The use of normal saline is ubiquitous in clinical practice. The history of saline resuscitation dates back to 1831 when the cholera epidemic swept across Britain leaving a trail of devastation. Out of sheer desperation, Thomas Latta, a young physician from Leith, near Edinburgh, administered intravenous saline solution to an elderly lady who was on the verge of death from extreme dehydration. He recounted his experience in a letter to The Lancet, “Having inserted a tube into the basilic vein, cautiously – anxiously, I watched the effects; she began to breathe less laboriously, soon the sharpened features, and sunken eye, and fallen jaw, pale and cold, bearing the manifest impress of death’s signet, began to glow with returning animation; the pulse, which had long ceased, returned to the wrist; at first small and quick, by degrees it became more and more distinct…”1 The solution that Latta administered was different in composition from the normal saline as we know it today. According to later analysis, one of the constituents of the solution was sodium bicarbonate. The Latta solution actually contained sodium 134 mmol/L, chloride 118 mmol/L, and bicarbonate 16 mmol/L. 2
The possible adverse effects of normal saline
Hyperchloremia and acidosis
Normal saline contains 154 mmol/L each of sodium and chloride. The serum chloride levels, normally 96–106 mmol/L, tend to rise with normal saline administration because of the excess chloride load. Hyperchloremia leads to metabolic acidosis, according to the Stewart physicochemical approach. Based on the Stewart equation, the pH of a solution depends on three variables: (1) the strong ion difference (SID), (2) the partial pressure of carbon dioxide (PCO2), and (3) the total concentration of weak acids.
SID = (strong cations) – (strong anions)
SID = (Na+ + K+ + Ca++ + Mg++) – (Cl– + lactate–)
The normal SID is 40–45 mEq/l. A higher value for SID denotes alkalosis, while a lower value denotes acidosis. How does this happen? As the chloride level rises, the SID decreases; this leads to the dissociation of water molecules to generate H+ ions to maintain electrical neutrality. The rise in H+ concentration results in acidosis. Thus, hyperchloremia due to excessive administration of normal saline may lead to metabolic acidosis.
Hyperchloremia and tubuloglomerular feedback
Hyperchloremia results in an increase in the chloride ions filtered through the glomeruli. This results in a higher chloride concentration in the distal convoluted tubule. The distal convoluted tubule is lined by specialized cells in the area called macula densa. The macula densa cells sense the high chloride content and trigger vasoconstriction of the adjacent afferent arteriole, thereby reducing glomerular filtration. This phenomenon is known as the tubuloglomerular feedback mechanism. Hyperchloremic afferent arteriolar vasoconstriction and the consequent reduction in glomerular filtration may lead to impairment of renal function.
Other possible harmful effects of hyperchloremia
Excessive chloride levels may lead to splanchnic hypoperfusion.3 High chloride levels and the consequent metabolic acidosis may also explain the increased blood loss, greater requirement for blood products, and the higher incidence of coagulation abnormities that may ensue following resuscitation with normal saline or hydroxyethyl starch suspended in saline.4
What’s the evidence against saline-based resuscitation?
The SPLIT trial
Young et al. compared normal saline with plasmalyte 148 in a multicentric study from four ICUs in New Zealand in a randomized controlled trial.5 Patients who required crystalloid fluid therapy according to the treating physician were included. If feasible, the study fluid was also administered in locations outside the ICU as required. A total of 2278 patients were enrolled; 1162 were assigned to the plasmalyte 148 group 1116 to the normal saline group. Most patients were admitted to the ICU after elective surgical procedures, mainly cardiovascular surgery. The primary outcome was the proportion of patients who fulfilled the criteria for injury or greater according to the RIFLE classification of acute kidney injury (AKI). There was no significant difference in the incidence of AKI fulfilling the criteria for Injury or higher according to the RIFLE classification between the two groups [102/1067 (9.6%) in the plasmalyte 148 group compared to 94/1025 (9.2%) in the saline group; (RR 1.04 95% CI: 0.8–1.36, p=0.77)]. Among the secondary outcomes, the number of patients who suffered AKI under each category of the RIFLE score was similar between groups. The requirement for renal replacement therapy (RRT) was also similar; 3.3% with plasmalyte 148 vs. 3.4% with saline. The difference between the baseline and the highest creatinine levels was also similar between groups. ICU and hospital mortality, the length of stay in ICU and hospital, and the requirement for and duration of mechanical ventilation were also similar. The risk of AKI among subgroups of patients with sepsis, trauma, APACHE score of 25 or more, cardiac surgical patients, and patients with traumatic brain injury (TBI) remained similar between groups.
The SPLIT trial was a feasibility study and hence not powered to evaluate clinical endpoints. The largest cohort of patients was postoperative cardiac surgical with relatively few patients with other underlying causes requiring ICU admission. However, this study suggested that normal the use of normal saline compared to a buffered crystalloid solution may not lead to adverse clinical outcomes.
The SMART trial
The SMART trial compared normal saline with a buffered crystalloid (Plasma-Lyte A or lactated Ringer’s solution) as intravenous fluid among critically ill patients admitted to the ICUs of a single center in the US (Vanderbilt University Hospital) in a randomized controlled trial.6 The choice of intravenous alternated between normal saline and buffered crystalloid on a monthly basis. The trial was coordinated between the ICUs, emergency department, and the operation theatres that enabled the same study fluid to be administered at different locations. The trial enrolled 15802 patients; 7942 patients received plasmalyte while 7860 patients received normal saline. The primary outcome was the proportion of patients who met one or more criteria for a major adverse kidney event, censored at 30 days (MAKE-30) including a composite of death, RRT, or persistent renal dysfunction. A major adverse kidney event occurred in a significantly higher number of patients who received normal saline compared to those who received buffered crystalloid (15.4% vs. 14.3%; OR: 0.91, 95% CI: 0.84–0.99, P=0.04). On pre-specified subgroup analysis, the difference in the primary outcome was greater among patients who received larger volumes of fluid. In patients with sepsis, the 30-d mortality was lower with a buffered crystalloid compared to normal saline. The hospital mortality was not different between the two groups. The requirement for RRT and the presence of persistent renal dysfunction were also not different between groups. Furthermore, there was no difference in the ICU-free days, ventilator-free days, vasopressor-free days, and AKI stage II or higher after enrolment between the two groups. The SMART-trial used either plasmalyte-A or lactated Ringer’s solution in the balanced crystalloid arm; hence the specific impact of each particular type of fluid could not be evaluated. Other limitations include the relatively small volume of fluid administered, with a median of 1.0 L in the crystalloid group and 1.02 L in the normal saline group. Besides, it was a single-center, unblinded study. The composite primary outcome may also have its limitations.
The SALT-ED trial
The SALT-ED trial also compared buffered crystalloids (plasma-Lyte A or lactated Ringer’s solution) with normal saline in a randomized controlled trial.7 The study included consecutive patients who received crystalloids in the emergency department of the Vanderbilt University Hospital and were hospitalized outside the ICU. Patients who were admitted to the ICU were included in the SMART study. The number of hospital-free days at 30 days, the primary outcome, was not significantly different between patients who received crystalloids compared to those who received normal saline. The incidence of major adverse kidney events at 30 days was significantly lower with a buffered crystalloid compared to normal saline (4.7% vs. 5.6%; OR: 0.82, 95% CI: 0.7–0.95 p=0.01). The hospital mortality was not significantly different between groups. Among patients who presented with AKI of KDIGO stage II or greater, a significantly lower incidence of a major adverse kidney event was observed with the use of a buffered crystalloid. This was again, an unblinded study with the drawbacks of a composite outcome. Furthermore, data were not available regarding the type of fluid administered after initial care in the emergency department.
The BaSICS trial
The BaSICS randomized controlled trial included patients from 75 ICUs in Brazil among patients who required fluid expansion at least on one occasion.8 In addition, one of the following criteria had to be met: age >65 years, hypotension, sepsis, respiratory support with mechanical ventilation (invasive or non-invasive ventilation or high-flow nasal cannula for at least 12 h), early signs of renal dysfunction, cirrhosis or acute liver failure. Those who were likely to require RRT within 6 h or already receiving RRT, and high (>160 mmol/L) or low (<120 mmol/L) serum sodium levels were excluded. Patients were randomized into four arms; plasmalyte 148 or normal saline was administered, each at 333 m/h or 999 ml/h. All fluid challenges, maintenance fluid, and drug dilution were carried out with the study fluid, for up to 90 days after randomization. The overall patient care was at the discretion of the treating physician. A total of 11052 patients were enrolled over approximately 3 years. The plasmalyte arm included 5230 patients, while 5290 patients were randomized to receive normal saline.
Patient characteristics were well balanced at baseline; 48.3% of patients were post-surgical. The primary outcome was the 90-d mortality. In the plasmalyte group, 1381/5230 (26.4%) patients had died at 90 days compared to 1439/5290 (27.2%) in the normal saline arm (hazard ratio: 0.97 95% CI: 0.9-1.05); the difference was not statistically significant. Among the secondary outcomes, the incidence of AKI requiring RRT on days 1,2,3, and 7 or during hospital admission was not different between groups. The composite outcome of AKI, stage II or greater by the KDIGO classification and mortality was not different on day 3 and day 7. The total SOFA scores on days 3 and 7 were also not different between groups. The SOFA scores under individual organ systems were not different on day 3. However, the incidence of a neurological SOFA score of >2 on day 7 was significantly higher with plasmalyte (32.1%) compared to normal saline (26%). This finding suggested a worse neurological status on day 7 among patients who received plasmalyte compared to normal saline. On pre-specified subgroup analysis, in patients with traumatic brain injury, the 90-d mortality was higher in the plasmalyte group compared to the normal saline group (31.3% vs. 21.1%). There was no significant difference in the 90-d mortality in subgroups of patients with AKI of KDIGO 2 or greater severity, sepsis as the underlying cause, surgical patients, and APACHE score of 25 or more. There was no difference between groups in the ICU or hospital mortality and the ICU or hospital length of stay. On post-hoc analysis, the composite outcome of 90-d mortality and the requirement for RRT was not different between groups. The authors concluded that among critically ill patients who require a fluid challenge, there was no difference in the 90-d mortality with plasmalyte compared to normal saline. There was a suggestion of harm among patients with traumatic brain injury, with higher neurological SOFA on day 7 and higher 90-d mortality.
The BaSICS study echoes the findings of the SAFE study, which also revealed higher mortality among patients with traumatic brain injury who received 4% albumin compared to normal saline.9
The BaSICS study is limited by the relatively small volume of fluid infused; a median of 1.5 L on day 1 and 2.9 L in the first 3 days. Perhaps, larger volumes of fluid may reveal a difference in outcomes. A large number of patients received the non-study fluid in the first 24 h prior to study enrolment which may have confounded the results. The sample size was calculated based on 35% 90-d mortality in the normal saline arm; however, the actual mortality was lower, which could have resulted in a slightly underpowered study.
- Normal saline has been extensively used as intravenous fluid for nearly two centuries; over the years, questions have been raised regarding possible adverse effects
- The high chloride content compared to plasma levels has been associated with hyperchloremia and metabolic acidosis
- Hyperchloremia has been implicated in the causation of AKI through the tubuloglomerular feedback mechanism; besides, concerns have also been raised regarding splanchnic hypoperfusion and coagulation abnormalities
- Randomized controlled trials that compared a buffered crystalloid solution with normal saline among critically ill patients have revealed conflicting results
- The multicentric BaSICS study does not reveal any significant benefit in clinical endpoints with a buffered crystalloid compared to normal saline among a general ICU population
- In traumatic brain injury, there is a consistent suggestion of possible harm with the use of a buffered crystalloid from the SAFE and the BaSICS trials. Although based on subgroup analysis, it may be appropriate to prefer normal saline over a buffered crystalloid in traumatic brain injury
1. Awad S, Allison SP, Lobo DN. The history of 0.9% saline. Clinical Nutrition. 2008;27(2):179-188. doi:10.1016/j.clnu.2008.01.008
2. Lewins, R. Injection of saline solutions in extraordinary quantities into the veins of malignant cholera. The Lancet. 1832;18:243-244.
3. Williams EL, Hildebrand KL, McCormick SA, Bedel MJ. The effect of intravenous lactated Ringer’s solution versus 0.9% sodium chloride solution on serum osmolality in human volunteers. Anesth Analg. 1999;88(5):999-1003. doi:10.1097/00000539-199905000-00006
4. Martin G, Bennett-Guerrero E, Wakeling H, et al. A prospective, randomized comparison of thromboelastographic coagulation profile in patients receiving lactated Ringer’s solution, 6% hetastarch in a balanced-saline vehicle, or 6% hetastarch in saline during major surgery. J Cardiothorac Vasc Anesth. 2002;16(4):441-446. doi:10.1053/jcan.2002.125146
5. Young P, Bailey M, Beasley R, et al. Effect of a Buffered Crystalloid Solution vs Saline on Acute Kidney Injury Among Patients in the Intensive Care Unit: The SPLIT Randomized Clinical Trial. JAMA. 2015;314(16):1701. doi:10.1001/jama.2015.12334
6. Semler MW, Self WH, Wanderer JP, et al. Balanced Crystalloids versus Saline in Critically Ill Adults. N Engl J Med. 2018;378(9):829-839. doi:10.1056/NEJMoa1711584
7. Self WH, Semler MW, Wanderer JP, et al. Balanced Crystalloids versus Saline in Noncritically Ill Adults. N Engl J Med. 2018;378(9):819-828. doi:10.1056/NEJMoa1711586
8. Zampieri FG, Machado FR, Biondi RS, et al. Effect of Intravenous Fluid Treatment With a Balanced Solution vs 0.9% Saline Solution on Mortality in Critically Ill Patients: The BaSICS Randomized Clinical Trial. JAMA. Published online August 10, 2021. doi:10.1001/jama.2021.11684
9. Finfer S, Bellomo R, Boyce N, et al. A comparison of albumin and saline for fluid resuscitation in the intensive care unit. N Engl J Med. 2004;350(22):2247-2256. doi:10.1056/NEJMoa040232