Porous Tantalum Implant in Early Osteonecrosis of the Hip: Preliminary Report on Operative, Survival, and Outcomes Results

Materials and Methods 

After obtaining approval from the institutional review board, data were collected on all patients receiving a porous tantalum implant. A minimum follow-up of 2 years was required for study inclusion.

The porous tantalum prosthesis population was obtained from a prospective, consecutive group of patients treated for UPenn stage I or II osteonecrosis [20] from June 1, 2000, through May 1, 2003. After surgical consent was obtained, each patient received a porous tantalum prosthesis (Trabecular Metal, Zimmer) (Fig. 1) in keeping with the implant surgical technique description. Postoperatively, patients were limited to touch down weight bearing and 2-point crutch ambulation for 6 weeks. After the initial protected weight-bearing period, patients began full weight-bearing as tolerated.

View Large Image Download to PowerPoint

Fig. 1. Postoperative radiograph of the porous tantalum implant.

The historical vascularized fibular graft subjects underwent surgery from April 1, 1990, through July 1, 1996. All patients were diagnosed with UPenn stage I or II osteonecrosis of the femoral head. A single team of orthopedic surgeons performed the core decompression, fibular harvesting, and graft insertion on all patients. The Urbaniak technique was used by an experienced microsurgeon (Fig. 2) [8]. Patients were limited to touch down weight bearing with 2-point crutch ambulation for 6 months postoperatively. Weight bearing was then advanced as tolerated after that period.

View Large Image Download to PowerPoint

Fig. 2. Postoperative radiograph of vascularized fibular graft.

The inclusion criteria for both treatment groups were as follows. Patients were between the ages of 18 and 60 years with a body mass index (BMI) of less than 40 (per National Institutes of Health determination of BMI) and at least 2 years of follow-up. Patients with stages more advanced than UPenn stage II receiving fibular grafts were excluded. Patients with a history of previous core decompression, bone grafting, proximal femoral osteotomy, or internal fixation in the affected hip were also excluded. Any patients who had undergone a previous treatment for avascular necrosis in the affected hip, such as electromagnetic and ultrasound stimulation, or taken medications intended to intervene in or treat the disease were also excluded.

Demographics including age, BMI, sex, contralateral disease, etiology, and UPenn classification stage were obtained for each patient. Surgical data including time of surgery, blood loss, and cell transfusions were recorded. Postoperative values were measured for hospitalization length as well as days requiring a patient-controlled analgesia (PCA) pump.

Time to failure was obtained for patients in both groups, with failure defined as patients undergoing a total hip arthroplasty (THA). The Harris Hip Score was used to measure outcomes for each patient. Patients were placed into 4 groups—excellent, good, fair, and poor—based on their outcome scores 21, 22. Femoral head collapse requiring THA was defined as a poor outcome and failure. Patients were excluded from outcome analysis if the patient was lost to follow-up or died.

Both populations consisted of all patients fitting the inclusion criteria over the stated time period. A comparison of differences between the 2 populations was made with the Student t test for continuous variables such as demographic data, surgical time, blood loss, hospitalization time, blood transfusion, and usage of PCA and transfusions. Using a Kaplan-Meier curve, we constructed a survival curve for each treatment group. Failure was defined as conversion to a THA. Confidence intervals of 95% were obtained, and a log rank test statistic 1.59 on 1 degree of freedom was performed for each curve. In all tests, an α level of .05 was deemed to be statistically significant. Outcome categories were obtained through the guidelines established with the Harris Hip Score questionnaire.

Results 

For the porous tantalum implant group, the average surgical time was 36 minutes with a mean blood loss of 70 mL. Average hospitalization was less than 1 day (0.8 days) with most patients being treated on an outpatient basis. There were no patients who required a PCA or blood transfusion (Table 1). In all implant cases, pain was controlled with oral analgesics and intravenous push.

Table 1. Intraoperative and Initial Postoperative Data for the Porous Tantalum Implant and the Vascularized Fibular Graft Populations

		Surgical time (min)	Blood loss (mL)	Hospital days (d)	Blood transfusions (U)	PCA pump usage (d)
	Porous tantalum implant	36 (15-71)	70 (25-200)	0.8 (0.5-1)	0 (0)	0 (0)
	Vascularized fibular graft	506 (425-687)	735 (250-1500)	9.5 (8-11)	1.4 (0-6)	4.0 (2-7)
	Difference	471	665	8.7	1.4	4
	P	<.00001	<.00001	<.00001	<.00001	<.00001

In comparison to the historical group, on average the fibular graft procedure required longer surgical time (471 minutes), larger blood loss (665 mL), longer hospitalization (8.7 days), more transfusions (1.4 U), and more PCA usage (4 days). All parameters were significant with a P value < .0001 (Table 1). One procedure in the fibular graft population was aborted because of poor vascular integrity intraoperatively. This patient's results were excluded from this analysis.

The 2 study populations were not significantly different in any demographic parameters. Age, BMI, sex, bilateral disease, etiology, and UPenn staging were compared (Table 2). There were a total of 30 fibular graft procedures during the study period. Nine hips were stage III osteonecrosis and were excluded.

Table 2. Patient Demographic Data

	(A) Demographics for 2 populations
	Demographics	n	Mean age (y)	Mean BMI	Sex (male)	Bilateral disease
	Porous tantalum implant	24	43.2 (28-60)	27.6 (22-37)	20 (83%)	24 (100%)
	Vascularized fibular graft	21	37.9 (19-50)	26.3 (19-32)	15 (71%)	11 (52%)
	(B) Etiology breakdown for each population
	Etiology	Idiopathic	Steroids	EtOH	Trauma
	Porous tantalum implant	8 (33%)	6 (25%)	10 (42%)
	Vascularized fibular graft	1 (5%)	14 (67%)	3 (14%)	3 (14%)
	(C) UPenn staging for each population
	UPenn stage	n	I	II	III	IV
	Porous tantalum implant	24	2 (8%)	22 (92%)	0 (0%)	0 (0%)
	Vascularized fibular graft	21	0 (0%)	21 (100%)	0 (0%)	0 (0%)

With a minimum of 2 years of follow-up and an average follow-up of 39 months (range, 27-59 months), the tantalum implant group had an 86% survival rate. Three of 22 had progressive pain and collapse requiring conversion to a THA. Two others were lost to follow-up after 6 months and were excluded (Table 3).

Table 3. Survival and Outcomes Data

	(A) Survival data at 39 months' follow-up
	Survival	Follow-up (m)	Loss to follow-up	Deaths	n	Failures (THA)	Survivors
	Porous tantalum implant	39	2	0	22	3 (14%)	19 (86%)
	Vascularized fibular graft	39	2	1	18	6 (33%)	12 (67%)
	(B) Clinical outcomes for both groups measured by Harris Hip Scores
	Outcomes	Average follow-up (m)	Failed	Poor	Fair	Good	Excellent
	Porous tantalum implant	39 (27-59)	3 (14%)	0 (0%)	0 (0%)	3 (14%)	16 (72%)
	Vascularized fibular graft	149 (121-180)	7 (39%)	7 (39%)	1 (6%)	2 (11%)	1 (6%)

Kaplan-Meier analysis at 39 months resulted in an 86% (95% CI, 73-100) and a 67% (95% CI, 48-92) survivorship for the tantalum implant and fibular graft, respectively. A log rank test statistic 1.59 on 1 degree of freedom was performed with a resulting P value of .21. There was not a statistical difference in survival at 39 months of follow-up.

Using the Harris Hip Score to characterize outcome, the implant group had 19 (86%) good-to-excellent results. The remaining 3 (14%) were classified as poor outcomes because they were converted to a THA. The fibular graft population had 3 (17%) good or excellent outcomes, and 14 (78%) were classified as poor at the 39-month follow-up point (Table 3).

There were no complications associated with the implant at the time of the study. There were no postoperative infections in either population. There were 4 complications in the fibular graft population. One patient failed intraoperatively and was aborted because of poor vasculature. One patient developed pneumonia secondary to overresuscitation. Two complications involved the harvest site, including a peroneal neuropraxia and a flexor hallucis longus contracture, which required a release.

Discussion 

The purpose of this paper was to examine the success of porous tantalum metal implants in the treatment of early-stage osteonecrosis. To date, vascularized fibular grafts have demonstrated the highest rates of success in treating early-stage osteonecrosis of the femoral head. Therefore, the implant population was also compared to a historical benchmark, the vascularized fibular graft population, from the same institute regarding operative data, as well as survivorship and outcome results.

Statistical analysis of operative data revealed significant differences in favor of the implant population. The operative time, blood loss, PCA usage, and hospitalization were significantly less for the implant group when compared to fibular grafting. More than half the tantalum implants were performed on an outpatient basis.

Previous vascularized fibular graft studies have shown success rates as high as 90% in early-stage osteonecrosis at 5 years 8, 15. Other reports have had poorer results with only a 52% rate of success at an average follow-up of 18 years [23]. A real concern with vascularized grafting is the significant complication rate. Moderate to severe pain has been reported as high as 15% of patients undergoing partial fibulectomy, whereas only 39% were described as asymptomatic 24, 25. A portion of patients have also described a feeling of instability (42%) and weakness (37%) at 59 months after surgery [26].

In our study, the tantalum implant had an 86% survival rate at an average follow-up of 39 months. All implant survivors were rated as good to excellent. The vascularized fibular grafting group had a 67% survival rate at 39 months (P < .21). But of the surviving fibular grafts, only 17% were classified as good or excellent outcomes. There was also a 22% complication rate associated with the fibular graft population with half pertaining to the harvest site. These vascularized fibular graft results are lower than some previously described results. Better outcomes at other institutions may be attributable to a higher volume and experience.

High success rates for vascularized fibular grafting are obtainable by a select few with extensive experience. Our data support the idea that the implant procedure is less invasive and far more accessible to the general orthopedic community with a smaller learning curve. There is no need for microsurgical expertise or harvesting with the implant either, which eliminates harvest morbidity and complications.

Revision to THA was required in both groups. There were no complications associated with revision from the porous tantalum implant to arthroplasty. The standard femoral cut using a standard oscillating saw and blade was made through the femoral neck and implant. The remaining lateral implant piece was accessed through the femoral cut and removed with thin osteotomes. Packing, copious irrigation, and suction were used to prevent dispersion of tantalum particles in the surgical wound. The femoral head was morselized and used for bone graft to fill any remaining bony defects laterally. At the time of submission, there have been no complications or needs for revision in the 3 patients requiring conversion to arthroplasty. There is a theoretical concern for third-particle wear, but this phenomenon has not been observed clinically up to 3 years postconversion.

The primary limitations of this study revolve around the use of a historical control group. Economic pressures and technological advances over the interval period could have helped contribute to the differences between the 2 groups. Because preoperative Harris Hip Scores were not obtained for the vascularized graft group, changes due to operative intervention could not be calculated and compared. Lastly, the size and location of the lesions were not recorded. Therefore, differences between the 2 groups may exist that were not appreciated regarding subsets of classifications.

With the current data, there are several conclusions that can be made. In patients with early-stage osteonecrosis of the hip, the porous tantalum metal implant offers a less invasive method for structural support of the subcondral bone after decompression. Without fibular harvesting and microsurgical anastemosis, the porous metal implant has less co-morbidity and a shorter postoperative recovery period. Preliminary data show the tantalum implant is at least equivalent to the vascularized fibular graft in terms of survival and outcomes. Trabecular metal is a straightforward and viable option that may offer more reliable and reproducible results. Long-term follow-up is needed before a final determination can be made.