Six Percent Hydroxyethyl Starch 130/0.4 (Voluven^®) versus 5% Human Serum Albumin for Volume Replacement Therapy during Elective Open-heart Surgery in Pediatric Patients

HUMAN albumin is still regarded as the preferred colloid in pediatric surgery, although this choice is not supported by robust published data.¹  In adult surgery, artificial colloids such as hydroxyethyl starch (HES) have replaced human albumin as the first choice for volume replacement therapy in many different settings.^2,3 

Tetrastarches have been developed to improve the pharmacokinetics of previous starches generations. As a result of a quicker achieved optimal in vivo molecular weight, plasma substitution effects and clinical efficacy of 6% HES 130/0.4 appear similar to other HESs,^4–6  whereas coagulation parameters are significantly less affected.^6–8 

However, there are only very limited data available from controlled studies on the use of tetrastarches in children. In addition, most of these data are obtained from low-risk population, and/or the volume of tetrastarch used was relatively low.^9,10 

Further evaluation of the efficacy and safety of tetrastarches in a surgical pediatric population will require a surgical procedure associated with the use of high volumes of a colloid and a relatively high risk of complications like bleeding and renal dysfunction.

The aim of this prospective, randomized, double-blind study was to compare 6% HES 130/0.4 with 5% human albumin for volume replacement therapy during elective open-heart surgery, in children aged between 2 and 12 yr. The primary objective was to demonstrate equivalence between the tetrastarch and the albumin solutions with regard to the total volume of colloid solution required in the intraoperative period, including priming of the extracorporeal circulation device (ECC). Secondary objectives included fluid balance, blood loss, and transfusion requirements, urinary biomarkers of acute kidney injury, patient outcome, and incidence of adverse events (AEs) up to postoperative day 28.

The study was part of a postmarketing commitment to the FDA and designed to evaluate the efficacy and safety of 6% HES 130/0.4 in children aged 2–12 yr according to International Conference on Harmonization guideline (International Conference on Harmonization—Guidance for Industry: E11 Clinical Investigation of Medicinal Products in the Pediatric Population) (ClinicalTrials.gov Identifier: NCT00860405).

This phase 4, prospective, randomized, double-blind, parallel-group clinical study was conducted at two centers (Hôpital Universitaire des Enfants Reine Fabiola, Brussels, Belgium and Algemeines Krankenhaus der Stadt Linz, Linz, Austria). It was approved by the Ethics Committee Providing Single Opinion of the Vrije Universiteit Brussels, Belgium, which supervises the local Hôpital Universitaire des Enfants Reine Fabiola Ethics Committee (reference code 2008/236) and by the Ethics Committee Providing Local Opinion of the two institutions (Brussels, Begium: 2008/43 and Linz, Austria: 2008/53). The study was assigned EudraCT trial no. 2008-006749-18 and ClinicalTrials.gov identifier no. NCT00860405.

After written parental informed consent and patient assent (if required) were obtained the day before surgery, patients aged 2–12 yr undergoing elective open-heart surgery were randomized to receive up to 50 ml/kg body weight (BW) of either 6% HES 130/0.4 (Voluven^®; Fresenius Kabi GmbH, Bad Homburg, Germany: HES group) or 5% human albumin (Baxter Deutschland GmbH, Unterschleißheim, Germany: human serum albumin [HSA] group) for intraoperative volume replacement including priming of the ECC.

Inclusion criteria were age 2–12 yr, American Society of Anesthesiologists physical status II to III, elective cardiac surgery for congenital heart disease requiring ECC. Exclusion criteria were: allergy to HES and human albumin preparations, a known contraindication against scheduled concomitant medication, total ECC volume less than 400 ml, American Society of Anesthesiologists physical status greater than III, abnormal hemostasis (international normalized ratio above 1.5, activated partial thromboplastin time 1.5 times above upper limit of normal, platelet count less than 100 × 10⁹/l, or fibrinogen less than 100 mg/dl), renal disease with oliguria or anuria unrelated to hypovolemia, need for preoperative dialysis treatment, intracranial bleeding, severe hypernatremia or hyperchloremia, a noncompliance risk, participation in a clinical drug trial in the last 2 months or concomitantly. Children were evaluated at the preoperative visit by the anesthesiologist in charge at least 1 day before the planned procedure. The protocol was proposed at that time, and the written informed consent was collected several hours later, at the latest on the morning of the surgery.

Anesthetic, surgical, and bypass techniques important for the study objectives are described in appendix 1. In particular, tranexamic acid was given as a bolus followed by a continuous infusion in all patients. Heparin was given to obtain and to maintain an activated clotting time above 440 s. Protamine was given at the end of ECC to recover the activated clotting time in a range of +20% of baseline value.

For intraoperative volume replacement including ECC priming, the patients could receive up to 50 ml/kg BW per day of study drug. Regarding the priming of the ECC, the dosage of the study drug depended on the patient’s BW and the ECC circuitry used. For intraoperative volume replacement before or after the ECC, the amount of the study drug not used for priming could be given, up to the maximum dosage for the individual patient, if needed. No specific algorithm for fluid administration was used. Infusion rates were adjusted to individual needs at the discretion of the anesthesiologist in charge of the patient, to maintain a mean arterial pressure within the range of 50–85 mmHg. If the maximum dose of the study drug was reached, 5% albumin solution could be used as rescue colloid. There was no daily dose limitation for the rescue colloid.

In the postoperative period, 5% albumin was the only colloid permitted in order to achieve hemodynamic targets. The use of inotropes and vasopressors was left at the discretion of the anesthesiologist and no specific algorithm was used.

Erythrocytes were transfused to maintain a hematocrit level between 20 and 25% during ECC and 25% after ECC. Fresh frozen plasma, platelets, and fibrinogen were left at the discretion of the anesthesiologist in charge according to the clinical situations.

Hemodynamics (heart rate, mean arterial blood pressure, central venous pressure, fluid input/output [plus calculated balance]) and the use of vasoactive and inotropic drugs were recorded at baseline (after induction of anesthesia), during surgery, and in the intensive care unit (ICU) until the morning of the second postoperative day.

The safety variables assessed comprised blood gas analyses, laboratory examination, hemostasis, and urinary biomarkers of acute kidney injury (α-1 microglobulin, neutrophil gelatinase–associated lipocalin, N-acetyl-β-glucosaminidase), and they were measured at baseline (after induction of anesthesia), and in the ICU on the morning of the second postoperative day. Measurements were performed by a central laboratory using commercially available methods/devices: α-1 microglobulin: turbidimetry/COBAS INTEGRA 700 Roche (Roche Diagnostics GmbH, Mannheim, Germany); neutrophil gelatinase–associated lipocalin: ELISA/ELISA photometer (ANthos Labtec Instruments, GmbH, Salzbourg, Austria); and N-acetyl-β-glucosaminidase: photometry/KoneLab 60 (Thermo Electron GmbH, Dreieich, Germany).

Calculated blood loss was determined on estimated blood volume, pre- and postoperative day 1, hematocrit, and volume of erythrocytes transfused.¹¹  Also, the amount of transfused blood products, duration of ECC and mechanical ventilation, type (risk) and duration of surgery, and length of ICU stay were documented. AEs were followed up until day 28 after surgery.

Adverse events and serious AEs (SAEs) were defined according to International Committee of Harmonization (ICH) guidance and criteria (E 2A) and graded according to Common Terminology Criteria for Adverse Events. For all AEs, the causal relationship to study medication and to study procedures was to be assessed by the investigator as related or not related. An AE was assessed as related, if there was a “reasonable causal relationship,” indicated by situations where it cannot be excluded that the study drug has caused or contributed to the AE/SAE or a plausible temporal relationship between AE/SAE and administration of the study drug and no alternative reason can be confirmed. Each assessment should have been shortly reasoned by the investigator; in case of “no reasonable causal relationship” the alternative cause should have been clearly stated.

Each patient who qualified for entry into the study was assigned a patient number consecutively in chronological order in each center. This number was used to identify the patient throughout the study. Randomization of patients was performed by means of a macro written in SAS^®, version 9.3.1. (SAS Institute Inc., Cary, NC). The computer program used the method of randomly permuted blocks. The block size was four, which was not revealed to the investigational sites before data base lock. Within each block, the number of patients allocated to each of the two treatments was equal. A block size of four was chosen to avoid possible imbalances between groups. Randomization was stratified according to the two different total ECC volumes (400 or up to 800 ml), which took into account the patient’s BW. Within each ECC group, the next patient eligible for randomization received the lowest available randomization number at the study site on the day before surgery. Each patient was given only the investigational product carrying his or her randomization number. The secrecy of the randomization codes was guaranteed by the fact that the statistician who made the randomization was different from the statistician doing the analysis. This ensured blinding of all study personnel until randomization code was broken after database closure.

Blinding of the study medication was performed by a third-party manufacturer in Germany in accordance with the relevant good medical practice requirements. Study drug was labelled with the corresponding randomization number, which allowed a double-blind administration to the patient. The allocation to the treatment groups was not known to any of the investigators, the surgeons, nurses, patients, or the sponsor (Fresenius Kabi Deutschland GmbH, Bad Homburg, Germany) until completion of the study. This also applied to the data manager (Omnicare Clinical Research Inc., King of Prussia, PA) and statisticians (DATAMAP GmbH, Freiburg, Germany) involved in the analysis of the study.

At the study site, the heart–lung machine was prepared by the perfusionist who did not participate in the study. Regarding the study fluids the glass bottles and the aluminium caps were overwrapped to disguise the different colors of the solutions. Brown/dark-yellow infusion sets were used. The heart–lung machine was located in such a way that the anesthesiologist could not look at the venous reservoir. The same unblinded perfusionist prepared those bottles of the study medication needed by the investigator for volume replacement. The volume requested by the investigator was administered through brown/dark-yellow syringes.

Blinded data were checked (100% source data) by monitors independent from the study sites. In addition, a blinded safety review was carried out. Database was kept blinded until all data were collected and validated. The blinding had not to been broken for any safety issue.

The aim of the study was to prove that the investigational drug (HES) had clinical equivalence in comparison with the control drug (HSA). Equivalence was defined as meeting both of the following two conditions:

• The volume of investigational drug was at most 45% smaller than the volume of control drug (i.e., volume of investigational drug at least 0.55 × volume of control drug).

• The volume of control drug was at most 45% smaller than the volume of investigational drug (i.e., volume of control drug at least 0.55 × volume of investigational drug).

It was desired to prove two-sided relative equivalence. In terms of the ratio of means investigational/control (plus any rescue colloid) this led to the equivalence range of 0.55–1.82 because 1/0.55 = 1.82. This ratio had been proposed by the FDA.

For the primary parameter, the volumes of intraoperative investigational drug, CIs were based on an ANOVA model that included fixed effects for treatment and center. The method of Fieller was used to calculate a confidence for the ratio of volumes.¹²  Equivalence is proven statistically if the 95% CI for the ratio of means was fully included in the equivalence range. This procedure has a type I error of 2.5%.

The statistical analysis plan was finalized before unblinding the database. Standard descriptive statistics were also used for all variables. Tests between treatment groups for categorical variables were performed with Fisher exact test or its generalization to tables larger than 2 × 2, the Freeman–Halton test. Numerical variables were analyzed by means of the Mann–Whitney U test or an ANOVA with treatment and center as effects. Numerical variables collected over time were analyzed by means of repeated measures analysis of covariance (ANCOVA), including treatment and center as effects and additionally the baseline value as covariate.

All P values were two sided; a P value of less than 0.05 was regarded as statistically significant for parameters other than the primary one. All effects in the analyses of variance and analyses of covariance were fixed effects. All analyses were carried out with the software SAS, versions 9.1.3. and 9.3 (SAS Institute Inc.).

Assuming a coefficient of variation of 0.363 and a desired power of 90% with a type I level of 2.5%, N = 11 patients per treatment group were needed. The power was calculated a priori by means of the software SAS, version 9.1.3, PROC POWER. Finally, 2 × 30 patients were planned to be included in order to get some safety information and to be able to finalize the study within the timeframe given by the FDA.

The following populations were distinguished in the analysis:

The intent-to-treat (ITT) population was defined as all patients who were randomized (receiving a randomization number). The safety population was defined as all patients who were treated with study medication. The per-protocol (PP) population was defined as all patients in the ITT population, without any major protocol deviation.

Equivalence was summarized for both the PP population and ITT population. The PP population was considered the primary population for analysis. Safety was summarized for the safety population.

Patients were enrolled in the two centers from March 31, 2009 (beginning of enrolment) to August 5, 2010 (last patient completed). In Brussels, 60 patients were screened, 40 patients were randomized, and 39 patients were treated. In Linz, 39 patients were screened and 21 patients were randomized and treated (fig. 1).

The most frequent major protocol deviation was that rescue colloid was administered before study drug had been given (one patient in the HES group and two patients in the HSA group). The other reported major violation was that for two patients in the HSA group rescue colloid had been administered before the maximum dosage of study drug had been administered. Furthermore, one patient in the HSA group received aspirin within 14 days before surgery and was excluded after randomization. One patient in the HES group had incomplete documentation of the intraoperative volume replacement.

The safety (2n = 60) and the ITT populations (2n = 61) were identical except for one patient who was randomized but not treated with the study medication. Fifty-five patients were included in the PP population. Fifty-two patients completed the study, because eight patients completed the treatment and the postoperative period but were then lost to follow-up (day 28).

Table 1 presents the demographic data, other baseline characteristics, and surgical data of the two treatment groups. No relevant differences for mean or median values were found between both groups, except for the durations of ECC and aortic cross-clamping, which were longer in the HES group. Eight patients in the HSA group and four in the HES 130/0.4 group were extubated within 4 h postoperation.

The mean volume of colloid solution required until end of surgery was 36.6 ± 11.8 ml/kg BW (mean ± SD) in the HES group and 37.0 ± 11.9 ml/kg BW in the HSA group (PP population; ratio of means HES/HSA = 0.98, 95% CI [0.84–1.16]; P < 0.0001; table 2). The CI for the ratio of means was within the equivalence range (0.55–1.82) for the PP population, thereby showing significant equivalence between the treatment groups. Analysis of the ITT population supported this result (table 2).

Three patients in the HES group required a rescue colloid in addition to the study drug until end of surgery, compared with five patients in the HSA group.

Intraoperative fluid balance tended to be less positive in the HES than in the HSA group (ITT analysis: P = 0.050; table 3). This was confirmed in the PP population analysis (15.4 ml/kg BW; range, −70.2 to 55.6 vs. 27.7 ml/kg BW; range: −3.4 to 77.2; P = 0.047).

There was no significant difference between the treatment groups in the course of the study regarding hemodynamic parameters (appendix 2).

From the end of the bypass up to arrival in the ICU 15 patients in the HES group and 21 patients in the HSA group received inotropes. In addition to this, seven patients in the HES group and four in the HSA group received vasopressors. As a consequence, nine patients in the HES group and four patients in the HSA group did not receive any of these drugs during this period.

There were no differences between both groups regarding transfused blood products and calculated blood loss (table 4).

No relevant differences were found between the treatment groups in the course of the study with regard to blood gases and parameters for hematology and clinical chemistry (data not shown) and hemostasis (appendix 3).

Renal biomarkers increased in all patients after surgery, but without significant differences between groups (table 5). Two patients in the HSA group and no patient in the HES group required renal replacement therapy during the course of the study.

All patients of the safety population, except one patient in the HES group, had at least one AE. In the HES group, 11 patients experienced 16 SAEs, whereas 7 patients of the HSA group experienced 20 SAEs (appendix 4). None of the SAEs were related to study medication. Multiorgan failure was reported for one patient in the HES group and for two patients in the HSA group (appendix 4). Specifically regarding the incidence of pruritus, this AE was reported only in two patients in the HSA group. However, scratching behavior of the patients was not routinely assessed by the nursing staff.

Duration of postoperative mechanical ventilation and ICU length of stay were not different between groups. Whereas time of mechanical ventilation was 0.6 days (median, range, 0.2–5.2) and length of ICU stay 3.1 days (median, range, 0.9–17.2) in the HES group, mechanical ventilation was 0.4 days (median, range, 0.7–18.0), and ICU length of stay was 3.1 days (median, range, 0.8–37.0) in the HSA group. No patient died during the 28-day follow-up period.

The current study demonstrates equivalence between 6% HES 130/0.4 and 5% human albumin with regard to the total volume of colloid solution required for intraoperative volume replacement therapy including priming of the ECC in children undergoing elective cardiac surgery. Intraoperative fluid balance was less positive in the HES group as compared with that in the HSA group. Blood loss, volume of blood products transfused, incidence of acute renal failure, and AEs were not different between the two groups. However, our study was too small to make any definitive statement about safety. Ruling out any safety concerns in this population would require hundreds if not thousands of patients.

These results are fully in line with those of the only other study that compared 6% HES 130/0.4 and human albumin in the same clinical setting.¹³  However, the latter study had some limitations: it was single center, not double blinded; all patients received aprotinin that has been since withdrawn from the market; and most of the observations were limited to the first 24 h after surgery. In addition, a 4% albumin solution was used, which is slightly hypooncotic. In contrast, the current study was a two-center, double-blinded trial, patients received tranexamic acid, and they were followed up for 28 days after surgery.

Some studies have reported that HES 130/0.4 may be associated with alterations of coagulation parameters in adult and pediatric cardiac surgery patients.^14–16  However, these ex vivo alterations do not necessarily translate into clinically relevant bleeding as shown by Chong et al.¹⁶  in children or by Schramko et al.^14,15  in adults. The current results did not reveal any differences between both treatments regarding transfusion of blood products and calculated erythrocyte loss. A recent systematic review reported that the use of tetrastarches is not associated with an increase in blood loss and erythrocyte transfusion in surgical patients and in particular in cardiac surgery.¹⁷ 

There have been concerns about the effect of starches on renal function in general.¹⁸  Osmotic nephrosis-like lesions, a decrease in tubular flow during glomerular filtration of colloids, and an accumulation of small molecules in the tubuli have been proposed to explain the deleterious effects of high–molecular-weight starches on renal function.¹⁹  However, a recent systematic review reported that there is no indication in the actual literature that the use of tetrastarches induces adverse renal effect in surgical patients.¹⁷  In addition, a prospective, multicenter, observational study concluded that moderate doses of tetrastarches appeared to be safe in children, even neonates and small infants.⁹  This study was one of the first to assess the effect of this tetrastarch on sensitive renal biomarkers in pediatric cardiac patients. Although high volumes of HES 130/0.4 were used, perioperative changes in these urinary renal biomarkers, as well as in plasma creatinine levels, were not different when compared with those observed with human albumin. Need for renal replacement therapy for up to 28 days after surgery was only observed in two patients of the HSA group but in no patient allocated to the HES group. On the basis of the results of the current study, HES 130/0.4 seems to be as safe as HSA 5% with regard to its effect on renal function.

The incidence of AEs was high in this study, which is not unexpected, if specifically documented in this high-risk population. However, it is important to consider that there was no difference between groups. Incidence of SAEs was also not different between the two groups, and none of them was considered to be related to study medication.

The results of this study should be interpreted taking into account the background of the following limitations: first, the study was not powered to allow for definite safety conclusions. Indeed, power analysis was done to assess equivalence in terms of colloid volume replacement in this trial. Second, anesthetic technique and ECC management were mostly left to the discretion of the senior anesthesiologist in charge of the children. However, given the double-blind design and the randomization by block size of four in each center, we do not think that this may have influenced our results. Third, no specific algorithm to infuse the colloids was used, because most of the study medication was used to prime the heart–lung machine. In addition, dynamic parameters like pulse pressure variation or stroke volume variation have not yet been validated in pediatric cardiac surgery and are not relevant during open-chest procedures. Our main hemodynamic goal was to maintain mean arterial pressure within the range of 50–85 mmHg, with fluids. Use of inotropes and vasopressors was left to the discretion of anesthesiologist in charge of the children. There was no predefined algorithm regarding when to add inotropes to reach this goal, as it was not possible to standardize the inotropic approach between the two centers. However, because of the randomization by center and the double-blind design of the study this should not have affected the results. Fourth, follow-up was limited to 4 weeks. However, renal dysfunction generally develops within days after surgery and therefore should have been covered by the 28-day follow-up period. Renal biomarkers were measured after arrival in ICU and on the second postoperative morning and did not show significant difference between the two groups. Measurement performed after arrival in the ICU correspond to approximately 3.5–4 h after initiation of the bypass, which are globally in line with the kinetics of neutrophil gelatinase–associated lipocalin measured by Krawczeski et al. in children undergoing cardiopulmonary bypass.²⁰ 

The current study demonstrates that 6% HES 130/0.4 and 5% albumin are equivalent with regard to the total volume of colloid solution required in the intraoperative period in children aged 2–12 yr undergoing cardiac surgery. The various safety analyses revealed no relevant differences between the treatment groups. However, the small sample size of the study precludes any firm conclusion regarding safety.