Fluid Resuscitation in Septic Shock: "There is nothing normal about Normal Saline" by Steven P. LaRosa, M.D.

So let me take you way back (at least for me) to medical school and residency. One of the first practical things that we got taught was about fluid management of patients. Hours were spent on teaching us the difference between a deficiency in free water requiring hydration versus volume depletion. We were taught that if a patient was intravascularly volume depleted that you needed to provide a solution isotonic with plasma. If you were a "flea" or aspiring "flea" (see Internal Medicine physician) this was done with 0.9% sodium chloride or what is known as Normal saline (NS) and if you were a surgical type with Lactated Ringers solution (LR). I cannot recall during residency or Chief residency a whole lot of discussion as to the merits of one type of crystalloid over the other and it felt like this was nothing more than a cultural phenomenon.

Fast forward now to my early days as an Infectious Disease Attending and aspiring Sepsis Researcher. I would attend many conferences, some in North America and some in Europe where heated debates would occur on the benefits of crystalloids versus colloids (albumin, dextran, gelatin, hydroxyethyl starch) as resuscitation fluids in septic shock. Again, it really appeared that this came down to a cultural phenomenon with the North Americans siding with crystalloids and the Europeans favoring colloids. Numerous studies and meta-analysis were performed without any evidence of superiority of crystalloids over colloids. Hydroxyethyl starch however found to be associated with poorer outcomes and an increased incidence of acute renal failure. Ultimately because of the lower cost, crystalloids are the preferred fluids by the Surviving Sepsis Guidelines.

The discussion of whether there is a superior type of crystalloid solution has been a relatively recent phenomenon. The development of 0.9% sodium chloride solution dates back to the late 1800s. This solution was considered to be physiologic because it caused less hemolysis of red blood cells than other concentrations. The chloride concentration in 0.9% sodium chloride (154mmol/L) is much higher than plasma (100mmol/L) and the balanced salt solutions Plasma-Lyte 148 (98mmol/L), and Lactated Ringers (LR) (112mmol/L). There is also a lower strong ion difference with 0.9%NS at 0 compared with plasma at 24 and LR at 28. The elevated chloride and low strong ion difference with 0.9% sodium chloride is associated with development of hyperchloremic acidosis. Animal and healthy volunteer studies have demonstrated renal arteriole constriction, decreased renal perfusion and GFR as a result of the hyperchloremia present in NS.

A number of publications comparing patient outcomes with the use of normal saline versus balanced salt solutions have appeared in the last 4 years. In 2014 Raghunathan and colleagues published the results of a retrospective cohort analysis with propensity matching comparing ICU patients who had received some balanced salt solutions versus those who did not receive any balanced salt solutions. In the 6730 patient sample patients who received balanced salt solutions had a statistically significant relative risk reduction of 14%. A dose response was observed with an incrementally lower mortality for every 10% increase in the percentage of fluids that were balanced salt solutions. In a retrospective cohort analysis by Zampieri and colleagues the use of Lactated Ringers was associated with a lower mortality and less acute kidney injury than normal saline. This effect was more pronounced with greater volumes of fluid given. In another study, patients who had received a large chloride load with at least 60 ml/kg of NS had an increased mortality. Walley and colleagues demonstrated an increased incidence of acute injury with an increase in the serum chloride by 5mmol/L during the ICU stay.
Proponents of the ongoing use of 0.9% sodium chloride as the primary resuscitative fluid point to results of the only reasonably sized randomized, controlled trial of balanced salt solutions versus NS. The SPLIT trial randomized 2278 ICU patients in New Zealand ICUs to receive NS or Plasma-Lyte when a crystalloid was required during their stay. No differences were observed in mortality, incidence of acute injury or the need for renal replacement therapy. The patients in this trial had a very low disease severity and received on average 2liters of fluid. This population does not even come close to approximating critically ill patients with septic shock.

So, where does this leave us with respect to crystalloid choice in septic shock? Fortunately at least 2 large RCTs are being conducted to definitively answer this question. In the meantime, based upon the non-physiologic nature of 0.9% sodium chloride, the animal and healthy volunteer data as well as the accumulated retrospective data, I would choose a balanced salt solution over NS. One cannot make a financial case over this as 1L of NS and LR cost $1. Plasma-lyte is about 2 bucks a bag, Hell, I'll pay the extra buck out of my own pocket!