Over the last 12 months, a number of high-quality randomized controlled trials (RCTs) addressing the safety and efficacy of currently used hydroxyethyl starch (HES) fluids—6 % HES (130/0.4–0.42) or tetrastarches—for resuscitation of critically ill patients have been published [1–3]. These RCTs include the Crystalloid vs. Hydroxyethyl Starch Trial (CHEST) [1] and the Scandinavian Starch for Severe Sepsis/Septic Shock study (6S) [2] that demonstrated that resuscitation with 6 % HES preparations was associated with increased 90-day mortality and use of renal replacement therapy (RRT) when compared with crystalloids.

These patient-centred outcomes are highly relevant to clinicians given the uncertainty about the safety of older, high-molecular-weight, highly substituted HES preparations that were associated with increased mortality and acute kidney injury (AKI), particularly in patients with sepsis. In addition, the retraction of a number of publications related to the safety of tetrastarch preparations raised further concerns amongst clinicians.

On the basis of the results of these RCTs, the Surviving Sepsis Campaign guidelines recommended against use of HES for fluid resuscitation of severe sepsis and septic shock for the first time in the latest iteration of guidelines.

A number of systematic reviews have been conducted following the publication of RCTs in 2012. In this issue of Intensive Care Medicine, Patel and colleagues [4] report the effects of resuscitation with tetrastarch preparations (both potato or waxy maize derivatives) compared with non-HES fluids in patients with severe sepsis. The primary outcome was 90-day mortality, and requirement for RRT was a tertiary outcome. From 6 studies that included 3,033 patients, tetrastarches were associated with a significant, 13 % relative increase in the risk of death at 90 days and 41 % relative increase in requirement for RRT compared with crystalloids. Three trials published in 2012 were adjudicated to be at low risk of bias and contributed 96 % of the patients [1–3]. A sensitivity analysis limited to the results of these trials produced similar results.

These results are consistent with recent systematic reviews published in 2013 comparing the effects of HES, both low- and high-molecular-weight preparations [5], on patient-centred outcomes versus a range of resuscitation fluids in acutely ill patients [6] and patients with sepsis [7] (Table 1).

Table 1 Summary table of systematic reviews and meta-analyses published in 2013 comparing hydroxyethyl starch preparations (HES) versus other resuscitation fluids in randomized controlled trials that reported mortality and use of renal replacement therapy (RRT) as outcome measures
Full size table

From these reviews, it appears that resuscitation with HES is associated with an increased relative risk of death and use of RRT regardless of the type of HES or the patient population.

Before the publication of the above papers, HES was the most commonly prescribed colloid globally, particularly in Europe [8]. It would be expected that the use of HES will decrease, not least in patients with sepsis or at risk of AKI, to accord with the publication of recent high-quality evidence.

Does this new evidence about the use of HES apply to other colloids? Whilst recent trials have demonstrated short-term haemodynamic differences with colloids, the relative volume-sparing effect compared with crystalloids is low [1–3, 9].

Specific patient populations are at increased risk of death when resuscitated with colloids—namely patients with traumatic brain injury and albumin [10]. Whether albumin confers a benefit in sepsis, as suggested in a sub-study of the Saline vs. Albumin Evaluation (SAFE) study, remains undetermined [11].

There is no substantive evidence to support the use of other semi-synthetic colloids, such as gelatin, as alternatives to albumin or HES. Recent observational data suggest that gelatin may be associated with development of nephrotoxicity similar to that observed with HES [12].

Whether the impact of these publications will change practice in critically ill patients remains to be seen. Further cross-sectional translational studies, such as that done following the publication of the SAFE study are required [8].

Tetrastarches continue to be widely used in surgical patients undergoing general anaesthesia. In the light of data in critically ill patients, there is an imperative to evaluate their safety and efficacy in large RCTs powered for patient-centred outcomes relevant to perioperative patients.

So, what does this mean for clinicians in 2013? That increased caution about use of colloids, particularly with HES, in critically ill patients is evident. However, crystalloids as alternative resuscitation fluids do not automatically confer increased safety to our patients.

Resuscitation with normal saline is associated with chloride excess that has been associated with adverse metabolic effects and potential nephrotoxicity [13].

To date, no high-quality RCT has been conducted comparing the effects of saline, the most commonly used crystalloid globally, versus “balanced” salt solutions. This is a further fundamental research question that needs to be addressed.

As the optimal type of resuscitation fluid remains uncertain, more attention needs to be given to the dose and volumes administered.

Fluids accumulate over time. This is associated with development of interstitial oedema and applies to both colloids and crystalloids. There is an increasing body of observational evidence suggesting that excess use of intravenous fluids is associated with adverse outcomes [14]. Accumulation of HES within the reticulo-endothelial system is implicated in adverse events and toxicity.

The role of bolus fluid resuscitation has been questioned following the Fluid Expansion as Supportive Therapy (FEAST) study that demonstrated increased mortality associated with bolus resuscitation with albumin and saline [15]. Whilst the generalizability of these results remains unclear, the principle of questioning the efficacy of bolus resuscitation, in terms of both volume and timing, needs to be carefully considered in all critically ill patients.

Clearly, the time has come to regard resuscitation fluid with the same degree of critical thought as that when administering a potentially toxic drug. The prescription of fluid needs to consider the specific indication: shock versus physiological measurements such as low urine output or central venous pressure; severity, type and source of insult: sepsis versus trauma; the time context in which fluid is administered: fluid resuscitation requirements change markedly over time; the constituents of the fluid: potential impact on metabolic and organ function; the potential for toxicity related to fluid excess and organ dysfunction that develops in the post-resuscitation period.

These considerations will also need to be considered by regulators when new formulations of semi-synthetic or non-physiological fluids are developed, so that efficacy and safety are determined in relevant pre-clinical models followed by high-quality investigator-initiated trials before registration and introduction into the market.

It is likely that this fundamental area of intensive care medicine will remain open to further intense debate and research that will ultimately change clinical practice in the future.