Eligibility criteria in paravertebral block meta‐analysis – a reply

Abstract

I thank Abdallah et al. for their interest in our study 1. A major consideration in designing studies on the long- or short-term effects of regional anaesthesia is blinding. If no sham block is performed, effective blinding is difficult to achieve because the patient is aware of a regional block being, or not being, performed. In case of continuous blocks, the outcome assessor is hard to blind as the presence of a catheter should be obvious. Ilfeld et al. 2 employed a sensible methodology: they performed the paravertebral block in both groups. Research ethics commands a consideration of the risks inflicted on the study participants. A paravertebral block carries the risk of serious complications including pneumothorax or high spinal anaesthesia and performing a sham block with no benefit for the patient may be considered unethical. It was conceivable that this or a similar consideration would lead to a design that involved an injection in both groups, the treatment group and the control group. After a first paravertebral injection in both groups, the control group received local anaesthetic via a perfusor while the control group received saline. Although Abdallah et al. are right that the Ilfeld et al. 2 study design differed from that of the other studies, we are not sure about the impact on our results. Firstly, the risk ratio (RR) of the Ilfeld et al. study was quite similar to that of Lee et al. 3 and Karmakar et al. 4 for pain after 3 months. Ifield et al.'s 12 months data are similar to those of Kairaluoma et al. 5. Omitting Ifield et al.'s data, the effect estimate remains not significant: at 3 months RR 0.67 (95%CI 0.32–1.42) including, vs. RR 0.75 (95%CI 0.48–1.15) excluding; and for pain 12 months after surgery, RR 0.42 (95%CI 0.15–1.16) including, vs. RR 0.52 (95%CI 0.09–3.02) excluding. Abdallah et al. contend that ‘local infiltration itself (is) an effective intervention in preventing chronic pain following breast cancer surgery’, but criticise local anaesthesia infiltration in the control group by Chiu et al. 6. We included Abdallah et al.'s study on the effects of paravertebral block on chronic pain in our meta-analysis, even though this study contained a control group for whom local anaesthetic was infiltrated 7.