Is Postoperative Function After Hip or Knee Arthroplasty Influenced by Preoperative Functional Levels?

The Journal of Arthroplasty Vol. 24 No. 7 2009 Is Postoperative Function After Hip or Knee Arthroplasty Influenced by Preoperative Functional Levels? Carlos Lavernia, MD,*yz Michele D'Apuzzo, MD,y Mark D. Rossi, PhD, PT, CSCS, and David Lee, PhDO Abstract: Our objective was to evaluate functional outcomes after surgery in a subgroup of patients presenting for hip and knee surgery who had low functional scores before surgery. One hundred twenty-seven unilateral total hip and knee arthroplasty patients were assessed preoperatively and 3 consecutive years after arthroplasty using: Western Ontario and McMaster University Osteoarthritis Index (WOMAC), Short Form 36 (SF-36), and the Quality of Well-Being index scales. Patients were placed into 2 groups based on preoperative WOMAC function scores; 51 points or more, worse functioning group, and less than 51 points, higher functioning group. Regardless of time, the worse functioning group in both procedures performed worse on the Quality of Well-Being index, SF-36 (function score), SF-36 (social score), and WOMAC total and pain scores (P.0001). The greatest change (range, 2%-638%) for all variables in both groups for both procedures occurred during the first year. Patients that had severe/extreme functional impairment had worse 3-year outcomes compared with patients getting surgery when their functional levels were better. Keywords: osteoarthritis, knee arthroplasty, functional outcomes, surgical timing, impairment. 2009 Elsevier Inc. All rights reserved. Numerous investigators have documented that individuals presenting with severe knee and hip joint osteoarthritis who undergo primary arthroplasty have decreased pain and increased function From the *Orthopaedic Institute at Mercy Hospital, Miami, Florida; yarthritis Surgery Research Foundation, Inc., Miami, Florida; zflorida International University, Miami, Florida; Department of Physical Therapy, Florida International University, Miami, Florida; and Department of Epidemiology and Public Health, University of Miami, Miami, Florida. Submitted June 20, 2007; accepted September 11, 2008. Conflict of interest statement: Financial support has been received by the corresponding author: Mercy Hospital, Miami, FL; Zimmer, Inc, Warsaw, IN; and Arthritis Surgery Research Foundation, Inc, Miami, FL. Royalties: Zimmer, Inc. Consultant: Zimmer, Inc. Reprint requests: Carlos J. Lavernia, MD, Orthopaedic Institute at Mercy, 3659 S. Miami Ave Ste 4008, Miami, FL 33133. 2009 Elsevier Inc. All rights reserved. 0883-5403/08/2407-0006$36.00/0 doi:10.1016/j.arth.2008.09.010 [1,2]. Despite these favorable surgical outcomes, some patients continue to report difficulty completing some functional tasks after surgery [3]. Variability in surgical outcomes has led some investigators to search for presurgical characteristics that are predictive of favorable functional outcomes after joint arthroplasty [4]. Moreover, are there presurgical characteristics that can influence functional outcomes? Identification of clinical and patient-reported characteristics that can help identify those individuals who will have low functional levels after surgery may help surgeons plan more appropriate interventions. The exact point in time in which a patient should have joint arthroplasty surgery is multifactorial. Severe pain decreased functional ability, x-ray findings, and failure to respond to conservative management can influence the timing of joint arthroplasty surgery. Fear of having a joint 1033

1034 The Journal of Arthroplasty Vol. 24 No. 7 October 2009 arthroplasty may cause patients with severe knee or hip joint osteoarthritis to delay their surgery, potentially leading to less favorable surgical outcomes [5]. Regardless of factor influencing outcome, there are people who have worse perceived function after arthroplasty surgery. There is limited literature examining the use of patient-oriented clinical measures to identify people who may have worse outcomes after hip or knee arthroplasty surgery [4,6-8]. Considering the expansion of joint arthroplasty surgeries over the next 20 years [9], it would seem important to identify those individuals who have poor outcomes because these people may place more demand on the health care system. Thus, our primary objective was to compare functional outcomes over a 3-year period in patients presenting for hip or knee arthroplasty between those having low functional scores before surgery with patients who presented with higher functional scores before surgery. We hypothesized that individuals who start out with poorer knee scores before surgery would end up with poorer scores years after surgery as compared with those patients who had higher functional levels before the arthroplasty. In other words, those individuals with poor functional scores before surgery would never catch up to their counterparts who had better functional levels before surgery, thus ending up with poorer scores. Materials and Methods Patient Selection Primary hip or knee arthroplasty patients with available preoperative, 1, 2, and 3 year postoperative evaluations were selected from a joint registry maintained by the senior author (N = 127) over a 5-year period. All patients gave written consent, which was approved by the institutional review board at Mercy Hospital. Inclusion criteria were those individuals having primary unilateral joint arthroplasty secondary to knee or hip joint osteoarthritis. Exclusion criteria were those individuals having another joint of the lower extremity diagnosed with osteoarthritis, bilateral musculoskeletal involvement including any fracture, dislocation, or joint arthroplasty within 1 year before surgery, individuals with terminal illness or those with any disease process that would negatively effect physical function. Patients were also excluded if there was clinically documented low back dysfunction. Mean patient age was 64.5 ± 14.21 years (±SD); 79 individuals within the sample were female. There were 62 individuals (35 women) in the total hip arthroplasty group with a mean age of 60.55 ± 16.11 years. There were 65 individuals (51 women) in the total knee arthroplasty group with a mean age of 68.99 ± 8.72 years. All subjects were evaluated ±2 weeks of the annual anniversary date. Measures The Quality of Well-Being Index. The Quality of Well-Being (QWB) index was developed by Kaplan et al [10]. This index has been widely used in assessing patients who have conditions such as cystic fibrosis [11,12], non insulin-dependent diabetes [13], and arthritis [14]. The QWB index has been validated in construct and content for the Anglo, Hispanic, and black populations [12]; in the present study, English and Spanish versions were administered as appropriate. Short Form 36. The Short Form 36 (SF-36), Short Form Health Survey, is designed to measure 3 broad quality-of-life domains, including functional status, well-being, and health perceptions. The SF-36 is composed of 8 subscales: physical function, bodily pain, mental health, social function, role limitation due to physical function, role limitation due to emotional problems, and vitality as well as general health perceptions. All the domains are scored separately on a 0- to 100-point scale, with higher numbers representing the better health status. These 8 domains also may be combined in a linear fashion to create a physical and mental summary score. Construct and content validity of the SF-36 has been thoroughly evaluated in the Anglo, Hispanic, and black populations [15]. A Spanish version has been validated in Cuban-American patients [16]. For this study, we used the SF-36 physical function and social function scores as outcome measures. Western Ontario and McMaster University Osteoarthritis Index. The Western Ontario and McMaster University Osteoarthritis Index (WOMAC) has become the current standard for evaluating outcomes after total hip and knee arthroplasty surgery. The WOMAC has been shown to be reliable and has been validated in both content and construct [17,18]. It is designed to provide information on the patient's perception of pain, stiffness, and physical function. The questionnaire consists of 24 questions (5 assessing pain, 2 assessing stiffness, and 17 assessing physical function), with each item scored as follows: 0 (none), 1 (mild), 2 (moderate), 3 (severe), and 4 (extreme).

Function Following Arthroplasty Lavernia et al 1035 Definition Of Worse Functional Levels At Presentation We operationalized our definition of worse functional levels at presentation using the 17-item physical function dimension of the WOMAC. Measured on a Likert scale, each item of the WOMAC is scored as follows: 0 = no difficulty, 1 = mild difficulty, 2 = moderate difficulty, 3 = severe difficulty, and 4 = extreme difficulty. Individuals with an item response of 3 (severe difficulty) or higher was used to define our group reporting worse physical function. Therefore, individuals with a WOMAC physical function subscale of 51 (17 items 3 = 51) or higher were placed in the subgroup with worse functional (WF) levels. Subjects with physical function subscale scores less than 51 made up the group reporting moderate or less difficulty performing functional tasks or reporting higher functional (HF) levels at presentation. We classified these groups for both the hip and knee cohorts. We also collected follow-up data on passive hip range of motion (flexion, abduction, adduction, internal, and external rotation) and passive knee range of motion (flexion and extension). Assistive device usage was also collected from both cohorts. Surgical Protocol All surgeries were performed by the senior author, and the same conservative protocol before the surgical intervention was followed. Conservative care included the appropriate use of 2 nonsteroidal (over 1 month each) medications, weight loss, offering the use of a cane, and patients with knee arthritis receiving at least 2 intra-articular injections. Patients not responding to conservative medical treatment were scheduled for surgery within a 2-month period. Statistical Analysis SPSS (Chicago, IL) software was used for the statistical analyses. The 5 dependent measures were the QWB, SF-36 physical function score, SF-36 social function score, WOMAC total score, and WOMAC pain dimension score. We completed two (hip and knee) 2 4 repeated measures analyses of covariance comparing our 2 groups (WF and HF) assessed at 4 periods (Pre, 1 year, 2 years, and 3 years). To control for age and sex, we used those variables as covariates in each analysis. A P b.05 was considered statistically significant. If violations of the assumption of sphericity were found, withinsubject effects were tested using a Greenhouse- Geisser adjustment. As appropriate, significant main effects were further reduced by pairwise comparisons using a least significant difference adjustment. Results Hip Twenty-four individuals had preoperative WOMAC physical function dimension scores of 51 or greater and made up the WF group (n = 24). The average age of this group was 59.67 ± 16.22 years; there were 12 women and 12 men. The average age of the HF group (n = 38) was 61.11 ± 16.23 years; 16 were women and 22 were men. Diagnoses for hip arthroplasty by group is presented in Table 1. Range of motion and frequency of assistive device usage before and after surgery, for both WF and HF groups, are presented as Tables 2 and 3, respectively. Analysis of variance (ANOVA) summary tables for each dependent variable are presented in Table 4 for both groups within total hip cohort. The Quality of Well-Being Index. There was no group-versus-time interaction and no main effect of time, regardless of group, the QWB index did not significantly change from pre through all annual follow-ups (Table 4). However, there was a main effect for group (P.0001), regardless of time, the WF subgroup scored poorly compared to the remaining cohort (HF) (Fig. 1). SF-36 Physical Function Score. There was no group-versus-time interaction, yet there were main effects for time and group (P.0001) (Table 4). Follow-up testing on the main effect of time indicated that the presurgical evaluation point was significantly different from all follow-up data points (P b.0001). Regardless of time, the WF subgroup scored poorly compared to the remaining cohort (HF) (Fig. 2). SF-36 Social Function Score. There was no group-versus-time interaction and no main effect of time, regardless of group, the SF-36 social function score did not significantly change from pre through Table 1. Primary Diagnosis for Total Hip Arthroplasty Cohort by Group Diagnosis WF (n = 24) HF (n = 38) Arthralgia hip pelvic region thigh 3 6 Arthritis ankylosing spondylitis 1 1 Avascular necrosis 4 2 Femoral neck fracture 1 0 Fracture acute femoral neck 1 0 Hip dislocation low 0 1 OA primary hip 13 25 RA 1 3 OA indicates osteoarthritis; RA, rheumatoid arthritis.

1036 The Journal of Arthroplasty Vol. 24 No. 7 October 2009 Table 2. Hip Passive Range of Motion (in Degrees, ±SEM) for Both Groups Over Time With Sample Sizes WF Pre (n = 22) 1 y (n = 19) 2 y (n = 12) 3 y (n = 8) Motion Flexion 75 ± 5.75 100.47 ± 4.62 103.75 ± 4.44 106.88 ± 5.42 Abduction 24.32 ± 3.09 39.47 ± 2.81 37.92 ± 2.49 38.75 ± 2.95 Adduction 17.5 ± 1.99 24.74 ± 2.77 33.33 ± 2.33 34.78 ± 3.19 Internal Rot. 7.27 ± 2.60 24.74 ± 4.21 28.33 ± 5.12 28.75 ± 4.60 External Rot. 22.5 ± 3.95 40.79 ± 2.17 38.33 ± 5.62 39.38 ± 4.37 HF Pre (n = 38) 1 y (n = 23) 2 y (n = 22) 3 y (n = 17) Motion Flexion 83.56 ± 3.22 111.03 ± 2.55 110.91 ± 3.01 112.06 ± 3.21 Abduction 27.76 ± 2.14 39.13 ± 1.43 39.77 ± 2.45 40.59 ± 2.22 Adduction 16.71 ± 1.57 35 ± 2.20 33.86 ± 2.27 31.77 ± 3.01 Internal rot. 11.18 ± 1.71 30 ± 2.16 26.82 ± 2.29 25.88 ± 3.95 External rot. 24.08 ± 2.73 35.87 ± 2.53 42.95 ± 2.75 43.24 ± 2.26 Pre indicates before surgery; 1 y, 1 year after surgery; 2 y, 2 y after surgery; 3 y, 3 y after surgery; rot., rotation. all annual follow-ups (Table 4). There was a main effect for group (P.0001). Compared to the remaining cohort (HF) and regardless of time, the WF subgroup scored poorly (Fig. 3). WOMAC Total Score. There was no groupversus-time interaction; however, there was a main effect for time and group (Table 4; P.0001). Thus, regardless of group, there was a difference in the WOMAC total scores between time intervals and regardless of time, the WF subgroup scored poorly compared to the HF group (Fig. 4). Follow-up testing indicated that the presurgical evaluation point was significantly different from all follow-up data points (P b.0001). WOMAC Pain Score. No group-versus-time interaction was reported, but there was a main effect of group and time. Regardless of time, the WF subgroup scored poorly compared to the HF individuals (P.0001; Table 4). The pattern of recovery was similar to the other variables. Followup testing on the main effect of time indicated that the presurgical evaluation point was significantly different from all follow-up data points (P b.005). Knee Thirteen individuals had preoperative WOMAC physical function dimension scores of 51 or greater and made up the WF group. The average age of this group was 69.23 ± 8.32 years; there were 11 women and 2 men. The average age of the HF group (n = 52) was 68.81 ± 8.9 years; 40 were women and 12 were men. Diagnoses for knee arthroplasty by group is presented in Table 5. Range of motion and frequency of assistive device usage before and after surgery, for both WF and HF groups, are presented Table 3. Assistive Device Usage for Both Groups Over Time Within THA Cohort With Sample Sizes WF Pre (n = 22) 1 y (n = 19) 2 y (n = 12) 3 y (n = 8) Assistive device 2 canes or 2 crutches 4 0 0 0 Cane full time 10 0 0 0 Cane occasionally 0 2 0 0 Cane most of the time 0 0 1 0 Walker 4 0 0 0 None 2 17 11 8 Not able to walk 2 0 0 0 HF Pre (n = 38) 1 y (n = 23) 2 y (n = 22) 3 y (n = 17) Assistive device 2 crutches or canes 2 0 0 0 Cane full time 15 0 0 1 Cane occasionally 1 0 1 0 Crutch fulltime 1 0 0 0 Walker 4 0 0 0 None 13 23 21 16 Pre indicates before surgery; 1 y, 1 year after surgery; 2 y, 2 y after surgery; 3 y, 3 y after surgery. THA indicates total hip arthroplasty.

Function Following Arthroplasty Lavernia et al 1037 Table 4. The ANOVA Summary Tables for Dependent Measures Within THA Group Source Df SS MS F p P QWB Group 1 0.20 0.20 18.36.0001* 0.98 Error 58 0.63 0.01 Time GG 2.53 0.01 0.007 2.29.097 0.52 Group Time GG 2.53 0.02 0.008 2.90.09 0.63 Error GG 146.79 0.42 0.002 SF-36 PF Group 1 26 090.30 26 090.30 15.94.0001* 0.97 Error 58 94 911.02 1636.39 Time 3 9581.32 3193.77 8.71.0001* 0.99 Group Time 3 1946.48 648.83 1.77 0.16 0.46 Error 174 63 841.50 366.91 SF-36 SF Group 1 26 574.76 26 574.76 20.73.0001* 0.99 Error 58 94 911.02 1636.39 Time 3 2674.26 891.42 1.34.26 0.35 Group Time 3 159.17 53.06.08.97 0.06 Error 174 115 797.96 665.51 WOMAC TOT Group 1 15 904.94 15 904.94 23.08.0001* 0.99 Error 58 39 977.87 689.27 Time 3 6076.89 2025.63 11.12.0001* 0.99 Group Time 3 781.45 260.48 1.43.24 0.38 Error 174 31 704.19 182.21 WOMAC P Group 1 616.90 616.90 20.09 b.0001* 0.99 Error 58 1780.91 30.71 Time 3 143.89 47.96 4.46.005* 0.87 Group Time 3 13.84 4.61.73.73 0.14 Error 174 1872.74 10.76 Df indicates degrees of freedom; SS, sum of squares; MS, mean square; F, F ratio, p, probability; P, power; THA, total hip arthroplasty; GG, Greenhouse-Geisser adjustment; SF-36 PF, SF-36 physical function score; SF-36 SF, SF-36 social function score; WOMAC TOT, WOMAC total score; WOMAC P, WOMAC pain dimension score. * P b.05. in Tables 6 and 7, respectively. The ANOVA summary tables for each dependent measure are presented as Table 8 for both groups in total knee arthroplasty cohort. The Quality of Well-Being Index. There was no group-versus-time interaction and no main effect of time. However, there was a main effect for group (P.0001; Table 8). Regardless of time, the WF Fig. 1. The QWB index for both groups within total hip arthroplasty group over time. RC indicates remaining cohort; DPG, delayed-presentation group. Fig. 2. The SF-36 physical function score for both groups within total hip arthroplasty group over time. RC indicates

1038 The Journal of Arthroplasty Vol. 24 No. 7 October 2009 Table 5. Diagnosis for Total Knee Arthroplasty Cohort by Group Diagnosis WF (n = 13) HF (n = 52) Arthralgia knee 0 2 Avascular necrosis 0 2 OA knee 9 32 OA traumatic 0 1 OA primary knee patella 4 14 RA 0 1 OA indicates osteoarthritis; RA, rheumatoid arthritis. Fig. 3. The SF-36 social function score for both groups within total hip arthroplasty group over time. RC indicates subgroup scored poorly compared to the remaining cohort (Fig. 5). SF-36 Physical Function Score. There was no group-versus-time interaction and no main effect for time. There was a main effect of group (P =.001; Table 8). Regardless of time, the WF subgroup scored poorly as compared with those individuals who perceived there function as higher (Fig. 6). SF-36 Social Function Score. No group-versustime interaction was reported; however, there was a main effect for time (P =.037; Table 8). Follow-up testing indicated that the presurgical evaluation point was significantly different from all follow-up data points (P.0001). There was also a main effect Fig. 4. The WOMAC total score for both groups within total hip arthroplasty group over time. RC indicates for group (P =.015; Table 8). Compared to the remaining cohort (HF) and regardless of time, the WF subgroup scored poorly (Fig. 7). WOMAC Total Score. There was no groupversus-time interaction and no main effect for time, yet, there was a main effect of group (P.0001; Table 8). Thus, regardless of time, the WF subgroup scored poorly compared to the reporting HF levels (Fig. 8). WOMAC Pain Score. No group-versus-time interaction was reported, and there was no main effect of time (Table 8). There was a main effect for group (P =.006). Regardless of time, the WF subgroup scored poorly compared to the HF group with a similar graphical representation as the other variables. Discussion Optimal timing of arthroplasty surgery will have an impact on costs as well as on outcomes. For example, poor preoperative functional status in patients with rheumatoid arthritis who are undergoing total hip or knee arthroplasty is predictive of longer hospital length of stay [6]. Escalante and Beardmore [6] also demonstrated that surgery performed before failure of conservative treatments may decrease length of hospital stay and minimize postsurgical loss of function. More recently, Fortin et al [7] noted that patients presenting for knee and hip arthroplasty with low preoperative function and pain scores did not improve 6 months postoperatively to the same magnitude as their counterparts with less pain and a HF capacity. These articles demonstrated that individuals who scored poorly before surgery never caught up to the individuals who had higher preoperative scores. Furthermore, these differences persisted 1 year after surgery [8]. The American Academy of Orthopedic Surgeons has projected that 3.48 million primary total knee arthroplasties and more than 500,000 hip arthroplasties will be performed by the year 2030 [9]. With

Function Following Arthroplasty Lavernia et al 1039 Table 6. Knee Passive Range of Motion (in Degrees, ±SEM) for Both Groups Over Time With Sample Sizes WF Pre (n = 11) 1 y (n = 7) 2 y (n = 5) 3 y (n = 5) Motion Passive ext. 1.64 ±.97 0.71 ±.71 0 ± 0 0 ± 0 Passive flex. 110.36 ± 3.48 107.14 ± 5.21 101 ± 7.31 106 ± 5.34 HF Pre (n = 44) 1 y (n = 27) 2 y (n = 22) 3 y (n = 20) Motion Passive ext. 3.41 ± 1.05 0.63 ±.41 0.68 ±.34 0.70 ±.46 Passive flex. 104.55 ± 2.04 107.93 ± 2.12 110.14 ± 2.56 107.65 ± 2.72 Pre indicates before surgery; 1 y, 1 y after surgery; 2 y, 2 y after surgery; 3 y, 3 y after surgery; ext., extension; flex., flexion. the continued increase in the number of arthroplasty surgeries, it is essential to determine the most cost-effective strategies for delivering these types of surgeries without diminishing quality patient care. Optimizing the timing of joint arthroplasty will improve both the outcomes and costs. The most important finding in our study was that hip and knee arthroplasty patients who presented for surgery with at least severe physical function disability as measured by the WOMAC physical function score had poorer outcomes on a wide range of clinical and well-being indicators relative to patients who had more favorable presurgical physical function scores. More importantly these differences persisted 1, 2, and 3 years after joint arthroplasty surgery. We found no evidence that our WF subgroup would catch up to those who presented for surgery with better preoperative functioning. A multiplicity of factors may explain the worse outcomes that persisted throughout time in our study. Permanent atrophy of the muscles, tendons, and perhaps ligaments around these joints may predispose these patients to WF outcomes after an Table 7. Assistive Device Usage for Both Groups Over Time With Sample Sizes WF Pre (n = 11) 1 y (n = 7) 2 y (n = 5) 3 y (n = 5) Assistive device None 3 3 3 4 1 cane 7 4 2 1 Walker 1 0 0 0 HF Pre (n = 44) 1 y (n = 27) 2 y (n = 22) 3 y (n = 20) Assistive device None 12 20 17 14 1 cane 28 5 4 5 1 crutch 1 0 0 0 2 canes 1 0 0 0 Walker 2 2 1 1 Pre indicates before surgery; 1 y, 1 y after surgery; 2 y, 2 y after surgery; 3 y, 3 y after surgery. arthroplasty. In addition to that is the permanent establishment of chronic pain for prolonged periods making pain relief more difficult to obtain. Another important finding in our study was that the period between the presurgical visit and the 1 year follow-up was the interval with the greatest improvement of perceived function and pain. The SF-36 physical and social function domains showed the greatest improvement by the subgroup from presurgery to the 1 year follow-up. Even with large improvements in self-reported functional measures, patients in the subgroup (patients who had low scores) were not able to catch up to the remaining cohort. In our study, patients that had higher preoperative scores had larger improvements both in pain (WOMAC pain dimension) and function (WOMAC physical function dimension) and overall quality of life (QWB index and total WOMAC score) when compared to those who started with lower scores. Although we would not expect all patients to have the same outcomes, our study does show that individuals who start out at lower functioning levels do improve yet do not reach the levels of their counterparts who started off with better perceived function. However, we are alerted to the fact that change over time in self-reported function and quality of life using scoring systems such as the WOMAC and QWB may represent age-related changes or the effects of some other condition over and above knee joint pathology. Severe joint pain and functional limitations are characteristic signs of end-stage arthritis. When conservative treatments have failed, joint arthroplasty has been a successful intervention. The intervention clearly decreases pain and improves physical function most patients. The decision of when to perform the surgical intervention in the course of conservative treatments is usually left to the patient. We feel that physicians should constantly monitor perceived outcomes during the course of conservative treatment, and waiting for

1040 The Journal of Arthroplasty Vol. 24 No. 7 October 2009 Table 8. The ANOVA Summary Tables for Dependent Measures Within TKA Group Source Df SS MS F P P QWB Group 1 0.13 0.13 13.91.0001* 0.96 Error 61 0.59 0.01 Time GG 2.84 0.002 0.001 0.23.87 0.09 Group Time GG 2.84 0.003 0.001 0.35.79 0.12 Error GG 183 0.51 0.003 SF-36 PF Group 1 14 010.66 14 010.66 11.06.001* 0.91 Error 61 77 282.29 1266.92 Time GG 3 1247.75 415.92 1.26.29 0.33 Group Time GG 3 833.49 277.83 0.84.47 0.23 Error GG 183 60 217.91 329.06 SF-36 SF Group 1 9684.30 9684.30 6.27.02 0.69 Error 61 94 158.74 1543.59 Time GG 3 5579.57 1859.86 2.89.04 0.68 Group Time GG 3 3727.25 1242.42 1.93.13 0.49 Error GG 183 117 906.70 644.30 WOMAC TOT Group 1 12 539.80 12 539.80 16.97.0001* 0.98 Error 61 45 088.66 739.16 Time GG 3 1088.72 362.91 1.74.16 0.45 Group Time GG 3 111.67 37.22 0.18.91.08 Error GG 183 38 267.95 209.11 WOMAC P Group 1 311.94 311.94 8.03.006* 0.80 Error 61 2370.68 38.86 Time GG 2.72 38.76 14.23 1.01.39 0.26 Group Time GG 3 57.42 19.14 1.49.22 0.39 Error GG 183 2352.84 12.86 Df indicates degrees of freedom; SS, sum of squares; MS, mean square; F, F ratio; p, probability; P, power; TKA, total knee arthroplasty; GG, Greenhouse-Geisser adjustment; SF-36 PF, SF-36 physical function score; SF-36 SF, SF-36 social function score; WOMAC TOT, WOMAC total score; WOMAC P, WOMAC pain dimension score. * P b.05. the patient to decide may have a downside especially when functional deficits are getting worse. More than 15 billion dollars are spent yearly in arthroplasty surgery. Furthermore, joint arthroplasty is one of the most cost-effective elective interventions in all of medicine. An easy strategy for Health Maintenance Organizations to keep their cost down is to keep the patients within the primary care setting and not allow them to see the specialist. The Fig. 5. The QWB index for both groups within total knee arthroplasty group over time. RC indicates remaining cohort; DPG, delayed-presentation group. Fig. 6. The SF-36 physical function score for both groups within total knee arthroplasty group over time. RC indicates

Function Following Arthroplasty Lavernia et al 1041 Fig. 7. The SF-36 social function score for both groups within total knee arthroplasty group over time. RC indicates senior author has seen in his environment multiple primary care doctors that have financial incentives to keep patients from arthroplasty surgery. This behavior, although unusual, occurs with some compensation schemes. The results of our study mirror closely the results reported by Fortin et al in patients undergoing total hip and knee arthroplasty in 2 major hospitals in Montreal and in Boston [7]. In their study, the physical function score of the SF-36 and pain and physical dimensions of the WOMAC were used to document perceived abilities in low- and highfunctioning groups. Although both groups reported improvement in functional abilities 6 months after surgery, the lower-functioning group continued to have worse outcomes [8]. Similar to our findings, the lower-functioning group was not able to catch up with those who presented for surgery with better functional status 2 years after surgery. Unlike the study by Fortin et al, we assessed our subjects out to 3 years where the pattern for both HF and WF groups persisted. Moreover, we did gather data on specific assistive devices usage and range of motion. Results from our study indicated that the difference in quality of life between groups for both procedures actually increased 1 year after surgery, and this difference did not change by 3 years after surgery. The greatest rate of change occurred from preevaluation to 1 year after surgery for both procedures and both groups. It appears that the first year after joint arthroplasty surgery is a critical period in optimizing functional performance, minimizing pain, and improving quality of life because positive changes are occurring at a fast rate. Although the number of respondents decreased in follow-up, it is clear that for both procedures and groups, individual usage for an assistive device decreased over time. In other words, as time increased, individuals came off their assistive device. Results of our study also showed that the greatest improvement in hip passive range of motion for tested ranges was from before surgery to 1 year after surgery with a plateauing over subsequent years. Knee joint flexion did not show this pattern of change over time probably due to the lack of follow-up. Multiple variables, which may influence the results of the surgical intervention, must be included in the analyses as independent predictors of surgical outcome. Our study used preoperative function as a measure of overall functional abilities before total hip or knee arthroplasty. However, other patientrelated factors are equally important. Current knee rating scores lack the inclusion of patient-oriented factors. When the total joint arthroplasty patient is assessed, overall quality of life, perceived function, and pain are poorer in those patients who have surgery when their functional status is indicative of greater disability. The optimal point for surgical intervention in arthritis is an important question. Intervening too early is often difficult because patients afraid of the surgical procedure will delay meeting with a surgeon as long as possible. Another factor that could delay the intervention is the fact that Health maintenance organizations may keep the patient away from the specialist as long as possible. Fig. 8. The WOMAC total score for both groups within total knee arthroplasty group over time. RC indicates

1042 The Journal of Arthroplasty Vol. 24 No. 7 October 2009 Future larger studies should evaluate the responsiveness of our clinical tools. Moreover, the financial cost of waiting too long (until outcome measures indicate substantial decline in function) needs to be addressed because allowing patients with end-stage arthritis to continue to lose function may ultimately be less beneficial than if surgical intervention is done sooner. Cutoff scores of our clinical outcome instruments needs to be developed that will aid the surgeon in determining the critical time point of surgical intervention without compromising cost yet promoting future self-abilities. For example, a prospective study that stratifies groups based on each of the 5 scoring levels of the WOMAC could aid clinicians in determining which individuals may require longer postoperative care. Moreover, preoperative education and primary care referral for orthopedic consultation may need to be addressed sooner and with greater emphasis. Finally, it should be noted that the optimal timing of joint arthroplasty surgery must ultimately be assessed in the context of a randomized clinical trial in which patients with varying levels of functional disability are randomly assigned to early versus date surgery. Sole reliance on findings from observational studies such as our own to determine optimal timing for joint arthroplasty surgery is not appropriate because we are unable to control for a variety of important potential confounders such as patient self-selection, patient questionnaire response tendencies, and date of first diagnosis of disease (onset). Furthermore, our sample sizes were small, and generalizations to the overall population can be tenuous; thus, a much larger prospective study would have enhanced the strength of our study. We also note that this study does not provide information regarding the outcome of early surgical intervention in the treatment of early arthritis. We also did not control for comorbidities that can influence function especially over the long term. Our follow-up for range of motion was less than expected; thus, our results for range of motion should not be considered conclusive but preliminary. Other limitations of our study were that we did not identify patient expectations or the length of time having disability before having surgery. Furthermore, we did not document the time of conservative care for each patient. Despite these limitations, our data raise the intriguing possibility that performing surgery sooner rather than later may result in better surgical outcomes, although we cannot confirm this from our study. We recommend much larger prospective studies exploring the relationship of diagnosis of disease (onset) to surgical day with outcome measures such as the WOMAC and QWB. We also recommend performance tests be used to gather objective data on physical abilities. Presently most joint arthroplasty surgeons recommend arthroplasty as last resort with their conservative management philosophy being the most important reason why patients delay surgery. Randomized clinical trials are therefore urgently needed to identify optimal surgical timing for patients with end-stage arthritis. References 1. Marx RG, Jones EC, Atwan NC, et al. Measuring improvement following total hip and knee arthroplasty using patient-based measures of outcome. J Bone Joint Surg Am 2005;87:1999. 2. Kane RL, Saleh KJ, Wilt TJ, et al. The functional outcomes of total knee arthroplasty. J Bone Joint Surg Am 2005;87:1719. 3. Weiss JM, Noble PC, Conditt MA, et al. What functional activities are important to patients with knee replacements? Clin Orthop Relat Res 2002;404:172. 4. Lingard EA, Katz JN, Wright EA, et al. Predicting the outcome of total knee arthroplasty. J Bone Joint Surg Am 2004;86-A:2179. 5. Clark JP, Hudak PL, Hawker GA, et al. The moving target: a qualitative study of elderly patients' decisionmaking regarding total joint replacement surgery. J Bone Joint Surg Am 2004;86-A:1366. 6. Escalante A, Beardmore TD. Predicting length of stay after hip or knee replacement for rheumatoid arthritis. J Rheumatol 1997;24:146. 7. Fortin PR, Clarke AE, Joseph L, 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. Fortin PR, Penrod JR, Clarke AE, et al. Timing of total joint replacement affects clinical outcomes among patients with osteoarthritis of the hip or knee. Arthritis Rheum 2002;46:3327. 9. Frankowski JJ, Watkins-Castillo S. Primary total knee and hip arthroplasty projections for the U.S. population to the year 2030. http://www.aaos.org/ wordhtml/research/stats/tjr_projections.pdf [Last accessed Sep 15, 2005]. 10. Kaplan RM, Bush JW, Berry CC. Health status: types of validity and the index of well-being. Health Serv Res 1976;11:478. 11. Orenstein DM, Nixon PA, Ross EA, et al. The quality of well-being in cystic fibrosis. Chest 1989;95:344. 12. Kaplan RM, Anderson JP, Wu AW, et al. The Quality of Well-Being scale. Applications in AIDS, cystic fibrosis, and arthritis. Med Care 1989;27(3 Suppl): S27. 13. Coffey JT, Brandle M, Zhou H, et al. Valuing healthrelated quality of life in diabetes. Diabetes Care 2002;25:2238.

Function Following Arthroplasty Lavernia et al 1043 14. Frosch DL, Kaplan RM, Ganiats TG, et al. Validity of self-administered Quality of Well-Being scale in musculoskeletal disease. Arthritis Rheum 2004;51: 28. 15. Lieberman JR, Dorey F, Shekelle P, et al. Outcome after total hip arthroplasty. Comparison of a traditional disease-specific and a quality-oflife measurement of outcome. J Arthroplasty 1997;12:639. 16. Arocho R, McMillan CA, Sutton-Wallace P. Construct validation of the USA-Spanish version of the SF-36 health survey in a Cuban-American population with benign prostatic hyperplasia. Qual Life Res 1998;7:121. 17. Bellamy N, Buchanan WW, Goldsmith CH, et al. Validation study of WOMAC: a health status instrument for measuring clinically important patient relevant outcomes to antirheumatic drug therapy in patients with osteoarthritis of the hip or knee. J Rheumatol 1988;15:1833. 18. Bombardier C, Melfi CA, Paul J, et al. Comparison of a generic and a disease-specific measure of pain and physical function after knee replacement surgery. Med Care 1995;33:AS131.