Response to 1L of normal saline

This chapter is relevant to Section I2(i) of the 2017 CICM Primary Syllabus, which expects the exam candidates to develop or at least fake "an understanding of the pharmacology of colloids and crystalloids". Specifically, the focus is the physiological response to the infusion of normal saline. Of the colloids and crystalloids, this is the fluid which has received the greatest attention in the CICM First Part Exam, among multiple SAQs:

• Question 17 from the first paper of 2018 (saline vs. dextrose)
• Question 1 from the second paper of 2015 (saline vs 4% albumin)
• Question 7 from the first paper of 2015 (saline solo)

As with magnesium sulfate and even water, the focus appears to be on the trainees' ability to force these IV fluids into a pharmacology answer template. Which is fine: structure is to be celebrated in all its forms. So:

Locally, normal saline is also discussed in several other chapters

• Normal saline, and why it is anything but "normal" is a description of the pharmaceutical of saline, with a focus on what's in the bag (i.e not much physiology)
• Normal saline vs balanced crystalloids as resuscitation fluid is something of a critical evaluation of the use of normal saline in critically ill patients, with advantages and disadvantages 
• Normal saline intoxication is a discussion of the normal anion gap acidosis which can occur with excess saline.

But realistically some peer-reviewed resources would be better. Hanh & Warner (2010), "Volume Kinetics for Infusion Fluids", is probably the best paper to describe not just the pharmacokinetics of these fluids but also the scientific rationale behind their movements. In fact pretty much anything by Robert Hahn is good.  If you can get a hold of it, the review of colloid and crystalloid pharmacology by Griffel and Kaufman (1992) is also excellent. 

Pharmacokinetics of normal saline

Presentation and administration

This 1 litre of 0.9% saline contains 150 mmol of sodium and 150mmol of chloride in sterile water. The pH of this fluid is often reported as something like 4.6, which is completely meaningless (Reddi, 2013).  The measured osmolality is 286 or so, for known reasons. It is designed for intravenous administration, but being a commonly available isotonic fluid makes it versatile, which means it can be given subcutaneously, as a neb, as a mouth wash, as an ocular rinse, and virtually in any other way you can think of. You can use it to flush an external ventricular drain. 

In their comments to Question 1 from the second paper of 2015, the examiners complained about nobody knowing how albumin was manufactured. As this was a compare-and-contrast question that asked for the trainees to juxtapose albumin and saline in a tabulated answer, the implication is that they also expected some sort of details about the industrial production of saline. Fortunately, that's a fairly straightforward one-step process. Sodium chloride has a molar weight of 58.44 grams, which means all you need to do is weigh out 150mmol of it (8.77 grams) and add that to a litre of sterile water. Or, more likely, add 877 grams of it to a ton of sterile water and then package it into litre containers in an automated packing machine.

Bioavailability and absorption

If administered orally, saline is rapidly and completely absorbed, as not only is it basically just water, but it contains sodium which is avidly absorbed in the proximal small bowel, creating the osmotic gradient necessary for water absorption. 

Distribution

The saline distributes rapidly from the intravascular to the interstitial compartment. According to Hanh & Warner (2010), this phase takes about 25-30 minutes to complete. 

And there, in the interstitial compartment, it is stuck. The cells have active pumps and the membrane is impermeable to sodium; similarly, the chloride is trapped. Because normal saline is isoosmotic with the extracellular fluid, water does not have any osmotic pressure to shift between compartments, and it distributes itself according to the proportional distribution of sodium (i.e. about 25% of it stays intravascular and 75% enters the interstitial fluid). To be frank this is not entirely accurate, as the osmolality of the extracellular fluid increases and some water undergoes a shift into the extracellular compartment. The new equilibrium point is 290.2 mOsm/L. To achieve this, a whole 25ml of water has to move up into the extracellular compartment. So, the intravascular volume increases not by 250ml, but by... 256ml.

Thus, in short, normal saline has a volume of distribution that resembles the volume of the extracellular fluid, which would be about 1/3rd of total body water, or about 0.2 L/kg.  Neither the sodium nor the chloride are especially protein bound, and the pKa of sodium chloride is 3.09 which means they should be fully dissociated at body fluid pH. 

Elimination

Water, sodium and chloride are all eliminated via the kidneys by mechanisms which are tightly regulated by the combined effects of vasopressin and angiotensin/aldosterone systems. In short, the elimination of the litre of saline will typically be rather rapid in the euvolaemic patient, as their homeostatic mechanisms will not perceive any need to hang on to the volume. The half-life of normal saline under these circumstances is usually described as 20-40 minutes (Hahn, 2016), though it is approximated as 2-4 hours for critically ill patients, and up to 8 hours in mechanically ventilated or shocked patients. 

Pharmacodynamic effects

Circulating volume expansion

Though it is said that normal saline distributes into the extracellular fluid and leaves behind only about 25% of its original volume, the effect of the infusion itself is somewhat larger. "Plasma volume expansion at the end of a brisk 30-min infusion is 50–75% larger than would expected if distribution of fluid [between compartments] had been immediate", report Hanh & Warner (2010). However, this is not a lasting effect.  Lobo et al (2003) infused a bunch of healthy males with 2 litres of saline over 1 hour, and observed that the fall in haematocrit and serum albumin produced by this took about 6 hours to resolve. After that, you are back to where you started from in terms of intravascular volume.

Circulatory reflex responses to a change in circulating volume

The intravascular compartment volume increases by 250ml or so – from 5000ml to 5250ml. The increase in intravascular volume is 5% - outside the volume receptor sensitivity threshold for ADH release, which is regulated mainly by the baroreceptors.  Obviously your various baroreceptors are sensitive to much smaller changes in pressure and vessel or atrial stretch (for example in case of the Bainbridge reflex, something between 50 and 100ml is enough to change the afferent firing rate), but the degree to which the arterial pressure will change depends on a whole host of parameters. What exactly will happen haemodynamically depends to a considerable extent on what is already happening, i.e. whether the circulatory system is stressed and compensating for hypovolemia, or whether it belongs to a healthy volunteer.

As a good rule of thumb it would be important to consider that all of these reflexes are designed to preserve homeostasis, i.e. their role is to keep everything unchanged as much as possible. This is well demonstrated in the data produced by Ukor et al (2017), who measured the haemodynamic effects of a 1L saline bolus in healthy volunteers (the purpose of the study was to compare it to a slow infusion); as you can see the haemodynamic parameters after two hours were essentially unchanged. 

Changes to the plasma osmolality, and their consequences

Plasma osmolality doesn't change much, because it has received a load of essentially isoosmolar fluid. You will notice that the osmolality of the compartments increases by 0.2 mOsm/kg. The osmoreceptor organs will definitely notice, as they are incredibly sensitive to changes probably lower than the  1% (2.8 mOsm/kg) minimum tonicity increment quoted in textbooks, but whatever minor adjustment they make to the vasopressin secretion rate will generally go unnoticed. For this reason, saline is viewed as a safe control fluid for experiments that look for vasopressin secretion changes in response to different hypertonic infusions, such as this study by Rittmaster et al (1987)

Changes to plasma oncotic pressure

If neither baroreceptors nor osmoreceptors react to the trivial changes in body volume and osmolality, does that mean that your organism simply neglects to notice this fluid bolus? Of course not. The detected change is not the change of osmolality or intravascular volume, but rather the oncotic pressure. Because intravascular protein concentration decreases, plasma oncotic pressure decreases.

Glomerulotubular balance ensures a return to homeostasis by increasing free water excretion (the mechanism is triggered by a decrease in peritubular capillary oncotic pressure; it decreases the rate of water resorption from the proximal tubule). This autoregulatory mechanism normally ensures that changes in the glomerular filtration rate don't alter the rate of sodium and water excretion. The peritubular capillary carries blood from the glomerulus, where ultrafiltration had concentrated the blood. The degree to which this blood was concentrated now determines the degree of water reabsorption.

Reabsorption leads to dilution of peritubular capillary blood, and thus oncotic homeostasis is maintained. If this blood happens to already be dilute (eg. after your saline bolus) the rate of water resorption will decline, and more water will be excreted in the urine. As more water is excreted the oncotic pressure will gradually return to normal.

Neurohormonal reactions

From reading the above, the reader may come to the impression that the 1L bolus of saline goes completely unnoticed by all the major homeostatic control systems, as if it has dropped into a bottomless void, and only the glomerulotubular feedback apparatus is held accountable for its ultimate removal from the body.  Of course basic logic indicates that this could not possibly be the case. To be sure, vasopressin secretion remains largely unchanged, but through the sensitivity of haemodynamic sensor mechanisms (eg. baroreceptor reflexes and atrial stretch receptors) other neurohormonal mechanisms do come into play. Without going into too much detail:

• Renin release decreases, as there is an increased delivery of sodium to whatever mysterious salt-sensing apparatus governs the release of renin (possibly, the macula densa)
• Angiotensin II synthesis decreases because it is linked to renin activity
• Aldosterone release decreases, as it relies on angiotensin II 
• As the result, the reabsorption of sodium and water in the distal nephron is decreased
• Natriuretic peptide release increases, as atrial stretch stimulates their release, which in turn stimulates natriuresis and diuresis

In short, the neurohormonal reactions to a saline bolus are the main medium-term compensatory mechanisms that work to restore the internal milieu to its pre-bolus state. These are all probably better off as a diagram. Again borrowing from the healthy volunteer data by Lobo (2010), and plotting it all on the same set of coordinates, the responses mentioned above can be represented like this:

Note that the vasopressin levels here do fluctuate more than one might expect, and this is possibly due to the fact that the renin angiotensin aldosterone axis has a role to play in modulating the secretion of vasopressin. Even though the baroreceptors might not command it, some change in vasopressin release may still occur because of these other factors.

Changes to plasma biochemistry

The sodium, which is slightly higher in saline than it is in normal blood, might be expected to rise slightly. To the total stores of extracellular sodium (lets say its concentration is 140 mmol/L, over 14 litres of ECF) we have just added another 150. Because we have also added 1000ml of water, the sodium concentration will only rise to 140.6 mmol/L, which is outside the error range of many laboratories. One may not even notice.  

Chloride, on the other hand, rises appreciably. If the chloride was 100 mmol/L, and we have just added 150 mmol, the extracellular chloride level will rise to 103. What's more, it's here to stay. Lobo's healthy volunteers still had elevated chloride at the end of the measurement period (8 hours). In order to maintain electroneutrality, with rising chloride levels more and more bicarbonate is lost (through the kidneys) and a normal anion gap metabolic acidosis develops.

With one bag, its not a big deal, but imagine for a second an emergency department with nothing but saline in stock, and a severely septic patient. After 6 litres of "goal directed" saline, one is left with an ECF volume expanded to 20 litres (little of which is intravascular, given the leaky capillaries). The serum sodium harmlessly rises to 143, but the chloride is now 115. An ICU consult is solicited for shock with metabolic acidosis refractory to fluid challenges. Hilarity ensues.

Inaccuracy in sodium calculations with saline infusion

The reader with some experience of managing sodium disturbances will at this stage raise some valid concerns. A couple of paragraphs above, this author's simplified calculations suggest that the serum sodium will rise by 0.6 mmol/L. However, if we use the well-weathered Adrogue-Madias formula, we will get a different value (140.3 mmol/L). If a higher concentration of sodium is used (eg. you give 1000ml of 3% saline, with 514 mmol/L of sodium) the discrepancy becomes even greater. This could be dangerous: underestimating the rate of replacement could give rise to all sorts of hideous neurological complications. Also, it throws doubt over the whole discussion: do you believe the calculations of a nameless intensivist blogger, or a well established tool for calculating sodium replacement?

Well, as it turns out, you should trust neither.  

Let us try to unravel the source of this discrepancy: As you throw a litre of saline into the system, it immediately increases the extracellular fluid volume by 1000ml, and the extracellular sodium by 150 mmol. At this stage, one might expect the sodium to remain exclusively in the extracellular fluid, intracellular sodium being a tightly controlled concentration. That would change the osmolarity of the extracellular fluid, causing some minor fluid shifts, which I have basically ignored.  

The Adrogue formula, however, expects the sodium and water to equilibrate between the compartments.

Surely, that can't be right. The mechanisms maintaining the Gibbs-Donnan equilibrium rely on the intracellular sodium concentration remining low, and though some sodium must sneak through into the cells, the cell membrane must surely act as more of a barrier-it should not be completely porous to the movements of this electrolyte.

Ergo, the Adrogue formula should underestimate the sodium increase following an infusion. But does it? In 2006, Liamis et al published a study looking at the sodium changes in 189 patients receiving saline (hypertonic in only a few cases), comparing the actual change with that which was predicted by the Adroque-Madias formula. Though the formula was good enough for government work, it really did seem to predict a much lower increase in sodium, particularly in the patients who had lower extracellular fluid volume: 

" ...in every subgroup, the achieved serum sodium was higher than the anticipated one, but the difference failed to reach statistical significance because of the low number of patients in each subgroup. The discrepancy was particularly marked and actually achieved statistical significance in the hypovolemic group, with the rise in serum sodium being two to three times larger than predicted by the formula. "

In short, it appears that faced with real clinical scenarios, the formula does really underestimate the sodium increase. There's a great article by Nguyen & Kurtz (2004) which goes through the limitations of the Adrogue-Madias formula with great matematical rigor, and basically their conclusions were that the equation fails to take into account several important factors such as the change of total body water with ongoing infusion and the normal expected ratio of intracellular to extracellular sodium. Not to mention factors completely external to the formula but still important clinically, like the urinary and GI losses of sodium, or gains of sodium through diet and medications, or the effects of turning off the ADH secretion by increasing volume. The effects of this inaccuracy would be amplified wherever the patient is so volume-depleted that the contributed infusion volume would contribute a significant fraction of the total body water, eg. where you've given 3000ml of saline to a 60kg dehydrated patient.

What's the upshot of all this? An editorial from the 2006 journal where Liamis published wisely counsels that "there is ultimately no substitute for the close monitoring of the serum sodium". In summary, no formula is sufficiently capable of predicting any individual sodium repletion with enough accuracy to allow you to prescribe a replacement regimen and then walk away irresponsibly.