This is the first study evaluating weight change after BS in patients who have had TKA either before or after BS. Further, this is the largest study evaluating the risk of revision after TKA depending on the sequence of surgery. We found no difference in weight change up to two years postoperatively depending on the sequence of surgery regarding patients who have had both TKA and BS. Furthermore, no statistically significant difference in risk of revision was found between the two cohorts.
Previous studies investigating weight change after TKA are inconsistent [10–12]. Teichtal et al observed 29 patients with TKA for 6 months following the surgery, with a mean BMI of 31.5 at the time of TKA. The majority of the patients (59%) lost weight (> 0kg). When using 5% as a threshold of weight change considered to be clinically significant, 38% of the TKA patients had lost weight after the procedure. Furthermore, 35% of the patients had gained weight after the TKA (> 0kg). The mean BMI change was − 1.07 (SD 1.80) corresponding to -3.3% (SD 5.7) reduction in weight [10]. In the current study, the median change of weight after TKA was demonstrated to be a gain of 5kg in the TKA-BS group. Ast et al reviewed 3,036 patients operated with TKA with a mean BMI of 30.2 at the time of TKA surgery. No change in BMI two years postoperatively was seen in 69% of the patients [11]. In the current study, the patients in the TKA-BS group had a median weight gain of 4%. Inacio et al assessed the weight patterns before and after TKA or total hip arthroplasty (THA). They demonstrated that most of the patients with TKA (68.5% of 20,060) had an unchanged weight after the procedure, when defining a change of 5% as clinically meaningful. The mentioned studies investigated patients with lower BMI compared to our study, since all of the patients who had TKA prior to BS in our study where candidates for BS [12]. Nearing et al (2017) evaluated the outcomes after TKA or THA in patients who have had BS either before or after their TKA or THA. They included 102 patients with either TKA or THA. TKA or THAs were performed at a mean of 4.9 years before and 4.3 years after BS. Information on obesity-related co-morbidities was available in their study, and these were similar between the groups. Patients who had TKA or THA before BS had an average increase of BMI 2.6 between the TKA or THA surgery and BS, which is in line with our results. However, there was a statistically significant lower BMI in patients with TKA or THA after BS one year after TKA or THA compared to patients with TKA or THA before BS [13]. Nonetheless, in our study, we compared one-year postoperative weight after BS. In a recent published randomized control trial analyzing change in BMI and weight one year after TKA, the intervention group who received BS prior to TKA had a significant greater BMI loss (-6) and weight loss (-16.5 kg) compared to the patients with “treatment as usual” before TKA. However, 2/41 patients did not have BS prior to their TKA, and 12/41 did not undergo any TKA in the intervention group, but were still included in the intention to treat analysis. Furthermore, the majority of the patient in the intervention group had undergone gastric banding which differ from the patients in our BS group [14].
In a relatively recent systemic review, obesity was shown to have a negative impact on TKA regarding risk of revision [7]. Sezgin et al. found that obesity was associated with overall risk of revision and revision for infection, but could not show the same relationship for revision for reasons other than infection [15]. Since BS is an effective method of obtaining long-term weight loss [3], it is reasonable to believe that BS prior to TKA could reduce the risk of revision. However, in a previous study we did not find any association between a reduced risk of revision in patients undergoing BS prior to TKA [16]. Risk of revision depending on sequence of surgery have also been studied, demonstrating similar results as our study [13, 17]. Nearing et al (2017) did not find any difference in risk of revision between the groups regardless of timing of TKA or THA in relation to BS. The mean follow-up after TKA or THA was 3.2 years in those who had TKA or THA after BS and 9.2 years in those who had TKA or THA before BS, in comparison to our study where we had a limit of two years between the surgeries [13]. In a retrospective study, Kulkarni et al (2011) evaluated one-year revision in 53 patients with TKA or THA before BS and 90 patients with TKA or THA after BS. No patients with TKA, whether with BS before or after TKA, were reported to have a revision within 1 year [17].
Even though the mentioned studies did not find any difference regarding risk of revision of TKA, and we did not find any difference in weight change postoperatively to BS regardless of sequence of surgery, there are other factors to take into consideration. A recent retrospective cohort study investigated the risk of medical complications after second surgery in patients with both BS and TKA/THA depending on the sequence of surgery. When adjusting for comorbidities, their results indicated that BS before TKA or THA were associated with improved postoperative outcomes. However, they did not include revision of TKA in their outcomes [18].
Although the present study carries the benefits of including a nationwide cohort based on prospectively collected data from two high-quality sources [8, 9] it is not without limitations. The study was an observational study why it only allows interpretation of associations and not causation. Furthermore, data on comorbidities were not available and therefore not included in the analyses. However, the SKAR obtains information on the American Society of Anesthesiologists (ASA) classification and the SOReg obtains information on obesity surgery mortality risk score (OS-MRS) but these variables were not adjusted for in the regression model, due to the interaction with BMI. Nevertheless, the patients included in the study have undergone elective surgery and have been optimized prior to the surgeries by both the surgeon and an anesthesiologist. The majority of the patients in both groups were classified as ASA 2 prior to the TKA surgery. In order to reduce the risk of other confounding health factors affecting the outcome, the cut-off time between the two surgical procedures was set to two years. The 95% CIs are relatively wide in our study which may indicate a potential problem with statistical power.