Hip resurfacing arthroplasty (HRA) with metal-on-metal bearings has become an established and viable hip arthroplasty option for the younger patient with higher activity levels due to bone preservation. This technique may also make revision surgery less complicated [1, 2]. In Europe, the rate of resurfacing has varied between 6% and 9% with 6% in France, 9% in Germany, and 7% in the UK [1, 3]. In Australia, the hip resurfacing accounts for 7.9% of all hip arthroplasty procedures. In some countries, hip resurfacing has been utilized in up to 50% of all hip arthroplasties in patients younger than 55 with a low revision rate of 2.8% at five-year follow-up post-operatively [4, 5].

The risk factors for stemmed total hip arthroplasty (THA) appear to be different than for HRA⁵, and many experts have advocated that HRA may be more advisable in certain subsets of patients with severe degenerative arthritis of the hip. Risk factors have been proposed that increase the risk for HRA [6–8]. Femoral head cysts are widely believed to increase the chances of early femoral failure in HRA; however, the only scientific data that exist now to support this idea is mainly from Beaule's study [6, 9, 10]. In their study, femoral head cysts were identified as a risk factor for early femoral failure after metal-on-metal HRA as a part of the proposed Surface Arthroplasty Risk Index (SARI) [6]. Cysts were found to be a significant risk factor (P = 0.028) for early femoral failure. Our concern is that the technique of managing cysts may be important in achieving a good outcome. In Beaule's study, cysts were filled with cement; our technique is to instead fill them with acetabular reamings prior to cementation or uncemented fixation.

We were not convinced that cysts affected the failure rate provided that they involved less than one third of the prepared femoral head and that they were bone grafted instead of being filled with cement. Because the scientific evidence to support cysts as an independent risk factor was limited, we have routinely used this approach. After many years of experience with these cases, we have now undertaken this study to independently analyze what the added risk of early femoral failure in HRA was when femoral head cysts were present and treated with bone grafting. Our hypothesis in this retrospective case-control study was that femoral heads with cysts involving less than 1/3 of the prepared femoral head did not significantly affect the early femoral failure rate after HRA.

Routine postoperative follow-ups were requested at six weeks, one year, two years, and every other year afterward. Harris hip score (HHS), UCLA activity score, and visual analogue scale (VAS) pain score were evaluated at every follow-up visit. Complications and failures were recorded. Anteroposterior and lateral radiographies were obtained at each follow-up (Figure 1). Radiolucencies, osteolysis, migration, reactive femoral lines, focal femoral neck narrowing, and heterotopic bone according to the Brooker scale[12] were evaluated.

The level of significance was set as 0.05 (α = 0.05) for all comparison tests in this study. The paired t tests were performed to compare the numeric variables between pre-operative and post-operative visits. The standard t tests were performed to compare the differences between numeric variables of the study and control groups. Chi-square tests were performed to evaluate the difference of categorical variables between these two groups. The Kaplan-Meier curves were used to analyze the survivorship rates using revision of femoral components as the end point among these two groups. The Chi-square tests were performed to approximate the results of the Wilcoxon tests in order to compare the differences of survivorship functions between groups. The null hypotheses of all of these tests were that the survivorship functions were the same between the two compared groups [13]. Also, the Pearson Chi-square tests were utilized to compare the differences of failure rates between groups without considering the time variable.

All patients in this study had a minimum follow-up of 24 months (Table 4). No patients died in the study group. Two patients died of causes unrelated to the hip surgery after two years in the control group. Both of them were included in this study. At the latest follow-up visits, there were three femoral failures (two in men and one in a woman) in the study group; there was no femoral failure in the control group (P = 0.08): two cases (1.7%) were revised due to femoral neck fracture prior to six months post-operatively; one (0.9%) was revised due to femoral component loosening (presumably due to osteonecrosis). Detailed information is listed in Table 5. The survivorship curves using revision of the femoral component as an endpoint are plotted in figure 2. At 60 months postoperatively, the survivorship rates of the femoral components were 97.4% in the study group and 100% in the control group. However, there was no significantly statistical difference of failure rates between these two groups without considering the time variable (P = 0.08) and there was no significantly statistical difference of survivorship functions between them (P = 0.09). In the cyst group, there was one femoral neck fracture among 53 uncemented femoral components; and there was one femoral neck fracture and one femoral component loosening that occurred among 64 cemented femoral components. There was no significantly statistical difference of the early femoral component failures between the fixation of femoral components (P = 0.67).

	Study Group -- with Cyst	Control Group -- Without Cyst	P-Value
Period of follow-up (months)	42 ± 11 (range: 24 to 61)	45 ± 12 (range: 24 to 65)	0.08
Pre-operative information
HHS score	54 ± 12 (range: 24 to 91)	55 ± 13 (range: 21 to 83)	0.2
Post-operative information
HHS score	97 ± 6 (range: 68 to 100)	95 ± 8 (range: 71 to 100)	0.22
UCLA score	8 ± 2 (range: 4 to 10)	8 ± 2 (range: 3 to 10)	0.95
VAS score in the regular day	0 ± 1 (range: 0 to 4)	0 ± 1 (range: 0 to 4)	0.59
VAS score in the worst day	1 ± 2 (range: 0 to 8)	1 ± 2 (range: 0 to 7)	0.27
Femoral radiolucency	0 (0%)	0 (0%)	1
Number of femoral failures (revisions)	3 (3%)	0 (0%)	0.08
Deceased	0 (0%)	2 (1.7%)	0.16

Time after surgery (Months)	Cyst size (cm3)	Femoral size (mm)	Primary diagnosis	BMI	Gender	Age	Reason of failure	Treatment of failure
0	1	48	Dysplasia	23	Female	49	Femoral Neck Fracture	Femur Revised
1	1	52	OA	35	Male	59	Femoral Neck Fracture	Femur Revised
37	2	58	OA	27	Male	53	Femoral Component Loosening	Femur Revised

Excluding the revised cases, the average post-operative HHS scores at the latest follow-up visit was 97 ± 6 in the study group and 95 ± 8 in the control group; both were improved significantly from the average pre-operative HHS scores, respectively (P < 0.001) (Table 4). There were no significant differences in the UCLA activity and VAS pain scores on the regular or worst days. Radiological analysis revealed that no hip showed evidence of femoral radiolucency or migration.

When comparing HRA to stemmed THA, the spectrum of complications is different. Considering that multiple bearing options are currently available for stemmed THA, the comparison between HRA and stemmed THA becomes even more difficult. Two complications that are unique to HRA are femoral neck fractures and postoperative femoral head osteonecrosis. We have therefore decided to focus on these. In combination, they represent early femoral component failures after HRA. Proximal femoral bone preservation in young active patients is the primary reason that metal-on-metal HRA was developed. However, if the risks of early femoral failures are particularly high in a certain group of patients, they may be considered poor candidates for HRA. If the alternative risks of amputating the femoral head and neck to perform a stemmed THA are much lower in this group, the theoretical advantage of bone preservation with HRA in younger patients may no longer be worthwhile. Numerous studies have focused on delineating risk factors for HSR to help the surgeon decide which patients may have too high a risk with HRA to make proximal femoral bone preservation worthwhile [6, 10, 14, 15]. Unfortunately, it is not always clear exactly why a certain risk factor is problematic. Does a smaller component size lead to more problems because of a small area of femoral fixation⁵? Or is the problem with smaller components primarily because of more adverse wear problems [10, 16, 17]. The present retrospective case-control study was specifically undertaken to assess one proposed risk factor for early femoral failures: Does the presence of femoral head cysts increase the risk of early femoral failure?

Cysts in the femoral head are areas where bone loss has occurred due to the arthritic process. Therefore, it is generally believed by experts that femoral head cysts negatively impact the success rate of HRA [6, 9, 10]. However, to our knowledge, only few papers have reported scientific evidence that femoral head cysts are a risk factor for HRA ⁵. Because it seems logical that cysts might affect femoral fixation, this belief has largely gone unchallenged, despite the fact that the evidence available is limited. Beaule et al ⁵ proposed a SARI on the basis of a study of 92 HRAs done in patients under 40 years of age. The average follow-up was 3 years (range: 2-5.6 years). Survivorship with revision for early femoral failure as an endpoint was 97% (two femoral neck fractures, one femoral loosening). There were two additional radiographically loose femoral components (migration) and eight additional possibly loose femoral components (complete stem radiolucency). This formed the problematic group (N = 13). A univariate analysis of multiple risk factors was done. Points were assigned to certain risk factors based on their odds ratio in this analysis. Two points were assigned for cysts > 1 cm³, 2 points for weight under 82 kg, one point for UCLA Activity score above 6, and one point for previous hip surgery. The maximum score was 6. The SARI was found to be significantly higher in the 13 problematic hips than in the remainder of the hips in the series (P < 0.001). Femoral head cysts were found in 53% of well-functioning hips while they were present in 92% of problematic hips (P = 0.028). Their data implicate the presence of femoral head cysts (>1 cm³) as a risk factor for HRA. It does not quantify the added risk for failure due to cysts. Also, the cysts in Beaule's study were managed by debridement and filling with cement.

In summary, our study, with a control group matched for other previously proposed risk factors for early femoral loosening, could not demonstrate that femoral head cysts were an independent negative risk factor for failure of the femoral resurfacing component. However, we caution that this may be due to the way we treat femoral head cysts with bone grafting, rather than filling them with cement. We therefore recommend that the presence of cysts within the femoral head, as long as they comprise less than 1/3 of the remaining prepared femoral head, be eliminated as a risk factor for HRA. We suggest that other surgeons consider bone grafting cysts rather than filling them with cement. Comparison studies to further compare these two techniques would be valuable.