Acceleromyography is the most commonly used quantitative neuromuscular monitoring method.  Current guidelines recommend calibration of acceleromyographic devices for scientific purposes by setting monitor gain and defining supramaximal current.  Calibration is rarely performed in clinical practice. Without calibration, the gain of the monitor may be set too low, and one or more responses to train-of-four stimulation may be outside the measurement window and will not be displayed, underestimating recovery.  Even in calibrated devices, however, acceleromyography may report baseline train-of-four ratios as high as 130 to 140%; therefore, comparison of final train-of-four ratio to the baseline train-of-four ratio (normalization) is also recommended.  Baseline calibration generally is performed after anesthesia induction but before the administration of neuromuscular blocking agents. In a rapid sequence induction and intubation setting, or during studies that involve awake patients, clinicians and investigators may opt to forgo calibration, to shorten the time that patients are at risk for pulmonary aspiration, or to avoid additional patient stress and discomfort. We investigated the validity and tolerability of calibration of the TOF-Watch SX monitor (Organon, Ireland) in awake patients compared to calibration in anesthetized patients.

Thirty-four consenting patients with American Society of Anesthesiologists (Schaumburg, Illinois) Physical Status I or II, aged 19 to 64 yr, undergoing elective surgery, provided written informed consent and were enrolled in this two-center (Geneva University Hospitals, Geneva, Switzerland; and Lugano Regional Hospital, Lugano, Switzerland) prospective study following good clinical research practice guidelines.  Ethics committee approval and registration (; NCT 03649672) were obtained before patient enrollment. All patients received intravenous sufentanil 0.2 µg · kg–1 or fentanyl 2 µg · kg–1 as premedication within 3 to 5 min of induction. The TOF-Watch SX monitors were attached to both arms using a hand adapter (Hand Adapter; Organon, Ireland); data were stored on two interfaced laptops. Awake calibration (Calibration 2 mode (CAL 2)) was first performed on a randomly assigned patients’ dominant or nondominant arm, and pain associated with neurostimulation was rated on a numeric rating scale (range: 0 to 10). After induction of anesthesia with IV propofol and loss of consciousness, calibration was performed on the opposite arm (anesthetized calibration). Continuous train-of-four stimulation was then started simultaneously on the two devices, and intravenous rocuronium (0.6 mg · kg−1) was administered for tracheal intubation.

The primary outcome was agreement between the two monitors in rocuronium total duration of block (time in minutes from injection of rocuronium to spontaneous return to a normalized train-of-four ratio 0.9 or greater). Secondary outcomes included pain scores during awake calibration, non-normalized baseline train-of-four ratio, total duration (non-normalized) and duration of deep block, time to recovery of train-of-four ratio to 25%, 50% and 75% of non-normalized baseline, and supramaximal current amplitude. Normalization was performed by dividing train-of-four ratio at recovery by the control train-of-four ratio obtained before administration of rocuronium.  We determined the mean difference between the time to normalized train-of-four ratio of 90% obtained with the TOF-Watch SX monitor using awake and anesthetized calibration including 95% CI.

The differences as a measure of bias were tested using the paired-samples t test. The agreement between the monitors was analyzed as the intraclass correlation coefficient with 95% CI. Finally, we constructed a Bland–Altman plot including limits of agreement to assess the difference between methods over the range of average values. 

The primary outcome was analyzed in 33 patients. Mean ± SD total recovery time was 50.7 ± 14.0 min for awake calibration and 51.0 ± 14.0 min for anesthetized calibration, with no significant difference between measurements (–0.4 min, 95% CI, –1.9 to –1.1 min, P = 0.624). The intraclass correlation between the measurements was 0.96 (95% CI, 0.92 to 0.98). Figure 1 shows the Bland–Altman plot of the mean values of both measurements (x-axis) against the difference (y-axis). Limits of agreement were –8.3 min (95% CI, –10.9 to –5.7 min) and 7.6 min (95% CI, 4.9 to 10.2 min). The mean ± SD numeric rating scale pain score during awake calibration was 3.2 ± 1.9. There was no difference in other recovery times (table 1).

Table 1.

Primary and Secondary Outcomes in Patients Undergoing Awake and Anesthetized Calibration

Primary and Secondary Outcomes in Patients Undergoing Awake and Anesthetized Calibration
Fig. 1.
A Bland–Altman plot of the differences between the times to a normalized train-of-four ratio of greater than or equal to 0.9 measured with awake calibration and with anesthetized (asleep) calibration (ordinate) versus the mean of the two measurements (abscissa). The solid line represents the bias, and the dashed lines represent the upper and lower limits of agreement.

A Bland–Altman plot of the differences between the times to a normalized train-of-four ratio of greater than or equal to 0.9 measured with awake calibration and with anesthetized (asleep) calibration (ordinate) versus the mean of the two measurements (abscissa). The solid line represents the bias, and the dashed lines represent the upper and lower limits of agreement.

There was agreement between the TOF-Watch SX monitor calibrations performed before (in awake patients) and after anesthesia induction. Awake calibration was well tolerated and might be considered before rapid sequence induction in both clinical and research settings. However, in the absence of opioid premedication, pain scores may be higher,7–9  and hand involuntary movements during calibration may create motion artifacts.10  The close agreement in the time to recovery from a single dose of rocuronium between the two groups suggests that meticulous care in applying the monitors to awake patients is important. In the clinical setting where awake calibration is necessary, obtaining consistent acceleromyographic responses in the awake patient is possible.