Shortening LOS can reduce the chance of cross-infection during hospitalization, especially during the novel coronavirus epidemic. It also can decrease the direct contact time between patients and doctors, which may reduce the potential risk of disease transmission[14]. Varying healthcare systems in different countries determine different LOS[3]. Most of the related researches were conducted in Europe and America, where Chinese patients with ERAS protocol were less covered. However, many factors may increase the LOS in TKA patients, which are still remain controversial. Therefore, it is important to investigate the clinical outcome of TKA patients with ERAS program in China and to identify the independent risk factors of prolonged LOS[15].
LOS is closely related to the prognosis of TKA in ERAS protocol, however, there are many factors that affect LOS after TKA remain controversial [6]. The major findings of this study were that LOS decreased linearly over time in TKA patients under ERAS protocol and was significantly associated with age, comorbidities, the operation time and the surgical day of week. Moreover, we are the first authors who focus on the surgical day’s difference in the relationship between LOS and TKA under ERAS protocol. We also found a longer stay was not significantly associated with gender, BMI, preoperative Hb, surgeons, ASA-PS class and blood transfusion.
In the comparison of baseline data, we found that 5 observation indicators were related to prolonged LOS, namely age, operation time, IHD, diabetes and the day of surgery. There is no consensus on whether gender and BMI contribute to prolonged LOS, and some investigators have found that women tend to have longer LOS than men due to higher rates of obesity, postoperative blood transfusions, and postoperative complications [16]. However, there are also some studies that believe that gender and BMI are not the main driving factors leading to prolonged LOS [17]. In our study, the difference between males and females was not significant, and the difference in BMI between the normal LOS group and the prolonged LOS group was not significant either. One result may be due to an unbalanced sex ratio (1:3) between males and females in the study. Therefore, more research is needed to investigate the effect of gender and BMI on LOS after TKA. Anis et al. [18] and Podmore et al. [19] found that patients with pre-existing comorbidities had variable effects on LOS after total knee and total hip replacement surgery, and in particular, diabetes mellitus were significantly associated with prolonged LOS. Our study found that patients with IHD were independent risk factors for prolonged LOS (OR = 4.917, 95%CI 1.046–23.114, P = 0.044), which was basically consistent with their findings. In addition to gender, BMI, smoking, preoperative Hb, ASA-PS classification, and blood transfusion status were not risk factors in our study (P > 0.05). These observational indicators were different from other studies on risk factors for prolonged LOS without ERAS intervention [20]. This may be due to the strict implementation of ERAS protocol in our department. After admission, patients with anemia were treated with erythropoietin and intravenous iron, and nutrition was enhanced. At the same time, the tranexamic acid was routinely used for anticoagulation to maintain the physiological balance of coagulation and fibrinolytic system, as well as good perioperative anesthesia management. Low perioperative global transfusion rate (8.9%) and high postoperative Hb level are guaranteed, which may minimize the impact of comorbid diseases on LOS, make patients more comfortable during and after surgery, and achieve early and rapid recovery. Similarly, the influence of surgeons on prolonged LOS after TKA was not positive, which was inconsistent with the findings of Monsef et al. [21]. This may be related to the fact that the two surgeons in our department have implemented the minimally invasive operation concept throughout the operation under ERAS management and complied with discharge standards after ERAS protocol.
In addition, a number of studies had found that age was an important predictor of LOS in patients with THA/TKA. Sibia et al. [22] found in their study of THA in ERAS mode that the ratio of prolonged LOS by more than 1 day in patients aged 70 years was 1.8 times higher than that in patients aged 60 years. In an age-stratified analysis of TKA patients in our ERAS mode, the risk of LOS prolongation in patients aged > 70 years was 3.32 times higher than that in patients aged > 60 years. This may be related to older patients' poor physical condition, higher likelihood of postoperative complications, and greater need for additional nursing support than younger patients. In the stratified analysis of the operation time, we found that patients with an operation time of more than 90 minutes had a 1.966-fold higher risk of prolonged LOS than those with an operation time of less than 90 minutes. The longer the operation time, the higher the risk of prolonged LOS. This is consistent with the study of Inneh et al. [23]. It is considered that the use of tourniquet for too long may cause ischemia-reperfusion injury of the affected limbs and thighs [24] and delay the early rehabilitation exercise of the patient. Another explanation may be the association with prolonged operation time and increased risk of infection after TKA [25].
Another strength in this study is that we performed a statistical analysis to control for the day of the week the surgery was done. There is still controversy about the influence of surgical day of the week on LOS after TKA. Edwards PK et al. [26] found that operation day is not associated with LOS after TKA. Recent study by Newman [27] et al. has shown that surgery day is associated with LOS after TKA/THA. In our analysis of surgical day selection, we also found that patients undergoing TKA on Monday had a higher risk of prolonged LOS than those undergoing TKA on Tuesday, Wednesday, and Thursday. This is inconsistent with the results of Muppavarapu et al. [28], but consistent with the results of studies such as Lilly et al. [29]. It may be that our average LOS after TKA is (4.6 ± 2.5) days, and most of patients undergoing surgery on Monday will be discharged on weekends. The relatively limited admission of new patients on weekends, fewer on-site staff to help with admissions on weekends, and fewer physiotherapy teams may all lead to longer hospital stays for patients expected to be discharged on weekends.
There are some limitations in this study. First, there is a lack of more survival data to explore, and further relevant follow-up data are needed to evaluate the postoperative complications and outcomes of TKA in the ERAS mode. Second, the lack of quantitative assessment of the specific implementation of the ERAS pathway may introduce additional variability. Third, it is a single-center retrospective cohort study, the sample size of the two groups is quite different, and no propensity matching analysis is performed, which will cause a certain error between the two groups of samples. Fourth, the sample size of this study was small, all of the patients analyzed were from our institution, thus, the result may not be representative of all hospitals in China. In addition, our selection of the 75th centile for defining prolonged LOS can be viewed as an arbitrary cut-off in the absence of a universal definition of prolonged LOS, the use of 75th centile has been done in literature before for a similar study [13, 30].