Predictors of reoperation after internal fixation of intra-articular distal humerus fractures

Abstract

Background

Despite good reported outcomes with open reduction and internal fixation of intra-articular distal humerus fractures, complication rates remain high. The objective of this work is to identify factors associated with reoperation.

Methods

Sixty-three patients treated with open reduction and internal fixation for intra-articular (Arbeitsgemeinschaft für Osteosynthesefragen type C) distal humerus fractures between 2004 and 2010 were identified using an institutional trauma registry, 62 of which were followed for a minimum of six months. Age, gender, fracture subclassification, open fracture presence, Injury Severity Score, time to definitive surgery, length of postoperative immobilization, and type of approach were recorded. Multivariate analysis was utilized to identify factors independently associated with reoperation.

Results

Complications requiring reoperation developed in 25 (40.3%) elbows. The most common reasons were wound dehiscence or infection in nine elbows (14.5%) and symptomatic hardware in six (9.6%). During multivariate analysis, only olecranon osteotomy remained an independent predictor for reoperation (P = 0.043).

Discussion

Despite improved internal fixation techniques, a high proportion of elbows require reoperation after open reduction and internal fixation for distal humerus fractures. Higher complication rates in fractures fixed through an olecranon osteotomy may reflect additional reoperations due to nonunion of the osteotomy or need to remove hardware from the ulna.

Level of evidence

Prognostic Level III.

Introduction

Open reduction and internal fixation (ORIF) of intra-articular distal humerus fractures is a technically complex procedure with many challenges, including difficulty obtaining adequate exposure of the injured joint surface, presence of comminution, and limited bone stock for stable fixation. Despite this, reported results in the last decade have shown good results with regard to union rates, range of motion, and functional outcome scores at mid- to long-term follow-up.^1–4 However, reoperation rates have been high in multiple case series, particularly as it relates to infection, stiffness, and need for hardware removal.^1,2,5–12

With the increasingly widespread utilization of total elbow arthroplasty for distal humerus fractures, the need to decide between fracture fixation and joint replacement has become a relatively common clinical scenario for elderly patients with complex fractures. To this end, data that can help identify prognostic factors for success or failure of these treatment alternatives and aid in decision-making are of interest to patients and practicing elbow surgeons.¹³

The objective of this study is to analyze the reoperation rate in a consecutive series of distal humerus fractures treated with ORIF and to quantify the relative contribution of patient (age and gender), injury (fracture subclassification, existence of open fracture, and Injury Severity Score (ISS)), and treatment (time to definitive surgery, length of postoperative immobilization, and type of approach) factors to the prognosis for reoperation.

Methods

Our institutional Trauma Registry was searched to identify patients with intra-articular distal humerus fractures treated with ORIF between 2004 and 2010. Fractures were limited to Arbeitsgemeinschaft für Osteosynthesefragen (AO) type C ¹⁴ distal humerus fractures only, resulting in inclusion of only complete articular fractures to ensure homogeneous fracture severity. Patients were excluded if they had not reached skeletal maturity or if concurrent proximal radial or ulnar fractures were present.

The query identified 63 elbow fractures in 63 patients. One patient was lost to follow-up prior to six months and was excluded; the remaining 62 patients were included for analysis. Initial consult notes, operative reports, and follow-up notes were reviewed in detail for demographics, surgery details, and complications. Range of motion measurements at latest follow-up were recorded. Patient radiographs were reviewed for evidence of nonunion and other radiographic findings at latest follow-up. Patients who did not have clinical and radiographic evidence of fracture healing at six months were considered to have a nonunion.

Patients

Of the 62 patients meeting inclusion criteria, 34 patients were female and 28 were male. Follow-up time ranged from 6 to 72 months (mean, 17 months). The mean age was 59 years old (range, 19–91). Patients presented with a mean ISS ¹⁵ of 10.4 (range, 4–34). Seventeen patients presented with open fractures, of which seven were type III, seven were type II, and three were type I based on the classification of Gustilo and Anderson. ¹⁶ Fourteen of these open fractures underwent an initial debridement prior to definitive fixation. Patients treated with an initial debridement were routinely immobilized in a plaster splint until definitive fixation. Time from injury to definitive fixation was a median of three days (range, 0–15 days).

According to the AO classification system, 15 fractures were considered type C1 (simple complete articular), 16 were type C2 (metaphyseal comminution), and 31 were type C3 (multifragmentary). Determination of the classification was made by the attending surgeon at the time of surgery.

Surgical technique

All patients were treated by one of seven surgeons with a minimum of five years of experience treating disorders of the upper limb. A midline posterior approach was used for all patients. After exposure and protection of the ulnar nerve, the fractures were reduced and fixed using an anatomic precontoured plating system. Forty-three of 62 (63%) patients underwent olecranon osteotomy that was subsequently fixed with a plate and screw construct in 25, a screw-and-wire construct in 13, and a tension band wire in 5. The surgeons involved in definitive fixation had different criteria for performing an olecranon osteotomy. Some surgeons performed an osteotomy in all intra-articular cases. Others had a selective approach, with the decision to proceed to olecranon osteotomy based on visualization of the intra-articular fracture fragments and ability to obtain reduction. The majority of osteotomies were performed as a chevron osteotomy, although two surgeons routinely used a transverse osteotomy. No olecranon osteotomy was performed in 19 patients. Eleven of these fractures were fixed through a triceps window or bilaterotricipital approach as described by Alonso-Llames, ¹⁷ five were fixed through a triceps-reflecting anconeus pedicle approach, ¹⁸ and three were fixed using a triceps tongue approach. Surgeons used a range of plating systems for definitive fixation. Plate configuration (90° or 180°) was determined by surgeon preference.

Surgeons followed differing postoperative protocols, but the majority performed a short course of immobilization. The majority of surgeons placed an anterior splint in extension, while one surgeon preferred a posterior-based splint. Patients were immobilized for a median of two days (range, 0–14 days). Only 11 of 62 patients were immobilized for seven or more days. None were immobilized for greater than 14 days. After immobilization was removed, patients initiated an active-assisted range of motion protocol.

Statistical analysis

Patient and surgical factors potentially associated with reoperation (age, gender, fracture classification, existence of open fracture, ISS, time to definitive surgery, length of postoperative immobilization, type of approach, and type of olecranon fixation for osteotomy approaches) were initially assessed using univariate analysis. The association between reoperation rate and categorical variables was compared using a chi-square test, and logistic regression was performed for continuous variables. Clinical outcome measures were compared between elbows that underwent revision surgery and those that did not, using a nonparametric Mann–Whitney U test. Statistical significance was presumed for P values <0.05.

Following this, variables were selected for multivariable analysis based on results of univariate testing (P value of <0.2). Presence of a comminuted type C3 fracture was also included to ensure that fracture severity was taken into account. A linear regression model with a nominal logistic personality was used. The final model was evaluated for area under the receiver operating characteristic curve.

Results

Reoperation rate

Acute or delayed complications requiring reoperation developed in 25 (40.3%) of 62 elbows. These included wound dehiscence or infection in nine elbows (14.5%), symptomatic hardware in six elbows (9.6%), heterotopic ossification (HO) in four elbows (6.5%), distal humerus nonunion in three elbows (4.8%), olecranon osteotomy nonunion in two elbows (3.2%), contracture in one elbow (1.6%), and ulnar neuropathy in one elbow (1.6%). Of the olecranon osteotomy nonunions, one of these was initially repaired with a tension band wire alone and one was repaired with a screw plus tension band wire. One patient underwent reoperation due to both distal humerus and olecranon osteotomy nonunion.

Patient and injury factors associated with reoperation

Increasing patient age was associated with a lower rate of revision surgery (P = 0.031) (Figure 1). The ISS was also directly correlated with greater reoperation rate (P = 0.038). With the numbers available, presence of an AO type C3 comminuted fracture was not correlated with a higher rate of reoperation (P = 0.796). In addition, gender (P = 0.501), history of surgery prior to definitive fixation (P = 0.618), and presence of any open fracture (P = 0.864) did not show a correlation with reoperation rate. There was a high rate of reoperation among the eight patients with a type 3 open fracture (five of eight, 62.5%); however, this did not reach statistical significance (P = 0.175).

Figure 1.

Reoperation rate by patient age.

Approach type and reoperation rate

Univariate analysis revealed that elbows treated with an osteotomy approach were more likely to be reoperated (21 of 43, 48.8%) compared to those treated with other approaches (4 of 19, 21%; P = 0.0345) (Figure 2). The four elbows without a prior osteotomy that underwent additional surgery were reoperated for nonunion (2), ulnar neuropathy (1), and HO formation (1). Among patients with an osteotomy, 14 of 25 (56%) fixed with a plate and screws were reoperated, of which 11 were for wound complications or symptomatic hardware. Seven of 18 osteotomies (39%) fixed with a tension band or screw-and-wire construct were reoperated, of which four were for wound complications or symptomatic hardware.

Figure 3.

Computed tomography scan of an intra-articular distal humerus fracture.

Figure 4.

Intraoperative photograph showing visualization of the joint surface achieved without olecranon osteotomy.

Figure 5.

Postoperative radiographs after fixation of an intra-articular distal humerus fracture.

Figure 2.

Reoperation rate by approach.

Other treatment factors and reoperation rate

Intraoperative usage of bone graft (0.493) and ulnar nerve transposition at the time of definitive fixation (P = 0.977) did not show a correlation with reoperation rate. Patients treated with definitive surgery at less than three days from the time of injury did not show a statistically significant difference in reoperation rate (P = 0.107). Postoperative immobilization of seven days or greater also did not correlate with a difference in reoperation rate (P = 0.703). Reoperation rates were not significantly associated with the surgeon of record.

Multivariate analysis

Factors included in the multivariate model based on a P value <0.2 included ISS, age, presence of a type 3 open fracture, time to surgery of less than three days, and use of an olecranon osteotomy. Presence of a comminuted type C3 fracture was also included to ensure that fracture severity was accounted for. In this multivariate analysis, only use of an olecranon osteotomy remained an independent predictor for reoperation (P = 0.043) (Table 1). Testing of the whole model showed the area under the curve to be 0.753, indicating adequate discriminatory power.

Table 1.

Multivariate analysis of predictors for reoperation.

Variable	Odds ratio	Confidence interval	P value
Injury Severity Score a	1.92	0.64–6.95	0.259
Age a	0.90	0.63–1.29	0.567
Type 3 open fracture	2.72	0.39–21.53	0.309
Time to surgery ≥3 days	2.96	0.91–10.67	0.072
Type C3 fracture	1.21	0.37–4.06	0.756
Olecranon osteotomy	3.88	1.04–18.01	0.043

^aOdds ratio is given per 10-unit change in regressor.

Clinical outcomes

Range of motion at latest follow-up is shown in Table 2. No differences were seen in final range of motion when comparing patients that had been reoperated to those who had not.

Table 2.

Reoperation and final range of motion following type C distal humerus open reduction internal fixation.

Outcome measure	Reoperated patients	Non-reoperated patients	P value	All patients
Flexion–extension range	103 ± 24°	99 ± 29°	0.801	101 ± 27°
Maximum extension	20 ± 13°	23 ± 17°	0.823	22 ± 15°
Maximum flexion	123 ± 14°	122 ± 18°	0.717	122 ± 17°
Supination–pronation range	143 ± 21°	142 ± 25°	0.736	143 ± 23°

Discussion

The purpose of this study was to identify factors associated with increased reoperation rates after ORIF of intra-articular C-type distal humerus fractures. We reviewed a wide range of patient factors (age and gender), injury factors (fracture subclassification, existence of open fracture, and ISS), and treatment factors (time to definitive surgery, length of postoperative immobilization, and type of approach). Multivariate analysis was performed to determine which factors were independently likely to be associated with reoperation and to quantify the strength of these associations.

Presence of an open fracture was associated with a higher risk of reoperation in a prior series, in which 6 of 14 patients with an open C3 distal humerus fracture underwent reoperation. ⁴ With the numbers available in our study, open fracture was not significantly associated with reoperation. However, the majority of open injuries included in the study by Sanchez-Sotelo et al. ⁴ were Gustilo–Anderson type III (8 of 13, 62%), compared to the present study, in which less than half were considered type III injuries.

Several prior studies have compared olecranon osteotomy to other approaches for fracture fixation. One recent study including patients older than 60 years of age with type C distal humerus fractures retrospectively compared 31 elbows that underwent a triceps-sparing approach to 36 elbows in which an olecranon osteotomy was used. ⁷ Reoperation was performed on 15 elbows in the osteotomy group (42%), but only three in the triceps-sparing group (10%); this difference remained significant even when limiting the analysis to patients with a type C3 comminuted fracture. Along the same lines, a high frequency of reoperation for removal of symptomatic hardware has been reported in several case series in which posterior plates were used for osteotomy fixation,^2,11,19,20 although one series reported a comparatively low reoperation rate of 12%. ²¹

Regarding olecranon osteotomy fixation methods, Coles et al. ²² reported on 67 type C distal humerus fractures in which olecranon osteotomy was used for exposure and fixed using either plate or intramedullary screw fixation. There was one case of delayed union and two early revisions of hardware, but the remainder of osteotomies progressed to union. At latest follow-up, 18 patients (29%) had undergone implant removal. There was no apparent difference in reoperation rates between the two fixation methods that were used. Another recent study compared screw and washer fixation to tension band, plate, and combined screw/tension band fixation. ²³ Tension band fixation showed a 3.7-fold higher odds of implant removal and 10.1-fold higher odds of osteotomy nonunion compared to screw/washer fixation. In our study, plate fixation was associated with a 56% reoperation rate compared to a 39% reoperation rate with tension band or screw fixation.

Other treatment factors beyond surgical approach have not been clearly demonstrated in the literature to affect complication or reoperation rates. Time to surgery was explored among open articular distal humerus fractures, with no difference in outcomes and complications seen between patients treated with early surgery (<6 h) and delayed surgery (mean delay, 4.6 days). ²⁴ Fracture fixation with parallel and orthogonal plating techniques has not shown differences in complications or outcome measures in case series.^19,25 In one study, use of locking plate fixation (n = 65) compared to unlocked constructs (n = 31) was shown to be no different with regard to rates of reoperation. ²⁶

One strength of this study is the use of an institutional trauma database, which ensures that details regarding complications and reoperations are recorded in the registry. In addition, the high rate of patient follow-up (98.4%) reduces the risk of selection bias. However, the study also presents certain limitations. First, the study is a retrospective review of prospectively gathered data, which introduces an inherent risk of confounding and bias. Identification of AO type C fractures was based on recorded data from the treating surgeon and no secondary assessment of radiographs was made. Second, the fact that surgeons chose an osteotomy approach may indicate that these injuries were more severely comminuted and unlikely to be easily fixed through a triceps-sparing approach. However, the consistent use of the AO subclassification allowed us to account for the presence of intra-articular comminution in all groups using multivariate analysis. The percentage of type C3 fractures was not significantly different between the osteotomy and triceps-sparing groups. Third, the inclusion of patients treated by multiple upper extremity specialists in the cohort introduces a risk of treatment bias compared to a single-surgeon series. Fourth, stratification of the data based on the variables under investigation reduces the likelihood of demonstrating statistical significance.

Despite improved internal fixation techniques, a high proportion of elbows require reoperation after ORIF for distal humerus fractures. With the numbers available to us, we determined that age, presence of open fracture, gender, ISS, time to treatment, and length of postoperative immobilization were not independently associated with reoperation rate. Use of an olecranon osteotomy was associated with a 3.6-fold higher rate of reoperation. Wound complications (n = 9) and symptomatic hardware (n = 6) were the most common reasons for reoperation. Patients should be counseled regarding the potential for wound complications and hardware removal. The higher rate of complications in fractures fixed through an olecranon osteotomy may reflect additional reoperations due to nonunion of the osteotomy or need to remove hardware from the ulna. In some cases, intra-articular fractures can also be treated successfully without osteotomy. Figure 3 shows a computed tomography scan of an intra-articular distal humerus fracture. The intraoperative view of the joint surface using a bilaterotricipital approach is shown in Figure 4. Figure 5 shows the immediate postoperative radiographs. When treating complex distal humerus fractures with olecranon osteotomy, surgeons should consider fixation methods that avoid implant prominence and soft-tissue irritation.