A nomogram for predicting PLOS in PACU for patients undergoing elective surgery was developed and validated with excellent discrimination and calibration in the current study. Combined with the nomogram, individualized prevention and intervention based on the risk factors can be achieved, thereby improving operating room turnover, and implementing protocols of ERAS.

Patients with endotracheal tube into PACU were included in this study for the purpose of consistency in the start time of anesthesia recovery. Some patients may be delayed entering the PACU after extubation in the operating room due to a shortage of beds. However, these patients need only a short stay to meet the PACU discharge, which makes the stay time in PACU not an accurate reflection of recovery quality. Therefore, in order to analyze the recovery more accurately, patients with endotracheal tube were included in this study. There is no definition of "PLOS". Some studies defined it as length of stay > 2 hours [4] or > 60 minutes [17]. Rodney A Gabriel et al defined it as greater than the third quartile range [4]. Subsequent similar studies have also adopted this definition [6]. For consistency, this approach was employed in this study. Based on clinical experience or related research reports, the following risk factors statistically analyzed in this study may not be surprising, but these factors were quantified in this study to predict PLOS in PACU and were used to establish a nomogram.

Age and BMI were analyzed as risk factors. Rodney A et al. found that age ≥ 65 was not a risk factor for PLOS in PACU [4]. However, Cao et al found the opposite result in the gallbladder surgery population [6]. The reason for choosing 65 as the cutoff point was not explained in their study. In order to make the diagnosis more accurate, we determined the diagnostic reliability of an age cutoff point of 55 by calculating statistical data including sensitivity, specificity, positive predictive value, negative predictive value, Cohen’s Kappa coefficient and AUC. Age ≥ 55 was a risk factor in the study. Most studies only used dichotomies for BMI [4, 6, 18]. Therefore, low body weight had been ignored. So, we distributed body weight in quartiles, using 25% and 75% as cut-off points. This study showed that low body weight (< 21 kg/m2 ) was a statistically significant risk factor. Some studies had reported that BMI ≥ 40 kg/m2 was a risk factor for delayed postoperative recovery, mainly because morbid obesity increase the risk of sleep apnea especially in the setting of opioid utilization [4, 19]. The reason why our study did not reach a consistent conclusion may be due to the characteristics of the patient population included in this study. In this study, the number of patients with BMI ≥ 40 kg/m2 was 13, accounting for 0.05%, which may be caused by regional dietary and living habits. One of the five diagnostic criteria for frailty syndrome is weight loss. These patients have a poor tolerance to anesthesia and surgery, therefore requiring longer recovery time after surgery [20]. In addition, Cao et al. had reported that the lower the BMI, the longer the postoperative length of stay in PACU [6].

It is not surprising that ASA, major surgeries, and operation duration are risk factors for PLOS in PACU. Major surgeries have been thought to cause inflammatory induced postoperative cognitive decline [21], which is associated with poorer postoperative recovery, increased medical costs and higher mortality [22, 23]. Postoperative delirium increased PLOS in PACU occurrence in the study, which consisted with above. Preoperative fluid fasting for more than 6 hours was reported as an independent risk factor for delirium [24]. Incidence of cognitive decline in elderly patients under general anesthesia at bispectral index 55-65 was higher than at 40-50 [25]. Chernov and his colleagues conducted neuropsychological tests pre-operation and post-operation, which determined the association between regional cerebral blood flow reduction and poor cognitive performance [26]. For major surgery, regional nerve block combined with general anesthesia could effectively prevent cognitive decline [21], which decreased incidence of PLOS in PACU. Whether the risk of delirium differs between general anesthesia with inhalants and total intravenous anesthesia remains controversial [27]. Perioperative use of dexmedetomidine, paracetamol and Nonsteroidal Anti-inflammatory Drugs (NSAIDs) has been suggested to prevent postoperative delirium by directly alleviating neuroinflammation [28, 29]. The most relevant medications with the onset of delirium are benzodiazepines, gabapentin, and scopolamine [29]. Hypothermia is also associated with complications such as PLOS in PACU, delirium, pain, and chills [30]. Strengthening intraoperative temperature management is beneficial to improve postoperative recovery.

This study showed that pain and naloxone use were risk factors of PLOS in PACU. Anesthesiologists use different opioids depending on surgery and individual patient. By converting different opioids into morphine and then standardizing them according to the patient's weight and length of operation, this would make nomogram complicated and impractical. So, this study collected data on the use of naloxone, indirectly reflecting opioid overdose. Since both insufficient analgesia and opioid overdoses can lead to PLOS in PACU, it is necessary to explore better methods of monitoring analgesia. In recent years, various monitors have been invented for quantifying pain, such as analgesia nociception index, skin conductance, pupillometry, nociceptive flexion reflex threshold, surgical pleth index and qNOX [31]. The nociception level index, a multi-parameter artificial intelligence-driven index, can objectively guide opioid administration, potentially leading to more appropriate analgesic regimen [32]. Additionally, incorporating a multimodal analgesia approach and reducing opioid use are considered key strategy [33]. Some studies have shown that dexmedetomidine is more effective than other agents in perioperative analgesia [34,35,36]. However, Bhiken I Naik et al. found no such positive effect after multistage deformation-correcting spinal surgery [37]. In addition, intraoperative intravenous lidocaine (1.5 mg/kg induction followed by 2 mg/kg/h) can reduce postoperative pain and promote recovery [38]. Meta-analysis shows that NSAIDs is beneficial, such as lornoxicam, pregabalin, ibuprofen, gabapentin, and acetaminophen [39]. In a network meta-analysis of 52 trials involving 2112 subjects, the pre-emptive analgesic effect of gabapentin was evaluated as significantly reducing postoperative pain scores, opioid consumption, and the incidence of PLOS in PACU [40]. In addition to adjuvant analgesics, combined nerve block and intrathecal anesthesia are also effective [41].

This study found that intraoperative transfusion of blood products prolonged postoperative recovery. Disruption of physiological homeostasis is the main reason. In addition, transfusion of blood products had been reported to cause postoperative residual neuromuscular blockade and affect neuromuscular functional recovery in the PACU [42]. It remains unclear whether blood transfusion also affects the pharmacokinetics and pharmacodynamics of other anesthetics. Regarding blood transfusion, the "yes" and "no" were used as cut-points in the study. If a cut-point "> x units" was used as a predictor, this may improve the nomogram's predictive power. However, the variety of blood component types and quantities were collected in this study. Treating each blood component as an independent variable and counting their respective quantities would make the resulting nomogram complex and inconvenient for clinical use, especially in the environment of busy clinical anaesthetic work. In addition, this study found that neostigmine use was not a risk factor for PLOS in PACU, possibly because neostigmine does not accurately reflect neuromuscular blocker overdoses. The amount and type of neuromuscular blocker were not collected in this study, mainly to consider that if it is a risk factor, it should be standardized according to drug type, operation duration and patient weight before conducting nomogram score, which will make the nomogram complex and less useful.

The study has several limitations. First, some variables that may contribute to improving the predictive ability of the model were not included, such as cerebral apoplexy etc. Second, we did not collect detailed usage and types of opioids and neuromuscular blockers. Third, the nomogram was not verified externally. In future clinical work, we will further verify the feasibility of the model.

In summary, we have developed a predictive model with excellent differentiation and clinical utility that can help identify patients with a higher probability of PLOS in PACU. The screening of high risk factors and the establishment of nomogram are helpful to improve the quality of ERAS and the efficiency of the operating room.