In today’s journal club we analysed an interesting paper looking at the role of a chloride restrictive approach to fluid therapy for minimising the risk of AKI. The authors start by discussing some of the experimental data that suggests the high amounts of chloride in some commonly used fluids impairs renal function with effects such as renal vasoconstriction and reduced urine output. They note that there was little clinical evidence of this harm at the time of the study (published in 2012) and hence embarked on this particular project. The full text is available at the link below (free) but I also advise having a quick look at the supplementary materials and letters.
Association between a chloride liberal vs chloride-restrictive intravenous fluid administration strategy and kidney injury in critically ill adults
Association between a chloride liberal vs chloride-restrictive intravenous fluid administration strategy and kidney injury in critically ill adults
What did they do?
Basically they carried out a ‘before and after’ cohort trial following a switch in the whole unit’s fluids at a single large (22 bed) ICU in Australia. They removed the ‘chloride rich’ fluids of 0.9% NaCl and Gelofusin, leaving the balanced crystalloids of Hartmann’s and Plasmalyte, and Human Albumin Solution as the colloid of choice (though if there was clinical indication for the old fluids they could still be used). Each period of the trial lasted 6 months (Feb to Aug) with a 6 months gap between to allow a phase out of the chloride-rich fluids along with educational interventions for staff. They looked at the impact of this by assessing for changes in creatinine levels, the incidence of AKI according to the RIFLE criteria, and the use of renal replacement therapy (RRT). They also assessed a number of baseline characteristics between the 2 cohorts, to reassure us as to their similarity, and commented on some other outcomes such as mortality.
What did they find?
They recruited a total of 760 patients in the control period, with 773 patients in the intervention period. Analysis of the predetermined patient characteristics are not notably different between the groups with very similar markers of baselines co-morbidity and acute illness severity. Unsurprisingly there was a significant chance in the fluid composition given to patients between the cohorts. They calculated that chloride administration dropped from a mean of 694 to 496 mmol/patient and sodium administration dropped from 750 to 623 mmol/patient. There were also rises in the amount of potassium and lactate given to patients.
In terms of the outcomes of interest the intervention group had a lower rise in their creatinine levels compared to the control group at 14.8 micromol/L, (95% CI 9.8 – 19.9 micromol/L) compared with 22.6 micromol/L (95% CI 17.5 – 27.7 micromol/L). There was a decrease in AKI as defined by the RIFLE criteria but this was only significant for the injury classification level at 3% (1.8 – 4.2) compared to 6.3% (4.5 – 8.1) of patients in the control group (p = 0.002). There was also a decrease in the use of RRT in the intervention group at 6.3% (4.6 – 8.1%) compared with 10% (8.1-12% in the control group (p = 0.005). Other differences in outcomes weren’t significant.
In terms of the outcomes of interest the intervention group had a lower rise in their creatinine levels compared to the control group at 14.8 micromol/L, (95% CI 9.8 – 19.9 micromol/L) compared with 22.6 micromol/L (95% CI 17.5 – 27.7 micromol/L). There was a decrease in AKI as defined by the RIFLE criteria but this was only significant for the injury classification level at 3% (1.8 – 4.2) compared to 6.3% (4.5 – 8.1) of patients in the control group (p = 0.002). There was also a decrease in the use of RRT in the intervention group at 6.3% (4.6 – 8.1%) compared with 10% (8.1-12% in the control group (p = 0.005). Other differences in outcomes weren’t significant.
Is it any good?
There are several problems arising from the design of the trial, with several areas for potential bias. Notably, they admit themselves it is a ‘package of care’, and included a change in several types of fluid, including gelatins and normal saline, clouding where the effect may be coming from. In addition external factors could actually be responsible for improving renal outcomes at this unit. Though the authors note that the results buck the trend of increasing RRT at that ICU, there is little description of the educational interventions performed as part of this switch in fluids. Was there a drive to improve renal outcomes at the same time, including a general awareness about the trial that might impact on clinician practice? A particular aspect of this is suggested by closer inspection of the ‘nested cohort’ - a group of 100 patients in each arm that they looked at in a bit more detail. A look at the volumes of fluid given to these patients suggests that the intervention group received significantly more fluid volume in comparison to the control group (a median volume 1840ml of Hartmann’s compared to 720ml of NaCl). Now to me that is a pretty big confounding factor that could have come about from these educational changes and an awareness on the unit of a big trial trying to improve AKI.
Final Thoughts
Overall there are some big areas for potential bias in this study from the factors just mentioned. However, as we discussed in the journal club, there is still some information to be gained from these pragmatic trials if the results are interpreted cautiously. In this case, they have taken a hypothesis with some decent preclinical evidence and applied a package of care intervention. Though it is difficult to unpick the exact cause of the (mild) benefits that they have observed, we can feel relatively comfortable that minimising the amount of chloride we give our patients isn’t going to be harmful. However, the more aggressive fluid therapy and (an assumed) cultural shift to focus on the importance of AKI in our patients, is probably also beneficial. Whilst not a game changing study on its own, it helps in some way to provide a bit more information about the best ways to manage these patients.
So in summary:
Tom Heaton
So in summary:
- It’s probably best to minimise the amount of chloride we give our patients, unless there is a clear indication otherwise.
- There are probably other interventions that are more important in minimising the impact of AKI on our patients than chloride load
Tom Heaton
References
- Yunos NM et al. Association between a chloride liberal vs chloride-restrictive intravenous fluid administration strategy and kidney injury in critically ill adults. JAMA. 2012. 308 (15). 1566-72.
RSS Feed
