To the Editor
We read with great interest the recently published randomized controlled trial describing the impact on anesthetic agent consumption after autonomic neural blockade (ANB) as a part of combined anesthesia protocol in laparoscopic sleeve gastrectomy (LSG). We commend Daes et al for studying the effect of the timing of the novel ANB, at the onset of LSG (early ANB) compared to performing the ANB at the end (late ANB). They reported a significant reduction in intraoperative remifentanil consumption when ANB was performed at LSG onset.
The sample size was calculated to detect a between-group difference of 200 µg (target effect estimate) in total remifentanil dose. The primary outcome during the study had been changed to remifentanil in µg/kg/min. However, the target effect estimate in the sample size calculation was not normalized to body weight and time. Although this change occurred during the course of the study, the sample size should have been recalculated for a minimum clinically important difference (initially defined as 100 µg total) in terms of µg/kg/min. In addition, in the data collection section, sevoflurane consumption was stated as a co-primary outcome but did not receive an a priori power/sample size calculation.
In the statistical analysis section, it is mentioned that the P values for multiple comparisons (of the primary outcomes) were Bonferroni adjusted, setting the significance threshold at P < .025. If this were the case, the sample size calculation should have been performed for an alpha of 0.025 for each primary outcome.
The strategy to modify anesthetic doses was primarily based on hemodynamics, and the results presented in Table 4 show a statistically significant difference in hemodynamics between the groups. This difference could be due to inadequate or overtitration of intraoperative remifentanil and sevoflurane in one of the groups. This reflects a potential failure of the anesthesia protocol. The independent investigator who collected the data was also responsible for ensuring that the opioid and sevoflurane protocols were followed. However, one of the 2 treating anesthesiologists administered more sevoflurane than remifentanil, as stated in the study. This indicates a failure of the anesthesia protocol. The protocol stated that a minimum alveolar concentration (MAC) concentration of 0.4 was used as the lower end for sevoflurane (which can place the patient at risk for awareness) was used as the lower end for sevoflurane, while in the strategy to modify anesthetic doses section, the lower end was mentioned as a MAC of 0.6. Although the bispectral index was monitored, the proportion of intraoperative time spent in the target range of 40 to 60 has not been presented. Had the titration of sevoflurane and remifentanil to hemodynamics adhered to the protocol, there should have been no significant difference in the hemodynamics.
The authors, however, choose to conclude that their intervention significantly stabilizes hemodynamics solely based on P values. The magnitude of these hemodynamic changes was very minor (a change in MAP by 8 mm Hg from 89 mm Hg or 0.5 mm Hg from 87 mm Hg is mostly clinically inconsequential). Interestingly, the difference from baseline appeared to be lower in the late ANB group, suggesting less hemodynamic fluctuation in this group. The hemodynamics could have been affected by other factors such as prior medication (considering that there were 10% and 17% hypertensives in each of the groups) and blood loss, which appear to have not been measured in the study.
The authors mentioned double blinding, but it is unclear who was blinded. Considering that ANB was performed intraoperatively, it is most likely that the patients were blinded, but it is less likely that the surgeon, the treating anesthesiologist, or the person collecting the opioid and sevoflurane consumption was blinded. The independent investigator collecting the study data was not present before the division of the greater omentum, and after the leak test, data were collected during these periods by the treating anesthesiologist, which potentially breaches the blinding. It was stated that the data manager had generated and had access to the randomization list throughout the study. These statements suggest a bias concerning randomization and blinding.
An examination of Table 1 shows a failure of randomization as the standardized differences were quite large for sex, body mass index, and hypertension. These factors are known to affect opioid requirements (and hemodynamics) and providing adjusted estimates would have been meaningful.
In conclusion, the authors claimed a significant reduction in opioid consumption with early administration of ANB. Although the P value for difference in remifentanil consumption is very small P < .001 the difference between groups is 0.04 µg/kg/min (95% confidence interval [CI], 0.02–0.06 µg/kg/min). This was less than the minimum rate of administration in the trial protocol (0.1 µg/kg/min). The between-group mean difference for the total dose comes to 182.9 (95% CI, 16.79–349) µg. The lower end of this CI suggests that the use of early ANB may produce only a minor difference in intraoperative remifentanil dose requirement. This article focuses on statistically significant P values, while the clinical significance of the effect estimate has been ignored.
Although the title of the article reads “Impact on Anesthetic Agent Consumption,” the article focuses mainly on remifentanil consumption, and sevoflurane takes much less space, probably because of P > .05. There appears to be a typo/mismatch in the values presented in the results section of the sevoflurane consumption (mean difference +0.17 mL/min, 95% CI, 0.003–0; P =.088). This should be a difference of 0.01 mL/min (95% CI, −0.010 to 0.03, P = .326 for a t test) based on data in Table 2.
Considering the expertise required to perform the block, the ambiguity around the results (hemodynamics, dosing protocol/practices, baseline differences, and data presentation), and the very small difference in the dose of this short-acting opioid, it would be difficult to conclude that early ANB produces a meaningful reduction in intraoperative opioid consumption or anesthetic agent consumption.
A few important questions arise from this treatment protocol: Does blocking the autonomic response to pain mean less intraoperative and postoperative pain? Does this warrant a reduction in intraoperative analgesic dose? Could the same hemodynamic stability (a lower heart rate/MAP) be replicated using beta-blockers, or does the analgesic effect of this ANB work mainly through a sensory block (essentially a visceral nerve block)?