The choice of colloid or crystalloid as the optimal intraoperative resuscitation fluid remains unresolved, and there is disagreement whether crystalloid or colloid is the optimal resuscitation fluid. There are inherent differences between colloids and crystalloids that may contribute to their effects (1–10). Several meta-analyses have failed to demonstrate a clear advantage in the use of colloid over crystalloid (11,12). Indeed, crystalloid resuscitation may be associated with less frequent mortality in trauma patients (13). Most clinical trials comparing colloid with crystalloid resuscitation have been inadequately powered to detect a difference in mortality as the primary end-point. A meta-analysis concluded that methodological limitations in these trials precluded any evidence-based recommendations of colloid versus crystalloid, and the numbers of patients in these studies were insufficient to detect small but potentially important clinical differences (13).
Although intraoperative and postoperative mortality is very infrequent in the elective surgical population, the incidence of morbidity during the postoperative recovery period is not. In a large surgical population undergoing a diverse group of routine, moderate-risk elective surgeries, Bennett-Guerrero et al. (14) demonstrated that the incidence of postoperative complications, defined as either in-hospital death or prolonged postoperative hospitalization (>7 days), was 27%. Almost all (98%) of these patients exhibited persistent dysfunction of at least one organ system. Gastrointestinal dysfunction, defined as the inability to tolerate an enteral diet for reasons including nausea, vomiting, and abdominal distension, was present in >50% of the cases. Given the frequency of postoperative nausea and vomiting (PONV) in that study and the association of large-volume crystalloid resuscitation with bowel edema, we tested the hypothesis that the intraoperative administration of colloids for patients undergoing moderate-risk elective noncardiac surgery is associated with a smaller incidence of PONV.
After IRB approval and written, informed patient consent, 90 ASA physical status I–III adult patients presenting for major elective general, gynecological, orthopedic, or urologic surgery with an anticipated blood loss of >500 mL were enrolled at Duke University Medical Center. Patients with the following conditions were excluded from the study: coagulopathy, significant hepatic (liver enzymes >50% upper limit of normal values) or renal (creatinine >50% upper limit of normal values) dysfunction, and congestive heart failure. Those who had received an investigational drug within the last 30 days and those with known hypersensitivity to hydroxyethyl starches were also excluded.
This was an investigator-initiated study of a previous Food and Drug Administration-guided Phase III parent trial examining the efficacy of hetastarch (HS)-normal saline (NS) versus HS-balanced salt (BS) (15). A third arm involving lactated Ringer’s solution (LR) was added to this study. Patients were randomized into one of three groups—HS-NS, HS-BS, or LR—by using a closed-envelope technique. Study investigators and research personnel were not directly involved in the care of these patients and hence were blinded to the treatment assignments. Each patient was premedicated with midazolam (1–2 mg) and fentanyl (1–2 μg/kg). Before the induction of anesthesia, an IV bolus of 7 mL/kg of LR solution was administered followed by an IV infusion of LR solution at a rate of 5 mL · kg−1 · h−1 throughout surgery. Anesthesia was induced by an IV technique involving thiopental (3–5 mg/kg), fentanyl (3–5 μg/kg), and succinylcholine (1–1.5 mg/kg) and was maintained with a balanced inhaled technique incorporating isoflurane, nitrous oxide, and oxygen, with neuromuscular blockade supplied by IV vecuronium. Additional fentanyl was titrated to effect as clinically indicated. Ventilation was adjusted to maintain Paco2 at 35–40 mm Hg, and temperature was maintained at >35.5°C throughout surgery. Patients receiving epidural analgesia for postoperative pain relief were given a 3-mL test dose of lidocaine 1.5% with 1:200,000 epinephrine during catheter placement. No intraoperative epidural drugs (local anesthetic or narcotic) were administered.
In each patient, arterial and central venous catheters were placed in addition to standard ASA monitors. Cardiovascular variables and urine output (UO) were recorded continuously during surgery. Postoperative heart rate, arterial blood pressure (BP), and UO were recorded hourly for 4 h then every 4 h until 24 h after surgery. Volumes of all types of fluids (including blood and blood products) were recorded during the perioperative period. Estimated blood loss and the volumes and doses of drugs given during the surgical procedure were recorded. If, in the judgment of the attending anesthesiologist, a tachycardia or low BP was due to light anesthesia, then the anesthetic was deepened by increasing the isoflurane concentration and giving additional opioid. If, in the judgment of the attending anesthesiologist, a tachycardia or low BP was due to hypovolemia, then intraoperative fluid administration was guided by a fluid algorithm (Fig. 1). Fluid boluses were given for a UO of <0.5 mL · kg−1 · h−1, a heart rate >110 bpm or >20% above baseline, and a systolic BP <90 mm Hg or 20% below baseline. For a central venous pressure (CVP) <15 mm Hg and hematocrit >21%, additional fluid boluses were given to maintain target hemodynamic goals. In hemodynamically unstable patients with a CVP >15 mm Hg, pulmonary artery catheterization and the use of vasoactive drugs were considered to support the circulation. For the management of surgical hemorrhage, blood was transfused when the hematocrit was ≤21% on evidence of myocardial ischemia with a hematocrit <30%. Patients were reassessed after each fluid challenge to determine whether the target hemodynamic goals were achieved. If fluid boluses did not achieve these goals, then CVP was reassessed to determine which limb of the fluid algorithm to follow.
For the treatment of nonsurgical bleeding, blood products (platelets, fresh frozen plasma, cryoprecipitate, or fibrinogen) were given for clinical evidence of bleeding that was supported by abnormal laboratory values. Coagulation measurements (obtained at baseline, at the end of surgery, 24 h after surgery, and at any time during surgery if the clinical picture warranted) triggering transfusion included platelet count <100,000/μL, prothrombin time >1.5 times control, activated partial thromboplastin time >1.5 times control, and fibrinogen <100 mg/dL. Patients were tracheally extubated either in the operating room or recovery room when the following criteria were met: adequate protective reflexes, adequate oxygenation, and stable hemodynamics.
Patients were visited daily in the postoperative period by independent research personnel unaware of the patient’s randomization until hospital discharge or death. A postoperative morbidity survey (Appendix 1) was conducted at baseline (before surgery) and daily after surgery. Specifically, the incidences of PONV, pain, and dependent or periorbital edema were assessed. Nausea, when solicited during assessment by the research personnel, was defined as the urge to vomit. It was scored with a four-point numerical scale from 0 to 3, with 0 = no nausea, 1 = mild nausea, 2 = moderate nausea, and 3 = severe nausea. Vomiting was defined as expulsion of stomach contents through the mouth. Retching was defined as an attempt to vomit, not productive of stomach contents. An emetic episode was defined as a single vomit or retch or any number of continuous vomits or retches. Droperidol 1.25 mg IV was administered if patients experienced two episodes of emesis per hour or when requested by the patient. Pain was assessed both at rest and when coughing on a three-point numerical scale from 0 to 3, with 0 = no pain, 1 = mild pain, 2 = moderate pain, and 3 = severe pain. Dependent and periorbital edema were assessed by using pressure applied by the examiner’s thumb for 15 s over the anterior aspect of the tibia proximal to the ankle, sacral prominence, and superior orbital ridge. Dependent edema required the presence of a resulting tissue indentation at both the ankle and sacral prominence.
Data were analyzed to compare all patients across the three groups on an intent-to-treat basis. An unadjusted univariate analysis using Fisher’s exact test, Wilcoxon’s two-sample ranked sum test, and logistic regression was performed to compare the three groups (HS-NS versus HS-BS versus LR) as well as between the two colloid groups and the LR group (HS-NS and HS-BS versus LR). A two-sample Wilcoxon’s ranked sum test and a univariate proportional odds model were used to study the ordinal variable nausea severity. A multivariate adjusted analysis was then performed by using logistic regression models and the proportional odds models, adjusting for the following covariates: age, ASA physical status, sex, study fluid, blood loss, and volume of LR. All tests of significance were two-sided, and a P value of <0.05 was considered statistically significant.
For estimation of sample size, the incidence of vomiting was used as a primary outcome. Pilot data revealed a baseline incidence of vomiting of 30%. Thirty patients per group were adequate to demonstrate a 50% reduction in the incidence of vomiting between two groups with α = 0.05 and β = 0.2.
Ninety patients were enrolled. One patient each from the HS-BS and HS-NS groups did not receive the study solution (one did not need colloid treatment, and the other had his surgery rescheduled). The three groups were well matched with regard to demographics and baseline characteristics (Table 1). There was no overall difference in estimated blood loss, hematocrit, or blood and blood product administration among the three groups (Table 2). Similarly there was no difference in postoperative requirements for the administration of fluids, blood, and blood products (Table 2). The volumes of study fluid (mean ± sd) administered were 1301 ± 1079 mL, 1448 ± 759 mL, and 5946 ± 1909 mL for the HS-NS, HS-BS, and LR groups, respectively (P < 0.05; HS-NS and HS-BS versus LR). In addition, the HS-NS and HS-BS groups used 3050 ± 1531 mL and 3242 ± 1308 mL of LR, respectively.
Patients who received LR had a significantly more frequent incidence and severity of nausea, vomiting, and use of rescue antiemetics, compared with the colloid groups (HS-NS and HS-BS). These patients also had more severe pain on coughing, periorbital edema, and double vision. Interestingly, the HS-NS group had a more frequent incidence of increased creatinine and oliguria compared with the HS-BS and LR groups, but this difference was not statistically significant (P = 0.65 and P = 0.15, respectively) (Table 3). The univariate logistic regression model revealed significance in all variables but dependent edema (P = 0.1604) (Table 4). The colloid group had smaller odds of nausea (P = 0.007), nausea severity (P = 0.003), emesis (P = 0.01), antiemetic use (P = 0.005), severe pain (P = 0.005), and periorbital edema and double vision (P = 0.03). The multivariate adjusted analysis revealed significance in the following: use of rescue antiemetics adjusted for study fluid (odds ratio, 0.28; 95% confidence interval, 0.11–0.07;P = 0.006).
Intraoperative fluid resuscitation with predominantly colloids appears to improve the quality of postoperative recovery compared with crystalloids. Specifically, colloid administration was associated with a less frequent incidence and severity of nausea, vomiting, and use of rescue antiemetics. Colloid-resuscitated patients also experienced less severe pain, periorbital edema, and double vision.
There has been debate since the 1960s regarding the ideal choice of fluid therapy for resuscitation. Although all sides agree that fluid resuscitation is fundamental in the management of hypovolemia, there is disagreement as to which solution should be used. Crystalloid proponents point to the hemostatic derangement, the increased incidence of adverse drug reactions, and the greater risk of fluid overload occurring with colloidal fluids (16). However, in a previous study by Gan et al. (15), HS-BS had a more favorable coagulation profile than HS-NS and was possibly a safer alternative when used in volumes up to 5 L. The colloid lobby focuses on the large volumes of crystalloid required to achieve adequate resuscitation and on the resultant tissue edema and reduction in tissue oxygen delivery (3).
Many studies have attempted to address this question in a number of different clinical settings. However, none had an adequate sample size to use mortality as an end-point for outcome (11,13,16). To demonstrate a 10% difference in treatment effect between crystalloid and colloid resuscitation (assuming a 15% baseline mortality rate and a two-tailed α of 0.05 and β of 0.2) would require a randomized clinical trial involving 5748 patients. If a 10% baseline mortality rate is assumed, then 9107 patients would be required to detect a 10% difference (13). Although there are no published 30-day perioperative mortality data in the United States, the equivalent mortality rate in the United Kingdom has been cited as 0.65% for 1999 (17).
Many colloid-crystalloid studies have included outcome variables such as pulmonary function (1,8), renal function (18), and coagulation (19,20). However, none has concentrated on gastrointestinal function. Bennett-Guerrero et al. (14) demonstrated that gastrointestinal dysfunction is the predominant cause (>50%) of prolonged hospital stay after noncardiac elective surgery. Although colloid or crystalloid use was not reported, the choice of resuscitation fluid may be one of the contributing factors to an increase in the postoperative complication rate.
Fluid composition will dictate its distribution after IV infusions. An isotonic solution of either LR or sodium chloride expands only the extracellular compartment and will increase intravascular volume by approximately one-fifth of the volume infused. Lamke and Liljedahl (21) demonstrated, in postoperative surgical patients, that 90 minutes after the infusion of 1000 mL of 6% HS, albumin, or NS, 75% of the HS and 50% of the albumin still remained in the intravascular space, whereas <20% of NS remained. The rest of the crystalloid is distributed in the interstitial compartment, and when large volumes are infused, this may result in interstitial edema.
Colloid solutions are made up of larger molecules and are maintained within the intravascular volume for a longer period, resulting in less interstitial edema. Colloid resuscitation has been associated with less endothelial swelling and less parenchymal injury in animal models compared with crystalloid (22). Prien et al. (9) demonstrated that, in patients undergoing a Whipple procedure, crystalloid resuscitation with LR resulted in a significant increase in the water content of the jejunal specimen compared with intraoperative resuscitation with HS or albumin. Intestinal edema is associated with impaired gastrointestinal function, an intolerance for enteral nutrition (23,24), an increased potential for bacterial translocation, and the development of multiple organ dysfunction syndrome (25,26). It is plausible that the decreased incidence of PONV in our colloid groups of patients was due to less bowel edema in these patients. In a study of 200 cardiac surgical patients that examined the effect of IV fluids on various clinical outcomes during surgery, Bennett-Guerrero et al. (27) found that patients administered starch-based fluids (HS-BS or HS-NS) exhibited significantly less edema, a decreased incidence of antiemetic use, and a faster return of bowel function.
An interesting finding in this study was the apparent increased frequency of oliguria and increased creatinine in the HS-NS group. This difference did not reach statistical significance. One possible explanation may be the large chloride content of this resuscitation fluid compared with the other two groups (145 versus 124 mmol/L). In the aforementioned study by Bennett-Guerrero et al. (27), all indices of renal function (serum creatinine, UO, and creatinine clearance) were inferior when NS-based solutions were used. All six patients in this cohort who required hemodialysis had been randomized to an NS-based solution. Wilcox (28), in a canine model, showed that hyperchloremia produced renal vasoconstriction and a decrease in glomerular filtration rate that was independent of renal innervation, enhanced by prior salt depletion, and related to a tubular reabsorption mechanism involving chloride.
There are several limitations to this study that should be considered along with the findings. We cannot exclude the possibility that the observed differences were due to factors other than the type of resuscitation fluid. However, it is a prospective study with an intent-to-treat design and with well defined end points. Peripheral dependent edema and orbital edema were assessed clinically. There is no objective method of quantifying the severity of edema. Weight gain has been shown to be inaccurate in this respect and is dependent on other variables. Brijker et al. (29) used a water displacement volumetry method to quantify the volume of fluid in the legs of patients with peripheral edema. However, this method is cumbersome and has not been well validated. It also cannot be used for sacral and periorbital edema. The observed differences are admittedly surrogate end points, because we did not design this study with sufficient sample size to detect a quantitative outcome difference, such as length of stay or mortality. However, the quality of recovery may be important to the well-being of patients recovering from surgery. PONV is one of the most unpleasant postoperative symptoms (30). Avoiding PONV is very important to patients, more so than avoiding postoperative pain (31), and patients have been willing to spend up to US$100 out of pocket for an effective antiemetic (32).
In conclusion, intraoperative resuscitation with colloid in patients undergoing elective, major noncardiac surgery is associated with less PONV and with a better overall postoperative recovery profile. Further studies are needed to investigate whether the use of colloid will lead to a quantitative improvement in postoperative outcomes.
1. Rackow EC, Falk JL, Fein IA, et al. Fluid resuscitation in circulatory shock: a comparison of the cardiorespiratory effects of albumin, hetastarch, and saline solutions in patients with hypovolemic and septic shock. Crit Care Med 1983; 11: 839–50.
2. Shoemaker WC, Schluter M, Hopkins JA, et al. Comparison of the relative effectiveness of colloids and crystalloids in emergency resuscitation. Am J Surg 1981; 142: 73–81.
3. Hauser CJ, Shoemaker WC, Turpin I, et al. Oxygen transport responses to colloids and crystalloids in critically ill surgical patients. Surg Gynecol Obstet 1980; 150: 811–6.
4. Hanken K, Radel C, Beez M, et al. Comparison of hydroxyethyl starch and lactated Ringer’s solution on hemodynamics and oxygen transport of critically ill patients in prospective crossover studies. Crit Care Med 1989; 17: 133–5.
5. Tranbaugh R, Elings V, Christensen J, Lewis F. Determinants of pulmonary interstitial fluid accumulation after trauma. J Trauma Inj Infect Crit Care 1982; 22: 820–6.
6. Gallagher JD, Moore RA, Kerns D, et al. Effects of colloid or crystalloid administration on pulmonary extravascular water in the postoperative period after coronary artery bypass grafting. Anesth Analg 1985; 64: 753–8.
7. Moss G, Lowe R, Jilek J, Levine H. Colloid or crystalloid in the resuscitation of hemorrhagic shock: a controlled clinical trial. Surgery 1981; 89: 434–8.
8. Lowe R, Moss G, Jilek J, Levine H. Crystalloid vs colloid in the etiology of pulmonary failure after trauma: a randomized trial in man. Surgery 1977; 81: 676–83.
9. Prien T, Backhaus N, Pelster F, et al. Effect of intraoperative fluid administration and colloid osmotic pressure on the formation of intestinal edema during gastrointestinal surgery. J C Anesth 1990; 2: 317–23.
10. Twigley AJ, Hillman KM. The end of the crystalloid era? A new approach to peri-operative fluid administration. Anaesthesia 1985; 40: 860–71.
11. Schierhout G, Roberts I. Fluid resuscitation with colloid or crystalloid solutions in critically ill patients: a systematic review of randomised trials. BMJ 1998; 316: 961–4.
12. Human albumin administration in critically ill patients: systematic review of randomised controlled trials—Cochrane Injuries Group Albumin Reviewers. BMJ 1998;317:235–40.
13. Choi PT, Yip G, Quinonez LG, et al. Crystalloids vs. colloids in fluid resuscitation: a systematic review. Crit Care Med 1999; 27: 200–10.
14. Bennett-Guerrero E, Welsby I, Dunn TJ, et al. The use of a postoperative morbidity survey to evaluate patients with prolonged hospitalization after routine, moderate-risk, elective surgery. Anesth Analg 1999; 89: 514–9.
15. Gan TJ, Bennett-Guerrero E, Phillips-Bute B, et al. Hextend, a physiologically balanced plasma expander for large volume use in major surgery: a randomized phase III clinical trial—Hextend Study Group. Anesth Analg 1999; 88: 992–8.
16. Velanovich V. Crystalloid versus colloid fluid resuscitation: a meta-analysis of mortality. Surgery 1989; 105: 65–71.
17. Callum K, Gray A, Hoile R. Then and now: the 2000 report of the national confidential enquiry into perioperative deaths. London: National Confidential Enquiry into Perioperative Deaths, 2000: 19–54.
18. Boutros AR, Ruess R, Olsen L, et al. Comparison of hemodynamic, pulmonary, and renal effects of use of three types of fluids after major surgical procedures on the abdominal aorta. Crit Care Med 1979; 7: 9–13.
19. London M, Franks M, Verrier E, et al. The safety and efficacy of ten percent pentastarch as a cardiopulmonary bypass priming solution: a randomized clinical trial. J Thorac Cardiovasc Surg 1992; 104: 284–96.
20. Martin G, Wakeling H, El-Moalem, et al. A prospective, randomized comparison of thromboelastographic coagulation profile in patients receiving lactated Ringer’s solution, 6% hetastarch in a balanced-saline vehicle, or 6% hetastarch in saline during major surgery. J Thorac Vasc Anesth 2002;16:441–6.
21. Lamke LO, Liljedahl SO. Plasma volume changes after infusion of various plasma expanders. Resuscitation 1976; 5: 93–102.
22. Morisaki H, Bloos F, Keys J, et al. Compared with crystalloid, colloid therapy slows progression of extrapulmonary tissue injury in septic sheep. J Appl Physiol 1994; 77: 1507–18.
23. Falk JL. Fluid resuscitation and colloid-crystalloid controversy: new thoughts on an old debate. Crit Care Med 1991; 19: 451–3.
24. Moss G. Plasma albumin and postoperative ileus. Surg Forum 1967; 18: 333–6.
25. Baker JW, Deitch ED, Ma LM, et al. Hemorrhagic shock induces bacterial translocation from the gut. J Trauma 1988; 28: 896–906.
26. Wilmore DW, Smith RJ, O’Dwyer ST, et al. The gut: a central organ following surgical stress. Surgery 1988; 104: 917–23.
27. Bennett-Guerrero E, Frumento RJ, Mets B. Impact of normal saline based versus balanced-salt intravenous fluid replacement on clinical outcomes: a randomized blinded clinical trial [abstract]. Anesthesiology 2001; 95: A147.
28. Wilcox CS. Regulation of renal blood flow by plasma chloride. J Clin Invest 1983; 71: 726–35.
29. Brijker F, Heijdra Y, Elshout FVD, et al. Volumetric measurements of peripheral edema in clinical conditions. Clin Physiol 2000; 20: 56–61.
30. Gan T. Postoperative nausea and vomiting: can it be eliminated? JAMA 2002; 287: 1233–6.
31. Macario A, Glenn D, Dexter F. What can the postanesthesia care unit manager do to decrease costs in the postanesthesia care unit? J Perianesth Nurs 1999; 14: 284–93.
32. Gan T, Sloan F, Dear G, et al. How much are patients willing to pay to avoid postoperative nausea and vomiting? Anesth Analg 2001; 92: 393–400.
© 2003 International Anesthesia Research Society
33. Gan TJ, Soppitt A, Maroof M, et al. Goal-directed intraoperative fluid administration reduces length of hospital stay after major surgery. Anesthesiology 2002; 97: 820–6.