbow dislocations with residual instability after fracture treatment are an uncommon problem. To the best of our knowledge, this is the first prospective study of this size with a highly structured followup design regarding complex elbow dislocations. The sample size did not allow analysis of a possible effect of fracture types on the patient-reported and clinical outcome measures. Second, because inter- and intraobserver reliability of testing elbow stability is unknown, the decision to use external hinged fixation was and will remain arbitrary. Likewise, some Mason Type II or III [14] fractures are treated with radial head prostheses, whereas others are treated with open reduction and internal fixation or a nonoperative approach. With the medial collateral ligament disrupted after most elbow dislocations, the radial head acts as the primary buttress against valgus stress. One can imagine the importance of stable fracture fixation or radial head replacement on elbow stability in these patients. It is not unlikely that the heterogeneous approach to radial head fractures could have contributed to a difference in outcomes among our patients. A similar discussion accounts for fractures of the coronoid process.

Third, 1 year of followup might not have been long enough to know the final patient-reported outcome measures and ROM, because the trends of the QuickDASH, Oxford Elbow Score, MEPI, and ROM all suggested additional improvement at 1 year. However, the role of osteoarthritis on the long-term outcome is unknown.

Finally, the different hospitals and surgeons in the current study, rather than the experience of one surgeon, might have been sources of multiple confounding factors, but this emphasizes the generalizability of our results.

Our series show very little disability after external fixator treatment of complex elbow fracture-dislocations. At 2 years, the median QuickDASH score of 7 is consistently lower than QuickDASH scores reported in previous articles on similar types of injuries treated with a hinged fixator (P₂₅–P₇₅, 15–28 points) [8, 13, 32, 34] or treated with ligament repair (P₂₅–P₇₅, 15–28 points) [6, 8, 10, 42]. The slight disability is paralleled by high scores on the additional patient-reported functional outcome measures. In the current study, most patients reported the maximum score (100 points) for the MEPI. MEPI scores in previous studies of patients with complex elbow injuries range between 75 and 93 [6, 11, 28, 32, 38, 41]. The flexion-extension arc result was better than expected. The 118° flexion-extension arc in our patients was in line with those reported for patients treated with ligament repair (P₂₅–P₇₅, 112°–117°) [6, 8, 10, 42], but consistently greater than for patients treated with a hinged external fixator (P₂₅–P₇₅, 93°–99°) [31, 32, 38, 41]. The latter could be because of the mean time to fixator placement. The most important differences in treatment between the current and previous studies were the use of early active mobilization, no collateral ligaments were reconstructed, and the short interval between trauma and surgery. However although it is likely that a combination of these factors played a role, their individual merit could not be extracted from current data.

Delay in treatment could be an important explanation for the superior results in our patients. The mean time to fixator placement reported in previous studies [31, 32, 38, 41] was between 26 days and 2 months versus 6 days in the current study. Although our study was not designed to define the window of opportunity for surgery, its results emphasize the importance of early reestablishment of a concentric and stable joint, which allows early movement. This is in concordance with Ruch and Triepel [34] who reported flexion-extension arcs of 120° and 84° in patients who underwent early versus delayed treatment with a hinged external fixator, respectively.

The fixator-related complication rate was relatively high in our study. The most frequent complication (five patients), which always resulted in fixator replacement, was joint incongruency. All other complications (pin-tract infection, pin-tract fractures, and redislocation) also have been reported at similar rates [3, 5, 31]. All surgeons had applied hinged external elbow fixators and attended a compulsory technique-oriented hands-on course before this study. Nevertheless, the most logical explanation for the high complication rate is underexposure of the surgeons to the procedure. This fuels the debate whether hinged elbow fixators should be used only by experienced surgeons. One patient reported moderate instability when evaluating the MEPI (this is the same patient who was treated with a lateral collateral ligament repair). No true recurrent dislocation was seen during the complete followup in any of the patients. This suggests that surgical repair of the collateral ligaments is not indicated as a standard procedure for adequate healing of the injured collateral ligaments. From experience with ligamentous injuries to the knee and ankle, it is known that ligaments have the ability to heal and to form a scar-like neoligament. Nevertheless, few data are available supporting a nonoperative approach to ligamentous injuries after complex dislocations of the elbow [8, 13, 15, 16].

This study confirmed that the hinged external elbow fixator provides enough stability to start early mobilization in patients with closed reduction or open treatment after an acute complex elbow dislocation with residual instability. This was reflected in good functional outcome scores and only slight disability despite a relatively high complication rate.