Introduction

The imaging findings of periprosthetic soft tissue lesions (or pseudotumours) have been typically defined in the context of newer, second-generation large-diameter metal-on-metal (MoM) hip arthroplasty. These adverse findings have been generally attributed to abnormal generation of metal particles at the articular interface and thus have not been commonly considered in the context of other bearing couples. However, similar lesions have recently been reported in the orthopaedic literature in cases of both primary and revision metal-on-polyethylene hip arthroplasty. While uncommon, emerging evidence suggests that cases such as these are potentially more prevalent that first thought and may be the result of corrosion at the modular head-neck and neck-stem interfaces, termed ‘trunnionosis’.

We report a case of delayed diagnosis of a periprosthetic soft tissue lesion related to a metal-on-polyethylene total hip arthroplasty (THA), postoperatively attributed to trunnionosis.

Case report

A 43-year-old male was referred to the orthopaedic oncology unit approximately 15 years following a right primary THA [titanium stem; a cobalt chromium (CoCr) femoral head and polyethylene acetabular liner in a titanium shell] after a CT demonstrated fluid collections and a mass-like soft tissue lesion encircling the prosthesis. He described a 4-month history of increasing stiffness, pain and reduced walking endurance. Core biopsy of the lesion demonstrated blood and necrotic material without neoplastic or otherwise atypical cells. No evidence of infection was elicited from microbiological analysis.

In the absence of infection or neoplastic disease, a reaction to polyethylene debris was considered to be the likely diagnosis. Even though the radiograph at this time did not show signs of periprosthetic osteolysis or eccentric femoral head positioning to suggest polyethylene wear (Fig. 1), ‘backside’ wear of the nonarticular surface of modular polyethylene liners can be a potential source of wear debris that does not result in radiographically apparent implant changes [1]. Given this possibility, the patient was offered and underwent revision surgery.

Fig. 1
figure 1

Radiograph taken approximately 15 years following index right hip arthroplasty demonstrated well-seated components with heterotopic ossification but no evidence of osteolysis or scalloping of the periprosthetic bone. Findings were unchanged compared to numerous prior postoperative follow-up radiographs (not shown)

Full size image

Both the stem and acetabular cup were found to be well fixed intraoperatively and thus only the femoral head and liner were revised to new CoCr and polyethylene components, respectively. A large haematoma (700 cm3) and granuloma (250 cm3) were removed. Pathological analysis was limited to gross assessment of excised tissue, with no significant findings, as an adverse local tissue reaction (ALTR) was not suspected at the time.

Although he initially did well, the postoperative course was complicated by recurrent wound drainage and ongoing swelling over several months. Radiographic evaluation was unremarkable (Fig. 2), and workup for infection remained negative. Ongoing pain and an enlarging groin mass prompted an MRI, which demonstrated a large complex cystic and solid mass lesion almost circumferentially encasing the right hip joint extending to the skin at the incision site, with similar size and appearance to that seen prior to the first revision procedure. It demonstrated a low to intermediate signal peripheral rim, with intermediate T2 signal soft tissue components (Fig. 3a). Cystic components demonstrated high T2 and intermediate to low T1 signal likely representing an element of proteinaceous fluid (Fig. 3b). Small peripheral foci of calcification were also present (Fig. 3c).

Fig. 2
figure 2

Radiograph taken approximately 1 year following excision of heterotopic ossification and revision of the femoral head and acetabular liner demonstrates stable components and no interval bony changes

Full size image
Fig. 3
figure 3

a Fourteen months following the first revision procedure, T2-weighted fat-suppressed axial acquisition demonstrates a large periprosthetic complex solid (white star) and cystic mass with peripheral low signal rim (white arrowhead) extending into the anterior compartment of the thigh. b T1-weighted axial acquisition demonstrates a large periprosthetic complex solid (white star) and cystic mass with peripheral low signal rim (white arrowhead) extending into the anterior compartment of the thigh. Cystic components demonstrate low to intermediate signal suggesting a component of proteinaceous fluid. c Contrast-enhanced axial CT acquisition demonstrates a large periprosthetic complex solid with a peripherally enhancing rim (white arrow head) extending into the anterior compartment of the thigh displacing the superficial femoral artery a and vein v. Small punctate foci of calcification are present (white open arrowhead)

Full size image

When the possibility of an ALTR to metal debris was raised, testing of whole blood metal ion levels was performed, revealing cobalt levels of 80.10 nmol/l (normal: <6 nmol/l) and chromium levels of 11.15 nmol/l (normal: <2 nmol/l) (Table 1).

Table 1 Frequently used terminology in the context of the generation of abnormal metal debris from hip arthroplasties and the associated radiographic findings a
Full size table

The patient underwent a second revision procedure. A large pseudotumour was excised. The stem and acetabular shell were again found to be well-fixed. Isolated exchange of the femoral head and acetabular liner was performed, with placement of a new ceramic femoral head component with an integrated titanium femoral neck sleeve (Fig. 4). No CoCr-containing components remained in situ following this revision procedure. Examination of the taper junction of the explanted CoCr femoral head revealed dark staining consistent with corrosion damage (Fig. 5). The patient was doing well at 2-month follow-up, with no evidence of recurrent wound drainage or hip swelling.

Fig. 4
figure 4

Postoperative radiograph taken 2 months following the second revision procedure, with excision of a large pseudotumour and femoral head exchange for a ceramic component with integrated titanium sleeve. No CoCr components remained in situ

Full size image
Fig. 5
figure 5

Black deposits consistent with corrosion products are visible on the taper of the explanted cobalt-chromium femoral head (white arrowheads)

Full size image

Discussion

Since development of the now historic monoblock design (Fig. 6), modularity of the femoral head-neck junction has become a ubiquitous feature of contemporary total hip arthroplasty components, allowing intraoperative optimisation of implant geometry (Fig. 7). Modularity at the neck-stem junction is a more recent design feature provided by various vendors, affording the surgeon additional versatility when attempting to restore normal biomechanical function (Fig. 8). In both cases, the interface consists of a large surface area friction-fit taper, designed to prevent motion once assembled (Fig. 9). Although current evidence is limited, it appears as though in some cases the forces across the junction may be sufficient to induce micromotion at the interface. This motion may ultimately result in the generation of metal particles and/or ions from the malfunctioning taper, in some cases producing ALTRs similar to those described with MoM hip arthroplasty, irrespective of the specific bearing couple used [2–5]. Given the high rate of poor postoperative outcomes reported following revision of cases with pseudotumour (up to 50% incidence of major complications in one cohort) [6], emphasis is now placed on early identification and surgical management of ALTRs [7], particularly in cases of soft tissue compromise.

Fig. 6
figure 6

The now historic Charnley monoblock total hip arthroplasty comprised a single cobalt-chrome femoral component a and a single all-polyethylene acetabular liner b

Full size image
Fig. 7
figure 7

Standard contemporary primary hip arthroplasty prostheses generally include a modular head-neck junction and typically consist of a femoral stem a, modular head b, acetabular liner c and acetabular shell d

Full size image
Fig. 8
figure 8

Dual-modular primary total hip arthroplasty prostheses include modular stem-neck and neck-head junctions, and typically consist of a femoral stem a, modular neck b, modular head c, acetabular liner d and acetabular shell e

Full size image
Fig. 9
figure 9

Coronal cross section through an assembled dual-modular primary hip arthroplasty illustrates the friction-fit modular neck-stem (black arrowheads) and head-neck (white arrowheads) interfaces. Both interfaces are potential sources of corrosion

Full size image

The release of metal ions from wear and corrosion can have both local and, in rare cases, systemic sequalae [8]. The cause of local reactions is incompletely understood. However, two overlapping histological patterns have been described [9]. Metal reactivity, characterised by high-bearing wear, elicits a predominately foreign body (macrophage/histiocytic) reaction to metal debris [10]. These particles are cytotoxic to the phagocytic cells and ultimately result in necrosis. Metal sensitivity is characterised by a predominately lymphocytic (immunological) reaction, thought to represent a type IV delayed hypersensitivity reaction to metal ion-protein complexes (haptens) [4], referred to as aseptic lymphocytic vasculitis-associated lesions (ALVAL). These soft tissue reactions appear to occur exclusively with cobalt-based alloys [11]. While exceptionally rare, a few cases of suspected systemic cobalt toxicity secondary to catastrophic wear-related failure of THA components—termed ‘arthroprosthetic cobaltism’—have been described, including one with fatal outcome [12–14]. Systemic effects of cobalt toxicity can include neurological (fatigue, weakness, poor coordination, cognitive dysfunction, depression, vertigo, visual and hearing impairment, and peripheral neuropathy), haematological (polycythaemia), endocrine (hypothyroidism) and cardiac (arrhythmias and cardiomyopathy) findings.

Various factors have been implicated in corrosion and wear at the modular interfaces including larger diameter heads, specific taper designs, inadequate manufacturing tolerances, contamination of the taper at the time of intraoperative assembly, the use of CoCr femoral stems and mixing of components from different manufacturers [2, 15, 16]. However, no consistent predictors have been identified to date, and ALTRs have been reported with a wide range of primary hip implant designs, including metal-on-polyethylene [2], ceramic-on-polyethylene [17], ceramic-on-metal [18] and metal-on-metal THAs [19], as well as metal-on-metal hip resurfacing arthroplasties [6]. Many contemporary revision hip systems include multiple taper junctions (specifically, between the head and neck, as well as between the body and stem), although their designs differ from those found in modular primary hip components. There have been several reports of catastrophic failure of modular revision hip components at the body-stem junction secondary to a high bending moment in patients with poor proximal femoral bone stock, with eventual metal fatigue failure [20, 21]. This failure mode has been subsequently addressed by design modifications to strengthen these junctions [21]. In vitro and retrieval studies have also identified corrosion and wear at modular junctions of revision hip arthroplasty components [22], although to the best of our knowledge, there have been no reported cases to date of this being associated with an ALTR. Thus, while ALTRs secondary to corrosion of modular junctions in revision hip arthroplasty components are theoretically possible, the relatively low burden of revision procedures combined with the paucity of in vivo reports suggests that such cases are very rare.

Retrieval analyses comparing explanted resurfacing prostheses (which have no trunnion) with their MoM total joint arthroplasty counterparts have enabled comparison of the wear debris volumes and thus potentially determine the contribution of modular interface. Multiple groups have demonstrated higher metal ion levels in patients with modular prostheses when compared to those with otherwise similar resurfacing prostheses [23–25]. These findings support the theory that taper corrosion contributes significantly to the release of potential biologically active metal byproducts. Langdon et al. postulated that abnormal wear at the trunnion may result in significantly worse soft tissue reactions than those elicited from bearing surfaces [26]. This may be explained by differences in the types of material generated; wear at the trunnion results primarily in in-situ corrosion with concomitant cobalt ion release, whereas an abnormal MoM-bearing function may be more likely to produce nanometer-sized CoCr alloy particles.

In this case, the patient appears to have had an ALTR related to corrosion at the taper junction of a titanium stem and CoCr femoral head, with gross changes consistent with corrosion damage visible on the explanted component (Fig. 5). Numerous experienced clinicians conducted extensive workup and treatment of this patient, (including assessment for malignancy) before the diagnosis of head/neck corrosion was considered.

No examination was performed at the time of the first revision procedure for signs of corrosion on the originally explanted femoral head. Unfortunately, the component was subsequently disposed of and was unavailable for further evaluation in support of the present report.

In retrospect, given the similarity in the cross-sectional imaging findings prior to both the first and second revision procedures, it is likely that corrosion at the femoral head-neck junction was responsible for the initial periprosthetic collection.

Imaging findings

The vast majority of literature describing the appearances of periprosthetic pseudotumours has been focussed on those associated with MoM resurfacing and total joint arthroplasties. Hauptfeisch et al. described a classification system for pseudotumours associated with MoM resurfacing [27]: type I, predominately thin-walled cystic lesions that communicate with the joint; type II, thick walled cystic lesions; type III, solid lesions. Types I and II were more typically anterior to the joint while the type III masses were more often posterior and associated more severe symptoms and higher revision rates. Alternate grading systems use bone marrow oedema, signal intensity of the cystic cavity, muscle atrophy/oedema, fascial penetration, tendon avulsion and periprosthetic fracture to differentiate findings [28].

While a number of assessment and grading systems have been reported, there is variable evidence concerning their clinical utility. Thomas et al. found a poor correlation between the MRI appearances of MoM pseudotumours and subsequent clinical findings, further suggesting that MRI tends to underestimate the extent of soft tissue change [29]. In contrast, Nawabi et al. found that their imaging protocols could predict the presence of ALVAL and the intraoperative findings of soft tissue damage based on the maximal thickness of the synovium and overall synovial volume [30].

Most of the cases of pseudotumour associated with modular wear and corrosion have been reported with little emphasis has been placed on detailed descriptions of their imaging characteristics. However, the available evidence suggests similar appearances to those seen in the context of failed MoM articulation, typically with normal or subtle and inconsistent findings on radiographs, even in the presence of marked pathology on MRI. Cooper et al. report that 1 (of 11) radiographs demonstrated subtle cortical scalloping of the medial femoral neck—the remaining radiographs in their series of modular neck corrosion were normal [3], mirroring our unremarkable radiographic findings (Figs. 1 and 2). However, their preoperative metal artefact reduction sequence MRI studies consistently demonstrated “large fluid collections and/or hypertrophic soft tissue reactions or pseudotumour formation”. Meftah et al. reported medial calcar and superior acetabular scalloping on radiographs as well as a distended joint pseudocapsule with synovial debris on MRI [31]. Scully et al. reported normal radiographs and a large, complex, thick-walled fluid collection at the hip joint and adjacent to the proximal femoral stem isointense to muscle on T1-weighted images but markedly hypointense on T2-weighted images [32]. In their assessment of modular neck stem corrosion cases, Molloy et al. found a significant correlation between calcar resorption on radiographs, positive MRI findings and elevated Co/Cr levels [33]. They graded radiographs as negative, mild (<2 mm of medial calcar regression), moderate (2.1–3 mm) and severe (>3 mm). All cases with moderate or severe radiographic findings had evidence of ALTR on MRI. They emphasise the importance of comparison with immediate postoperative imaging to identify medial calcar regression (Figs. 10a and b). However, other reasons for calcar bone loss, such as stress shielding and polyethylene wear, are much more common in the overall THA population and thus cannot be considered pathognomonic for taper junction failure.

Fig. 10
figure 10

Taper junction corrosion has been reported to result in interval calcar resorption and scalloping (a, arrowheads) when compared to immediate postoperative radiographs (b). (Images from Molloy et al. [33], used with permission from Rockwater, Inc.)

Full size image

In all, these findings mirror those of pseudotumour associated with MoM; however, limited systematic evaluation of these lesions by a musculoskeletal radiologist precludes comprehensive imaging descriptions and comparisons with MoM-related lesions.

Summary

The modular head-neck and neck-stem junctions of modern CoCr hip arthroplasty components can, in rare cases, result in metal particle and/or ion release that produces a local soft tissue pseudotumour. Wear at modular junctions has further been identified in explanted stemmed total knee arthroplasty components [34], with at least one reported case of a potential associated ALTR [35]. ALTRs should be included on the differential of all periprosthetic lesions, even in the context of non-MoM implants, as trunnionosis is becoming increasingly recognised as a cause of postoperative symptoms. This phenomenon is a topical issue among arthroplasty surgeons, and knowledge of this entity should help radiologists facilitate early diagnosis and ensure early management of this potentially serious complication.