Osteosynthesis and radial head arthroplasty achieve comparable and favorable results in terrible triad injury: a systematic review using propensity score matching

Abstract

The terrible triad literature reports data across wide ranges of patients ages and the treatment allocation for the radial head is commonly based on fracture severity. These observations raise the question of selection bias which can hinder comparative efficacy. Propensity score matching analysis may provide better control of patient variables and treatment allocation. We convey two conclusions from this analysis. Firstly, clinical outcome metrices were favorable for both radial head arthroplasty and open reduction internal fixation following terrible triad injury. This finding supports the recent literature which describe satisfactory outcomes for acute surgical management of these injuries. Secondly, the analysis suggests that revision rates may be lower for both treatment options than previously reported in the literature.

Keywords

Elbow dislocation radial head arthroplasty radial head ORIF terrible triad unstable elbow Level of evidence IV systematic review

Introduction

Terrible triad (TT) injuries of the elbow are complex injury patterns which have historically been problematic for surgeons due to high rates of complications and a lack of consensus on treatment.^1,2 Hotchkiss initially used this nomenclature to describe elbow ligament injury with concomitant radial head and coronoid fractures.³ This triad of disruption produces loss of the primary and secondary stabilizers of the elbow.⁴ The most common injury mechanism involves forceful ground contact with an outstretched arm. Force transmits proximally leading to increased axial, rotary and valgus loads at the elbow causing catastrophic failure of the bony and capsuloligamentous structures.

It is now more widely accepted that surgical treatment that addresses each component of the triad of disruption improves outcomes.^5–10 The radial head is a secondary stabilizer of the elbow, providing buttress support against valgus stress.^11,12 Thus, repair or replacement of the radial head in TT injury is crucial in restoring elbow stability.² Algorithms for surgical treatment of TT injuries have been developed in order to facilitate standardized surgical intervention which have produced increasingly favorable outcomes.^6,13,14

Utilization of radial head arthroplasty (RHA) is increasing,^15–17 with substantial increases seen in the presence of traumatic elbow instability.¹⁸ The driving forces behind this increase in arthroplasty utilization may include the favorable outcomes of RHA in older adults and the short-term failure rates reported for open reduction internal fixation (ORIF).^5,18–21 Although short term findings for RHA are compelling, concerns remain for long term complications due to the prosthesis.^5,21

Based on aggregate findings, RHA and ORIF yield equal outcomes in TT injuries^22–24 but these findings may be confounded by two important selection biases. The first is inclusion of patients from a broad age range (mean 20–77, SD 14.25, N = 403).^{1,5,7,14,21,22,25} These reports pool results from patients with widely different postoperative functional demands. The second is the treatment allocation of ORIF to fractures of lower severity and RHA to those of higher severity. A review by Kyriacou et al. concluded a significantly higher number of Mason III fractures in the RHA (79%) group compared to the ORIF group (48%) (P = 0.0001).²²

These observations raise the question of comparative efficacy if studied in better controlled circumstances. Therefore, the objective of this systematic review was to utilize propensity score matching to compare the results RHA and ORIF in TT injury. This analysis will provide better control of patient variables and treatment allocation.

Materials and methods

Data sources

Ethics approval was not required for this work. Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines were used to structure the format. PubMed database was queried for publications with the following terms: radial head fracture AND Mayo, terrible triad AND Mayo, terrible triad AND outcomes, elbow dislocation AND Mayo, elbow dislocation AND outcomes. English language articles published prior to July 15, 2021, were available for inclusion.

Study selection

Study inclusion required stages of screening. The initial requirement was a table with individual patient reporting. Within the table, the following data were required for inclusion: patient age, patient gender, follow up term, radial head fracture classification and the reporting of a numeric Mayo Elbow Performance Score (MEPS). If the specific value of the MEPS was not provided, the study was excluded. These covariates provided the basis for propensity score matching. Further matching by injury characteristics and treatment variables were not possible due to heterogeneity of reporting. The terrible triad literature commonly classifies the radial head fracture using the Mason classification (I–IV). This variable was reported as described in the included articles. Selection was independently assessed by 2 authors.

Data extraction

The outcome measures evaluated were MEPS, Disabilities of the Arm, Shoulder and Hand (DASH), failure/revision of the surgical construct, and the functional parameters of elbow arc of flexion/extension and forearm arc of rotation. Revision data represents all-cause revision. Additionally, revision cases were excluded therefore the outcomes analyzed represented primary surgical intervention.

Propensity score matching

Propensity score matching (PSM) provides statistical analysis based on individual patient covariates in order to better equalize potential confounding factors.²⁶ By limiting selection bias, causal effects can be derived from the data which strengthens the conclusions. PSM application has increased due to the monetary cost and time investment associated with prospective randomized study designs.²⁷ Covariates used to match patient pairs included age, gender, term of follow up, and fracture classification. Matching was done using the matchit package within R software (Lucent Technologies, Murray Hill, NJ, USA). The method used was nearest neighbor with caliper set to 0.1.²⁸ The caliper determines the allowable difference between values of the covariates. The caliper used for this analysis maintained a narrow range of allowable variance between matched pairs for the covariate values.²⁹

Data synthesis

Pearson's chi square with Yates’ continuity correction was used to determine trends in treatment based on patient gender. Fisher's exact test was used to determine trends in treatment based on radial head fracture classification. Wilcoxon rank sum test was used to compare the follow up term between treatment groups given the non-normal distribution of the sample. Pearson's chi square test, Fisher's exact test and Student t-test were used to compare matched pair data. Significance level was set at P < 0.05.

Results

Following appropriate exclusions, the database query generated 15 articles which met the inclusion criteria (Figure 1). Within the included articles, analysis was based on 77 cases of RHA (mean age 49 years, mean follow up 43 months) and 97 cases of ORIF (mean age 42 years, mean follow up 36 months).

Figure 1.

Flowchart depicting literature search and article selection process with exclusion criteria.

Propensity score matched pairs

Sixteen (16) patients with Mason III fractures were matched from the RHA and ORIF groups for outcome comparison (Table 1). When matched for age, gender and follow up term, MEPS (P = 0.90) and DASH (P = 0.80) comparisons between RHA and ORIF were not significant. Consistent with the finding across the entire sample, patients with Mason III fractures who were treated with RHA were significantly older than patients with Mason III fractures who were treated with ORIF (P = 0.02).

Table 1.

Covariate and outcome comparison between RHA and open reduction internal fixation for Mason III fractures in terrible triad injury.

Mason III fractures	Total			Propensity score matched for age, gender, fracture classification and follow-up term
	RHA^a	ORIF^a	P value	RHA^a	ORIF^a	P value
Sample size (N)	55	21	-	16	16	-
Age	50.2 ± 14.4	40.8 ± 14.8	0.02	44.3	46.4	0.68
Gender (m)	52.7%	76.2%	0.07	68.8%	75%	1.00
Follow up (mth)^b	42 ± 32	38 ± 37	0.57	38.3	41.9	0.78
MEPS^b^,^c	90.9 ± 11.5	91.7 ± 7.6	0.75	91.3	91.7	0.90
DASH^b^,^c	12.4 ± 13.7	8.0 ± 7.2	0.20	9.9	8.7	0.80
FE arc of motion^b^,^c	120 ± 25	118 ± 20	0.70	115	117	0.81
PS arc of motion^b^,^c	150 ± 31	141 ± 23	0.22	143	143	0.95

RHA = radial head arthroplasty, ORIF = open reduction internal fixation.

Mean values.

FE = flexion extension, PS = pronosupination, MEPS = Mayo Elbow Performance Score, DASH = Disabilities of the Shoulder and Hand score.

Overall sample

Mean age was significantly higher in the RHA group compared to the ORIF group (P = 0.002). Evaluating treatment method and gender, males were significantly more likely to be treated with ORIF while females were significantly more likely to be treated with RHA (P = 0.01) (Table 2). Evaluating treatment method and radial head fracture classification, patients were significantly more likely to be treated with RHA with increasing levels of Mason fracture classification (P < 0.01) (Table 2). Comparisons between the RHA and ORIF groups are shown in Table 3. Comparison between ORIF and RHA for categorical MEPS is shown in Figure 2 and for DASH scores in Figure 3.

Figure 2.

Mayo elbow performance scores for RHA and open reduction internal fixation in terrible triad injury.

Figure 3.

Disabilities of the arm, shoulder and hand scores for RHA and open reduction internal fixation in terrible triad injury.

Table 2.

Trends for radial head treatment method in terrible triad injury based on gender and fracture classification.

	RHA^a	ORIF^a	P value
Female	35	26	0.01635
Male	42	71	0.01635
Mason I	0	5	<0.0001
Mason II	10	69
Mason III	55	21
Mason IV	12	2

RHA = radial head arthroplasty, ORIF = open reduction internal fixation.

Table 3.

Covariate and outcome comparison between RHA and open reduction internal fixation in terrible triad injury for all included articles.

	RHA^a	ORIF^a	P value
Sample size (N)	77	97	-
Gender, m	42 (55%)	71 (73%)	<0.001
Age^b	49 ± 15 years	42 ± 13.6 years	0.002
Follow up term^b	43 ± 31 months	36 ± 28 months	0.04
FE arc of motion^b^,^c	121 ± 22 degrees	113 ± 24 degrees	0.04
PS arc of motion^b^,^c	149 ± 29 degrees	144 ± 23 degrees	0.18
MEPS^b^,^c	91.2 ± 10.6	89.4 ± 11	0.29
DASH^b,c	13.2 ± 13.9 (N = 54)	15.4 ± 18 (N = 54)	0.49
Revision/removal	3 (4%)	2 (2.5%)	0.45

RHA = radial head arthroplasty, ORIF = open reduction internal fixation.

Mean values.

FE = flexion extension, PS = pronosupination, MEPS = Mayo Elbow Performance Score, DASH = Disabilities of the Shoulder and Hand score.

Discussion

Propensity score matching is a statistical method that reduces the confounding potential of selection bias.^28,30 This novel application of PSM in a systematic review generated 16 matched pairs for Mason III radial head fractures in TT injury. Our overall results demonstrate favorable outcomes for RHA and ORIF which differ from some existing reports.²⁵ Further, functional outcomes and MEPS were not significantly different between RHA and ORIF. While the conclusion of similar clinical outcomes between RHA and ORIF is consistent with some reports,²² our results showing low rates of ORIF failure are in contrast with some of the literature.^5,18,19,21 In totality, these results demonstrate that prospective randomized trials for this complex fracture pattern can be confidently and ethically initiated due to the favorable outcomes between these 2 treatment methods.

The literature has demonstrated a shift in radial head fracture management which may not be completely validated by the currently available data. Utilization of RHA is increasing, due in part to favorable outcomes in RHA and reports of short-term failure in ORIF.^5,21 Across the pooled data of 97 ORIF cases, 1 case was revised due to malunion, and 1 case required reoperation due to failure of the LCL repair, equating to a reoperation rate of 2.5%. Therefore, our results suggest a low rate of ORIF reoperation at a mean patient age of 42 and a mean follow up of 36 months. This finding is in contrast to prior reports of high reoperation rates for ORIF following TT injury, ranging up to 33% (mean age 42 years, mean follow up 41 months) and 44% (mean age 48 years, mean follow up 18 months).^5,21 Reoperations following ORIF may be due to the complexity of the injury pattern,^1,31 disruption of the vascular anatomy of the radial head,^32,33 and the surgical dissection required for anatomical reduction and fixation.^1,31

Although RHA continues to gain momentum due to favorable outcomes, these findings have been less demonstrable in the terrible triad literature. Our pooled data across 77 arthroplasties yielded a 4% rate of revision at a mean follow up of 43 months. This finding is in contrast to a previous review which calculated rates of revision exceeding 40% for RHA in TT injury (mean age 50 years, mean follow up 36 months).³⁴ Arthroplasty failures are due to inappropriate prosthesis size and position which may lead to painful aseptic loosening.^35,36 Lengthening of the radius with the prosthesis—termed joint overstuffing—can lead to alterations in ulnohumeral and radiocapitellar pressures which can predispose to degenerative changes.³⁷ Symptoms due to overstuffing include pain, stiffness, loss of motion, and aseptic loosening. Rates of loosening may vary based on polarity and stem fixation.^36–38 The aggregated results for RHA and ORIF reinforce the increasing body of evidence which describe satisfactory outcomes following acute surgical management of terrible triad injury.^{1,6,25,36,39,40}

Weaknesses of this review include the inherent limitations of amalgamating the work of others into a cohesive summary. Measurement bias from multiple practitioners and the variation in diagnostic and surgical skill may confound the current work. Matching cases by fracture classification allowed analysis on a specific fracture pattern but this classification does not account for displacement. Mason III fractures represent a wide spectrum of injury severity. The magnitude of injury severity may be a more important variable than fracture pattern when determining surgical treatment. The inclusion criterion of individual case reporting and MEPS was necessary for application of propensity score matching but reduced the available sample of literature and may bias the results. Further, this statistical analysis yielded a small sample for comparison of ORIF and RHA. Changes in the parameters within propensity score matching may alter the sample size and the similarity of the covariates. Revision/reoperation was aggregated for the entire sample however due to insufficient reporting this variable was not included in the propensity score matched analysis. Further, revision/reoperation was not a component of the inclusion criteria thus, our conclusions therein are limited. Finally, the current work compared treatment options for radial head fracture in terrible triad injury. Clinical outcomes for these injury patterns are also influenced by the magnitude of coronoid fracture and the subsequent treatment of this component of the injury. Strength of the current work include the broad search parameters which returned an appropriate volume of literature. The applied statistical analysis homogenizes the outcomes and provides a reproducible foundation for future review work. The intended consequence of a systematic review is to derive generalizable conclusions from a large albeit variable sample. Utilizing propensity score matching, a more specified conclusion may be obtained. Matching patients for fracture classification, age, gender and follow-up term allowed better control of some of the confounding variables which are inherent to systematic reviews.

Conclusion

In summary, we convey two conclusions from this analysis. Firstly, clinical outcomes were favorable for RHA and open reduction internal fixation following terrible triad injury. This finding supports the recent literature which describe satisfactory outcomes for acute surgical management of these injuries. Secondly, the analysis suggests that revision rates may be lower for both treatment options than previously reported in the literature.

Footnotes

Acknowledgments

The authors appreciate the support of Rachel Elaine York.

Declaration of conflicting interests

The authors would like to declare the following conflict of interest with respect to the research, authorship, and/or publication of this article: Deana Mercer, speakers bureau Skeletal Dynamics, Axogen.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

ORCID iDs

John J Heifner

Asdrubal E Rivera Dones

References

Ring

Jupiter

Zilberfarb

. Posterior dislocation of the elbow with fractures of the radial head and coronoid. J Bone Joint Surg Am 2002; 84: 547–551.

Josefsson

Gentz

Johnell

, et al. Dislocations of the elbow and intraarticular fractures. Clin Orthop Relat Res 1989; 246: 126–130.

Hotchkiss

. Fractures and dislocations of the elbow. In: Rockwood

Green

Jr Bucholz

Heckman

(eds) Rockwood and Green’s fractures in adults. 4th ed, 1. Philadelphia: Lippincott-Raven, 1996, pp.929–1024.

O’Driscoll

King

Jupiter

, et al. The unstable elbow. Instr Course Lect 2001; 50: 89–102.

Leigh

Ball

. Radial head reconstruction versus replacement in the treatment of terrible triad injuries of the elbow. J Shoulder Elbow Surg 2012; 21: 1336–1341.

Papatheodorou

Rubright

Heim

, et al.

Terrible triad injuries of the elbow: does the coronoid always need to be fixed?

Clin Orthop Relat Res 2014; 472: 2084–2091.

Giannicola

Calella

Piccioli

, et al.

Terrible triad of the elbow: is it still a troublesome injury?

Injury 2015; 46: S68–S76.

Beingessner

Stacpoole

Dunning

, et al. The effect of suture fixation of type I coronoid fractures on the kinematics and stability of the elbow with and without medial collateral ligament repair. J Shoulder Elbow Surg 2007; 16: 213–217.

Zeiders

Patel

. Management of unstable elbows following complex fracture-dislocations–the “terrible triad” injury. J Bone Joint Surg Am 2008; 90: 75–84.

10.

Lindenhovius

Jupiter

Ring

. Comparison of acute versus subacute treatment of terrible triad injuries of the elbow. J Hand Surg Am 2008; 33: 920–926.

11.

Morrey

Stormont

. Force transmission through the radial head. J Bone Joint Surg Am 1988; 70: 250–256.

12.

van Riet

Van Glabbeek

Baumfeld

, et al. The effect of the orientation of the radial head on the kinematics of the ulnohumeral joint and force transmission through the radiocapitellar joint. Clin Biomech (Bristol, Avon) 2006; 21: 554–559.

13.

Mathew

Athwal

King

. Terrible triad injury of the elbow: current concepts. J Am Acad Orthop Surg 2009; 17: 137–151.

14.

Pugh

Wild

Schemitsch

, et al. Standard surgical protocol to treat elbow dislocations with radial head and coronoid fractures. J Bone Joint Surg Am 2004; 86: 1122–1130.

15.

Klug

Gramlich

Wincheringer

, et al. Epidemiology and treatment of radial head fractures: a database analysis of over 70,000 inpatient cases. J Hand Surg Am 2021; 46: 27–35.

16.

Stirling

Malhas

Rymaszewski

, et al. The changing epidemiology of radial head replacement over a 22-year period in Scotland. Ann R Coll Surg Engl 2021; 103: 612–614.

17.

Foroohar

Prentice

Burfeind

, et al. Radial head arthroplasty: a descriptive study of 970 patients in an integrated health care system. J Shoulder Elbow Surg 2022; 31: 1242–1253.

18.

Kupperman

Mitchell

. Treatment of radial head fractures and need for revision procedures at 1 and 2 years. J Hand Surg Am 2018; 43: 241–247.

19.

Yan

Song

, et al.

Radial head replacement or repair for the terrible triad of the elbow: which procedure is better?

ANZ J Surg 2015; 85: 644–648.

20.

Hackl

Wegmann

Hollinger

, et al. Surgical revision of radial head fractures: a multicenter retrospective analysis of 466 cases. J Shoulder Elbow Surg 2019; 28: 1457–1467.

21.

Watters

Garrigues

Ring

, et al.

Fixation versus replacement of radial head in terrible triad: is there a difference in elbow stability and prognosis?

Clin Orthop Relat Res 2014; 472: 2128–2135.

22.

Kyriacou

Gupta

Bains

, et al. Radial head replacement versus reconstruction for the treatment of the terrible triad injury of the elbow: a systematic review and meta-analysis. Arch Orthop Trauma Surg 2019; 139: 507–517.

23.

Chen

Wang

Cao

, et al. Comparison between radial head replacement and open reduction and internal fixation in clinical treatment of unstable, multi-fragmented radial head fractures. Int Orthop 2011; 35: 1071–1076.

24.

Schindelar

Sherman

Ilyas

. Comparison of surgical techniques for fixation of terrible triad injuries of the elbow: a meta-analysis. Orthopedics 2020; 43: 328–332.

25.

Klug

Nagy

Gramlich

, et al. Surgical treatment of the radial head is crucial for the outcome in terrible triad injuries of the elbow. Bone and Joint Journal 2020; 102-B: 1620–1628.

26.

Kim

. Propensity score analysis in non-randomized experimental designs: an overview and a tutorial using R software. New Dir Child Adolesc Dev 2019; 2019: 65–89.

27.

D’Agostino

RBJ

. Propensity score methods for bias reduction in the comparison of a treatment to a non-randomized control group. Stat Med 1998; 17: 2265–2281.

28.

Baumbach

Urresti-Gundlach

Braunstein

, et al. Propensity score-matched analysis of arthroscopically assisted ankle facture treatment versus conventional treatment. Foot Ankle Int 2021; 42: 400–408.

29.

Austin

. Optimal caliper widths for propensity-score matching when estimating differences in means and differences in proportions in observational studies. Pharm Stat 2011; 10: 150–161.

30.

Jeong

Kim

Rhee

, et al. Risk factors for and prognosis of folded rotator cuff tears: a comparative study using propensity score matching. J Shoulder Elbow Surg 2021; 30: 826–835.

31.

Faraj

Livesly

Branfoot

. Nonunion of fracture of the neck of the radius: a report of three cases. J Orthop Trauma 1999; 13: 513–515.

32.

Koslowsky

Schliwa

Koebke

. Presentation of the microscopic vascular architecture of the radial head using a sequential plastination technique. Clin Anat 2011; 24: 721–732.

33.

Yamaguchi

Sweet

Bindra

, et al. The extraosseous and intraosseous arterial anatomy of the adult elbow. J Bone Joint Surg Am 1997; 79: 1653–1662.

34.

Vannabouathong

Venugopal

Athwal

, et al. Radial head arthroplasty: fixed-stem implants are not all equal-a systematic review and meta-analysis. JSES Int 2020; 4: 30–38.

35.

Viveen

Kodde

Heijink

, et al. Why does radial head arthroplasty fail today? A systematic review of recent literature. EFORT Open Rev 2019; 4: 659–667.

36.

Schnetzke

Jung

Groetzner-Schmidt

, et al. Long-term outcome and survival rate of monopolar radial head replacement. J Shoulder Elbow Surg 2021; 30: e361–e369.

37.

Van Glabbeek

van Riet

Baumfeld

, et al. Detrimental effects of overstuffing or understuffing with a radial head replacement in the medial collateral-ligament deficient elbow. J Bone Joint Surg Am 2004; 86: 2629–2635.

38.

van Riet

Sanchez-Sotelo

Morrey

. Failure of metal radial head replacement. J Bone Joint Surg Br 2010; 95: 661–667.

39.

Liu

, et al. Operative treatment of terrible triad of the elbow with a modified pugh standard protocol: retrospective analysis of a prospective cohort. Medicine (Baltimore) 2018; 97: e0523.

40.

Sanchez-Sotelo

Morrey

. Complex elbow instability: surgical management of elbow fracture dislocations. EFORT Open Rev 2016; 1: 183–190.