Surgical trials for pain relief: in search of better answers.
Surgical trials for pain relief: in search of better answers.
复制标题
缓解疼痛的手术试验：寻找更好的答案。
DOI：
10.1016/j.pain.2012.07.017
复制
发表时间：
2012
期刊：
影响因子：
7.4
通讯作者：
Ching,Alex
中科院分区：
文献类型：
--
作者：
Deyo,RichardA;Ching,Alex
Surgical interventions are an important aspect of pain management. In the US, rates of cervical spine surgery for neck pain, and especially fusion procedures, have increased rapidly in recent years [1]. Though randomized controlled trials (RCTs) of surgery are more challenging than RCTs of drug treatments, they are feasible and at least as important, given the risks and costs that attend invasive interventions. Unfortunately, because of the complexities and costs, RCTs of surgical interventions are relatively uncommon, tend to be small, and have methodological challenges. The systematic review of cervical spine surgery by van Middelkoop et al. in this issue of Pain, epitomizes some of the frustrations of reviewing such randomized trials [2]. The authors identified 10 randomized trials of decompression alone versus decompression plus fusion for patients with neck pain, with or without neurologic signs. The results suggested no advantage of adding a fusion to decompression, but the conclusion is mitigated by the observation that overall, the evidence is low to very low in quality. The internal validity of the trials was generally uncertain, and the external validity (generalizability) even more so. Why are surgical RCTs so difficult? There are several important differences from drug trials. First, the intervention cannot be standardized like identical pills. Each patient’s pathoanatomy is slightly different from the rest, and each surgeon has a particular combination of skills and preferences that vary from the rest. Nonetheless, many features of surgical technique and patient selection can be standardized. Second, unlike a drug trial where treatment can be stopped for adverse effects, surgical trials are like ‘‘crossing the Rubicon’’. Once the intervention is done, there is no going back to the original anatomy. There is greater finality in the intervention. Perhaps for this reason, it is often harder to find both patients and surgeons with true equipoise, equally willing to pursue two alternative treatments. In RCTs comparing surgical with nonsurgical treatments, blinding may be impossible. However, in comparing one surgical treatment to the other, blinding patients and outcome assessors should often be feasible. Among the trials analyzed by van Middelkoop et al., most were small (average size of the smallest arm= 29 subjects) and none attempted blinding of patients or outcome assessors. In some cases, loss to follow-up was substantial, and in some, the timing of outcome assessments was highly variable. One might wish for more consistent reporting of certain outcomes. In studies of neck pain, especially with myelopathy, it would be helpful to have routine reporting of neurological outcome. The occurrence of repeat surgery and complication rates are other important outcomes that go missing here, but are essential for comprehensive comparison. Two-year follow-up is often regarded as a standard in trials of spinal surgery, yet this may often be insufficient [3]. Here, the authors defined any follow-up greater than 12 months as longterm. Yet surgeons regard this as the minimum time to diagnose non-union in the case of a fusion operation. Of the 10 studies included here, only 6 had 24 months or more of follow-up. Given the healing time, it may be unsurprising that decompression alone and decompression with fusion have similar success rates in a relatively short (< 24 months) time frame. Another long-term concern is the potential for evolution of cervical kyphosis, a complication of decompression alone that may affect neck pain and neurologic recovery. This problem may become apparent over 3–5 years of follow-up. In fact, one study in this synthesis reported a higher rate of …
