Pendulum Swings Again
Authors: M. C. Kelleher, D. J. Buggy
Br J Anaesth. 2014;113(3):335-337.
Since the 1980s, evidence has been accumulating that survivors of major surgery generate a higher postoperative cardiac output than non-survivors, and that achieving a supra-normal oxygen delivery to the body’s major organs (DO2, normally greater than 1000 ml O2min) in the perioperative period, using supplementary i.v. fluids combined with inotropes, could improve survival in high-risk patients.
A recent Cochrane Systematic Review identified 31 studies (including 5292 participants) that studied the perioperative increase in global blood flow to explicit defined goals. Of these, 28 were single-centre and included less than 200 participants. One large multi-centre study dominated the results of the analysis. While no difference in mortality was found, colloid-based, individualized goal-directed therapy (GDT) was associated with a reduced rate of major complications. No difference was observed between interventions using fluids alone compared with those using a combination of fluids and inotropes to achieve oesophageal Doppler-guided blood flow targets. Similar conclusions have been made with regard to the use of GDT in cardiac surgery. A randomized controlled trial of GDT in patients with proximal femoral fractures was concluded early because of slow recruitment, but demonstrated a clinically important trend towards favourable results with colloid-based GDT.
However, colloid-based GDT apparently does not benefit ‘fit’ patients [who achieved anaerobic threshold (AT) greater than 11 ml O2 kg−1 min−1] compared with ‘unfit’ patients (AT 8.0–10.9 ml O2 kg−1 min−1) in a randomized controlled trial of 179 patients undergoing major colorectal surgery. In this population, ‘fit’ GDT patients had a longer length of hospital stay compared with controls, whereas there was no difference for ‘unfit’ patients. GDT patients received an average of 1360 ml additional colloid compared with controls, indicating a detrimental effect of this additional colloid load despite GDT, although DO2 was not directly measured.
When considering the optimum fluid choice to achieve supra-normal DO2, colloids have a theoretical and intuitive advantage. Being composed of molecules greater than 35 kDa, colloids have traditionally been thought to have greater fluid efficacy (i.e. they remain in the intravascular space for longer, thus contributing a greater effect on cardiac output) compared with crystalloids. These theoretical benefits of colloids led to their routine use in GDT.
Studies in healthy volunteers have demonstrated that crystalloids tend to expand the intravascular volume by ~20% of the infused volume; however, when physiological endpoints have been used in clinical studies, the fluid efficacy of crystalloids increases up to 60%. This discrepancy may be explained by a number of factors, including the time taken for fluid to equilibrate throughout the extracellular fluid, the vasodilating effect of anaesthetic-induced hypotension on vascular capacitance, and the impairment of fluid elimination because of the surgical stress response.
While colloid has been the default fluid choice in most GDT studies to date, recent evidence from both the perioperative anaesthetic setting and from intensive care have raised concerns that these agents may produce adverse effects on renal function and coagulation. Recent studies have sought to compare the use of crystalloid solutions with colloids for GDT.
Senagore and colleagues randomized 64 patients undergoing laparoscopic colectomy to receive standard therapy or oesophageal Doppler-guided GDT using either crystalloid (lactated Ringer’s) or colloid (hetastarch). While patients randomized to crystalloid GDT received slightly more intraoperative fluids (mean 3800 ml vs 3300 ml colloid), their length of hospital stay was similar to patients in the hetastarch group (72 h vs 76 h).
In a German pilot study, patients (n=50) undergoing cytoreductive surgery for ovarian cancer were randomized to receive either balanced crystalloid or balanced hydroxyethyl starch (HES) colloid. Oesophageal Doppler monitoring was used to guide fluid administration to optimize stroke volume. While patients in the colloid group achieved marginally better numerical haemodynamic stability, there was no difference in complication rates or length of intensive care unit or hospital stay.
A recent study from York compared the use of Hartmann’s crystalloid solution with a balanced starch solution in 202 medium- to high-risk patients undergoing elective colorectal surgery, and using a LiDCO Rapid monitor (LiDCO, Cambridge, UK) to guide GDT. The primary outcome studied was the ability to tolerate enteral diet on the fifth postoperative day. No difference was seen between the crystalloid and colloid group in the primary outcome. While the crystalloid group received more fluid, again there was no difference in the incidence of complications between the groups, thus supporting the use of crystalloids as a safe and significantly less expensive alternative to colloids in the perioperative setting.
The incidence of postoperative death and acute kidney injury (AKI) associated with the use of 6% HES in the perioperative period has been the subject of a recent systematic review and meta-analysis. No differences were identified in the incidence of death or AKI in surgical patients receiving 6% HES, leading the authors of the meta-analysis to conclude that the use of 6% HES should not be recommended in surgical patients because of the absence of clinical evidence of benefit and higher cost.
In the critical care setting, the use of colloids as resuscitation fluid has been the subject of a number of important recent publications, which have led to the withdrawal of HES by regulatory authorities in Europe and the USA. The efficacy of volume substitution and insulin therapy in severe sepsis (VISEP) study, the Scandinavian starch for severe sepsis/septic shock (6S) trial, and the crystalloid versus hydroxyethyl starch (CHEST) trial compared the use of starches for resuscitation with modified Ringer’s lactate, Ringer’s acetate, and sodium chloride 0.9%, respectively. All three trials showed not only no survival benefit with the use of starch but also increased risk of AKI.
These studies have been included in an updated systematic review and meta-analysis on the use of colloidvs crystalloids for fluid resuscitation in critically ill patients, which has again concluded that the use of expensive synthetic colloids is difficult to justify given the failure to identify a mortality benefit associated with their use.
What about the use of non-starch colloids in the critically ill? The CRISTAL study was a pragmatic, randomized, open-label multicentre clinical trial comparing the use of colloids vs crystalloids in critically ill patients presenting with hypovolaemic shock. Some 2857 clinically shocked patients were randomized to receive either colloids or crystalloids for initial resuscitation, with the choice of particular fluid used determined by the individual investigator, depending on local availability. Colloids used included HES, gelatins, and albumin, while Ringer’s solution and isotonic saline were the principal crystalloids used. No difference in mortality was observed at 28 days, but 90-day mortality was lower in the colloid group. It is possible that the different fluids used for resuscitation in this study masked a detrimental effect of colloids as shown elsewhere. It is also plausible that older gelatin colloids are not as harmful, but if the endothelial glycocalyx is impaired (as it is in severe sepsis), then intravascular retention of any colloid may be no better than that of crystalloid.
Gelatins have not been shown to be superior to crystalloid in clinical outcome studies, and have also been previously associated with AKI, coagulopathy, a significant incidence of anaphylaxis, and high cost. Albumin has been studied in a large, multi-centre, randomized trial in the critical care setting, and was found to result in similar outcomes at 28 days compared with 0.9% saline. The cost of albumin means that its use may not be justified based on the available evidence. A subgroup analysis found that patients with a traumatic brain injury had an increased mortality when resuscitated with albumin.
Not all crystalloids are equal either. The role of chloride in influencing the adverse effects of colloids on renal function has been recently questioned, and a prospective, open-label, sequential period pilot study was undertaken at an Australian intensive care unit. After a control period, during which standard fluid therapy was administered, and a 6-month phase-out period, the use of chloride-rich i.v. fluids (saline 0.9%, succinylated gelatin 4%, or albumin 4%) was restricted to specialist approval only. Balanced crystalloid and colloid solutions were allowed during the intervention period. The implementation of this strategy was associated with a significant decrease in the incidence of AKI and use of renal replacement therapy. It remains to be seen whether this effect can be reproduced in large-scale trials of the use of balanced solutions.
In summary, while intuitively colloid fluids ought to have greater fluid efficacy in expanding fluid volume and enhancing oxygen delivery during GDT regimens, recent evidence has shown evidence of detrimental outcomes particularly on renal function using starch colloids. Guidelines on i.v. fluid therapy published by the National Institute for Health and Care Excellence (NICE) recommend the use of crystalloid solutions containing a sodium concentration in the range of 130–154 mmol litre−1 for i.v. fluid resuscitation, and recommend against the use of tetrastarch for this purpose. Human albumin solution 4–5% should be considered only in patients with severe sepsis.
Older gelatin colloids have not been subject to the same level of scrutiny as other colloids, but available evidence suggests that they too are causally linked to worse outcomes in cardiac surgery and in the intensive care setting. Further research comparing gelatin colloid with balanced salt crystalloid both in the intensive care and perioperative GDT context is warranted.
When using crystalloid fluids, chloride-rich saline 0.9% should be avoided unless there are specific circumstances, but emerging evidence from both the intensive care unit and the perioperative GDT setting suggests balanced salt solutions (e.g. Hartmann’s solution) provide at least equivalent outcomes compared with synthetic colloids, without adverse effects and at considerably less cost. It is clear that the current balance of evidence from quality, randomized trials indicates that the pendulum of fluid therapy has swung decisively away from colloid towards balanced salt solutions of crystalloid.