Balanced Crystalloid versus Saline Solution in Critically Ill Patients: Is Chloride the Villain?

To the Editor:

Semler and Kellum present a thorough and scholarly review of studies (theirs and others) comparing saline and balanced crystalloid solutions for intravenous fluid therapy in critically ill patients, and make a good case for the superiority of balanced crystalloid solutions over saline with respect to mortality and adverse renal events (1). This issue will hopefully be definitively settled by the results of two large randomized controlled trials in almost 20,000 patients that are currently underway and hopefully will involve the administration of larger volumes than the 1–2 L studied to date. In large part, Semler and Kellum highlight the deleterious effects of hyperchloremia and associated mild metabolic acidosis arising from saline administration. However, the argument that modest elevations in serum chloride after saline administration are entirely responsible for these worse outcomes is too simplistic. Much of the putative blame attached to chloride rests on the widely cited experiments of Wilcox (2), which involved isolated blood perfusion of dog kidneys with various hypertonic fluids at a chloride concentration of 126 mM. The kidney’s sudden exposure to an instantaneous almost 20-mM rise in chloride (and the resulting hypertonicity) led to a degree of vasoconstriction and release of thromboxane that Semler and Kellum and others cite as the cause for chloride’s vasoconstrictive and proinflammatory effects in patients requiring fluid resuscitation. This rationale, however, does not necessarily carry over to far lesser and more slowly developing 2- to 4-mM plasma chloride elevations as the cause of renal injury and increased mortality among critically ill patients given saline. It is important to note that Wilcox did no dose-response experiments within the range of chloride elevations that are more typically found in saline-treated critically ill patients. Other differences in the composition of balanced crystalloids beyond changes in chloride concentration could be playing a protective role in the outcomes that appear to be consistent across multiple trials. Semler and Kellum do suggest that there may be benefits to the provision of lactate or other metabolized anions in balanced solutions, as there is emerging evidence that lactate functions as an important fuel in the central nervous system and heart under stressed conditions. Likewise, small changes in potassium and calcium concentrations might also be beneficial. One way to potentially absolve or condemn chloride would be to test “normal” saline against saline with a one-to-one replacement of 24-mM bicarbonate for chloride. Given the present lack of equipoise regarding chloride, this experiment is unlikely to be performed, but until such time, chloride should be presumed innocent and not yet guilty as charged. In analogy to the arguments that arose concerning the original goal-directed bundled therapy for sepsis resuscitation proposed by Rivers and colleagues (3), balanced crystalloid solutions are a “bundle,” and we do not know which element(s) is the most critical—less chloride or its replacements. Although one sometimes hears the casual statement that saline may kill, millions of patients saved might otherwise disagree.