1 Introduction

Radial head fractures comprise approximately one out of three elbow fractures.¹ Nationally, there are 2–3 cases annually per 10,000 people.^2,3 Fractures of the radial head can cause instability about the elbow due to their action as a secondary valgus stabilizer of the elbow.^4,5 Fractures of the radial head can be classified according to the Mason classification system and describe type III-IV patterns as often requiring surgical intervention.⁶

Despite recent trends towards radial head arthroplasty (RHA) versus open reduction and internal fixation (ORIF), the best treatment for Mason type III fractures remains unclear.³ The does exist some evidence that patients undergoing RHA have lower complication rates and higher patient satisfaction scores.^7,8 Despite this, malpositioning, aseptic loosening, and poor prosthesis sizing have been associated with RHA.^9,10

Outcome measures such as PROMIS have been developed to evaluate outcomes of various medical conditions and procedures.^11,12 Patient reported outcomes, specifically PROMIS, has been insufficiently studied in the setting of radial head fractures despite the success with other orthopaedic conditions. The purpose of this study is to compare the outcomes of RHA versus ORIF using PROMIS.

2 Methods

This study was granted institutional review board (IRB) approval with a waiver of consent. The PROMIS data was collected between 2015 and 2022 as part of the routine clinical encounter. CPT and ICD-10 codes were utilized to identify the patients with radial head fractures. Radiographs were reviewed by a group of orthopaedic surgeons to determine the Mason classification. Patients between the ages of 18–75 who underwent RHA, ORIF, or radial head excision were included. Open reduction and internal fixation was performed with plate and screw fixation construct. All other fracture patterns were excluded regardless of treatment type. Patients additionally needed to have at least one pre-operative and one 6-week post-operative clinic visit with recorded PROMIS scores. Poly trauma patients, patients with prior elbow instability or trauma, and those with prior elbow surgery were excluded.

The data was securely stored and de-identified. PROMIS scores, range of motion, and demographic variables were analyzed at the preoperative and subsequent post-operative appointments. A combination of T-test and chi-square testing was used to compare the treatment groups. Responsiveness was determined using Cohen's Standardized Response Mean (SRM). These values were used to determine a small, moderate, or large response.¹³ Statistical significance was set at p < 0.05.

3 Results

We identified a total of 143 patients who underwent operative treatment for Mason III radial head fractures. Among them, 89 patients underwent radial head arthroplasty (RHA), 47 underwent open reduction and internal fixation (ORIF), and 7 underwent radial head excision. Table 1presents descriptive characteristics of the included patients. The RHA group had a significantly higher mean age (54 years) compared to the ORIF group (41 years; p < 0.001). Overall, 54% of patients were female (n = 74), and the average BMI was 28.7 kg/m2 (SD: 7). No other statistically significant demographic differences were observed between the two groups.

PROMIS data was completed by 114 patients (68 RHA, 37 ORIF, 5 excision) at the initial visit and by 78 patients (49 RHA, 28 ORIF, 1 excision) at the 12-month follow-up. The PASS questionnaire was completed by 98 patients at the initial visit, 63 at the 6-month follow-up, and 57 at the 12-month follow-up. Among the RHA group, 59 patients (66.3%) completed the PASS question at the initial visit, and 19 (38.8%) completed it at 12 months. In the ORIF group, 25 patients (53.2%) completed the PASS question at the initial visit, and 13 (46.4%) completed it at 12 months (Table 2 and 3).

At the initial pre-surgical visit for the RHA group, the average PROMIS scores for Physical Function (PF), Pain Interference (PI), Upper Extremity (UE), and Depression were 34.75 (SD: 2.4), 63.44 (SD: 2.9), 32.47 (SD: 2.4), and 49.8 (SD: 1.9), respectively. At 12 months, the corresponding scores were 33.42 (SD: 4.1), 61.95 (SD: 4.0), 34.7 (SD: 4.4), and 50.1 (SD: 3.1). Among patients who completed the PASS question, 22.0% reported an acceptable symptom state at the initial visit, increasing to 84.2% at 12 months (p < 0.01).

For the ORIF group, the average PROMIS scores at the initial visit were 38.02 (SD: 2.1) for PF, 63.52 (SD: 2.1) for PI, 33.01 (SD: 2.9) for UE, and 48.1 (SD: 3.3) for Depression. At 12 months, the corresponding scores were 41.24 (SD: 2.9), 61.97 (SD: 4.1), 35.1 (SD: 3.7), and 50.3 (SD: 2.4). Among patients who completed the PASS question, 24.0% reported an acceptable symptom state at the initial visit, increasing to 69.2% at 12 months (p < 0.001).

Comparing RHA and ORIF groups, there were lower PROMIS PF scores among the RHA group at the initial visit (34.75 vs 38.02, p < 0.01) and at 12 months (33.42 vs 41.24, p < 0.001). There were no statistically significant differences in PROMIS PI or UE at the initial visit or at 12 months. PROMIS Depression scores were higher in the RHA group at the initial visit (49.8 vs 48.1, p < 0.001) but were similar at 12 months. Using the ½ standard deviation calculator for mean clinically important difference (MCID), none of the PROMIS PF, PI, or Depression scores demonstrated clinically significant differences at 12 months. Only PROMIS UE showed clinically significant differences at 12 months but not at earlier time points (Fig. 1).

Close

Fig. 1 PROMIS physical function (PF) and pain interference (PI) for radial head arthroplasty (RHA) versus open reduction and internal fixation (ORIF).

Export to PPT

For the excision group, the average PROMIS scores at the initial visit were 37.12 (SD: 3.5) for PF, 65.01 (SD: 3.0) for PI, 32.98 (SD: 2.8) for UE, and 47.2 (SD: 3.7) for Depression. At 12 months, only one patient completed the PROMIS questionnaire and the corresponding scores were 39, 66, 32, and 48. The pass questionnaire was not completed by any patients in this cohort at 12-months. There were no reported complications within this cohort. Range of motion data was not available at final follow-up for this patient. This cohort was excluded from further comparisons to RHA and ORIF due to insufficient follow-up.

PROMIS PI, PF, depression and UE between initial visit and 12-months significant improved in the ORIF group. In the RHA group, there were statistically significant improvements in PROMIS PI and UE between the initial visit and 12 months (all p < 0.05). PROMIS Depression was not significantly different between the initial visit and all follow-up time points. PROMIS PF showed a significant decrease from the initial visit to 12 months but an increase at 6 months (p < 0.01). Using the ½ standard deviation for MCID, only PROMIS PF and UE demonstrated clinically significant differences at 6 and 12 months.

Complications were rate in both groups and there was no observe difference in heterotopic ossification, revision surgery, or infection. Elbow flexion, extension, supination, and pronation were statistically similar between the RHA and ORIF patients with available physical exam data (43 and 22 respectively) (Table 4 and 5).

4 Discussion

Comparing the RHA and ORIF groups, there was no difference in PROMIS PI or UE scores nor was there a clinically significant improvement at the 6- or 12-month mark. The ORIF group demonstrated improved PROMIS PF at all follow-up periods and did show a clinically significant improvement. Patient Acceptable Symptom State (PASS) correlated only with PROMIS UE at 6- and 12- months for both groups.

For upper extremity conditions included carpal tunnel syndrome, distal biceps tendon repair, elbow ulnar collateral ligament reconstruction, and elbow arthroscopy, PROMIS has been found to be an effective patient reported outcome.¹⁴ SF-36 and DASH scores have been found to correlated with PROMIS PF and UE noting that PRIOMIS UE has a significantly ceiling effect in healthy patients.¹⁴

In assessments of PROMIS as a patient-reported outcome for radial head fractures, previous research has indicated that both treatment approaches exhibit enhanced QuickDash and PROMIS UE scores during intermediate and long-term monitoring periods.^15,16 Our data similarity suggests PROMIS UE as well as depression, PF and PI improved at intermediate and long term follow-up however this was only clinically significant in both groups for PROMIS UE at 12-months.¹⁷ PROMIS UE was also the only measurement that correlated with PASS in our population for both groups. This suggests that PROMIS UE may be sensitive enough to detect clinically significant change in functionality for both RHA and ORIF. PROMIS PF, PI, and depression conversely may not be sensitive enough to detect a clinically significant improvement.

When comparing RHA to ORIF for Mason Type III radial head fractures, studies on short and long-term patient-reported outcomes have indicated positive long-term results for both approaches. While the available data slightly favors RHA, ongoing research is needed to confirm this trend. At 6-month and 12-month follow-up, similar scores were observed for PASS, range of motion, complication rates, and PROMIS PI, UE, and depression. The only notable difference within our cohort was in PROMIS PF, where the ORIF group showed significantly better results, possibly due to their younger and healthier demographic. However, PROMIS PF's lack of sensitivity in detecting changes over time raises questions about its applicability in assessing improvement for operative treatment of Mason III radial head fractures. This discrepancy highlights a potential limitation of PROMIS, which might not adequately capture recovery and functional improvement for this specific condition. Additionally, factors such as sleep interference may play a role in outcomes but are not addressed by PROMIS metrics.

As previously mentioned, PROMIS UE data demonstrated clinically significant recovery following operative intervention for radial head fractures in patients treated with ORIF and RA. While a score of 50 with a standard deviation of 10 is typically considered the average in all PROMIS domains for a healthy population, recent research indicates that the average PROMIS UE score for healthy patients under 40 years old was 55.9 ± 6.6, and for patients over 40 years old, it was 51.2 ± 8.2.¹⁸ Hence, despite demonstrating modest improvement between the initial and 1-year follow-up, the final measurements still fell between one and three standard deviations below the normal population, varying depending on patient age and the defined population and standard deviation averages. This indicates that radial head fractures requiring surgical intervention constitute significant injuries to the elbow, resulting in compromised function even one year post-surgery. This understanding equips surgeons with a valuable tool during preoperative discussions to establish realistic expectations for patients undergoing operative intervention.

This study is noted to have several limitations. The completion rate for PROMIS and PASS as well as the clinical follow-up variability prevents complete, direct comparisons between each group at identical time points. Decrease of the response rates over time could have yielded unintended bias in the analysis. PROMIS scores were collected pre-operatively and did not truly represent a baseline “healthy” score but rather the post-injury baseline. Having more complete pre-injury data could have substantially added to the project and tracking of the recovery trajectory back to a true baseline. Age was not controlled for in this comparison which may have yielded unintentional bias in the analysis. Other patient reported outcomes were not available for cross-reference but may have added to the analysis of PROMIS as a effective patient reported outcome.

Complication rates among this population was rage and likely insufficiently powered to draw meaningful complications about comparative rates. Lastly, variability in physical exam documentation and retrospective analysis prevents meaningful clinical conclusions from being made about range of motion specifically.

Funding statement

The authors received no funding for this project.

Guardian/patient consent

Patient consent was not required for this retrospective project.

Ethical statement

There were no procedures performed for this project.

CRediT authorship contribution statement

Thomas John Carroll: Conceptualization, writing, editing. Akhil Dondapati: writing, editing. Jordan Cruse: writing, editing, data collection. Jonathan Minto: Conceptualization, editing. Warren C. Hammert: Conceptualization, editing, . Bilal Mahmood: Conceptualization, editing.