Anatomic Monopolar Press-fit Radial Head Arthroplasty; High Rate of Loosening at Mid-Term Follow Up

Abstract

Introduction

Radial head arthroplasty (RHA) is used for the management of unstable or unreconstructable injuries of the radial head. Our aim was to investigate clinical and radiographic outcomes in patients treated with the Acumed anatomic radial head press-fit system for trauma.

Methods

Clinical and radiographic assessment of RHAs undertaken for trauma with minimum 2-year follow-up.

Results

16 consecutive patients, mean age 53 (21–82) and 66 month ± 27 (26–122) clinical follow-up were included. There were marked radiographic changes with 11/16 showing periprosthetic lucent lines and 13/16 showing subcollar osteolysis. Radiographic changes occurred early post-surgery. Stem loosening was associated with larger cantilever quotients (0.47 vs 0.38, p  =  0.004). Overall survivability was 81.2%, with 3 RHAs removed. Clinical outcomes for the retained RHAs were acceptable with mean flexion 134°, extension deficit of 10°, pronation of 82°, and supination of 73°. Mean VAS scores were 8.5 ±  14.4, QuickDASH 13.8 ±  18.9, Mayo Elbow Performance Scores were 91.5 ±  12.5 with no poor scores.

Conclusion

Mid-term clinical functional outcomes following the Acumed anatomic RHA are acceptable in most cases. However, in view of the extensive periprosthetic lucencies and surgical removal due to loosening, patients should be cautioned when consented for implantation of the prosthesis, especially if a large collar is anticipated.

Introduction

Up to one third of elbow fractures involve the radial head and a third of these are associated with significant ligamentous and/or other osseous injury and instability.^1,2 Radial head arthroplasty (RHA) is utilised in cases of highly comminuted fractures where reconstruction is not feasible and in the setting of significant elbow instability. ³

Various types of RHA options exist but there is limited high quality evidence to support one type of radial head prosthesis over another.^4,5 Rapid and marked radiographic changes may be observed following RHA including heterotopic ossification (HO), loosening, and radiocapitellar degeneration ⁶ which may be more common with press fit monopolar prostheses. ⁷ The Acumed anatomic Radial Head (Acumed, Hillsboro, OR, USA) is a monopolar, press fit device designed to recreate native regional anatomy. ⁸ Reported short term outcomes of this prosthesis have been variable, with some studies reporting high loosening and osteolysis rates.^9–13 However, there is paucity in the literature with regards to longer term outcomes.^9–11,13 The aim of this study was to assess the mid-term clinical and radiographic outcomes in patients treated with this prosthesis for elbow trauma.

Materials & methods

Patients

All patients who had RHA for trauma at a single institution in the UK, were identified using theatre databases and in house computer software. Clinical and operative notes were reviewed. Preoperative radiological imaging (radiographs and/or CT) were used to determine the injury pattern and classify the radial head fracture (using the Charalambous classification ³ and the Johnston modification of Mason's classification system. ¹⁴ ) Local approval for running this study was obtained from our institution's Research and Development department (as a service evaluation).

Surgical technique

All patients were operated on between January 2009-July 2015 using the Acumed Anatomic Radial Head system (Acumed, Hillsboro, OR, USA) which is routinely used at our institution. This is a modular implant with several head, collar and stem size options. It has a grit blasted porous coated titanium stem designed to promote osseous integration in a press-fit design and an anatomic cobalt chromium head designed to replicate native anatomy. ¹² Although several sizing options exist, the head length is fixed at 10 mm and stem length at 25 mm.

Surgery was performed by senior surgeons at our institution experienced in dealing with complex elbow trauma. Surgery was performed using either the Kaplan's approach ¹⁵ or the Wrightington posterolateral approach. ¹⁶ The choice of the surgical approach was based on the surgeon's preference and experience in using that approach. Following fracture exposure, loose radial head fragments were excised and used to estimate sizing of the native head. The medullary canal was sequentially rasped until a tight canal fit was obtained. A trial prosthesis was inserted and screened under fluoroscopy intraoperatively to confirm adequate sizing and elbow stability. The prosthesis was inserted ensuring a press-fit as per the manufacturer's instructions. In patients requiring concurrent ligament reattachment, TWINFIX^TM DURABRAID (Smith & Nephew, London, UK) and Mitek LUPINE^TM BR Anchor (Raynham, MA, USA) suture anchors were used.

Elbow immobilization ranged from 10 days-4 weeks with above elbow plaster. Patients were reviewed in clinic according to the discretion of their attending physician. Following immobilisation patients were mobilised under the guidance of physiotherapy, avoiding heavy lifting activities for up to 12 weeks.

Patients were contacted to attend our institution for clinical assessment and collation of patient recorded outcome measures for the purposes of this evaluation. At final clinical follow-up, range of motion was measured using a goniometer in the affected and contralateral elbow. Functional assessment was performed using the Mayo Elbow Score (MEPS) and Disability of the Arm, Shoulder and Hand (DASH) scores. ¹⁷ Elbow pain was recorded using the Visual Analogue Scale (VAS) with patients marking their pain on a scale ranging from 0 no pain to 100 severe pain. Clinical assessments were carried out by healthcare professionals not involved in the surgical treatment of the patients.

Radiographic assessment

Initial radiographs were taken at the first post-operative follow-up and subsequent clinic visits as well as when patients were assessed for this study if patients had not had recent radiographs. Serial post-operative radiographs were reviewed independently by an experienced musculoskeletal radiologist, who had not been involved in the patients’ care. At least 2 postoperative radiographs were assessed to determine the presence of and progressive changes in the following parameters:

• Periprosthetic Lucencies: these were classed as discrete lines surrounding the stem recorded with reference to the affected Zones on both the AP and lateral x-ray images adopted from Popovic et al. ¹⁸ in Figure 1. The maximal perpendicular distance of these lucencies from the stem to the endosteal bone was measured and recorded in each zone. Stems were considered well fixed if there were lucent lines in ≤1 zones in either AP/lat or if lucencies were visible only in one AP or lateral image.

• Subcollar Osteolysis: this was measured in anterior/posterior (AP) and lateral images as the maximal perpendicular distance between the collar of the prosthesis and nearest point of bony contact with the stem (Figure 2). The 10 mm head length of the prosthesis was used to calibrate radiographic measurements.

• Capitellar Erosion/Osteopenia: these were recorded as present or absent and the appearance of the capitellum described qualitatively noting the appearance of osteophytes, surface flattening of the capitellum and/or proximal migration of the prosthesis.

• Cantilever Quotient: this was measured by recording the total length of the prosthesis on a lateral radiograph, divided by the exposed portion of the implant not contained within the bone, adopted from Shukla et al. ¹⁹

• Degenerative changes in ulnohumeral joint: these were recorded according to the system devised by Broberg and Morrey ²⁰ ranging from Grade 0 (none), Grade I (mild), Grade II (moderate), and Grade III (severe).

• HO: this was recorded in relation to the RHA as anterior, anterolateral, lateral or diffuse, and described as bridging or non-bridging.

Figure 1.

Radiographs demonstrating the zones of radiographic changes used for this study in accordance with Popovic et al. ¹⁸ along with incidence of periprosthetic lucency by zone at radiographic follow-up. Case 5 (Final AP and lateral images at 83 month follow-up). Lucent lines surrounding the stem in all zones with a loose stem at final follow up.

Figure 2.

Radiographs demonstrating measurements of subcollar osteolysis in AP and lat views used for this study. Image demonstrates a typical stress shielding response with subcollar osteolysis with relative preservation of the endosteal cortex relative to the periosteal cortex. Case 10 Final follow up (45 months), note appearance of HO.

Statistical analysis

Results were collated and analysed using Microsoft excel and SPSS (version 20.0, IBM) software. Normality tests for data distribution were performed. Non-Parametric testing was carried out using the Mann-Whitney U-Test and parametric comparisons using the Independent Sample T-Test Fisher's exact test was used to ascertain differences between nominal variable groups in subgroup analysis. A two-tailed 5% significance level was utilised.

Results

16 consecutive acute trauma cases treated with the Acumed Anatomic Radial Head System between 2009–2015 were included. There were in 11 females and 5 males with a mean age of 53 years ± 13 (range 21–82), with 8/16 cases involving the dominant arm. Injuries involving the radial head were categorised as per the Charalambous classification ³ with 12 displaced complete articular fractures, 2 displaced comminuted partial articular fractures, and 2 displaced radial neck fractures (Table 1). 7 patients had Mason ¹⁴ type III fractures and 9 had type IV fractures with associated elbow instability according to the Johnston modification of the Mason classification system. ¹⁴

Table 1.

Patient demographics, injury characteristics and acumed anatomic radial head system implant sizes.

Case	Sex	Age	Dominant Arm	Mason Class	Charalambous Class + Associated Injury	RHA Head Size	RHA Collar Size	RHA Stem Size
1	F	62	No	4	Displaced Radial Neck + Elbow Dislocation	22	2	7
2	F	56	No	4	Displaced Comminuted, Partial Articular + Elbow Dislocation + Olecranon Fracture	20	4	6
3	F	52	Yes	3	Displaced, Complete Articular	20	4	8
4	M	58	Yes	3	Displaced, Complete Articular	26	2	10
5	F	50	No	4	Displaced, Complete Articular + Elbow Dislocation + Coronoid Fracture	22	0	7
6	F	56	No	3	Displaced, Complete Articular	22	4	7
7	F	57	Yes	4	Displaced, Complete Articular + Elbow Dislocation + Coronoid Fracture	20	4	8
8	M	21	Yes	4	Displaced Complete Articular + Coronoid Fracture	24	0	8
9	F	60	Yes	3	Displaced, Complete Articular	22	0	7
10	M	33	Yes	3	Displaced, Complete Articular + Coronoid + Olecranon Fracture	26	6	8
11	F	61	No	4	Displaced Radial Neck + Elbow Dislocation	26	4	10
12	F	82	No	4	Displaced Comminuted, Partial Articular + Elbow Dislocation + Coronoid + Olecranon Fracture	22	0	9
13	M	55	Yes	3	Displaced, Complete Articular + Olecranon Fracture	26	8	10
14	F	50	Yes	4	Displaced, Complete Articular + Elbow Dislocation	24	2	7
15	F	49	No	4	Displaced, Complete Articular + Elbow Dislocation	24	2	8
16	M	48	No	3	Displaced Complete Articular	26	4	9

For implanted RHA, head sizes ranged from 22–26, collar sizes from 0–8, and stem sizes from 6–10, (Table 1). Olecranon fractures of the Monteggia-type were treated with plate osteosynthesis in 3 and tension band wiring in 1 case. Ligament avulsions were reattached using suture anchors in 10 cases. There were no cases of interosseous membrane disruption. The Wrightington approach ¹⁶ was performed in 8 and the Kaplan's approach in 7 cases. The choice of the surgical approach was based on the surgeon's preference and experience in using that approach. All cases were operated on within 6 weeks of injury.

At mean clinical follow-up of 66 months ( + /-27, range 26–122), the survival of the prosthesis was 81.2% (13/16) with 3 cases undergoing removal of the RHA (2 due to loose painful stems and 1 for stiffness). Of the cases removed, 1 had removal of the prosthesis for painful capitellar erosion and stem loosening at 18 months but continued to have ongoing pain at final follow-up and was referred for chronic pain management. One case had painful RHA loosening along with capitellar erosion and ulnohumeral arthritis at 71 months and was revised to a linked total elbow arthroplasty with resolution of pain and improvement in motion at last follow-up. A further patient had a well-fixed RHA removed as part of open arthrolysis to address stiffness with HO. Cases where the prosthesis was removed had no evidence of infection with normal inflammatory markers. There were no post-operative infections or neurovascular complications following surgery in this series.

Functional outcomes

13 cases with a retained prosthesis had full outcomes measured at final follow-up (Table 2). One patient was unable to attend hospital for follow-up, with outcome scores completed over the telephone.

Table 2.

Final radiographic changes and outcome scores in cases.

Case	Radiographic Follow-up (months)	Cantilever Quotient	Time to Visible Change (months)	Progressive Changes	Capitellar Erosion	Degenerative Changes (Broberg-Morrey)	HO	Max Subcollar Osteolysis (mm)	Zone Lucencies AP	Zone Lucencies lat	Stem Well Fixed	Secondary Surgery	Range of Movement (degrees)				VAS	Quick DASH	MEPS
													Flex	Ext Loss	Pro	Sup
1	91	0.36	2	Yes	No	Grade I	No	13.0	4	4	Yes	No	140	8	90	90	5	0	100
2	51	0.46	2	No	No	Grade II	Yes	0	0	0	Yes	Yes** RHA removed. Arthrolysis 71 months	-	-	-	-	-	-	-
3	122	0.34	82	Yes	No	Nil	No	7.7	0	0	Yes	No	140	18	90	90	10	0	95
4	81	0.40	1	Yes	Yes	Nil	Yes	0	1,2,3,4,5,6,7	4,5,6,7	No	No	120	15	90	80	0	0	100
5	83	0.52	2	Yes	No	Nil	No	5.2	1,2,3,45,6	1,2,3,4,5,6,7	No	No	135	0	90	90	20	27.3	85
6	46	0.36	2	Yes	No	Nil	No	7.6	4	0	Yes	No	145	5	80	80	0	9.1	100
7	26	0.40	26	N/A	No	Nil	No	3.8	0	7	Yes	No	140	15	85	80	0	9.1	100
8	18	0.48	1	Yes	Yes	Nil	No	3.9	1,2,3,4,5,6,7	1,2,3,4,5,6,7	No	Yes RHA removed 18 months	-	-	-	-	-	-	-
9	56	0.37	2	Yes	No	Nil	No	1.9	0	0	Yes	No	-	-	-	-	20	4.5	65
10	45	0.39	4	Yes	No	Nil	Yes	7.4	0	0	Yes	No	130	20	90	75	5	27.3	80
11	51	0.43	4	Yes	No	Nil	No	1.2*	0	0	Yes	No	130	10	30	30	0	4.5	100
12	85	0.30	85	No	No	Grade I	Yes	0	0	3,4*	Yes	No	130	10	80	45	50	54.5	70
13	65	0.49	60	Yes	Yes	Grade II	Yes	2.0	2,3,4,5	2,3,4,5	No	Yes 70 months, RHA removed TER	-	-	-	-	-	-	-
14	56	0.35	3	Yes	No	Nil	No	6.6	0	3,4*	Yes	No	140	0	90	90	0	2.3
15	30	0.43	3	Yes	No	Grade I	Yes	7.2	0	1,3*	Yes	No	130	10	80	45	0	40.9	95
16	51	0.39	1	Yes	No	Nil	No	1.7	4,5,6,7*	0	Yes	No	130	5	85	85	1	0	100

Case 9 was unable to attend follow-up for formal ROM assessment, telephone consult only.

Case 13 moved out of area and followed up in local hospital where RHA removed converted to linked Total Elbow Replacement (TER), no outcome measures, preoperative radiographs transferred for assessment.

Case 7 had only 1 postoperative x-ray taken 27 months following surgery, therefore unable to assess progressive x-ray changes. **Case 2 was revised at 71 months for marked stiffness with an otherwise well-fixed stem.

*changes visible in one view only 1 plane.

Mean range of motion was flexion of 134° ± 7.0° (range 120–145°), extension deficit 10° ± 9.7° (range 0–20), pronation 82° ± 16.8° (range 30–90°), supination 73° ± 21.0° (range 30–90°). No patient had an extension deficit greater than 20°. Mean MEPS was 91.5 ± 12.5 (range 65–100) with 11/13 cases recording good to excellent scores with 2 fair scores. Mean QuickDash score was 13.8 ± 18.9 (range 0–54.5), and mean VAS score was 8.5 ± 14.4 (range 0–50). Subgroup analysis revealed no statistically significant relation between clinical outcomes and age, sex or fracture patterns. Three patients had a change of occupation following injury.

Radiographic outcomes

Serial post-operative radiographs were reviewed for all, one case had only a single post-operative radiograph at final follow-up, and thus progression could not be assessed. Mean radiographic follow-up for cases retaining prosthesis was 60 months ± 27 (range 18–122) (Table 2).

All 16 patients had visible radiographic changes as described below:

• Periprosthetic lucency: At final radiographic follow-up, lucent lines were detected in 11/16 cases, with mean periprosthetic lucency for cases measuring 0.86 mm ± 0.71 mm (range 0.3–2.12 mm). Only 4 cases had lucent lines surrounding the stem in ≥2 zones visible in both AP and lateral images and were thus considered loose, the 12 remaining cases were well fixed (Table 2). Two of the loose stems were removed. Incidence of periprosthetic lucencies by zone are shown in Figure 2. Lucencies were most commonly visible in zone 4, 38% (6/16) cases on AP and 44% (7/16) on lateral images with mean lucency per zone ranging from 1.20 mm in zone 7 on lateral images to 3.26 mm in zone 4 on lateral images. Lucencies visible around the stem often occurred early within the first 3 months in and progressed within the first 12–18 months (Table 3, Figure 3). There was no association between the appearance of lucent lines with elbow instability injuries (p-0.859) or with cases requiring ligamentous repair (p-0.2335).

• Subcollar osteolysis: Thirteen of 16 cases had subcollar osteolysis (Figure 1) with a mean extent of 5.3 mm ± 3.4 (range 1.2–13 mm). Of the 12 well fixed cases 75% (9/12) had evidence of subcollar osteolysis. In 7 cases with ≥1 radiograph in the first year following surgery, subcollar osteolysis occurred on average at 2.4 months ± 1.0. Three cases had >4 mm of subcollar osteolysis measured in either AP or lateral images within the first 6 months.

• Cantilever quotient: Eight (50%) cases had cantilever quotients ≥0.4 in this series. Mean cantilever quotient was 0.40 ±  0.06 (range 0.30–0.52). Larger cantilever quotients were associated with loose stems, 0.47 vs 0.38 (Mann-Whitney-U test p-0.004).

• HO – this was seen in 6 of 16 (38%) cases and was located anteriorly in 4, and anterolaterally in 2. HO was non-bridging in 5 and bridging in 1 case. Appearance of HO was associated with a significant reduction in supination (Mann Whitney U-Test, 50° vs 80°, p-0.040) but had no significant effect on other recorded outcome variables.

• Capitellar erosion: Three (19%) cases had visible capitellar erosion, all 3 of these cases had associated periprosthetic loosening with 2 of these requiring revision surgery.

• Degenerative changes – 5 cases had visible osteoarthritic changes in the ulnohumeral joint at final follow-up, with 3 Broberg-Morrey Grade I and 2 Grade II changes. Both cases with grade II changes were revised.

Table 3.

Lucent line subgroup, radiographic changes with time.

Case	RHA Removed	Radiographic Follow-Up	Cantilever Quotient	Mean Lucency <3months (mm)	Zones affected	Mean Lucency 12–18months (mm)	Zones affected	Mean Lucency Final Follow-Up (5 years + ) (mm)	Zones affected
4	No	81	0.4	1.5 ±  0.76	14,5	2.2 ±  0.69	1,2,4, 56,7	2.3 ±  0.19	1,2,3,4, 56,7
5	No	83	0.52	1.37 ±  0.43	1,2,3, 46,7	2.1 ±  0.65	1,2,3, 4,5,6,7	2.3 ±  0.79	1,2,3,4, 56,7
8	Yes 18 months	18	0.43	1.48 ±  0.45	1,2,3, 4,5,6,7	2.04 ±  1.54	1,2,3, 4,5,6,7	N/A	-
13	Yes 70 months Revised to TER	65	0.49	0	0	No available radiographs	-	2.88 ±  1.05	2,3,4,5

Figure 3.

Case 5. lateral (upper row) and AP (lower row) radiographs at 3 months, 12 months, and final follow-up (83 months). Progressive and early changes with lucent lines surrounding the stem in all zones with a loose stem at final follow-up.

Discussion

Fractures of the radial head not suitable for internal fixation can be treated with excision or RHA. RHA is preferred for restoration of elbow stability and kinematics^5,21 especially when there is associated ligamentous disruption.^1,21 The Acumed Anatomic Radial Head system is a modular implant with a grit blasted porous coated, press-fit titanium stem and an anatomic cobalt chromium head designed to replicate native anatomy. ⁸ The anatomic design has been shown to evenly distribute radiocapitellar contact pressures through joint conformity ²² with minimal changes in radiocapitellar joint pressure and interosseus membrane tension when inserted in an optimal position. ²³ Furthermore, the anatomically designed articular dish aims to promote radiocapitellar stability resisting translational forces and subluxation. ²³ Conforming to the normal radiocapitellar shape using anatomic implants is hypothesised to improve tracking and congruency, reducing joint contact pressures and thus capitellar wear.^22,23

We report acceptable clinical mid-term clinical outcomes in most patients with the Acumed monopolar press-fit stem in accord with previous reports of the same implant.^5,9–13 Previous studies have urged caution with use of monopolar press-fit designs over concerns with radiological loosening.^24–26 However, although radiographic changes have been previously observed with this prosthesis, their impact on longer term clinical outcomes have been less clearly defined.^9–13 Sallakh et al. ¹² at mean follow-up 42 months, reported all 12 patients with good-excellent clinical outcomes with no significant radiographic changes with regard to proximal migration, radiocapitellar arthritis or capitellar osteopenia, and no cases of prosthesis removal. At 30 months mean follow-up, Levy et al. ¹³ showed that 6/15 cases with radiographic follow up had loose stems with 2 requiring removal. They suggested a possible association between loosening with injuries requiring ligamentous repair. Mou et al. ¹¹ reported good mid-term clinical results with no radiological issues in their series of 12 cases. However, substantial subcollar osteolysis was visible on radiographs included in their published article. In our series, 3/16 (18.8%) cases underwent secondary revision surgery at 18, 70 and 71 months post-surgery with 2 cases revised for painful loosening at 18 and 70 months. Rodriguez-Quintana et al. ²⁴ prospectively compared the Acumed monopolar press fit stem with a smooth stemmed implant and reported that at 2 years, lucent lines appeared in 5/14 patients in the Acumed monopolar press fit group with 2 of these requiring removal. ²³ Similarly, Cloke et al. ²⁵ found 6/17 patients with a loose Acumed monopolar prosthesis and 5/17 requiring removal for symptomatic loosening within 24 months of surgery (Table 4). Our series provides the longest mean clinical and radiological follow-up for the Acumed prosthesis currently in the literature, with an overall survivorship of the prosthesis of 81.2% at 66 months mean follow-up.

Table 4.

Comparison of studies for acumed anatomic radial head system (acumed, hillsboro, OR, USA).

Author	Year of Publication	Indication	No of cases	Follow-Up (months)	RHA removal rate % (cases removed/total cases)
Cloke et al. 25	2010	Acute Trauma and Revision of loose cemented implant	17	<24	29% (5/17)
Berschback et al. 7	2013	Acute and Chronic Trauma	13	29.4	7.7% (1/13)
El Sallakh et al. 12	2013	Acute Trauma	12	42	0%
Mou et al. 11	2015	Acute Trauma	12	60.8	Not stated
Levy et al. 13	2016	Acute Trauma	15	26	13.3% (2/15)
Rodriguez-Quintana et al. 24	2017	Acute Trauma	14	24	14.3% (2/14)
Tarallo et al. 10	2017	Acute Trauma	31	30	Not stated
Samra et al. (current study)	2020	Acute Trauma	16	66	18.8% (3/16)

Loosening of a press fit prosthesis may be related to an initial failure of osseointegration. Stem movement <100 µm is needed for osseointegration to take place^9,19 with micromotion >150 µm leading to the development of fibrous tissue at the stem-bone interface. ²⁷ Shukla et al. demonstrated that stability of a radial head prosthesis is affected by the length of the stem outside the bony canal relative to the total implant length, described as the cantilever quotient, and showed that cantilever quotients >0.4 were deemed at risk of excessive micromotion >100 µm compromising stability. ¹⁹ The Acumed monopolar implant is a modular implant with a fixed width 10 mm head along with a 25 mm stem. Large collars ( + 8) or improper sizing with a greater amount of exposed stem can produce cantilever quotients of ≥0.4 and thus may place the prosthesis at risk of excessive micromotion due to cantilever bending forces, inhibiting osseointegration. Moon et al. 2019 ²⁸ assessed cadaveric implant fixation with varying neck cuts and stem lengths and demonstrated that stem lengths ≥25 mm had significantly reduced micromotion with both 10 mm and 25 mm neck cuts. In our series, larger cantilever quotients were associated with stem loosening, which may be related to inadequate osteointegration (although a definite cause effect relation cannot be established).

Moon et al. ²⁹ demonstrated that appropriate stem diameter is crucial to obtaining greater stem stability and limiting micromotion, with submaximally sized stems impairing osseointegration. It is of note that in our series, the median stem diameter for cases that developed loosening was 9 mm (7.25–10) and for well-fixed stems 8 mm (7–8.75), with 2/4 loose cases having had a 10 mm stem. Berschback et al. ⁷ suggested that movement of a grit blasted stem may cause erosion of endosteal bone compared with the permissible movement that occurs in some smooth stemmed implants. Furthermore, grit blasted titanium surfaces may exhibit inferior osseointegration compared with Hydroxyapatite (HA) coated orthopaedic implants.^30,31 Series comparing smooth stemmed radial head implants with monopolar press fit designs have suggested greater osteolysis in the latter.^24,32

Subcollar osteolysis was widely observed in this series and was typical of stress shielding. Thirteen of 16 cases had evidence of subcollar osteolysis, with 9 of these occurring in well-fixed cases (Table 2). Stress shielding in a radial head prosthesis occurs when load is transferred across the radiocapitellar joint preferentially through the prosthesis, bypassing the proximal periosteal bone. ³³ The balance of load transfer and thus amount of stress shielding will depend on the interface bonding characteristics and relative stem stiffness in relation to that of bone. ³³ Implant stiffness depends on the elastic modulus and diameter of the stem. The titanium stem of the Acumed prosthesis has an elastic modulus close to that of cortical bone theoretically mitigating stress shielding. Less stiff stems have been shown to result in more uniform load transfer and reduced stress shielding in hip arthroplasty ³⁴ though the impact on RHA has not been demonstrated. Nevertheless, stress shielding with press-fit stems in RHA may be of low clinical significance.^9,13 In our series, those patients with a visible stress shielding had good clinical outcomes and none required revision. Similar to previous reports, resorption often occurred early and progressed slowly. Five cases of subcollar osteolysis >7 mm were seen in well-fixed stems, hence we consider subcollar resorption to be reflective of adequate osseointegration at mid-term. However, there is concern that progressive subcollar bone loss may eventually lead to insufficient mechanical support of the short-stem prosthesis and the long-term consequences are not yet known. Further long-term follow-up is required to ascertain the true clinical significance.

The 37.5% incidence of HO in this series is comparable with previous reports.^10,24 A recent meta-analysis has shown that half of all removed radial head prostheses are removed for HO. ³⁵ In one of our cases the prosthesis was removed as part of open arthrolysis for stiffness. Although HO cases showed a significant reduction in supination, there was no effect on outcome scores. Prophylaxis against HO was not used in our series. Some have advocated the routine use of indomethacin, ³⁶ however no clear consensus yet exists in prophylaxis for complex elbow trauma, with recent evidence demonstrating no difference in HO in cohorts of patients receiving prophylaxis and those without. ³⁷

Limitations of our study include the retrospective design and small sample size, though this reflects the overall low incidence of radial head injuries requiring RHA for trauma and is comparable with most published series. A recent epidemiological study for traumatic radial head replacement demonstrated an incidence of 0.65/100,000/year in Scotland ³⁸ comparable with our sample size over the study period. We could not obtain baseline outcome measures for patients in this series due to its retrospective design. Furthermore, clinical outcome measures were not taken at set time points during the clinical follow-up period. However, we believe the outcomes recorded at final follow-up using multiple validated patient recorded outcome measures in this series gave an accurate representation of the mid-term results for this particular prosthesis. We found no association between injury complexity and clinical or radiological outcomes in our series, this is consistent with recent evidence which has demonstrated no effect of concomitant elbow injuries on long term clinical or radiological outcomes for traumatic radial head arthroplasty in a large prospective study. ³⁹ However, variability in post-operative rehabilitation protocols in this series may have influenced the final outcome measures and range of motion. Early mobilization of simple radial head fractures has been shown to improve patient related outcome measures ⁴⁰ though for surgically treated complex injuries the evidence is less clearly defined. Additionally, assessment of lucencies is limited by radiographic projection and quality and bony changes may be underestimated as significant bone loss must occur to be visualised on radiographs. CT scanning may better allow visualisation of peri-prosthetic lucencies but this is not utilised in routine clinical practise. We attempted to improve assessment of lucencies by reviewing multiple radiographs over time in both AP and lateral views. Radiographs in this series were obtained only for the ipsilateral elbow for radiographic follow-up, as a result any assessment of the DRUJ and ulna variance at the wrist could not be made.

Conclusion

Our study provides the longest follow-up in the literature for the Acumed anatomic monopolar press-fit prosthesis used in elbow trauma. Despite acceptable clinical outcomes in most cases, in view of the appearance of extensive lucent lines and the prostheses removed due to loosening in this study, patients should be cautioned when consented for implantation with this prosthesis. Patients with large cantilever quotients should be observed closely for loosening.