Elbow hemiarthroplasty vs. open reduction internal fixation for acute Arbeitsgemeinschaft für Osteosynthesefragen/Orthopaedic Trauma Association (AO/OTA) type 13C fractures-A systematic review.

Background: Open reduction and internal fixation (ORIF) is the standard treatment for multifragmentary intra-articular distal humeral fractures. Fractures not amenable by ORIF are treated with total elbow arthroplasty (TEA). In recent years, elbow hemiarthroplasty (EHA) has been used as an alternative to TEA, as weight bearing restrictions and risk of component loosening are lower. We systematically reviewed the literature reporting functional outcomes and complication rates after either EHA or ORIF for Arbeitsgemeinschaft für Osteosynthesefragen/Orthopaedic Trauma Association (AO/OTA) type 13C fractures.
Methods: We searched PubMed, Embase, The Cochrane Library, and Scopus. The inclusion criteria were at least 5 patients, aged ≥50 years, AO/OTA type 13C fracture treated with ORIF or EHA, and evaluation with the Mayo Elbow Performance Score. Literature screening and data extraction were conducted according to the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) statement. The results were synthesized qualitatively using weighted means. No comparative statistical analyses were done.
Results: We included 27 articles, which included 96 patients treated with EHA and 535 patients treated with ORIF. We identified 1 randomized controlled trial and 26 case series. The weighted mean Mayo Elbow Performance Score was 86.9 (n = 89) in the EHA group and 84.7 (n = 535) in the ORIF group. There were 26 (33%) complications (n = 78) in the EHA group and 103 (38%) complications (n = 270) in the ORIF group. Complication rates were generally high in both groups.
Conclusion: We found comparable results of EHA and ORIF, which indicate that EHA is a viable treatment option for AO/OTA type 13C fractures not amenable by ORIF. Because of high risk of bias, interpretation of the results should be done with caution.

The treatment of displaced intra-articular distal humeral fractures is still a challenge for the orthopedic surgeon. Open reduction and internal fixation (ORIF) remains the gold standard., Good and reliable results can often be achieved, but the risk of complications is high.,, In the elderly population, rigid stabilization is not always possible mainly because of poor bone quality, which may increase the risk of nonunion. Total elbow arthroplasty (TEA) has been widely used in elderly patients or for fractures not amenable by ORIF. In 2009, McKee et al found TEA to be superior to ORIF in elderly patients with multifragmentary intra-articular distal humeral fractures in a small group of 42 patients. In case series, TEA has proven to be an efficient treatment for distal humeral fractures where stable fixation cannot be achieved because of comminution of the fracture, severe osteoporosis, or both. Weight bearing restrictions and the risk of component loosening make linked TEA less suitable for active patients and patients with weight bearing demands., In recent years, elbow hemiarthroplasty (EHA) has been used in an attempt to avoid these complications.,, However, the documentation of safety and efficacy is limited,, and there is no existing systematic review reporting the functional results and complications after EHA. To gain a better understanding of the benefits and harms of EHA in comparison to ORIF, we conducted this systematic review of literature reporting functional outcomes and complication rates after either EHA or ORIF for Arbeitsgemeinschaft für Osteosynthesefragen/Orthopaedic Trauma Association (AO/OTA) type 13C fractures.

Materials and methods

Eligibility criteria

We included randomized controlled trials (RCTs), observational studies, and case series. We included studies with at least 5 patients, AO/OTA type 13C1, 13C2, and/or 13C3 fractures, aged ≥50 years, evaluated with the Mayo Elbow Performance Score (MEPS).

We excluded studies of unoriginal data such as reviews, meta-analyses, and educational material, studies of other fracture types (AO/OTA 13A or 13B) or mixed fracture types where data for AO/OTA type 13C fractures could not be extracted, conference abstracts, studies on pathological fractures, fracture sequelae, and cadaver and biomechanical studies. We also excluded studies applying fracture classification systems other than AO/OTA, studies that did not use the MEPS as outcome, studies in other languages than English, and studies of ORIF without the use of at least one plate.

Search strategy

We searched PubMed, Embase, The Cochrane Library, and Scopus.

The database searches were done on April 19 and 21, 2020. The searches were repeated on March 14, 2022, where 3 additional papers were found eligible for inclusion,,, and 1 paper with updated results from an already included paper was identified. These were added to the study, and the tables, figure, and references were updated accordingly.

We conducted a search for studies comparing EHA and ORIF for distal humeral fractures and did not find eligible studies. We then applied 2 different search strategies: one for EHA and one for ORIF. We did not filter or limit the search strategy for EHA. For ORIF, we limited the search strategies to studies published in 2005 and forward. The first author (A.F.N.) developed a search strategy for PubMed, which was then adapted for use with the other 3 databases. The search strategy for EHA was based on the search terms elbow joint, distal humerus, and hemiarthroplasty. The search strategy for ORIF was based on the search terms elbow joint, distal humerus, humeral fractures, intra-articular fractures, internal fracture fixation, and plate fixation. These search strategies were then evaluated by all authors in plenum, and minor adjustments were made before the final search strategies were reached. We used the same search strategies for both the database searches in April 2020 and March 2022. For the complete search strategies, see Supplementary Appendix S1.

Data extraction

One reviewer (A.F.N.) conducted the database searches and extracted references into the review software program (Rayyan QCRI, Cambridge, MA, USA). Two reviewers (A.F.N. and A.A.H.) then independently screened titles and abstracts for relevant studies and read the full-text articles to assess eligibility. The 2 reviewers were blinded to each other’s evaluation. If disagreements would occur, we planned to discuss and resolve these by consensus, but no disagreements occurred. The first author (A.F.N.) then screened the reference lists of included studies where 2 additional eligible studies were identified and included. Data were extracted into a preformed spreadsheet and were then reviewed by the other authors. We planned to discuss and resolve disagreements with all authors in plenum, but no disagreements occurred. Figure 1 shows the exact screening process.

Figure 1

Screening process.

Because of differences in reporting, it was not possible to extract data on complications from all included studies. Therefore, we excluded some studies from analyses on complications.

We did not extract data on radiographic adverse events.

Operational definitions

We defined a case series as a study of patients treated with EHA or ORIF but without a defined control group. Two studies classified as cohort studies, were reviewed and reclassified as case series because our inclusion criteria only allowed data extraction for part of the study population and because 2 surgical approaches for ORIF were compared. In this review, we did not distinguish between surgical approaches, and the 2 groups were therefore merged into 1 case series.

Risk of bias assessment

In accordance with the recommendations of the Cochrane Collaboration, we used the Cochrane risk of bias tool for randomized trials version 2 to assess the risk of bias of the included RCT. In a recent review, Ma et al suggested the Joanna Briggs Institute Critical Appraisal Checklist for Case Series for quality assessment of case series. Therefore, this was used for risk of bias assessment of case series.

Data synthesis and statistics

The results were collected and presented in tables and were then described qualitatively. Sample sizes varied in the included papers. Therefore, weighted means were estimated for range of motion (ROM), age, length of follow-up, and MEPS.

We included 27 articles and extracted data from 96 patients treated with EHA and 535 patients treated with ORIF. The number of patients in each study ranged from 5 to 181 (Table I). There was 1 RCT and 26 case series1, 2, 3,6, 7, 8, 9, 10,,,20, 21, 22,,,,,,,,44, 45, 46, 47,, (Table II). Eight case series,,20, 21, 22,,, were not clearly classified but were viewed as case series for the purpose of this review.

Table I

Results.

Elbow hemiarthroplasty
Study	AO/OTA classification	Mean age (range)	Sex (f/m)	Mean flex/ext arc in degrees (range)	Mean pron/sup arc in degrees (range)	Mean MEPS (range)	Complications
							Ulnar nerve	Infection requiring revision	Periprosthetic fracture or loosening requiring revision	Others	Formalized rehabilitation program (yes/no)
Al-Hamdani (2019)2	13C2 and 13C3	65.4 (53-80)	20/3	108 (60-140)	154 (115-180)	83 (50-100)	3	0	0	4	Yes
Burkhart (2011)7	13C3	75.8 (66-88)	5/0	103 (75-120)	160 (130-180)	92.6 (85-100)	1	0	0	3	No
Hohman (2014)18	13C3	68.4 (59-75)	-/- (5)	101 (70-130)	167 (155-180)	79.4 (67-95)	1	0	0	3	Yes
Leung (2022)27	13C3	65.4 (54-76)	4/4	86 (20-130)	-	78.8 (30-100)	1	0	0	3	-
Nestorson (2015)36	13C3	72.9 (57-84)	-/- (18)	98 (60-140)	179 (160-180) (reported as loss in degrees) (n = 17)	89.7 (55-100)	3∗	0	1	6∗	Yes
Phadnis (2015)38	13C2 and 13C3	78.9 (60-90)	-/- (15)	116 (80-140)	171 (160-180)	89.2 (85-100)	1	0	0	0	No
Schultzel (2022)44	13C2 and 13C3	71.9 (56-81)	9/1	84 (-) (n = 5)	130 (105-150) (n = 5)	88 (75-100) (n = 5)	0	0	1	1	Yes
Smith (2013)47	13C3	66.3 (50-85)	-/- (12)	112 (100-130)	176 (170-180) (n = 11)	94 (70-100) (n = 10)	2	0	0	2	-	Total complications: 26 (n = 78)
Summarized:		70.4 (50-90)	96	104 (20-140) (n = 91)	165 (105-180) (n = 81)	86.9 (30-100) (n = 89)	9 (12%)	0	1 (1%)	16 (21%)		Complication rate: 33% (n = 78)

AO/OTA, Arbeitsgemeinschaft für Osteosynthesefragen/Orthopaedic Trauma Association; ORIF, open reduction and internal fixation; MEPS, Mayo Elbow Performance score.

Gray cells indicate exclusion from analysis.

Unclear if only type C fractures.

Reoperations for stiffness, postoperative nerve affection of the radial and median nerves, infections not requiring revision, resection of heterotopic ossification, etc.

Not exclusively type C fractures.

Calculated from averages.

Table II

Study characteristics.

Elbow hemiarthroplasty
Study	Country	Study design	Selection of cases	Mean time to surgery/days (range)	Implant/prosthesis	Surgical technique	(1 = all, 0 = none, S = some)			Mean follow-up/months (range)
						Approach	Triceps splitting	Olecranon osteotomy	Ulnar nerve transposition
Al-Hamdani (2019)2	Denmark	Case series	Retrospective	-	Latitude	Posterior	1	0	0	25.6 (12-70)
Burkhart (2011)7	Germany	Case series	Prospective	8 (2-18)	Latitude	Bryan-Morrey	1 (unclear)	0	1	11.6 (7-19)
Hohman (2014)18	USA	Case series	Retrospective, consecutive	14 (-)	Latitude	Posterior or lateral	S	S	1	36 (-)
Leung (2022)27	UK and Canada	Case series	Prospective, consecutive	12 (4-22)	Latitude	-	-	-	-	29.8 (14-48)
Nestorson (2015)36	Sweden	Case series∗	Prospective, consecutive	5 (0-11)	Latitude	Posterior	1	0	S	34.5 (24-61)
Phadnis (2015)38	UK	Case series	Prospective, consecutive	13 (2-27)†	Latitude	Posterior, triceps on	1	0	0	35.7 (24-79)
Schultzel (2022)44	USA	Case series	Retrospective	-	Latitude	Bryan-Morrey	Unclear	1	1	115.2 (96-144)
Smith (2013)47	Australia	Case series	Prospective	-	Sorbie & Latitude	Posterior	S	S	S	78.2 (25-133)
										44.9 (7-144)

RCT, randomized controlled trial; ORIF, open reduction and internal fixation.

Reclassified as case series.

Not exclusively type C fractures.

Not clearly classified.

Elbow hemiarthroplasty

Eight papers met our inclusion criteria. All 8 papers were case series: 3 retrospective series,, and 5 prospective series.,,,, Weighted mean age was 70.4 (50-90; n = 96). The weighted mean follow-up was 44.9 months (7-144; n = 96). The mean time to surgery was reported in 5 studies,,,, and ranged from 5 to 14 days. The weighted mean MEPS was 86.9 (30-100; n = 96). One study did not report exact supination/pronation ROM for each patient but reported loss of degrees of ROM for all affected patients as a group, and one study did not report data on supination/pronation ROM. The weighted mean flexion/extension arc was 104 (20-140) degrees (n = 91). The weighted mean supination/pronation arc was 165 (105-180) degrees (n = 81).

We were unable to extract complications for only type C fractures in one study. Therefore, we did not include data from this study in the analysis of complications.

There were 26 (33%) complications (n = 78). For details, see Table I.

Open reduction and internal fixation

Nineteen articles met our inclusion criteria. There was one RCT, 14 retrospective,,,,,,20, 21, 22,,,,,, and 3 prospective,, case series. One study included both a prospective and a retrospective series. Weighted mean age was 74.2 (50-99) years (n = 511). Weighted mean follow-up time was 36.3 months (6-174; n = 511). Because of inconsistent reporting, the mean time to surgery was not analyzed. Weighted mean MEPS was 84.7 (30-100; n = 535). All studies except one reported flexion/extension. Five studies did not report pronation/supination.,,,, One study reported no limitations in postoperative pronation/supination. Weighted mean flexion/extension arc was 100 (30-155) degrees (n = 526), and weighted mean pronation/supination arc was 149 (90-180) degrees (n = 253).

There were 103 (38%) complications (n = 270). For details, see Table I.

Risk of bias

There was generally a moderate to high risk of bias and low internal validity in the case series. The one RCT was judged to have a low risk of bias in 3 domains, high risk of bias in 1 domain, and showed some concerns in the last domain. The high risk of bias was mainly caused by deviations from the intended intervention. For details, see Table III.

Table III

Quality assessment of case series.

Quality assessment of case series
Joanna Briggs Institute critical appraisal tool—Checklist for case series		Clear inclusion criteria	Condition measured in standard, reliable way for all	Valid methods used for identification of condition for all	Consecutive inclusion	Complete inclusion	Clear reporting of demographics	Clear reporting of clinical information	Clear reporting of outcomes or follow-up results	Clear reporting of study site's demographics	Was statistical analysis appropriate	Sum
EHA	Al-Hamdani (2019)2	Yes	Yes	Yes	Yes	Yes	No	No	Yes	No	N/A	6/9
	Burkhart (2011)7	No	Yes	Yes	Unclear	Unclear	No	No	Yes	No	N/A	3/9
	Hohman (2014)18	No	Yes	Yes	Yes	Unclear	Yes	Yes	Yes	No	N/A	6/9
	Leung (2022)27	Yes	Yes	Yes	Yes	Unclear	No	No	Yes	No	Yes	6/10
	Nestorson (2015)36	Yes	Yes	Yes	Yes	Unclear	No	No	Yes	No	Yes	6/10
	Phadnis (2015)38	No	Yes	Yes	Yes	Unclear	No	No	Yes	No	Yes	5/10
	Schultzel (2022)44	No	Yes	Yes	Unclear	Unclear	No	No	Yes	No	Yes	4/10
	Smith (2013)47	Yes	Yes	Yes	Unclear	Unclear	No	Yes	Yes	No	Yes	6/10
ORIF	Al-Hamdani (2022)1	Yes	Yes	Yes	Yes	Yes	No	Yes	Yes	Yes	Yes	9/10
	Ansari (2020)3	Yes	Yes	Yes	Unclear	Unclear	No	Yes	Yes	No	Yes	6/10
	Biz (2017)6	Yes	Yes	Yes	Yes	Unclear	Yes	Yes	Yes	No	Yes	8/10
	Celli (2008)8	No	Yes	Yes	Unclear	Unclear	Yes	No	Yes	No	N/A	5/9
	Chen (2011)9	No	Yes	Unclear	Unclear	Unclear	No	No	Yes	No	Yes	3/10
	Clavert (2013)10	Yes	Yes	Yes	Unclear	Unclear	No	No	Yes	No	Yes	5/10
	Ducrot (2013)14	Yes	Yes	Yes	Yes	Unclear	No	No	Yes	No	Yes	6/10
	Huang, J. (2011)20	Yes	Yes	Yes	Yes	Unclear	No	No	Yes	No	Yes	6/10
	Huang, T. (2005)21	No	Yes	Yes	Yes	Unclear	No	No	Yes	No	N/A	4/9
	Iwamoto (2017)22	No	Yes	Yes	Unclear	Unclear	No	No	Yes	No	N/A	3/9
	Kaiser (2011)24	No	Yes	Yes	Yes	Unclear	No	No	Yes	No	N/A	4/9
	Liu (2009)28	Yes	Yes	Yes	Yes	Unclear	No	No	Yes	No	N/A	5/9
	Luegmair (2008)30	No	Yes	Yes	Yes	Unclear	No	No	Yes	No	N/A	4/9
	Sanchez-Sotelo (2007)42	Yes	Yes	Yes	Yes	Unclear	No	No	Yes	No	Yes	6/10
	Serrano-Mateo (2014)45	Yes	Yes	Yes	Unclear	Unclear	No	No	Yes	No	N/A	4/9
	Shannon (2018)46	Yes	Yes	Yes	Yes	Yes	No	No	Yes	No	N/A	6/9
	Xie (2014)50	Yes	Yes	Yes	Unclear	Unclear	No	No	Yes	No	Yes	5/10
	Zhang (2014)51	Yes	Yes	Yes	Unclear	Unclear	No	No	Yes	No	Yes	5/10

N/A, not applicable; EHA, elbow hemiarthroplasty; ORIF, open reduction and internal fixation.

Discussion

The early design of an acrylic hemiarthroplasty was described by Mellen and Phalen in 1947. Nevertheless, according to Rangarajan et al, interest in EHA decreased as a consequence of the success with the linked, cemented TEA during the 1970s and 1980s. This might be a reason for the limited number of articles reporting the results of EHA. ORIF remains the gold standard for distal humeral fractures with no restrictions on weight bearing or risk of prosthesis failure., This might be a reason for the higher number of articles reporting the results of ORIF.

The MEPS is a surgeon-administrated instrument for evaluating the functional outcome after elbow surgery. Patient-reported pain and range of motion are important to the patient and have a high impact on total MEPS. Furthermore, we chose to use MEPS because it is a widely used score for the evaluation of distal humeral fractures, although it has not been validated for patients with fractures. Other outcome measures were used in some studies, but we chose only to report the results for MEPS and ROM for simplicity of presentation and because it is one of the most used elbow performance scores.

When analyzing functional outcome, it is relevant to account for the surgical approaches and implants used. As shown in Table II, data on surgical approach were generally not clearly described. In addition, there is no consensus on the best surgical approach for exposing the elbow joint in neither ORIF nor TEA. Therefore, we did not find it fitting to distinguish between surgical approaches in this review. Regarding the used implants, all studies in the EHA group used the Latitude system, but various implants and plating systems were used in the ORIF group. This difference is a potential confounder and a weakness of this study, but there is no consensus on the best implants for ORIF of type 13C fractures, and we did therefore not find it reasonable to distinguish between implants within the scope of this review.

Complications were generally inconsistently reported with different definitions and terminology. This poses a problem when data on complications are merged, interpreted, and compared with other publications. In their registry-based study from 2019, Nestorson et al underline the same problem with inconsistent reporting of complications. For this reason, they chose to avoid the term “complication” and instead applied “revision” and “adverse events” with an adverse event defined as “readmission to the hospital due to reasons related to the elbow implant or the elbow procedure but not necessitating revision surgery of the prosthetic components.” This is also the case regarding findings on radiographic imaging, as there is no consensus on the definition of radiological complications or adverse events, and therefore, no clear way of distinguishing complicating radiographic findings and unexpected radiographic findings. Therefore, we chose not to extract data on findings on radiographic imaging.

In a meta-analysis from 2018, Jordan et al compared TEA and ORIF for acute distal humeral fractures in patients aged ≥60 years. They identified 5 comparative studies (n = 330), 14 ORIF (n = 777), and 8 TEA case series (n = 200). They did not describe individual complications but estimated pooled complication rates of 34% (5.9-67) for ORIF and 25% (10-41.6) for TEA with no significant differences. This huge variation might be caused by differences in patient selection, but it could also be caused by differences in reporting of complications.

Furthermore, Jordan et al found a mean MEPS of 91.5 after TEA and a mean MEPS of 82.8 after ORIF but do not report results for ROM. Although there are inherent differences in the complications associated with TEA and EHA, we found similar results for MEPS and complications for EHA and ORIF in our review when compared with the results for TEA and ORIF reported by Jordan et al. Although we did no formal statistical testing, there is a tendency in the results in favor of arthroplasty. To our knowledge, the minimal clinically important difference of MEPS has not yet been reported, although it has been approximated to 10 points. Regardless, the relatively small difference we found in MEPS might not be clinically relevant.

Generally, the included case series were considered to have low to medium quality assessment scores and therefore a high risk of bias, especially selection bias. All studies, except one, were case series where patients were selected based on having undergone either EHA or ORIF for fracture of the distal humerus. The indication for not only surgery, but also the type of surgery, was often not reported, and information about the outcome of nonoperative treatment of type C distal humeral fractures in elderly patients with low functional demands is sparse. Some studies show acceptable functional outcomes with nonoperative treatment in elderly, low-demand populations.,, This could indicate that surgical treatment with ORIF, EHA, or TEA might not necessarily be the best treatment choice in this selected group of patients. In addition, there is a high variation in length of follow-up among the included studies ranging from 6 to 174 months. Longer follow-up times increase the time at risk for complications, mainly secondary arthrosis for ORIF and loosening or ulnar wear for EHA. Functional outcome could also be affected with longer follow-up times, but in shorter follow-up times, patients might not have fully completed their rehabilitation program and might therefore still have the potential for improvement of function.

Conclusion

We systematically reviewed the published literature that reported outcomes after EHA and ORIF for acute AO/OTA type 13C fractures. We found reliable and comparable functional results for both EHA and ORIF but also high complication rates. Owing to high risk of bias, interpretation of the results should be done with caution, and randomized clinical trials comparing EHA with ORIF are needed before safe recommendations can be made.

Disclaimers:

Funding: No funding was disclosed by the authors.

Conflicts of interest: J.V.R. reports grants from Zimmer/Biomed, grants from Depuy Synthes, and speaker fees from Depuy Synthes, without relation to the submitted work. B.S.O. reports grants from Zimmer/Biomed, grants from Depuy Synthes, and speaker fees from Zimmer/Biomed, without relation to the submitted work. The other authors, their immediate families, and any research foundation with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.