Predictive Study of Pharmacological Reversal for Residual Neuromuscular Blockade, Early and Late Postoperative Pulmonary Complications: A Prospective, Observational, Cohort Study

Monitoring and Pharmacological Reversal.


Results
A total of 714 patients were analyzed. According to our model, none of the studied comorbidities were a predisposing factor for an increase in residual neuromuscular blockade. However, the use of rocuronium, pharmacological reversal with sugammadex, and, particularly, neuromuscular monitoring during surgery markedly decreased the probability. When using rocuronium, avoidance of neuromuscular monitoring and pharmacological reversal led to an incidence of residual neuromuscular blockade of 41.27%; conversely, the use of neuromuscular monitoring and pharmacological reversal with sugammadex decreased the probability to 2.17%.

Conclusions
In our study, intraoperative neuromuscular monitoring and pharmacological reversal with sugammadex, unlike neostigmine, were the factors that most effectively reduced the risk of residual neuromuscular blockade as well as early postoperative pulmonary complications in PACU (upper airway obstruction and desaturation) and late postoperative pulmonary complications (pneumonia or atelectasis) within 30 days.

Background
More than 400 million people receive neuromuscular blocking agents annually to paralyze skeletal muscle groups, facilitate tracheal intubation, allow for controlled mechanical ventilation and achieve optimum relaxation conditions for surgery [1,2].
The possibility of residual neuromuscular blockade (RNMB) after the use of neuromuscular blocking agents has been known for some time. However, in recent years, there has been an increase in the number of publications showing its high incidence, its relationship to postoperative pulmonary complications (POPC), and, therefore, to increase potential healthcare costs [3][4][5][6]. POPC include upper airway obstruction, oxygen desaturation, bronchoaspiration, pneumonia, atelectasis, and reintubation for severe respiratory failure requiring an unplanned admission to an intensive care unit (ICU) [2,3,[5][6][7].
Numerous studies and multiple international organizations have suggested that every patient receiving nondepolarising neuromuscular blocking drugs should have at least qualitative, and preferably quantitative intraoperative monitoring of the neuromuscular blockade (NMB) and assessment of the pharmacologic antagonism of NMB [8][9][10][11]. Nevertheless quantitative measurements of drug-induced NMB and the adequacy of pharmacologic reversal have not been widely utilized by anaesthesia professionals [3][4][5]. The rates of intraoperative neuromuscular monitoring (NMM) vary according to each center and do not cover the entire surgical patients with general anaesthesia and neuromuscular blocking agents [11][12][13].
In addition, in recent years, some studies have generated controversy since they contradict most previous studies, concluding that intraoperative monitoring of the NMB and pharmacological reversal do not contribute in any way to the reduction of postoperative RNMB or pulmonary complications [2,14].
The primary objective of this study was to assess the incidence of RNMB and POPC according to spontaneous or pharmacological neuromuscular reversal, comparing neostigmine and sugammadex. The secondary objective was to present a prognostic model to predict the probability of having RNMB depending on the patient's comorbidities and the intraoperative management of the neuromuscular blocking agents.

Study Design and Setting
A single-center, prospective, observational, cohort study was designed that included patients undergoing elective or emergency surgery at Miguel Servet University Hospital in Zaragoza from January 2016 to December 2019. The reporting of this study conforms to the STROBE statement.

Ethics
The study was rst approved by the Ethical and Research Committee of Miguel Servet University Hospital, Zaragoza, Spain, with registration code 06/2014 (Chairperson J.M. Larrosa Poves) and subsequently it was reauthorized by the Regional Ethics Committee of Aragón (CEICA), with number CAB-SUG-2019-01 (Chairperson M. Gonzalez Hinjos) as requested by regional guidelines. Written informed consent was obtained from all subjects.

Inclusion/Exclusion Criteria
The inclusion criteria were de ned as: patients with ASA physical status I to III, age over 18 years, and signed informed consent. Exclusion criteria included patients with ASA physical status IV to V, known neuromuscular disease, diabetes mellitus with diagnosed neuropathy, pregnancy or lactation, known allergy to neuromuscular blocking agents, cardiac surgery, or planned admission to surgical ICU with mechanical ventilation. The patients were selected prior to the surgery having signed the consent form for inclusion in the study.

Patient Population and Anaesthesia
Four patient groups were established according to the neuromuscular blocking agents and its spontaneous or pharmacological reversal, based on current practice: group 1 cisatracurium without pharmacological reversal, group 2 cisatracurium and neostigmine antagonism, group 3 rocuronium without pharmacological reversal, and group 4 rocuronium and sugammadex antagonism.
The recruited patients were those who received neuromuscular blocking agents under general anaesthesia. Neuromuscular blockade was performed according to standard clinical practice with cisatracurium (0.1 -0.2 mg/kg) or rocuronium (0.6 -1.2 mg/kg) for anaesthetic induction at the choice of the anaesthesiologist in charge of the patient who was blinded to the patient's inclusion in the study. Similarly, anaesthetic maintenance, intraoperative NMM, repeated doses of the neuromuscular blocking agent, or the need for pharmacological reversal at the end of surgery depended on the clinical criteria of the same anaesthesiologist. If used, patients in the rocuronium group with pharmacological antagonism received sugammadex (2 -4 mg/kg), and those in the cisatracurium group with antagonism were administered neostigmine (0.03 -0.05 mg/kg) and atropine (0.02 mg/kg) according to routine clinical practice.

Measurements and Data Handling
Patient demographic data included age, weight, gender, ASA physical status and comorbidities (chronic obstructive pulmonary disease, obstructive sleep apnea syndrome, restrictive lung disease, asthma, acute myocardial infarction, heart failure, high blood pressure, anaemia, chronic renal failure, diabetes mellitus, dyslipidemia, hyperthyroidism, hypothyroidism, chronic liver disease, dementia and fragility).
The type of surgery (general surgery, maxillofacial surgery, otolaryngology, urology, vascular surgery, and others), emergency or elective procedure, quantitative NMM with acceleromyography during surgery and repeated doses of neuromuscular blocking agent were recorded as intraoperative data.

Primary and Secondary Outcomes
The main outcome was the presence of postoperative RNMB, de ned as a TOF ratio < 0.9 at admission to the PACU. A single TOF measurement (four stimuli of 0.2 ms in duration at a frequency of 2 Hz) with an intensity of 40 mA using a TOF-Watch-SX® acceleromyography device [Organon, Oss, The Netherlands] calibrated in the operating room prior to the rst dose of NMB. It was performed in 100% of the patients at admission to the PACU by the research staff, who was blinded and was not involved in the intraoperative care of the patient.
The secondary outcomes were the POPC, as de ned in other studies like ARISCAT [15] or PERISCOPE [16]. Early POPC were considered as at least one of the following respiratory events in the PACU: upper airway obstruction, desaturation below 92%, bronchoaspiration or need for reintubation for severe respiratory failure of the patient. Late POPC were de ned as at least one of the following respiratory events in the 30 days following surgery: pneumonia or atelectasis.
For this, the patient's electronic clinical history was consulted, recording any clinical event, laboratory test, radiological study, and primary care or emergency room consultation reports during hospital admission or 30 days after surgery con rming this type of POPC.

Sample Size
Assuming an incidence of residual neuromuscular blockade with rocuronium and cisatracurium of 13% [17] and 34,1% [7] respectively, with a signi cant level of 5% and a 95% of power, a sample size of 103 patients was calculated using the EPIDAT v. 4.1. software. To account for dropouts we included at least 110 patients per group. Patient recruitment was performed through a sequential review of cases in a recruitment period from January 2016 to December 2019.

Statistical Analysis
To perform data analysis, a descriptive analysis was completed using the mean, standard deviation and quartiles to summarize quantitative data according to normal distribution. For qualitative data, frequency and percentages were used. For qualitative variables, a χ2 test and a Fisher's test were used, and when proportions were compared for different groups, a difference in proportions test was used. A Kruskal-Wallis test and an analysis of variance (ANOVA) test were used to study the relationship of a qualitative variable in a quantitative variable.
To predict RNMB, a Binomial Generalized Linear Model was performed using the demographic data (age, weight, gender, ASA), the comorbidities (chronic obstructive pulmonary disease, obstructive sleep apnea syndrome, restrictive lung disease, asthma, acute myocardial infarction, heart failure, high blood pressure, anaemia, chronic renal failure, diabetes mellitus, dyslipidemia, hyperthyroidism, hypothyroidism, chronic liver disease, dementia and fragility) and the variables of the neuromuscular blockade management (neuromuscular blocking agent, intraoperative NMM, pharmacological reversal) as previously detailed.
For this, the Likelihood-Ratio test was used to select the variables of the Binomial Generalized Linear Model with Logistic Regression (logit link) that were part of the nal model. The modeling process was carried out in stages, eliminating in each stage the variables with a lower signi cance or equivalently with a higher p-value for the Likelihood-Ratio Test.
Differences for which the p-value was < 0.05 were considered signi cant. The analysis has been developed with

Results
During the study period, 735 patients were included, 21 of whom were excluded as detailed in Fig 1. STROBE patient ow diagram. Patients were divided into the four groups detailed in the methodology by type of neuromuscular blocking agent and spontaneous or pharmacological reversal (Fig. 1). The groups were homogeneous, and there were no differences between the groups in patient demographic data or comorbidities (Table 1).
Intraoperative NMM was used in 30.3% (n = 216) of patients, with no statistically signi cant differences in the four groups (p = 0.98, χ2 test). The percent administration of neuromuscular reversal agents was 36.7% (n = 262) of all patients.
If we analyze the in uence of intraoperative NMM and RNMB, patients not monitored intraoperatively had an incidence of RNMB of 35.7% (n = 178), however, when monitored the incidence decreased to 10.2% (n = 22) with p < 0.001, χ2 test (Table 3).
With regard to the respiratory events, a total of 15.27% (n = 109) of all patients had some type of early POPC in the PACU. Of the total patients, 10.92% (n = 78) presented oxygen desaturation and 4.34% (n = 31) presented upper airway obstruction. There were no cases of bronchoaspiration or reintubation for severe respiratory.
The analysis of the in uence of RNMB upon both the early and late POPC shown that all of them had a statistically signi cant relationship (p < 0.001, χ2 test) and both POPC were decreased when intraoperative NMM was performed (Table 3); early POPC (p = 0.023, χ2 test) and late POPC (p = 0.011, χ2 test).

Predictive Model for Residual Neuromuscular Blockade
All those variables with a positive coe cient estimate contributed to increase the incidence, while those with a negative coe cient decreased it ( Table 4). None of the demographic data and none of the comorbidities added to the model predisposed to having more RNMB. However, the type of neuromuscular blocking agent used, TOF monitoring during surgery and pharmacological reversal did have a signi cant effect.
When using rocuronium, avoidance of intraoperative NMM and neuromuscular blockers antagonism led to an incidence of RNMB of 41.27%; conversely, the use of monitoring and pharmacological reversal decreased the probability to 2.17% (Table 5).

Discussion
This prospective, observational, cohort study was intended to clarify certain questions arising in recent years from several international studies on RNMB. According to the available literature, this article is one of the few analyzing the patient's demographic data, comorbidities and the intraoperative management of the neuromuscular blocking agents in a single predictive model for RNMB, comparing between neostigmine and sugammadex.
We showed that the incidence of RNMB in our study was 28.3%, i.e. approximately 1 out of every 3 patients under balanced general anaesthesia presented this complication. This gure still appears to be high today; however, it did not differ from the data from the most recent studies, where the incidence ranges from 14% to 32% [18][19][20][21][22].
According to our results, RNMB was signi cantly decreased when monitoring was performed and the NMB was reversed pharmacologically with sugammadex. As seen in several studies, such as Togioka et al. [5] or Fuchs et al. [22], we showed that when an RNMB existed there was an increase in the POPC.
In cases where intraoperative acceleromyography was used, the expected probability of RNMB decreased by a little over 25%. When pharmacological reversal was used with sugammadex, this probability decreased by 17%.
But more importantly, when we performed both techniques in the same intervention, the probability decreased by more than 30%.
In this case, the incidence of RNMB in the group with sugammadex was 5.3%, since it could probably be explained by clinical error. The dosing of sugammadex should be based on actual body weight [23][24][25]. For moderate NMB, de ned as 1 to 2 twitches, the dose is 2 mg/kg, but for deep NMB, de ned as post tetanic count of 1 to 2, it is up to 4 mg/kg, and if no monitoring is performed, the degree of NMB cannot be known. Many specialists routinely use 200 mcg of sugammadex [12], and this probably often leads to overdosing, but also to underdosing, particularly in patients weighing more than 100 kg [25] . The same occurs during emergency surgery. The dose of rocuronium is usually doubled, i.e. 1.2 mg/kg, when a rapid sequence intubation is used.
Specially in these cases, as always, it is critical to use NMM and, if necessary, to use the correct dose of sugammadex per the patient's real weight and degree of blockade [24][25][26][27].
With regard to the percentage of intraoperative NMM at our center, we found that in 30.3% of the surgical procedures quantitatively monitor NMB was used; this gure is similar to those of other studies and centers [7,12,13]. As stated by Naguib et al., the percentage of anaesthesiologists who rely solely on clinical signs for extubation remains very high [28]. According to our results, this lack of intraoperative NMM led to an increase in both early and late POPC. However, of them, this statement is only true for desaturation and atelectasis since, in our sample, we have not been able to demonstrate that intraoperative NMM decreased the incidence of postoperative pneumonia and obstruction. In fact, we have distinguished between early complications in the PACU and late pulmonary complications 30 days after surgery, since the pathophysiological mechanism and the involved factors between them are very different [15,16]. The incidence for early POPC was 15.27%, and for late POPC was 8.12%, similar to those reported by Kheterpal et al. or Ledowsky et al. [3,29] After analyzing the NMM, we should see what happened with neuromuscular reversal and its in uence on POPC.
The controversy [31][32][33] lies with some studies, such as Grosse-Sundrup et al. [2] or POPULAR [14], which have reported that reversal, with one reversal agent and another, was not able to decrease these postoperative complications, and moreover, they questioned the utility of quantitative monitoring "the use of reversal agents or neuromuscular monitoring could not decrease this risk." [14].
In our study, we found that pharmacological reversal with sugammadex was associated with a lower risk of suffering early POPC in the PACU. When rocuronium and sugammadex were used instead of cisatracurium and neostigmine, the incidence of desaturation decreased by approximately 12% and, in the case of upper airway obstruction, by up to 2%.
With regard to late POPC and pharmacological reversal, we also showed that sugammadex was able to reduce them. In fact, reversal with neostigmine did not appear to decrease, but to subtly increased, the incidence of these complications, which paradoxically aligned with the conclusions of recent studies [2,22,[34][35][36]. Thus, when rocuronium was used with sugammadex, instead of cisatracurium and neostigmine, these complications decreased by up to 7%.
One of our limitations was that detection of late respiratory complications, of both pneumonia and atelectasis, was based on clinical and laboratory criteria, and it may underestimate the complication rate. We reviewed the clinical history of the patient, without performing systematic X-ray in all cases, since the patients can develop well-tolerated clinical postoperative atelectasis, which would imply unnecessary radiological exposure to all study patients. Perhaps a more effective and improved method for future research would be systematic examination with pulmonary ultrasound. It currently provides similar results to chest CT and chest X-ray for the evaluation of pneumonia and atelectasis [37].
In addition, our results were based on clinical management under real-life conditions. We have not analyzed other factors, such as mechanical ventilation parameters, recruitment maneuvers, multimodal anaesthesia techniques, opioid doses, uid therapy, duration of surgery and others that are known to contribute to increasing these complications and are probably factors that need to be assessed in subsequent studies and reviews [38,39].

Conclusion
Thus, based on the results of our study, it may be concluded that RNMB derived from the use of neuromuscular blocking agents was associated with higher risk of having both early and late POPC.
Intraoperative NMM was one of the factors that most effectively reduced the risk of all these postoperative complications and, furthermore, pharmacological reversal of NMB with sugammadex, unlike neostigmine, was