*4.1. Incidence of Delirium*

In this study, RNB reduced the incidence of delirium. It is widely accepted that pain reduction is important for the prevention of delirium [6], and we hypothesized that postoperative pain control could prevent delirium after hip fracture surgery. Our study results proved the effect of RNB on the incidence of postoperative delirium after hip fracture surgery. Herein, we excluded patients with a high risk of developing delirium, including those with psychiatric illness and cognitive impairment, which was based on a study by Mouzopoulos et al. [22]. They classified patients into intermediate or high risk for postoperative delirium, and FICB did not affect the incidence of delirium in patients with high risk after hip fracture surgery. In contrast, a significant reduction in the incidence of delirium (FICB group vs. placebo group: 2.4 vs. 16.9%) was observed among patients with an intermediate in their study.

The results of RNB and delirium in patients with hip fracture are inconsistent. RNB was reported to be associated with less postoperative analgesia, a lower incidence of delirium, and shorter inpatient stay [24,25]. In contrast, Guay et al. presented that there were no differences in the incidence of acute confusional state in their Cochrane review based on seven trials with 676 participants [26]. However, the study has a limitation of heterogeneity and lack of risk stratification. Unneby et al. reported that femoral nerve block (FNB) did not reduce the incidence of postoperative delirium in patients with hip fracture [27]. They focused on patients with dementia, and a large proportion developed delirium (FNB group vs. placebo group: 50/52 vs. 55/57) regardless of FNB use. This suggests that patients with major risk factors, such as cognitive and psychiatric disorders, are highly prone to developing delirium regardless of pain control. In the present study, we could prove the effect of RNB on the incidence of delirium by excluding high risk patients with the inclusion of a relatively large number of patients as a single study. Additionally, we performed a multivariable analysis with other known risk factors of delirium [21–23]. Age and RNB were presented to be associated with postoperative delirium, which means an appropriate risk stratification of present study. Therefore, we believe the RNB helps to prevent postoperative delirium after hip fracture surgery except in high risk elderly patients.

#### *4.2. Effect on Pain Intensity and Opioid Consumption*

Opioids are useful in reducing pain after surgery but have limitations regarding side effects and drug poisoning [28]. Elderly patients are known to be vulnerable to the side effects of opioids, with reductions in renal and hepatic blood flow [29]; in our study, one of our purposes was the reduction in opioid consumption with the effect of nerve block. Thompson et al. reported that preoperative fascia iliaca block significantly decreased postoperative opioid consumption [30]. They reduced the amount of tramadol by 43% and morphine by 98% with fascia iliaca block. However, contrary to our expectations, there was no significant difference in the amount of opioids consumed despite early pain reduction in our study. The first possible reason is that diverse types of opioids were prescribed with a retrospective feature. Although we calculated the equianalgesic dose of each opioid into milligrams of oral morphine [18–20], the results could be influenced by the diversity of the opioid types prescribed. The second possible reason is that all patients received

intravenous PCA because it is a customary procedure desired by patients. In addition to additional opioids injected by nurses, rescue opioids were also administered through PCA, which were not included in the calculation of the quantity of opioids consumed. These reasons may have influenced the outcome for the quantity of opioids.

#### *4.3. Functional Recovery*

Based on the results reported by Marino et al., who demonstrated that continuous lumbar plexus block provided pain reduction during physical therapy [31], we expected that functional recovery could be encouraged with nerve block. However, there was no significant difference observed in the time to wheelchair ambulation between the two groups, which can be explained by some reasons. A consistent rehabilitation protocol was applied to the patients in both groups. In addition, Kim et al. demonstrated that the postoperative ambulatory capacity after hip fracture surgery is decided not by only a single factor but by multiple factors, including age, sex, preoperative ambulatory capacity, and combined medical diseases [32]. Since the postoperative ambulatory capacity is significantly associated with preoperative factors, it is possible that the reduction in pain itself could not affect the short-term functional recovery.

#### *4.4. Subgroup Analysis According to Type of RNB and Anesthesia Method*

In clinical practice, we usually perform LPB after induction of general anesthesia because it should be conducted in the lateral decubitus position. In contrast, FICB is usually performed before induction of spinal anesthesia because it could be conducted in the supine position, which could reduce pain for positioning of spinal anesthesia. Since LPB and FICB could block both anterior innervations of the hip joint and some surgical incision site, similar pain reduction and delirium prevention could be expected in both blocks (Figure 2). In our study, the subgroup analysis between RNBs did not demonstrate significant differences in the postoperative pain score and amount of opioids consumed, but time to wheelchair ambulation of PENG block presented significant differences compared to that of FICB and LPB. The possible reason for these results was that PENG block was introduced as a potential motor sparing analgesic block [14,33], which could encourage patients to ambulate early. However, investigation for PENG block is very limited, and validation to propose motor sparing and analgesic benefit is needed [33]. Additionally, with a relatively small number of patients included in each RNBs in the present study, detailed analysis according to block type requires careful interpretation.

There could be concerns regarding anesthesia method. Previous studies showed controversial results. Choi et al. reported general anesthesia was an independent predictor of immediate delirium [23]. In contrast, Patel et al. concluded that there was no evidence to suggest that anesthesia type influence postoperative delirium in their systematic review [34]. In the present study, there were no significant differences in the incidence of delirium between general and spinal anesthesia in subgroup analysis, and multivariate analysis showed no effect on postoperative delirium according to anesthesia method.

#### *4.5. Procedure-Related Complications*

Falls and procedure-related nerve injuries are important complications of RNB in patients with hip fracture [35,36]; in our study, no complication was observed. While wheelchair ambulation was encouraged from the day after surgery, the protocol of our institution was to involve at least three individuals (nurse, caregiver, and paramedics) in the transfer of a patient to a wheelchair. At 2 days postoperatively, the patients were permitted to bear weight as tolerable. Considering that the duration of a single shot of bupivacaine 0.5% (20 mL) is 22 (range, 15–32) hours [37], it is unlikely that a fall would occur during ambulation owing to RNB. Further, all RNBs were performed by experienced anesthesiologists under ultrasound guidance. We believe that our rehabilitation protocol and technique for RNB could ensure the safeness of RNB in patients with hip fracture.
