MARS MRI Abnormalities in Asymptomatic Revision Total Hip Arthroplasty Patients With a Titanium Sleeved Ceramic Head

Abstract

Background

Titanium sleeves are the gold standard to protect ceramic heads placed on a previously used trunnion. However, this introduces another articulation that may lead to local soft tissue reaction. The purpose of this study is to determine the prevalence of abnormalities on metal artifact reduction sequence magnetic resonance imaging (MARS MRI) in asymptomatic patients with a titanium sleeve in revision total hip arthroplasty (rTHA).

Methods

Sixteen asymptomatic patients (17 hips) with a minimum of 1-year follow-up after rTHA with titanium sleeves in a ceramic head were enrolled in this study. The inclusion criteria included the absence of hip pain. All patients underwent an MARS MRI scan interpreted by a fellowship-trained musculoskeletal radiologist to determine the presence of fluid collections in asymptomatic patients with titanium sleeves.

Results

The average follow-up from rTHA to MRI was 5.58 ± 4.61 years (range 1-13.1). Fluid collections were observed in 13 (76%) of the 17 asymptomatic hips in this cohort. Most of these collections were extra-articular in nature over the greater trochanter, with only 3 having direct intracapsular communication. All these collections except 2 were a type 1 lesion based on the Hart classification, whereas the remaining were type 2a. The average pseudocapsule thickness was 11.2 mm in the patients with fluid collections, in comparison to 7.5 mm in those who did not. There were no signs of osteolysis at the most recent follow-up.

Conclusions

Fluid collections are not uncommon after r-THA with a titanium-sleeved ceramic femoral head. Unlike primary THA, synovial thickening may be present in this population. The clinical importance of these findings remains unknown, but this provides information to providers who may obtain MARS MRIs after r-THA.

Level of Evidence

Diagnostic level IV.

Introduction

Total hip arthroplasty (THA) utilization continues to rise dramatically. In a landmark study in 2007, researchers predicted that the prevalence of THA will steadily increase over the years at rates over 170% from 2005 to 2030 [1]. Not only have these predictions been validated, but more recent models have shown even faster growth, with approximately 635,000 procedures expected in 2030 [2]. This growth will also affect the incidence of revision total hip arthroplasty (rTHA) at the same time, resulting in up to 85,000 rTHA annually in the United States [3]. Current revision rates are estimated between 3.2% and 17% at 15 years postoperatively through the New Zealand Joint Registry. Approximately one-third of these operations involve only revision of the prosthesis head and acetabular component, as retaining a well-fixed stem is often desired [4,5].

Retention of a well-fixed stem offers several advantages, including less blood loss, preservation of femoral bone stock, and potentially faster recovery, and often yields good long-term clinical outcomes [[6], [7], [8]]. However, there are also risks associated with this procedure. Removing the femoral head ball using an extraction tool can damage the stem taper surface, resulting in fretting and crevice corrosion [9]. Of particular concern is the risk of catastrophic head fracture, which can occur when a single scratch causes high local stresses under the head when there is an elevation of the metal surface of the taper, as this affects the strength to failure of the ceramic head by up to 96% [10,11].

To mitigate this risk, titanium sleeve adapters have been created to “cushion” the surface between the metal stem taper and the ceramic head. This effectively allows surgeons to bypass unfavorable femoral taper defects that could lead to ceramic head fracture [12]. Instead, the sleeve adapter creates uniform contact and lower local peak stresses and prevents premature fracture. However, this creates another modular junction and may be a mode of potential failure. Despite their common use in rTHA when the femoral component is retained, limited data exist regarding the potential adverse effects associated with titanium adapter sleeves [13].

Metal artifact reduction sequence magnetic resonance imaging (MARS MRI) is the gold standard for detecting adverse local tissue reactions after primary THA [14]. While these scans have the ability and sensitivity to detect lesions and fluid collections, they may overestimate the prevalence of clinically important lesions [15,16]. Given the widespread use of titanium adapter sleeves in revision procedures with retained femoral stems, it is vitally important to understand the prevalence of MRI-documented fluid collections in an asymptomatic population. Establishing this baseline would allow for more accurate interpretation of imaging findings and differentiation between clinically relevant pathology and incidental findings. It is therefore the primary objective of this study to determine the frequency and type of fluid collections noted on MARS MRI in asymptomatic patients with an adapter sleeve in place. It is our hypothesis that there will be fluid collections noted on MARS MRI scans.

Material and methods

This is a single institutional study, and our institutional review board approved this study prior to study initiation. To identify possible study participants, our joint arthroplasty database was queried to identify individuals with at least 1-year follow-up with an adapter sleeve in place under a ceramic head. Only individuals who underwent isolated aseptic head and liner exchanges or had an acetabular component revision were included in this study. Individuals also needed to have absent groin or thigh pain, no pain with range of motion, and radiographs within 2 years before the enrollment period to ensure there were no obvious radiographic abnormalities (eg, eccentric polyethylene wear) that had yet to become symptomatic. Additional exclusion criteria were patients who were unable to have an MRI because of incompatible metal components (ie, pacemakers). All the rTHAs were performed between 2010 and 2023 by 1 of 6 surgeons using a variety of implants and approaches. Each patient had inflammatory markers obtained prior to revision to rule out the presence of an infection. If elevated, an aspiration was performed.

Potential patients were sent a recruitment letter and later followed up by a telephone call by our research staff. A telephone script was used during these encounters to properly discuss the risks and benefits of participating in the study. Additionally, patients were recruited in the office and hospital if they were noted to fit the inclusion criteria. Those who participated provided informed consent prior to any research activities and were compensated $50 after their MRI was complete. A total of 809 head and liner exchange with/without acetabular component revision operations were performed during the study period. Eighty-two patients (10.1%) were identified in our institutional database to have an adapter sleeve in place. Sixty-nine of these patients (84.8%) were noted to have asymptomatic hips, defined as a Harris Hip Score greater than 90. We recruited 19 interested patients (23.2%) between April 1, 2023 and June 3, 2023 to complete this study. Three patients (3.7%) were unable to perform their MARS MRI scans because of scheduling conflicts. Therefore, 16 patients (19.5%) with 17 THAs were the final patient population enrolled in our study, limited by convenience and financial resources (Fig. 1). A majority of patients either declined to participate or could not be reached by telephone.

Figure 1.

Screening and inclusion of study participants.

An MRI scanner (1.5-T; General Electric Healthcare) was used to perform MARS MRI utilizing the same protocol in all patients, as described by Potter et al. [14]. To obtain accurate and standardized images, the 8-channel cardiac coil was centered over the study hip in the supine position. Sequences included axial T-1 weighted (repetition time/echo time [TR/TE] = 784/min, echo train length [ETL] = 5, matrix size = 384 × 224, and number of excitations [NEXs] = 5, with 7-mm slice thickness); axial T2-weighted (TR/TE = 6434/64 ms, ETL = 31, matrix size = 384 × 224, and NEX = 6, with 7-mm slice thickness); axial short tau inversion recovery (TR/TE = 8200/55 ms, inversion time = 135 ms, ETL = 17, matrix size = 320 × 192, and NEX = 4, with 7-mm slice thickness); coronal T1-weighted (TR/TE = 800/min, matrix size = 384 × 125, and NEX = 5, with 5-mm slice thickness); coronal T2-weighted (TR/TE = 7934/125 ms, ETL = 31, matrix size = 320 × 224, and NEX = 6, with 5-mm slice thickness); and coronal short tau inversion recovery images (TR/TE = 7100/55 ms, inversion time = 135 ms, ETL = 17, matrix size = 320 × 224, and NEX = 4, with 5-mm slice thickness). The receiver bandwidth for all sequences was 125 MHz.

All scans were reviewed by 1 fellowship-trained musculoskeletal radiologist in this cohort, and therefore, no interobserver or intraobserver reliability analysis was performed. The scans were interpreted for a variety of factors, including the presence of a lesion, the location and size of the lesion, the type of lesion morphology, and the size of pseudocapsular thickening. The radiologist also noted the presence or the absence of osteolysis. If the pseudocapsule thickness was greater than 6 mm, it was considered pathologic, based on other studies examining aseptic lymphocyte–dominated vasculitis-associated lesions [17,18]. The synovium was measured on coronal views where the thickness was deemed to be the greatest. Lesions were also classified based on the Hart’s manuscript, which has been described in metal-on-metal THA and used in asymptomatic patients with metal-on-poly, ceramic-on-poly, and ceramic-on-ceramic bearing surfaces [19] (Table 1). Descriptive statistics were used to describe the surgical characteristics and fluid collections. No statistical analyses were performed because of the small sample size.

Table 1.

Method of classification of lesions with use of MARS MRI according to Hart et al.

Pseudotumor type	Wall	Contents	Shape
1	Thin walled	Fluid like: hypointense on T1, hyperintense on T2	Flat, with walls mainly in apposition
2a	Thick walled or irregular	Fluid like: hypointense on T1, hyperintense on T2	Not flat, with >50% of the walls not in apposition
2b	Thick walled or irregular	Atypical fluid: hyperintense on T1, variable on T2	Any shape
3	Solid throughout	Mixed signal	Any shape

From Hart et al.[18].

MARS, metal artifact reduction sequence; MRI, magnetic resonance imaging.

Results

Surgical characteristics

The mean duration of follow-up from rTHA to MRI was 5.58 ± 4.61 years (range 1-13.1). The mean time to revision from the index surgery was 8.22 ± 7.45 years (range 0.25-17.6). There were 2 direct anterior approaches, and the remaining 15 were posterior approaches. Eleven patients had isolated femoral head and acetabular liner exchanges. Six patients had acetabular component revisions. Indications for rTHA varied. Seven patients had polyethylene wear resulting in synovitis (4 with ceramic-on-polyethylene articulations, 3 with metal-on-polyethylene articulations). This group included the bilateral rTHA patient who had identical acetabular cup, polyethylene liner, and femoral stem components but had a 36 mm + 6 head size on the right and a 36 mm + 3 head size on the left. Five patients were found to have aseptic loosening of their acetabular component that was revised in addition to a head and liner exchange (4 with ceramic on polyethylene articulations, 1 with a metal-on-polyethylene articulation). Two of these were revised to a dual mobility construct. Two hips had painful metal-on-metal articulations with elevated metal ions (patient 1: cobalt 2.3, chromium 1.6; patient 2: cobalt 55.8, chromium 38.7) that were revised to ceramic on poly. Neither of these patients had noted pseudotumor formation beforehand. One patient had multiple dislocation events that were treated with acetabular component revision and placement of a dual mobility construct. The last patient had a failed hemiarthroplasty secondary to acetabular wear, with the femoral component maintained and the placement of an acetabular cup. The range of implants used, sizes, and approaches are listed in Table 2.

Table 2.

Implants in patient cohort.

Cup
Continuum	1
G7	2
OsseoTi	1
Mallory head	5
Zimmer TM	4
Tritanium	1
Wright Medical	1
Pinnacle	1

Stem
Accolade I	1
Summit	1
Perfecta	1
Taperloc	4
Mallory head stem	4
Mallory head revision	2
SecurFit	3

Liner type
ALTRX	1
Longevity	6
Duramer	1
TridentX3	2
Vivacit E	3
G7 DM	2

Head size (mm)
32	4
36	10
28	2

Offset
0	2
2.5	1
3	3
5	3
6	6
7	1

Prevalence and characteristics of fluid collections

Fluid collections were observed in 13 (76.5%) of the 17 asymptomatic hips in our patient cohort. Eight of these collections were extra-articular, located over the greater trochanter. Three collections extended into the intra-articular space. The average size of these lesions was 18.46 mm (standard deviation [SD] ±22.54) in the anterior-posterior dimension, 23.72 mm (SD ± 21.18) in the transverse plane, and 50.09 (SD ±29.38) in the craniocaudal plane. Based on the Hart classification, 11 of the 13 lesions (84.6%) were type 1. Two patients had type 2a lesions. Of the patients with MRI abnormalities, 9 patients had a head size of 36 mm, 2 patients had a head size of 32 mm, and 1 patient had a head size of 28 mm (dual mobility construct). The patient who underwent bilateral rTHA had a type 1 lesion on the side with the 36 mm + 6 femoral head. There was no lesion on the side with the 36 mm + 3 femoral head. The mean time in situ for the patients with abnormalities was 6.3 ± 4.8 years.

Pseudocapsular thickness and additional MRI findings

Five patients had thickened synovium and were all noted to have pseudocapsule thickness measuring greater or equal to 6 mm (6, 7, 11, 13, and 17 mm). The remaining 8 patients had small elongated fluid collections without significant synovial thickening, but 7 of 8 still had pseudocapsular thickness greater than 6 mm. The average pseudocapsule thickness was 10.6 ± 4.7 mm in the patients with fluid collections, in comparison to 7.5 mm ± 2.4 mm in those who did not. None of the hips in this series had radiographic or MRI findings suggestive of osteolysis. Remaining characteristics of these MRI abnormalities are listed in Table 3.

Table 3.

Dataset for asymptomatic hips with titanium sleeves that had fluid collections.

Case	Approach	Time in situ (y)	Hart classification	Capsular thickening
1	Posterior	4.7	1	4
2	Posterior	11.5	1	11
3	Posterior	2.8	1	6
4	Posterior	13.1	1	9
5	Posterior	1.0	2a	6
6	Posterior	1.0	1	15
7	Posterior	9.4	1	12
8	Posterior	13.0	1	7
9	Posterior	3.4	2a	7
10	Posterior	1.9	1	16
11	Posterior	10.4	1	20
12	Anterior	3.2	1	12
13	Posterior	6.9	1	13

Discussion

In this study of asymptomatic patients who underwent rTHA using a titanium sleeve adapter on a retained femoral stem, we found that the majority of hips demonstrated fluid collections on MARS MRI. Most of these collections were small and elongated without any synovitis or intra-articular extension. Because sleeves are widely used in rTHA when the femoral component is retained but create another modular junction that may be a mode of potential failure, understanding the prevalence of MRI-identified fluid collections in asymptomatic individuals is important. Our findings suggest that fluid collections are common in this specific revision population. However, distinguishing the relative contribution of the revision surgery itself, prior pathology, and the titanium sleeve-trunnion interface remains challenging.

Multiple studies have examined the prevalence of fluid collections in asymptomatic THA. A study examining ceramic heads on highly cross-linked polyethylene liners for 50 asymptomatic patients found that 18% of patients had notable fluid collections, including 4% that had thickened synovium [20]. A similar publication looking at ceramic-on-ceramic articulations observed that 19% of asymptomatic patients had noted fluid collections, with 3 of these patients having thickened synovium [21]. A recent work looking at asymptomatic metal on polyethylene total hip replacements noted abnormalities for 28% of patients with well-functioning THAs [22]. While all these studies demonstrate that postoperative fluid collections could be a common finding in asymptomatic THA, all these rates were significantly lower than the 75% fluid collections noted in this report. It should be noted that all the patients in this present study were revision cases and therefore underwent multiple surgeries to the hip joint, possibly making it more likely to develop these postoperative changes. One patient was found to have an absence of the posterior-lateral structures intraoperatively, whereas the rest underwent repair of the short external rotators and posterior capsule upon closure. However, these structures are often not as robust in the revision setting, which may contribute to the development of fluid collections and their intracapsular/extracapsular extension.

It could also be hypothesized that high rates of fluid collections occurred in this series because of the nature of these revisions, where a partially damaged trunnion interacted with a titanium sleeve. It is known that titanium particles are not as toxic to the human body as cobalt or chromium; however, it is not known if these particles can result in a reactive fluid response [11]. Still, trunnion corrosion could contribute to cystic responses. There have been case series of “silent” trunnion corrosion, where pseudotumor defects occurred in asymptomatic metal on polyethylene patients. Even though 40% of this cohort had no pain, eventually they needed revisions as soft tissue defects resulted in instability events [23]. As there are no longitudinal data evaluating asymptomatic patients with fluid collections, it is difficult to determine if these collections wax and wane in size or if they continue to progress and hold potential for soft tissue damage that could necessitate intervention in the future. The observed fluid collections in this study, although asymptomatic at the time of imaging, could be an early indicator of pathological changes that may progress to adverse local tissue reactions.

Recent literature has investigated the mid-term to long-term outcomes of rTHA with titanium sleeves. A study of 516 rTHAs with femoral component retention found no reoperations or failures for ceramic head fracture or taper corrosion at a mean 4-year follow-up, including 12 patients who underwent rerevision for adverse local tissue reaction [24]. A long-term study with 316 patients demonstrated similar results with no failures at the femoral head/neck junction after rTHA with titanium sleeves at a mean 8-year follow-up [25]. Both authors conclude that the use of titanium sleeves is safe; however, neither reported on the presence of fluid collections. A future study that includes imaging data to assess the prevalence and progression of fluid collections in a large cohort of patients would be essential to better understand their clinical significance and potential association with clinical outcomes. Additional studies with a comparison cohort of hip revision surgeries who are symptomatic and another with a control group that did not require a titanium sleeve would also be ideal to see if fluid collections are present in these patient populations.

Limitations

There are a few limitations in this study that need to be addressed. First, there are significant financial costs of obtaining an MRI for asymptomatic patients. Therefore, we were only able to obtain a small patient population for review, and no statistical analyses were performed. Selection bias is also present, as only a small proportion of eligible patients ultimately had an MRI performed. Additionally, the heterogeneity of revision indications, implant designs, and bearing surfaces makes it difficult to draw significant conclusions regarding the etiology of the observed fluid collections. Furthermore, all MRI scans were interpreted by a single fellowship-trained musculoskeletal radiologist. While this ensured consistency in image interpretation, the absence of interobserver or intraobserver reliability analysis limits generalizability. Last, there were 2 patients who were revised for metal-on-metal articulations resulting in pain. Of note, both these patients had MRIs of their hips before revision without signs of pseudotumor or fluid collections before their revision surgery, although their metal ion levels were elevated and could possibly contribute to an ALRT postoperatively.

Conclusions

Fluid collections are commonly noted after rTHA in asymptomatic patients when a titanium-sleeved ceramic femoral head is used. Synovial thickening was noted in our entire patient population, although it was more prevalent in those who had a fluid collection. The clinical importance of these findings is still unknown and could result in possible complications in the future. Further longitudinal studies are needed to identify if these findings could contribute to pathology. Still, these data will provide a baseline for providers who may obtain MARS MRI after rTHA.

CRediT authorship contribution statement

Adam S. Driesman: Writing – review & editing, Writing – original draft, Investigation. Nicholas C. Aprey: Writing – review & editing, Investigation. Gregory Czuczman: Writing – review & editing, Methodology, Investigation. Madison C. Hulst: Writing – review & editing, Project administration, Data curation. Roseann M. Johnson: Writing – review & editing, Formal analysis, Data curation. Douglas A. Dennis: Writing – review & editing, Investigation, Conceptualization. Jason M. Jennings: Writing – review & editing, Supervision, Investigation, Conceptualization.

Conflicts of interest

The authors declare there are no conflicts of interest.

For full disclosure statements refer to https://doi.org/10.1016/j.artd.2026.102017.