We conducted an open, randomized, and prospective study to determine the effect of hypertonic saline on the secretion of antidiuretic hormone (ADH) and aldosterone in children with severe head injury (Glasgow coma scale <8). Thirty-one consecutive patients at a level III pediatric intensive care unit at a children’s hospital received either lactated Ringer’s solution (Ringer’s group, n = 16) or hypertonic saline (Hypertonic Saline group, n = 15) over a 3-day period. Serum ADH levels were significantly larger in the Hypertonic Saline group as compared with the Ringer’s group (P = 0.001; analysis of variance) and were correlated to sodium intake (Ringer’s group:r = 0.39, R2 = 0.15, P = 0.02; Hypertonic Saline group:r = 0.42, R2 = 0.18, P = 0.02) and volume of fluids given IV (Ringer’s group:r = 0.38, R2 = 0.15, P = 0.02; Hypertonic Saline group:r = 0.32, R2 = 0.1, P = not significant). Correlation of ADH to plasma osmolality was significant if plasma osmolality was >280 mOsm/kg (r = 0.5, R2 = 0.25, P = 0.06), indicating an osmotic threshold for ADH release. Serum aldosterone levels were larger on the first day than during Days 2 and 3 in both groups and inversely correlated to serum sodium levels only in the Ringer’s group (r = −0.55, R2 = 0.3, P < 0.001). This group received a significantly larger fluid volume on Day 1 (P = 0.05, Mann-Whitney U-test) than did patients in the Hypertonic Saline group, indicating hypovolemia during the first day. Head-injured children have appropriate levels of ADH. They may be hypovolemic during the first day of treatment, especially if they receive lactated Ringer’s solution.
*Intensive Care Unit, †Department of Pediatric Surgery, and ‡Department of Pediatrics, Children’s Hospital, Zurich, Switzerland; §Intensive Care Unit, Children’s Hospital, Innsbruck, Austria; and ∥Institute of Medical Statistics and Epidemiology, Munich, Federal Republic of Germany
November 2, 2000.
Implications: In head-injured patients, we recommend fluid restriction to avoid inappropriate secretion of antidiuretic hormone. In a prospective, randomized, and controlled study in 31 children, we were able to show that the antidiuretic hormone levels are appropriate in response to hypovolemia, sodium load, or both.
Address correspondence and reprint requests to Burkhard Simma, MD, Children’s Hospital, 6800 Feldkirch, Austria. Address e-mail to email@example.com.
Hyponatremia is common in children with severe head injury (1). The syndrome of inappropriate secretion of antidiuretic hormone (SIADH) is one of the most frequent causes (2–4). Inappropriate secretion of antidiuretic hormone (ADH) is characterized by increased water retention, resulting in hyponatremia, and by impaired water excretion with concentration of the urine and increased urinary sodium concentration (5–7). In adults with head injury, SIADH occurs in 5%(3) to 33%(4) of patients. In pediatric patients, one report has demonstrated a 25% prevalence of SIADH in mild head injuries with a Glasgow Coma Scale score (GCS) > 7 in eight children (8). Although the rationale of fluid restriction in head injury has been discussed recently (9), water restriction to two thirds of maintenance intake is still recommended to avoid SIADH (2,5,6).
However, trauma patients need large volumes of fluids, which may intensify the effect of inappropriate secretion of ADH, leading to cerebral edema and worsening neurologic outcome (10). In patients with serum sodium levels <125 mmol/L caused by SIADH, the administration of hypertonic saline is recommended (1,2). Hypertonic saline has the additional advantage of restoring arterial blood pressure and cardiac output with a smaller volume of fluid (11,12) and may lead to decreased intracranial pressure and improved cerebral perfusion pressure (13).
We analyzed the data from a prospective, randomized, and controlled study (13) to determine the effect of hypertonic saline on the secretion of ADH and aldosterone—as an indicator of intravascular volume—in patients with severe head injury (GCS < 8).
We studied 35 consecutive children with severe traumatic brain injury who were admitted to a pediatric intensive care unit (ICU) in Zurich from August 1, 1992 to November 30, 1995. The study was approved by the IRB of the hospital, and informed consent was obtained from the parents. The study protocol was described in detail elsewhere (13). To be enrolled, the patients had to have a GCS of <8 at the scene of the accident and be younger than 16 yr old. The patients entered the study at the time of admission to the ICU. Three patients had to be excluded because of a GCS ≥8 and one patient because of diabetes insipidus. Patients were randomly assigned to receive either lactated Ringer’s solution (Ringer’s group) (sodium 131 mmol/L, potassium 5 mmol/L, calcium chloride 1.8 mmol/L, lactate 27 mmol/L, 277 mOsm/L) or hypertonic saline (Hypertonic Saline group) (sodium 268 mmol/L, potassium 4 mmol/L, chloride 218 mmol/L, sodium bicarbonate 54 mmol/L, 598 mOsm/L) over 72 h. The aim in the Hypertonic Saline group was to reach a serum sodium level of 145–150 mmol/L. In both groups, maintenance fluid was restricted to 1200 mL per square meter of body surface area per day. In all patients, intracranial pressure was measured to calculate cerebral perfusion pressure. Low cerebral perfusion because of increased intracranial pressure was treated with a predefined sequence of therapeutic interventions, including additional analgesics and sedatives, mannitol, and thiopentone. Low cerebral perfusion pressure because of low mean arterial pressure was treated with a bolus of crystalloids (lactated Ringer’s solution, isotonic saline) (20 mL/kg body weight). Routine care was standardized (14) and included head positioning at 30°, normothermia, analgesia and sedation, volume-controlled ventilation (Paco2 3.5–4.0 kPa), optimal oxygen delivery, and phenobarbital.
ADH (arginine-vasopressin) and aldosterone levels were determined by a commercially available radioimmunoassay (Nichols Institute; CIS Bio International, San Jose, CA). All procedures were performed according to the manufacturer’s specifications. Reference values are reported elsewhere (15).
Mean arterial and central venous blood pressure were monitored continuously and documented hourly. For 3 days, serum sodium was measured every 4 h, and serum osmolality, twice daily. Blood samples for ADH and aldosterone determination were collected every day in the morning—namely, the first sample 8–12 h after admission and then daily for another 2 days. Differences in continuous variables were calculated by means of analysis of variance (ANOVA) for repeated measures. Differences in baseline values were assessed by using t-tests; differences in sodium and fluid administration and urine output were tested according to the Mann-Whitney U-test. In testing the correlation of two continuous variables over time, the partial correlation coefficient was used with the patient (dummy coded) as the control variable. The resulting correlation coefficient has the interpretation of a mean of all correlation coefficients of all patients within each group. Values are expressed as mean ± sd. P < 0.05 was considered significant, and all stated P values are two sided. Statistical computations were performed with SPSS™ for Windows Version 7.5 (SPSS, Chicago, IL).
The mean age of 31 patients was 87 ± 42 months (range, 12–173 months), and 16 patients (50%) were boys. No statistically significant differences with respect to age, sex ratio, initial GCS (5.8 ± 1.6 vs 5.5 ± 1.4) and laboratory findings, the mechanism of injury, or the computer tomography examination findings of the head were observed between the two groups.
On admission, no difference was seen in serum sodium or in serum or urine osmolality levels. According to the design of the study, patients of the Ringer’s group received significantly less sodium as compared with the Hypertonic Saline group (7.8 ± 4.5 vs 11.5 ± 5.0 mmol · kg−1 · d−1, P < 0.05, Mann-Whitney U-test) (Table 1). Over time, the difference in serum sodium and serum osmolality levels became statistically significant for the groups (P < 0.05, ANOVA) (Fig. 1). In both groups, correlation between serum sodium and osmolality was statistically significant (Ringer’s group:r = 0.46, R2 = 0.21, P < 0.005; Hypertonic Saline group:r = 0.72, R2 = 0.52, P < 0.001). Patients in the Ringer’s group received a significantly larger fluid volume on Day 1 as compared with patients in Hypertonic Saline group (2860 ± 1530 vs 2180 ± 770 mL per square meter of BSA per day, P = 0.05, Mann-Whitney U-test) (Table 1). However, over the 3-day period, no significant difference in fluid volume was observed between the groups (Table 1). Mean arterial and central venous pressure did not differ between the two groups over the study period.
Serum ADH levels were significantly smaller in the Ringer’s group as compared with the Hypertonic Saline group over the whole study period (P = 0.001; ANOVA) (Fig. 2). Correlation of ADH to sodium intake was significant in both groups (Ringer’s group:r = 0.39, R2 = 0.15, P = 0.02; Hypertonic Saline group:r = 0.42, R2 = 0.18, P = 0.02). ADH levels positively correlated to plasma osmolality in the Hypertonic Saline group when plasma osmolality was >280 mOsm/kg, indicating an osmotic threshold for ADH level (Fig. 3). There was a positive correlation between ADH and volume of fluids in both groups. This correlation was significant only in the Ringer’s group (Ringer’s group:r = 0.38, R2 = 0.15, P = 0.02; Hypertonic Saline group:r = 0.32, R2 = 0.1, P = not significant). ADH did not correlate with the severity of head injury defined by initial GCS.
Levels of serum aldosterone, as an indicator of intravascular volume, showed no difference within or between the two groups (P = 0.09, not significant; ANOVA) (Fig. 4). However, in both groups, aldosterone levels were larger on the first day than during the continued study. In the Ringer’s group, aldosterone inversely correlated to serum sodium levels (r = −0.55, R2 = 0.3, P < 0.001), whereas in the Hypertonic Saline group, this correlation was not significant. No significant correlation was observed between aldosterone and the amount of sodium intake or volume of administered fluids.
In this study, we found that plasma levels of ADH are appropriate in children with severe head injury. In those receiving lactated Ringer’s solution, ADH levels remained in normal range (15). In those receiving hypertonic saline, ADH levels were appropriately increased because of an increase of plasma osmolality, demonstrating that an osmotic threshold for ADH release was preserved in these patients. Corresponding to the larger ADH levels, these children need fewer fluids to achieve a similar circulatory volume—indicated by the same aldosterone level—than those receiving lactated Ringer’s solution.
In a previous study in the same patient cohort (13), we were able to demonstrate that increasing serum sodium levels by hypertonic saline significantly correlates with decreased intracranial and increased cerebral perfusion pressures. Patients of the Hypertonic Saline group required fewer interventions to keep the intracranial pressure <15 mm Hg, and they had fewer complications and stayed a shorter time in the ICU.
The appropriate response of ADH to changes in osmolality and sodium intake in our patients stands in contrast to previous clinical studies about secretion of ADH in head-injured patients (2–4,8). These papers suggest that head injury leads to an inappropriate response of ADH in a large percentage of such patients, and therefore fluid restriction is still recommended as standard therapy (5). Kamoi et al. (16) studied the ADH response to hypertonic saline in six patients with SIADH related to central nervous system disorders. They found a persistent ADH secretion with a loss of hypotonic suppression. It is thought that continued ADH release despite small serum sodium levels and hypotonicity are caused by head trauma per se, and the degree of resulting hyponatremia is said to correlate with the extent of neurologic sequelae (8). Therefore, fluid restriction is still recommended because this may decrease the risk of hyponatremia and cerebral edema and consequently improve neurological outcome (1,3,8). In contrast to the report of Kamoi et al. (16), the number of patients examined in our study is larger, and the result shows an appropriate response of ADH to sodium load. Because of this, however, the existence of SIADH cannot be excluded in our patients.
Release of ADH into the circulation is influenced by a number of stimuli. Powell et al. (7) investigated ADH levels in children with meningitis receiving either maintenance plus replacement fluid therapy or a restricted fluid intake of two thirds of maintenance. They showed that ADH positively correlated with the amount of sodium intake. Their data suggest that in the patients receiving two thirds of maintenance fluids, hypovolemia is responsible for the increased ADH levels rather than high osmolality.
In our patients, ADH and aldosterone levels were larger on the first day than during the rest of the study period. This corresponds to a larger volume of fluid administered during the first 24 hours after head injury and, surprisingly, to an inverse correlation of aldosterone to serum sodium level in patients receiving lactated Ringer’s solution. These data suggest that children with head injury may be hypovolemic during the first day of treatment, especially if they receive lactated Ringer’s solution.
The administration of hypertonic solutions will restore plasma volume (11,12), theoretically leading to decreased ADH levels. The influence of increasing plasma osmolality by means of hypertonic saline, however, will interact with this effect and increase ADH secretion. In healthy subjects, the plasma osmolality at which ADH release is initiated can be determined during infusions of hypertonic saline and was found to range from 280 to 290 mOsm/kg (17). In our study, this osmotic threshold for ADH release may be preserved in the patients with severe head injury, supporting the evidence of appropriate secretion of ADH to sodium load.
Measurements of ADH are not routine practice at an ICU because it takes at least several days to obtain results. Large ADH levels are clinically diagnosed in patients with hyponatremia, oliguria, and concentrated urine. To define SIADH, other conditions leading to appropriately large ADH levels—such as hypovolemia, mechanical ventilation, hypothyroidism, pain, and several drugs (18–20) —must be excluded. We feel that the two patient groups in this study are homogeneous and comparable because care of these traumatized patients was standard, according to the literature (14). Pain control and sedation were provided on a routine basis, and all patients were mechanically ventilated in a volume-controlled mode. Increased ADH levels in the patients studied on the first day of therapy are appropriate and the result of a physiological stimulus, namely, hypovolemia. After this time, ADH secretion is stimulated by increased plasma osmolality. As described by Ichikawa (20), the body gives the priority to defending plasma volume and then to plasma osmolality.
The data of our study show that hypertonic saline in children with severe head injury significantly increases ADH because of increased plasma osmolality. The fact that ADH positively correlates with sodium intake suggests that head-injured children appropriately respond to a sodium load. The fact that both groups show a positive correlation of ADH and fluids indicates that head-injured patients appropriately release ADH when they are hypovolemic. These findings are of clinical importance in that the practice of fluid restriction may not be appropriate in all children with severe head injury.
1. Khilani P. Electrolyte abnormalities in critically ill children. Crit Care Med 1992; 20: 241–50.
2. Haycock GB. The syndrome of inappropriate secretion of antidiuretic hormone. Pediatr Nephrol 1995; 9: 375–81.
3. Doczi T, Tarjanyi J, Huszka E, Kiss J. Syndrome of inappropriate secretion of antidiuretic hormone (SIADH) after head injury. Neurosurgery 1982; 10: 685–8.
4. Born JD, Hans P, Smitz S, et al. Syndrome of inappropriate secretion of antidiuretic hormone after head injury. Surg Neurol 1985; 23: 383–7.
5. Moses AM, Streeten DHP. Disorders of the neurohypophysis. In: Wilson JD, Braunwald E, Isselbacher KJ, et al, eds. Harrison’s principles of internal medicine. New York: McGraw-Hill Inc, 1998: 2003.
6. Robertson GL. Syndrome of inappropriate antidiuresis. N Engl J Med 1989; 321: 538–9.
7. Powell KR, Sugarman LI, Eskenazi AE, et al. Normalization of plasma arginine vasopressin concentrations when children with meningitis are given maintenance plus replacement fluid therapy. J Pediatr 1990; 117: 515–22.
8. Padilla G, Leake JA, Castro R, et al. Vasopressin levels and pediatric head trauma. Pediatrics 1989; 83: 700–5.
9. York J, Arrillaga A, Graham R, Miller R. Fluid resuscitation of patients with multiple injuries and severe closed head injury: experience with an aggressive fluid resuscitation strategy. J Trauma 2000; 48: 376–9.
10. Al-Zahraa Omar F, Al Bunyan M. Severe hyponatremia as a poor prognostic factor in childhood neurologic disease. J Neurol Sci 1997; 151: 213–6.
11. Velasco IT, Pontieri V, Rocha e Silva M, Lopes OU. Hyperosmotic NaCl and severe hemorrhagic shock. Am J Physiol 1980; 239: 664–73.
12. Rocha e Silva M, Velasco IT, Nogueira da Silva RI, et al. Hyperosmotic sodium salts reverse hemorrhagic shock: other solutes do not. Am J Physiol 1987; 253: H751–62.
13. Simma B, Burger R, Falk M, et al. A prospective, randomized, and controlled study of fluid resuscitation in children with severe head injury: lactated Ringer’s solution versus hypertonic saline. Crit Care Med 1998; 26: 1265–70.
14. Safar P. Cerebral resuscitation after cardiac arrest. Circulation 1986; 74: 138–53.
15. Nicholson JF, Pesce MA. Reference ranges for laboratory tests and procedures. In: Behrmann RE, Kliegman RM, Arvin AM, eds. Nelson textbook of pediatrics. Philadelphia: WB Saunders Co, 2000; 2188–90.
16. Kamoi K, Toyama M, Takagi M, et al. Osmoregulation of vasopressin secretion in patients with the syndrome of inappropriate antidiuresis associated with central nervous system disorders. Endocr J 1999; 46: 269–77.
17. Robertson GL, Mahr EA, Athar S, Sinka T. Development and clinical application of a method for the radioimmunoassay of arginine-vasopressin in human plasma. J Clin Invest 1973; 52: 2340–52.
18. Moses AM, Miller M, Streeten HP. Pathophysiologic and pharmacologic alterations in the release and action of ADH. Metabolism 1976; 25: 697–721.
19. Leslie GI, Philips JB, Work J, et al. The effect of assisted ventilation on creatinine clearance and hormonal control of electrolyte balance in very low birth weight infants. Pediatr Res 1986; 20: 447–52.
© 2001 International Anesthesia Research Society
20. Ichikawa I. A bridge over troubled water: mixed water and electrolyte disorders. Pediatr Nephrol 1998; 12: 160–7.