What Factors Are Associated with Reoperation After Operative Treatment of Terrible Triad Injuries?

Abstract

Background

Terrible triad injuries of the elbow, consisting of posterior ulnohumeral joint dislocation with associated fractures of the radial head and coronoid process, are challenging injuries due to the difficulty in restoring stability to the joint surgically while also attempting to allow early ROM to prevent stiffness. Furthermore, complications are both debilitating and relatively common, frequently requiring reoperation.

Questions/purposes

(1) What patient-, injury-, or surgery-related factors are associated with reoperation after surgical treatment of terrible triad injuries of the elbow? (2) What are the most common causes of reoperation after these injuries?

Methods

Between January 2000 and June 2017, we identified 114 patients who had surgery for terrible triad injuries at two tertiary-care referral centers. Of those, 40% (46 of 114) were lost to follow-up before 1 year, and an additional 5% (6 of 114) were excluded because they underwent the index surgery at an outside institution (n = 4) or underwent closed reduction with or without percutaneous pinning (n = 2). That left 62 patients for analysis in this retrospective study with a minimum of 1-year follow-up (median 22 months; range 12 to 65) or who met the endpoint of reoperation before 1 year. During the study period, indications for surgical treatment of terrible triad injuries of the elbow included joint incongruity or instability precluding early ROM. In our study cohort, 45% (28 of 62) underwent reoperation. Indications for reoperation after surgical treatment included stiffness that interfered with activities of daily life, symptomatic prominent hardware, ulnar neuropathy, instability of the elbow joint at rest or with range of motion, and infection. Patient-related (such as age, sex, race), injury-related (for example, ipsilateral extremity fracture, open fracture), and surgery-related factors (for instance, time to surgery, radial head treatment) as well as outcomes were collected by the treating surgeon at the time of follow-up and ascertained using chart review. The primary outcome measure was reoperation after surgical treatment of a terrible triad injury of the elbow. Bivariate analysis was used to assess whether explanatory variables were associated with reoperation after surgical treatment of terrible triad injuries of the elbow.

Results

Of the patient-, injury-, and surgery-related factors that were analyzed, only radial head treatment was associated with an increased reoperation risk (p = 0.03). No other variable met criteria for inclusion in our multivariable logistic regression model (p < 0.10), and therefore, a multivariable logistic regression model was not performed. The most common indication for reoperation was stiffness (21% [13 of 62 patients]), followed by symptomatic hardware (18% [11 of 62 patients]), nerve symptoms (ulnar neuropathy 16% [10 of 62 patients] and incisional neuroma 2% [1 of 62 patients]), instability (6% [4 of 62 patients]), and wound problems (infection 2% [1 of 62 patients]).

Conclusion

The reoperation risk after surgical treatment of terrible triad injuries of the elbow is high. No patient- or injury-related factors were associated with the reoperation risk. Based on our finding, we recommend fixation of radial head fractures in these injuries when feasible and compatible with early postoperative motion, and we suggest the use of radial head excision or arthroplasty as a secondary options.

Level of Evidence

Level III, therapeutic study.

Introduction

Terrible triad injuries of the elbow, consisting of posterior ulnohumeral joint dislocations with associated fractures of the radial head and coronoid process, present a challenge in restoring joint stability to allow for early ROM, and they are historically fraught with complications and poor functional outcomes [9, 19]. Although surgery is frequently indicated, the choice of surgical technique is controversial. In general, radiocapitellar contact is restored and the lateral collateral ligament complex is repaired [1]. Controversy remains regarding the role of anterior capsule and/or coronoid process repair as well as medial collateral ligament repair [6, 17]. Regardless of the surgical technique, multiple studies have reported complication rates as high as 50% [16-18]. Complications of surgical treatment include acute and chronic instability of the elbow, posttraumatic arthrosis, pain, stiffness, and infection [16, 18, 24, 25], and the reoperation risk is in the range of 20% to 40% [1, 5, 6]. Independent reoperation risk factors after operative treatment of these injuries may alter surgeon practice or influence patient counselling and prognostication; however, such factors are not well-described. Moreover, characterizations of the types of reoperation after a large series of operatively treated terrible triad injuries of the elbow are limited.

Given these gaps in our current understanding, we sought to answer two questions: (1) What patient-, injury-, or surgery-related factors are associated with reoperation after surgical treatment of terrible triad injuries of the elbow? (2) What are the most common causes of reoperation after these injuries?

Patients and Methods

Patient Identification

With institutional review board approval, we retrospectively analyzed all patients of at least 18 years of age who underwent surgical treatment for a terrible triad injury of the elbow at two academic tertiary referral centers from January 2000 to June 2017. Using our institution’s research patient data registry, we identified patients who had ICD-9 and ICD-10 codes for the diagnoses of elbow joint dislocation (832.00, 832.01, 832.02, 832.03, 832.04, 832.09, 832.1, 832.10, 832.11, 832.12, 832.13, 823.14, and 832.19), a fracture of the coronoid process of the ulna (813.02 and 813.12), and a fracture of the head or neck of the radius (813.05, 813.15, 813.06, and 813.16). One-hundred fourteen patients were identified in the initial query (Fig. 1). Of those, 40% (46 of 114) were lost to follow-up prior to 1 year, and an additional 6% (6 of 114) were excluded because they underwent surgical treatment at an outside institution (n = 4) or underwent closed reduction with or without percutaneous pinning (n = 2). That left 62 patients for analysis in this retrospective study with a minimum of 1-year follow-up (median 22 months; range 12 to 65) or who met the endpoint of reoperation before 1 year. We included patients who were lost to follow-up at 1 year if a reoperation occurred before that time.

Fig. 1.

Patient identification flowchart that demonstrates initial patient identification and subsequent inclusion/exclusion of patients to reach final study cohort.

Surgical Indications

During the study timeframe, our indications for surgical treatment of terrible triad injuries of the elbow included a malreduced ulnohumeral or radiocapitellar joint; a comminuted, displaced or angulated radial head fracture; a coronoid fracture measuring greater than 10%; or an unstable elbow that would not allow for early ROM. Surgical technique varied at the discretion of the treating surgeon; however, patients with highly comminuted radial head fractures generally underwent radial head arthroplasty and those with less comminuted fracture types underwent open reduction internal fixation. Repair of the coronoid fracture or anterior capsule, lateral collateral ligamentous complex, and medial collateral ligamentous complex was predicated upon elbow stability. Of those treated surgically, 45% (28 of 62) underwent reoperation. Our general indications for reoperation after surgical treatment included stiffness that interfered with activities of daily life, symptomatic prominent hardware, ulnar neuropathy, instability of the elbow joint at rest or with ROM, and infection.

Data Collection

Medical record data were collected through our institution’s research patient data registry by investigators other than the treating surgeons (PJO, MAT, MJH, DZ). Data collected by record review included patient demographics (age, sex, race, hand dominance, BMI, workers compensation status, and Area Deprivation Index). Race was self-identified by patients in our electronic medical records. The Area Deprivation Index is a validated, neighborhood-based tool for assessing social deprivation comprised of 17 United States census variables that account for metrics such as poverty, housing, employment, and education [11, 20]. We present the Area Deprivation Index by United States national percentile, such that lower percentiles represent more affluent populations, to give readers context. The operations were performed by 11 orthopaedic surgeons. Medical comorbidities were assessed by the Charlson Comorbidity Index, diabetes mellitus, alcohol abuse, and smoking status. Alcohol abuse and smoking status were binary variables assessed at the time of presentation. Injury-related factors were assessed and included other concurrent fracture, ipsilateral extremity fracture, and open fracture. Through a review of the operative reports, we studied the following surgery-related factors: time from injury to surgery, coronoid treatment (fixation or no fixation), lateral collateral ligament treatment (suture anchor repair or transosseous suture repair), radial head treatment (partial excision or no treatment, fixation, or arthroplasty), medial collateral ligament treatment (repair or no repair), and decade of treatment (2000-2009 or 2010-2017). For each patient, all subsequent clinic notes and operative reports were reviewed for any complications and/or reoperations. For all patients who underwent a reoperation, the type of procedure performed and the indication for reoperation were recorded.

Statistical Analysis

Continuous variables are presented as the mean and SD or median and interquartile range, and categorical variables as percentages. We used a t-test to compare parametric variables, a Mann-Whitney U test for nonparametric variables, and a Fisher’s exact test for categorical variables. We used bivariate analyses to screen for variables associated with reoperation with the intent of using multivariable logistic regression analysis to control for the effects of confounding. Our criteria for inclusion in our multivariable logistic regression model was defined a priori as p < 0.10; if less than two variables met our inclusion criteria, then no multivariable logistic regression analysis would be performed and the bivariate analysis would be the most definitive statistical analysis. Subgroup analyses were performed for separate indications for reoperation (reoperation for nerve-related complications, stiffness, instability, hardware complications, and wound complications or infection). We used a convenience sample. The standard significance criterion of α = 0.05 and standard power criterion of (1-β) = 0.80 was used for all statistical tests. A post hoc power analysis showed 80% power to detect an approximately 35% difference in the risk of reoperation between categorical variables.

Results

Factors Associated with Reoperation

Of the patient-, injury-, and surgery-related factors that we analyzed, only radial head treatment was associated with reoperation in the bivariate analysis (p = 0.03) (Table 1). Reoperations were performed in 47% (26 of 55) patients who underwent radial head arthroplasty, 2 of 3 patients who underwent radial head excision, and 0 of 4 patients who underwent radial head open reduction internal fixation. No other variables met criteria for inclusion in the multivariable logistic regression model, and therefore, we did not perform a multivariable logistic regression. Subgroup analyses based on the indication for reoperation, including nerve symptoms, stiffness, symptomatic hardware, recurrent instability, and wound complications or infection did not identify additional factors that were associated with reoperation.

Table 1.

Bivariate analysis of patient-, injury-, and surgery-related factors associated with reoperation after terrible triad injuries of the elbow^a

Variable	Reoperation (n = 28)	No reoperation (n = 34)	Effect sizec	p value
Patient-related variables
Sex (% male)	57% (16)	59% (20)	0.93 (0.34 to 2.57)	0.99
Race (% white)b	75% (21)	79% (27)	0.78 (0.24 to 2.56)	0.76
Age (years), average ± SD	46 ± 14	52 ± 16	6.01 (-1.45 to 13.47)	0.11
Dominant hand injured	47% (13)	41% (14)	1.24 (0.45 to 3.40)	0.99
BMI (kg/m2), average ± SD	32 ± 11	30 ± 7	-2.83 (-8.39 to 2.72)	0.31
Charlson Comorbidity Index, average ± SD	1 ± 1	1 ± 2	0.38 (-0.54 to 1.30)	0.77
Diabetes mellitus	4% (1)	9% (3)	0.38 (0.04 to 3.90)	0.99
Alcohol abuse	4% (1)	3% (1)	1.27 (0.09 to 21.27)	0.99
Current smoker	18% (5)	9% (3)	2.35 (0.51 to 10.87)	0.45
Area Deprivation Index (U.S. percentile)	22 ± 15	22 ± 23	-0.50 (-10.42 to 9.42)	0.92
Workers compensation	14% (4)	15% (5)	0.97 (0.23 to 4.01)	0.99
Injury-related variables
Open fracture	4% (1)	3% (1)	1.22 (0.07 to 20.47)	0.99
Concurrent fracture	14% (4)	9% (3)	1.72 (0.35 to 8.44)	0.69
Ipsilateral upper extremity fracture	7% (2)	9% (3)	1.25 (0.20 to 8.11)	0.99
Surgery-related variables
Coronoid repaired	86% (24)	79% (27)	1.55 (0.40 to 5.97)	0.74
Lateral collateral ligament fixation
Suture anchor	88% (23)	91% (30)	reference	0.99
Transosseous	11% (3)	9% (3)	1.30 (0.24 to 7.07)
Radial head treatment
Radial head arthroplasty	93% (26)	85% (29)
Open reduction and internal fixation	0% (0)	12% (4)		0.03d
No treatment	7% (2)	3% (1)
Medial collateral ligament repaired	11% (3)	9% (3)	1.24 (0.23 to 6.69)	0.99
Time to surgery (days), average ± SD	6 ± 6	6 ± 7	-0.71 (-4.17 to 2.75)	0.68
Decade of treatment
2000–2009	46% (13)	41% (14)	reference	0.80
2010–2017	54% (15)	59% (20)	0.81 (0.29 – 2.22)

^aAll analyses were performed using available data; variables with partially complete datasets included race (98% complete), alcohol abuse (98% complete), smoking status (98% complete), lateral collateral ligament fixation (95% complete), time to surgery (94% complete), and BMI (74% complete).

^bRace was self-identified by the patient.

^cEffect size is given as odds ratios (95% confidence interval) for dichotomous variables and mean difference (95% confidence interval) for continuous variables.

^dStatistical significance, p < 0.05.

Indications for Reoperation

The most common indication for reoperation was stiffness (21% [13 of 62 patients]), followed by symptomatic hardware (18% [11 of 62 patients]), nerve symptoms (ulnar neuropathy 16% [10 of 62 patients] and incisional neuroma 2% [1 of 62 patients]), instability (6% [4 of 62 patients]), and wound problems (infection 2% [1 of 62 patients]) (Table 2). Of the 13 patients who underwent reoperation because of stiffness, heterotopic ossification was identified in three patients on radiographs, with the remaining procedures being attributable to either progression of arthritis or soft-tissue contractures. Ulnar neuropathy accounted for 10 of 11 reoperations for nerve symptoms, with the remaining patient returning for a painful incisional neuroma that was excised on reoperation. The diagnosis of ulnar neuropathy was supported by electrodiagnostic studies in all 10 patients before reoperation. Nine patients underwent ulnar nerve transposition and one underwent in situ decompression. Of the 11 patients who underwent reoperation for symptomatic hardware, five had deep implants removed because of chronic pain without other etiology, four had removal of the radial head implant alone for aseptic loosening seen on plain radiographs, and two had deep sutures removed because of localized pain.

Table 2.

Indications for reoperation

Indication	Reoperation proportion (n = 62)
Stiffness	21% (13)
Nerve symptoms
Ulnar neuropathy	16% (10)
Incisional neuroma	2% (1)
Symptomatic hardware	18% (11)
Instability	6% (4)
Infection or wound problems	2% (1)

Discussion

Terrible triad injuries of the elbow are challenging injury patterns that require reconstitution of elbow stability, allowing for early ROM to prevent long-term stiffness. Unfortunately, complications are quite common and outcomes may be functionally limiting and require reoperation. Our study found that 45% of patients followed for at least 1 year after this injury undergo reoperation. The indications for reoperation and risk factors associated with reoperation after terrible triad elbow injury surgical treatment are not well-established. With this in mind, we sought to identify any patient-, injury-, or surgery-related factors associated with reoperation and define the reoperation risk based on indication. In this study, only radial head treatment was associated with reoperation in the bivariate analysis; proportions of reoperation were higher after radial head excision and radial head arthroplasty than radial head open reduction internal fixation. Patients may be counseled that stiffness, symptomatic hardware, and ulnar neuropathy are the leading reasons for reoperation.

Limitations

Our study has several important limitations. First, our study has the potential for transfer bias, as roughly 40% of our patient population was lost to follow-up before 1 year. The group lost to follow-up may have a higher reoperation risk than our inclusion cohort, which would lead to an underestimation of the actual reoperation risk. Second, our study was conducted over a 20-year period, across two tertiary academic centers, and had 11 different surgeons. This heterogeneity allows for the potential of selection bias. However, although surgical technique may vary among surgeons, the overall principle of restoring a stable and concentrically reduced elbow throughout the physiologic arc of motion remains constant. Although surgical techniques, rehabilitation protocols, and postoperative follow-up were not standardized, this is true to clinical practice and adds external validity to our findings. Third, our study may be affected by assessment bias, as the sole outcome was reoperation. Other endpoints such as patient-reported outcome measures or ROM may add to a more complete assessment of overall function after terrible triad injuries of the elbow. Although the decision to undergo reoperation is complex and predicated on shared decision-making between the surgeon and the patient, we feel the indications for reoperation in this retrospective study were within the standard of care and as such, are applicable to clinical practice. Lastly, we also recognize that over the timeframe of our study, the treatment of complex elbow fracture-dislocations has evolved, and consequently, it is possible that the reoperation risk may have decreased over time. In addition, although surgical technique and understanding certainly evolved over this timeframe, so did implant design. Specifically in regard to radial head prostheses, studies have shown that there is a higher complication risk with certain implants [21], which we were unable to specifically examine in this study.

Factors Associated with Reoperation

The only factor that was associated with a higher reoperation risk was radial head treatment, a high proportion of which occurred after radial head arthroplasty. The most common indications for reoperation in patients who underwent radial head arthroplasty were aseptic loosening of the radial head, stiffness due to progressive radiocapitellar arthritis, and instability. These findings may be attributable to inherent complications of radial head arthroplasty, such as the potential for future loosening. In addition, proper sizing of the radial head intraoperatively can be challenging, with improper size leading to instability. Alternatively, patients who underwent radial head arthroplasty may have had different injury properties, such as higher-energy injuries and more soft-tissue damage, that may have predisposed them to undergo a future reoperation [8, 15]. Although concomitant injury properties were studied (for example, open fracture, concomitant other fracture, or concomitant ipsilateral upper extremity fracture), there are properties of the injury such as energy and mechanisms that remain unobserved in this study. Based on the findings of our study, we recommend open reduction and internal fixation of the radial head in terrible triad injuries of the elbow when stable osteosynthesis can be feasibly achieved and is compatible with early ROM, reserving radial head arthroplasty and radial head excision as secondary options. Although we did not identify any patient- or injury-related factors that contributed to a higher reoperation risk in our study, prior studies on patients who sustain orthopaedic trauma suggest that higher BMI, higher Charlson Comorbidity Index scores, and workers compensation insurance status portend worse outcomes [7, 12, 23]. In contrast to a prior study of orthopaedic elbow trauma [3], we did not find an association between social deprivation and reoperation after surgical treatment of terrible triad injuries of the elbow.

Indications for Reoperation

The most common indications for reoperation in our study were stiffness, symptomatic hardware, and ulnar neuropathy, whereas less common indications for reoperation were recurrent instability and wound complications. Our study supports prior evidence on the notable risk of stiffness and neuropathy after surgical treatment of this injury. In our study, 21% of patients underwent reoperation for stiffness, alone or in combination with another indication. This is slightly higher than the results of previous studies that showed a reoperation risk between 9% and 14% after surgical treatment for a terrible triad injury of the elbow due to stiffness [6, 13, 14). The reoperation risk for ulnar neuropathy in our study was 16%, which is consistent with previously reported risk of 8% to 18% [6, 10,14). The time to secondary ulnar nerve surgery was just over 18 months. Earlier studies examining secondary ulnar neuropathy have theorized that it may develop because of traction neuritis caused by post-traumatic adhesions [2, 4]. Adhesions and fibrosis that develop around the ulnar nerve may not clinically manifest until the patient gains a certain amount of motion. Furthermore, once adhesions surrounding the ulnar nerve develop and cause compression, the patient may be more likely to experience stiffness because of increased arcs of motion that produce more traction on the already compressed nerve. This link between stiffness and ROM may explain why 10% of patients underwent reoperation concurrently for elbow stiffness and ulnar nerve symptoms. The reoperation risk for instability was low and occurred in only 6% of patients. Because of the low number of reoperations for instability, we did not perform quantitative statistics. Two studies have reported higher instability risk after fixation of radial head fractures than after radial head arthroplasty [13, 22]. By contrast, in our study, three of four patients who underwent reoperation for instability underwent radial head arthroplasty at the time of the initial surgery and the remaining patient underwent open reduction and internal fixation. Of the four patients who underwent reoperation for instability, three had dislocations within the first 2 months after surgery and one had late instability that developed approximately 5 years postoperatively. Our findings support an evolution of our understanding and management of terrible triad injuries of the elbow since their initial description when treatment was fraught with recurrent instability [18]. Although surgical techniques were heterogeneous in our study, our findings indicate that with current surgical management, recurrent instability is uncommon [25]. Future studies should focus on comparisons of surgical techniques and rehabilitation protocols to minimize stiffness and ulnar neuropathy after surgical treatment of these injuries.

Conclusion

In a retrospective study of 62 patients with terrible triad injuries of the elbow with minimum 1-year follow-up, we found a high risk of reoperation of 45%. Radial head treatment was the only variable associated with reoperation in our bivariate analysis. Patients may be counseled that the most common indications for reoperation are stiffness, symptomatic hardware, and ulnar neuropathy. In contrast with early reports of this injury, current risks of recurrent instability are low. Based on the findings of our study, we recommend open reduction and internal fixation of the radial head in terrible triad injuries of the elbow when stable osteosynthesis can be feasibly achieved and is compatible with early ROM, reserving radial head arthroplasty and radial head excision as secondary options. Future studies on terrible triad injuries of the elbow may focus on factors associated with patient-reported outcome measures and ROM as well as comparisons of surgical techniques and postoperative rehabilitation protocols to minimize stiffness and ulnar neuropathy.