According to the results, the knee flexion angles of patients in groups C1 and C2 on the 7th day and 3 months after surgery were better than those of patients in the control group, and the pairwise comparison also showed a significant difference. At this time, the knee flexion angle was obtained on the basis that no local anesthetic was used after removing the femoral nerve catheter, and the VAS scores of the three groups showed no significant difference, which could be considered as an evaluation in a completely natural state and of high value for the analysis of recent knee joint function. In group NS, the standard deviation of the flexion angle on the 7th day after surgery was relatively large, and some patients could achieve an angle of 120°, while some patients could only achieve an angle of about 90°. One of the reasons was the individual’s pain tolerance during training. The standard deviation of the flexion angle at 3 months after surgery did not decrease, which may be strongly associated with the influence of periarticular adhesion and psychological factors during training. Therefore, during postoperative rehabilitation training, femoral nerve block before training can improve the effect of rehabilitation training, especially in patients with poor pain tolerance. Escramilla et al. believed that knee flexion angle above 100° is considered to be able to squat, which enables people to complete the movements of kneeling and cross-legged sitting, which can greatly improve the standard of daily life. Thus, in our study, we further compared the number of patients with a knee flexion angle above 100° in the three groups, and we found that there were more patients in groups C1 and C2 than in the control group on the 7th day and at 3 months after surgery. AKS scoring is a commonly used method to evaluate knee joint function [7,8]. In our study, AKS scores on 7th day and at 3 months after surgery showed no significant difference among the three groups, which was consistent with the results reported by Sakai et al. . We took into account that active pain occupies a greater weight in the AKS knee score, and the functional score mainly focuses on daily basic functions (walking and stairs) without squatting and other actions. If there is little difference in active pain among the patients of different groups or the knee flexion angles reach above 90°, then the function of the knee joint cannot be further distinguished. Therefore, the evaluation of the effect of knee rehabilitation training after surgery should have a specific scale to make the evaluation more scientific and reasonable. Previous studies [9,10] have focused on continuous femoral nerve block (CFNB) continuous infusion within 3 days after surgery. It is believed that the enhancement effect of CFNB on postoperative training is mainly through reducing postoperative pain, postoperative knee swelling, and muscle spasms.
Local anesthetics play an important role in the recovery of knee joint function by reducing peripheral inflammation[11,12] and inhibiting the release of pro-inflammatory mediators[13,14]. Combined with the results, the VAS score at 3–6 days after the operation still showed active pain; thus, it was more significant to continue the use of nerve block to reduce the inflammatory response at this stage to achieve a better training effect.
A possible decline in the quadriceps femoris muscle strength is considered as a defect of CFNB [15,16]. Ilfeld et al.  believed that continuous infusion will cause a decrease in the quadriceps muscle strength, while other researchers [18,19] believed that it can improve the postoperative walking training ability and achieve the rehabilitation standards earlier. However, if local anesthetics affect the quadriceps muscle strength for a long time, it will not only increase the risk of falling, but it is also not conducive to early joint function exercise. Therefore, in this study, chloroprocaine with an onset time of 6–10 min and a duration of action of about 45–60 min was used as a local anesthetic. The results showed that one patient in group C1 and six patients in group C2 had difficulty in active knee joint movement during training due to the quadriceps femoris muscle weakness. It interfered with the training by waiting for the quadriceps muscle strength to recover before the training was performed during active joint movement. All of them returned to normal levels about 1 h after administration. However, it remains to be explored whether it is possible to further reduce the concentration of chloroprocaine combined with a certain amount of non-steroidal drugs to avoid the impact on quadriceps muscle strength while maintaining optimal analgesia. Adductor canal block (ACB) is considered to be a blocking method that can retain femoral nerve analgesia and has little influence on quadriceps femoris muscle strength . However, some authors believed that femoral nerve block (FNB) was associated with the perception of less severe pain after TKAs . Kim et al.  reported that FNB may be superior to ACB in analgesic intensity. From the results, the active pain is still very severe during rehabilitation training, and some patients achieve the degree of severe pain. The effect of ACB may not be as good as that of FNB, and further study is needed.
Long-term CFNB is considered to have a risk of infection [9,10] ; thus, there are very few literary works on CFNB for more than 3 days. In this study, it took 6 days to remove the femoral nerve catheter from the insertion of the catheter, and none of the patients had local or systemic infections. This is closely related to strict aseptic operation during catheterization, and two sessions of disinfection during rehabilitation training and replacement of the film.
A total of 90 patients were enrolled in this study, among which 4 cases were lost to follow-up, 1 case developed femoral nerve catheter blockage, 2 cases pulled out the catheter unexpectedly, 1 case withdrew due to postoperative delirium, and 2 cases refused training. Other shortcomings were that only relatively old (65–80 years old) and ASA I-II grade patients were taken as subjects, and there was no further study of postoperative rehabilitation in the elderly and patients with poor physical condition. In addition, obese patients were not included in the study, and the effect of this method on postoperative recovery in this population could not be analyzed. In the next step, we can continue to study these two populations.