Respiratory muscle weakness and atelectasis secondary to abdominal surgery are major factors contributing to the development of postoperative pulmonary complications (PPCs).1 Residual neuromuscular block may increase the risk of PPCs.2,3,4 Reversal with sugammadex is associated with a lower incidence of residual paralysis at postanesthesia care unit (PACU) admission than reversal with neostigmine.5 Moreover, recent studies suggest that electromyographic activity of the diaphragm may be impaired after the use of neostigmine compared with sugammadex.6

Our hypothesis was that reversal of neuromuscular block with sugammadex would result in less postoperative pulmonary dysfunction than reversal with neostigmine. The primary objective of this study was to compare the change in forced vital capacity (FVC) one hour after reversal with neostigmine or sugammadex among patients undergoing major abdominal surgery. A secondary objective was to compare area of atelectasis size as assessed by planimetry in postoperative lung ultrasound images.

Methods

We carried out a randomized-controlled trial at Hospital Universitario La Princesa, a tertiary care university hospital in Madrid, Spain. The study was approved by the Human Research Ethics Committee (Chairperson Prof. Francisco Abad) on 19 November, 2014 and registered in the EU clinical trials register (EudraCT: 2014‐005156‐26). The trial was conducted from July 2015 to July 2016.

All patients scheduled for major abdominal surgery (liver resection, pancreatectomy, gastrectomy, or any type of colectomy) were included. Use of postoperative epidural analgesia was also an inclusion criterion. Exclusion criteria included: refusal to participate, admission to postoperative recovery unit under mechanical ventilation, hypersensitivity to any of the drugs, severe asthma or mild asthma under treatment, myocardial infarction or coronary occlusion three months prior to surgery, myasthenia gravis, emergency surgery, pulmonary fibrosis, or very severe chronic obstructive lung disease (GOLD IV). Informed consent was obtained by the residents on duty, the night before surgery.

Study interventions

Before entering the operating room, basal spirometry (KoKo® Legend, nSpireTM) and lung ultrasound (LUS) (Sonosite M Turbo with a P21x Phased Array Probe, Sonosite, Bothell, WA, USA) were performed by an independent researcher. Spirometry was performed following international recommendations7 but with patients in a semi recumbent position (40º) as recommended for postoperative position. Forced vital capacity, forced expiratory volume in the first second (FEV1), and forced expiratory flow 25–75% were measured. Changes in lung aeration were studied by LUS with the patients in the same position. Sagittal sections were performed at three areas in each lung: ventral, medial, and posterior, corresponding to three predefined locations (parasternal, medial axillary line, posterior axillary line).8 One to two complete respiratory cycles were recorded in each location for offline analysis. Collapsed areas were defined by sonography as the presence of a condensation “tissue-like” (hepatization) ultrasound pattern. For the offline analysis, a single frame corresponding to end expiration was selected from the video file. Brightness was adjusted setting the brightest level in the bony surface of ribs and the brightness level in the acoustic shadow behind ribs. After this adjustment, the scale in cm2 was set according to image resolution in pixels. Collapsed areas were then outlined and measured by planimetry. The sum of the six explored lung areas was considered the total collapsed area for the study. We used the software ImageJ (ImageJ, U. S. National Institutes of Health, Bethesda, MD, USA) for LUS image analysis.9

Combined epidural and general anesthesia were carried out by anesthesiologists as routine clinical practice. Neuromuscular block was performed with rocuronium and monitored by train-of-four (TOF) kinetomyography with Datex-Ohmeda MechanoSensor™. Anesthesiologists were free to maintain TOF level according to usual criteria. After the surgical resection was completed, patients were randomly assigned to receive either sugammadex 4 mg·kg−1 or neostigmine 40 µg·kg−1 in combination with atropine 10 µg·kg−1 using sealed opaque envelopes. An extra dose of reversal was permitted by protocol, when needed, at the discretion of the attending anesthesiologist. Unrestricted blocked randomization was previously carried out by an independent contract research organization with the program M.A.S (sampling and scheduled randomizations) Glaxo SmithKline version 2.1.

Reversal medications were unblinded to the anesthesiologist in the operating room but blinded for the patient and the researcher testing pulmonary function. Separated case report forms and databases were used for clinical and pulmonary data.

One hour after the patient was extubated, new spirometry and lung ultrasound explorations were performed in the postoperative recovery room with the same procedure and position used in the preoperative determinations. Postoperative clinical management was conducted according to clinical preferences. The day after surgery, prior to ward discharge from the postoperative recovery unit, or 24 hr after extubation, additional pulmonary tests were performed.

Outcomes and data collection

Clinical data

Patient variables including age, height, weight, American Society of Anesthesiologists score, history of chronic obstructive pulmonary disease, congestive heart failure, basal peripheral capillary oxygen saturation (SpO2), smoking, weight loss > 10%, functional class, and respiratory infection in the last month were collected at baseline.

Surgical variables such as duration of surgery (skin-to-skin), open vs laparoscopic approach, type of surgery, and anesthetic variables including tidal volume, positive end-expiratory pressure (PEEP), fraction of inspired oxygen (FiO2), need for alveolar recruitment (under anesthesiologist criteria), epidural analgesia, and amount of fluids administered were recorded during surgery.

Characteristics of the neuromuscular block including neuromuscular blocking drug, depth of block prior to reversal, reversal drug used, and last TOF percentage registered immediately before awakening were noted. Oxygenation was assessed by the pO2/FiO2 ratio at the first hour after surgery and before discharge while SpO2 with ambient air was documented after spirometry. Hospital length of stay, Clavien–Dindo classification for postoperative complications, pneumonia, need for mechanical ventilation (invasive or non-invasive), and death were documented.

Variables confounding assessment of FVC such as body mass index, ARISCAT scale for postoperative risk of pulmonary complications, basal FVC and FEV1, residual neuromuscular block, and postoperative pain visual analogue scale (VAS) before and after spirometry were recorded.

Outcomes

Difference in reduction of FVC in the first hour between groups was the primary outcome. Other spirometry values were analyzed one hour and 24 hr after surgery. Lung atelectasis size was measured by planimetry on ultrasound images as a secondary outcome. Differences in the incidence of hypoxemia (defined as SpO2 ≤ 92% with FiO2 21% after spirometry, or partial pressure of oxygen [pO2]/FiO2 < 300 with an FiO2 28–31%) were also included in the secondary analysis. Incidence of nausea and vomiting in both groups was included as an exploratory analysis.

Statistical analysis

Sample size

In a previous unpublished study developed in our institution (2013) we found a mean loss of 18.1% and a standard deviation of 14.2% for the difference between basal FVC and FVC 30 min after surgery. We could not find data of previous studies in these settings reporting the clinical relevance of changes in FVC. For these reasons we estimated this relevance based on studies which analyzed changes in FVC and FEV1 after treatment with bronchodilators.10,11 According to these studies, we considered as relevant a difference of 7% between groups. For a power of 80% with an α level (probability of a type I error) of 5%, the sample size needed was 64 patients per group.

Variable analysis

Continuous variables were described by their measures of central tendency (mean or median) and dispersion (standard deviation or interquartile range). Homoscedasticity was tested with Levene’s test and normality with Shapiro–Wilk test. Differences in confounding factors between groups were analyzed with t-test in case of quantitative variables and χ2 test for categorical variables. Corresponding non-parametric tests were used when needed. Differences in mean changes from baseline were analyzed with repeated measures analysis of variance (ANOVA) with treatment as a principal factor. Bonferroni post hoc test was used when P ≤ 0.05 in ANOVA. If significant deviations from normality or the assumption of homoscedasticity were observed, a nonparametric test was used.

Results

One hundred and thirty patients were randomized 1:1 to receive either neostigmine 40 µg·kg−1 + atropine 10 µg·kg−1 or sugammadex 4 mg·kg−1. One patient was excluded from analysis in the neostigmine group because of probable recurarization. This patient had one count in the TOF at reversal and four counts with a TOF ratio of 96% immediately after extubation. Twenty minutes after arrival to PACU he suffered dyspnea, severe hypoxemia, and muscular weakness, which was treated with 200 mg of sugammadex without monitoring muscular status. He was excluded from analysis because of protocol violation. In the sugammadex group one patient refused to participate before the one-hour spirometry, another patient could not have an adequate spirometry because of postoperative pain, and in one patient spirometry was not performed on time because of spirometer failure. Figure 1 shows the study flow chart.

Fig. 1
figure 1

Consort statement

Full size image

Baseline and clinical characteristics

Demographic and clinical baseline characteristics were similar in both groups (Table 1). Surgical and anesthetic management are described in Table 2. At the moment of neuromuscular block reversal, the proportion of patients with deep neuromuscular block (TOF count equal to 0) was three of 64 (5%) in the neostigmine group compared with 13 of 62 (21%) in the sugammadex group (P = 0.007). Nevertheless, only 19 of 62 (31%) of those reversed with sugammadex had TOF < 90% measured immediately after extubation compared with 45 of 64 (71%) of those in the neostigmine group (P < 0.001).

Table 1 Characteristics of participants
Full size table
Table 2 Surgical and anesthetic management
Full size table

Spirometry outcomes

Values of FVC at baseline, one, and 24 hr following surgery, expressed in litres, are shown in Table 3. The decrease in FVC one hour after surgery, expressed as a percent of baseline, was 16% (95% CI, 12 to 20) in the neostigmine group and 13% (95% CI, 9 to 18) in the sugammadex group (P = 0.40). The day after surgery, decreases in FVC, expressed as a percent of baseline, were 24 (95% CI, 20 to 29)% and 24 (95% CI, 20 to 29)%, respectively (P = 0.94). Therefore, decreases in FVC one hour and one day after surgery were similar in both groups. FEV1 values in litres (not shown) decreased proportionally to FVC; relative differences were 15% one hour (95% CI, 11 to 18) and 25% (95% CI, 21 to 29) one day after surgery. Again, there were no differences were between groups.

Table 3 Forced vital capacity at baseline, and one hour and 24 hr following surgery
Full size table

Atelectasis on lung ultrasound

Examples of ultrasound technique and measurements are shown in Figs 2-4. Three patients were excluded from the lung ultrasound analysis after one hour in the neostigmine group because of basal abnormalities (consolidation or pleural effusion). Forty-two patients had visible new consolidations in the first hour (Table 4). Twenty-four of them (39%) were in the neostigmine group and 18 (30%) were in the sugammadex group. Median atelectasis size was 9.7 cm2 in those patients who received neostigmine and 6.7 cm2 when sugammadex was employed for neuromuscular block reversal (Table 4).

Fig. 2
figure 2

Probe located in last intercostal space, posterior axillary line in a sagittal orientation

Full size image
Fig. 3
figure 3

Arrow = border of the rib with acoustic shadowing beneath it. Asterisk = lung consolidation

Full size image
Fig. 4
figure 4

Arrow = border of the rib with acoustic shadowing beneath it. Asterisk = lung consolidation. Consolidation area is outline using manual tracing tools in ImageJ software. Area is calculated after adjusting scale with the same software

Full size image
Table 4 Atelectasis on lung ultrasound
Full size table

The number of patients with lung consolidations at the first day after surgery increased to 43 (74%) in the neostigmine group and to 39 (66%) in the sugammadex group. Median atelectasis size was 17.0 cm2 and 13.6 cm2, respectively. No statistically significant differences were found during the early postoperative period.

Postoperative complications

Incidence of postoperative hypoxemia (SpO2 ≤ 92), measured without oxygen after a spontaneous recruitment maneuver (spirometry) or with oxygen when pO2/FiO2 rate was lower than 300, was similar in both groups (Table 5). The day after surgery, one (1%) patient in the neostigmine group and 10 (16%) patients in the sugammadex group had a SpO2 with ambient air after spirometry of less than 92%. There were no differences between groups in respiratory complications, other postoperative complications, need for mechanical ventilation, or death (Table 5).

Table 5 Postoperative events
Full size table

Discussion

We found no differences in postoperative spirometry at one and 24 hr after surgery in patients who received sugammadex or neostigmine to reverse neuromuscular block. FVC decreased by approximately 15% compared with baseline one hour postoperatively in both study groups, an amount lower than expected for upper abdominal surgery.4,12,13 Previous data describing FVC loss after surgery are heterogeneous and were published many years ago. After upper abdominal surgery, a loss of at least 30% of FVC is expected.14 We believe that good postoperative analgesia and patient cooperation was responsible for the reported preservation of FVC.15

Atelectasis may occur in up to 90% of anesthetized patients16 and is suspected when oxygenation is impaired after surgery. The rate of postoperative atelectasis, especially one hour after general anesthesia is not well documented. In our population, incidence of atelectasis measured by LUS one hour after surgery was lower than expected. Intraoperative low tidal volume (Vt) ventilation with PEEP may reduce the incidence of atelectasis and pulmonary complications.17,18,19 In our patients, the mean Vt was 8 mL·kg−1 ideal body weight with very low dispersion and PEEP above 4 cmH2O was applied in every patient. Lung ultrasound is a novel and accurate technique for postoperative lung collapse diagnosis,20 but may be less sensitive than computed tomography.

Although there is not a unanimous definition for PPCs, clinically relevant pulmonary complications after elective major abdominal surgery are uncommon.1 Such was the case in our study with only four patients requiring mechanical ventilation and two developing pneumonia. Our study focused on process markers like changes in FVC and atelectasis that might be a starting point for clinically relevant pulmonary complications. Surprisingly, both FVC loss and incidence and size of atelectasis were higher the day after surgery. Protective mechanical ventilation may have influenced the benefit found in the first hour. Lack of early mobilization, positive fluid balance, or inconsistent chest physiotherapy may have increased atelectasis as time passed.21 It is unlikely to be related to the agent used for neuromuscular block reversal.

Residual neuromuscular paralysis is a frequent complication of general anesthesia with an incidence of up to 30%3,22 reported in the literature. Our study, despite the open use of quantitative neuromuscular monitoring, was no different. Despite protocolized monitoring of neuromuscular block, and the possibility of using rescue doses of neostigmine, none of the clinical researchers used more than 40 μg·kg−1. More than two thirds of patients in the neostigmine group and nearly one third of those in the sugammadex group had a TOF ratio ≤ 90% immediately after extubation. Although we determined TOF ratio immediately after extubation instead of in the PACU, as commonly done,23 the proportion of patients with an inadequate reversal was high. Previous research showed an association between residual neuromuscular block and PPCs.3 Our study was not designed to evaluate this association.

A recently published multicentre prospective observational study including 22,803 patients did not show any advantage of sugammadex over neostigmine with respect to preventing pulmonary complications.24 Moreover, extubation with a TOF ratio > 90% was not associated with a lower risk of postoperative respiratory complications. Authors concluded that using neuromuscular agents during anesthesia is associated with an increased risk of postoperative pulmonary complications irrespective of dose. Diaphragm impairment after surgery, immobilization, and inflammatory response may have a more relevant influence in postoperative pulmonary outcomes than reversal agents.

There are limitations to our study. Use of recruitment maneuvers was not controlled and 27 (44%) patients in the sugammadex group received this intervention during surgery compared with 15 (23%) in the neostigmine group. The PROVHILO trial did not show differences in pulmonary outcomes when a recruitment maneuver followed by a PEEP of 12 cmH2O was performed.25 The iPROVE trial also failed to prove better oxygenation three hours after surgery or a lower rate of atelectasis during the first day with an intraoperative open lung strategy with individual PEEP assessment.26 We believe this had little impact on our results. Secondly, anesthesiologists in the operation room were not blinded to the study drug and managed neuromuscular block under their own criteria. This could have biased results because of different rocuronium dose regimens. Total rocuronium dose was 10% higher in the sugammadex group but time of surgery was also slightly higher for these patients. The rate of TOF = 0 before reversal was significantly higher in the sugammadex group. We did not record post tetanic count before reversal when TOF = 0. These patients might have been at risk of symptomatic residual neuromuscular block but with respect to our findings and the main point of our results, none of these factors influenced postoperative lung function one hour after surgery.

In conclusion, we found no differences in spirometry after reversal with sugammadex or neostigmine in a high-risk population with intraoperative lung-protective ventilation, neuromuscular block monitoring, and epidural analgesia. No difference in the incidence of atelectasis, area of atelectasis, or pulmonary complications was noted. The benefit of sugammadex might be related to the reduction of infrequent critical respiratory events during the immediate postoperative period due to severe residual neuromuscular block.