Abstract
Dr. McDonald and colleagues have attempted to address the role of tranexamic acid (TXA) in reducing bleeding complications from percutaneous nephrolithotomy (PCNL). 1 They have provided a systematic summary of body of evidence that largely supports the use of TXA to reduce major bleeding events and is in line with observations from other types of surgery. However, many key questions remain about when to use TXA, which patients benefit most? What is the optimal timing? Does delayed postoperative initiation work as well as preoperative? Would less invasive PCNL with lower bleeding risk retain any benefit to TXA?
The downstream consequence of major bleeding should not be underappreciated. The Vascular events In noncardiac Surgery patIents cOhort evaluatioN (VISION) study, which included >40,000 patients, identified major bleeding as strongly associated with 30 days mortality rate. 2 The attributable fraction (potential proportion of deaths attributable to this complication) was 17% in this cohort, with an adjusted hazard ratio 2.6, 95% confidence interval (CI) 2.2–3.1. Thus, the effect of major bleeding on mortality rate is substantial. Although there is too little evidence to demonstrate this effect among a population of patient undergoing PCNL, the VISION study is international, multi-institutional, multispeciality, and demonstrated consistent effects across most surgery types. As such this at least provides strong indirect evidence for the importance of reducing major bleeding complications in general.
The major limitation of the study presented by MacDonald et al. is the limited reporting of patient important outcomes available in the literature and studies of PCNL. In this case data for <1000 patients, judged to be of sufficiently low risk of bias, were available overall. Only a subset of this for any given outcome. Unfortunately, this precluded evaluating the impact of TXA on mortality rate or other important outcomes, such as embolization or surgical intervention for bleeding, which occur in about 1% of this population. 3
The authors were able to calculate a reduction in transfusion with wider CIs 0.34 (95% CI [0.19–0.61]), or somewhere between 40% and 80% decreased risk. This effect size seems quite large and probably reflects the smaller amount of data available. Although it is possible that TXA in the PCNL population somehow has a much larger effect than is typically seen in other patient populations, it is difficult to be certain from the available data. In contrast, a large meta-analysis (N > 10,000) by Ker et al. in other patient populations suggest a reduction in transfusion closer to 1/3, 4 which is consistent with other large trials of postpartum hemorrhage 5 and coronary artery bypass. 6 The effect might be even less in populations with lower risk of major bleeding. 7,8
It is clear that bleeding is an important potential harm in PCNL and that TXA does reduce the need for transfusion and possibly more important outcomes. It is important for our field to continue to elucidate which patient populations benefit most and to report our harms data well for the outcomes that matter most to our patients.