The strength of this report is a single-unit cohort study on a long-term base. Both validated clinical assessment and objective radiographic survey are performed with a comparison between two different types of injuries after RHA. Our study demonstrates encouraging results of RHA for irreparable radial head fractures at an average of nine-year follow up in both valgus injury cases and fracture dislocation cases. None receive subsequent removal or revision of radial head prosthesis. With further comparison of patient characteristics between two groups, statistically significant difference is noted in sex distribution, lateral collateral ligament involvement and time to RHA, which may extend substantial influence in functional and radiographic outcomes and warrant further investigation and comparison between two types of patients on a long-term base.
Biomechanical studies anecdotally analyzed stabilizing influence of radial head replacement based on valgus type injuries with medial collateral ligament insufficiency [14, 15]. However, little clinical information has been addressed on the long-term outcomes after RHA in irreparable radial head fractures with traumatic valgus instability. Harrington IJ, et al reported early experience on RHA between 1966 and 1979; eligible criteria was radial head fractures combined with elbow dislocation, fracture of the proximal ulna, fracture of a major portion of the coronoid process and the medial ligament tear [16]. Subsequent publication with a mean follow-up of 12 years confirmed the long-term function of RHA including two patients of valgus type injury [17]. Another publication in mid- to long-term results after bipolar RHA also reported excellent to good outcome while with a heterogeneity of injury patterns [18].
In the index study, surgical outcomes of RHA is analyzed and compared between valgus injury group and fracture dislocation group at an average of postoperative nine years. Significantly better results regarding elbow motion range and QuickDASH scores are found in valgus injury group than in fracture-dislocation group. While higher MEPS score and lower VAS are also found in valgus injury group, the difference is insignificant. This could be partially attributed to late treatment in more patients of valgus injury group. Based on comparison of patient characteristics between two groups, significantly longer time interval from injury to RHA is found in valgus injury group than in fracture dislocation group. There is a trend for the patients in valgus injury group to have treatment delay, which could be a negative influence and lead to less remarkable difference in MEPS. Postoperative complication is comparable in both groups, and not directly related to the prosthesis implantation. Therefore, early decision making and selection of proper candidates for RHA is recommended in optimization of treatment outcomes.
Late complications with RHA are commonly reported in the literature including arthrosis, pain, elbow stiffness, and heterotopic ossification. In our cases, no significant difference is noted regarding overall complication between two groups. There is an equal patient number of arthrosis in each group, and none of the radiographic analysis presented capitellar erosion. Literature review showing capitellar erosion and arthrosis has been a general concern following RHA [19, 20]. Recent biomechanical studies comparing different prosthesis design suggest greater radiocapitellar contract in monoblock prosthesis may better resist instability and lessen cartilage attrition [21, 22]. While the reported outcome favors monoblock design in our series as well as other clinical studies [8], avoidance of cartilage erosion and arthrosis could be attributed to many factors and technical demands in addition to prosthesis profiles [23].
Periprosthetic osteolysis has been commonly observed around smooth-stemmed radial head prosthesis with loose-fit implantation. In our study, the incidence is significantly higher with fracture dislocation group than valgus injury group. By measuring the osteolysis area from latest radiographs, radiolucency score is also higher in the fracture dislocation group while the difference is not significant. PCC reveals no correlation is found in radiolucency score with VAS and QuickDASH in both groups, and with MEPS in fracture dislocation group. This is comparable with short- to mid-term reports in previous publication [24, 25]. However, analysis between radiolucency score and MEPS in valgus injury group exhibits low to moderate correlation, which warrants extended cohort study with long-term survey.
Several limitations in our study include relatively small sample size with heterogeneity of trauma severity. Pre-injured status of the involved limb could not be fully assessed owing to the retrospective nature. Finally, the missing data in the cases excluded due to insufficient medical records may exert a possible influence on the statistical outcomes.