Fluid resuscitation is a mainstay of the medical management of haemorrhagic hypovolaemia. However, there is continuing uncertainty about the most appropriate fluid (Krausz 1995). Isotonic crystalloid solutions are often used to replace blood loss until a blood transfusion can be administered, but the wish to administer large volumes (advanced trauma life support [ATLS] guidelines suggest two litres of isotonic crystalloid), particularly in the pre-hospital phase when there may be problems with venous access, has stimulated the development of alternative approaches. One such approach is the use of hypertonic saline. Hypertonic solutions are considered to have a greater ability to expand blood volume and thus elevate blood pressure, and can be administered as a small volume infusion over a short time period (Krausz 1995). Infusion of hypertonic saline is believed to act by causing an osmotic shift of fluid from the intracellular and interstitial spaces to the extracellular compartment. The resulting auto-transfusion of fluid increases blood pressure and circulating volume. The use of hypertonic solutions has the potential to provide rapid volume resuscitation but with less interstitial oedema than with isotonic saline solutions (Shackford 1983).
It has also been suggested that hypertonic solutions may be the fluid of choice in hypovolaemic patients with head injuries (Khanna 2000; Peterson 2000; Walsh 1991). Cerebral perfusion pressure (CPP) depends on both intracranial pressure (ICP) and mean arterial blood pressure. (CPP = mean arterial blood pressure - mean ICP.) Patients in hypovolaemic shock who have head injuries may require rapid blood pressure elevation to maintain cerebral perfusion pressure, but excessive fluid and salt administration may result in brain swelling with an increase in intracranial pressure. Hypertonic solutions, however, are believed to reduce intracranial pressure by establishing an osmotic gradient across the blood brain barrier that draws water from the brain tissue into the vascular space (Fisher 1992). Hypertonic solutions, therefore, have the potential to restore blood pressure rapidly, but without increasing intracranial pressure. Hypertonic solutions are also thought to be beneficial in preventing the 'water logging' effect when there is interstitial lung injury, for example as occurs both in elective surgery and in trauma.
On the other hand, the use of hypertonic solutions for volume replacement may also have important disadvantages. In situations where haemorrhage is on-going, hypertonic solutions may result in continued bleeding from injured vessels. A potential problem in head injuries is that, in patients with a disrupted blood brain barrier, excess sodium may leak into brain tissue drawing water with it, thus worsening cerebral oedema. At present, there are no clinical ways to assess the integrity of the blood brain barrier. Furthermore, not only could the integrity of the blood brain barrier vary among patients with head injury, but it might also vary in different parts of the brain in a single patient. The possibility that hypertonic fluids may worsen outcome following head injury cannot therefore be dismissed (Krausz 1995, Shenkin 1976).
To determine whether hypertonic crystalloid decreases mortality in patients with hypovolaemia with and without head injuries, we conducted a systematic review of randomised controlled trials.
Criteria for considering studies for this review
Types of studies
Randomised controlled trials. Crossover trials were excluded.
Types of participants
Patients with trauma, burns or surgery. Trials in both the pre-hospital and hospital setting were included.
Types of interventions
Trials that compare hypertonic to isotonic and near isotonic crystalloid. Trials of hypertonic crystalloid with an add-on colloid (e.g. hypertonic saline and dextran) are not included. This comparison has been dealt with in a previous systematic review by the Cochrane Injuries Group (Perel 2007).
Types of outcome measures
The principal outcome measure is mortality from all causes and disability assessed at the end of the follow-up period scheduled for each trial. Disability was assessed using the Glasgow outcome scale (Jennett 1975) which includes the following categories: death, persistent vegetative state, severely disabled, moderately disabled and good recovery. For the purpose of this review, the scale was dichotomised with death, persistent vegetative state and severe disability denoting a poor outcome, and moderate disability and good recovery denoting a good outcome.
Intermediate physiological outcomes were not used for several reasons. Such outcomes are subject to intra and inter-observer variation, they have no face value to patients and relatives, and those seen as appropriate are not stable over time. Also, there would need to exist a strong predictive relationship between the variable and mortality.
Search methods for identification of studies
The search was last updated in October 2007.
We searched the following electronic databases;
- Cochrane Injuries Group's specialised register
- National Research Register
The original search strategies were based on the terms listed below. The full search strategies for the most recent search update are listed in Appendix 1.
- Saline solutions hypertonic (MeSH)
- Isotonic solutions (MeSH)
- Hypertonic or isotonic or hyperosmotic or hyperoncotic
- Hypotensive resuscitation
- #1 or #2 or #3 or #4
- double-blind-procedure (MeSH)
- #6 or #7
- #5 and #8
Data collection and analysis
Selection of studies
One reviewer (FB) examined the electronic search results for reports of possibly relevant trials and these reports were retrieved in full. Two reviewers applied the selection criteria independently to the trial reports, resolving disagreements by discussion.
Data extraction and management
Two reviewers independently extracted information on the following: study quality, number of randomised patients, type of participants and the interventions. The outcome data sought were number of deaths and disability. The reviewers were not blinded to the authors or journal when doing this, as evidence for the value of this is far from conclusive (Berlin 1997). Results were compared and any differences resolved by discussion.
For each trial the relative risk of death and 95% confidence interval were calculated. The relative risk was chosen as it is more readily applied to the clinical situation.
The groups of trials were examined for statistical evidence of heterogeneity using a chi squared test. As there was no obvious heterogeneity on visual inspection or statistical testing, pooled relative risks (RR) and 95% confidence intervals (CIs) were calculated using a fixed effects model.
Assessment of risk of bias in included studies
Since there is evidence that the quality of allocation concealment particularly affects the results of studies (Schulz 1995), two reviewers scored this quality on the scale used by Schulz (Schulz 1995) as shown below, assigning C to poorest quality and A to best quality:
A = trials deemed to have taken adequate measures to conceal allocation (i.e. central randomisation; numbered or coded bottles or containers; drugs prepared by the pharmacy; serially numbered, opaque, sealed envelopes; or other description that contained elements convincing of concealment).
B = trials in which the authors either did not report an allocation concealment approach at all or reported an approach that did not fall into one of the other categories.
C = trials in which concealment was inadequate (such as alternation or reference to case record numbers or to dates of birth)
In addition, data was extracted on blinding and loss to follow-up. Where the method used to conceal allocation was not clearly reported, the author was contacted, if possible, for clarification. We then compared the scores allocated and resolved differences by discussion.
Description of studies
Eighteen randomised controlled trials were identified by the searches. However, four (Gunn 1989; McGough 1990; Younes 1988a; Younes 1988b) did not provide data on the outcomes specified in the review. Details of these studies are also reported in the table of included studies for completeness.
In the 14 trials reported in the meta-analysis, patients with burns were included in three (n=72) (Bortolani 1996; Caldwell 1979; Jelenko 1978), patients undergoing surgery in five (n= 230) (Croft 1992; Cross 1989; Jarvela 2002; Shackford 1983; Shackford 1987) and trauma patients in six (n= 654) (Cooper 2004; Simma 1998; Vassar 1990; Vassar 1993a; Vassar 1993b; Younes 1992).
Eleven trials compared hypertonic saline versus Ringer's lactate (Bortolani 1996; Caldwell 1979; Cooper 2004; Croft 1992; Jelenko 1978; Shackford 1983; Shackford 1987; Simma 1998; Vassar 1990; Vassar 1993a; Vassar 1993b), and the rest compared hypertonic saline with normal saline.
For more detailed descriptions of individual studies, please see the table of included studies. No additional studies were identified for this latest update.
Risk of bias in included studies
Allocation concealment was judged to be adequate in five trials (Cooper 2004; Simma 1998; Vassar 1990; Vassar 1993a; Vassar 1993b), inadequate in three (Caldwell 1979; Shackford 1983; Shackford 1987), and unclear in the rest. Five trials reported the use of identical bags or containers for the fluids, so that staff were blinded to the identify of the solutions (Cooper 2004; Cross 1989; Vassar 1990; Vassar 1993a; Vassar 1993b).
Effects of interventions
Death was reported either in the paper, or the information was obtained by contacting the researcher, in 14 studies. Data on death were not obtained for four trials (Gunn 1989; McGough 1990; Younes 1988a; Younes 1988b). Data on disability was obtained from the author of one trial (Cooper 2004).
Due to the clinical heterogeneity of the different patient groups it was felt to be inappropriate to pool them; therefore, only the results for the subgroups are given. The pooled relative risk for death in trauma patients was 0.84 (95% CI 0.69 to 1.04), for patients with burns 1.49 (95% CI 0.56 to 3.95) and for patients undergoing surgery 0.51 (95% CI 0.09 to 2.73). Only one trial gave data on disability (Cooper 2004) and the relative risk for a poor outcome was 1.00 (95% CI 0.82 to 1.22).
This review does not give us enough data to be able to say whether hypertonic crystalloid is better than isotonic crystalloid for the resuscitation of patients with trauma or burns, or those undergoing surgery. However, the confidence intervals are wide and do not exclude clinically significant differences between hypertonic and isotonic crystalloid. A previous review (Perel 2007) found there was a trend towards a favourable effect on mortality for colloids in hypertonic crystalloid compared to isotonic crystalloids. However, those results are compatible with the play of chance.
We chose not to pool the results of the burns, surgery and trauma patients, as we felt these groups were too clinically heterogeneous. Bleeding and fluid management in patients undergoing elective surgery would tend to be more controlled and, therefore, different to that in trauma patients.
Most of the trials are small and quality was judged to be adequate in only five of them. There was variation in the type of participants, and length of follow-up, and little standardisation in terms of fluid regimes. Also some of the trials were old. Although older trials will not necessarily be of poorer quality, it may be that treatment protocols have subsequently altered, making these trials less relevant to current clinical practice. Indeed in the 1970s and 1980s there were few protocols on fluid resuscitation in the critically ill. Since the late 1980s, there have been more clear guidelines and standardisation of fluid resuscitation regimes, although many areas of contention still exist.
Mortality was selected as the main outcome measure in this systematic review for several reasons. In the context of critical illness, death or survival is a clinically relevant outcome that is of immediate importance to patients, and data on death are reported in many of the studies. Furthermore, one might expect that mortality data would be less prone to measurement error or biased reporting than would data on pathophysiological outcomes. The use of a pathophysiological end-point as a surrogate for an adverse outcome assumes a direct relationship between the two, an assumption that may sometimes be inappropriate. Finally, when trials collect data on a number of physiological end-points, there is the potential for bias, due to the selective publication of end-points showing striking treatment effects.
Hypertonic solutions have been proposed as the fluid of choice in patients with head injuries (Walsh 1991), as they may maintain cerebral perfusion pressure without causing brain swelling with an increase in intracranial pressure. However, we found only one small trial (Simma 1998) among people with head injuries.
Implications for practice
This review does not provide any evidence that hypertonic crystalloid is better than isotonic crystalloid, but it does not rule out clinically important differences.
Implications for research
Further trials are needed comparing hypertonic to isotonic crystalloid. These trials should be multi-centre prospective randomised controlled trials, that are large enough to detect a clinically important difference. Clinically relevant outcomes such as mortality should be used and trials should specify the type and amount of fluid used.
Thanks to Reinhard Wentz for help with the searches and to Phil Alderson for overseeing the editorial process. Thanks also to E Akpa for assistance during the review process.
Data and analyses
- Top of page
- Authors' conclusions
- Data and analyses
- What's new
- Contributions of authors
- Declarations of interest
- Sources of support
- Index terms
Appendix 1. Search strategy
Lactated Ringer's not isotonic
Lactated Ringer's solution is not a truly isotonic fluid. In one report (Tommasino C, Moore S, Todd MM. Crit Care Med 1988;16 p867) the measured osmolality was stated to be approximately 254 mosm/l while the calculated osmolality was 273 mosm/l.
The treatment of traumatized patients will include the infusion of many literes of crystalloids during the first hours. In comparison, the 250 ml of lactated Ringer's or saline used at intervention in the studies concerned probably does not matter very much. The problem addressed by the review, rather than one of "isotonic versus hypotonic", may be more precisely formulated as something like "early supplementation or not" of hypertonic fluid to the continued use of many liters of a weakly hypotonic fluid.
The hypothesis that 250 ml of hypertonic fluid is beneficial, may easily lead to the idea that many litres of a hypotonic fluid is detrimental. Or is the hypertonic fluid of benefit only when it is added to adjust for the hypotonic one? Will a test with really isotonic crystalloid do better than the hypotonic one and show the supplementation with hypertonic fluid not only to be unnecessary, but even harmful?
The reports often conceal the true nature of the fluids used behind designations like "conventional isotonic solutions" or "standard of care", and the amount of fluid given after arrival in hospital may not be stated.
Implications for research is that studies with the continued use of truly isotonic solutions have to be done to decide whether hypertonic or weakly hypotonic solutions are beneficial or detrimental. The nature and amount of fluids used in future studies should be clearly stated.
I certify that I have no affiliations with or involvement in any organisation or entity with a direct financial interest in the subject matter of my criticisms.
We agree that lactated Ringer's is not a truly isotonic fluid and have, therefore, changed the title of the review to reflect this. The title is now 'Hypertonic versus near isotonic crystalloid for fluid resuscitation in critically ill patients'.
We also agree that the nature and amount of fluids used in future studies should be clearly stated, and have included a statement to this effect in the conclusions.
Comment by Per Størset (anesthesiologist), December 2002.
Reply from Frances Bunn, May 2004.
Last assessed as up-to-date: 14 October 2007.
Protocol first published: Issue 2, 2000
Review first published: Issue 4, 2000
Contributions of authors
FB screened citations for eligibility, obtained references, contacted authors, extracted data, entered data and wrote up the review. IR helped to write the review. RT commented on the protocol and review. DT screened citations for eligibility.
Declarations of interest
Sources of support
- University of Hertfordshire, UK.
- NHS Research and Development, UK.
Medical Subject Headings (MeSH)
*Plasma Substitutes; Critical Illness; Hypertonic Solutions [*therapeutic use]; Hypovolemia [mortality; *therapy]; Isotonic Solutions [*therapeutic use]; Randomized Controlled Trials as Topic; Rehydration Solutions [*therapeutic use]
MeSH check words
* Indicates the major publication for the study