Selection Committee Member: Charlie Wray @WrayCharles
Charlie Wray is an academic hospitalist at the San Francisco VA Medical Center and Assistant Professor of Medicine at the University of California, San Francisco. As a clinician-investigator, his research program focuses on social risk factors’ effects on health outcomes and resource utilization.
Writer: Pascale Khairallah @Khairallah_P
Pascale Khairallah is a second-year nephrology fellow at Columbia University Medical Center. She is part of the inaugural class of AJKD Editorial Interns. She is interested in studying the relationship between mineral bone disorders and cardiovascular disease in kidney disease patients.
Competitors for the Hospitalist Region
Perioperative ACEi/ARB vs PRN Hydralazine
This matchup pits two very common questions that come up in the hospital:
- Is there any need to hold angiotensin-converting enzymes inhibitors (ACEi) or angiotensin receptor blockers (ARB) prior to, or immediately after, surgery?
- What is our target blood pressure (BP) for hospitalized patients? And is giving “prn” (”as needed”) doses of IV hydralazine for systolic BPs over 170 helpful to patients?
ACEi, specifically lisinopril, are the most prescribed antihypertensive medications in the US with ARB (losartan) coming in third (just behind amlodipine). In one large study of non-cardiac surgery, about one third of patients were on these medications preoperatively. While the benefits of starting and maintaining ACEi/ARB for treatment of hypertension and management of CKD (especially with proteinuria) is well-known, whether ACEi/ARB should be continued perioperatively is less known. A similar debate exists about advanced CKD. Moreover, does the type of surgery (cardiac versus non-cardiac) inform perioperative management of ACEi/ARB?
When physicians are surveyed about their perioperative management of ACEi/ARB, one thing is clear: there is wide variability among physicians regarding when to stop ACEi/ARB preoperatively and when to restart them postoperatively. This is not surprising as the guidelines do not give a consistent answer:
- The 2014 ACC/AHA Guidelines state that continuation of ACEi/ARB is reasonable perioperatively and that it is reasonable to restart them postoperatively as soon as clinically feasible.
- The Canadian Cardiovascular Society Guidelines on perioperative cardiac risk assessment and management for patients who undergo noncardiac surgery recommend withholding ACEi/ARB 24-hours before noncardiac surgery and restarting ACEi/ARB on the second day after surgery, if the patient is hemodynamically stable.
- The 2014 European Society of Cardiology gives somewhat more detailed recommendations, favoring continuation of ACEi/ARB during non-cardiac surgery in stable patients with heart failure and LV systolic dysfunction, considering initiation of ACEi/ARB at least 1 week before surgery in cardiac-stable patients with heart failure and LV systolic dysfunction, and transient discontinuation of ACEi/ARB before non-cardiac surgery in hypertensive patients.
The reason for these diverse recommendation is that the evidence base was composed primarily of retrospective and observational studies.
In a meta-analysis of 25 studies including 6,022 patients undergoing non-cardiac surgery, there was no difference in mortality between patients who withheld or continued ACEi/ARB preoperatively. However, the risk of intraoperative hypotension was significantly increased in patients who continued ACEi/ARB with an incidence of nearly 30% (although the definitions of hypotension differed among the studies and the duration of hypotension was not reported). Continuation of ACEi/ARB did not affect postoperative hypotension rates. While we would all be very interested in knowing whether continuation or discontinuation of ACEi/ARB affects the rates of major cardiovascular (CV) events, myocardial infarctions, congestive heart failure, and acute kidney injury (AKI) postoperatively, the studies that have evaluated this are underpowered, preventing any conclusions.
The VISION (Vascular events In noncardiac Surgery patIents cOhort evaluatioN Prospective Cohort) study is an international prospective cohort analysis of 4,802 participants undergoing non-cardiac surgery at 12 centers in 8 countries. In a secondary analysis of patients on chronic ACEi or ARB, when the drug was held for 24 hours preoperatively there was a relative risk of 0.82 for 30-day all-cause mortality, stroke, or myocardial injury. Participants additionally had a 20% relative reduction in the risk of intraoperative but not postoperative hypotension. The investigators noted that they could not find consistent clinical factors that predicted whether a patient continued or held ACEi/ARB so the decision was likely an expression of physician preference.
The PREOP-ACEI (Prospective Randomized Evaluation of Preoperative Angiotensin-Converting Enzyme Inhibition) study is one of the few randomized controlled trials (RCTs) that sought to answer the question of whether holding ACEi preoperatively favorably affected BP both intra- and postoperatively in non-cardiac surgery. 275 patients were randomized to either stopping or continuing ACEi preoperatively. Rates of both intraoperative hypotension (defined as SBP < 80 mm Hg) and postoperative hypotension (defined as SBP < 90 mm Hg) were significantly lower in the group that stopped ACEi (P = 0.03 and P = 0.02, respectively). There were no patient deaths in either group.
More recently, a randomized trial of patients undergoing nonemergent cardiac surgery evaluated the effect of continuing versus holding ACEi/ARB for 24 hours preoperatively. These were 121 patients undergoing CABG or valve surgery and who had been on ACEi/ARB for at least 7 days prior to surgery. They were randomized to either continue or stop their ACEi/ARB. Postoperative shock, vasodilator use, vasoplegic shock (defined as a mean arterial pressure < 60 mm Hg requiring vasopressor administration for at least 4 hours and a central venous pressure ≥ 8 mm Hg), median duration of vasopressor or inotropic agent use, and intravenous (IV) vasodilator requirements were not significantly different between the two groups. Moreover, no differences were observed in the incidence of AKI, stroke, or mortality. Interestingly, the authors performed a sensitivity analysis stratifying results by ACEi/ARB’ half-life, which did not alter the findings.
A 2016 Cochrane analysis attempting to answer the question of continuing versus holding preoperative ACEi/ARB before either cardiac or non-cardiac surgery could not give definitive conclusions since the included studies had poor methodology, high risk of bias, and lack of power to answer the question.
The only real conclusion we can make out of this evidence is that we need large RCTs to understand the effects of ACEi/ARB continuation and discontinuation in the setting of both cardiac and non-cardiac surgery on intraoperative and postoperative hypotension, adverse renal and cardiac outcomes, and mortality. We also need to learn whether ACEi/ARB with different half-lives have a different effect on these outcomes, and whether it is the ACEi/ARB themselves or their combination with other antihypertensives that may be deleterious. If we decide to stop the ACEi/ARB, then we need to know when to resume them.
Stopping ACEi/ARB in the perioperative period is an easy and cost-free intervention that may have substantial effects on the outcomes of people undergoing surgery everyday. Whether it is the right thing to do or not is unfortunately still unclear. We need to have better answers about this for our patients.
Many of us are guilty of prescribing as-needed (PRN) IV or oral medications for acute control of elevated BP in patients admitted to the hospital. The PRN antihypertensive du jour most likely reflects institutional preference, availability of the medicine on the floor (vs. having to order it from a central pharmacy), and national drug shortages at the time of prescription. In this section, we aim to answer a number of questions regarding the practice of prescribing PRN antihypertensives:
- Why do we prescribe PRN antihypertensives?
- What are we treating?
- Do we have a certain BP threshold at which we prescribe PRN antihypertensives?
- How do we choose our treatment thresholds?
- Are we treating numbers (and nurses and physicians) or are we treating an important clinical condition?
It is well-established that effective treatment of chronic hypertension results in improved CV and neurovascular outcomes. The last several decades have witnessed impressive reductions in cardiovascular disease related to effective treatment of elevated BP. The original Veterans Administration Trial published in 1967 established the importance of BP control. More recently, the SPRINT trial published almost 50 years later continues to show how mortality is improved with lowering BP. This science has convincingly shown that comprehensive outpatient management of high BP for prolonged periods of time reduces catastrophic CV end-points and mortality. But none of this science has addressed whether treating an isolated elevated high BP discovered at 1:30 AM in a hospitalized patient with a dose of hydralazine (or any other drug du jour) has any benefit at all.
While hypertension is primarily relegated to the outpatient clinic, hypertensive emergencies necessitate hospital admission. So, what is a hypertensive emergency anyway? A hypertensive emergency is defined as a systolic blood pressure (SBP) > 180 mm Hg and/or diastolic blood pressure (DBP) > 120 mm Hg associated with evidence of new or worsening target organ damage. The association with end-organ damage is essential for the diagnosis.
Hypertensive emergencies are associated with high mortality rates in the first year of diagnosis and a median survival of 10.5 months when left untreated. Patients with hypertensive emergency typically require the use of IV antihypertensive therapy to decrease the BP acutely. Recommendations (not based on high-quality evidence) are to decrease BP by no more than 25% from baseline within the first hour, as any further drop can result in organ hypoperfusion, worsening organ damage. There is no consensus as to which drug should be used as first-line agent for the treatment of hypertensive emergency. Calcium channel blockers, direct vasodilators, nitric-oxide dependent vasodilators, adrenergic blockers, and ACEi/ARBs can all be used.
However, there are contraindications for use of certain agents in specific situations. For example:
- Nitroglycerin has a fast-onset of action and is easily titratable but is contraindicated in patients with volume depletion due to risk of hypotension.
- Hydralazine has a quick onset of action, but its hypotensive effect can be unpredictable and difficult to regulate. A single dose can last up to 12 hours, making it difficult to titrate in response to BP changes.
- Labetalol is effective in lowering BP, but it can lead to life-threatening bradycardia. Moreover, it is contraindicated in patients with acute decompensated heart failure or bradyarrhythmias.
In contrast to hypertensive emergency, hypertensive urgencies are defined as elevated BP > 180/120 mm Hg without evidence of organ damage. While guidelines and expert opinion agree that hypertensive emergencies should be treated quickly with IV medications to reduce the BP, patients with hypertensive urgency generally do not need to be hospitalized for management of their elevated BP. The American College of Emergency Physicians (ACEP) recommends against giving IV meds to treat asymptomatic uncontrolled hypertension. However, a significant number of patients presenting to the ED continue to receive IV antihypertensives to acutely decrease their BP. What are the reasons for doing this? Is it because physicians are obsessed with attaining an acceptable BP? Are they worried the patient won’t be accepted to an inpatient floor bed? Are they concerned about litigation?
Most housestaff have been called at some point about inpatients with elevated BP. Greater than 40% of providers prescribe PRN medications to inpatients with elevated BP. What is particularly worrisome is the randomness of this practice. In one study evaluating the use of IV hydralazine in hospitalized patients, only 7.5% of the patients who were prescribed hydralazine were actually evaluated by a provider prior to the medication being prescribed (and of those evaluated, only 2% had hypertensive emergency symptoms). This means that the providers who prescribed the medication did not confirm that a hypertensive emergency (vs. urgency) was occurring, and simply gave the medication to treat a number. In that same study, the mean BP of patients at the time hydralazine was prescribed was 175/82 ± 25/16 mm Hg, therefore half of the patient’s’ BP did not meet hypertensive urgency definitions. In a different study evaluating the same question, 84.5% of patients who received IV antihypertensive therapy had SBP <180 mm Hg. This indicates that the BP thresholds for which providers prescribe PRN medications are variable, inconsistent, and do not adhere to guideline recommendations.
As evidenced by these and other studies, not only are the thresholds for prescribing PRN antihypertensives haphazard, but the formulations (oral vs IV) and doses of the prescribed medications are disparate as well. Beyond the indications for these drugs, the effectiveness of this treatment is questionable. In a review of over 2 million office visits, where nearly 60,000 (4.6%) met criteria for hypertensive urgency (mean BP 182/96), propensity matching to look at the outcomes of patients who went to the emergency department (ED) versus those who were sent home showed no difference in the rate of major adverse cardiac events at 7 days, 30 days, or 6 months. Sending the patient to the ER for urgent treatment for their hypertensive urgency didn’t help the heart.
The inpatient setting is a minefield of possible causes of acute hypertension. These include holding antihypertensives at the time of admission, pain, anxiety, and volume overload among others. Poor inpatient BP control may also simply reflect poor outpatient BP control. In one study of patients who were given PRN medications for BP control, 40% had not been continued on their home medications in the hospital, while 32% of patients who were continued on their home regimen had them restarted only 2 or more days after admission.
Avoiding unnecessary treatment of isolated elevated BP levels in ED or hospitalized patients is particularly important because:
- There is no evidence to suggest that lowering BP in the acute setting is clinically beneficial (except in hypertensive emergencies, which was discussed earlier).
- Acute treatment may be harmful. Up to 8% of doses of IV hydralazine are associated with adverse events, with hypotension being the most common. Rarely, patients need vasopressor therapy to support the BP. Patients also report dizziness and lightheadedness, even when they are not overtly hypotensive. Even modest drops in BP in patients who are chronically hypertensive can have deleterious effects due to poor organ perfusion (cerebral, myocardial, renal etc.). One can also speculate about other side effects that can result from dizziness and hypotension, including falls, poor mobility, delayed rehab, etc.
So when is PRN antihypertensive therapy justified for modest isolated asymptomatic elevations of the BP in the ED or hospitalized patients? Maybe never. Severe hypertension, especially with symptoms of end-organ damage, requires acute treatment, of course. As for severe, but asymptomatic hypertension, we should:
- Assess the patient for hypertensive emergency symptoms. This includes the often forgotten fundoscopic examination and urinalysis.
- Ensure that outpatient antihypertensive regimens have been continued or resumed unless contraindicated.
- Optimize volume status.
- Intensify the regimen being used in the hospital if indicated.
- Address medication compliance.
- Address other situational causes of elevated BP such as stress, anxiety, or drug use.
Finally, it is important to ensure that patients have adequate outpatient follow-up prior to their discharge from the ED or hospital. The STAT (Studying the Treatment of Acute Hypertension) cross-sectional study showed that only 35% of patients admitted with acutely elevated BP had follow-up scheduled prior to discharge or followed up with a provider after discharge.
It is important for healthcare providers, particularly residents and nurses, to be aware of the lack of benefit and the potential harm in prescribing PRN medications for acute BP lowering when it isn’t really necessary. Given the widespread practice of administering PRN agents to lower BP in asymptomatic patients, more research is needed and medical professional societies should address this issue in their guidelines to raise awareness and hopefully eliminate this potentially harmful practice.
Lactated Ringer’s vs Normal Saline
This is the classic matchup, and has been done before in NephMadness (we even had an editorial from Skeptical Scalpel). Now we have data from the SALT-ED and SMART trials looking at outcomes in critically ill and non-critically ill patients receiving normal saline or lactated Ringer’s solution. Internists have traditionally been fans of normal saline, but the chloride in normal saline may actually be a bad thing. This matchup needs to look at the data carefully.
After determining that crystalloids are at least as good as, if not superior, to colloids in septic patients, the more recent question has been whether one type of crystalloid solutions is better than another. There are two types of crystalloid fluids: saline and balanced solutions (e.g., lactated Ringer’s, Plasma-Lyte A, Hartmann’s solution). Balanced crystalloids have sodium, chloride, and potassium concentrations close to those of extracellular fluid and contain a buffer source such as lactate or acetate. They may also have physiologic concentrations of magnesium and calcium.
Let’s look at the two head-to-head competitors…First, the challenger:
Lactated Ringer’s (LR)
There have been several observational studies linking normal saline (NS) to a higher risk of AKI and hospital mortality compared to balanced crystalloid solutions. This difference has been attributed to the higher chloride concentration of NS and the hyperchloremia that results from its use.
Concerns about the use of IV saline first arose in the early 1980s when the administration of NS was linked to the development of metabolic acidosis. The link between NS and metabolic acidosis is best explained by The Stewart Approach. According to Stewart, acid-base status in extracellular fluids is dependent on the partial pressure of CO2, the strong-ion difference (SID), and the total amount of weak acids such as albumin. SID indicates the excess between strong positively charged ions and strong negatively charged ions, SID = strong cations – strong anions. The positively charged ions include Na+, K+, Ca2+, Mg2+. The negatively charged ions include Cl–, and lactate and ketones when those are present.
SID determines unmeasured anions in a solution. The SID of extracellular fluids is always positive at 40 mEq/L. A lower SID suggests the presence of an acidosis, and a higher SID indicates the presence of an alkalosis. The SID of NS is zero because the sodium (positive ions) and chloride (negative ions) concentrations are equal. Adding NS to the extracellular fluid would then shift the positive SID of extracellular fluid towards zero resulting in an acidosis. This is in contrast to LR whose SID is 28 mEq/L and therefore lowers the extracellular fluid SID to a much lower extent than NS, largely avoiding the development of non-gap metabolic acidosis. The metabolic acidosis that ensues from the lowering of SID can result in vasodilation which may worsen shock, and can induce inflammation which results in coagulopathy and multiorgan damage.
The supraphysiologic concentration of chloride in NS is additionally reported to be associated with several deleterious effects. Drops in mean arterial pressure (MAP) in sepsis have been found to correlate more strongly with hyperchloremia than with the degree of pH reduction. Moreover, chloride is associated with kidney dysfunction. Early experiments showed that increased chloride delivery to the macula densa in the distal tubule results in subsequent release of adenosine and vasoconstrictors from the macula densa cells. This “tubuloglomerular feedback” leads to afferent arteriolar vasoconstriction and a reduction in GFR. Chloride also induces thromboxane release which can increase the resistance of both afferent and efferent arterioles, reducing GFR further. Animal studies also suggest that hyperchloremia enhances angiotensin II intrarenal vasoconstriction. MRI imaging of the kidneys during fluid administration shows that saline but not balanced crystalloid infusions lead to significantly decreased renal blood flow and cortical perfusion. All this put together suggests that in patients with low GFR secondary to hypotension, shock, or renal failure, supraphysiologic extracellular chloride concentrations can negatively impact GFR through lowering MAP and through afferent arteriolar vasoconstriction.
More recent observational studies of critically ill patients showed that patients receiving balanced crystalloids had lower rates of AKI and were less likely to receive renal replacement therapy (RRT) compared to patients who received NS. Further studies found that this higher chloride load also resulted in increased rates of mortality, even after adjustment for volume of fluids given and severity of illness. While some of these observational studies suffered from bias and other limitations, they did lay the groundwork for a host of clinical randomized trials. Most of these earlier clinical trials were done in surgical patients. While the risk of development of acidosis and hyperchloremia with saline was clear in those studies, they failed to show a reduction in AKI or mortality with balanced solution. However, more recently, a double-blinded randomized trial comparing acetate-balanced crystalloid to saline in patients undergoing major abdominal surgery was terminated early because of a significant increase in the number of patients receiving saline who needed pressors as compared to those who received balanced crystalloids (97% vs 67%, P = 0.03).
Yunos et al performed a prospective open label pilot study in ICUs at a tertiary care center in Australia. During the intervention period, patients were given a balanced crystalloid unless the primary physician favored saline. Primary outcomes included increase in creatinine from baseline to peak ICU level and incidence of AKI per the RIFLE criteria. Secondary post hoc analysis outcomes included the need for RRT, length of stay in ICU and hospital, and survival. Baseline mean creatinine of participants was 1 mg/dL. Patients had a relatively low mean Apache II scores of 15-16. Use of balanced crystalloids was associated with a decrease in risk of kidney injury (P < 0.001) and need for RRT use (P = 0.004). ICU length of stay and hospital length of stay were not different between the groups. Finally, mortality and need for RRT after hospital discharge were not statistically different between the control and the intervention groups.
The SMART trial, published in the NEJM in 2018, was a cluster-randomized, multiple-crossover trial conducted in five intensive care units at Vanderbilt University Hospital. 15,802 patients received either 0.9% sodium chloride or IV isotonic crystalloid. The balance crystalloids included lactated Ringer’s solution or Plasma-Lyte A. Baseline characteristics were comparable between patients, including baseline median creatinine which was 0.89 mg/dL. Patients received a median volume of 1000 ml (IQR, 0 to 3210mL) of balanced crystalloids and a median volume of 1020 ml (IQR, 0 to 3500 mL) between ICU admission and hospital discharge or 30 days (whichever came first). 14.3% of patients in the balanced-crystalloids group and 15.4% in the saline group had a major adverse kidney event as defined by a composite of death, new RRT, or persistent renal dysfunction (defined as final serum creatinine concentration ≥ 200% of the baseline value) at hospital discharge or at 30 days after the index emergency department (ED) visit. Mortality and the need for RRT were not statistically different between the groups.
The SALT-ED trial looked at this comparison between balanced crystalloids and saline in patients outside the ICU. The trial was a single center, unblinded trial performed in the emergency department. The type of crystalloid administered was determined by the calendar month. SALT-ED enrolled a total of 13,347 patients over 16 months. Patients with brain injury or hyperkalemia received saline based on the provider’s clinical judgement regardless of month. Baseline characteristics were comparable between patients. Patients received a mean of 1.6 +/- 1 L of balanced crystalloids and a mean of 1.6 +/- 1.1L of saline. The main outcome was major adverse kidney events within 30 days (MAKE 30) which was a composite of death, new RRT, or persistent renal dysfunction (final serum creatinine concentration, ≥ 200% of the baseline value) at the earliest of hospital discharge or 30 days after the index ED visit. Patients who received balanced crystalloids had a lower incidence of MAKE 30 within 30 days of admission as compared to patients who had received saline (4.7% vs. 5.6%; adjusted odds ratio, 0.82; 95% CI, 0.70 to 0.95; P=0.01). Patients with serum creatinine ≥ 1.5 mg/dl or hyperchloremia had the largest benefit from balanced crystalloids. Other clinical outcomes including hospital-free days, stage 2 or 3 AKI, and in-hospital death did not differ between groups.
The SMART and SALT-ED trials have limitations. Similar to some previous trials, the amount of fluids administered to patients was small. While MAKE 30 was significantly lower (though absolute difference was ~1%) in the balanced crystalloid vs. saline group in SMART, it does not inform us about the individual components of MAKE 30, nor does it allow us to make inferences about whether hyperchloremia, other electrolyte imbalance, or other factors altogether (such as inflammation, coagulopathy, volume overload, etc.) could have contributed to this observed effect.
Physicians fear using LR in certain situations, particularly in hyperkalemic patients and in lactic acidosis. Let us address these concerns:
- Hyperkalemia: Lactated Ringer’s has a potassium concentration of 4 mEq per liter. Total body potassium is approximately 55 mEq/kg (3850 mEq in a 70 kg man). The vast majority (99%) of that potassium is intracellular with only an odd 50 mEq or less found extracellularly. The potassium administered in LR will almost all distribute to the intracellular compartment. This is similar to what happens when patients are given IV infusions of KCl to correct hypokalemia. Giving 40 mEq of KCl raises the serum potassium by 0.4 mEq. By using standard estimates for total body water and distribution between the intracellular and extracellular compartment, an increase in serum potassium of 0.4 requires only 5.6 mEq of potassium to remain in the extracellular compartment. 86% of the potassium moves intracellularly. If we presume LR acts the same way (and we have empiric data to suggest this is the case), only 0.6 mmol of potassium would remain in the extracellular compartment for every liter infused.
Contrast this with NS, which induces a normal anion gap metabolic acidosis, which may cause a shift of potassium from the intracellular compartment to the extracellular space, potentially worsening hyperkalemia.
- As for the lactic acidosis argument, one should not confuse lactic acid with lactate. Hospital assays evaluating blood lactic acid levels measure lactate levels.
Lactate is a byproduct of anaerobic glycolysis; it is the accompanying hydrogen ion in lactic acid, not lactate itself which causes the acidosis. Lactate is largely metabolized by lactate dehydrogenase to pyruvate. 60% of this metabolism occurs in the liver and around 30% occurs in the kidneys. In the absence of severe liver dysfunction, L-lactate can be metabolized at rates of 100 mmol/hour or more. LR contains lactate, not lactic acid. When healthy volunteers are administered LR and NS, they have a similar small increase in lactate levels. In liver failure, there may be a more significant increase in serum lactate levels after LR administration. This does not have adverse effects per se but may confuse the interpretation of lactate testing (i.e. even in well-perfused states, patients may have falsely elevated lactate levels).
Now let’s hear from the traditionalists…
Normal Saline (NS)
IV fluids are thought to have been developed during the European cholera pandemic in the early 1800s. The original solution contained 134 mEq/L of sodium,118 mEq/L of chloride, and 16 mEq/L of bicarbonate. The term “normal saline” was first introduced in 1888 to describe a solution containing 150 mEq/L of sodium, 128 mEq/L of chloride, and 27 mEq/L of bicarbonate. The composition of these solutions was modified through the years.
It was not until the early 1900s that H.J. Hamburger did in vitro studies in humans and mammals, and concluded that the now known 0.9% saline solution was isotonic and physiologic. His conclusion was accepted by the medical community who continues to use NS to this day. NS is one of the most commonly administered solutions. It is cheap to produce, is widely available, and is compatible with blood product infusions as it does not contain any calcium. However, NS has supraphysiologic concentrations of both sodium and chloride, leading to much criticism about its widespread use.
Is NS a safe crystalloid? And is its widespread use justified?
The SPLIT randomized clinical trial (Effect of a Buffered Crystalloid Solution vs. Saline on Acute Kidney Injury Among Patients in the Intensive Care Unit) trial was published in JAMA in 2015. It was a cluster randomized, double-crossover pilot trial comparing 0.9% saline vs. Plasma-Lyte in 2,278 patients admitted to ICUs in New Zealand. The study excluded patients who were on RRT or were expected to require RRT within 6 hours of admission to the ICU. Most patients who entered the SPLIT trial had been admitted to the ICU following elective (mostly CV) surgery. Moreover, the APACHE scores of the patients were quite low at 6.9 and 6.7. Both groups received a median of 2 L of fluids. The primary outcome was the development of AKI (KDIGO AKI stage 3). AKI occurred in 9.6% of patients receiving balanced crystalloid and 9.2% of those receiving saline (P = 0.77). 3.3% patients receiving balanced crystalloid vs. 3.4% of patients receiving saline required RRT (P = 0.91). In-hospital mortality and mortality at 90-day follow-up was not significantly different between groups. The APACHE score of patients in this trial predicts an ICU mortality of only around 8%. Similarly, the rate of AKI is <10%, also much below the 20% rate seen in ICUs in general. So one has to wonder whether SPLIT’s patient population is adequately representative of the typical academic medical center ICU population. Moreover, the low amounts of fluids delivered could have masked the effects of 0.9% saline and hyperchloremia.
The SALT trial followed. SALT was a cluster-randomized, multiple-crossover trial of 974 patients admitted to the ICU. 25% vs. 28.6% of patients assigned to saline vs. balanced crystalloids had sepsis or septic shock. 21.9% vs. 24.4% of patients were on vasopressors. By day 7 of hospitalization, patients had received a median of 1,250 ml [390–3,000] of saline vs. a median of 1,320 ml [435–3,139] of balanced crystalloids (P = 0.38). Serum chloride levels were significantly different between groups (median [IQR], 109 mmol/L [105–113] in saline group vs. 108 mmol/L [104–112] in balance crystalloids groups, P = 0.03), though in clinical practice one wonders about the significance of this result. There were no differences between groups in incidence of AKI stage 2 or greater or receipt of RRT. Similar to SPLIT, the amount of IV fluids administered was lower than expected in an ICU setting, possibly preventing us from seeing any deleterious effects of hyperchloremia.
A more recent blinded, 2-arm parallel allocation, randomized controlled trial aimed to investigate whether the choice of NS vs. LR in stable patients in the ED influences patient reported outcomes. 157 participants were enrolled. Patients received 2 L of NS or LR. Patient-reported outcomes were collected following ED discharge using Quality of Recovery–40, a validated survey. At 24 hours post-discharge, the results of Quality of Recovery–40, patient comfort, emotional state, physical independence, and pain were similar between both groups. Seven days post-discharge, the number of patients who returned to the ED, who required care by other providers, or who filled a prescription was non-significantly different between groups. The study was small, with 40% dropout and mostly young patients who had few comorbidities (including only 11% having abnormal BUN and Cr), but it does provide some reassurance for the use of NS especially in a population similar to the study’s population.
Are there specific clinical situations when NS is needed?
Fluid management in patients with both hypotension and brain injury is difficult as physicians have to balance raising systemic BP with minimizing the rise in intracranial pressure (ICP). Avoiding hypotonicity is key in order to prevent cerebral edema. Hyperosmolar therapy for elevated ICP is associated with improved survival. NS with an osmolarity of 308 mOsm/L is more hyperosmolar compared to LR with an osmolarity of 278 mOsm/L, making it the more intuitive choice for physicians treating unstable patients with high ICP.
In the SAFE study, saline was superior to albumin in reducing mortality in patients with traumatic brain injury. Further, in hemodynamically stable patients with brain injury, neurologic outcomes at 6 months were not statistically different between the group that received saline and the group that received a hyperosmolar sodium lactate solution despite the former’s chloride load being significantly higher. Based on all this, the neurology and neurosurgery world has generally favored the use of NS as the resuscitation fluid of choice in hemodynamically unstable patients with brain injury. While the choice of NS over LR in these patients is based mainly on physiologic rationale and not on randomized trials, perhaps a head-to-head comparison between these two types of fluids is not warranted due to the expected neurologic harm that can be caused by the osmolarity of LR.
While we await the results of PLUS, BaSICS, the ASTRAU, and other trials to provide more definite answers to the balanced crystalloids vs. normal saline battle, who will our blue ribbon panel choose to move forward in this round?
How to Claim CME and MOC
US-based physicians can earn 1.0 CME credit and 1.0 MOC point for reading this region.
- Register/log in to the NKF’s Professional Education Resource Center (PERC).
- Review the activity and accreditation information.
- Click “Continue” and review Course Instructions.
- Complete Post-Test. Please note: By selecting “Yes” to the participation questions for each region, the corresponding Post-Test questions will appear. Click “Save Draft” to save your responses and finish later. When you are ready to submit your answers, click “Preview” to review all responses, then click “Submit.”
- Click “Next” to complete the Evaluation form, then click“Submit.”
- Claim 1.0 CME credit and 1.0 MOC point per region (up to 8.0 total for 8 regions of NephMadness).
- Print your certificate.
- Review the Post-Test answers and rationale.
The CME and MOC activity will expire on June 15th, 2019.