A Critical Reassessment of the Kidney Risk Caused by Tetrastarch Products in the Perioperative and Intensive Care Environments

Purpose: To reassess the results of former meta-analyses focusing on the relationship between novel HES preparations (130/0.4 and 130/0.42) and acute kidney injury. Previous meta-analyses are based on studies referring to partially or fully unpublished data or data from abstracts only. Methods: The studies included in the former meta-analyses were scrutinized by the authors independently. We completed a critical analysis of the literature, including the strengths, weaknesses and modifiers of the studies when assessing products, formulations and outcomes. Results: Both the published large studies and meta-analyses show significant bias in the context of the deleterious effect of 6% 130/0.4–0.42 HES. Without (1) detailed hemodynamic data, (2) the exclusion of other nephrotoxic events and (3) a properly performed evaluation of the dose–effect relationship, the AKI-inducing property of 6% HES 130/0.4 or 0.42 should not be considered as evidence. The administration of HES is safe and effective if the recommended dose is respected. Conclusions: Our review suggests that there is questionable evidence for the deteriorating renal effect of these products. Further well-designed, randomized and controlled trials are needed. Additionally, conclusions formulated for resource-rich environments should not be extended to more resource-scarce environments without proper qualifiers provided.


Introduction
In clinical practice, one of the most common interventions is volume expansion in those with perceived hypovolemia.Intravenous fluid administration is easily performable with crystalloid and colloid infusions or with various blood products.In the current era, the isotonic but non-physiologic 0.9% saline and balanced solutions are available as crystalloid infusions, whereas the 6% hydroxyethyl starch (HES) (130/0.4 or 0.42) and the 5% or 20% human albumin are available as colloids, respectively.
Formerly, dextrans, gelatin and early generations of HES were also available as well.Dextran products are high  molecular weight polymers of glucose produced by bacteria (Leuconostoc mesenteroides) in sucrose-rich environments [1].Their volume expansive effect is quite significant.Unfortunately, the risk of life-threatening allergic reactions to dextran products is prohibitively high.While these reactions are preventable by the administration of hapten (1 kDa dextran) a few minutes before the infusion, this property of dextran makes it unsuitable for use in acute situations.An alternative, gelatin infusion was manufactured by partial hydrolysis and chemical modifications after extraction from animal (pig, calf, fish) bones, skin and tendon (molecular weight: 30-35 kDa, concentration: 3-5%) [2].Their volume-expanding effect is limited and their administration carries the risk of prion-mediated disease transmission.HES preparations are plant-derived products featured at various concentrations (6%, 10%), molecular weights (450 kDa, 200 kDa, 130 kDa) and molar substitutions (0.7, 0.6, 0.5, 0.42, 0.4) [3,4].This latter property needs some explanation for further interpretation.A molar substitution of 0.7 means that on average there are 7 hydroxyethyl groups for 10 glucose molecules.The evolution of HES generations is as follows: (1) hetastarch-6% HES 450/0.7,(2) hexastarch-6% HES 200/0.6,(3) pentastarch-6%/10% HES 200/0.5 and (4) tetrastarch-6% HES 130/0.4.Other properties such as the C2:C6 hydroxylation ratio, or whether it is made from potato or waxy maize, are generally not labeled on the infusion bottle.The C2:C6 hydroxylation ratio-which potentially affects the elimination of the molecule or its blood coagulation compromising effect-has shown an increasing tendency in commercial products over the years (9:1 in currently available solutions) [5,6].All dextrans, gelatins and older generation HESs are now removed from the market for various reasons [3,4].More recently, the use of 6% HES (130/0.4 or 0.42) has been restricted by the European Medicines Agency and the U.S. Food and Drug Administration as its deleterious effects on kidney function came to light [3,7].On 24 May 2022, the European Commission issued a suspension of the marketing authorizations of HES solutions for infusion in the EU (https://www.ema.europa.eu/en/news/hydroxyethyl-starch-solutions-infusion-recommended-suspension-market (accessed on 11 February 2022, updated on 26 July 2022).The opportunity was given for the individual EU Member States to delay the suspension for no longer than 18 months and keep HES solutions on the market.However, conclusions derived from resource-rich environments, such the EU is, should not be extended to more resource-scarce environments without proper qualifiers.Albumin, the ideal "volume expander", remains expensive and its supply is ultimately limited.While current methods are safe for preventing the transmission of prion-like illnesses with human albumin preparations, all these are contingent on resource investment and societal wealth to support them [8,9].
Early hemodynamic stabilization can be crucial in the prevention of AKI regarding the short warm ischemic time of the kidneys [10,11].A promising tool to discriminate between hypovolemic and normovolemic patients is the hypovolemic index (values between 0 and 1) [12].This parameter is capable of separating these groups of patients (threshold: 0.5), but its validation is still in progress.The first step for hemodynamic stabilization is to achieve euvolemia, which is a wide gray zone without clear boundaries between the volume-sensitive and volume-resistive circulatory states [13,14].Interstitial accumulation of intravenously administered fluids can increase the renal parenchymal pressure dramatically, and therefore the fluid resuscitation with crystalloids is only a question under debate [15].The evaluation of kidney perfusion by ultrasound can aid in finding the right balance between fluids and vasoactive drug therapy.
At the same time and over the past several years, the definition of acute kidney failure has become increasingly precise, fostering earlier diagnosis and standardization across the world.The first systematic, universal definition of acute kidney injury (AKI) was accepted in 2002 (RIFLE criteria) and has been followed by three other generally established ones (AKIN, KDIGO, KDIGO with biomarkers) [16][17][18][19].The studies conducted with thirdgeneration HES show wide differences in the definition of deteriorating renal function, as discussed further below.The severity stages of AKI do not correspond equivocally between the AKI definitions, making it harder to generate a robust comparison [20].AKI itself has multiple possible causes and is featured by different microhemodynamics and humoral/cellular changes depending on the underlying pathological processes [21].Two meta-analyses on this topic were performed in 2013, which also included a few studies conducted with the older generation of HES culminating in harmful renal consequences [22,23].Two other meta-analyses were conducted in recent years to demonstrate the advantages and disadvantages of the administration of 6% 130/0.4-0.42HES in surgical and trauma patients, proving it safe and favorable in terms of hemodynamic properties [24,25].
It is important, however, to recognize that modern HES products may have value due to their low cost, easy storage and represent a meaningful potential alternative in resource-scare environments.Albumin, although an ideal volume expander, remains expensive and its supply is ultimately limited.While current methods are safe for preventing the transmission of prion-like illnesses with human albumin preparations, all these are contingent on resource investment and societal wealth to support them [8,9].To further complicate the scenario, we also recognize that the use of plasma expanders may not entirely come from the expansion of plasma volume.A quantity of 250 milliliters of 5% albumin is really 12.5 mL of albumin, which is a syringeful; it is unlikely to only work by expansion of the intravascular space [26].Shimizu K. et al. have shown in an elegant study that the injection of 20 mL of "plasma expander" hypertonic saline or hypertonic glucose increased blood pressure by suddenly increasing endogenous vasopressin even though plasma volume only increased by 2.3%.The injection of 200 mL of isotonic saline, while expanding plasma volume by 12.7%, did not increase vasopressin levels.
The aim of our narrative review is to conduct a critical re-assessment of the literature on the safety and efficacy of one specific product, the currently used 6% HES (130/0.4 or 0.42), with regard to renal function, independently of any industrial ties or potential conflicts of interest.Although the indication of tetrastarch is also a clinically relevant point, we have not discussed it due to the limited length of the manuscript.Only safety and efficacy concerns are conferred.

The Brief Pathophysiology of AKI
In high-income countries, the three main forms of AKI are the postoperative, the septic and the AKI of cardiac origin, except for forms caused by nephrotoxic agents [21,27].After noncardiac surgeries, the leading cause of renal dysfunction is the ischemic-reperfusion injury due to general or local hemodynamic instability, transport hypoxia due to blood loss and increased intraabdominal pressure [27].In cardiology patients, venous congestion and with on-pump cardiac surgery, the activation of the immune system is added to these confounders as a significant contributing factor [28].Hypovolemia and congestive cardiac insufficiency are accompanied by the high activity of the renin-angiotensin-aldosterone system (RAAS) in contrast to the low activity of the RAAS due to hypervolemia, resulting in absolutely different renal microcirculation.Given the presence of renal capsules in in situ kidneys, with venous congestion, fluid overload and third-spacing, the interstitial pressure can exponentially rise within the kidney parenchyma [29].It is to be understood that from an evolutionary biological standpoint, one would expect fewer escape mechanisms to evolve for surviving fluid overload than coping with hypovolemia.However, septic AKI is characterized by a different intrarenal hemodynamics: the dilatation of the efferent glomerular arteries and the increased patency of shunt vessels produce a low-pressurehigh-flow state, consequently dropping the filtration rate in the glomeruli [30].Besides circulatory changes, several inflammatory mediators play a crucial role in the progression of septic AKI.However, the main contributor of AKI is hemodynamic instability with a potential contribution of nephrotoxic agents, as described recently [31].

The Diagnostic Uncertainties of AKI
The worsening of kidney function represents a continuum.Since no clear boundaries can be observed between physiological and pathological conditions, it is difficult to define infliction points.Despite several known pitfalls, most generally accepted diagnostic systems employ the rise of serum creatinine and the amount of urine output as the basis for detecting AKI [16][17][18][19].However, serum creatinine concentration is considered a 'slowreacting parameter': serum creatinine levels follow clinical changes with an outstanding delay.Moreover, the definition of the perceived "baseline" serum creatinine level further qualifies the perceived frequency and severity of AKI [32].Using eGFR (as suggested by the Acute Dialysis Quality Initiative [ADQI]) or minimum inpatient serum creatinine levels as the baseline inflated the incidence of AKI in comparison to the most recent outpatient serum creatinine levels between 7 and 365 days prior to admission (38.3%, 35.9% vs. 25.5%,p < 0.001, respectively) [33].However, the first admission serum creatinine level underestimated the incidence of AKI compared to the most recent outpatient serum creatinine concentration (13.7% vs. 25.5%,p < 0.001) [33].In this study, the main differences (both false positive and negative) were in the AKIN 1 stage.Based on data from the Beginning and Ending Supportive Therapy for the Kidney (BEST Kidney) study, the estimated serum creatinine (Modification of Diet in Renal Disease [MDRD]) leads to a bidirectional misclassification of patients at enrollment (false negative for Risk: 7.3%; false positive for Failure: 18.7%, false positive for all AKI: 11.7%) and at admission to ICU (false positive for Injury, Failure, all AKI: 5.5%, 14%, 18.8%, respectively) [34].Muscle wasting, sarcopenia, racial differences and fluid overload are important qualifier to interpret serum creatinine values in the ICU settings [32].In an attempt to overcome these difficulties, newer markers (e.g., cystatin C, NGAL, TIMP2 × IGFBP7) are implemented, but their general usefulness is debated [21,27].Urine output is an important parameter contributing to the diagnostic frequency and severity of AKI, but administering diuretics blurs the diagnostic reliability [35].

Studies Conducted with 6% HES 130/0.4 or 0.42 Analyzing Its Renal Effects
The designs of studies conducted to evaluate the deleterious renal effects of 6% HES 130/0.4 or 0.42 (the different molar substitution value represents products of different manufacturers) are listed in Table 1.Four large meta-analyses were published aiming to evaluate the relationship between the administration of HES and the development of AKI [22][23][24][25].Two investigators (CsK, TG) scrutinized all the studies included in the systematic reviews independently.The outcomes and the investigators' critical remarks can be found in Table 2. Two trials (Safety and Efficacy of a 6% Hydroxyethyl Starch Solution vs. an Electrolyte Solution in Trauma Patients (TETHYS); Safety and Efficacy of 6% Hydroxyethyl Starch Solution vs. an Electrolyte Solution in Patients Undergoing Elective Abdominal Surgery (PHOENICS)) are officially registered (NCT03338218, NCT03278548), but no results have been published to date [36,37].
One of the meta-analyses published in 2013 was based on 10 studies concerning RRT with extremely low reported heterogeneity (τ 2 = 0, I 2 = 0%) [22].The largest included study (CHEST, weight: 51.8%; n = 6651) was conducted in hypovolemic patients at any time in the ICU, but the percentage of septic patients was about 23-25% in both groups [38,39].They found a significantly lower incidence of AKI at either the Risk or Injury stage and a non-significant difference in the Failure stage.The only finding referring to kidney damage was the slightly higher rate (7.0% vs. 5.8%, p = 0.04) of RRT in the HES group, but the initiation of RRT was based on the clinicians' discretion; objective criteria were not communicated.The authors' opinion was that this did not affect the study results since the clinicians were unaware of study-group assignment.The second larger study (6S, weight: 20.8%; n = 798) was conducted in septic patients and the diagnostic criteria of AKI were different from the generally accepted systems [40].Nevertheless, the relative risk of AKI was similar in the intervention and the control groups.It is to be noted that patients with AKI at the time of randomization were included with equal frequency in the two groups.There was another study conducted on septic patients (CRYSTMAS, weight: 3.9%; n = 196) which showed no significant difference in the incidence of AKI [41].Similar courses of serum creatinine and biomarkers were observed in both study groups.One small study from China included in the research was designed to demonstrate the effect of HES on intraabdominal pressure [42].The diagnosis of AKI was based on urine output.No data were reported about renal replacement therapy in this study.Another small study contained no data on renal function [43].One of the included studies was in abstract form; four others were conducted with the second generation of HES (the sum of the weight of these five studies was altogether 22.7%) [44][45][46][47].
Generally speaking, hemodynamic data and nephrotoxic agents are reported only in a few studies.Interestingly, the recommended dose of tetrastarch was exceeded in many studies, but this point has been underemphasized in the systematic reviews to date, and none of them investigated a dose-side-effect relationship between the 6% HES 130/0.4 or 0.42 and AKI.Another critical aspect is the possibility of overcorrection of hypovolemia.To date, several methods have been described and employed to detect volume status, but none of them can guide the fluid therapy precisely [48].Despite the fact that the role of fluid overload in developing AKI is well known, it is not mentioned in any study conducted with tetrastarch.
In the Cochrane library, a systematic review was performed to analyze the effects of HES on kidney function [23].This review is from studies referring partially [38,44,45,[49][50][51][52][53][54] or entirely [43,55,56] to unpublished data or data from abstracts only [57].Certain studies included in the research are from published data only [40,41,46,47,[58][59][60][61][62][63][64][65][66][67][68] and one of them contains no data on renal function [43].The high (≥200 kDa) and lower than nowadays commercially available (70 kDa) molecular weight HES solutions are also included.Renal outcomes were determined according to the RIFLE criteria, need of RRT or by the authors' definition.We included in our analysis all the studies conducted with 6% 130/0.4-0.42HES; the details can be found in Tables 1 and 2. The included trials were not selected based on patients' subgroups.Only a statement made in the main text of the article indicates that non-septic patients had fewer adverse effects, but the divergent types of HES make it hard to draw a relevant conclusion for everyday practice in 2023.This systemic review was helpful at the time of writing, but several new data have emerged since then.
One recent meta-analysis (heterogeneity for both AKI and RRT: τ, I 2 = 0%) reported that a 6% 130/0.4HES is safe against different comparator fluids in various subgroups of patients [24].The authors included three studies for demonstrating AKI in cardiac and eleven trials in non-cardiac/mixed surgery patients.One of the cardiac surgery [69] and one of the non-cardiac surgery [70] trials were designed as noninferiority studies, and two other cardiac surgery [71,72] and ten non-cardiac surgery [70,[73][74][75][76][77][78][79][80][81] trials were observational.One of the cardiac trials [62] (weight: 0.5%, total weight of cardiac studies: 6.8%) and two of the non-cardiac ten [76,82] did not report the renal function appropriately (weight: N/A); one applied a 24 h follow-up [75] only, while another one compared HES derived from maize and from potato (weight: 17.5%) [83].The sample size of these studies is less than 100 patients, with two exceptions [73,78] (n = 386 and 534).Both cardiological and non-cardiological surgery studies consider HES to be at least non-inferior regarding renal safety parameters for crystalloids, gelatin and 5% human albumin.
The planning process of the studies has several methodological problems, which can exert a significant impact on the results.A good example is the second largest study (CRISTAL) conducted on septic patients [84].In this trial, crystalloids were administered to only one fourth of the patients in the colloid group.In everyday practice, the first intravenous fluid administered is a crystalloid of any kind (0.9% saline, balanced or hypotonic solution), and colloids are considered second-line drugs [85].In certain studies, a proportion of patients received HES before randomization, an aspect that remained unanalyzed [40,41,86].Hemodynamic instability itself can lead to impaired kidney function and may be an ongoing issue in sepsis or postoperative states [87,88].Surprisingly, hemodynamic data (e.g., the duration and severity of the hypoperfusion period, any organspecific cessation of renal blood flow during surgery, ultrasonographic data about intrarenal blood flow and venous congestion, etc.) were not reported in most of the studies (Table 2).The results are confusing from the perspective of renal detrimental effects as well.The implemented definition of deteriorated kidney function varies in a wide range between decreasing urine output and fulfilling KDIGO criteria with biomarkers.Finally, nephrotoxic mediators and agents are very common in sepsis patients and during intensive care therapy.However, these factors or the lack of these factors are usually not indicated.
We must also consider discriminating among the trials according to different patient subpopulations because of the previously mentioned distinct patho-mechanisms of AKI.
The largest studies (over 1000 patients per group) were conducted only in septic patients, while some middle-sized (500-1000 patients per group) studies were steered in patients who had undergone abdominal surgery and only small studies are available in cardiac surgery patients (Table 1).In multi-center studies, a significant heterogeneity can be observed among data produced by different centers, but the results are not provided according to investigator sites.A further shortcoming is the missing logistic regression analysis.If we assume that HES is an independent influencing factor in the development of kidney failure, it is then critical to be verified by a multivariate logistic regression analysis.In the studies where the harmful impact of HES on kidney function was referred to, no such analysis was carried out in any but one of these; this particular study failed to identify any relationship between the worsening of kidney function (AKI) and the administration of tetrastarch [89].A major shortcoming of all the studies assuming the kidney-damaging effect of HES is that none of them performed a dedicated dose-side-effect analysis.If a drug is harmful, it can be rightly assumed that side-effects occur more often with higher doses and longer use.Based on the data, there would have been an opportunity for this in many investigations, but in no case was such an analysis carried out.As a consequence of this methodological deficiency, in the opinion of the authors of this review, any current meta-analysis is inevitably distorted.
Studies performed in cardiac patients do provide more hemodynamic data but show no significant difference in AKI or the need for RRT [50,52,62,69,90,91].Studies conducted on postoperative patients after abdominal surgery compared HES with other colloids [63,64,70,83,92], while others did so against crystalloids [73,74,78] and some against both [53].These studies proved that HES is not inferior to other colloids or crystalloid infusions.Even in cases where the risk of AKI seems to be higher in the HES group, the 95% CI saddles on 1.0, indicating that relative risk is uncertain.

Studies Supporting the Beneficial Hemodynamic Effects of HES
Although it was not always their primary endpoint, several of the mentioned studies reported the favorable hemodynamic effects of HES relative to crystalloids [41,42,53,73,82,91] or other colloids, [52,69,95], while a few studies are against the favorable circulatory effects of HES compared to crystalloids [64,94].
A large multi-center controlled randomized study conducted by Gondos et al. found that 6% HES 130/0.4 is a valuable alternative to other colloids [96]; 200 mixed postoperative ICU patients were investigated in this multi-center study.After the baseline hemodynamic evaluation was carried out, 10 mL/kg of lactated Ringer's solution, succinylated gelatin 4% w/v, 130/0.4hydroxyethyl starch 6% w/v (HES) or human albumin 5% w/v was administered over 30 min.Hemodynamic measurements were performed at 30, 45, 60, 90 and 120 min.Their findings were supported by Toyoda et al. [97].These studies clearly showed that both tetrastarch and albumin have significant hemodynamic effects even at 120 min, while the hemodynamic effect of crystalloids disappears within 20 min.Another controlled randomized single-center study conducted in 57 severe sepsis patients compared the hemodynamic effects of 6% (130/0.42)HES (250 mL every 6 h) and 20% human albumin (100 mL every 12 h) [56].The administration of a crystalloid solution was allowed as it was considered necessary.The hemodynamic goals were MAP > 65 mmHg, intrathoracic blood volume index (ITBVI) > 850 mL × m −2 and cardiac index > 3.5 L × min −1 × m −2 .The most common source of sepsis was ventilator-associated pneumonia.The decrease of the alveolar-arterial oxygen gradient (AaDO 2 ) was significantly better in the HES group in the first 72 h, with no significant differences in hemodynamic indices.Renal effects were not investigated.

The Role of Hyperchloremia in the Development of AKI
A substantial bias and debate have emerged about whether we should differentiate the solutions based on their chloride content and how this effect further modifies potential interactions with source colloid materials and the plant they are derived from.One should keep in mind that isotonic saline can lead to both hyperchloremia and a significant increase in total body sodium content.Only one liter of 0.9% NaCl contains three times the recommended daily sodium intake.The entire topic is not discussed here in detail for reasons of limited space, but we cite the study conducted on twelve healthy adult male volunteers [98].Renal artery blood flow velocity and renal cortical perfusion were compared by magnetic resonance imaging at 0, 30, 60, 120, 180 and 240 min after starting a 30-min intravenous administration of one liter 6% 130/0.4maize-derived HES in 0.9% NaCl and 6% 130/0.4potato-derived HES in a balanced solution.The authors found similar mean peak serum chloride levels, blood volume, strong ion difference, serum creatinine to serum NGAL ratios and mean renal artery flow velocities between groups, albeit renal cortical perfusion was significantly increased (7% from the baseline) after the infusion of potato-derived HES in a balanced solution, compared with a 2.5% decrease from the baseline in the case of maize-derived HES in 0.9% saline.The authors reported significant hyperchloremia (109 mmol/L vs. 104 mmol/L, p < 0.0001), a greater expansion of extracellular fluid (1484 mL vs. 1155 mL, p = 0.029) and the deterioration of both renal artery blood flow velocity (a 13% decline from the baseline, p = 0.045) and renal cortical perfusion (an 11.7% reduction from the baseline, p = 0.008) after the infusion of two liters of 0.9% NaCl compared with a balanced solution (raised renal circulatory parameters) by the same method [99].At the end of the four-hour observational period, 14% and 12% of saline and balanced solutions remained in the intravascular compartment, respectively.

Conclusions
Summarizing these results, it is the opinion of the authors that the administration of HES is safe and effective if the recommended dose is respected.Restoring circulating plasma volume is essential to prevent renal hypoperfusion.Crystalloid solutions alone fill the extravascular and interstitial space, whereas colloids retain a longer intravascular effect duration.The tissue deposition of HES can be minimized by adherence to the manufacturer's proposal.
Some renal benefits can be achieved by potato-derived HES in a balanced solution.To date, both the published large studies and the meta-analyses show significant bias in the context of the deleterious effect of 6% 130/0.4-0.42HES.The 6% HES (130/0.4 or 0.42) can have a better hemodynamic profile than crystalloid infusions used alone, but its deleterious effect on kidney function remains questionable.Without (1) detailed hemodynamic data, (2) the exclusion of other nephrotoxic events and (3) a properly performed evaluation of the dose-effect relationship, the AKI-inducing property of the 6% HES 130/0.4 or 0.42 could not be accounted for as evidence.We need some well-designed randomized controlled trials to appropriately explore and reflect on clinical problems.