An unfiltered rant about the most under-rated electrolyte
First of all, there is the hyperchloremic metabolic acidosis usually caused by excessive normal saline, which unlike LR has a supraphysiological ratio of chloride to sodium (though as we will discuss, there are contexts when you should use normal saline). Then there is this theory that chloride actually causes AKI.
What is the likely mechanism of AKI? It probably has to do with tubuloglomerular feedback (TGF). Recall that TGF involves a negative feedback loop that aids in autoregulation of glomerular filtration. See animation with this concept for review: Chloride and TGF. Notably, chloride rather than sodium is the electrolyte that is primarily sensed as a marker of tubular flow. You can therefore imagine that excess filtered chloride in the context of hyperchloridemia could similarly cause a reduction in GFR. This is an easy mechanism to remember, though probably it is more complicated; interstitial chloride along with tubular chloride may also play a role [1]. It may be that chloride-induced reduction in GFR via TGF may contribute to clinically significant AKI when there are additional superimposed hemodynamic insults to the kidneys.
What is the clinical evidence for chloride-induced AKI? Initially, in a very small (N=12) study of healthy subjects randomized to a 2L bolus of NS vs. plasma-lyte (a balanced solution), a significant reduction in renal perfusion was observed in those receiving NS compared to plasma-lyte [2]. This first study in humans redemonstrated findings that had previously been seen in prior animal studies [1], and notably, no change in urine biomarkers of tubular injury between the two groups was seen [2], suggestive that the mechanism did not inherently involve direct tubular injury. Data from much larger trials have been mixed: among other trials, both the SMART trial [3], of over 15 thousand ICU patients, and the SALT-ED trial [4], of over 13 thousand non-critically ill patients in the ED, demonstrated a reduce risk of adverse kidney events (though not mortality) with balanced crystalloid fluids rather than NS. Keep in mind that in both of these trials, each with a different heterogeneous study population, the number needed to treat with respect to reducing adverse kidney events is on the order of ~100. The most recent data from the PLUS trial, published in NEJM earlier this month, are inconsistent with these earlier trials [5]: in this trial of over five thousand ICU patients with clinical indication for crystalloid, patients were randomized to either normal saline or plasma-lyte; serum chloride level was significantly higher in the NS group, but there was no statistically significant difference in longitudinal serum creatinine values, adverse kidney events, or in the primary endpoint of death at 90 days [5].
When it comes to this chloride-related AKI, it seems the jury is not out yet! Studies are ongoing, and the heterogeneity of these large trials is obviously a limiting factor, as harm from normal saline probably depends on clinical context. Avoiding, for example, an iatrogenic hyperchloremic metabolic acidosis from NS in a patient who is already acidemic should be self-evident! So then in what other contexts might NS actually be beneficial?
If the chloride load from normal saline is best avoided in the context of hyperchloremia or acidemia, then chloride must be a good thing in the opposite context, metabolic alkalosis, right? That seems to be the case!
Recall that unlike metabolic acidosis, a metabolic alkalosis requires not only a primary inciting cause but also maintenance factors that limit the expected renal correction of this problem. Why these maintenance factors? Because the kidneys exert so much energy reabsorbing filtered bicarbonate in order to maintain homeostasis and defend the body pH, it is conversely quite easy, energetically speaking, for the kidneys to just dump that filtered bicarb (simply not reabsorb it) when necessity compels in the context of an alkalemia. Recall the three classic maintenance factors that prevent the kidneys from dumping bicarb in the context of alkalemia: hypovolemia (mechanism likely relates to RAAS activation), hypokalemia (mechanism beyond scope of this email and my own fund of knowledge here!), and, most importantly, hypochloremia! Some authorities have even suggested that it is the hypochloremia that may actually be the principal maintenance factor in metabolic alkalosis (based on evidence that chloride can maintain metabolic alkalosis even without volume contraction or hypovolemia) that we should abandon the term contraction alkalosis and instead start saying
So how does a low chloride state help maintain a metabolic alkalosis? As with any renal tubular process, it is always more complicated in reality but practically speaking it is easiest to just think of that one proposed mechanism that helps us remember how it works clinically (analogous to our highly practical but potentially unrealistic framework for HAGMA vs. NAGMA +/- delta/delta, which is another topic for a separate rant). I personally like to think through it based on the following (perhaps now outdated) mechanism: distal tubular flow of sodium is of course of great clinical relevance; it is how med students often remember the reason loop diuretics cause hypokalemia and alkalosis, since more sodium flowing through the distal nephron means more sodium reabsorption, for which a potassium or hydrogen ion may be secreted to maintain tubular electroneutrality. Since this distal sodium reabsorption in exchange for distal hydrogen ion secretion depends on electronegative potential generated by a sodium reabsorbed without a corresponding anion, image what would happen if chloride was also reabsorbed along with that sodium; no proton could be secreted into the distal tubule! If this is confusing, here is another animation to help visualize this: Chloride and Metabolic Alkalosis. In summary, low chloride in blood -> low chloride filtered in kidney -> less tubular chloride that can be reabsorbed distally together with distal sodium -> reduced proton secretion distally. In reality, there are multiple other likely mechanisms that explain why chloride helps maintain the alkalosis relating to RAAS, TGF, and other distinct distal tubular exchange processes involving chloride and bicarb directly [6].
So when is this relevant clinically? Considering the physiology of a metabolic alkalosis (inciting cause along with maintenance factors), here is the perfect example of a patient I recently saw as a phys consult: we were consulted for a patient with two days of serum bicarb >46 (above assay) and a very low chloride. This patient was admitted to surgery with an SBO and had been vomiting profusely for several days, with substantial NGT output. The cause of the alkalosis was multifactorial: loss of stomach acid from the emesis, including by the way loss of chloride in the vomit, that had resulted in substantial hypovolemia. Fortunately, the patient had two working kidneys, so to fix the problem, we needed only to remove the maintenance factors, thereby allowing the kidneys to dump that excess bicarb. In order to correct both the hypovolemic-maintenance factor and the hypochloremic maintenance factor for the alkalosis, the crystalloid solution of choice was obvious: good old normal saline! Once euvolemia was restored, subsequent chloride was administered not with a volume expander but via oral potassium chloride (she also had low potassium iso vomiting). Before the consult note was even finished at the end of the day, the bicarb had already returned to her baseline (which admittedly was actually quite high, since she also had underlying COPD).
Here is one final pearl on chloride: with all this talk about serum chloride, what is the urine chloride good for? Usually we look at urine sodium as a marker of tubular avidity and RAAS activation. Of course, this can be unhelpfully high even in a RAAS active state if loop diuretics have been administered (hence FEUrea instead). Additionally, urine sodium can be unhelpfully high in the context of a metabolic alkalosis even in a RAAS active state, since the excreted bicarbonate will cause sodium to follow. Urine chloride, however, is not obligated to chase the bicarbonaturia through the collecting ducts, making a low urine chloride a more sensitive marker of RAAS activation than urine sodium in the specific context of metabolic alkalosis.