Printer Friendly

Severe dehydration and acute renal failure associated with external ventricular drainage of cerebrospinal fluid in children.


We report three paediatric cases of severe dehydration and hyponatraemia with circulatory compromise associated with the use of external ventricular drainage of cerebrospinal fluid. Two of the children had cardiac arrests. All were successfully resuscitated. While there were additional factors that contributed to other fluid losses, and fluid balance data are incomplete, these cases highlight a need for increased vigilance when managing children with external ventricular drains.

Key Words: dehydration, hyponatraemia, renal failure, cerebrospinal fluid, external ventricular drain, intensive care, paediatric

External ventricular drainage (EVD) of cerebrospinal fluid (CSF) is commonly used to control intracranial pressure in paediatric intensive care (PICU) and neurosurgical units. We report three cases of life-threatening dehydration from fluid losses associated with EVD. Their biochemical derangements were corrected rapidly with no detectable adverse sequelae. Policy changes in our units have been introduced to prevent future occurrences.


These three cases presented as unexpected crises in children who were clinically stable for days preceding the event. All three children were preverbal and none had clinical signs of dehydration detected before they decompensated, needing resuscitation. The table shows their blood biochemistry and total daily fluid balances before and after the adverse events. Vital signs (heart rate, blood pressure and temperature) were in the normal ranges for all three, and no noticeable changes occurred to alert us to the level of dehydration that these children ultimately exhibited.

Case 1

A 6-week-old, 3.8 kg girl with Pfeiffer's syndrome had been admitted previously, aged three days, with central and obstructive apnoea from severe choanal stenosis. Symptoms improved with nasal airway insertion and surgical palatal split. She also had a feeding gastrostomy inserted. Magnetic resonance imaging showed caudal cerebellar herniation, Chiari malformation type 1, obstructive hydrocephalus and cranial synostosis. She underwent cranial vault reshaping and posterior fossa decompression, with planned postoperative PICU admission. Intraoperatively she was transfused 2.5 blood volumes of crystalloid, packed red cells, fresh frozen plasma, cryoprecipitate and concentrated platelets. Postoperatively she had arterial, central venous, peripheral venous (IV) and urinary catheters. She failed extubation on day three. Subsequent CT scan showed hydrocephalus, and she returned to the operating theatre for the insertion of an EVD, set at the level of the tragus. She was extubated, uneventfully, two days later. Gastrostomy feeding was established as IV fluids were weaned and ceased, and we removed her arterial, urinary and central venous catheters. She remained in PICU for EVD monitoring and antibiotic prophylaxis with cephalothin. She was irritable, a common finding in infants following posterior fossa surgery.

Twelve days after the original surgery, she developed mild vomiting and diarrhoea that was managed without IV fluids. Her vital signs remained within normal limits. In retrospect some of her urine output" may have been fluid stool. On the 14th postoperative day a doctor could not place a peripheral IV cannula for maintenance fluids while she was fasted for the planned removal of the EVD. Skin turgor was noted to be poor, but her vital signs were unchanged on baseline. Fasting commenced at 0400h and there was no IV access obtained before her arrest at 0800, when she had a monitored episode of sustained ventricular tachycardia with hypotension. Resuscitation included endotracheal intubation and ventilation, intra-osseous saline 40 ml/kg 0.9% and reinstitution of invasive monitoring. She reverted to sinus rhythm after intubation, without cardioversion. Serum potassium measured during the arrest was greatly elevated (11.8 mmol/1) and serum sodium was very low (112 mmol/1). Urea and creatinine were also high (Table 1). Calcium chloride and sodium bicarbonate were given. The potassium corrected rapidly with fluid administration, insulin and dextrose. The hyponatraemia was deliberately corrected more slowly with IV fluids (Table 1).

She was extubated to spontaneous ventilation in room air the day after her arrest. The EVD was then raised to 10 cm and was removed 48 hours later. She made a full recovery, undergoing further craniofacial surgery a week later without need for an EVD.

Case 2

A 22-month-old, 10.6 kg girl with Crouzon's disease was admitted electively to PICU after cranial vault reshaping and fronto-orbital advancement. A similar operation had been performed at age ten months. During the second operation she been transfused four blood volumes of crystalloid, packed red cells, cryoprecipitate, fresh frozen plasma and concentrated platelets. She remained intubated and ventilated postoperatively and was admitted to the PICU. Monitoring included intra-arterial and central venous pressures, urinary output via catheter, ECG and oxygen saturation. An EVD was set at 5 cm above the tragus. She was extubated on day two and remained in PICU for EVD management pending evaluation for a ventrccuao-peritoneal (VP) shunt.

IV vancomycin and cefotaxime were given for four days, followed by cephalexin for the duration of the use of the EVD. By day five postoperatively she was sitting up, active and alert, playing and smiling, eating solids and drinking milk. The urinary and central venous catheters were removed. EVD drainage was 15 to 20 ml per hour. On day six she developed vomiting and diarrhoea, postponing the VP shunt insertion. IV fluids were not deemed necessary, enteral feeding continued with human milk and the symptoms subsided within 48 hours. Blood tests were not performed. On day eight she was playful and interactive. During the night, at 0400h and without any warning, she had an ECG-monitored asystolic cardiac arrest. She was resuscitated with external cardiac massage, endotracheal intubation, ventilation, IV adrenaline and IV 0.9% saline 15 ml/kg. Sinus rhythm was restored within three minutes. Blood gas analysis showed hyperkalaemia (7.6 mmol/1) and she was given calcium chloride and sodium bicarbonate. Other significant biochemical abnormalities are detailed in Table 1 and these normalized quickly with IV fluid administration. A VP shunt was inserted two days later. She was discharged from hospital two weeks later with no detectable adverse sequelae.

Case 3

A 12-month-old, 7.4 kg boy presented with a two-week history of fever. His medical history included Beare-Stevenson syndrome, hydrocephalus and VP shunt, choanal atresia, laryngomalacia, vesicoureteric reflux, tracheostomy and gastrostomy. His temperature was 40[degrees]C, he had a cough and increased tracheal secretions. There was no vomiting or diarrhoea and his hydration state and fontanelle level were both normal. VP shunt aspiration produced blood-stained CSF with an elevated cell count and gram positive cocci on microscopy. Staph. epidermidis grew in culture. Blood tests showed hyponatraemia with normal urea and creatinine (Table 1). He was treated with IV fluids, ceftriaxone and vancomycin. Twelve hours after admission the distal end of his VP shunt was externalized and connected to a closed drainage system. A shunt valve remained attached. In the immediate postoperative period the collecting chamber was level with the tragus. The following day it was lowered a further 5 cm to improve drainage.

For the first six days he was managed in the neurosurgical ward. The fever dissipated within 48 hours, so IV fluids were stopped and he was fed 140 ml/kg/day via gastrostomy. IV antibiotics were continued. He vomited frequently after bolus feeds, but had no diarrhoea. A fluid balance chart was updated hourly. The recording of "fluid in" included IV fluids and feeds. The recording of "fluid out" included the volumes of CSF drainage and gastric loss from the vomiting. Volumes of urine and stool were not recorded. On day five IV fluids were recommenced because of frequent vomiting. On day six plasma electrolytes, measured for the first time following externalization of the shunt, showed hyponatraemia, hyperkalaemia, hypochloraemia and markedly elevated urea and creatinine (Table 1). He had lost 400 grams, equivalent to 6% of his body weight. He was transferred to PICU and given 20 ml/kg of 0.9% saline as an IV bolus, followed by an infusion of 0.9% saline with 5% dextrose. Gastrostomy feeding was stopped. Electrolyte abnormalities corrected over 48 hours and renal function normalized over four days. His subsequent clinical course was uneventful. He was discharged after the insertion of a new VP shunt.


By reporting these cases we render ourselves open to criticism for our apparent lack of vigilance with fluid management. All three cases are examples of insidious onset of acute renal failure with dramatic changes in serum biochemistry that were not suspected clinically. We could find only one previous report of a similar event in association with EVD management (1), and one description of hyponatraemia associated with repeated CSF drainage in hydrocephalus (2). On this basis we thought these cases worthy of report.

We considered the associated renal failure to be pre-renal in origin, caused by severe dehydration from EVD and gastrointestinal losses that were underestimated. Other potential causes, such as drug toxicity or sepsis, are unlikely in the absence of abnormal urinary sediment and in view of the rapid recoveries.

Pathophysiology of EVD fluid losses

These children hadEVD insertion for the prevention or treatment of raised intracranial pressure. In cases 1 and 2 this was done as part of extensive cranial vault surgery with dural perforation, to accommodate perioperative swelling, subarachnoid blood and debris which may temporarily impair CSF circulation. Case 3 had an existing shunt externalized because of infection. The volume of CSF lost to open drainage was considerable and continued in the presence of dehydration. The control of EVD fluid losses is related to the level at which the EVD is set. Other factors that influence EVD output include age, weight, site of EVD insertion, level of activity and gender (3,4) In these cases the EVD was placed at a level lower than normal CSF pressure to encourage drainage. Sodium ions are actively transported across the epithelia of the choroid plexus and ependyma in the production of CSF, with secondary active transport of chloride and water because of electrical and osmotic gradients. The electrolyte composition of CSF is remarkably constant, with a sodium concentration of approximately 150 mmol/1, regardless of serum levels (5). Excessive EVD losses will therefore deplete total body sodium, chloride and water.

The hyponatraemia, hyperkalaemia and uraemia exhibited by these children are consistent with severe dehydration and sodium loss. Measured EVD losses were substantial, but all three cases had additional gastrointestinal fluid losses, most of which escaped accurate measurement. The relative contribution of gastrointestinal losses to their deterioration is unquestionable, but there is inadequate information available to draw specific conclusions because of the incomplete fluid balance measurements. High levels of serum antidiuretic hormone (ADH) have been reported in children with gastroenteritis and hyponatraemia, and this may have contributed in cases 1 and 2 (6). Stool specimens were sent, but no pathogens were found. In all three cases the vomiting may have resulted from intracranial pressure disturbances from excessive CSF losses, but this was not specifically investigated.

Syndromes of inappropriate antidiuretic hormone and cerebral salt wasting are associated with cranial vault reconstructive surgery (7-9). Polyuria is a feature of cerebral salt wasting, while in the syndromes of inappropriate antidiuretic hormone, oliguria is more common, although not universal. In both syndromes the renal fractional excretion of sodium is elevated (7-9). Cerebral salt wasting may have contributed to early sodium depletion in cases 1 and 2, but spontaneous resolution is usual within five days from the time of operation (8). Both children received 0.9% saline as their principal IV fluid postoperatively, and serum sodium was normal for the first five postoperative days. Up until the 24-hour period prior to arrest these children had normal urine flow rates. They were both oliguric on the day of collapse. Their arrests were distant from the time of surgery, being day 15 post-surgically for case 1, and day 10 for case 2. All three cases had low urinary sodium levels at the time of their recognised crisis. We think that the marked hyponatraemia and hypochloraemia in these children reflects the losses of these ions by combined cerebral salt wasting drainage and gastrointestinal output, and were not related to cerebral salt wasting. The absence of neurological signs in the presence of profound hyponatraemia would suggest gradual development.

Acidosis and hyperkalaemia are the most likely causes of the cardiac arrhythmias in cases 1 and 2. In case 3 the bicarbonate was elevated initially and may reflect acid loss from protracted vomiting. In all three cases, serum potassium was low once total body water had been replenished, necessitating additional potassium supplementation. We attribute this finding to urinary potassium losses under the influence of aldosterone, and additional gastrointestinal losses in the preceding days. Timely IV fluid resuscitation averted the need for renal replacement therapy in all cases.

Conspiratorial nature of adverse events

Adverse events often result from a series of small oversights that are individually insignificant but collectively disastrous. Systematic errors and human factors may occur concurrently, but in most circumstances they will not all occur together in the one patient, so the potential for catastrophe is averted (10,11). We confined our analysis of cases 1 and 2 to local unit audit until we learned of case 3 in another hospital, signifying a need for wider reporting to heighten general awareness of the potential adverse outcomes associated with EVD in small children. A number of system-based events and human factors contributed to the final outcomes:

* EVD drainage was extended longer than necessary in cases 1 and 2, due to surgical schedules and a public holiday period.

* EVD levels were possibly set too low, encouraging CSF loss. Partial absorption may occur if set higher, reducing total losses.

* Combined CSF and gastrointestinal fluid losses were underestimated and reliance on enteral fluid intake in this setting was probably unwise.

* Fluid losses occurred slowly so that compensatory physiological changes minimized the clinical signs, until no further compensation was possible.

* Measured daily fluid balances were positive in each case on every day except one (case 2) in the week preceding the crisis, indicating their relative inaccuracy.

* Our reluctance to take blood samples frequently in infants who look well is based on avoidance of sampling anaemia and pain.

Policy changes may minimize future risk

Following these cases we have implemented the following changes:

* EVD losses greater than 0.5 ml/kg/h will be replaced intravenously with 0.9% saline + 5 mmol/1 KCI, independently of maintenance fluids.

* Urea, creatinine and electrolytes are to be measured at least twice weekly in all children with EVD, and more frequently if other fluid losses are noted.

* We will remove EVDs as soon as clinically indicated.


These three cases demonstrate that life-threatening dehydration and biochemical derangements can occur in association with EVD use in paediatrics. Concurrent gastrointestinal fluid losses were undoubtedly contributory to the overall clinical syndromes and the inaccuracy of measured fluid balance monitoring in these cases was misleading. Of note there were no standout premonitory clinical signs detected. Other systematic factors may have contributed to the events reported. A high index of suspicion needs to be maintained when supervising EVDs in young children to avoid serious adverse events.

Accepted for publication on May 23, 2006.


(1.) Tobias JD. Cerebrospinal fluid losses through ventricular catheters leading to hyponatremia in two children. South Med J 1991; 84:279-280.

(2.) MacMahon P, Cooke RW Hyponatraemia caused by repeated cerebrospinal fluid drainage in post haemorrhagic hydrocephalus. Arch Dis Child 1983; 58:385-386.

(3.) Drake JM, Sainte-Rose C, DaSilva M, Hirsch JE Cerebrospinal fluid flow dynamics in children with external ventricular drains. Neurosurgery 1991; 28:242-250.

(4.) Yasuda T, Tomita T, McLone DG, Donovan M. Measurement of cerebrospinal fluid output through external ventricular drainage in one hundred infants and children: correlation with cerebrospinal fluid production. Pediatr Neurosurg 2002; 36:22-28.

(5.) Venkatesh B, Scott P, Ziegenfuss M. Cerebrospinal fluid in critical illness. Critical Care and Resuscitation 2000; 2:42-54.

(6.) Neville KA, Verge CF, O'Meara MW, Walker JL. High antidiuretic hormone levels and hyponatremia in children with gastroenteritis. Pediatrics 2005; 116:1401-1407.

(7.) Lee SJ, Huh EJ, Byeon JH. Two cases of cerebral salt wasting syndrome developing after cranial vault remodeling in craniosynostosis children. J Korean Med Sci 2004; 19:627-630.

(8.) Byeon JH, You G. Cerebral salt wasting syndrome after calvarial remodeling in craniosynostosis. J Korean Med Sci 2005; 20:866-869.

(9.) Levine JP, Stelnicki E, Weiner HL, Bradley JP, McCarthy JG. Hyponatremia in the postoperative craniofacial pediatric patient population: a connection to cerebral salt wasting syndrome and management of the disorder. Plast Reconstr Surg 2001;108:1501-1508.

(10.) Allnutt MF. Human factors in accidents. Br J Anaesth 1987; 59:856-864.

(11.) Runciman WB, Sellen A, Webb RK et al. The Australian Incident Monitoring Study. Errors, incidents and accidents in anaesthetic practice. Anaesth Intensive Care 1993; 21:506-519.

S. SIMPSON *, M. YUNG [dagger], A. SLATER [double dagger]

Paediatric Intensive Care Unit, Women's and Children's Hospital, Adelaide, South Australia and Queensland Paediatric Intensive Care Service, Royal Children's Hospital, Brisbane, Queensland, Australia

* M.B., B.S., D.A.(UK), F.A.N.Z.C.A., F.F.P.M.A.N.Z.C.A., Senior Registrar, Paediatric Intensive Care Unit, Women's and Children's Hospital, Adelaide, South Australia.

[dagger] M.D., F.R.A.C.P., F.J.F.I.C.M., Consultant, Paediatric Intensive Care Unit, Women's and Children's Hospital, Adelaide, South Australia.

[double dagger] F.R.A.C.P., F.J.F.I.C.M., Director, Queensland Paediatric Intensive Care Service, Royal Children's Hospital, Brisbane, Queensland.

Address for reprints: Dr S. Simpson, Department of Critical Care Medicine, Hospital for Sick Children, 555 University Avenue, Toronto M5G 1X8, Canada.



PICU Day 9 12 13 14

INPUT (ml)

IV fluids/
drugs 48.8 37.9 4.7

PEG/ Oral daily total 479 609 449

OUTPUT-measured (ml)

EVD/CSF daily total 134 122 94

Urine ml/kg/h 0.89 1.97 0.56

 daily total
 (ml) 82 180 51

OUTPUT - unmeasured

No of vomits x11 x4

No of stools 121ml x4 x2

Fluid (as measured) 191 345 311


Bloods RR (FOR AGE)

Na+ 130-145 mmol/l 135

K+ 35-6.5 mmol/l 6.1

Cl- 100-111 mmol/l 100

HCO3- 18-26 mmol/l

Anion Gap 4-18

Creatinine 0.02-0.07
 mmol/l 0.027

Urea 1.8-7.1 mmol/l 1.8

Calcium 22-2.7 mmol/l

Mg++ 0.65-1.05

P04 1.3-2.4 mmol/l

Hb 120-180 g/l 135

Platelets 150-450
 x10*9/l 348

WBC 6-15 x 10*9/l 7.16


Na+ mmol/l

K+ mmol/l

Cl- mmol/l

Osmolality mmol/kg


PICU Day ARREST 16 17 1

INPUT (ml)

IN fluids/
drugs 435 470 192

PEG/ Oral daily total 275 156 292

OUTPUT-measured (ml)

EVD/CSF daily total 38 34 8

Urine ml/kg/h 1.43 4.92 4.32

 daily total
 (ml) 123 423 372

OUTPUT - unmeasured

No of vomits x6

No of stools x3 x1

Fluid (as measured) 549.9 169 104


Bloods RR (FORAGE)

Na+ 130-145 mmol/l 112 128 135 144

K+ 35-6.5 mmol/l 11.8 35 3.6 3.5

Cl- 100-111 mmol/l 69 95 104 109

HCO3- 18-26 mmol/l 16.4 20.6 22.3 23.6

Anion Gap 4-18 33.3 15.9 12.3 14.9

Creatinine 0.02-0.07
 mmol/l 0.221 0.05 0.03 0.035

Urea 1.8-7.1 mmol/l 25.3 12.7 1.8 3.6

Calcium 22-2.7 mmol/l 218 2.17

Mg++ 0.65-1.05
 mmol/l 0.75 0.52

P04 13-2.4 mmol/l 2.88 1.47

Hb 120-180 g/l 138 85 86 116

Platelets 150-450
 x10*9/l 1310 886 701 112

WBC 6-15 x 10*9/l 19.6 16.3 7.1 534


Na+ mmol/l 15 10

K+ mmol/l 75 NA

Cl- mmol/l 15 15

Osmolality mmol/kg 343 306

 Case 2

PICU Day 5 6 7 8

INPUT (ml)

IN fluids/
drugs 209 0 6.7 5.2

PEG/ Oral daily total 580 544 645 610

OUTPUT-measured (ml)

EVD/CSF daily total 339 408 369 329

Urine ml/kg/h 1.76 0.72 0.5 0.26

 daily total
 (ml) 422 173 120 63

OUTPUT - unmeasured

No of vomits

No of stools x3 81ml

Fluid (as measured) 27.6 -37 81.7 223


Bloods RR (FORAGE)

Na+ 130-145 mmol/l

K+ 35-6.5 mmol/l

Cl- 100-111 mmol/l

HCO3- 18-26 mmol/l

Anion Gap 4-18

Creatinine 0.02-0.07

Urea 1.8-7.1 mmol/l

Calcium 22-2.7 mmol/l

Mg++ 0.65-1.05

P04 13-2.4 mmol/l

Hb 120-180 g/l

Platelets 150-450

WBC 6-15 x 10*9/l


Na+ mmol/l

K+ mmol/l

Cl- mmol/l

Osmolality mmol/kg

 Case 2

PICU Day 9 ARREST 11 12

INPUT (ml)

IN fluids/
drugs 8.6 671 1566 987

PEG/ Oral daily total 452 63 167 408

OUTPUT-measured (ml)

EVD/CSF daily total 232 269 428 381

Urine ml/kg/h 0.41 1.71 3.31 1.84

 daily total
 (ml) 105 435 842 467

OUTPUT - unmeasured

No of vomits x11 x3

No of stools x1 x1

Fluid (as measured) 313 30 463 527


Bloods RR (FORAGE)

Na+ 130-145 mmol/l 126 139 138

K+ 35-6.5 mmol/l 7.6 2.6 3.7

Cl- 100-111 mmol/l 88 112 107

HCO3- 18-26 mmol/l 15.4 20 20.2

Anion Gap 4-18 30.2 9.6 14.5

Creatinine 0.02-0.07
 mmol/l 0.17 0.028 0.033

Urea 1.8-7.1 mmol/l 23.7 4.9 1.2

Calcium 22-2.7 mmol/l 2A2 2.16

Mg++ 0.65-1.05
 mmol/l 1.17 0.8

PO4 13-2.4 mmol/l 2.72 0.91

Hb 120-180 g/l 92 72 71

Platelets 150-450
 x10*9/l 1180 980 807

WBC 6-15 x 10*9/l 22.1 15 11.6


Na+ mmol/l 11 182

K+ mmol/l 95 37

Cl- mmol/l 22 230

Osmolality mmol/kg



 1 4 5 EVENT 7 8 9

 354 105 40 700 662 48

 360 1225 994 100 226 789

OUTPUT - measured (ml)

EVD/CSF 215 279 276 163 52 94

Gastric 15 nr 420 45 417 133
loss (ml)

No of wet
nappies x2 x4 x8 x4

OUTPUT - unmeasured

No of vomits x3 x14 nr x8 0 x1

No of stools x1 x2 0 x2 x4 x2

 484 inc 338 592 419 610



133-143 mmoW 128 123 130 138 135

32-4.5 mmol/l 3.3 6 33 2.9 4.1

97-108 mmol/l 76 64 84 100 98

22-33 mmol/l 35 29 23 19 16

4-13 mmol/l 16 30 23 19 16


mmol/l <0.03 0.205 0.1 0.05 0.04

2.0-7.0 mmol/l 6.2 36 33.4 21.1 10.8

mmol/l 2.41 238 2.41


1.0-1.6 mmol/l 1.47 3.17 1.88

120-180 g/l 101 102 93

150-450 x10*9/l 1113 1509 1205

6-15 x 10*9/l 26.9 16.9 16




COPYRIGHT 2006 Australian Society of Anaesthetists
No portion of this article can be reproduced without the express written permission from the copyright holder.
Copyright 2006 Gale, Cengage Learning. All rights reserved.

Article Details
Printer friendly Cite/link Email Feedback
Title Annotation:Case Report
Author:Simpson, S.; Yung, M.; Slater, A.
Publication:Anaesthesia and Intensive Care
Date:Oct 1, 2006
Previous Article:Association of obesity with increased mortality in the critically ill patient.
Next Article:Short term low-dose propofol anaesthesia associated with severe metabolic acidosis.

Related Articles
Patients with severe renal dysfunction ok to receive VAD as bridge to heart transplant.
Heart failure.
Role of albumin peritoneal dialysis for bilirubin removal after complicated liver transplant.
Posterior fossa haematoma following sudden decompression of acute hydrocephalus resulting from neuroendoscopy.

Terms of use | Privacy policy | Copyright © 2019 Farlex, Inc. | Feedback | For webmasters