|Year : 2022 | Volume
| Issue : 1 | Page : 75-80
Predicting short-term outcome of Metal-on-Metal Hip Resurfacing (MOMHR): A multivariate analysis using 14 independent variables
Amrit Goyal1, William Macaulay2, Jeffrey A Geller3, Wenbao Wang4, Jonathon D Nyce5
1 Department of Orthopaedics, SMMH Medical College, India
2 Chief, Division of Adult Reconstruction, NYU Langone Health, NY, USA
3 Chief, Division of Hip & Knee Reconstruction, Columbia University Medical Center, NY, USA
4 Physician, Baylor Scott & White Physical Medicine and Rehabilitation, TX, USA
5 Icahn School of Medicine Mount Sinai, NY, USA
|Date of Submission||24-Mar-2021|
|Date of Acceptance||10-Nov-2021|
|Date of Web Publication||15-Jun-2022|
Dr. Amrit Goyal
Department of Orthopaedics, SMMH Medical College, Ambala Road, Saharanpur - 247 232, Uttar Pradesh
Source of Support: None, Conflict of Interest: None
Introduction: The aim of this study was to research factors affecting the short-term outcome of metal-on-metal hip resurfacing (MOMHR) and develop a multivariate regression model that may predict outcome. Materials and Methods: This was a prospective study of 154 patients who underwent MOMHR and were followed for a minimum of 1 year. Fourteen independent variables (age, gender, diagnosis, co-morbidities, body mass index (BMI), pr-operative Western Ontario and McMaster Universities Osteoarthritis (WOMAC) physical component/stiffness (S)/pain (P), short form 12 (SF-12) physical (SP), SF-12 mental (SM), acetabular and femoral component sizes, operative time, and estimated blood loss) were analyzed using correlation and multivariate regression analyses. Multivariate regression model was tested by using an independent cohort for validation. Results: Correlation analyses found four variables that significantly influence short term MOMHR outcome. These include comorbidities (C, P = 0.0001), preoperative SF-12 mental (SM, P = 0.0004), BMI (P = 0.0006), and gender (G, P = 0.0454). By multivariate analysis, the subsequent regression model was obtained with an R2 value of 0.3816: Outcome = G*4.72 ‒ BMI*0.70 ‒ C*0.11 + SM*0.31 + 87.44. The average predicted outcome using this equation did not differ significantly from the observed WOMAC physical function outcome at a minimum of 1 year postoperatively. Conclusion: To the best of our knowledge, this study is the first reported multivariate analysis of factors affecting MOMHR and confirms the correlation of some of the previously proposed factors such as gender, BMI, comorbidities, and preoperative function. The multivariate regression equation can be used to predict the short-term outcome of MOMHR.
Keywords: Arthroplasty, hip resurfacing, outcome, patient factors
|How to cite this article:|
Goyal A, Macaulay W, Geller JA, Wang W, Nyce JD. Predicting short-term outcome of Metal-on-Metal Hip Resurfacing (MOMHR): A multivariate analysis using 14 independent variables. J Orthop Traumatol Rehabil 2022;14:75-80
|How to cite this URL:|
Goyal A, Macaulay W, Geller JA, Wang W, Nyce JD. Predicting short-term outcome of Metal-on-Metal Hip Resurfacing (MOMHR): A multivariate analysis using 14 independent variables. J Orthop Traumatol Rehabil [serial online] 2022 [cited 2022 Jun 26];14:75-80. Available from: https://www.jotr.in/text.asp?2022/14/1/75/347359
| Introduction|| |
Metal-on-metal hip resurfacing (MOMHR) arthroplasty had gained popularity in the past decade as a treatment for degenerative osteoarthritis, especially for younger active patients. The purported advantages over total hip replacement (THR) are supposedly increased joint motion range, better proprioception and less chances of dislocation. Further, because the proximal femoral bone stock is preserved, revision to a THR is easier if required.,
Yet despite the advantages of hip resurfacing arthroplasty, the results are not uniform in all patients. Various identifiable preoperative factors play a role in the outcome after surgery. In many THR studies, outcomes have been shown to be affected by preoperative patient variables such as age, gender, body mass index (BMI), preoperative comorbidities, and activity status.,,,,,,,, However, there is still disagreement regarding patient selection and factors that would improve clinical outcome for surface replacement arthroplasty. In this study, we attempt to find and quantify the patient factors governing the predictability of outcome of surface replacement. This would help to optimize patient selection and also help surgeons predict a realistic outcome for their high demand patients. To the best of our knowledge there is no study in current literature which used multivariate analysis to find significant factors affecting the outcome of MOMHR.
This study attempts to prospectively analyze 14 independent factors, find their association with postoperative functional outcome using univariate and multivariate regression analysis, and find an equation which can aid surgeons in the prediction of expected outcome after the surgery. The factors: age, gender, diagnosis, presence of preoperative comorbidities, BMI, preoperative Western Ontario, and McMaster Universities Osteoarthritis (WOMAC) physical function (PF), pain and stiffness, short-form 12 (SF-12) physical score (SP), SF-12 mental score, acetabular cup size, femoral head size, surgical time, and estimated blood loss (EBL) were examined using correlation and multivariate regression analysis.
| Materials and Methods|| |
All data for this study were prospectively collected under the Center for Hip and Knee replacement joint registry which has full Institutional Review Board approval. Patients were consented and enrolled in the study if they underwent an elective Birmingham MOMHR between 2006 and 2009 done by two senior surgeons here at our center. All patients enrolled in the study were initially asked to complete preoperative WOMAC and SF-12 questionnaires. In addition, age, gender, diagnosis, presence of preoperative comorbidities, BMI, acetabular cup size, femoral head size, surgical time, and EBL were collected preoperatively and intraoperativerly by the surgeon or fellow. Comorbidities included alcohol dependency, smoking, cancer, cardiac disease, endocrine disorders such as diabetes mellitus, hypothyroidism, gastrointestinal, hematologic and hepatobiliary diseases, hypertension, infectious disease, neurological disease, osteoporosis, Parkinson's disease, psychiatric disease, respiratory disease, use of steroids, thromboembolis disease, and vascular disease.
Patients completed postoperative WOMAC and SF-12 questionnaires at standard office follow-up visits: at 3 months (±1 month), 1 year (±3 months), and 2 years (±3 months) postoperatively. All data collection and maintenance was performed using Patient Analysis and Tracking System (PATS 4.0) software (Axis Clinical Software, Portland, Ore).
WOMAC PF score at the end of 1 year was the main outcome measure. The outcome WOMAC PF score was tested for correlation to each independent variable using a nonsimple correlation analysis (significance level of 0.05). Multivariate regression analysis was then used to analyze those independent variables that were found to have a significant effect on outcome WOMAC PF score. Variables that did not meet criteria for significance in multivariate regression analysis (P > 0.15) were identified as confounding variables. Multivariate regression analysis was used to find the desired equation for predicting the outcome WOMAC PF score. All significant factors were used to form the equation using a stepwise procedure. All quantitative variables (age, BMI, preoperative WOMAC score, length of stay, surgery time, and EBL) were analyzed using their original values. Codes were used for the qualitative variables. Microsoft Office Excel (Microsoft, Redmond, WA) and SAS 9.1 software (SAS Institute, Cary, NC) was used to find the correlation and do multivariate linear regression analysis.
| Results|| |
A total of 154 patients with resurfacing arthroplasty were included in the study. All patient data analyzed used follow-up at 1 year. The study group was predominantly male, 109 (70.7%) [Table 1]. The mean age (and standard deviation) of the patients was 50.2 ± 8.55 years with a mean BMI of 27.99 ± 4.72. The preoperative diagnosis was primary degenerative osteoarthritis in 112 (72.7%) patients, avascular necrosis 17 (11%) patients, dysplasia 13 (8%) patients and inflammatory arthritis in the rest of the patients. A total of 72 patients (46.7%) had more than one preoperative comorbidity as specified earlier. All patients received a Birmingham hip resurfacing implant (Smith and Nephew) with uncemented cup and a cemented femoral component. Regional spinal anesthesia was used in all cases. The operative time was 109 ± 21.2 min with an EBL of 383 ± 161 ml.
|Table 1: Analysis of Independent variables having a significant effect on outcome WOMAC PF using multivariate regression|
Click here to view
The average acetabular component size was 54.74 ± 3.95 with an average femoral component size was 48.14 ± 4.09. The preoperative SP and SF mental scores were 33.2 ± 7.72 and 50.19 ± 11.63, respectively. The average preoperative WOMAC pain, stiffness, and function were 48 ± 21.2, 43 ± 22.1, and 50 ± 20.9, respectively. At 1 year follow-up, the WOMAC PF average was 91.36 ± 13.86.
Nonsimple correlation analysis identified eight factors which affected the postoperative SP function outcome: BMI, femoral component size, preoperative diagnosis, operative time, preoperative comorbidities, EBL, preoperative SF-12 mental score, and preoperative WOMAC pain score. A summary of the univariate analysis results can be found in [Table 1].
Multivariate regression analysis was performed and only four factors: gender (G), BMI, preoperative comorbidities (CMB), and preoperative SF mental function (SF [MF]) score were found to have a correlation with the outcome at 1 year with a significance level of 0.05. Using the stepwise regression analysis, femoral component size, preoperative diagnosis, operative time, EBL, and preoperative WOMAC pain scores were found to be confounding factors affecting MOMHR outcomes. Gender, not originally significant in the univariate analysis, became a significant factor in the multivariate regression analysis after eliminating the confounding variables. The following regression equation was obtained:
Outcome score = 87.43 + (G*4.72) ‒ (BMI*0.69) ‒ (CMB*0.114) + (SF [MF]*0.309) where Gender is defined by female (1) or male (2) and comorbidity by absent (1) or present (2). Regression analysis of predicted versus observed WOMAC PF scores at 1 year of follow-up resulted in an R2 value of 0.3816.
Comorbidities were found to be most significant factor effecting postoperative outcome (P < 0.0001) followed in significance by SF (MF) (P = 0.0004), BMI (P = 0.0006), and gender (P = 0.0454). The predicted outcome score did not differ significantly from the observed score when compared at 1 year follow-up [Figure 1], [Figure 2], [Figure 3], [Figure 4]. The predicted outcome score had an average value of 91.39 ± 5.35 in comparison to the average WOMAC observed value at 1 year of 91.36 ± 13.86.
|Figure 1: Short form 12 mental score versus observed physical function score|
Click here to view
When compared to an independent cohort of patients [Figure 5], the equation exhibited positive correlation between predicted and observed 1 year outcomes. It demonstrated an average predicted score of 91.39 compared to the average observed score of 91.36. The selection bias between the small independent cohort and the rest of our study population was negligible.
|Figure 5: Comparison of predicted and observed outcomes in an independent study cohort|
Click here to view
| Discussion|| |
We used multivariate regression analysis to prospectively study the factors affecting outcome of the surface replacement arthroplasty. After initial univariate analysis of our data, 8 of 14 factors were found to significantly affect short-term functional outcomes. Upon further multivariate analysis of all 14 factors, only four factors were found to significantly alter predicted outcomes at 1 year of follow-up: gender, BMI, preoperative comorbidities, and preoperative SF (MF) score while the rest were eliminated as confounding variables. The equation generated from the multivariate regression analysis, outcome score = 87.43 + (G*4.72) ‒ (BMI*0.69) ‒ (CMB*0.114) + (SF [MF]) * 0.309), takes into account the varying degrees of impact each factor has to predict the short-term outcomes. With an R2 value of 0.3816, the equation can be used to accurately predict the short-term functional outcomes of patients who have undergone a MOMHR arthroplasty.
The impact of gender on the risk of femoral neck fracture, prosthesis survival, and poor outcomes has been documented in MOMHR and THA studies.,,,,,,, Katz et al. suggested that women at the time of hip surgery had worse preoperative status and Holtzman et al. found that female gender was associated with a poor outcome after THR compared to the male gender in their respective studies. However, Gabriel et al. found no difference of gender for THA outcome. MacWilliam et al. attributed a higher risk of osteoporosis in women as a negative factor for THA outcome. When controlling for 12 independent factors, Wang et al. found that female gender was associated with lower postoperative outcomes.
Amstutz et al. found a greater percentage of revision in their female patients in surface replacement arthroplasty compared to the male patients. Mcbryde et al. found similar results in their study showing an increased incidence of revision and femoral neck fracture after MOMHR in female patients. They hypothesized that this difference was due to smaller femoral component size and was not related to the difference in gender. To the contrary, Khan et al. did not find any difference in terms of prosthesis survival between genders but found that female patients undergoing resurfacing had lower postoperative scores especially in mobility scores compared to their male counterparts. In a review of 50 MOMHR cases of periprosthetic femoral neck fractures, Shimmin et al. found that the incidence of femoral neck fracture was 1.94 the relative risk in females as compared to males; a result duplicated by Marker et al. In contrast to McBryde's hypothesis, they did not find any correlation between smaller femoral component head size and femoral neck fractures in women. In attempt to clarify the contradicting results from these previous studies, our study attempted to control for any demographic and clinical characteristics between genders so as to eliminate any potential confounding variables. Further, component size was not found to significantly affect the short-term outcome in our multivariate analysis. Our study helps to clarify other conflicting reports showing that gender, independent of any confounding factors, ultimately affects the short-term functional outcome of MOMHR patients. Mcbryde et al. did not find any association between factors such as sex, age, surgeon experience, surgical approach or component size and outcome hip scores.
Paralleling the results seen in many THR studies, our study identifies preoperative mental status as a predictor of short-term outcomes. Studies done by Holtzman et al., Fortin et al., and MacWilliam et al. all concluded that patients with a poor preoperative status had a poorer outcome compared to the patients with a better preoperative status.,, The indications for surgical intervention in progressive hip disease can be vague and subjective when it comes to determining patient pain and disability. These results indicate that it would be in the best of the patient to consider early surgery as their long-term worsening condition may lead to lower mental status which in turn could lead to poorer outcomes.
A number of THA studies suggested that overweight patients were more susceptible to poor outcomes after surgery.,,, Wang et al. found increased chances of perioperative complications in patients with a higher BMI. Young et al. and Katz et al. both reported better functional outcomes in patients of THR with lower BMI and better preoperative functional status. Marker et al. and Shimmin et al. found that overweight patients had higher risk of femoral neck fracture after MOMHR in their respective studies., Our study also found patients with higher BMI to have lower postoperative WOMAC PF score.
Lubekke and Wang et al. reported that presence of preoperative comorbidities and higher ASA scores were associated with lower postoperative outcome scores after primary THR., In an analysis of demographic factors affecting long term outcome in 25,990 THA, Roder found that patient comorbidities had the most profound effect on overall postoperative functional status. In our study preoperative comorbidities (P < 0.0001) were found to have the most significant effect on short-term outcomes in MOMHR patients.
The four predictive factors of short-term MOMHR arthroplasty outcomes (preoperative comorbidities, SF-12 mental status, gender, BMI) determined through multivariate regression analysis can be used to calculate an expected patient functional outcome. When compared to an independent cohort of patients, the equation exhibited a positive correlation between predicted and observed 1 year outcomes. With an average predicted score of 91.39 compared to the average observed score of 91.36, the selection bias between the small independent cohort and the rest of our study population was negligible. To confirm these results, though, a larger independent cohort study is required at a separate center. With these consistent results, our regression equation in hands of surgeons can prove to be a valuable tool in terms of patient selection and managing high patient expectations after MOMHR.
It is important to note, that our study had a few limitations. We included only 14 factors but still many factors remain to be studied. While our average predicted scores were relatively consistent with average actual outcome scores, not all individual scores matched. If more factors are included, other significant factors can be found which could help further hone the equation. Another limitation was the size and demographic of our patient population. Currently, only 154 patients in our center's registry had comprehensive data. A larger cohort of patients would help in developing a stronger regression equation to predict outcomes. Further study is required using a larger cohort of patients, with a different demographic profile, and at different institutions to validate the equation further.
| Conclusion|| |
Our study helps confirm many of the significant predictive factors while filtering out some confounding factors reported in earlier literature. Many predictive factors of both short- and long-term outcomes for THA are the same as those predicting short-term outcome of MOMHR. Further long-term studies for MOMHR can help us predict the significant factors affecting the long-term outcome.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
dela Rosa MA, Silva M, Heisel C, Reich M, Schmalzried TP. Range of motion after total hip resurfacing. Orthopedics 2007;30:352-7.
Kishida Y, Sugano N, Nishii T, Miki H, Yamaguchi K, Yoshikawa H. Preservation of the bone mineral density of the femur after surface replacement of the hip. J Bone Joint Surg Br 2004;86:185-9.
Fortin PR, Clarke AE, Joseph L, Liang MH, Tanzer M, Ferland D, et al
. Outcomes of total hip and knee replacement: Preoperative functional status predicts outcomes at six months after surgery. Arthritis Rheum 1999;42:1722-8.
Holtzman J, Saleh K, Kane R. Gender differences in functional status and pain in a Medicare population undergoing elective total hip arthroplasty. Med Care 2002;40:461-70.
Holtzman J, Saleh K, Kane R. Effect of baseline functional status and pain on outcomes of total hip arthroplasty. J Bone Joint Surg Am 2002;84:1942-8.
Katz JN, Wright EA, Guadagnoli E, Liang MH, Karlson EW, Cleary PD. Differences between men and women undergoing major orthopedic surgery for degenerative arthritis. Arthritis Rheum 1994;37:687-94.
Lubbeke A, Katz JN, Perneger TV, Hoffmeyer P. Primary and revision hip arthroplasty: 5-year outcomes and influence of age and comorbidity. J Rheumatol 2007;34:394-400.
MacWilliam CH, Yood MU, Verner JJ, McCarthy BD, Ward RE. Patient-related risk factors that predict poor outcome after total hip replacement. Health Serv Res 1996;31:623-38.
Roder C, Parvizi J, Eggli S, Berry DJ, Muller ME, Busato A. Demographic factors affecting long-term outcome of total hip arthroplasty. Clin Orthop Relat Res 2003;417:62-73.
Young NL, Cheah D, Waddell JP, Wright JG. Patient characteristics that affect the outcome of total hip arthroplasty: A review. Can J Surg 1998;41:188-95.
Katz JN, Phillips CB, Baron JA, Fossel AH, Mahomed NN, Barrett J, et al
. Association of hospital and surgeon volume of total hip replacement with functional status and satisfaction three years following surgery. Arthritis Rheum 2003;48:560-8.
Amstutz HC, Le Duff MJ, Campbell PA, Dorey FJ. The effects of technique changes on aseptic loosening of the femoral component in hip resurfacing. Results of 600 conserve plus with a 3 to 9 year follow-up. J Arthroplasty 2007;22:481-9.
Khan M, Kuiper JH, Edwards D, Robinson E, Richardson JB. Birmingham hip arthroplasty: Five to eight years of prospective multicenter results. J Arthroplasty 2009;24:1044-50.
Marker DR, Seyler TM, Jinnah RH, Delanois RE, Ulrich SD, Mont MA. Femoral neck fractures after metal-on-metal total hip resurfacing: A prospective cohort study. J Arthroplasty 2007;22 7 Suppl 3:66-71.
McBryde CW, Theivendran K, Thomas AM, Treacy RB, Pynsent PB. The influence of head size and sex on the outcome of Birmingham hip resurfacing. J Bone Joint Surg Am 1992;1:105-12.
Shimmin AJ, Back D. Femoral neck fractures following Birmingham hip resurfacing: A national review of 50 cases. J Bone Joint Surg Br 1987;2005;463-4.
Wang W, Morrison TA, Geller JA, Yoon RS, Macaulay W. Predicting short-term outcome of primary total hip arthroplasty a prospective multivariate regression analysis of 12 independent factors. J Arthroplasty 2010;25:858-64.
Gabriel SE, Wenger DE, Ilstrup DM, Lewallen DG, Melton LJ 3rd
. Lack of evidence for gender bias in the utilization of total hip arthroplasty among Olmsted County, Minnesota residents with osteoarthritis. Arthritis Rheum 1994;37:1171-6.
[Figure 1], [Figure 2], [Figure 3], [Figure 4], [Figure 5]