Safety of Albumin and Hydroxyethyl Starch as Priming Fluid for Cardiopulmonary Bypass in Cardiac Surgery: A Meta-analysis of Randomized Controlled Trials

Introduction: Hydroxyethyl starch (HES) has been widely used for volume expansion, but its safety as priming uid for cardiopulmonary bypass has been questioned recently. The aim of this meta-analysis is to compare the safety of albumin and hydroxyethyl starch as priming uid for cardiopulmonary bypass. Methods: Pubmed, Embase database and Cochrane Library were searched for randomized controlled trials (RCTs) involving patients who received HES or albumin as priming uid for cardiopulmonary bypass in cardiac surgery published up to October 2019. Two reviewers independently extracted the valid data, including the length of ICU stay, ventilator time, the length of hospital stay, crystal volume, fresh frozen plasma, platelet input, blood loss, blood platelet count. hemoglobin value, brin, APTT, PT, urea, creatinine and urine volume. Meta-analysis was performed with revman version 5.3. Results: Total 9 RCTs involving 452 patients were included in this meta-analysis. Compared with albumin, HES had similar effects on the length of ICU stay(MD = 0.70 (cid:0) 95%CI (cid:0) -0.14 to 1.55 (cid:0) P = 0.10 (cid:0) I 2 = 89%);ventilation time(MD = 2.31 ;95%CI-3.93 to 8.55;P = 0.47,I 2 = 60%); the length of hospital stay(MD = -0.31;95% CI:-2.00 to 1.37; P = 0.71 ;I 2 = 0%); crystal volume(SMD = 0.26;95% CI:-0.09 to 0.61; P = 0.15;I 2 = 0%); fresh frozen plasma(SMD = 0.25 (cid:0) 95%CI (cid:0) -0.08 to 0.59 (cid:0) P = 0.66 (cid:0) I 2 =0%); platelet input(SMD =-0.17;95% Cl:-0.59 to 0.26;P =0.45;I


Introduction
Currently, cardiopulmonary bypass (CPB) is an important adjunct to cardiac surgery because it improves the success of cardiac surgery. By providing low blood ow perfusion, CPB prolongs the safe operation time and provides basic blood ow supply for various organs of the human body. However, CPB can lead to blood dilution, decrease plasma osmotic pressure, damage vascular endothelial cells (VEC), increase vascular permeability, and cause systemic complications such as postoperative tissue edema, cardiac insu ciency, pulmonary exudation,infection, liver and renal insu ciency [1] . At present, how to select priming uid for CPB is a focus of controversy. CPB priming solution directly affects the change of colloid pressure and the incidence of VEC damage after cardiac operation. Albumin has become the preferred uid for patients undergoing CPB in many centers. Compared with other liquids, albumin is a natural colloid that effectively provides plasma colloid osmotic pressure [2] , protects vascular endothelium, delays brinogen adhesion in circulation pipeline, and reduces surface activation and platelet adhesion [3] . Large randomized controlled trials comparing albumin with other uids have not been conducted in patients undergoing cardiac surgery. Engelman et al. [4] reported that preoperative hypoalbuminemia (< 25 g·L-1) is an independent risk factor for various adverse outcomes after cardiac surgery. Fritz et al. [5] found that hypoalbuminemia was even better than EuroSCORE score in predicting mortality after cardiac surgery, and the cut-off point for hypoalbuminemia was 18 g·L-1. Several other studies have evaluated the effectiveness of albumin as priming uid for CPB.
A retrospective cohort study by Sedrakyan et al. [6] enrolled 19,578 CABG patients, the results showed that compared with the older generation of arti cial colloid, albumin was associated with a reduction in mortality (OR = 0.8, 95% CI: 0.67 to 0.96), and compared to crystal liquids, the albumin group required less priming uid. In addition, the use of albumin as priming uid can reduce postoperative blood loss.
Onorati et al. [7] retrospectively compared low dose albumin with pure crystal as priming uid solution in 377 patients, and patients who received albumin need less blood transfusion and had less blood loss after surgery.
Hydroxyethyl Starch (HES) is another plasma substitute widely used in the clinic practice since 1960s, which is made of high molecular weight amylopectin through degradation, hydroxyethylation and further processing. HES is mainly used to prevent or treat intraoperative hypotension and hemodilution [8,9] . On June 14, 2013, due to safety concerns, the European Medicines Agency (EMA) and Drug Risk Assessment Committee (PRAC) recommended the delisting of HES. At the end of June 2013, U.S. o cials pointed out that HES may increase the death rate of patients, cause the decline of hemostatic function of patients [10] and have serious negative effects on the kidneys [11] , and issued a warning to HES [12] . A meta-analysis by Myburgh et al. [13] showed that HES can increase the amount of blood transfusion during reoperation due to bleeding. These negative effects are not directly related to the average molecular weight and substitution level of HES. However, a recent study showed that the use of arti cial colloid as priming uid for CPB has not raised the risk of organ injury [14] . Moreover, many randomized controlled clinical studies showed that for patients undergoing cardiac surgery with normal organ function and coagulation function, the use of arti cial colloid as priming uid for CPB has not raised the risk of organ injury. At present, arti cial colloidal uids are still widely used for priming uid of CPB, but there is no de nite clinical evidence to support the safety or even superiority of arti cial colloid priming uid of CPB. Therefore, we performed this meta-analysis to compare the safety of albumin and HES as priming uid for CPB to provide reference for clinical application.
Materials And Methods 1.1 inclusion criteria and exclusion criteria 1.1.1 Included in the study: The types of literature included in the study are randomized controlled trial (RCT) regardless of whether blind method and allocation concealment are used. The language is limited to English.
1.1.2 The inclusion criteria were: (1) RCTs; (2) The subjects of the study were heart disease patients; (3) HES was used in the test group and albumin was used in the control group.(4) All were extracorporeal circulation operations.

Search strategies
PubMed, Embase and the Cochrane Library were searched using the following key words and related free words: 'cardiopulmonary bypass', 'albumin', 'Hydroxyethyl Starch'. The relevant clinical trials were those published before October 2019 that met the above criteria. The search was limited to 'randomized controlled trials', 'human', and the language was restricted to English.

Study selection and data extraction
Two reviewers (Ye and Zang) independently screened the search results and obtained the full texts according to the inclusion and exclusion criteria, and independently extracted the valid data. Data extraction and analysis were performed under the supervision of an experienced statistician (Mo).

Quality assessment
Risk of bias assessment was performed using the'Risk of bias' tool in the Cochrane Handbook for Systematic Reviews of Interventions [15] . We assessed each study according to the quality domains of random sequence generation, allocation concealment, blinding of participants and personnel, incomplete outcome data, selective outcome reporting and other bias.

statistical analysis
Data were analyzed by Review Manager (5.3 RevMan, Cochrane Collaboration). The continuous effect data were assessed by mean difference (MD). When median and extreme values were presented in the original articles, these data were converted into mean and standard deviation according to relevant formulas [16] . Non-continuous data were assessed by the risk ratio (RR). Statistical heterogeneity of the data was analyzed quantitatively by the I 2 -test [17] . The decision of to use xed effect or random effect is based on whether we believe that the true effect is the same in all studies, i.e. all of the studies are identical and the true effect size is the same across all studies. If this is believed to be true (which is a very strong assumption, rarely found), then xed effect should be used; if not, random effect should be used instead. Publication bias was tested by funnel plots. Two-sided tests were performed with a signi cant difference de ned at P <0.05.  According to Cochrane's systematic evaluation manual, the methodological quality of the 9 clinical studies included in the study was evaluated for items such as random method, allocation concealment, blind method, withdrawal from missed visits and selective reporting Figure 2). heterogeneity among the studies (P = 0.10 I 2 = 89%),and the random effect model was selected for the analysis. The results showed no signi cant difference in the length of ICU stay between the HES group and the albumin group MD = 0.70 95%CI -0.14 to 1.55 P =0.10 ( Figure 3A). Therefore, the use of albumin and HES as priming uid for CPB in cardiac surgery had a similar effect on the length of stay in ICU. (Figure 3A).  Figure 5C).

Discussion
The use of HES in priming solution for CPB is still controversial. This meta-analysis included 9 RCTs that met the inclusion criteria to compare the impact of albumin and HES on the prognosis of patients. The results showed that compared with albumin, HES group had no statistical difference in the length of ICU stay, ventilator time, the length of hospital stay, crystal volume, fresh frozen plasma, platelet input, blood loss, blood platelet count. hemoglobin value, brin, APTT, PT, creatinine and urine volume, while HES group had lower urea than the control group.
The effects of HES on renal function were reported in the included studies, but the renal function test indexes were different, including blood creatinine, glomerular ltration rate, urine volume, urea, renal replacement therapy and other indexes. Because urea, blood creatinine and urine volume were common indexes reported in previous studies, we chose urea, blood creatinine and urine volume for analysis.
Previous basic research and clinical trials have shown that HES has potential nephrotoxicity. A study found that HES was an independent risk factor for acute kidney injury after cardiac surgery [27] . However, other studies suggested that HES as priming uid for CPB had no side effect on renal function [28] . Our results showed that HES had no different effects on serum creatinine and urine volume compared to albumin. The inconsistent results may be due to the molecular weight of HES. In their study they used macromolecular HES [29] . As the molecular weight of HES decreases, its effect on renal function becomes less. These data suggest that low molecular HES is safe for priming solution during CPB. But HES group had lower urea than the control group,may be the sample size included in this meta-analysis is relatively small.
HES may change the coagulation function in vivo and in vitro by prolonging the coagulation time, reducing the intensity of blood clots, and increasing blood loss of patients [30][31] .However, this metaanalysis found that HES had no different effect on brinogen value, platelet input, platelet count, APTT, PT and did not increase blood loss compared with albumin. It may be related to the new generation HES130/0.4 used in the most of studies enrolled in this meta-analysis. It was reported that the third generation HES130/0.4 did not increase blood loss [32] , because its molar replacement ratio was less than 0.5 and it was a rapidly degradable HES solution [33] . However, the sample size included in this metaanalysis is relatively small, and our conclusion needs to be con rmed by large-sample RCT study.
There was no difference between HES group and albumin group in the amount of liquid required during the operation, indicating that the volume expansion effect of HES was not different from that of albumin. However, large sample RCT studies are needed to con rm these results. Our results also found that HES group had no signi cant difference in the length of ICU stay and ventilation time compared with albumin group, indicating that HES treatment could not shorten the hospitalization time of patients. Since our data were estimated from the original data, our conclusion remains to be con rmed.
This meta-analysis is designed and implemented according to the requirements of Cochranne Collaboration Network. We strictly followed the inclusion criteria and exclusion criteria. Two independent authors chose articles, collected data and evaluated bias risks. Although this meta-analysis tried to use high-quality studies such as RCTs, we cannot completely eliminate the confounding and bias in independent studies, which inevitably lead to certain bias in the results. For example, some studies did not use blind methods, some studies did not report the hidden schemes and the generation of random sequences in detail, and some articles cannot trace other bias sources. In addition, only English literatures were included in this study, which may have publication bias. Moreover, the dosage of HES need to be considered, because the dosage used in different studies was different. In addition, the sample size of some studies was small. More RCTs with large samples and multiple centers are needed to con rm our conclusion.
Taken together, HES had no obvious side effects on renal function, coagulation function and did not prolong the length of ICU stay and ventilation time of patients. These results indicate that HES may be safe to be used as priming uid for CPB. Even with some problems such as insu cient samples, different follow-up time and different dosage of HES, we think that the evidence is credible and the quality grade is judged as medium quality because all the included studies are RCTs, most of the studies used the random grouping and concealment methods, reported complete outcome indicators, had little heterogeneity among the experiments, and had consistent conclusions for all indicators.

Conclusion
This meta-analysis provides evidence that HES as priming uid for CPB in cardiac surgery will not damage renal function, coagulation function or increase bleeding volume and intraoperative infusion volume as well as the length of ICU stay and ventilation time of patients, compared with albumin.

Declarations Author Contributions
Xuming Mo Zhaocong Yang conceived and designed the study. Mingtang Ye edited the manuscript text and performed the statistical analyses. Xiaodong Zang Peicheng Ding analyzed and interpreted the data. Ruonan Wang collected data. Feng Chen, Jirong Qi prepared tables and gures. All authors read and approved the nal version of the manuscript.

Con icts of Interest
The authors declare that they have no competing interests.