现在的位置: 首页研究点评, 进展交流>正文
[Lancet作者回复]:急性肾损伤的肾脏替代治疗
2021年01月28日 研究点评, 进展交流 暂无评论

CORRESPONDENCE

Renal replacement therapy in acute kidney injury – Authors' reply

Stéphane Gaudry, David Hajage, Agnès Dechartres, et al

Lancet 2020; 396: 1975-1976 DOI:https://doi.org/10.1016/S0140-6736(20)32673-8

We disagree with Vincenzo Sepe and colleagues' comment about the timing and prescription of renal replacement therapy (RRT). In the randomised controlled trials included in our systematic review and individual patient data meta-analysis,1 the protocol was precise; it mandated starting RRT as soon as Kidney Disease: Improving Global Outcomes acute kidney injury stage 2 or 3 was present in the early strategy, and it had stringent criteria for initiating RRT in the delayed strategy, such as severe hyperkalaemia or acidosis, among others. Obviously, clinicians had some degree of freedom in the interpretation of such criteria, which is scientifically and ethically desirable. If we follow Sepe and colleagues' reasoning, all randomised controlled trials in the field would be subject to the same criticism, including STARRT-AKI,2 which is the largest trial to date.

我们不同意Vincenzo Sepe及其同事有关肾脏替代治疗(RRT)时机及处方的评论。在我们的系统回顾以及个体病例数据meta分析中,纳入的随机对照试验的试验方案非常详尽;早期策略组要求急性肾损伤的KDIGO分级2或3级时尽快开始RRT,延迟策略组开始RRT也有严格标准,如严重高钾血症或酸中毒。很明显,临床医生在解读这些标准时具有一定的自由度,从科学和伦理角度都是允许的。如果我们遵从Sepe及其同事的推理,同样的批评意见适用于这一领域中所有的随机对照试验,包括迄今为止最大样本的临床试验即STARRT-AKI研究。

With regards to Sepe and colleagues' second comment, indeed, most patients with severe acute kidney injury have multiple system organ failures and require care intensive care.3 The COVID-19 pandemic is a tragic example of this situation. Do they consider that a simple renal unit or an internal medicine ward would be able to save the lives of patients with severe acute respiratory distress syndrome, septic shock, and acute kidney injury? Intensive care units (ICUs) probably do not care for the same patients as Sepe and colleagues care for, and we would be interested in having more data from their own experience. We believe that not taking care of patients who have multiple system organ failure in an ICU is not safe in a high-outcome country with modern care. Interestingly, in the only study done in a country that could not offer ICU beds for all patients, results were in favour of a delayed RRT strategy.4

有关Sepe及其同事的第二个评论,事实上,多数严重急性肾损伤患者同时有多器官功能衰竭,因而需要收入ICU。COVID-19疫情就是这种情况的一个典型例证。他们是否认为单纯肾脏内科病房或内科病房就足以挽救重度ARDS、感染性休克和急性肾损伤患者的生命。ICU收治的患者可能与Sepe及其同事治疗的患者不同,我们有兴趣了解根据他们自己经验的所总结的数据。我们相信,在高收入国家中,将多器官功能衰竭患者放在ICU之外接受治疗并不安全。有趣的是,在无法为所有患者提供ICU床位的国家所做的唯一研究支持延迟RRT的策略。

We also do not agree with the contention by Zhenxing Lu and colleagues that considering absolute effect would modify our conclusion. We have already reported the results for mortality at day 28, our primary endpoint in terms of absolute risk difference.1 Regardless of the endpoint definition (ie, mortality at day 28, day 60, hospital mortality), the absolute risk difference is close to 0, and the 95% CI is narrow (0·01 [95% CI −0·04 to 0·06] for mortality at day 28; 0·00 [–0·05 to 0·05] for mortality at day 60; −0·01 [–0·06 to 0·03] for hospital mortality). Our meta-analysis provides a fairly accurate estimate of the absolute difference in mortality between the groups of patients receiving delayed and early RRT. Zhenxing Lu and colleagues express these differences in number of deaths per 1000 patients, which might exaggerate the impression of imprecision (it would be even worse if expressed by millions of patients), yet this does not change our results. We wonder what degree of precision would have been considered sufficient to Zhenxing Lu and colleagues; however, we want to highlight that many non-inferiority trials in ICUs admit a non-inferiority margin for mortality equal to or higher than 0·05.5

我们也不同意Zhenxing Lu及其同事的意见,他们认为绝对效应可能改变我们的结论。我们报告了主要研究终点即28天病死率的绝对效应。无论主要研究终点的定义如何(28天,60天及住院病死率),其绝对风险差异均接近0,95% CI范围很窄(28天病死率0·01 [95% CI −0·04 to 0·06];60天病死率 0·00 [–0·05 to 0·05];住院病死率−0·01 [–0·06 to 0·03])。我们的meta分析对于接受延迟和早期RRT患者病死率的绝对差异提供了较为准确的评价。Zhenxing Lu及其同事将这一差异表示为每1000例患者的死亡人数,这可能增加对于结果不准确的印象(如果将结果表示为每1000000例患者将会更糟糕),然而这并不改变我们的结果。我们想知道Zhenxing Lu及其同事认为精确度达到何种程度才足够;然而,我们希望指出的是,很多ICU的非劣效试验将病死率的非劣效界值等于或高于0.05。

给我留言

您必须 [ 登录 ] 才能发表留言!

×
腾讯微博