Best Evidence Topics
  • Send this BET as an Email
  • Make a Comment on this BET

Do hydroxyethyl starch colloids increase the incidence of renal failure in patients with sepsis?

Three Part Question

In [septic patients], do [hydroxyethyl starch colloids] increase the incidence of [renal failure]?

Clinical Scenario

Both crystalloids and colloids are commonly used in both the emergency department and ICU in the fluid resuscitation of patients with sepsis. The use of hydroxyethyl starches is controversial, and improved hemodynamic parameters compared to crystalloids and other colloids must be balanced against growing evidence of nephrotoxicity. This appraisal suggests that hydroxyethyl starch, when used in the management of patients with sepsis, increases the risk of acute renal failure.

Search Strategy

A Medline search was conducted using the exploded MESH terms “sepsis”, “renal insufficiency” and “hetastarch”. A similar EMBASE search was conducted using the exploded terms “sepsis”, “kidney failure”, and “hetastarch”. Both searches were limited to English language articles involving human subjects.
Forty-three citations were found, which included 39 unique articles. Of these, 2 textbooks were excluded, and 14 articles not relevant to the primary research question were excluded. Eight letters and 9 editorials were also excluded, as well as one article that could not be obtained through the university library system.
Three potentially relevant randomized controlled trials were found using the search criteria. Of these three, one RCT was excluded, as it studied a heterogenous patient population in which septic patients were a minority. References were searched in one systematic review of the use of HES for fluid management in sepsis, although this review was not specific to acute renal failure as an endpoint. One additional relevant RCT was found by the review of references in the systematic review. In total, three RCTs were selected for further analysis.

A search of the Cochrane database of systematic reviews was conducted, and a review of crystalloids versus colloids in the resuscitation of critically ill patients was found. This was not limited to sepsis patients, and acute renal failure was not a specific outcome measure. In the review, patients receiving hydroxyethyl starches were analysed separately.

Relevant Paper(s)

Author, date and country Patient group Study type (level of evidence) Outcomes Key results Study Weaknesses
Boldt., J., Muller, M., Mentges, D., Papsdorf, M., Hempelmann, G.
150 patients with sepsis secondary to major surgery. Patients with renal failure or shock at the start of the study were excluded.Patients were randomized to receive 5 days of volume therapy with either HES or albumin to maintain a PCWP of 12-15. The physicians involved in managing volume therapy were blinded to the infusion fluid. Primary outcomes were hemodynamics and organ morbidity. The two groups were similar in age, septic source, and incidence of severe sepsis. Acute renal failureNo differenceShort duration of follow-up.
Schortgen, F., Lacherade, J.C., Bruneel, F., Cattaneo, Hemery, F., Lemaire, F., Brochard, L
129 patients with severe sepsis in three ICUs and three medical/surgical units.Patients were centrally block randomized to either of the study fluids, and physicians opened an opaque envelope indicating treatment assignment immediately after a patient met study criteria for severe sepsis or septic shock. There was no limit on the allowed cumulative volume of gelatin, while the cumulative allowed infusion of HES was limited to 80mL/kg. Any further fluid infusions were given as crystalloid. Dose differences between the study fluids precluded the blinding of physicians. Baseline creatinine was higher in the HES group than in the gelatin group, and more patients were admitted for a surgical reason in the HES group than in the gelatin group. The groups were similar in age and severity of illness scores. Acute renal failure (ARF) after 30 days of follow-upIncreased ARF in HES group (OR 2.32)
Brunkhorst, C.E., Bloos, F., Meier-Hellmann-Meier, A., et al.
537 patients with severe sepsis in academic ICUs.Compared Pentastarch to Ringer’s lactate, and intensive insulin therapy to standard therapy. The study was a randomized two-by-two factorial design. All were treated for three weeks or until discharge. Physicians were not blinded to the assigned study fluid. 90 day mortality was one of several secondary endpoints. Groups were similar at study outset with regards to age, APACHE scores, comorbidities, and site of sepsis. Organ morbidity, 28 day and 90 day mortalityincrease in 90 day mortality (p=0.09), renal failure (p=0.02), and need for renal replacement therapy (p=0.001) in the Pentastarch group. This translated to an absolute mortality of 33.9% in the RL group versus 41.0% in the Pentastarch group.


Major treatment goals in sepsis include the restoration of adequate intravascular volume, blood pressure, and end-organ perfusion. Surrogates such as central venous pressure and pulmonary capillary wedge pressure are frequently used in goal –directed therapy in the early management of such patients. The choice of fluid for resuscitation is controversial, and may include a crystalloid or a colloid. The potential drawbacks of over-resuscitation with crystalloid solutions include peripheral and pulmonary edema, since two-thirds of sodium-containing resuscitation fluids should redistribute from the intravascular space to the interstitial space. Normal saline has been criticized further as physiologically unsound, in that the low pH of 5.3 could potentially cause a nonanion gap hyperchloremic metabolic acidosis. Ringer’s lactate, in contrast, can potentially cause a metabolic alkalosis due to the metabolism of lactate to bicarbonate. The theoretical benefits of colloid fluid resuscitation include rapid intravascular volume expansion with decreased pulmonary and peripheral edema. Hydroxyethyl starches are artificial colloids with a high molecular weight that results in an improved ability to remain intravascular when compared with gelatins and albumin. Biochemically, they are glucose polymers derived from amylopectin corn starch. In this critical appraisal, a focused Medline and Embase search was performed for evidence of a nephrotoxic effect of hydroxyethyl starch colloids in the resuscitation of patients with sepsis. A total of three randomized controlled trials were found, two of which suggested a significantly increased risk of acute renal failure and renal replacement therapy in patients treated with one hydroxyethyl starch. Furthermore, the largest study also demonstrated a trend toward increased mortality in the group treated. Comparison of the three randomized controlled trials is complicated by the fact that a different comparator fluid was used in each of the trials. Furthermore, it is worth noting the lack of homogeneity in any group of patients with sepsis, due to the multitude of potential etiologies. In particular, the results of the Schortgen study are tempered by baseline differences in creatinine and surgical sources between the treatement and comparator groups. Several mechanisms have been suggested to explain the observed nephrotoxic effects of HES solutions. Increased colloid oncotic pressure may further decrease the glomerular filtration rate in patients who are already fluid-depleted. Uptake of HES by the reticular endothelial system, including macrophages in the kidney, may lead to toxicity6. Finally, hyperosmolar chloride has been known to reduce renal flow, urine output, and sodium retention, and it has been proposed that the normal saline used as a carrier solution for HES may play a role in its nephrotoxicity.

Clinical Bottom Line

While the hard evidence for a benefit to the use of hydroxyethyl starches is lacking, there is growing evidence of a dose-dependent nephrotoxic effect in the context of volume depletion. Until the safety of such agents is established for this indication, it seems prudent to avoid the use of HES in the early fluid resuscitation of patients with sepsis.


  1. Boldt., J., Muller, M., Mentges, D., Papsdorf, M., Hempelmann, G. Volume therapy in the critically ill: is there a difference? Intensive Care Medicine 1998; 24: 28-36
  2. Schortgen, F., Lacherade, J.C., Bruneel, F., Cattaneo, Hemery, F., Lemaire, F., Brochard, L. Effects of hydroxyethylstarch and gelatin on renal function in severe sepsis: a multicentre randomised study. Lancet 2001; 357: 911-16
  3. Brunkhorst, C.E., Bloos, F., Meier-Hellmann-Meier, A., et al. Intensive insulin therapy and pentastarch resuscitation in severe sepsis New England Journal of Medicine 2008. 358 (2): 125-139