Postoperative Hyponatremia despite Near-Isotonic Saline Infusion: A Phenomenon of Desalination

Methods

Patients with Acute, Fatal Postoperative Hyponatremia

We evaluated the medical records of five patients who died in hospitals in Canada (n = 3) and the United States (n = 2) to discover why severe hyponatremia occurred after surgery. We used data from these patients to define the problem rather than rigorously document the mechanism of postoperative hyponatremia. The volume and composition of intravenous and oral fluid intake, the plasma sodium and glucose concentrations at the time of cardiac arrest [4], and body weight had been measured in all patients. Urine had not been analyzed for electrolytes in any patient. We calculated the positive balance for electrolyte-free water that was necessary to induce the degree of hyponatremia that was seen using the following assumptions: 1) Preoperative total body water was 50% of body weight, and intracellular fluid volume was two thirds of body water [5] and 2) the number of particles in the intracellular fluid did not change in this acute setting. According to these assumptions, a woman weighing 70 kg would have 35 L of total body water before surgery and 9800 “effective” milliosmoles in her body (2 × 140 × 35). If the plasma sodium concentration was 120 mmol/L after surgery, total body water would be 40.8 L [9800/(2 × 120)], or a gain of 5.8 L of electrolyte-free water.

Patients Who Received Only Near-Isotonic Saline

Because fatal hyponatremia after surgery occurs predominantly in young women [1-3], we studied 22 randomly selected women (weight, 70 ± 4 kg; age, 42 ± 1 years) who had gynecologic (uterine) surgery (duration, 2.2 ± 0.7 hours) under general anesthesia. Management of each patient was determined by the patient's anesthesiologist and surgeon. The study protocol was approved by the ethics committee for studies in human subjects at St. Michael's Hospital (Toronto, Ontario), and informed consent was obtained from each patient. Patients had no significant medical illness aside from the condition requiring surgery (fibroids, menorrhagia, or pelvic inflammatory disease) and were not given a diuretic. All patients received only intravenous near-isotonic fluid (half as isotonic saline [sodium chloride, 154 mmol/L] and half as Ringer lactate [sodium, 130 mmol/L; potassium, 4 mmol/L]. Patients did not ingest any fluids orally; each patient had an indwelling Foley catheter that remained in place for the 24 hours of observation (which is routine practice at our hospital). Samples of venous blood were obtained at the time of induction of anesthesia and 24 hours later. Urine samples were collected every 2 to 4 hours for the first 24 hours after surgery. All volumes of ingested and excreted fluids were carefully recorded. The surgeons estimated that less than 0.5 L of blood was lost.

Analytic Techniques

Urine and plasma osmolality and concentrations of sodium, potassium, chloride, creatinine, urea, and glucose were measured as described elsewhere [6].

Statistical Analysis

Results are reported as the mean ± SE. Statistical analysis was done on group mean values using the Student t-test. A P value less than 0.01 was considered to be statistically significant.

Results

Patients with Acute, Fatal Postoperative Hyponatremia

Data on all patients whose medical and nursing records were examined are reported in Table 1. Patients in whom electrolyte-free water gain was the only cause of hyponatremia received considerably less electrolyte-free water than would be required to induce the degree of hyponatremia that was seen. Rather than postulate errors that may have occurred in the documentation of each case, we surmised that hypertonic urinary losses could also be an important reason for developing such a severe degree of hyponatremia. However, pertinent data (that is, urinary electrolyte levels) were not available to test this hypothesis in any of these patients.

Patients Who Received Only Near-Isotonic Saline

Plasma sodium concentration decreased in 21 of the 22 patients (average decrease, 4.2 ± 0.4 mmol/L) (Table 2); the lowest plasma sodium concentration was 131 mmol/L (in 2 patients). The average volume of isotonic saline (2.6 ± 0.3 L) and Ringer lactate (2.8 ± 0.3 L) that were infused for 24 hours was 5.3 ± 0.2 L (3 L during surgery plus 100 to 125 mL of isotonic saline per hour thereafter). The average urine volume was 2.5 ± 0.3 L, resulting in a net water gain of 2.9 ± 0.3 L; however, the entire retained volume did not represent electrolyte-free water (Table 3). Almost every urine sample from these patients was hypertonic (defined as a combined sodium and potassium concentration greater than 150 mmol/L) for the first 100 minutes after surgery (Figure 1). During this time, the maximum sodium plus potassium concentration in urine was 294 ± 9 mmol/L. In the subsequent 8 hours, the degree of hypertonicity of urine samples remained the same in some patients but decreased (sodium plus potassium concentration as low as 30 mmol/L) in others.

View larger version:

• In this window
• In a new window

Figure 1. A sodium plus potassium concentration greater than 150 mmol/L ( ) represents hypertonic urine and thus contributes to generation of hyponatremia. In contrast, values below this level represent excretion of hypotonic urine and thus a period when hyponatremia was being corrected. In samples obtained after 1000 minutes (from 16 to 24 hours), the urine was hypertonic in some patients and hypotonic in others. Urinary excretion of sodium and potassium after surgery.dotted line

The balance was positive for sodium (367 ± 50 mmol) and negative for potassium (90 ± 7 mmol) (Table 3), yielding a gain of 277 ± 50 mmol of sodium plus potassium for the 24-hour period of observation. The balance for water during this period was positive (2.9 L) (Table 3). Subdividing the gain of water and electrolytes into isotonic and electrolyte-free water results in a net gain of 1.8 L of isotonic saline (which expanded the extracellular fluid volume) and a gain of 1.1 L of electrolyte-free water (which caused the hyponatremia) (Figure 2 and Table 3).

View larger version:

• In this window
• In a new window

Figure 2. Hatching represents isotonic saline in the extracellular fluid. The rectangle on the left represents the preoperative setting with a normal plasma sodium concentration (140 mmol/L). The rectangle on the right represents the postoperative state during which hypotonicity and edema have resulted from partial desalination of the infused near-isotonic saline. Expansion of extracellular fluid volume is caused by retention of isotonic saline. The desalination process involves administration of near-isotonic solutions and loss of hypertonic urine. Antidiuretic hormone causes retention of electrolyte-free water, which forms as a result of hypertonic urinary losses. Thus, two volumes are retained: The first volume (larger hatched rectangle) represents retained isotonic saline, which would not change the plasma sodium concentration. The second volume (white rectangle) represents retained electrolyte-free water (generated by excretion of hypertonic urine), which leads to hyponatremia and cell swelling. The desalination process.

Discussion

The physiologic principles that explain the development of hyponatremia after surgery are well characterized in the literature. A source of electrolyte-free water must be added to the extracellular fluid, and antidiuretic hormone must act to prevent excretion of this electrolyte-free water [7]. The usual source of electrolyte-free water is administration from exogenous sources (5% dextrose in water or hypotonic saline intravenously or orally). It is also possible, however, for the kidney to generate electrolyte-free water. For this to occur, the urine must have high concentrations of sodium plus potassium. For the kidney to generate a large volume of electrolyte-free water in body fluids, there must be a high rate of excretion of a sodium salt, potassium salt, or both in hypertonic urine. However, it is not always obvious why some patients have a greater degree of natriuresis than others. Several possibilities are considered below.

Our study was unique because it showed that administration of electrolyte-free water retained in the body (0.36 L in Ringer lactate) is only partly responsible for acute postoperative hyponatremia. Most of the electrolyte-free water (0.74 L) was generated by the kidney. Excretion of this electrolyte-free water was prevented by action of antidiuretic hormone, as evidenced by the elaboration of hypertonic urine (Figure 1 and Table 3). Our patients had a positive balance for fluid (2.9 L) and electrolytes (276 mmol of cations) that resulted from partial desalination (that is, excretion of much of the sodium chloride that was administered with disproportionately less water than was infused) of the near-isotonic saline that was administered (Figure 2). In quantitative terms, these balances represent a net gain of approximately 1.8 L of isotonic saline (which would expand the extracellular fluid volume but not change the plasma sodium concentration) plus a gain of approximately 1.1 L of electrolyte-free water (which would lead to hyponatremia and cell swelling). This amount of electrolyte-free water represents approximately 3.0% of the total body water and accounts for the decline in plasma sodium concentration that we saw (Table 3).

At the same time, antidiuretic hormone must be present to prevent excretion of electrolyte-free water. Release of antidiuretic hormone was not triggered by the physiologic stimuli of hypertonicity or a low effective circulating volume but could represent a response to any combination of pain, drugs, stress, or nausea. The duration of action of antidiuretic hormone is important. Some patients (8 of 22) began to excrete electrolyte-free urine after 16 hours, whereas others (14 of 22) continued to excrete hypertonic urine for the entire observation period (Figure 2). Hyponatremia is being corrected in the former set of patients, but the latter group is generating more electrolyte-free water and is in danger of developing more severe hyponatremia.

The degree of postoperative hyponatremia in our patients resulted from the large volume of hypertonic urine that was excreted. One factor that contributed to the high rate of excretion of electrolytes was the amount of saline that was infused to maintain adequate blood pressure after induction of anesthesia. In our patients, 2 to 3 L of Ringer lactate was infused during surgery with the expectation that retained near-isotonic saline would not change tonicity; however, tonicity decreased because of the excretion of hypertonic urine. The most important stimulus for excretion of sodium was overexpansion of the extracellular fluid volume. The ratio of infused isotonic saline to extracellular fluid volume also affects the severity of hyponatremia. Hence, a smaller volume should be given to smaller patients because their risk for hyponatremia and desalination increases with each liter infused. Whether isotonic fluid is retained in extracellular fluid before anesthesia, as may occur premenstrually in some women or in anyone who has consumed a high-salt diet, should also be taken into account. Such retention could make a larger volume available for desalination without causing a low effective in extracellular fluid volume.

It follows from the above reasoning that a “good urine output” is no longer reassuring but rather a potential worry with respect to postoperative hyponatremia if such output is hypertonic. The degree of natriuresis and hyponatremia could be exaggerated further if natriuretic factors, either endogenous (such as in cerebral salt wasting [8, 9]) or pharmacologic (such as when thiazide diuretics are given) are present. Loop diuretics, in contrast, can compromise the concentration process and thereby help to prevent excretion of hypertonic urine. In quantitative terms, a urinary loss of 300 mmol of sodium plus potassium in 1 L of urine would be equivalent to the desalination of 2 L of isotonic saline, yielding a positive balance of 1 L of electrolyte-free water in the body. In a patient who weighs 70 kg, this occurrence would cause an almost 4-mmol/L decrease in plasma sodium concentration if all such water is retained in the body.

Another point that merits emphasis is potassium excretion [5, 10-12]. It is well known that potassium in the urine is important for tonicity [13]. As with sodium excretion, excretion of potassium without water represents a loss of particles, which leads to a decrease in the concentration of sodium in plasma. Nevertheless, the effect of potassium loss on the intracellular fluid volume may differ from that of sodium loss [14]. If potassium is lost with chloride and the potassium is replaced in the intracellular fluid with sodium, it represents a net loss of 2 particles from the extracellular fluid (the same as with loss of sodium chloride). This loss is associated with expansion of intracellular fluid volume and contraction of extracellular fluid volume. If potassium is lost with phosphate, however, water is shifted out of cells (that is, intracellular particles are lost) because both particles are ultimately of intracellular origin. In such cases, hyponatremia is accompanied by contraction of global intracellular fluid volume and expansion of extracellular fluid volume [5, 11, 12, 14, 15]. However, potassium represented only 20% of the major cations excreted in our patients, which makes loss of potassium unlikely to have contributed substantially to the hyponatremia.

Another factor to consider is the ability of the renal medulla to increase the sodium concentration in urine under the influence of antidiuretic hormone. For example, if the urine has an inadequate concentration of sodium because of recent obstructive uropathy or administration of a loop diuretic, the patient would be expected to excrete a large volume of urine in which the sum of the concentration of sodium and potassium is almost 150 mmol/L even when antidiuretic hormone is acting. In this situation, hyponatremia would not be anticipated to occur despite natriuresis if only isotonic fluids were being administered.

Most articles that described postoperative hyponatremia [1, 3, 16-19] imply that this derangement is primarily the result of administration of hypotonic intravenous solutions, but quantitative analysis could not be done from the data available in these studies. The fatal postoperative hyponatremia that developed in five patients we evaluated would have required a positive balance of 4.1 L of electrolyte-free water if the condition developed on the basis of the administration of electrolyte-free water alone; in fact, these patients received only 2.7 L of electrolyte-free water intravenously (Table 1).

Four studies [20-23] that evaluated the influence of administration of isotonic fluid on the incidence of postoperative hyponatremia 24 hours after a surgical procedure are pertinent here. Only one [20] showed a substantial decline in natremia. Unfortunately, no data on urinary electrolytes were reported to help establish the mechanism that was involved in this decline. A second study [21] reported mean urine outputs and sodium excretion rates, but not values for potassium excretion (plasma osmolality did not change in the 24-hour period, but this measure is not sensitive enough to evaluate whether there was a small decline in the plasma sodium concentration). In two other studies [22, 23], the plasma sodium concentration did not decline during a 24-hour period. Nevertheless, such key data as the sodium plus potassium concentration in the urine were not reported. The urine output of the patients in these two studies was much lower than that of our patients. It is important to recognize that lower volumes of hypertonic urine yield lower volumes of electrolyte-free water (Figure 1). To explain why our patients had a high urine output is difficult without more information. Our patients were young, had elective surgery, and received a large infusion of fluid (5.3 L in 24 hours); therefore, extensive natriuresis might be anticipated. Moreover, the same quantity of electrolyte-free water generated in a small patient causes a greater degree of hyponatremia because of the smaller volume of total body water, as noted.

Our study findings should not be misinterpreted as suggesting that infusing hypotonic solutions has a risk similar to that of infusing isotonic solutions; in fact, the risk is larger with hypotonic infusions. For example, if 2 L of isotonic saline were infused and 1 L of urine with a sodium plus potassium concentration of 300 mmol/L was excreted, the positive electrolyte-free water balance would be 1 L. In contrast, if 2 L of half-normal saline (77 mmol/L) were infused, all the sodium infused would now be excreted in 0.5 L with a concentration of sodium plus potassium of 300 mmol/L and lead to a positive electrolyte-free water balance of 1.5 L.

One results suggest that standard clinical perioperative fluid and electrolytic management should change. Hypotonic intravenous fluids should not be given during or immediately after surgery; they should only be given if a patient is hypernatremic. The minimum volume of isotonic fluid needed during and after surgery to maintain hemodynamics should be infused. Plasma sodium concentration should be checked if more than 2 to 3 L of hypertonic urine (specific gravity > 1.020) has been excreted in the first 24 hours after surgery in a person weighing 70 kg. If hyponatremia is present and a large volume of hypertonic urine continues to be excreted, two therapeutic options are available to prevent a further decrease in the plasma sodium concentration. One option is to infuse a saline solution with the same tonicity and at the same flow rate as the urine; this requires that a hypertonic solution be infused until the concentration of sodium plus potassium in the urine decreases. The second option is to decrease the concentration of sodium plus potassium in the urine by administering a loop diuretic [24, 25] or an osmotic diuretic (such as urea) [26]. A greater degree of hyponatremia can be anticipated if the patient has also consumed water, received hypotonic solutions, or retained saline before surgery (high salt intake); if a natriuretic agent (nonloop diuretic) has been acting; or if antidiuretic hormone actions (that is, pain, nausea, drugs, and stress) persist longer than 24 hours after surgery.

In summary, postoperative hyponatremia occurred in patients infused with isotonic saline. The degree of hyponatremia depends largely on the volume of intravenous fluid given relative to body size. Postoperative hyponatremia is the result of two factors: 1) addition of electrolyte-free water by infusion, renal generation, or both and 2) the presence of antidiuretic hormone to prevent excretion of electrolyte-free water. Many factors may explain the degree of desalination in an individual patient. With a better understanding of the mechanisms involved and more rational management, such tragedies as those described in Table 1 can be prevented.

Dr. Feldman: University Hospital, Room 60F, 11-339 Windermere Road, London, Ontario N6A 5A5, Canada.