Ethical approval for this retrospective cohort study was provided by the Ethical Committee of the Hospital Italiano de Buenos Aires in Buenos Aires, Argentina (Chairperson: Dr. Augusto Pérez, Ethical Committee N° 2832) on June 10, 2016. This study was conducted at the Italiano de Buenos Aires Hospital, a tertiary referral university hospital in Buenos Aires, Argentina. A medical record review of 240 consecutive patients who had undergone elective surgical procedures requiring NMB was performed. Scheduled recovery in the intensive care unit was the only exclusion criterion. Acceleromyography (AMG) in the PACU was routine clinical practice during 2015 as part of a clinical audit to improve patient care. Cases between June 2015 and December 2015 were reviewed. The jurisdictional Institutional Review Board that approved the study waived the need for informed consent from the patients due to the retrospective nature of the study.
The patient variables recorded included sex, age, the American Society of Anesthesiologists (ASA) physical status classification score, weight, height, and body mass index. The surgery-related data assessment included duration of surgery, NMBD used (compound, total dose, and time from last dose to TOFR in the PACU), antagonist for NMBD (sugammadex or neostigmine), and the use of quantitative neuromuscular function monitoring. All data were obtained via electronic medical records.
Two different types of NMB monitoring were used due to limited availability of monitors. Intra-operative TOFR measurements were determined via a Philips IntelliVue NMT Module 865383 (Phillips Healthcare, Amsterdam, the Netherlands), attached to multiparameter monitors in every operating room as it is not portable. A different monitor, TOF-Watch® SX (Organon Ireland Ltd., a division of MSD, Dublin, Ireland), was available as a portable device for PACU measurements. Although both monitors use the same measuring principle (AMG), there is no published evidence on the similarity of measurements between them. No qualitative monitors were used during the study. The site of monitoring as well as the results of quantitative monitoring (depth of block and TOFR prior to tracheal extubation) were not manually recorded nor automatically transferred to the electronic anaesthesia record. The retrospective nature of the study precluded possible methods for obtaining this information. Results from TOFRs before tracheal extubation were verbally communicated to the physician performing TOF monitoring in the PACU in cases in which NMB monitoring or reversal agents had been used. Results from TOFRs in the PACU were added manually in the postanaesthetic subsection of the electronic anaesthesia record.
Within 5 minutes of arriving in the PACU, all patients were monitored via AMG of the adductor pollicis muscle (TOF-Watch® SX). TOFR measurements were performed by anaesthesia trainees who had previous training in NMB monitoring. Surface electrodes were placed 3 cm apart over the ulnar nerve as routine practice. The skin was prepared using alcohol cleansing pads to decrease resistance to stimulation. Uncalibrated TOF stimulation was applied and free movement of the thumb was ensured in the monitored extremity. Electrode placement for TOF monitoring was standardised, and it was not manually recorded. Nonetheless, it is highly likely that NMB monitors were applied to the thumb opposite existing peripheral venous or arterial catheters in both the operating room and the PACU.
The stimulation current was set at 50 mA (4 pulses of 0.2 ms duration at a frequency of 2 Hz) and 3 consecutive stimuli were applied to each patient, separated by 15 seconds. Baseline settings of the TOF Watch SX monitor include a stimulating current of 50 mA. The mean of the 3 resulting TOFRs was used for decision making. An NMBD antagonist (sugammadex or neostigmine) was administered if the TOFR was < 0.9.
The primary outcome was the presence of RNMB, defined as a TOFR of < 0.9. The secondary outcomes were the associations between RNMB and potentially related variables as per a multivariable regression model.
Statistical analysis
We assumed a 45% incidence of RNMB13 and calculated a required sample size of 222 patient records. This sample size allowed an incidence estimation with a 95% confidence interval (CI) margin of error that did not exceed 7%. Eventually, 240 electronic medical records were analysed after considering a missing data rate of 10%.
Descriptive analyses were performed using the mean ± standard deviation for continuous variables and numbers and percentages for categorical variables. Qualitative variables derived from each group were compared using the chi-square test or Fisher’s exact test in cases involving low expected counts. The Student’s t-test was used to analyse normally distributed quantitative data, while the nonparametric Wilcoxon rank-sum test was used to analyse non-normally distributed quantitative data.
In the first step, we evaluated the association between the different potentially predictive variables and the outcome (RNMB) through a bivariate analysis. Since we had a total of 58 outcomes, we predicted that we could build a multivariable model with a maximum of 5 potential predictors as recommended by Norman et al.14
The multivariable model was built using the 4 variables (use of monitoring, type of reversal agent, duration of surgery, and the time between the last administration of NMBD and the presence of a TOFR in the PACU) that had shown a statistically significant association (P < 0.1) in the bivariate analysis. In addition, we forced the entry of the NMBD type as a variable for clinical reasons.
Once the model was built, we tested whether the elimination of those variables with no statistical significance (P > 0.05) modified the odds ratio (OR) of the other variables in the model in a substantial way (more than 10%). For this reason, the duration of surgery was retained as a variable in the model.
The final model was established using the following variables: use of NMB monitoring, type of reversal agent used, duration of surgery, time between the last administration of NMBD and TOFR in the PACU, and type of NMBD used.