Secondary Logo

Journal Logo

SECTION I: SYMPOSIUM: Papers Presented at the 2005 Meeting of the Knee Society

Total Knee Replacement Outcome and Coexisting Physical and Emotional Illness

Ayers, David C MD*; Franklin, Patricia D MD, MBA, MPH*; Ploutz-Snyder, Robert PHD; Boisvert, Catherine B MD*

Editor(s): Laskin, Richard S Guest Editor

Author Information
Clinical Orthopaedics and Related Research: November 2005 - Volume 440 - Issue - p 157-161
doi: 10.1097/01.blo.0000185447.43622.93
  • Free


Total knee replacement (TKR) is a high-volume, effective procedure predicted to increase in use, cost, and importance throughout the next decade. In 2002, more than 350,000 primary TKRs were done in the United States.2 Medicare funded 60% of the procedures with aggregate charges of more than $9.8 billion.2 Age and obesity are directly related to an increased risk of knee osteoarthritis, so the demand for TKR is likely to grow as the United States population ages, obesity rates increase, and treatment options become available.3

Expanding use of TKR is a result of the overwhelming evidence that the procedure improves quality of life for patients with knee arthritis.11 Surgical techniques and prostheses have improved in the last 20 years with uniformly good results.10 The 2003 National Institutes of Health consensus panel reported that TKR effectively eliminates pain from degenerative joint disease, and improves physical function for patients with advanced osteoarthritis, rheumatoid arthritis, and other joint pathologies.11 Despite this surgical success, variation persists in functional outcome after TKR.

A subset of patients has suboptimal postoperative improvement in physical function after TKR.2,9 The TKR Patient Outcomes Research Team reported that 30% to 33% of patients with poorer preoperative emotional health [Short Form-36 Mental Component Score (SF-36 MCS) < 50] have poorer 6-month and 12-month postoperative physical function.9 A preoperative MCS < 50 is associated with a mixed psychological picture that includes clinically defined trait anxiety, mild depression, poor social supports, and limited coping skills.2

In addition to emotional health conditions, coexisting physical illness is prevalent among older adults, including patients who have TKR.3 Among patients with osteoarthritis, the presence of comorbid physical conditions has been associated with variation in functional status.4,5 In addition, poorer baseline physical function has been associated with less functional improvement after TKR.6 Therefore, it is important to determine to what degree physical coexisting diagnoses may be associated with variations in functional return after TKR surgery.

Our aim is to quantify the relative contributions of emotional and physical coexisting conditions to the variation in improvement in 12-month postoperative TKR physical function. We hypothesize that the presence of an increasing number of coexisting physical diagnoses will be associated with smaller pre-to-post TKR functional improvement as measured by the SF-36 and Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC). We hypothesize that the addition of emotional health diagnoses will further reduce the degree of postoperative functional improvement. We will offer a better understanding of the roles that emotional and physical health play in functional recovery after TKR in order to guide optimal patient selection and preparation for surgery.


We conducted a secondary analysis of an existing registry that includes demographic, clinical, and patient-reported functional status measures collected prospectively for a consecutive cohort of patients having primary TKR at one academic total joint replacement center. All patients in the TKR registry were included in the analysis with the exception of those who were scheduled emergently for surgery; those who needed revision of previous total joint prostheses; or those who were having surgery because of malignancy, infection, or complications. In developing the registry, all patients having TKR were enrolled consecutively with the single exception of those patients who were unable to answer functional status questions in English (ie, patients who did not speak English or those patients with cognitive impairment.) The 165 study patients identified were representative of the center’s patients having primary TKR [62% women, mean age of 68 years; standard deviation (SD) = 9.8].

The registry was created to integrate demographic and clinical variables with patient-reported outcomes. All patients scheduled for TKR surgery at our center routinely complete standardized general health status (SF-36) and knee-specific assessments (WOMAC) at a preoperative office visit.8,13 The preoperative SF-36 subscores [MCS and Physical Component Score (PCS)] and WOMAC physical function (WOMAC-PF), pain, and stiffness scores are calculated. For this analysis, health status measures and demographic variables (age, gender, and marital status) were extracted from the registry. Measurement of 12-month postoperative function (SF-36, WOMAC) was completed at an office visit or by telephone by the research assistant, and then entered into the study database.

The International Classification of Diseases 9 (ICD-9) diagnosis codes for active medical conditions were obtained from the hospital discharge database for all 165 study patients. Diagnoses were aggregated by organ system (ie, cardiovascular, pulmonary, endocrine, rheumatic, lower extremity-related, and mental health) (Table 1). The ICD-9 codes and organ system categories were added to the registry data.

Table 1
Table 1:
Categorization of Hospital Diagnoses in Clinical Condition Categories

We developed a multivariate prediction model to evaluate the contribution of preoperative demographic variables and physical and emotional health conditions to variation in 12-month physical function after TKR. Blocked multiple regression analyses were run twice; once using a general measure of physical function (SF-36 PCS), and once using a knee specific measure of functional outcome (WOMAC-PF). In the first regression analysis, the initial block was adjusted for age, gender, and preoperative TKR physical function (SF-36 PCS). This was followed by a block including the presence or absence of each coexisting medical diagnosis. Then the pre-TKR MCS was added. The second regression analysis replaced the preoperative and 12-month PCS with the preoperative and 12-month WOMAC-PF scores. The second model allowed us to verify the contribution of physical and emotional conditions (MCS) when predicting an independent measure of joint-specific physical function (WOMAC-PF).

To estimate the effect of the total burden of coexisting illness on physical function, a variable was created that summed the number of organ systems with coexisting diagnoses for each patient. The blocked regression analysis was repeated using this sum of coexisting conditions variable to replace the individual conditions. The initial block adjusted for age, gender, and preoperative TKR physical function (PCS, WOMAC). This was followed by a block including the sum of coexisting medical conditions and the preoperative MCS. Statistical significance was defined as p ≤ 0.05 for all tests.


Eighty four percent of patients had at least one comorbid illness. Sixty-one percent of patients had a cardiovascular disease diagnosis, 22% of patients had an endocrine problem, 18% of patients had nonarthritis lower-extremity conditions, 8% of patients had pulmonary conditions, 8% of patients had rheumatoid arthritis, and 4% of patients had a mental health diagnosis (Table 2). Women had a higher (p < 0.019) prevalence of lower-extremity-related (nonarthritis) diagnoses than men (23% and 8%, respectively). There were no other gender differences in the prevalence for diagnostic categories. There was also no difference in the rates of coexisting physical diagnoses among patients with high and low pre-TKR emotional health (MCS ≥ 50; MCS < 50, respectively). The mean SF-36 PCS improvement from pre-TKR to 12 months post-TKR was similar for patients with and without comorbid diagnoses (9.3 versus 12, respectively) (Table 3).

Table 2
Table 2:
Demographics and Clinical Summary
Table 3
Table 3:
Mean Change in Physical Function (SF-36 PCS) by Clinical Condition

Mean WOMAC-PF improvement was similar for patients without any comorbid diagnosis and for patients with one or more diagnosis (24 versus 22.1, respectively). However, for patients with and without a cardiovascular comorbidity, there was a marginal improvement (p = 0.06) in WOMAC-PF score (mean = 19.6 and 27, respectively) (Table 4).

Table 4
Table 4:
Mean Change in WOMAC Physical Function Score

Adding coexisting conditions to age, gender, and baseline physical function did not improve the model’s ability to explain variation in the 12-month physical function when measured by either SF-36 PCS or WOMAC (Table 5). However, emotional health remained a predictor of 12-month improvement in physical function after adjusting for physical coexisting illnesses. The addition of the MCS explained almost ½ of the total variation predicted in the PCS model (r2 = 0.357) and in the WOMAC-PF model (r2 = 0.273). Pulmonary diagnoses were the only conditions that approached significance in predicting (p < 0.063) 12-month WOMAC physical function in the multivariate model.

Table 5
Table 5:
Prediction of Variation in 12-month Physical Function: Regression Results

Multivariate models using the summed number of coexisting diagnoses had results similar to the analyses specifying each condition (r2 = 0.231 with WOMAC-PF; r2 = 0.331 with SF-36 PCS). The sum of coexisting medical conditions did not improve the prediction of 12-month physical function whereas the MCS improved the model.


We did not detect a relationship between coexisting physical conditions and the degree of 12-month post-TKR functional improvement. However, consistent with previous research, lower preoperative emotional health was associated with smaller improvements in physical function as measured by the SF-36 PCS and WOMAC physical function scores.

Previous research in patients who have not had TKR has shown that multiple comorbid conditions reduce functional status to a degree that is more than the additive impact of the individual conditions.5,12 For example, Fried et al7 found that arthritis and hypertension is associated with an increased risk of reduced function and self care [Odds Ratio (OR) = 7.5] more than either condition alone (arthritis OR = 2.4; hypertension OR = 0.9). Comorbidity has also been suggested to have a relationship with WOMAC and SF-36 PCS scores, and some recommend that coexisting conditions be considered when interpreting orthopaedic physical function scores.4 Patients with four or more comorbid conditions have lower 3-month SF-36 PCS scores.14

The lack of a relationship between the presence of coexisting medical diagnoses and 12-month physical function in this study is important for patients and surgeons. As expected, more than 80% of the patients who had TKR had active coexisting medical conditions. It is possible that patients with the most severe medical disease were excluded from TKR because of surgical risks. In contrast to Wasielewski et al,14 we assessed 1-year function (as opposed to 3 months) and found no relationship between medical conditions and functional outcome. This difference may suggest that short-term postoperative TKR function is affected more by coexisting conditions than function at 1 year. In addition, our multivariate model adjusted for baseline physical function, which may have eliminated the contribution of coexisting diagnoses as a predictor of postoperative TKR function. A severity adjusted assessment of coexisting illness may have discriminated among patient illness burden better than the present/absent determination possible with discharge data. However, our findings suggest that if a patient’s medical condition does not prohibit anesthesia and the TKR procedure, the medical conditions will not limit 12-month physical function.

These data support that emotional health status as measured by the MCS remains a predictor of 12-month SF-36 PCS and WOMAC PF after adjusting for comorbid physical conditions, age, gender, and preoperative physical function. Only six of 165 patients had a mental health diagnosis in the medical record whereas approximately ⅓ of patients reported an MCS < 50. It is possible that depression or anxiety was present in the hospital but was omitted from the diagnoses in error. However, it is likely that the MCS is a sensitive screen for the coexistence of an array of symptoms of anxiety, depression, and poor social support. Each of these symptoms may exist as a subclinical condition, but may be important enough to limit optimal functional return.

Our analysis reaffirms that patients with lower pre-TKR emotional health are at risk for suboptimal functioning at 1 year postoperatively. This finding persists despite a skilled operative procedure using prostheses of proven design without complication. The variation in functional outcome was not explained by the coexisting medical conditions listed in the hospital discharge database. The association between preoperative TKR emotional status (SF-36 MCS) and physical function at 12 months postoperatively persisted for general (SF-36 PCS) and knee-specific (WOMAC) physical function. Future research may include clinically refined assessments of physical and emotional coexisting conditions. Our results emphasize the need to identify effective perioperative clinical interventions to support emotional health needs to enhance postoperative physical function.


1. Agency for Healthcare Research and Quality: HCUPnet, Healthcare Cost and Utilization Project. Rockville, MD. Accessed on June 12, 2005.
    2. Ayers DC, Franklin PD, Trief PM, et al: Psychological attributes of preoperative total joint replacement patients: Implications for optimal physical outcome. J Arthroplasty 19:125-130, 2004.
    3. Center for Disease Control and Prevention: Prevalence of Disabilities and Associated Health Conditions Among Adults-United States, 1999. MMWR 50:120-125, 2001.
    4. Dunbar MJ, Robertsson O, Ryd L: What’s all that noise? The effect of co-morbidity on health outcome questionnaire results after knee arthroplasty. Acta Orthop Scand 75:119-126, 2004.
    5. Ettinger WH, Davis MA, Neuhaus JM, et al: Long-term physical function in persons with knee osteoarthritis from NHANES-1: Effects of co-morbid medical conditions. J Clin Epidemiol 47:809-815, 1994.
    6. 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 42:1722-1728, 1999.
    7. Fried LP, Bandeen-Roceh K, Kasper JD, et al: Association of Comorbidity with disability in older women: The Women’s Health and Aging Study. J Clin Epidemiol 52:27-37, 1999.
    8. Hawker G, Melfi C, Paul J, et al: Comparison of a generic (SF-36) and a disease specific (WOMAC) instrument in the measurement of outcomes after knee replacement surgery. J Rheumatol 22:1193-1196, 1995.
    9. Heck DA, Robinson RL, Partridge CM, et al: Patient outcomes after knee replacement. Clin Orthop Relat Res 356:1-18, 1998.
    10. Katz JN, Barrett J, Mahomed NN, et al: Association between hospital and surgeon procedure volume and the outcomes of total knee replacement. J Bone Joint Surg 86A:1909-1916, 2004.
    11. National Institutes of Health Consensus Development Panel on Total Knee Replacement: National Institutes of Health consensus statement on total knee replacement. December 8-10, 2003. Final statement. Rockville, MD. US Department of Health and Human Services Feb 17, 2004. (DHHS)
    12. Stewart AL, Greenfield S, Hays RD, et al: Functional status and well-being of patients with chronic conditions: Results from the Medical Outcomes Study. JAMA 262:907-913, 1989.
    13. Ware J, Kosinski M, Keller SD: SF-36 Physical and Mental Component Summary Scales: A User’s Manual. Boston, The Health Institute, New England Medical Center, 1994.
    14. Wasielewski RC, Weed H, Prezioso C, et al: Patient comorbidity: Relationship to Outcomes of total knee arthroplasty. Clin Orthop Relat Res 356:85-92, 1998.
    © 2005 Lippincott Williams & Wilkins, Inc.