Article Text
PDF
Abstract
Objective To compare the difference in plasma sodium at 16–18 h following major surgery in children who were prescribed either Hartmann's and 5% dextrose or 0.45% saline and 5% dextrose.
Design A prospective, randomised, open label study.
Setting The paediatric intensive care unit (650 admissions per annum) in a tertiary children's hospital in Brisbane, Australia.
Patients The study group comprised 82 children undergoing spinal instrumentation, craniotomy for brain tumour resection, or cranial vault remodelling.
Interventions Patients received either Hartmann's and 5% dextrose at full maintenance rate or 0.45% saline and 5% dextrose at two-thirds maintenance rate.
Main outcomes measures Primary outcome measure: plasma sodium at 16–18 h postoperatively; secondary outcome measure: number of fluid boluses administered.
Results Mean postoperative plasma sodium levels of children receiving 0.45% saline and 5% dextrose were 1.4 mmol/l (95% CI 0.4 to 2.5) lower than those receiving Hartmann's and 5% dextrose (p=0.008). In the 0.45% saline group, seven patients (18%) became hyponatraemic (Na <135 mmol/l) at 16–18 h postoperatively; in the Hartmann's group no patient became hyponatraemic (p=0.01). No child in either fluid group became hypernatraemic.
Conclusions The postoperative fall in plasma sodium was smaller in children who received Hartmann's and 5% dextrose compared to those who received 0.45% saline and 5% dextrose. It is suggested that Hartmann's and 5% dextrose should be administered at full maintenance rate postoperatively to children who have undergone major surgery in preference to hypotonic fluids.
Statistics from Altmetric.com
Introduction
The optimal management of fluid and electrolytes in hospitalised children has been the subject of recent debate.1,–,3 There are numerous case reports of severe hyponatraemia following elective surgical procedures in children, which have resulted in long-term neurological sequelae.4 This potential for postoperative hyponatraemia and its associated morbidity has challenged the traditional approach to paediatric intravenous fluid therapy which advocated the use of hypotonic dextrose/saline solutions.5 6 A systematic review of fluid therapy in hospitalised children confirmed that current prescribing practices are based on limited evidence.7
Antidiuretic hormone (ADH) is secreted by the hypothalamus and controls water metabolism by increasing free water absorption at the distal renal tubules and collecting ducts.8 Several studies have found that administration of hypotonic dextrose/saline (traditional paediatric intravenous therapy) predisposed postoperative patients to the development of hyponatraemia associated with excess ADH release, whereas isotonic saline protected patients from hyponatraemia.9,–,11 A study of 20 children undergoing posterior spinal fusion surgery observed the syndrome of inappropriate ADH secretion (SIADH) in five patients (25%), although no child had a serum sodium lower than 130 mmol/l.12 In another study of 30 children following spinal fusion, SIADH was diagnosed in 10 children (33%) on clinical criteria and serum ADH levels were highest in these patients.13
What is already known on this topic
▶ Children in hospital with various medical and surgical illnesses have raised serum antidiuretic hormone levels, which predisposes them to the development of hyponatraemia.
▶ There are reports of hyponatraemic encephalopathy in hospitalised children receiving hypotonic intravenous fluids. Some authors suggest that children should receive isotonic fluids, despite concerns of hypernatraemia.
What this study adds
▶ The plasma sodium fell less when Hartmann's and 5% dextrose was administered compared with 0.45% saline and 5% dextrose.
▶ Significantly more children who received 0.45% saline and 5% dextrose became hyponatraemic.
▶ No child receiving Hartmann's and 5% dextrose at full maintenance rate became hypernatraemic.
There has recently been a change in maintenance fluid administration in many paediatric intensive care units on the basis of case reports of severe symptomatic hyponatraemia, but there is little evidence to support these changes.4 Our aim was to investigate whether the change in postoperative plasma sodium levels in children undergoing surgery would be different in children who received Hartmann's and 5% dextrose (at full maintenance rate) compared with children who received 0.45% saline and 5% dextrose (at two-thirds maintenance rate).
Methods
The study was conducted at the Royal Children's Hospital (RCH), Brisbane, Australia between September 2006 and December 2008. RCH is a tertiary paediatric referral centre servicing a population of approximately 1.5 million. The paediatric intensive care unit (PICU) has eight beds and admits approximately 650 patients annually from all subspecialties except cardiac surgery, and including oncology, burns, trauma and liver transplantation.
Children were eligible for enrolment if they were admitted to the PICU following spinal instrumentation for correction of scoliosis, craniotomy for excision of brain tumours and cranial vault remodelling. Children were not recruited if they were undergoing lengthening only of spinal instrumentation rods, insertion or revision of ventriculoperitoneal shunts, intracerebral cyst fenestration, or were previously enrolled in the study.
Participants were allocated to receive one of the maintenance fluids under investigation: either Hartmann's and 5% dextrose solution (AHB2074, Baxter, Sydney, Australia) at full paediatric fluid maintenance requirements or 0.45% saline and 5% dextrose solution (AHK6028, Baxter, Sydney, Australia) at two-thirds of paediatric fluid maintenance requirements. In all cases intravenous fluids were continued over the first postoperative night; however, if patients tolerated oral fluids, the intravenous fluid rate was decreased accordingly. Children were consistently managed and cared for according to spinal and craniotomy clinical care pathways, which standardise all aspects of medical care (see supplemental web file). A checklist outlined parameters to guide postoperative fluid bolus administration including capillary refill time, heart rate, blood pressure, hourly urine output and wound or drain losses (see supplemental web file).
Our primary outcome was the plasma sodium concentration 16–18 h postoperatively. Sodium was measured in whole blood using a potentiometric technique with a sodium ion selective electrode on the Radiometer ABL725 Blood Gas Analyser in the PICU (normal range plasma Na 135–145 mmol/l). At a sodium level of 126 mmol/l, the assay coefficient of variation is 0.36%, giving 95% confidence limits of 126±1 mmol/l. At a sodium level of 160 mmol/l, the assay coefficient of variation is 0.27%, giving 95% confidence limits of 160±1 mmol/l. Secondary outcomes included serum ADH levels 1–4 h postoperatively and the number and volume of fluid boluses (10 ml/kg) administered to 16–18 h postoperatively. ADH levels were measured on EDTA plasma by double antibody radioimmunoassay (Bühlmann Laboratories, Schönenbuch, Switzerland). We collected demographic, clinical and laboratory data. Plasma sodium levels were measured preoperatively (usually on admission the day before surgery), within 1–4 h postoperatively and 16–18 h postoperatively. Urine electrolytes (sodium and potassium), creatinine and osmolality were collected from an indwelling urinary catheter over 1 h, within 1–4 h postoperatively and 16–18 h postoperatively. Due to a miscommunication with the laboratory, urine from the first 20 children was collected and tested for urine electrolytes and creatinine, but not for osmolality. Sodium was measured in urine using a potentiometric technique with a sodium ion selective electrode on the Beckman Coulter Synchron DXC800 automated chemistry analyser at the central pathology laboratory. At a sodium level of 86 mmol/l, the assay coefficient of variation is 0.88%, giving 95% confidence limits of 86±2 mmol/l. At a sodium level of 179 mmol/l the assay coefficient of variation is 1.15%, giving 95% confidence limits of 179±4 mmol/l.
Participants were allocated using computer generated random numbers in blocks of 10. Randomisation was stratified by type of surgery (spinal or cranial). Treatment allocations were contained in consecutively numbered opaque envelopes. Patients and caregivers were masked to treatment. Families were approached prior to surgery by a study team member, and informed, written consent was obtained from the caregivers. The study received approval from the Human Research Ethics Committee of the Royal Children's Hospital, Brisbane, and was registered with the Australian New Zealand Clinical Trials Registry (ACTRN12607000116426).
To detect a difference in sodium levels between groups of 2.0 mmol/l, with α=0.05, power=80%, and assuming an SD of 3.0 mmol/l, we required 36 individuals in each treatment group. We compared biochemical data between treatment groups at 16–18 h postoperatively. We used a repeated-measures regression analysis to adjust for biochemical data at the start of the trial (1–4 h postoperatively). We included time and treatment group as covariables, and included a time-by-treatment interaction term to determine whether the magnitude of change over time differed between treatment groups. We compared non-parametric variables using the Wilcoxon rank test and categorical variables using Fisher's exact test. Analysis was conducted on an intention to treat basis. Significance values were based on two-tailed tests, with p<0.05 considered significant.
Results
All participants who were randomised received their assigned study fluid, and the treatment of all children conformed to their prescribed clinical pathway. No patients withdrew after entering the study. Of 135 patients screened, 104 met the eligibility criteria for the study (figure 1); 82 agreed to participate, 4 refused, and for 18 we were unable to obtain consent in time (16 had a diagnosis of brain tumour and were scheduled for urgent surgery). The 22 children who did not participate were similar to the participants in terms of age and sex (48% of participants were male with a median (IQR) age of 136 (60–168) months; 36% of non-participants were male with a median (IQR) age of 78 (48–144) months; p=0.47 for sex; p=0.12 for age; figure 1).
Participant flow diagram.
Baseline demographics and clinical characteristics for all 82 randomised patients are shown in table 1 and were balanced across the treatment groups. None of these patients were invasively or non-invasively ventilated postoperatively (table 1).
Baseline characteristics by postoperative fluid type
Table 2 shows postoperative biochemistry and haematology data. Three patients were excluded from analysis (Hartmann's group: n=2; 0.45% saline group: n=1) as they were discharged early to the ward. Sodium levels in the Hartmann's group (mean group difference = 1.4 (95%CI 0.4 to 2.5)) were significantly higher at 16–18 h postoperatively than sodium levels in the 0.45% saline group. In the 0.45% saline group, 7 patients (18%) were hyponatraemic (Na <135 mmol/l) at 16–18 h postoperatively, whereas in the Hartmann's group no patient was hyponatraemic (p=0.01). The number needed to treat with Hartmann's to prevent one case of hyponatraemia was 6 (95% CI 3 to 20). Of children who became hyponatraemic, three had plasma sodium levels of 134 mmol/l, three had plasma sodium levels of 133 mmol/l and one had a plasma sodium of 130 mmol/l. No child in either group became hypernatraemic (Na >145 mmol/l) at 16–18 h postoperatively. The type of surgery did not have a significant effect on postoperative sodium levels for either fluid regimen.
Plasma biochemistry values measured at 16–18 h by postoperative fluid type
ADH levels rose from pre- to post-surgery in both the Hartmann's (median (IQR)=1.0 pmol/ml (0.4 to 3.2)) and 0.45% saline groups (1.2 pmol/ml (−0.2 to 4.6)), but ADH levels did not differ significantly between treatment groups 1–4 h postoperatively (median (IQR)) in Hartmann's group was 2.0 (1.7 to 5.6) vs 2.4 (1.5 to 6.0) in the 0.45% saline group; p=0.79). There were no differences between groups for plasma or urine osmolality or sodium (table 3).
Fluid and tonicity balance data at 16–18 h postoperative by fluid type
Table 3 shows fluid and tonicity balance data results by postoperative fluid type. There was no difference in the proportion of patients receiving any fluid boluses (46% in Hartmann's group vs 43% in 0.45% saline group; p=0.82), or in the type or total amount of postoperative bolus fluid received, calculated either as a volume (ml/kg) (p=0.16) or amount of sodium (mmol/kg) (p=0.10).
Discussion
Our study shows that children undergoing spinal instrumentation surgery, craniotomy or cranioplasty who received Hartmann's and 5% dextrose were less likely to become hyponatraemic. This group of postoperative patients has traditionally been observed in our PICU on the first postoperative night, when patients are closely monitored, with particular attention to cessation of bleeding, careful assessment of haemodynamic status, monitoring of urine output and provision of adequate analgesia. The spinal and craniotomy clinical care pathways were developed to standardise and optimise the postoperative care of these patients. A major strength of this study is that all children received standardised care prescribed by the clinical care pathway (see supplemental web file). The medical and nursing staff were not masked to the postoperative fluid prescribed; however, the primary outcome measure, postoperative plasma sodium measured at 16–18 h, is objective. In our study plasma sodium levels were not routinely monitored following PICU discharge to the ward; however, there were no cases in which the clinical situation dictated measurement of further postoperative electrolytes, suggesting that there were no instances where children became seriously hyponatraemic following discharge to the ward. We chose to administer the 0.45% saline and 5% dextrose at two-thirds of traditional maintenance rate14 because the routine practice in our PICU was to limit the administration of hypotonic fluids in ‘sick patients’ to prevent hyponatraemia secondary to the water retaining effects of stress hormones.15 However, we believe that it makes little sense to limit fluids to a non-ventilated child who has just experienced major blood loss during surgery, so we chose to deliver the Hartmann's and 5% dextrose at full maintenance rate.
A secondary outcome measure was the amount and number of fluid boluses administered, and a checklist was used to ensure appropriate decision-making for postoperative fluid bolus administration (see supplemental web file). The extent of the intraoperative fluid and blood replacement was similar between groups, and further, both groups received a similar number of postoperative crystalloid (normal saline) and colloid (4% normal serum albumin, Commonweatlh Serum Laboratories, Melbourne, Australia) fluid boluses (10 ml/kg). Indeed, despite the prescription of full maintenance fluid rate for the Hartmann's group and 70% (two-thirds) maintenance fluid rate for the 0.45% saline group, both the total fluid input and 16–18 h fluid balance were similar between the two groups. This underscores the challenge of restricting fluid input to prevent hyponatraemia even in the closely monitored PICU setting and suggests the administration of a ‘saltier’ isotonic fluid is a safer and more practical approach, particularly on the general paediatric ward. Analysis of the urine sodium and urine osmolality data revealed similar urine sodium excretion and urine osmolality between the two groups of patients, suggesting that the type of intravenous fluid administered is more important for maintenance of normonatraemia than urinary sodium losses. These findings contrast with children treated with either normal saline or half normal saline intravenously for viral gastroenteritis who excreted more sodium when normonatraemic at presentation.16
The recent debate on postoperative fluid management in children has raised concerns about the development of hypernatraemia if isotonic solutions are administered; however, we did not observe hypernatraemia in our study.17 18 This study was conducted in children undergoing major surgical procedures with significant surgical stress and considerable expected blood loss; therefore ensuring an adequate non-osmotic stimulus for secretion of ADH and other stress hormones. Indeed, following spinal surgery or craniotomy during which there is major surgical stress and transfusion of blood and blood products, it could be argued that the release of stress hormones, including non-osmotic ADH secretion, is not inappropriate, and classifying postoperative hyponatraemia due to inappropriate and chronic causes of raised ADH is not logical. We observed a preoperative to postoperative rise in ADH levels, which confirmed our patients were subjected to significant surgical stress. In one study, the fall in serum sodium matched the numbers of spinal segments operated on, and thus appeared to be in proportion with the surgical stress.19 This is likely due to a rise in the plasma levels of ADH postoperatively, which peak within the first few postoperative hours.13 20 An observational study in an Australian children's hospital noted that ADH levels are high in children admitted with viral gastroenteritis,21 and a subsequent prospective randomised study in the same hospital found less hyponatraemia in children admitted with viral gastroenteritis, who received 0.9% saline compared with 0.45% saline.16 The same authors randomised 124 hospitalised children undergoing surgery to receive either normal saline (0.9% saline) at 100% or 50% maintenance rate, or half normal saline (0.45% saline) at 100% or 50% maintenance rate in the postoperative period. The plasma ADH concentrations rose two- to four-fold during surgery, and multivariate regression analysis demonstrated that fluid type, rather than rate, determined the postoperative risk of hyponatraemia.22 In a more recent open label study, hospitalised children receiving intravenous fluids for at least 24 h were randomised to one of three groups: (A) 0.9% saline (normal saline) at standard maintenance rate; (B) 0.18% saline and 5% dextrose at standard maintenance rate; or (C) 0.18% saline and 5% dextrose at two-thirds maintenance rate. The incidence of hyponatraemia (plasma Na <130 mmol/l) was 1.7% (1/58), 14.3% (8/56) and 3.8% (2/53), respectively, and although eight patients developed asymptomatic hypernatraemia, this did not differ between fluid groups.23
The findings in this study represent those of a single PICU in a specific population of children undergoing routine surgical procedures and may not be applicable to children admitted to hospital with medical illnesses or undergoing other types of surgical procedures. However, a study of intravenous maintenance fluids in another Australian PICU, which included postoperative patients, found that dextrose/0.18% saline administration resulted in a greater fall in sodium than dextrose/0.9% saline, and that fluid administration at full maintenance rate resulted in a greater fall in sodium than a restricted fluid rate.24 A further PICU-based study randomised 122 children to receive isotonic or hypotonic fluids at full maintenance rate. The two groups were similar at study commencement; however, in the 46 patients followed up at 24 h, 13/23 of the hypotonic fluid group were hyponatraemic compared with 3/23 of the isotonic fluid group.25 A recent observational study of 81 paediatric intensive care patients administered hypotonic fluid (Na 40 mmol/l + K 20 mmol/l) for >12 h showed that 21.0% (17/81) became hyponatraemic (Na ≤135 mmol/l) at 12 h and 31.3% (15/48) were hyponatraemic at 24 h.26 The findings of our study combined with recent studies in hospitalised children suggest that hyponatraemia is less common if isotonic intravenous fluids are administered.
In summary, the administration of Hartmann's and 5% dextrose (at full maintenance rate) compared with 0.45% saline and 5% dextrose (at two-thirds maintenance rate) resulted in a smaller postoperative fall in plasma sodium at 16–18 h and a significantly smaller proportion of children becoming hyponatraemic. We suggest that Hartmann's and 5% dextrose should be administered at full maintenance rate to children who have undergone major surgical procedures.
Acknowledgments
The authors gratefully acknowledge the support of the Henry Blackwood Trust. The authors thank Cate Barnett and Dominic Crowley for the development of the clinical care pathways; Dr Lyndsay Cheater for motivation; Tanya Cienfuegos, Desley Horn, Tara Williams, Kylie Pearson and Sue Kendall for data collection; and Dr Helen Jeays for editorial advice.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Files in this Data Supplement:
- Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
- Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
- Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
- Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
- Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
- Web Only Data - This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
-
Competing interests None.
-
Ethics approval The study received approval from the Human Research Ethics Committee of the Royal Children's Hospital, Brisbane, and was registered with the Australian New Zealand Clinical Trials Registry (ACTRN12607000116426).
-
Provenance and peer review Not commissioned; externally peer reviewed.