Outcomes After Metal-on-Metal Hip Resurfacing: Could We Achieve Better Function?

Methods 

Consecutive patients who had received a Conserve MOM hip resurfacing arthroplastya from October 2003 were invited to a single follow-up assessment 1 year after surgery. Most improvement after arthroplasty surgery occurs in the first year and then most patients plateau; so a review at 1 year or more is thought to be representative of mid- to long-term outcomes.15 Clinical data about a minimum of 100 operated hips was sought. Five patients had been revised to THA before the study start; they were noted as failures but excluded from further follow-up. Six patients were involved in a separate radiologic study and were unavailable for follow-up by this study. Those unable (eg, moved overseas) or unwilling to attend a review were asked to complete postal questionnaires to try to gain as much data as possible. Ethics approval was provided by the institutional research committee.

To be eligible for MOM hip resurfacing, patients met specific criteria. All had been diagnosed with end stage symptomatic osteoarthritis of the hip. Success in MOM hip resurfacing is closely linked to good femoral bone quality, usually found in younger active adults.10 Consequently eligible patients were relatively active, for example, able to walk independently and undertake some nonsedentary occupation, or leisure pursuit and relatively young, typically less than 60 years of age, although some older, active people were included. The surgery was performed in lateral decubitus position. In 111 cases, a posterior approach was used, and in the remaining 14 cases a Hardinge (anterolateral) approach was used. After exposure, the hip was dislocated, capsule divided circumferentially, femur and acetabulum prepared, and appropriate implant size selected. After the components were implanted, the hip was reduced and stability checked. The wound was closed in layers over a drain.

Postoperatively patients followed a standardized care pathway to guide mobilization and return to function prior to discharge. As in other centers, there had been no change to the existing hip arthroplasty rehabilitation protocol,14 but rather a protocol was used that was based on the current evidence-based literature and developed for THA procedures.16 It assumes a hospital stay of 5 days and all members of the multidisciplinary team provide care. Patients started walking on the first postoperative day and followed an incremental progression of full weight-bearing gait re-education, range of motion (ROM) and muscle strengthening exercises as described by Enloe et al.16 This program includes active hip flexion, extension and abduction exercises, advice to follow THA hip precautions, and the provision of adaptive equipment (ie, toilet raise, shoe-horn) for use at home for the first 3 months.

Patients were discharged when able to independently perform transfers from bed to standing, sit to stand, walk independently using an appropriate walking aid, and able to ascend and/or descend stairs.15 They were discharged with a standard physiotherapy (PT) information and exercise leaflet including details about progressing exercises and mobility. All patients were offered an outpatient review 6 weeks postoperatively, when progress could be monitored and exercises and gait training progressed as needed. At this stage, if problems were identified, a patient might be referred for additional outpatient PT where an individualized treatment program would be developed, but the majority would receive no further rehabilitation.

The review assessments were carried out in a dedicated research outpatient clinic by 2 independent observers, both physiotherapists with over 10 years of experience. Neither research physiotherapist had seen the patients prior to the 1 year follow-up. After giving informed consent, patients completed a number of self-report outcome measures of functional status. These included the Oxford Hip Score (OHS),17, 18, 19 Hip disability and Osteoarthritis Outcome Score (HOOS),20, 21 and UCLA Activity Score12, 22 (table 1). Two screening questions about the impact of hip pain on sexual activity and social life were included; these items were adapted from the Oswestry Disability Index, a validated and reliable self-report measure typically used in populations with spinal conditions.23 For these questions scores ranged from 1 (sex/social life normal and causes no extra pain) to 6 (pain prevents any sex/social life).23 The UCLA Activity Score was used with modifications for the British population.4 Total HOOS and the individual HOOS subscales were used. Any missing data were treated as such; 1 missing value was substituted with the average value for the dimension. If 2 or more items were omitted then the dimension was considered invalid.20

Table 1. Outcome Measures

	Measure	OHS17, 18, 19	HOOS20, 21	UCLA Activity Score4, 10, 22
	Population	Adult hip surgery	Adult hip disability with/without OA.	Adult hip, knee surgery
	Structure	About last month. 12 items about pain, mobility, and function.	About last week, 40 items in 5 domains; pain, other symptoms, function in ADLs, sport and recreation, QOL.	10 descriptive activity levels with activity type, intensity, and frequency considered.
	Administration	Self-administered	Self-administered	Self-administered
		Takes 5min	Takes 7–10min	Takes 2−3min
		Clinic or postal	Clinic or postal	Clinic or postal
	Psychometric properties	Validated. Reliability (ICC>.88). Sensitive to change.	Validated. Reliable (ICC>.78). Sensitive to change.	Validated.
	Scoring	Items 1–5 on ordinal scale. Responses are totaled. Range, 12 (best) to 60 (worst).	Items 0–4 on an ordinal scale. A normalized score (0 worst to 100 best) calculated for subscales and total HOOS. WOMAC calculable from HOOS.	1–10: inactive (1), mild (3), moderate (5), active (7), regular impact sports (10).

Abbreviations: ADLs, activities of daily living; ICC, intraclass correlation coefficient; OA, osteoarthritis; QOL, quality of life; WOMAC, Western Ontario and McMaster Universities osteoarthritis index.

Patients were asked about the status of the opposite hip and a Charnley class24 (A, B, or C) was determined. A clinical examination was completed and hip ROM for the resurfaced hip and isometric muscle strength, ability to perform a single-leg stand, and Trendelenburg test were assessed for both hips. Goniometry was used to record ROM for the resurfaced hip; this procedure has acceptable intra and interrater reliability.25 Hip strength was measured using manual muscle testing and graded according to the Medical Research Council (MRC) scale of motor recovery, a widely used clinical procedure.23 Walking and stair climbing ability and a timed 10-m walk were recorded.26 The 10-m walk is a valid, reliable measure of walking ability associated with functional independence and used in orthopedic populations.27 Then patients were simply asked whether they felt the resurfaced hip was satisfactory and about postoperative complications using a standard checklist that included unexplained pain lasting for 3 months after surgery. Hospital notes and operation records were also checked for complications. Finally patients attended radiology where anteroposterior pelvis and lateral hip radiographs were taken.

Results 

From July 2005 to April 2006, 126 operations, performed on 120 patients from October 2003 to March 2005, were reviewed at a mean of 14.9±2.9 (range, 9–22) months after surgery. This comprised 67 (53.2%) right and 59 (46.8%) left hips of which 105 were reviewed clinically and 21 through postal questionnaire. Of the remaining surviving hip resurfacings, 1 patient with bilateral hip resurfacings had progressive Parkinson’s disease confounding outcomes and was excluded, and 16 patients did not respond or declined to participate. Thus in a cohort of 137 patients with 144 hip resurfacings, 126 (87.5%) resurfaced hips were reviewed.

Of the 120 patients, 71 (59.2%) men and 49 women (40.8%) participated with a mean age of 56±9 (range, 24–76) years. Sixty-one percent reported that a single hip joint was affected (Charnley class A), 29% that both hips were affected (Charnley class B), and 10% that multiple systems were affected (Charnley class C). Of those with opposite hip symptoms, 16% had received no treatment but 24% had undergone bilateral hip arthroplasties. Of these 16% reported a successful MOM resurfacing procedure on the opposite side, 4% an unsuccessful hip resurfacing, 3% a successful THA, and 1% a problematic THA.

Complications 

Complications are given in table 2. Five patients had undergone revision surgery prior to this study due to a fractured neck of femur in 4 cases and to avascular necrosis (AVN) in 1 case. Of the surviving hips examined clinically, 91 (86.7%) experienced no complications and 92 (87.6%) were described as satisfactory. Fifteen complications that did not lead to revision were noted. Thirteen of these occurred in the perioperative period. In addition, 21 patients (20%) reported hip pain that persisted for more than 3 months after surgery. Pain reported at 3 months was moderately correlated with the HOOS pain subscale (ρ=.52, P<.001). Other complications were not correlated with pain, satisfaction, mobility, or functional scores at 1 year. One dislocation occurred in the series. It was reduced by closed reduction and at 1 year the patient had a UCLA Activity Score of 7 (active), an OHS of 22 and reported that the hip was satisfactory.

Table 2. Complications

	Complication	Surviving Hips (n=105)	95% CI for Population Percentage	Revised Hips (n=155)	95% CI for Population Percentage
	Hemorrhage	6(5.7)	1.3–10.1
	Superficial wound infection	3(2.9)	0.0–6.0
	Chest infection	1(1.0)	0.0–2.8
	Dislocation	1(1.0)	0.0–2.8
	Nerve injury	2(1.9)	0.0–4.5
	Heterotrophic ossification	1(1.0)	0.0–2.8
	Cardiac arrhythmia	1(1.0)	0.0–2.8
	Trendelenburg positive	11(10.4)	4.6–16.3
	Pain at 3mo	21(20)	12.3–27.6
	Femoral neck fracture			4(2.6)	0.1–5.1
	Avascular necrosis			1(0.6)	0.0–1.9
	Further hospital treatment	3(2.9)	0.0–6.0	5(3.2)	0.4–6.0

NOTE. Values are n (%) or as indicated. Abbreviation: CI, confidence interval.

Self-Reported Function 

Function and activity levels are shown in table 3. The median OHS was 15, with 75% of hips scoring less than 20 on the OHS; indicating few problems for most patients. The median score for all individual questions within the OHS was 1 (not at all), other than for question 4 of the OHS, which had a median score of 2; for example, the population was more likely to find more difficulty in putting on shoes and socks than in any other activity.

Table 3. Self-Report Function and Activity Questionnaires

	Questionnaires	Hips (n)	Median	Range	IQR (25%–75%)
	UCLA Activity Score (1–10)	124	7.0	2–9	5–8
	Total OHS (12–60)	126	15	12–55	13–20
	Total HOOS (0–100)	122	88.2	9.32–100	75.4–95.4
	HOOS subscales (0–100)
	HOOS pain	126	97.5	12.5–100	86.8–100
	HOOS symptoms	126	90	15–100	80–95
	HOOS ADL	126	93.4	19.12–100	83.8–98.5
	HOOS sport	122	84.4	0–100	56.25–100
	HOOS QOL	126	75	0–100	62.5–93.75
	Sex life (1–6)	99	1.0	1–6	1–2
	Social life (1–6)	118	1.0	1–5	1–2

Abbreviation: IQR, interquartile range.

On the UCLA Activity Score most patients described themselves as active with a median score of 7, indicating that they regularly participated in activities such as gardening, cycling, or going to the gym. Seventy-five percent of the participants were at least moderately active, sometimes exercising, and unlimited in their ability to perform activities such as shopping and housework.

The median HOOS total score of 88.2% describes a high level of function for most patients after resurfacing with 75% scoring more than 75.4%. This highlights the positively skewed distribution that might be hoped for after surgery designed to improve pain and function. Subscale analysis indicated fewest problems with pain and greatest negative impact on quality of life (QOL) and the ability to participate in sport; in fact 4 patients were unable to complete any part of the HOOS sport and recreation section. There was a considerable range in both the total and each separate subscale HOOS score. Low minimum scores of less than 40 persisted for only 5 patients. Activities more frequently identified by patients as being difficult included bending to the floor to pick up an object, putting on and taking off shoes and socks, and squatting. Two questions accounted for most missing values: one about running and one about taking a bath. Performance as measured by the HOOS and OHS was significantly and highly correlated (ρ=.86, P<.001). The UCLA Activity Score only correlated moderately with OHS (ρ=.47, P<.001) and HOOS (ρ=.5, P<.001); reflecting the dimension of higher physical activities covered in more detail by the UCLA Activity Score.

Mobility 

For most patients, walking distance was unlimited, no walking aids were used, and stairs could be negotiated without difficulty (table 4). The mean 10-m walking time was 7.76±1.26 (range, 5.67–11.58) seconds, or a walking speed of 1.29m a second, which fell within the normative adult range.28 However, 25.7% of patients reported some limitation of walking distance, 7.6% used a walking stick, and 10.5% required a rail to negotiate the stairs. Limited walking distance correlated with a slower 10-m walk time (ρ=.41, P<.001), difficulty on stairs (ρ=.48, P<.001), and worse OHS (ρ=.46, P<.001).

Table 4. Mobility

	Measures	Hips (n)	n (%)
	Stair use	105
	(1) Normal		87(82.9)
	(2) One step at a time		7(6.7)
	(3) Down with rail		4(3.8)
	(4) Up & down rail		7(6.7)
	(5) Unable		0(0)
	Walking aids	105
	None		97(92.4)
	Single stick		8(7.6)
	Other		0(0)
	Walking distance	105
	(1) Unlimited		78(74.3)
	(2) 500m–1km		20(19)
	(3) 100–500m		5(4.8)
	(4) <100m		2(1.9)
	(5) Housebound/unable		0(0)

Physical Measures 

The results are given in table 5. The ROM at the hip 1 year after surgery was significantly less than the normative ranges reported for this age group for hip flexion, extension, and abduction.25 At 1 year a significant difference in the strength of hip flexors, extensors, and abductors on the reviewed side remained (P<.001), although 75% of patients had at least MRC grade 4 on the reviewed side and most patients had normal (grade 5) strength on both sides. Most patients were able to perform a single-leg balance on the reviewed side, and able to hold this position for over 10 seconds. A subset of patients had some problems performing this activity, and there was a correlation between poor single-leg balance and a positive Trendelenburg test (ρ=.49, P<.001).

Table 5. Hip Range of Movement, Power, and Balance

	Measures	Hips	Reviewed Side	Contralateral Side	P⁎
	Hip ROM (deg)			ND	NA
	Hip flexion	105	94.46±12.71
	Hip abduction	105	25.67±9.24
	Hip extension	105	7.86±8.96
	Power (grade 0–5)
	Hip flexion	103	5 (3–5)	5(3–5)	<.001
	Hip abduction	103	5(3–5)	5(3–5)	<.001
	Hip extension	102	5(4–5)	5(3–5)	<.001
	Single-leg balance (grade 1–5)	101	5(2–5)	5(2–5)	.60

NOTE. Values are n, mean ± standard deviation, or median (IQR). Single-leg balance scores are: (1) unable; (2) tries, unable for 3 seconds; (3) able for 3 seconds; (4) able 5 to 10 seconds; and (5) able more than 10 seconds. Abbreviations: NA, not applicable; ND, no data.

⁎ Wilcoxon signed-rank test (P<.05).

Discussion 

Overall, this review of MOM hip resurfacing patients showed high levels of function, activity, and satisfaction with the outcome of surgery. The patients that are reported here have a higher level of bilateral and multiple joint disease and higher mean age than the group described in a comparable study by Daniel et al4 in their report of activity after Birmingham hip resurfacing. This may explain the lower mean UCLA Activity Score level found in our study.

Fractured neck of femur was the most common reason for revision of the surgery. This agrees with the findings of other studies.6, 10, 12 AVN of the femoral head which may lead to fracture has been recognized and associated with the posterior surgical approach, which can devascularize the femoral head.3, 4, 12 In our study, 1 patient developed AVN and the surgery in this case was done through a posterior approach, but no patients with hips undertaken through a Hardinge approach had AVN. The small numbers in this study allow no meaningful subgroup analysis but further comparison of outcomes in relation to surgical approach in a larger study is warranted.

Reports of pain at 3 months were of interest, especially because this correlated with pain in the week prior to review (HOOS pain). If this is considered alongside the correlation between the HOOS pain subscale and the OHS, not only is the link between pain and poorer functional outcome emphasized, but it also appears that problems did not resolve independently over the year. It suggests that further intervention may be needed for patients with persisting pain at 3 months.

The strong correlation between overall OHS and HOOS total scores does raise questions about the need to use both questionnaires; suggesting that both measure similar dimensions of pain and function. The completion rate for the HOOS was lower due to the difficulty 4 patients had completing the sport dimension, which included items such as running, squatting, and pivoting. Two of these patients had bilateral hip resurfacings. Reporting a modified HOOS without the sport dimension has been suggested and could be considered. However, the HOOS did identify a wider range of activities that this relatively high functioning population found difficult and arguably the sport items were most sensitive to restricted function.20 It also considered QOL specifically. Because QOL was the subset of the HOOS that scored lowest, it was valuable in this study for highlighting the wider impact of hip problems. To avoid duplication, if may be better in future studies to only use 1 functional questionnaire, an activity measure, and to explore QOL in greater depth using a specific measurement tool.

The majority of patients reported no problems with either their sexual activities or social life. However, a small subset of patients did indicate problems, and these could be severe. Both these areas are rarely discussed, perhaps due to understandable anxiety on behalf of patients or clinicians or a perceived lack of opportunity within clinics. However, these results suggest that monitoring this area is important and that access to suitable support may be needed. Further research using qualitative methodology to explore the problem, including any perceived barriers to seeking help, would improve understanding of this area.

Walking continued to be limited for 25% of patients with a small number requiring walking aids and rails to negotiate stairs at 1 year. Findings of pain, reduced ROM, strength, and proximal stability are likely to have contributed to restriction and reduced activity scores but reduced aerobic fitness, weight, and confidence negotiating outdoor enviroments may also have had a role. Further investigation of such underlying factors and analysis of gait patterns late after hip resurfacing would be valuable.

At 1 year after surgery, there were persisting muscle strength differences compared with the contralateral side and 25% of patients did not score full strength (grade 5) for the 3 muscle groups tested. This finding supports the work of Shih et al29 who reported that although hip strength improved considerably in the year after total hip replacement and improved from the preoperative values, it remained around 80% of the asymptomatic side.

It is known that hip abductor weakness affects pelvic stability.30 The number of patients who had a positive Trendelenburg test at 1 year was notable and this was associated with poorer high level balance. Hip abductor weakness is a risk factor for falls and has been linked to abnormal gait and reduced mobility in patients a year after THA.31 Thus the finding of long-standing weakness after hip resurfacing has important implications for the content of any future rehabilitation. As suggested by Shih,29 rehabilitation should encompass a focus on strength training for the first year after surgery.

A weakness of this study is that it recorded muscle strength using manual muscle testing23 as part of the overall examination—a method that is widely used clinically but is known to be less accurate than techniques such as isokinetic dynamometry. We recommend that future studies investigate strength changes in more detail using dynamometry. This information is important to analyzing the impact of both rehabilitation protocols and surgical techniques; for example, whether the use of a Hardinge versus posterior approach is linked to abductor weakness at 1 year.

A further finding of note was poor recovery of hip flexion and abduction ROM at 1 year. In fact the mean value for hip flexion was only 94.46°±12.71° with a lowest value of 55°, values which are some way below the age-related norms of 120°±14°.25 Lack of hip flexion was associated with lower function. Higher ranges of hip flexion would be key to activities such as putting on shoes and socks or bending to the floor,32 both highlighted as more difficult by patients in this study. Restricted ROM may reflect a rehabilitation strategy that has overemphasized the importance of hip precautions and the risks of dislocation in extreme ranges of flexion, reflecting the higher risk of dislocation associated with THA. In fact the dislocation rate of 1% recorded in this study agrees with low rates of dislocation reported by others4 and contrasts with reported rates of dislocation after conventional THA of up to 5% in the first year.5 The results contrast with those of dela Rosa et al.33 They investigated hip ROM changes pre- and post-Conserve MOM hip resurfacing in a cohort of 56 patients with 64 resurfaced hips at 1 year and reported mean hip flexion of 110°. It may partly be explained by the higher mean age of our cohort at 56 versus 48 years and by the higher incidence of bilateral disease. Also, in the study by dela Rosa,33 postoperatively patients had been given no specific ROM exercises until 6 weeks after surgery but at this stage unlimited stretches to hip flexors, abductors, and external rotators were recommended.

Poor recovery of ROM may be due to other factors, for example, heterotrophic ossification (HO) in 1 case. Back et al11 recorded HO in 59% of their 230 Birmingham MOM resurfacings, with 3 patients requiring further surgery to improve ROM and pain. However, in contrast to our patient group, they reported mean hip flexion ROM to be 110° postoperatively.11 A stronger focus on regaining ROM in early postoperative rehabilitation may explain these outcomes. Immobilization followed by forcible mobilization has been suggested as an important trigger of HO in trauma and arthroplasty patients.34 Dela Rosa33 reported that 12 hips (18%) in 11 patients developed HO postoperatively and these had significantly less external rotation ROM at 1 year. Dela Rosa also noted that factors such as limited preoperative hip ROM, valgus positioning of the femoral component, and the presence of bilateral hip disease were associated with significantly less hip ROM postoperatively. These factors highlight the need to examine both rehabilitation and surgical protocols in analyzing outcomes.

This study was set up retrospectively; consequently, no preoperative assessment was conducted that would have allowed measurement of change within subjects and could have strengthened the findings considerably. In addition, 12% of patients were lost to follow-up, limiting generalizability of the study. Monitoring of a larger cohort of MOM hip resurfacing patients including complications, physical measures, functional outcomes, and radiographic follow-up appears warranted and with this in mind current patients undergoing Conserve MOM resurfacing at this hospital are being invited to participate in a review preoperatively and at 1 year.

Like others we used a rehabilitation protocol derived from that used after THA surgery. There has been no attempt to tailor the rehabilitation to the needs of younger patients or to the differing surgical technique used in MOM hip resurfacing. Consequently, it is possible that a contributory factor to the suboptimal outcome of a small number of patients was postoperative rehabilitation that failed to meet the different demands of this patient group. Furthermore, it may be that some weaknesses in the agreed core strategies developed for hip protocols16 are highlighted by this hip resurfacing population. Previous work suggests that after arthroplasty there is frequently persistent muscle weakness and poor functional mobility and that ongoing rehabilitation may be required to optimize recovery.35 A second phase of rehabilitation postdischarge emphasizing weight bearing, strengthening, higher-level balance, and gait activities may be beneficial for hip arthroplasty patients in terms of improving strength, postural stability, and function.35, 36

The provision of a tailored rehabilitation protocol for the MOM hip resurfacing population requires the identification of factors uniquely associated with resurfacing. Preliminary enquiry reveals that patients undergoing this procedure are younger, more physically active, have greater expectations, and can be more adherent (to the point of overzealous behavior) than standard THA patients. Some of these factors might lead to suboptimal rehabilitation. High activity levels in the medium term are not associated with poorer outcomes,10 but the timing of more intense activity may need consideration to reduce risks of pain, fracture, or HO. A protocol that acknowledges the different patient characteristics of those undergoing hip resurfacing seems important. In early inpatient protocols advice to follow hip precautions could be modified and instead could emphasize the need to introduce activities in a phased manner to avoid overstressing tissues. A later phase of outpatient rehabilitation should occur and focus on regaining ranges of hip movement, hip extensor, and abductor strength, include retraining functional activities (eg, putting on shoes and socks) and progress to higher-demand balance and mobility activities.

Conclusions 

Overall, the cohort obtained a very satisfactory outcome after their hip resurfacing procedures. The results are in line with other research with the majority of patients returning to high levels of activity, including sport and heavy occupational demands. However, our deeper investigation of function has revealed that a proportion of hip resurfacing patients continue to experience some difficulties as late as 1 year after surgery. We have postulated that this may be improved, to some extent, by a rehabilitation program that is more specifically focused on the differing biomechanics of the resurfacing prostheses and the different demands and expectations of the hip resurfacing patient population.