Spare the Chloride

Fluid therapy is essential in ICUs and not surprisingly there is still much controversy about which fluid to use, how much and when. Nephrologists often roll their eyes at other subspecialty's preferences, e.g. surgeon's preferences for Ringers, citing the risk of hyperkalemia in renal failure patients given Ringers. I learned that normal saline is the preferred agent unless there is a special consideration such as acidemia necessitating alternatives. Now chloride, the partner of sodium that gets considerably less attention most of the time, enters the stage.

Yunos et al in JAMA suggest that too much of chloride increases acute kidney injury (AKI) episodes in tertiary ICUs and increases the need for renal replacement therapy (RRT) but does not affect mortality.

The physiological rationale for the detrimental effect of chloride on the kidney is described as vasoconstriction mediated by chloride in dog experiments and a possible role of tubuloglomerular feedback mediated vasoconstriction as well as decrease in GFR caused by increased distal chloride delivery. Furthermore they cite thromboxane mediated vasoconstriction caused by chloride and enhanced responsiveness to vasoconstrictor agents as possible physiological sequelae of chloride administration.

The authors of the JAMA article conducted a prospective, open-label sequential pilot study of patients admitted consecutively to the ICU. Initially patients were treated with chloride-rich IV fluids (0.9% saline, 4% succinylated gelatin solution or 4% albumin solution) and after that initial control period a chloride-restricted strategy was implemented with lactate (Hartmann solution), a balanced solution (Plasma-lyte 148) or chloride-poor 20% albumin as preferred agents.

The results were a lower increase in serum creatinine levels and fewer episodes of RRT in the chloride-restricted group but no differences in mortality, hospital or ICU length of stay or need for RRT after discharge.

How does this study affect our choice of ICU fluids? Certainly, these results are hypothesis generating and important but need to be viewed as preliminary given the design of the study. An accompanying editorialby Waikar mentions the Hawthorne effect as potential major concern. Clearly these important preliminary data need follow up in a controlled prospective trial. 

Posted by Florian Toegel