The Effect of Colloids versus Crystalloids for Goal-Directed Fluid Therapy on Prognosis in Patients Undergoing Noncardiac Surgery: A Meta-Analysis of Randomized Controlled Trials

Background Goal-directed fluid therapy (GDFT) contributes to improvements in intraoperative fluid infusion based on objective parameters and has been widely recommended in clinical practice. In addition, increasing evidence reveals that GDFT can improve the prognosis of surgical patients. However, considering the individual characteristics of colloids and crystalloids in clinical use, it is uncertain as to which type of fluids administered is associated with better outcomes in the condition of GDFT. Objectives To evaluate the effect of colloids versus crystalloids under GDFT on prognosis in patients undergoing noncardiac surgery. Data Sources. Randomized controlled trials (RCTs) from PubMed, EMBASE, Ovid MEDLINE, CNKI, Cochrane library, and reference lists of relevant articles. Methods Two investigators independently screened and reviewed studies for inclusion and performed data extraction. Our primary outcome was a composite of postoperative complications. The secondary outcomes were (1) mortality at the follow-up duration; (2) postoperative complications of several organ systems, including cardiac, pulmonary, digestive, urinary, nervous system, and postoperative infection events; and (3) hospital and ICU length of stay. Heterogeneity was assessed by the I2 and chi-square tests. The odds ratio (OR) of the dichotomous data, mean difference (MD) of continuous data, and 95% confidence intervals (CI) were calculated to assess the pooled data. Results Of 332 articles retrieved, 15 RCTs (involving 2,956 patients undergoing noncardiac surgery) were included in the final analysis. When the data were pooled, patients in the colloids and crystalloids group revealed no difference in postoperative composite complications (OR = 0.84, 95% CI = 0.51–1.38, P=0.49) under GDFT. Regarding the secondary outcomes, patients in the colloids group were associated with fewer digestive system complications (OR = 0.64, 95% CI = 0.41–0.98, P=0.04). However, no difference was found in mortality (OR = 1.37, 95% CI = 0.72–2.58, P=0.34), complications of the cardiac system (OR = 1.49, 95% CI = 0.66-3.37, P=0.34), pulmonary system (OR = 0.89, 95% CI = 0.62–1.28, P=0.53), urinary system (OR = 1.05, 95% CI = 0.61-1.80, P=0.87), nervous system (OR = 1.04, 95% CI = 0.55–1.98, P=0.90), postoperative infection events (OR = 0.89, 95% CI = 0.75–1.07, P=0.22), length of hospital stay (difference in mean = −0.71, 95% CI = −1.49–0.07, P=0.07), and ICU stay (difference in mean = −0.01, 95% CI = −0.20–0.18, P=0.95) between patients receiving GDFT with colloids or crystalloids. Conclusion There is no evidence of a benefit in using colloids over crystalloids under GDFT in patients undergoing noncardiac surgery, despite its use resulting in lower digestive system complications.


Introduction
An important duty of anesthesiologists is to maintain the stability of intraoperative hemodynamics by suitable fuid infusion, which can meet the needs of sufcient organ perfusion.Previous works confrmed that fuid therapy decisions can infuence the clinical prognosis of patients [1].Optimization of perioperative fuid treatment often results in enhanced postoperative outcomes, reduced perioperative complications, and shortened hospitalization [2].Moreover, fuid administration based on objective indicators of individual responsiveness to liquid therapy can reduce postoperative complications and enhance patient recovery from surgery [3,4].Tis method of fuid administration based on a variety of objective hemodynamic parameters is called goal-directed fuid therapy (GDFT).It operates on fowbased dynamic hemodynamic parameters such as pulse pressure variation (PPV), stroke volume (SV), and stroke volume variation (SVV) [5].In 1997, Sinclair et al. found that intraoperative intravascular volume loading to optimal SV guided by esophageal Doppler resulted in a more rapid postoperative recovery and a signifcantly reduced hospital stay [6].Since then, GDFT has attracted much attention and proved its positive role in various clinical situations [7][8][9].However, which type of fuid infusion favors better outcomes in the condition of GDFT is still the subject of debate and remains unclear.In general, two main types of fuid are used intraoperatively in clinical practice: colloids and crystalloids.Crystalloid solutions are the most commonly used fuids because they are inexpensive, readily available, and relatively nontoxic [10].However, crystalloid solutions have a short half-life and leave the intravascular space within minutes, thus providing little hemodynamic support.It also accumulates in tissues, including the lungs and incision sites, and promotes edema, weight gain, and prolonged recovery, which all worsen the prognosis of surgery patients [10][11][12].In contrast, colloids might be more efcient than crystalloids in expanding plasma volume because they are retained within the intravascular space and maintain colloid oncotic pressure longer than crystalloids, thereby reducing tissue edema, but their cost has limited their use [10,13].Moreover, colloids may also afect coagulant function and increase the risk of acute kidney injury (AKI) and permeability of the microvasculature under conditions of surgery, infammation, or trauma.Te leakage of colloidal macromolecules into interstitial tissues can also increase interstitial colloid osmotic pressure (which is consistent with its role in the blood vessels), thus exacerbating edema.Previous clinical studies that focused on the efcacy of colloids and crystalloids in the condition of GDFT have yielded diferent conclusions [14][15][16][17][18][19][20][21][22][23][24][25][26][27][28] and gained no consensus on the results; thus, we performed a meta-analysis to evaluate the efect of colloids versus crystalloids under GDFT on prognosis in patients undergoing noncardiac surgery.

Methods
Te present meta-analysis adhered to the preferred reporting items for systematic reviews and meta-analysis guidelines [29].All analyses were based on previously published studies; thus, no ethical approval or patient consent was needed.It has been registered in the international prospective register of systematic reviews (CRD42020153043).

Selection Criteria.
Te selection criteria are as follows.

Intervention.
Intraoperative fuid infusion decision of GDFT with colloids or crystalloids.
2. 1.3.Comparison.Patients who received colloids and those who received crystalloids in the condition of GDFT were compared.
2. 1.4.Outcomes.Te primary outcome was a composite of postoperative complications.Te secondary outcomes were (1) mortality at the follow-up duration; (2) postoperative complications of several organ systems, including cardiac, pulmonary, digestive, urinary, nervous, and postoperative infection events; and (3) length of stay at hospital and ICU.Any study containing the primary outcome or secondary outcomes was screened and then included in the fnal analysis based on its full text.

Data Sources.
RCTs from PubMed, EMBASE, Ovid MEDLINE, CNKI, and Cochrane library were reviewed (last updated in May 2023) to identify eligible studies that met the inclusion criteria by two investigators who searched independently.Moreover, we referred to the references of the included studies for potential trials.

Date Extraction.
After completing the frst step of literature screening to confrm the fnal data study, two members of the team began to extract data according to the outcome indicators.In order to ensure the accuracy of data extraction, the two members of the data extraction do not know each other's data results.After the data extraction is completed, two members check the data extraction results.

2
Anesthesiology Research and Practice Data abstracted from each individual trial included the population, intervention, comparison, outcome, study type, and other detailed characteristics, such as frst author name, year of publication, goal-directed strategy, basal fuid therapy, sample size, and conclusion (Table 1).Our secondary outcomes include postoperative complications of several systems, including cardiac, pulmonary, digestive, urinary, nervous system, and postoperative infection.Among them, complications of the cardiac system include acute heart failure, myocardial infarction, ischemia, arrhythmia, and acute coronary syndrome.Complications of the pulmonary system include pulmonary embolism, pulmonary edema, respiratory failure, pneumonia, pulmonary pleural efusion, and pulmonary efusion.Digestive complications include bowel and surgical anastomosis stricture, anastomotic leak, internal or external fstulas, efusion, gut paralysis, peritonitis, gastrointestinal tract dysfunction, intestinal obstruction, GI failure/small-bowel obstruction, and postoperative nausea and vomiting.Complications of the urinary system include renal dysfunction requiring dialysis, progressive renal insufciency, and AKI.Complications of the nervous system include transient neurologic disease, stroke, and postoperative confusion.Te complications mentioned above and relevant data were all derived from the fnal included studies.
A total of 332 articles were retrieved by searching PubMed, EMBASE, Ovid MEDLINE, CNKI, the Cochrane library, and reference lists of relevant articles.We excluded 128 duplicate articles and an additional 182 articles that did not meet the selection criteria after reviewing the abstracts.We further excluded seven articles after reviewing the full text.Finally, 15 articles were included in the fnal analyses.Figure 1 shows how we identifed the required studies according to the PRISMAfow diagram.
We also assessed the methodological quality of the included studies by the Jadad score, which is considered the optimal valid and reliable tool to assess the methodological quality of a clinical trial [30].Te result of the assessment is shown in Figure 3.
2.8.Sensitivity Analysis for Primary Outcome.STATA16.0 software was used for performing the sensitivity analysis for the primary outcome and the result is presented as Figure 4. Te results of four studies were stable, and the results of two studies were slightly unstable.Overall, the results are relatively reliable.
2.9.Statistical Analysis.We used review manager (RevMan for Windows, version 5.3; Cochrane Collaboration, Oxford, UK) to perform most analyses.We calculated the odds ratio (OR) for dichotomous (postoperative complications and mortality) and mean diferences (MDs) with 95% CIs for continuous (hospital and ICU length of stay) data.We quantifed statistical heterogeneity using the I 2 and chisquare statistics and considered heterogeneity to be substantial if the I 2 value was greater than 50%.Te results of the study are presented in the form of a random-efects model.We considered a P value < 0.05 to be statistically signifcant.If the median and interquartile range (IQR) were reported in a study, we assumed that the median of the outcome variable was equal to the mean response and that the interquartile range was approximately 1.35 standard deviations in width [31].We also performed a sensitivity analysis for the primary outcome using STATA16.0 software.

Discussion
In the present meta-analysis, we included 15 RCTs involving 2,956 patients undergoing noncardiac surgery under the condition of GDFT with colloids or crystalloids.Te results indicated that GDFT with colloids reduces digestive system complications compared with crystalloids.However, in terms of the composite of, mortality and postoperative complications, including cardiac system, pulmonary system, urinary system, nervous system, and postoperative infection events, and length of hospital or ICU stay, colloids under the GDFT protocol do not ofer any additional benefts over crystalloids in this population.Adequate fuid management during surgery is of utmost importance to maintain adequate perfusion and oxygen delivery to tissues.Traditional fuid therapy protocols based on static hemodynamic targets such as central venous pressure and delayed volume status indexes, such as blood In contrast, GDFT relied on sophisticated dynamic measures of volume status and may serve as a better choice for perioperative fuid administration.Horgan et al. [32] reported a shorter hospital stay and decreased morbidity in patients undergoing elective colorectal resection with a protocol-based fuid optimization programme using   Anesthesiology Research and Practice intraoperative oesophageal Doppler monitoring.In a wellconducted meta-analysis of eight RCTs, Javier et al. [33] reported a signifcant reduction in mortality associated with GDFT compared with conventional fuid therapy in adult noncardiac surgery patients.Another meta-analysis of 76 trials [34] suggested that GDFT during general anesthesia might decrease mortality, length of stay at hospital, and several postoperative complications.However, there is no consensus regarding which type of fuid is associated with better outcomes among patients who undergo GDFT.
Colloids and crystalloids all have their own advantages and disadvantages; for example, colloids have longer intravascular persistence and volume expanding efects, thus resulting in lower volume requirements and less

Total (95% CI)
Total events Heterogeneity: tau 2 = 0.07; chi 2 = 9.18, df = 6 (P = 0.  Anesthesiology Research and Practice extravascular edema but causing more anaphylaxis and adverse renal and coagulation efects [35].In contrast, crystalloid solutions are much cheaper, conveniently acquired, and relatively nontoxic but provide little hemodynamic support and promote edema and weight gain, thus prolonging recovery.In 2015, Ripollés et al. published a meta-analysis [36] that included six RCTs concerning the efect of colloids versus crystalloids for GDFT on prognosis in patients undergoing noncardiac surgery and observed no positive conclusions.However, this research failed to explore the detailed complications of individual systems, and the data pooled in this meta-analysis are extremely limited (only two studies included information about relevant outcomes).Tus, we performed the present meta-analysis to explore the detailed prognosis, which included a composite of postoperative  Te fndings of our meta-analysis suggest that patients who received colloids to achieve GDFT were associated with fewer digestive complications (P � 0.04).Te reason why colloid infusion is associated with fewer digestive complications may be because intraoperative GDFT with accurate targeting of fuid may prevent excessive fuid administration.Research shows that 16% of colloids and more than 68% of the saline solution escaped into the extravascular fuid compartment 1 h after the infusion [37].However, large amounts of infused fuids were less efective and impaired gastrointestinal function.Edema of the intestines and other tissues may be responsible for poor tissue oxygenation and postoperative gut dysfunction, thus contributing to increased digestive system complications.Te present metaanalysis found a lower length of hospital stay in the colloid group.Zhang et al. [23] also reported that GDFT with colloids decreased the length of hospital stay compared with crystalloids because their study proved a lower number of postoperative complications in the colloid group.However, this meta-analysis found a similar risk of postoperative complications between the two groups and a high heterogeneity (I 2 � 76); thus, this result must be carefully interpreted as a negative result.
It is worth noting that although the present metaanalysis revealed no signifcant diference in complications of the cardiac system, the results of the subgroup analysis indicated that colloids under GDFT were associated with a higher risk frequency of cardiac complications (P � 0.002) with no heterogeneity (I 2 � 0) during the inhospital period.When our team combed through the data from each study, we found that the results were probably infuenced by Yates et al.'s fndings: in this study, 34 adverse cardiac events occurred in the colloid group, while only 14 occurred in the crystalloid group, far exceeding the incidence seen in other studies.Unfortunately, the Yates's study does not analyze and explain the reason of diferences in the cardiac system complications.Considering large molecular weight and difculty crossing the endothelium of colloids, solutions are expected to remain longer in the intravascular space by using colloids than crystalloids [38].For example, McIlroy and Kharasch [37] found colloids were associated with an intravascular volume expansion efect twice as large as that of the crystalloid solution.What's more, results from a subanalysis of a large multicenter randomized trial by Kabon et al. [39] revealed that the cardiac index increased signifcantly more immediately after a colloid bolus administration as compared to a crystalloid bolus administration.However, their results suggested that the time-weighted average in the cardiac index was only 200 mL higher in the colloids group as compared to the crystalloids group which means a statistically signifcant but clinically unimportant diference in the fow-driven parameters was found.In this context, the net efect of intraoperative goaldirected hydroxyethyl starch administration on the cardiac index is too small to have a relevant impact on clinical outcome, such as complications of the cardiac system.Also,  Anesthesiology Research and Practice in the trial conducted by Feldheiser [18], the benefcial intraoperative efect by the HES colloid solution included a higher cardiac output compared with crystalloids.However, there were no signs of clinical beneft in terms of a reduction in postoperative complications by using a goaldirected hemodynamic algorithm to optimize stroke volume with balanced starch compared with balanced crystalloid.Terefore, it seemed reasonable that the administration of fuids under GDFT protocol might be more important, whereas the type of fuid administered during surgery plays a much smaller role than that was initially assumed.
Although the intensive care setting reported potential nephrotoxic efects and kidney injury in critical and sepsis patients with colloids [40][41][42], studies performed in a surgical context have not been reported, and the specifc efect of colloids on kidney function is still uncertain.In this metaanalysis, we extracted information about urinary system complications in seven RCTs [14,15,17,22,24,26,28] of 15 included trials and observed no signifcant diference between the colloid and crystalloid groups under GDFT.It is worth mentioning that our included trials reported a low rate of AKI, and the situation of GDFT is not a critical and septic population; thus, our results indicated that more research is needed to explore the efects of colloid infusion on long-term kidney function in surgical conditions.Previous publications have reported that providing between 1.75 and 2.75 L of crystalloid fuid during open abdominal surgery will not increase the total number of postoperative complications [43].However, increasing the crystalloid fuid load to 6-7 L in colon surgery increases the risk of pulmonary complications [44].Teoretically, a large amount of crystalloids is usually needed to achieve hemodynamic stability to meet GDFT.Considering that the adverse efects of crystalloid fuids are usually related to their preferential distribution to specifc interstitial areas, such as the lungs, we explored the diference in pulmonary complications between the colloid and crystalloid groups and found no signifcant diference.Te reason for this result may be that the studies included in this meta-analysis did not use such a large amount of total liquid.Finally, it is important to note that the benefts of the GDFT strategy have been demonstrated by numerous clinical studies, and relevant societies have recommended the GDFT strategy during surgery.Terefore, in the case of perioperative GDFT implementation, patients will theoretically not present hypovolemia, which means that the differences that will be caused by the total amount of fuid input and the type of fuid selected under the premise of GDFT will not be sufcient to produce clinically signifcant statistical diferences.

Limitation
Several limitations of this meta-analysis should be mentioned.First, current evidence lacks RCTs of high quality, and the sample sizes of the included studies are generally small; thus, the pooled results potentially lack credibility.Secondly, the secondary outcomes of this meta-analysis were postoperative complications, which were calculated as the frequency of complications occurring in each included study, and this counting method may resulting double counting.For example, if a patient experiences pulmonary edema and respiratory complications at the same time, we will include them as postoperative pulmonary complications twice.In addition, the complications reported in various studies were possibly explored using a diferent defnition, and the possibility of inaccurate data could not be ruled out.Last but not least, the clinical use of HES will be greatly limited due to the range of adverse reactions that HES are prone to, such as renal impairment, allergic reactions, and coagulation dysfunction.Tus, given the aforementioned limitations, the conclusion of this meta-analysis should be considered carefully.

Conclusion
Tere is no evidence of a beneft in using colloids over crystalloids under GDFT in patients undergoing noncardiac surgery, despite their use resulting in lower digestive system complications.

Figure 3 :
Figure 3: Quality of included studies by Jadad score.

Figure 4 :
Figure 4: Sensitivity analysis for primary outcome.

Figure 6 :
Figure 6: Mortality in the follow-up period.

Table 1 :
Basic information of each individual trial.

1.14.1 Length of hospital stay
Figure 13: Length of hospital stay.