Improved technology and a greater understanding of the biomechanics of metal-on-metal bearings have led to a resurgence in the use of metal resurfacing techniques for hip replacement procedures1,2. Metal-on-metal resurfacing is an attractive procedure that allows bone conservation and may lead to increased stability and hip motion because of the large femoral head size3. These properties are especially advantageous for young, physically active patients3,4.
One of the potential concerns with metal-on-metal resurfacing is metal ion release5. Elevated cobalt and chromium ion levels have been documented in the blood, serum, and urine of patients who have undergone hip replacement procedures with metal articulations1,6-30. The levels detected vary depending on the implant system, the time since implantation, the analytical technique, and the medium used to determine metal ion levels31. Currently, the clinical importance of these elevated ion levels remains unclear, although recent studies have shown a link between highly elevated metal ion levels, implant dysfunction, and local adverse soft-tissue reactions32,33. Case reports of systemic toxicity as well as benign and even malignant tumor development have been reported34-38. The carcinogenic and mutagenic properties of chromium and cobalt have been examined in the literature39-41. Several epidemiological studies have examined the risk of cancer following the implantation of metal-on-metal hip implants; however, these results have not been consistent, and controversies persist42-52.
The purpose of the present study was to prospectively determine serum, erythrocyte, and urinary levels of cobalt and chromium in patients who underwent hip resurfacing with the Conserve Plus total resurfacing hip system (Wright Medical Technology, Arlington, Tennessee).
All patients who were eligible for hip resurfacing arthroplasty because of noninflammatory degenerative joint disease between July 2004 and June 2006 were considered for inclusion. Patients were excluded if they were pregnant or lactating, had other metal implants, had known metal allergies, had renal dysfunction, or had neurologic or musculoskeletal disease that may adversely affect gait or weight-bearing. Three surgeons in two centers performed all procedures. A total of ninety-seven patients were enrolled. Eighty-five procedures were performed by two surgeons with previous hip resurfacing experience. The third surgeon, who performed the remaining twelve procedures, had limited hip resurfacing experience prior to the study. Blood and urine samples were obtained preoperatively and at three, six, twelve, and twenty-four months after surgery. All patients completed a health survey at the six, twelve, and twenty-four-month follow-up visits.
Hip Prosthesis
All eligible patients were managed with the Conserve Plus metal-on-metal hip resurfacing system. This system consists of an acetabular component and femoral component made of a high-carbon cast cobalt-chromium alloy conforming to ASTM F75 and containing 27% to 30% chromium, 50% to 60% cobalt, and 5% to 7% molybdenum. Nickel, iron, carbon, silicon, and manganese compose the remainder of the alloy. The castings undergo two heat treatment regimens prior to final machining and polishing. Hot isostatic pressing is conducted in order to eliminate tiny voids left in the castings during the cooling process. Solution annealing is conducted for the dissolution of large blocky carbides into the matrix. The arc of coverage of the inner articular bearing surface of the acetabular shell ranges from 159° to 164°, depending on cup size. The wall thickness of cups used in this series was 3.5 mm at the equator.
Sample Collection
Three 10-mL syringes of blood were collected from each subject. Each syringe was labeled to indicate the sequence of collection. Prior to sample collection, all collection containers, syringes, and apparatus were triple acid-washed with Ultrex-grade chemicals and were verified for the absence of contamination by flushing with double-deionized water, followed by analysis for residual trace metals. The samples of blood were maintained at room temperature until they had fully clotted (approximately twenty minutes) and then were centrifuged at 1850 × g for thirty minutes. Serum and clot fractions were then separated, and the fractions were frozen and stored at —80°C in polypropylene tubes until analysis. All manipulations of the specimens were carried out in a class-100 environment with a Sterilguard Hood (Baker, Sandford, Maine) and class-100 gloves.
Urine samples were collected from the patients by providing acid-washed polypropylene containers and detailed instructions for the collection and storage of urine in a manner that avoided contamination. After discarding the first morning urine on the first day, patients collected all urine, including the first morning urine on the next day, in the container. Female patients were provided with acid-washed Specipans (Covidien, Mansfield, Massachusetts) from which urine was transferred into the polypropylene containers. Samples of urine were labeled, frozen, and stored at —80°C before analysis.
Analysis of Samples
As a precautionary measure, the first 10 mL of blood drawn was used to rinse the needle; this blood was used for determination of total serum protein values. The remaining two syringes of serum and erythrocytes were shipped to the Trace Elements Laboratory at the University of Western Ontario for elemental analysis with use of a high-resolution inductively coupled mass spectrometer (HR-ICPMS). Blood and urine samples were analyzed for creatinine, cobalt, and chromium levels.
Other Data
Activity level was measured using the University of California, Los Angeles (UCLA) activity score53 preoperatively and at six, twelve, and twenty-four months after surgery. Functional outcome was assessed with use of the Harris hip score at the same time intervals54. All patients had a standard anteroposterior pelvic radiograph as well as a cross-table lateral radiograph of the affected hip to evaluate the cup abduction angle and femoral component orientation at these same time intervals55-57.
Statistical Analysis
The mean and median levels of blood and urinary chromium and cobalt ions were calculated before surgery (to provide a baseline indicator of the concentrations of these ions) and at three, six, twelve, and twenty-four months after surgery. Because the Shapiro-Wilk test indicated a non-normal distribution of ion levels at baseline and follow-up, nonparametric tests were used. Wilcoxon signed-rank tests were used for longitudinal comparisons of the ion levels at three, six, twelve, and twenty-four months after surgery with the baseline levels obtained before surgery. Spearman rank-order correlation coefficients (?) were used to examine correlations between erythrocyte, serum, and urinary ion levels. The Kruskal-Wallis test was used to determine whether cobalt and chromium ion levels differed by patient characteristics (sex and body-mass index [BMI]). A two-tailed p value of 0.05 was used for all statistical tests. All analyses were performed with use of SAS version 9.1 (SAS Institute, Cary, North Carolina)58. Ethics approval was obtained from the Ottawa Hospital Research Ethics Board.
A total of ninety-seven patients were recruited into the study, including seventy-six men and twenty-one women with a mean age (and standard deviation) of 48 ± 8.8 years (range, eighteen to sixty-five years) and a mean BMI of 27.6 ± 4.1 kg/m2 (range, 19.1 to 39.9 kg/m2). The majority of patients (eighty-four patients; 86.6%) had an underlying diagnosis of osteoarthritis. Three patients underwent revision surgery because of adverse events (cup or femoral component loosening) during the two-year postoperative follow-up period. Twenty-one patients were lost to follow-up. One patient died of an unrelated cause at three months. The remaining reasons for attrition are listed in Table I.
Radiographic analysis demonstrated a mean cup abduction angle of 46.0° ± 5.4° (range, 26° to 62°). The mean femoral stem-shaft angle was 136.9° ± 5.9° (range, 120° to 157°) on the anteroposterior pelvic radiograph and 10.3° ± 6.8° of anteversion (range, 0° to 35° of anteversion) on the cross-table lateral radiograph. The mean size of the acetabular component was 56 ± 3.6 mm (range, 48 to 62 mm). The mean UCLA activity score improved from 6.0 ± 2.6 (range, 2 to 10) preoperatively to 7.1 ± 2.1 (range, 2 to 10) at two years of follow-up (p = 0.002). There was no significant change between the mean preoperative and postoperative urinary creatinine levels (9.6 ± 4.4 mmol/L [range, 2.1 to 21.6 mmol/L], compared with 9.7 ± 4.4 mmol/L [range, 3.0 to 21.3 mmol/L]) (p > 0.05). The median (and mean) levels of cobalt and chromium ions in erythrocytes, serum, and urine are presented in Table II. Box plots for chromium and cobalt levels in serum, erythrocytes, and urine are provided in Figures 1-A through 1-F.
Figs. 1-A through 1-F Box plots showing serum, erythrocyte, and urine levels of chromium (Figs. 1-A, 1-B, and 1-C) and cobalt (Figs. 1-D, 1-E, and 1-F) preoperatively (time = 0) and at three, six, twelve, and twenty-four months after the implantation of a Conserve Plus metal-on-metal hip resurfacing device. The inferior and superior horizontal lines of the box represent the first and third quartiles and the line inside the box represents the median. The ends of the whiskers correspond to the limits of the data, to a distance of at most 1.5 interquartile ranges, beyond which values are considered anomalous. Outliers are represented by circles (°) and extreme outliers are represented by asterisks (*).
Serum Ion Levels
The median preoperative serum cobalt and chromium levels were 0.08 µg/L (range, 0.03 to 0.86 µg/L) and 0.19 µg/L (range, 0.10 to 0.93 µg/L), respectively. Both serum cobalt and chromium levels were significantly elevated postoperatively (p < 0.001), with cobalt levels being thirteenfold and fourteenfold higher at one and two years (median, 1.04 and 1.08 µg/L, respectively) and chromium levels being elevenfold and ninefold higher at one and two years, respectively (median, 2.00 and 1.64 µg/L, respectively).
Erythrocyte Ion Levels
The median preoperative erythrocyte cobalt and chromium levels were 0.10 µg/L (range, 0.04 to 0.41 µg/L) and 1.0 µg/L (range, 0.50 to 4.10 µg/L), respectively. The postoperative median erythrocyte cobalt levels were significantly elevated at all time intervals (p < 0.0001) and increased by approximately eightfold at both one and two years. The median erythrocyte chromium levels were slightly higher in the postoperative period compared with preoperative levels; however, the increase was significant only at three months postoperatively (median, 1.30 µg/L) (p = 0.008).
Urinary Ion Levels
The median preoperative urinary cobalt and chromium levels were 0.27 µg/L (range, 0.03 to 3.48 µg/L) and 0.19 µg/L (range, 0.01 to 1.26 µg/L), respectively. The postoperative median urinary cobalt levels were significantly elevated at all time intervals (p < 0.0001), rising thirty-one-fold higher than preoperative levels at one year and thirty-sevenfold higher at two years postoperatively. The postoperative median levels of chromium in urine were significantly elevated at all follow-up time points (p < 0.0001), being eighteenfold higher than preoperative levels at one year and twenty-twofold higher at two years of follow-up.
Ion Level Relationships
There was a strong correlation between serum and urinary cobalt levels at one and two years after surgery (? = 0.7 [p < 0.0001] and ? = 0.8 [p < 0.0001], respectively). A similar although somewhat weaker correlation between serum and urinary chromium levels was also observed at one year (? = 0.5, p = 0.0007) and two years (? = 0.8, p < 0.0001) after surgery. A strong correlation between erythrocyte and urinary cobalt levels was also apparent at one year (? = 0.5, p = 0.0002) and two years (? = 0.7, p = 0.0002) after surgery. No significant association between erythrocyte and urinary chromium levels was seen. There was a strong correlation between serum and erythrocyte cobalt levels (? = 0.9, p < 0.0001) at one and two years. However, a relatively weak correlation was observed between serum and erythrocyte chromium levels at one year (? = 0.4, p < 0.002) and two years (? = 0.4, p < 0.01). Ion levels were not correlated with other individual characteristics, including demographic factors (age, sex, height, weight, and BMI), UCLA activity level, shell size (outer diameter), acetabular cup inclination, and stem-shaft angle as measured on anteroposterior and cross-table lateral radiographs (p > 0.05).
Clinical Results
The mean Harris hip score was 54.31 ± 11.47 (range, 29 to 75) preoperatively and 89.94 ± 13.97 (range, 40.0 to 100.0) at the two-year follow-up. The mean UCLA activity level score was 6.0 ± 2.6 (range, 2 to 10) preoperatively and 7.2 ± 2.0 (range, 2 to 10) at the two-year follow-up. Three patients (3.1%) required revision within the first two postoperative years because of acetabular loosening (two patients) and femoral neck fracture (one patient). There were no radiographic signs of loosening or implant migration in the remaining patients. No pseudotumors or other adverse metal reactions were encountered in any patient.
Note: Daniel Krewski, PhD, MHA, is the Natural Sciences and Engineering Research Council of Canada Chair in Risk Science at the University of Ottawa.