I write to offer a comment and a suggestion on the systematic review and meta-analysis by Hussain et al. of local anesthetics used for fascial plane block.  This report compared analgesia between liposomal bupivacaine and plain bupivacaine, the primary outcome being rest pain scores over the 24-to-72 h postoperative interval expressed as an area under the curve.  Using data from 874 patients in 13 studies and adopting a Hartung-Knapp-Sidik-Johnkman random-effects model with the pain severity difference between treatments transformed as a standardized mean difference, the authors report in the abstract and the body that “No difference was observed between the two groups for area under the curve scores, with a standardized mean difference (95% CI) of –0.21 cm.h (–0.43 to 0.01; P = 0.058; I² = 48%).” The authors used excellent statistical methods and software within frequentist statistical methodology; inferences were made by null hypothesis significance testing. As emphasized by statisticians and epidemiologists observing a P value > 0.05 does not demonstrate that no effect or no difference was observed or that absence of an effect was demonstrated. Neither is the P value the probability of the null hypothesis.  Nor does the 95% CI have a 95% probability of including the true effect size.  A correct and more cautious description for the result of a P = 0.058 for this meta-analysis would be that the null hypothesis of no difference was not rejected.

Statistical approaches using Bayesian theory, methods and software can provide an alternative for doing this meta-analysis. This random-effects meta-analysis can be expressed in the parameters μ (true mean of the population standardized mean difference) and τ² (variance between studies). Using the 13 study values displayed in the article’s figure 3,  I ran a Bayesian random-effects model comparing the standardized mean difference between liposomal and plain local anesthetic using 32,000 iterations and four chains in the Markov chain Monte Carlo algorithm of the R package brms (2.21.6). The prior distributions were Normal(0, 1) for μ and HalfCauchy(0, 0.5) for τ using guidance for choosing “weakly informative” prior distributions for a Bayesian random-effects meta-analysis.  All sampling convergence characteristics of the fitted model (Gelman-Rubin convergence statistic = 1.00, large effective sample size, small Monte Carlo standard error, etc.) were satisfactory. Sensitivity analyses with other priors for both μ and τ gave very similar, almost identical estimated values and credible intervals. The model fit using the normal and half Cauchy priors was subjected to power scaling using the priorsense (1.0.1) package; there was no sensitivity of the model fit to either strengthening or weakening the prior distribution.  Estimated parameter values were reported as means, standard deviations and 95% credible intervals.

The estimated parameter value for μ (standardized mean difference) was –0.21, identical to the mean value (–0.21) reported by Hussain et al.  However, the 95% credible intervals (–0.43, –0.01) no longer crossed the zero line of no effect. With 98% probability, the reduction in rest pain scores was larger for liposomal bupivacaine than plain bupivacaine. Also, the estimated variability between studies (τ²) was 0.06 with the 95% credible intervals (0.01, 0.26).

The standardized mean difference used in meta-analysis is also known as Hedges’ g. It is a unitless, dimensionless metric. Heuristics for interpreting the absolute magnitude of effect has been offered with 0.2, 0.50, and 0.8 proposed as showing small, medium, and large effects.  With either frequentist or Bayesian models, the summary standardized mean difference was about –0.20; this is a small, clinically unimportant effect. Using the Bayesian model, the probability that the standardized mean difference has a medium sized effect of –0.5 or better is only 1%.

Bayesian methods (1) allows direct modeling of the uncertainty of differences between studies (τ²); (2) produces the full posterior distribution of the mean and variance (μ, τ²) parameters (fig. 1); (3) permits the estimation of 95% credible intervals for μ and τ² that have a 95% probability of containing the parameter values; and (4) allows hypothesis testing of the probability of interesting population values.  All statistical models and inferences, frequentist and Bayesian, are subject to the validity of the assumptions. The use of Bayesian methods does offer some additional tools for meta-analysis.

We thank Dr. Pace for selecting our publication highlighting the lack of difference between liposomal and plain bupivacaine, and presenting alternative findings that are based on the differences between the frequentist and Bayesian approaches for meta-analyses. We present in this reply a few of our thoughts on the different analyses.

First, while we have limited awareness of the use of the Bayesian method as a mainstream approach in meta-analyses, we are familiar with the nascent  Bayesian statistics that are slowly gaining popularity as an alternative analytic method.

Second, for acute pain meta-analyses comparing two different interventions where the mean of a group of data (i.e., randomized trials) is most important, a frequentist analysis may be better suited for the clinician. In contrast, a Bayesian model may be more applicable for those clinicians wishing to evaluate the probability that a certain intervention will be better for an outcome. On a broad scale, clinicians assessing the results of a single outcome (e.g., rest pain at 24 h) may be more interested in knowing whether one treatment is “better” than another treatment, rather than the “probability” of being better; this is more consistent with a frequentist approach.

Third, we would caution readers against an over-reliance on P values for interpreting clinical research. While the threshold for statistical significance does provide some insights into the results of a clinical research question, there should always be interpretation based on clinical importance and the minimal clinically important difference. For even if we utilize the Bayesian model to accomplish a statically significant 95% credible interval of –0.43 cm.h to –0.01 cm.h, this alternative result is still not clinically important, especially when the clinical importance threshold is 3.0 cm.h. Indeed, the differences between the results of Bayesian and frequentist approaches were trivial, and the effect size of liposomal bupivacaine was very small.

Finally, interpretation of research findings is contextual. Using the frequentist method, we have already demonstrated that liposomal bupivacaine is not superior to plain bupivacaine when used in perineural nerve blocks, periarticular infiltration and surgical field infiltration.  Our current findings in the setting of fascial plane blocks are consistent, and the alternative Bayesian analysis does not alter this finding in a clinically meaningful way.

The recent article by Mashour et al. on consciousness and the dying brain caught my eye, as I have had such a personal experience.

Once I suffered from gastro-esophageal reflux after eating spicy meals. I occasionally experienced bile reflux during my sleep, would wake up coughing, and observed the pillow next to my face was wet with green gastric fluid. One night, the same happened again, and I developed total laryngospasm. I could not cough or breathe. I got out of bed and stood facing the bed, leaning onto it with my hands. I was vividly awake and realized I was in a serious situation with total laryngospasm. Soon, my feelings of panic soothingly switched into feelings of indescribable calmness and pleasure. It was an experience unmatched by anything in my life before. I felt peace, calmness, contentment, satisfaction, happiness, and joy, all to a high degree for which there are no adequate superlative adjectives in the English language. The experience lasted a minute. Then, everything gradually became dark for me and ended.

My wife had woken up, too, and looked at me, bewildered. She later described how I slowly lapsed forward facedown onto the bed. I lay still for a minute, then slowly started making respiratory movements and little coughs. The coughs became more energetic, and finally, I stood up, coughing vigorously.

I was fine once my lungs were cleared out of the gastric bile. I was shocked to have nearly died. Thirty years later, my memories of that intriguing experience are clear. My recollections of the extremely pleasurable happiness I experienced before losing consciousness are vivid. Additionally, I was highly motivated to address all my health factors and never again had a near-death experience. I never aspirated bile again. I fear not my final death later in my life if that one-time hypoxic cerebral pleasurable experience is to be repeated. It felt divine.

As an experienced anesthesiologist with extensive hands-on experience with laryngeal masks and teaching novices, I have seen total laryngospasms in many anesthesia patients. I have often observed that hypoxia finally resolves the total laryngospasm with the development of total flaccidity. Of course, we always provided patient airway support and reoxygenation immediately upon the laryngospasm resolving itself or with succinylcholine administration. We never saw brain damage. 

In 2017, Chawla et al. described research using electroencephalographic research in intensive care unit patients.  End-of-life electrical surges are observed in about 50% of patients without preceding diagnoses of brain death after their cessation of heartbeat. In already brain-dead diagnosed patients, no end-of-life electrical surges were ever seen after cessation of heartbeat. It is probable those end-of-life electrical surges seen on electroencephalogram correlate with emotional experiences from hypoxia-induced large surges of cerebral cortical dopamine and serotonin neurotransmitter release, causing experiences like I had with my near-death hypoxia and survived to describe.

The study by Mashour et al. that I am responding to focuses on the neurobiology of consciousness in the dying brain. The end-of-life cortical electrical surges are associated with the activation of glutamate-gated neuron channels. Brain electrical surges start within 3 to 6 min after losing measurable blood pressure. There is an “explosive surge” in global brain connectivity before cortical isoelectricity finally develops. The effects are similar to what psychostimulants can induce, like ketamine. Animal research correlates the near-death cortical electrical spike with explosive surges in released dopamine, adenosine, serotonin, and noradrenaline in the frontal and occipital cortices. Interestingly, full patient resuscitation is sometimes achievable even after the cortical isoelectric stage is reached during hypoxia.

Thank you, Mashour et al., your study review is fascinating. We can finally possibly understand the perceived out-of-body, probably hypoxic, near-end-of-life experiences that people have reported for thousands of years and vividly recall.