I read the review by Singh et al. [1] with interest. They assessed the influence of local anaesthetic wound infiltration and regional anaesthesia techniques on outcomes related to pain following caesarean delivery. It is my concern, however, that the presentation of the findings has increased the potential for confusion and misinterpretation.

I agree with Singh et al. that “… compared with placebo, the injection or infusion of local anaesthetic into the caesarean wound or around nerves that supply the wound reduced morphine dose and pain in the first 24 postoperative hours. Wide credible intervals and different orders of efficacy for different outcomes precluded us from concluding which injection site was best.” Evaluation of their network league tables for the primary outcome, morphine- equivalent consumption at 24 h, confirms that statistical differences were only present between control vs. local anaesthetic wound infiltration and regional anaesthesia modalities, and statistical differences were absent between local anaesthetic wound infiltration and the various regional anaesthesia techniques. Importantly, differences in the magnitude of effect between control vs. local anaesthetic wound infiltration and regional anaesthesia modalities are not a substitute for comparisons derived from direct and indirect estimates between local anaesthetic wound infiltration and regional anaesthesia techniques themselves. It is not obvious what the minimal clinically important difference was set at for the various outcomes, but this is a necessary prerequisite for the evaluation of certainty or quality of evidence through the confidence in network meta-analysis approach. The assessment of this involves several domains, one of which is imprecision and is judged in relation to the credible intervals and minimal clinically important difference. In the presence of few differences between interventions and serious imprecision, it is erroneous, and not robust in my opinion, to rank them in order of efficacy as this would not account for the uncertainty and increase the risk of false conclusions. Unfortunately, this might represent exactly what has been performed, with ilio-inguinal block ranked first and ilio-inguinal-iliohypogastric block ranked second for the reduction of morphine equivalent consumption at 24 h. Further, other rankings of interventions for particular outcomes have been prominently presented. In the network league tables for the static and dynamic pain score at 4–6 h, static pain score at 24 h and time to rescue analgesia, few differences were present and the ranking of interventions was, in my view, inappropriate. The results of a re-analysis that examined the extent of uncertainty in interventions from previous network meta-analyses provide some insights [2]. In 90% and 71% of comparisons in these 58 network meta-analyses, no differences were revealed between the best ranked intervention and second and third ranked intervention, respectively, and in 39 network meta-analyses with six or more interventions, the median probability that one of the top two interventions was in fact among the bottom two ranked interventions was 35%.

What should the conclusions of this systematic review be? If the minimal clinically important difference was to be set at 10 mg for morphine equivalent consumption at 24 h, local anaesthetic wound single shot injection and infusion as well as ilio-inguinal-iliohypogastric, quadratus lumborum and transversus abdominis plane block are clinically superior to no intervention. Ilio-inguinal-iliohypogastric block may be a new kid on the block in the armamentarium of the anaesthetist for analgesia subsequent to caesarean section. The evidence would be insufficient to reach firm conclusions with regard to erector spinae plane blocks or transversalis fascia blocks vs. no intervention. Moreover, with the use of this same minimal clinically important difference, we are able to conclude that local anaesthetic wound single shot injection and infusion and quadratus lumborum and transversus abdominis plane blocks are all clinically equivalent to each other. The credible intervals for all remaining comparisons of local anaesthetic wound infiltration and regional anaesthesia techniques are wide and, despite the lack of statistical differences, clinically important differences might be present but still hidden with the strength of evidence we have at this time. In previous meta-analyses, most of the desirable effects of local anaesthetic wound infiltration and regional anaesthesia techniques vs. control were lost in the presence of long-acting intrathecal opioid [3, 4], and this relationship was not fully analysed by Singh et al [1], precluding further conclusions.

Network meta-analyses are understandably complex, and summarising the relative effects of multiple interventions can be challenging with the number of pairwise comparisons involved as the number of interventions increases. I would, therefore, encourage the increased reporting and utilisation of minimal clinically important differences for all systematic reviews to facilitate interpretation, even though the evidence for its selection is limited [5], and welcome the publication of results related to the imprecision in this particular systematic review.

我阅读了 Singh 等人的评论。[ 1 ] 感兴趣。他们评估了局部麻醉伤口浸润和区域麻醉技术对剖宫产后疼痛相关结果的影响。然而，我担心的是，研究结果的呈现增加了混淆和误解的可能性。

我同意 Singh 等人的观点。“……与安慰剂相比，在剖腹产伤口或供应伤口的神经周围注射或输注局部麻醉剂可减少术后前 24 小时的吗啡剂量和疼痛。广泛的可信区间和不同结果的不同疗效顺序使我们无法得出最佳注射部位的结论。”对主要结果（24 小时吗啡当量消耗量）的网络排名表的评估证实，仅在对照组与局部麻醉伤口浸润和区域麻醉方式之间存在统计学差异，而局部麻醉伤口浸润和局部麻醉之间不存在统计学差异。各种区域麻醉技术。重要的是，控制与局部麻醉伤口浸润和区域麻醉方式之间的影响大小差异不能替代局部麻醉伤口浸润和区域麻醉技术本身之间的直接和间接估计的比较。对各种结果设定的最小临床重要差异是什么尚不清楚，但这是通过网络荟萃分析方法的信心评估证据的确定性或质量的必要先决条件。对此的评估涉及多个领域，其中之一是不精确的，并根据可信区间和最小的临床重要差异进行判断。在干预措施之间几乎没有差异和严重不精确的情况下，在我看来，按疗效顺序对它们进行排名是错误的，而且不可靠，因为这不会考虑不确定性并增加错误结论的风险。不幸的是，这可能正好代表了已执行的操作，髂腹股沟阻滞排名第一，髂腹股沟-髂腹下阻滞排名第二，以减少 24 小时吗啡当量消耗。更远，突出显示了针对特定结果的其他干预措施排名。在 4-6 小时的静态和动态疼痛评分、24 小时的静态疼痛评分和抢救镇痛时间的网络排行榜中，几乎没有差异，在我看来，干预措施的排名是不合适的。一项重新分析的结果检查了先前网络荟萃分析中干预措施的不确定性程度，提供了一些见解[2 ]。在这 58 项网络荟萃分析的 90% 和 71% 的比较中，排名最高的干预措施与排名第二和第三的干预措施之间分别没有差异，在 39 项包含六项或更多干预措施的网络荟萃分析中，中位概率前两项干预措施之一实际上是排名倒数的两项干预措施之一，这一比例为 35%。

这个系统评价的结论应该是什么？如果将 24 小时吗啡当量消耗量的最小临床重要差异设置为 10 mg，则局部麻醉伤口单次注射和输注以及髂-腹股沟-髂腹下、腰方肌和腹横肌平面阻滞在临床上优于无干涉。髂腹股沟-髂腹下阻滞可能是剖宫产术后镇痛麻醉师设备中的一个新手。证据不足以得出关于竖脊肌平面阻滞或横筋膜阻滞与不干预的确切结论。此外，利用同样的最小临床重要差异，我们可以得出结论，局部麻醉伤口单次注射和输液与腰方肌和腹横肌平面阻滞在临床上都是相互等效的。所有剩余的局部麻醉伤口浸润和区域麻醉技术比较的可信区间很宽，尽管缺乏统计学差异，但可能存在临床上重要的差异，但仍然隐藏在我们目前拥有的证据强度下。在以前的荟萃分析中，在长效鞘内注射阿片类药物的情况下，局部麻醉伤口浸润和区域麻醉技术与控制相比的大部分理想效果都丧失了 [... 所有剩余的局部麻醉伤口浸润和区域麻醉技术比较的可信区间很宽，尽管缺乏统计学差异，但可能存在临床上重要的差异，但仍然隐藏在我们目前拥有的证据强度下。在以前的荟萃分析中，在长效鞘内注射阿片类药物的情况下，局部麻醉伤口浸润和区域麻醉技术与控制相比的大部分理想效果都丧失了 [... 所有剩余的局部麻醉伤口浸润和区域麻醉技术比较的可信区间很宽，尽管缺乏统计学差异，但可能存在临床上重要的差异，但仍然隐藏在我们目前拥有的证据强度下。在以前的荟萃分析中，在长效鞘内注射阿片类药物的情况下，局部麻醉伤口浸润和区域麻醉技术与控制相比的大部分理想效果都丧失了 [...3, 4 ]，而 Singh 等人 [ 1 ] 没有充分分析这种关系，排除了进一步的结论。

可以理解，网络荟萃分析很复杂，并且随着干预数量的增加，涉及的成对比较的数量可能具有挑战性，总结多种干预的相对影响可能具有挑战性。因此，我鼓励对所有系统评价增加报告和利用最小的临床重要差异，以促进解释，即使其选择的证据有限 [ 5 ]，并欢迎发表与这一特定的不精确性相关的结果系统评价。