Lactated Ringer's solution (LR) and 0.9% sodium chloride solution (NS) are commonly used as IV fluids. The osmolarity of LR is 273 mOsm/L. In dilute physiological solutions, the values of osmolality and osmolarity are interchangeable . However, when measured by the depression of freezing point, the osmolality of LR is 254 mOsm/kg [2,3]. This discrepancy is due to the incomplete ionization of the solutes in LR. However, the measured osmolality of NS (mOsm/kg), which is more completely ionized, is similar to the calculated osmolarity of 308 mOsm/L. Thus, the osmolality of LR is lower than, and that of NS is equal to or higher than, the osmolality of normal serum (285-295 mOsm/kg) . In animal studies of isovolemic hemodilution with large volumes of IV LR, the serum osmolality decreased [1,2,5-9]. Because the blood-brain barrier (BBB) allows the passage of water along osmotic gradients , serum osmolality is a determinant of brain water content, so that low serum osmolality may contribute to cerebral edema [5-7,11,12]. Based on these data, it has been recommended that IV LR be administered cautiously to neurosurgical patients [6,13].
Studies demonstrating changes in osmolality associated with the administration of large volumes of IV LR and NS have been performed only in animals [1,2,6,7]. We undertook the current investigation in healthy human volunteers to determine whether large volumes of IV LR or NS would result in changes of serum osmolality.
The protocol was approved by our institutional review board. Twenty healthy human volunteers aged 20-48 yr were enrolled in the study. After giving written, informed consent, each volunteer was randomly assigned to one of two groups using sealed envelopes. The investigations were performed before noon. There was no restriction on oral intake before the IV infusions, and each subject urinated and was weighed immediately before the infusion. Initially, one group received LR and the other received NS. The alternate solution was infused during a second study period at least 3 days after the initial infusion. The subjects and the laboratory staff (performing the measurements) were blinded as to which solution was administered. However, the investigators administering the IV fluids, taking blood samples, and monitoring the volunteers were not blinded. After local anesthetic infiltration, a peripheral IV cannula (16 or 14 g) was inserted in the antecubital fossa. The subjects were seated or lying with head elevation, except to walk approximately 30 yards to urinate. They read, conversed, or slept as they desired. There was no oral fluid intake during the study period, and 50 mL/kg study solution was infused IV over the first hour.
Venous samples were taken through the same cannula in the antecubital fossa without a tourniquet. The first sample (T1) was taken before the IV infusion tubing was attached. The second sample was taken at the end of the infusion (T2): the IV infusion was disconnected, the first 5 mL of blood was discarded, and a sample taken and used for measurements. The cannula was then flushed with 5 mL of the study fluid, and over the next hour, 25 mL was infused to keep the cannula patent. One hour after the end of the infusion (T3), the third sample was taken in a manner identical to the second sample, and the IV cannula was removed.
If requested by the subject, IV furosemide 5-10 mg was administered via the cannula and flushed in with 5 mL of study solution after the third sample was taken and before removal of the cannula. The objective of furosemide administration was to promote a diuresis, which would decrease the duration of sensations associated with hypervolemia, such as the feeling of distension of the head and neck, ankles, and abdomen.
Throughout the study period, blood pressure, heart rate, peripheral oxygen saturation, and respiratory rate were recorded at 10-min intervals. The volunteers were also closely observed for symptoms of discomfort or abnormal physical signs, and the subjects' comments were noted by the observers who administered the infusions.
Serum sodium, glucose, osmolality, and pH were measured from each venous blood sample. Serum sodium concentration was measured using an ion-selective electrode. Serum glucose was determined using the oxygen rate method with an oxygen electrode. Serum osmolality was measured by determination of freezing point depression, and whole blood pH was measured with a pH-sensitive glass electrode.
We also recorded the length of time from the start of the infusions until the subjects first urinated.
The data for venous serum osmolality, pH, glucose, and sodium concentrations were each analyzed with a two-solution (LR vs NS) by three-episode (T1, T2, T3) mixed-model repeated-measures analysis of variance. Time to urination was analyzed as a two-solution paired t-test. Forward stepwise logistical regression was used to assess whether the incidence of mental changes or abdominal discomfort was associated with any of the solution variables.
Two subjects withdrew from the study after the first infusion. The first subject experienced abdominal discomfort after the first infusion (NS). The other subject withdrew because of causes unrelated to the study. The results of the remaining 18 subjects (6 women and 12 men) were used for analysis. The measured venous blood variables are displayed in Table 1.
When LR was infused, serum osmolality (Table 1) decreased by 4 +/- 3 mOsm/kg from T1 to T2, whereas it showed an insignificant increase with the NS infusion. This difference between the infusions was statistically significant (P < 0.05). At T3, serum osmolality returned to baseline after LR and remained unchanged after NS. When NS was infused, serum sodium concentration (Table 1) increased 1 +/- 2 mEq/L between T1 and T2 and decreased insignificantly when LR was infused, but these differences were statistically significantly different from each other (P < 0.05). Whole blood venous pH (Table 1) increased from T1 to T2 by 0.04 +/- 0.04 after the LR infusion and decreased by 0.04 +/- 0.04 after NS. This difference was significant (P < 0.001). The pH at T3 after the administration of NS remained the same as T2. After LR infusion, the pH at T3 insignificantly trended toward baseline. Serum glucose concentration (Table 1) showed no significant difference between the two infusions and did not change over time.
The time from T1 until first urination (Table 2) was significantly different between the two infusions (P < 0.001): NS 106 +/- 11 min compared with LR 75 +/- 10 min. Women demonstrated a 53-min difference in time to urination with NS (133 +/- 39 min) compared with LR (80 +/- 14 min). This differed significantly from the 10-min difference (69 +/- 40 min with LR and 79 +/- 11 min with NS) in men (P < 0.05). Furosemide was administered to 10 subjects 1 h after each infusion.
Blood pressure, heart rate, and peripheral oxygen saturation were stable throughout the study in all subjects. With both solutions, all subjects reported mild discomfort associated with swelling of the head and neck, abdominal girth, and extremities. However, lassitude and a perceived difficulty in abstract thinking (such as mental arithmatic, reading medical journals, or replying to anesthesiology board questions) were noted by 13 subjects with NS and none with LR. These subjective mental changes did not correlate with any other measured variable. Ten subjects experienced abdominal discomfort with NS, one subject experienced it after both LR and NS, and seven did not experience it after either LR or NS. Logistical regression revealed no correlation between abdominal discomfort and any measured variable, except that pH at T2 was associated with abdominal discomfort in the NS group (P < 0.01).
These results indicate that, in healthy humans, the infusion of large volumes of LR but not NS tends to decrease serum osmolality. This is in agreement with previous animal studies [6,11], although those studies examined the infusion of large volumes of crystalloid to replace blood loss, whereas our study did not remove large volumes of blood. However, our study also shows that, within an hour of ending the infusion of LR, serum osmolality returns toward baseline. This is to be expected, because changes in serum osmolarity are sensed in the hypothalamus by neurons that terminate in the posterior pituitary . Decreased osmolality inhibits the release of antidiuretic hormone from these neurons, and the resulting diuresis of hypotonic urine causes the serum osmolality to return to normal.
In the experimental literature, decreased serum osmolality after infusion of LR is associated with increased cerebral water content. When LR was used to resuscitate a porcine model of head injury, Walsh et al.  and Shackford et al.  demonstrated a decrease in osmolarity (from 290 +/- 4 to 277 +/- 10 mOsm/L and from 291 +/- 3 to 286 +/- 5 mOsm/L) with a worsening of cerebral water content and intracranial pressure (ICP) in both. In a rabbit model of hemodilution with LR, Tommasino et al.  showed that there was a transient decrease in osmolality (from 279 +/- 9 to 275 +/- 2 mOsm/kg) accompanied by a transient increase in ICP and cerebral water content. An increase in cerebral water content and ICP was also demonstrated in a rabbit model of head injury by Zornow et al.  when the infusion of LR resulted in a decrease of osmolality of 5 +/- 3 mOsm/kg.
Theoretically, a difference of 1 mOsm across a semipermeable membrane, such as the BBB, exerts a pressure of >19 mm Hg in osmotic pressure , so that even small changes in osmolality can be expected to have a marked effect on water transfer. Evidence to support this is provided by Hyodo et al. , who reported increased cerebral water content in dogs hemodiluted with LR after middle cerebral artery occlusion. The osmolarity of these animals was 295 +/- 1 mOsm/L, compared with 296 +/- 1 mOsm/L in the control animals. However, when Feldman et al.  administered LR to a rat model of head injury, there was no increase in cerebral water content, despite a decrease of serum osmolality from 293 +/- 11 to 287 +/- 6 mOsm/kg.
In our study, the difference in osmolality after LR infusion was small (4 +/- 3 mOsm/kg), and the human volunteers showed no clinical signs or symptoms of increased ICP or cerebral edema after LR. This may indicate the presence of compensatory mechanisms that protect the brain. Unfortunately, these mechanisms may not be intact in the clinical situation in which large volumes of IV crystalloid infusion are required for patients with intracranial pathology. Therefore, care should still be taken to ensure iso- or hyperosmolarity in the presence of possible brain injury.
NS did not change serum osmolality but resulted in a decrease in venous blood pH. Although the pH in venous blood is lower than arterial pH, consistent correlation between the two has previously been demonstrated  when the samples are taken from the same site, with subjects in the resting state as they were throughout the study. The decrease in pH associated with NS infusion has been observed before and is attributed to hyperchloremic metabolic acidosis [17,18]. We presume that this was the cause, but we did not measure serum chloride. It was interesting that the pH changes did not return to baseline 1 h after the infusion. Furosemide has been recommended to treat hyperchloremic acidosis . However, although it was administered to a number of the volunteers, no further blood samples were taken in our study to evaluate the efficacy of this treatment.
The symptoms of abdominal discomfort and mental changes were unexpected observations. There was no systematic quantification of these observations, and they were simply recorded as comments by the subjects. However, except for one subject who had abdominal pain with both NS and LR, this was only noted after the NS infusion. Drowsiness or decreased mental capacity for performing tasks such as reading and mental arithmatic was observed in 13 subjects only after NS. Although the possibility of bias cannot be excluded from these findings because the observers were not blinded, it is an interesting speculation that early recovery from anesthesia might be modulated by the type of IV fluid if large quantities have been administered. However, further investigation with a double-blinded protocol and a structured mental state examination would be required to verify these observations.
The longer time before urination after NS versus LR implies that NS may be associated with more fluid retention than LR. Inhibition of ADH secretion caused by the lowered osmolality  observed after LR infusion is one possible explanation of this finding. Although we did not measure it, increased serum chloride concentration, which could have resulted from the infusion of NS, decreases renal blood flow and glomerular filtration rate , thereby causing urinary retention. We observed more prolonged urinary retention in women than in men with both types of fluid. These observations may have clinical implications in that the type of IV fluid, as well as the volume, may influence urinary output, especially in women. The administration of furosemide had little effect on the time at which the subjects first urinated. It was administered 120 min after the start of the infusion, by which time all subjects-except four after NS and one after LR administration-had already urinated.
In summary, LR 50 mL/kg administered IV over 1 h to healthy human volunteers resulted in a transient small decrease in serum osmolality. An equal volume of IV NS administered over 1 h did not affect the serum osmolality but was associated with a mild acidosis, which had not changed 1 h after the infusion was finished.
The authors thank Dr. Michael Shaw for technical assistance, Dr. Joseph Lucke for statistical assistance, and Dr. Glenn Gravlee for helpful discussion of this manuscript.
1. Greger R. Units used in physiology and their definition. In: Greger R, Winhorst U, eds. Comprehensive human physiology from cellular mechanisms to integration. Berlin: Springer, 1996:2421-6.
2. Feldman Z, Zachari S, Reichenthal E, et al. Brain edema and neurological status with rapid infusion of lactated Ringer's or 5% dextrose solution following head trauma. J Neurosurg 1995;83:1060-6.
3. Zornow MH, Scheller MS, Shackford SR. Effect of a hypertonic lactated Ringer's solution on intracranial pressure and cerebral water content in a model of traumatic brain injury. J Trauma 1989;29:484-8.
4. Normal reference laboratory values. N Engl J Med 1980;302:37-48.
5. Korosue K, Heros RC, Ogilvy CS, et al. Comparison of crystalloids and colloids for hemodilution in a model of focal cerebral ischemia. J Neurosurg 1990;73:576-84.
6. Tommasino C, Moore S, Todd MM. Cerebral effects of isovolemic hemodilution with crystalloid or colloid solutions. Crit Care Med 1988;16:862-8.
7. Hyodo A, Heros RC, Tu Y, et al. Acute effects of isovolemic hemodilution with crystalloids in a canine model of focal cerebral ischemia. Stroke 1989;20:534-40.
8. Walsh JC, Zhuang J, Shackford SR. A comparison of hypertonic to isotonic fluid in the resuscitation of brain injury and hemorrhagic shock. J Surg Res 1991;50:284-92.
9. Shackford SR, Zhuang J, Schmoker J. Intravenous fluid tonicity: effect on intracranial pressure, cerebral blood flow, and cerebral oxygen delivery in focal brain injury. J Neurosurg 1992;72:91-8.
10. Gregoire N. The blood-brain barrier. J Neuroradiol 1989;16:238-50.
11. Zornow MH, Todd MM, Moore SS. The acute cerebral effects of changes in plasma osmolality and oncotic pressure. Anesthesiology 1987;67:936-41.
12. Kaieda R, Todd MM, Cook LN, Warner DS. Acute effects of changing plasma osmolality and colloid oncotic pressure on the formation of brain edema after cryogenic injury. Neurosurgery 1989;24:671-8.
13. Zornow MH, McQuitty C, Prough DS. Perioperative fluid management of the neurosurgical patient. In: Albin MS, ed. Textbook of neuroanesthesia with neurosurgical and neuroscience perspectives. New York: McGraw-Hill, 1997:117-48.
14. Koizumi K. The role of the hypothalamus in neuroendocrinology. In: Greger R, Winhorst U, eds. Comprehensive human physiology. Berlin: Springer-Verlag, 1996:379-401.
15. Zornow MH, Scheller MS. Intraoperative fluid management during craniotomy. In: Cottrell JE, Smith DS, eds. Anesthesia and neurosurgery. St. Louis, MO: Mosby, 1994:247-59.
16. Tobias JD, Meyer DJ, Jr, Helikson MA. Monitoring of pH and PCO2
in children using the Paratrend 7 in a peripheral vein. Can J Anaesth 1998;45:81-3.
17. Veech RL. The toxic impact of parenteral solutions on the metabolism of cells: a hypothesis for physiological parenteral therapy. Am J Clin Nutr 1986;44:519-51.
18. McFarlane C, Lee A. A comparison of Plasmalyte 148 and 0.9% saline for intra-operative fluid replacement. Anaesthesia 1994;49:779-81.
19. Sebastian A, Schambelan M, Sutton JM. Amelioration of hyperchloremic acidosis with furosemide therapy in patients with chronic renal insufficiency and type 4 renal tubular acidosis. Am J Nephrol 1984;4:287-300.
20. Wilcox CS. Regulation of renal blood flow by plasma chloride. J Clin Invest 1983;71:726-35.