First, let me apologize for the long delay between posts. We had a bit of a snafu in the blog, but things should be better now.
On to a discussion about isotonic crystalloids in our critically ill patients. Administration of intravenous fluids represents the single most common intervention in critically ill patients. And most of that iv fluids is given in the form of isotonic crystalloids. So we are all on the same page, isotonic crystalloids include 0.9% NaCl (i.e. Normal Saline), Lactated Ringer’s solution, and Plasmalyte. The use of one of these fluids is almost ubiquitous in our care of patients.
Intravenous fluids were first developed in the 1830’s for the treatment of Cholera. They were refined in the 1880’s when Ringer’s solution was developed to bathe cardiac muscle ex vivo for physiology studies and 0.9% NaCl was found to avoid RBC lysis in vitro. Use in patients began shortly thereafter and the argument about which iv fluid was best began.
In practice, clinicians believe in either Normal Saline or balanced intravenous fluids, largely based on dogma and their training. In the United States, in general, surgeons prefer balanced intravenous fluids while internists prefer Normal Saline. In general, Normal Saline uses chloride exclusively as the anion while the balanced solutions have a lower concentration of chloride and are balanced with other anions (see table 1 below for composition of common isotonic intravenous fluids compared to plasma). However, despite being available and administered to patients for over a century, studies comparing the different isotonic intravenous fluids were very limited until recently.
Recently, data began emerging suggesting that Normal Saline may be associated with worse outcomes. It isn’t really debated that the higher chloride load from Normal Saline results in a hyperchloremic metabolic acidosis (Yunos NM, et al. CCM 2011;39(11):2419-2424.) (Table 2). What is debated is whether or not that hyperchloremic acidosis makes any difference in the clinical outcomes of our patients.
Observational studies suggested an association between administration of Normal Saline and higher chloride intravenous fluids and both renal failure and mortality (Raghunathan K, et al. CCM. 2014;42(7):1585-1591.; Shaw AD, et al. Intensive Care Med. 2014;40(12):1897-1905.; Raghunathan K, et al. Anesthesiology. 2015;123(6):1385-1393). Then a before-after study showed a reduction in the incidence of renal failure and need for renal replacement therapy by altering practice to severely restricting the use of iv fluids with higher chloride content, such as Normal Saline (Yunos NM, et al. JAMA. 2012;308(15):1566-1572).
While this all sounds very provocative, subsequent studies demonstrated conflicting results and clouded the picture. First, Yunos and colleagues followed their before-after study with an additional follow-up period and saw some regression back towards their incidence of renal failure and renal replacement therapy in the before group, despite continuing restricted NS access (Yunos NM, et al. Intensive Care Med. 2015;41(2):257-264). And other studies failed to find any detrimental effect of Normal Saline administration (Krajewski ML, et al. Br J Surg. 2015;102(1):24-36).
As is often the case, the ANZICS group decided to study this in a double-blind, cluster randomized, crossover trial in 4 ICUs which they called the SPLIT study. In the SPLIT study, ICUs were randomized to Normal Saline or Plasmalyte for 7 week blocks and then crossed over to the other iv fluid three times throughout the study (Young P, et al. JAMA. 2015;314(16):1701-1710). However, patient data were only used if consent could be obtained, which occurred in 2278 of the 3346 admitted patients. There was no difference in development of acute renal failure (primary outcome), need for renal replacement therapy, change in serum creatinine, ICU or hospital lengths of stay, or ICU or hospital mortality.
Some thought these data definitively answered the question. However, others were concerned about potential bias in excluding the patients who consent could not be obtained from and the relatively low volume of fluid given to study patients overall.
Last month, our group published a similarly designed trial (cluster randomized, crossover design) in medical ICU patients (Semler MW, et al. AJRCCM. 2016; Epub Ahead of Print, October 17, 2016). Similar to the SPLIT study, we didn’t see any difference in renal failure, new dialysis, or mortality overall. However, an analysis by volume of iv fluids received demonstrated a very interesting signal. Patients who received higher volume of isotonic intravenous fluid had worse renal function and higher mortality when they received Normal Saline as that fluid. The signal was not present for patients who received less fluid, especially less than 2 liters of fluid during their ICU stay.
Well, well, well. These data are incredibly provocative. And may make some sense. If Normal Saline is harmful (and I still don’t think we have the definitive data to tell us if it is or not), then it makes sense that we would see a bigger signal of harm with higher volumes of Normal Saline administered.
Overall, I think this question is still up for debate. Large trials are ongoing which will provide more data and information toward better answering this question. Until those trials are completed, however, it makes sense to me that switching to, or continuing to favor, balanced intravenous fluids makes sense. Although the signal for benefit of balanced fluids has been mixed, no studies have suggested worse outcomes when balanced IV fluids are used. As far as cost, both balanced and Normal Saline IV fluids are cheap and similar in price. Given this, and the currently available data, I think the preferred use of balanced iv fluids (and I think Lactated Ringer’s and Plasmalyte are equally acceptable) is a reasonable approach in our critically ill patients.
Let me know if you have other thoughts.