1 Introduction

Radial head fractures (RHFs) account for nearly 33% of elbow fractures and 4% of all fractures.¹ Treatment options for RHFs vary based on the extent of injury to the radial head, classified according to a classification system created by Mason et al. in 1952 and later modified by Johnston et al. in 1962.^2,3 Minimally displaced fractures (Mason type I) often undergo minimal treatment with weight bearing restrictions or short term immobilization at most while fractures with major displacement and/or comminution (Mason type III or IV) often necessitate surgical treatment.^4,5 The optimal treatment strategy for more severe radial head fractures remains inconclusive.^6–8 When reconstruction is infeasible due to comminution, radial head resection or replacement is often utilized.^9,10 Two treatment methods that have been widely used in the surgical management of RHFs are radial head arthroplasty (RHA) and open reduction and internal fixation (ORIF) with no clear reported advantage over one another.^9,11 Many of the aforementioned studies are comprised of relatively small sample sizes, which limits the estimation of actual complication rates. Furthermore, with reported revision rates amongst available studies ranging from 0% to 38.9%, it is important to quantify postoperative complications for the treatment of RHFs.^10,12–15 Therefore, the purpose of this study is to use propensity score matching to assess the incidence of and risk factors for various short-term complications following RHA versus ORIF for the treatment of RHFs.

2 Methods

3 Results

Prior to matching, there were significant differences in demographics and patient-specific risk factors between the two treatment groups (Table 1). Those patients who underwent RHA were found to have significantly higher rates of class III obesity (p < .001), smoking (p < .001), hypertension (p < .001), and ASA class III physical status classification (p < .001). Hospital length-of-stay greater than 2 days occurred for 9.2% of RHA treated patients and 3.7% of ORIF treated patients (p = .007) (Table 2). No significant differences were observed for unplanned readmission, reoperation, non-home discharge, mortality, surgical complications (infection, bleeding, wound dehiscence), or medical complications (reintubation, pulmonary embolism, renal insufficiency, deep venous thromboembolism). Perioperative factors for both treatment groups were analyzed to assess duration of surgery and length-of-stay. RHA treated patients had a longer mean total operative time than ORIF treated patients (111.6min ± 59.2min vs 95.4min ± 48.8min, p < .001). Also, patients in the RHA cohort reported a significantly longer mean length-of-stay (1.0d ± 1.6d vs .4d ± 1.1d, p < .001) (Table 3).

After propensity matching, there were no significant differences in sex, age groups, BMI, ASA classification, major comorbidities (e.g., diabetes, smoking, COPD, ventilator, CHF, hypertension, dialysis, renal failure, chronic steroid use, bleeding disorder, and recent transfusion), or ASA class between the two treatment groups (Table 4). A total of 435 patients were included in the final matched analysis with 250 who underwent RHA and 185 who underwent ORIF for the repair of a RHF. Overall, complication rates for both procedures were markedly low. Comparison of matched-pair analysis revealed RHA patients experienced a significantly higher rate of length-of-stay greater than 2 days (9.2% vs 3.8%, p = .028) (Table 5). No significant differences were found between the two groups for unplanned readmission, reoperation, non-home discharge, overall surgical complications and overall medical complications. Despite relatively low and similar complication rates between the two cohorts, patients who underwent RHA had a significantly longer operative time (111.6min ± 59.2min vs 99.6min ± 54.4min, p = .031) and length of stay (1.0d ± 1.6d vs 0.6d ± 1.3d, p = .003) (Table 6).

Multivariate logistic regression identified several risk factors for various 30-day complications across both cohorts (Table 7). A history of COPD (LR+: 3.993 [1.149–13.875]) and ASA class III (LR+: 3.912 [1.922–7.961]) were significant risk factors for any complication regardless of treatment group. ASA class III (LR+: 5.454 [2.025–114.685]) was observed to be a risk factor for increased rate of non-home discharge. Additionally, a history of bleeding disorder (LR+: 5.192 [1.768–10.031]) and ASA class III (LR+: 4.211 [1.768–10.031]) were significant risk factors for length-of-stay greater than 2 days. No significant risk factors were identified for unplanned readmission or reoperation.

4 Discussion

We investigated the comparative safety and efficacy of performing RHA and ORIF for the treatment of RHFs using the ACS-NSQIP database and a propensity-score matched design. Our study revealed there were no significant differences in 30-day perioperative surgical or medical complications from either surgical procedure; however, RHA treated patients were met with a significantly longer length-of-stay and longer duration of operating time.

Numerous studies have compared the usage rates and utility of RHA to ORIF in treating radial head injuries. Motsi et al. found a significant trend increase in annual percent utilization of RHA (0.22% per year) and a significant trend decrease in the annual utilization of ORIF in treating RHFs (−1.0% per year) between 2007 and 2014.²³ The potential reasons for this finding are multifactorial although significant recent increases in RHA usage should be noted. Currently, there is no clear consensus regarding the superiority of RHA versus ORIF of radial head fractures in terms of perioperative complications.

The growing support for RHA has been most pronounced for comminuted fractures involving 3 or more fragments or when RHFs are present in combination with other injuries such as coronoid fracture, elbow dislocation, or proximal ulna fracture.^1,24 In a retrospective review of 2981 patients, Kupperman et al. found after initial surgical treatment, 12.7% and 14.4% of RHFs that underwent ORIF required a revision surgery at 1 and 2 years, respectively, compared with 8.6% and 10.7% of RHAs.²⁴ Higher reoperation rates in ORIF treated patients have also been reported in the literature.^25,26 Although not statistically significant, we found ORIF treated patients had higher reoperation rates (1.6% vs 0.8%). Chaijenkij et al. conducted a meta-analysis of 526 patients across twelve studies and found patients who underwent RHA reported significantly higher Mayo Elbow Performance Index (MEPI) scores relative to ORIF.⁷ Positive postoperative outcomes and patient satisfaction rates were further substantiated in the meta-analysis by Sun et al.⁸ The group concluded that when compared to ORIF, RHA afforded a significantly higher patient satisfaction rate with better MEPI scores. Contrary to our study, they also noted significantly shorter operation times for patients treated with RHA. While similar results have been demonstrated by additional studies, limitations regarding sample size and quality of evidence exist.^8,25,27,28

Conversely, Klug et al. retrospectively analyzed 88 patients with Mason type III radial head fractures and found patients who underwent RHA had significantly worse outcomes and higher perioperative complication rates than those patients treated with ORIF.²⁹ Kusnesov et al. further found more favorable outcomes in the usage of ORIF as compared to RHA in 77 US military service members. They reported RHA resulted in significantly higher rates of implant failure and overall complication rates.³⁰ The low statistical power and small sample sizes from these two studies hinders the direct conclusions that should be made.

Similar to our analysis, many available studies report comparable outcomes and complication rates between the two treatment procedures. Burdeni et al. concluded that the two techniques had both comparable functional outcomes when using the QuickDASH metric and perioperative complication rates.¹² Similarly, Ryu et al. studied 42 patients with Mason type III or IV RHFs and found no significant differences in clinical outcomes between RHA and ORIF. The group did note that younger patients had better long term functional outcomes when treated with ORIF.³¹ After propensity matching, our results found no significant differences in 30-day readmission, reoperation, non-home discharge, or overall surgical and medical complications. Reoperation rates for RHA (0.8%) and ORIF (1.6%) are lower than reports in current meta-analysis, but given the shorter data collection period this can be expected.^6–8 It is important to recognize the discrepancy in the literature as to the efficacy of both treatment options. Also, understanding that multiple variations of each treatment type exist may further lead to results across studies that are incomparable. For example, a retrospective study of 66 patients with RHFs compared the midterm (16–64mo) outcomes of RHA using short-stemmed bipolar to unipolar long-stemmed osseointegrated rigidly fixed prosthesis. The report found that while both can result in a good outcome, a monopolar prosthesis is preferred due to lower rates of painful loosening.³² This is important as various studies in the literature do not report which method of RHA is used, which may alter comparisons across treatment options.

There are a few important limitations to our study. First, inherent to large national database studies that are retrospective in nature, our analysis was limited by the variables reported in the ACS-NSQIP. This database was formed to document common complications across all major surgical procedures, which precludes the acquisition of certain procedure-specific variables. Furthermore, all RHFs are not the same, therefore additional metrics such as specific fracture type and time before repair would have added to the value of this study. Likewise, many RHFs are not amenable to a certain procedure type. Second, only 30-day outcome measures are collected by the ACS-NSQIP. This database is primarily used for short term complication rates and does not provide information on mid-term or long-term complications and thus likely under-represents the actual rate of complications. Third, surgeon experience and surgical facility type play a role in the management of radial head fractures. Given this is an anonymous national dataset, matching individual surgeons to procedure outcomes was feasible in this study. Taking into account Despite conflicting evidence regarding the efficacy and safety of these two commonly used procedures, it is important to highlight that the treatment of RHFs is multifactorial and short-term complication rates are just one of the many variables in the treatment decision. These findings may be useful to clinicians for both surgical planning, decision-making, and discussions of risk with patients undergoing either treatment modality.

5 Conclusion

This matched cohort study revealed there were no significant differences in 30-day perioperative surgical or medical complications from either surgical procedure for radial head fractures, however; RHA treated patients were met with a significantly longer length-of-stay and longer duration of operating time. We also identified significant risk factors such as ASA class III, a history of a bleeding disorder, and COPD that were independently associated with higher rates of complications regardless of treatment type.