Reply to: adverse effects of hyperchloraemic solutions
We appreciate Professor Priebe's interest1 in our study.2 We would like to comment on the issues and questions raised in his letter. We agree with Dr Priebe's first statement that some factors may have modified or obscured a possible treatment effect of preoperative intravenous hydration. Specifically, as we addressed in the discussion section, the pragmatic trial design with uncontrolled selection of the type and volume of intraoperative and postoperative fluids may have acted as a confounding factor.
We understand that the selection of 0.9% normal saline as our study fluid can be considered controversial for many readers of the journal at the present time. However, our study was designed over 10 years ago when the clinical evidence regarding the potential adverse effects of chloride-rich solutions was scarce or lacking. This is what we state in the discussion. We support this statement with two publications, the work by Shaw et al.3 published in 2012 (6 years after the inclusion of our first patient), and a position paper (pro–con debate) by Lobo and Awad4 published in 2014. Both manuscripts, as Professor Priebe correctly claims in his letter, highlight the adverse effects of normal saline. Our purpose in citing these articles was to draw readers’ attention to the dates of these publications. We apologize if this was not clear in the text. Furthermore, our statement about the rare clinical consequences of hyperchloraemia and acidosis after resuscitation with normal saline is based on the opposite point of view held by Ince and Groeneveld5 in the same pro–con debate. It is our opinion that the reader interested in this controversy should start by reading these two articles.
We disagree with Professor Priebe's opinion about the robustness of the association between the use of chloride-rich fluids and adverse outcomes. This assessment is at least debatable. The 0,9% Saline vs Plasma-Lyte 148(PL-148) for ICU fluid Therapy (SPLIT) randomized clinical trial,6 compared over 2000 ICU patients (over 70% of them postsurgical) and found no difference in terms of safety (acute kidney injury, mortality) between those treated with normal saline and those treated with a buffered crystalloid solution with a physiologic chloride concentration (Plasmalyte A 148, Baxter SL, Ribarroja del Turia, Valencia, Spain). Similarly, the meta-analysis by Krajewski7 cited in Professor Priebe's letter concludes that ‘these findings underscore the need for large, well designed randomized controlled trials that are adequately powered to detect differences in outcomes such as mortality and morbidity including AKI’. There is no doubt that large amounts of normal saline administered over short time periods cause hyperchloraemia and acidosis. However, it is a long way from the 2 l of normal saline infused to healthy volunteers over 1 h in Chowdhury's work,8 cited by Professor Priebe, and the 1.5 ml kg−1 h−1 infused to American Society of Anesthesiologists II to IV patients over 12 h in our trial.
Regarding other questions posed by Professor Priebe, hyperchloraemia in our study was defined as a value greater than 110 mmol l−1. Although it is true that differences in postoperative hyperchloraemia incidence cannot be assessed without knowing preoperative values, there is no reason to believe that randomization did not work for this particular electrolyte.2 We report the postoperative incidence of hyperchloraemia as an adverse event2 at the request of one reviewer, but it was not included in the original plan. Over the 5-year period during which the trial was conducted (2006 to 2011), Hartmann's solution (chloride concentration 109 mmol l−1) was the first choice for intraoperative maintenance fluid at our institution. This may explain the postoperative incidence of hyperchloraemia.