But Is It Safe? Hydroxyethyl Starch in Perioperative Care

Hydroxyethyl starch (HES) has been used for >40 years as an intravascular volume expander.1 As a result, much of the pharmacology of HES has since been determined, including coagulopathic,2 anaphylactoid side effects,3 and the size-dependent retention of HES molecules in liver, skin, and nervous tissue.4 To address these issues, more modern HES formulations have progressively smaller molecular weights and molar substitutions (average number of hydroxyethyl residues per glucose subunit), a change known to facilitate faster clearance from the body.5 Because so much is already known about HES solutions, the possibility of increased nephrotoxicity with IV HES administration in acute care medicine has been surprising. Beginning with nephrosis-type lesions observed in transplanted kidneys from donors given HES,6 researchers have proceeded to identify associations between HES and renal failure in critically ill patients,7 and (most recently) a causal relationship in large scale trials of HES 130/0.4 versus crystalloids in septic patients.8,9 This line of inquiry has been sufficiently convincing that in 2013 both the European Medicines Agency and the United States Food and Drug Administration (FDA) issued warnings against HES use in septic patients in the intensive care unit (ICU).a,b The FDA further warned providers about the potential for excessive bleeding in cardiac surgical patients with the use of HES.a The possibility that a similar nephrotoxic effect may exist with perioperative HES use has stirred considerable controversy among anesthesiologists. In the September issue of this journal alone, 2 of The Open Mind articles10,11 and 1 Editorial12 extensively reviewed the complex mix of physiological data, clinical experience, and large outcome trials that currently inform the safety of perioperative HES use. Together, these well-written analyses are an excellent primer for anesthesiologists interested in the >40-year history of HES use in clinical medicine and 20-year history of renal toxicity. Not unsurprisingly, they also come to different conclusions with one concluding that a blanket ban on HES use is not justified,10 and another accepting the potential perioperative nephrotoxicity of HES but calling for more research to better define risk/benefit ratios.11 To a non-HES user, this controversy is fascinating. How can clinicians use a drug for almost half a century and not already have clearly identified a perioperative complication acute kidney injury (AKI) that can more than double postoperative mortality?13 With an ostensibly simple question (drug versus control), a defined period (perioperative), a clear end point (AKI), and >3000 citations in PubMed alone,c how can we know so much about a drug and yet not know enough to decide whether a patient undergoing a radical cystectomy should receive crystalloid or HES when acute bleeding causes hypotension? To the many thoughtful points raised by The Open Mind articles of Irwin et al. and Raghunathan et al., we add the ongoing and related controversy regarding perioperative hemodynamic optimization. Arguments that we give too much fluid perioperatively14 coexist uneasily with observations that preemptive perioperative hemodynamic optimization utilizing more advanced monitors leads to both more fluid administration and better outcomes.15 If perioperative clinicians cannot agree regarding the optimal fluid state for surgery, how can we distinguish between resuscitation and maintenance use of HES? Furthermore, in the absence of agreed-upon optimal end points for resuscitation, will any therapy (HES, crystalloid, or other colloids) consistently result in a positive outcome? What both Irwin et al. and Raghunathan et al. do agree on is more research to better assess the risk of perioperative HES use. In this month’s issue of Anesthesia & Analgesia, 2 studies shed more light on this question. Kancir et al.16 measured creatinine clearance, urinary NGAL (a protein released into the urine during renal injury), and urine output in 40 patients randomized to receive either HES 130/0.4 or 0.9 normal saline for perioperative fluid therapy. No difference in any measure of renal function was observed for 15 postoperative days. Kancir et al. did find indirect evidence of better volume expansion with HES (versus 0.9 normal saline), reflected in lower plasma concentrations of renin, angiotensin, and vasopressin. Patients receiving HES in this group also showed a trend toward higher blood loss. In the second paper,17 Hand et al. report the results of a hospital-mandated switch from albumin to 6% HES (130/0.4) for fluid therapy during orthotopic liver transplant. In 174 liver transplant recipients receiving albumin, HES, or both, Hand et al. observed an odds ratio of 1.18 for AKI as measured by Risk Injury Failure Loss End-Stage Kidney Disease (RIFLE) criteria for HES use (versus albumin). The authors also found a dose-dependent effect of HES on the odds of renal injury, with the adjusted odds ratio for renal failure in patients receiving HES only (2.94) nearly 3 times that for patients receiving albumin only. Although these 2 studies have apparently disparate findings, they differ in ways that highlight why the safety of HES is both controversial and difficult to establish. While Kancir et al. studied patients undergoing radical prostatectomy, Hand et al. focused on considerably sicker liver transplant patients. Hand et al.’s patients were more likely to be transfused (90% vs 2.5%), and at significantly higher risk for renal injury (22% of patients that did not receive HES sustained renal injury). Neither study distinguished between ongoing fluid maintenance (to replace routine fluid loss) and volume resuscitation to treat acute blood loss or other volume deficits. As some have argued,18 because HES was designed as a volume expander with greater efficacy than crystalloid, its use should be limited to acute hypovolemic resuscitation. By this logic, using HES for maintenance fluid therapy merely increases the toxicity of HES without any benefit. The small size of Kancir et al.’s study (N = 40) and complex, retrospective nature of Hand et al.’s study may also have limited their ability to find an effect of HES on renal failure or to accurately estimate the magnitude of any effect. After propensity adjustment, for example, the odds ratio reported by Hand et al. for HES use fell from 2.74 (using multivariate logistical regression) to 1.18. Although the patients in Kancir et al.’s study may not have been at sufficiently high risk for renal failure to show a difference between groups, Kancir et al. did find a (nonsignificant) trend toward higher blood loss in the HES group. This trend is consistent with the known coagulopathic effects of HES and the recent FDA warning in cardiac surgery patients. Hand et al. did not find that HES increased bleeding, but their study was not powered to detect differences in bleeding, and any coagulopathic effect may have been difficult to identify in a liver transplant population with a 90% transfusion rate at baseline. How should the clinician interpret these findings? One possibility is that patients at high risk for renal failure are more susceptible to the potential perioperative nephrotoxicity of HES. Such an interpretation would mirror nephrotoxic effects of HES in septic ICU patients. As noted by Raghunathan et al., the “Scandinavian Starch Severe Sepsis/Septic Shock” or 6S trial studied patients in severe sepsis as defined by a known focus of infection, at least 2 systemic inflammatory response syndrome criteria, and at least 1 organ with a Sequential Organ Failure Assessment (SOFA) score >2. Over a 90-day period, patients receiving HES had a relative risk of 1.17 for the primary outcome of death or dependence on dialysis. More patients in the HES group also suffered severe bleeding (RR = 1.52) and required renal replacement therapy (RRT) (RR = 1.35). However, these results differ from those observed in the CHEST study (Crystalloid versus Hydroxyethyl Starch Trial) published 6 months later. Entry criteria for this trial consisted of patients admitted to the ICU judged by the treating clinician to require fluid resuscitation with one accompanying physiological measurement. Patients in CHEST may thus have been less critically ill (mean Acute Physiology and Chronic Health Evaluation II [APACHE II] score in 6S = 17; mean SOFA score in CHEST = 7).8,9 Unlike 6S, the CHEST trial found less nephrotoxicity with HES (greater requirement for RRT in the HES group but no difference in mortality or renal failure). As with the 2 perioperative studies in this month’s issue of Anesthesia & Analgesia, a greater degree of physiological stress appeared to enhance the nephrotoxicity of HES in the ICU. Is HES safe for perioperative use? As no large randomized trials comparable to 6S or CHEST exist for perioperative use, a definitive answer is not possible, and existing studies regarding the perioperative nephrotoxicity of HES use are mixed.19–22 In fact, Hand et al. first conceived their study in response to Mukhtar et al.23 who found no adverse effect of HES in liver transplant. Clinicians might consider some general observations. First, the longevity of HES in clinical use suggests that for most patients, the magnitude of any nephrotoxic risk is likely to be small and a greater perioperative concern may be the risk of coagulopathy or anaphylactoid reaction. Second, any nephrotoxic effect is likely to be greater in patients at greater risk for perioperative renal injury. Identifying preoperative risk factors for nephrotoxicity might then assist in determining which patients, if any, might receive HES safely. Finally, existing data suggest that the magnitude of any increased overall risk attributable to HES alone (and not to factors related to HES administration such as the need for acute resuscitation) is small. Even in patients with severe sepsis8,9 or undergoing liver transplant,17,23 the attributable adjusted or multivariate odds ratio for renal injury or RRT is <1.4. Unanswered questions still remain regarding the perioperative use of HES. The deleterious effects of using HES for patients undergoing higher risk procedures (pancreatic cancer, multilevel spinal fusion,) as compared with lower risk procedures (hip arthroplasty, urologic procedures) remain unclear. In addition, few data address outcome end points beyond the immediate postoperative period when comparing HES 130/0.4 versus other “real-world” options such as crystalloids, blood, and albumin. Trials addressing these issues face considerable methodological challenges. They should distinguish between use of HES for acute volume expansion and HES use for maintenance fluid therapy. They should also address the potential confounding effect of hyperchloremic versus balanced carrier solutions. We look forward to such research in the future! DISCLOSURES Name: Steven Greenberg, MD. Contribution: This author helped design the study and write the manuscript. Attestation: Steven Greenberg approved the final manuscript. Conflicts of Interest: Steven Greenberg reported no conflicts of interest directly related to this manuscript. Steven Greenberg has served as consultant to CASMED in 2013. Name: Avery Tung, MD. Contribution: This author helped design the study and write the manuscript. Attestation: Avery Tung approved the final manuscript. Conflicts of Interest: This author declares no conflicts of interest. RECUSE NOTE Dr. Avery Tung is the Section Editor for Critical Care, Trauma, and Resuscitation for Anesthesia & Analgesia. This manuscript was handled by Dr. Steven L. Shafer, Editor-in-Chief, and Dr. Tung was not involved in any way with the editorial process or decision.