Dear Editor,

We thank Professor Bennett Andrassy et al. for their interest in our recent article “Deep Neuromuscular Block Attenuates Chronic Postsurgical Pain and Enhances Long-Term Postoperative Recovery After Spinal Surgery: A Randomized Controlled Trial” [1]. They raised a meaningful question which is of great significance to our research. So we hope to help readers better understand our study through this reply.

First of all, we must admit that the concern highlighted by Andrassy et al. is reasonable. In patients diagnosed with lumbar degenerative diseases, the comparative efficacy of lumbar decompression alone versus lumbar decompression plus fusion on postoperative pain, disability scores, and other associated outcomes is controversial and inconclusive [2,3,4,5]. The proportions of these two surgery types between the deep neuromuscular block (DNMB) and moderate neuromuscular block (MNMB) groups in our study were indeed not consistent. Although only seven patients more in the MNMB group received lumbar fusion than the DNMB group, and the difference was not statistically significant (P = 0.292), the surgery type may still serve as a confounding factor within this study.

However, we consulted a statistician, and the impact of this confounding factor should be handled by regression instead of subgroup analyses. So we conducted a logistic regression which included the intervention (DNMB vs MNMB) and surgery type (decompression alone vs decompression plus fusion) as independent variables, and the incidence of chronic postsurgical pain (CPSP) as the dependent variable. The result indicated that, after adjusting for the influence of different surgery type proportions among the two groups, the protective effect of DNMB on CPSP remained unchanged (odds ratio, 0.54; 95% CI 0.30–0.96; P = 0.035). Also the incidence of CPSP was similar among the two surgery types (odds ratio, 1.09; 95% CI 0.54–2.19; P = 0.816) (Table 1).

Table 1 Results of logistic regression
Full size table

Subgroup analyses, mentioned by Andrassy et al. in their letter, are often conducted to investigate the effect of interventions in different patient subsets. Because of false negatives induced by inadequate statistical power, subgroup analyses can mostly only provide exploratory results [6]. Nonetheless, considering the potential difference among the two surgery approaches, we also performed subgroup analyses based on Prof. Andrassy’s advice. In patients who underwent lumbar decompression plus fusion surgery, the incidence of CPSP was significantly lower in the DNMB group than in the MNMB group (26.6% vs 44.2%, P = 0.02). While in patients who underwent lumbar decompression alone, the incidence of CPSP was comparable among the groups (36.0% vs 36.8%, P = 0.95). This is a result that can be expected, because only 44 out of 209 recruited patients in our study received lumbar decompression surgery. In this context, even though the difference truly existed, it could not be found because of the inadequate statistical power. Furthermore, our result also indicated that the interaction between intervention and surgery type was insignificant (P interaction = 0.297), thus we believe that deep neuromuscular block could help alleviate the CPSP in the total population (including patients who received lumbar decompression surgery) (Table 2).

Table 2 Results of subgroup analyses
Full size table

In summary, after conducting logistic regression and subgroup analyses, we found that different surgery types exerted little impact in this study, and demonstrated the robustness of the primary result.

Once again, we deeply thank Prof. Andrassy et al. for their contribution to this discussion. Their questions and insights about our study have helped us to improve our research and make it even more meaningful.