This is the first study to report that ipsilateral lower extremity joint involvement is associated with significantly higher odds of moderate to severe index arthroplasty pain and moderate to severe functional limitation at 2 years and 5 years after index THA and TKA, both in primary and revision cases. The findings were robust and effect sizes were consistent across the type of arthroplasty (hip or knee), primary and revision arthroplasty, and the two follow-up time-points. Sensitivity analyses that adjusted for additional covariates (anxiety and depression), or restricted to the patients with osteoarthritis confirmed the findings. Several findings in this study deserve further discussion.
A key finding of our study was that concomitant ipsilateral knee/hip involvement increased the risk of moderate to severe index arthroplasty joint pain, unequivocally an undesired outcome of THA/TKA, at both 2 years and 5 years after primary THA/TKA. In the absence of any previous studies, these data add new knowledge. Potential mechanisms include referred pain from their involved ipsilateral knee/hip to the index THA/TKA, altered biomechanics and more weight bearing on the index THA/TKA, and limited ability to do adequate rehabilitation and strengthening due to concomitant ipsilateral joint involvement. A causal relationship cannot be inferred due to the assessment of ipsilateral involvement and pain/function outcomes cross-sectionally.
Several underlying conditions can lead to the involvement of the ipsilateral knee/hip, such as  osteoarthritis or other arthritis in multiple joint in patients with primary THA and TKA;  a failing primary or revision arthroplasty in the ipsilateral joint; and  diseases of periarticular structures, such as bursitis or tendinitis, that lead to articular and periarticular symptoms. Future studies should assess whether the treatment of ipsilateral joint involvement leads to improvement in outcomes related to index TKA/THA joint. Whether treatment of activity limitation related to ipsilateral knee/hip with physical therapy, surgical (that is, arthroplasty) or other modalities can improve index THA/TKA outcomes remains to be seen.
Ipsilateral knee/hip involvement had an even stronger relationship with functional limitation following primary THA/TKA than its association with moderate to severe pain. This finding is not unexpected. The higher the number of involved joints in the lower extremity, the more likely it is that a patient will have moderate to severe activity limitation, since these limitations are specific to lower extremity joint and muscle function. To the best of our knowledge this is the first study to assess the impact of ipsilateral joint involvement on TKA/THA pain and function outcomes, using robust analyses. This finding adds to the recent findings that concomitant contralateral knee pain is associated with poorer post knee replacement function, both when present preoperatively  and postoperatively . These observations indicate that it is important to pay attention to other joint involvement in patients with suboptimal outcome after THA/TKA. This study does not answer a critical question whether the presence of ipsilateral knee/hip involvement leads to the suboptimal outcome in index THA/TKA, which needs to be examined in future studies. Whether properly addressing concomitant contralateral or ipsilateral joint involvement might improve the pain and functional outcome of the operated joint remains to be seen. Thus, these findings have implications for improving care and potentially outcomes of patients undergoing THA/TKA.
The association of ipsilateral knee/hip involvement with poor pain and function outcomes noted in primary THA/TKA was also noted in patients who underwent revision THA/TKA. In particular, the strength of association after revision THA/TKA was similar to that noted in patients undergoing primary THA/TKA. In addition, we noticed little or no attenuation of the association noted at 2-year follow-up at the longer 5-year follow-up. These findings support the robustness of this association.
The study findings must be interpreted considering study strengths and limitations. Study strengths include a large cohort, prospective standardized data collection by dedicated clinical registry staff, adjustment for important covariates and confounders, and the robustness of findings across several sensitivity analyses. Our study also has several limitations, however. Non-response and referral bias may limit our ability to generalize these findings to other populations. However, patient demographics are similar to previously published studies of primary and revision THA and TKA. The response rate at 2 years is similar to the mean response rate of 60%, typical for large surveys of this size . The response rate at 5 years at 48% is low, and therefore these findings should be interpreted with caution. In general, non-responders were more likely to be younger, obese, higher comorbidity, higher ASA class and live at a greater distance from the Mayo Clinic, characteristics associated with poorer pain and/or functional outcomes after THA/TKA. However, it is unlikely that the association of ipsilateral joint involvement with index THA/TKA pain and function outcomes differed by these characteristics, in absence of any such published data. Therefore, the direction of impact of non-response bias on our findings is unclear.
The joint registry does not provide detailed data on disease pathology in all other joints, and may miss interval arthroplasty in the ipsilateral joint if performed at another institution and this was not reported by the patient in their mailed survey response or telephone interview. Therefore, we are unable to comment on the underlying disease/pathophysiology responsible for ipsilateral joint involvement. Another limitation is that diagnoses of comorbidities were based on the presence of respective ICD-9 codes, making underdiagnosis and misclassification bias likely. However, misclassification would bias our findings towards null; therefore our estimates are conservative. We made an ‘a priori’ decision to combine moderate and severe categories based on our clinical judgment of what would be considered suboptimal by operating surgeons, but also to have enough events to analyze predictors of poor outcomes. Despite our efforts to include several important variables, residual confounding is possible.