**4. Discussion**

Degenerative diseases including osteoarthritis and spinal stenosis are serious public health concerns globally because of the severe pain and disability they cause [14]. Specifically, lower back pain and osteoarthritis were the first ranked and 12th ranked, respectively, global burden of diseases that cause disability from a systemic analysis in 2016 [15]. Moreover, these chronic conditions lead to multi-morbidity, which limit function and cause pain and disability [14,15]. However, the impact of multi-morbid conditions has not been extensively studied ye<sup>t</sup> [14]. In an arthroplasty study, the impact of total hip arthroplasty in spinal fusion was reported in hip-spine syndrome, but there is a relative lack of evidence for that of TKA [4]. Therefore, this study aimed to identify the impact of preoperative TKA in spinal fusion for patients with concurrent severe KOA and degenerative LSD.

Regarding preoperative radiological parameters, our results showed that LL and sagittal spinopelvic parameters were worse in the TKA group. There were attempts to elucidate the association between radiological factors of the spine and flexibility of the knee [6,16,17]. Flexion contracture of the knee was associated with not only loss of LL, but also poor sagittal spinopelvic parameters [16,17]. Kim et al. suggested that lumbar flexibility is important for spinal and lower limb alignment following TKA [7]. However, the studies reported that removal of flexion contracture by TKA could not compensate for sagittal global imbalances [5,6]. The results have similar preoperative aspects of worse LL and sagittal spinopelvic parameters, which support the finding that TKA does not compensate for these parameters. Our results sugges<sup>t</sup> the patients that require both TKA and spinal fusion have relatively worse preoperative radiological outcomes in LL and sagittal spinopelvic parameters. Therefore, sagittal spinopelvic parameters could consider one of the factors for surgical decision-making in the patients with severe KOA and degenerative LSDs.

The pelvic morphology, which is influenced by sagittal malalignment, was significantly different in elderly patients with concurrent KOA and degenerative LSDs compared to patients with LSD only [18]. Increased sagittal malalignment with a lack of LL was caused by double-level listhesis (i.e., spondylolisthesis and/or retrolisthesis) and greater knee flexion [19]. Although decompression with short-segment fusion at less than three levels can yield improvement of clinical outcomes, corrective lumbar surgery alone may be insufficient for radiological outcomes because of greater pelvic retroversion (high PT) and, worse sagittal spinopelvic alignment [20,21]. Kohno et al. reported that surgical strategies in concurrent degenerative knee and LSDs may be necessary to restore sagittal spinopelvic alignment, followed by decreased pelvic retroversion [18]. In our study, patients with preoperative TKA exhibited greater pelvic retroversion than patients with KOA, and more often required fusion surgery for correction of sagittal spinopelvic alignment. The optimal values of sagittal spinopelvic parameters that need to be corrected was under-estimated by compensatory mechanism of spine from knee stiffness in non-TKA group. Therefore, preoperative TKA could be a benefit for in proper correction of sagittal spinopelvic alignment by spinal fusion.

Schwab et al. showed a PI/LL mismatch that reflected the disharmony between spine and pelvis correlate with increase in ODI [22]. From our result, the preoperative TKA group (i.e., the patients who needs to both spinal fusion and TKA) showed worse ODI values. Because TKA with worse sagittal spinopelvic parameters is associated with poor range of motion, it led to dissatisfaction and did not improve disability [6]. For significant improvement of ODI in the TKA group, preoperative TKA may have contributed to more vigorous activity by resolution of neurogenic claudication. The most important thing in our study was that complementing compensatory mechanisms by preoperative TKA gave

a chance for better correction of sagittal spinopelvic parameters, which has a significant impact on improving disability. The value of ODI reflects pain as well as activities of daily living affected by knee discomfort [4]. Lee et al. reported that the presence of preoperative KOA and multi-level fusion were poor prognostic factors in lumbar spinal surgery, and Lee et al. also showed worse ODI scores in the patients who underwent TKA before spinal fusion on retrospective case analysis [23]. However, considering that our study included patients with spinal fusion at less than three levels, preoperatively worse spinopelvic sagittal parameters as well as lower lumbar lordosis contributed to a higher ODI level in the preoperative TKA group compared to the non-TKA group [24]. If the case of long-level spinal fusion and instrumentation, this can clearly affect balancing and lumbar spine alignment by nonunion and/or instrumentation failure. Therefore, in order to minimize this effect and evaluate the impact of preoperative TKA, we assessed only the patents who underwent spinal fusion at less than three-level (i.e., short-level fusion). Preoperative TKA in spinal fusion at less than three levels could be helpful for predicting disability and pain in the case of worse sagittal spinopelvic parameters.

Lower back pain is affected by various factors, and has a broad spectrum of symptoms that requires differential diagnosis based on degenerative, congenital, and traumatic causes [25]. Escobar et al. reported the preoperative absence of lower back pain in TKA as a predictor of a good quality of life in a multi-center prospective study conducted in 2007 [26]. Pivec et al. also suggested that the presence of spinal stenosis was associated with worse clinical outcomes following TKA [27]. However, little is known about the clinical relevance between back pain and preoperative TKA for fusion surgery in patients with KOA. In our study, back VAS was not significantly different between the two groups, which indicates that preoperative TKA in spinal fusion does not seem to have much impact on lower back pain. Preoperative TKA in spinal fusion showed better clinical outcomes in terms of leg VAS, which means significantly improved pain. Lumbar radiculopathy by nerve root compression from L3 to L5 is a typical clinical presentation of spinal stenosis, which share the same portion in anterior knee pain by joint degeneration [28]. Furthermore, the origin of pain from knee and/or spine could be impact on determining clinical outcomes [29]. Therefore, preoperative TKA in the case of short-level spinal fusion significantly impacts improvement by eradicating the pain source.

There were several limitations to our study. First, the number of patients was relatively small and we used a retrospective design. Future trials would be needed by large sample in multicenter study and/or meta-analysis. Secondly, this study did not reflect the morphology and clinical scales of the knee. It also included the limitation of being a retrospective study, which suggests the need to evaluate radiological factors and clinical function of the knee in future trials. However, our study focused on comparing radiological factors, function, and pain measures limited to the spine. Large multi-center prospective studies should be needed to perform to confirm our results. Nonetheless, our study suggested that preoperative TKA in spinal fusion (less than three levels) have significantly impact on lumbar radiculopathy and disability.
