Fluid management of patients undergoing abdominal surgery – more questions than answers^*

Abstract

 

The ‘wet vs. dry’ philosophy in patients undergoing abdominal surgery is a subject of substantial debate. It has been suggested that restricting fluid input would significantly reduce complications and improve outcome following abdominal surgery. Keeping the patients dry may be a two-edged sword because the resulting hypovolaemia may result in compromised organ perfusion and poor tissue oxygenation. A review of the literature from 1990 to 2004 revealed that only very few studies on this subject have been published. Unfortunately, most of the ‘dry’-supporting studies used fixed amounts of volume instead of a fluid concept adapted to the patients' need (‘goal-directed’) and there is no generally accepted definition of ‘restricted’, ‘dry’ or ‘overload’. Not only the amount but also the kind of administered fluid appears to be important. Current evidence indicates that using crystalloids exclusively may cause overloading of the interstitial compartment with considerable negative sequelae, whereas using colloids may improve microperfusion and tissue oxygenation. This review shows that the meagre literature on a restricted volume replacement strategy in abdominal surgery patients cannot clearly support the ‘dry’ approach. Further well-performed studies are necessary to elucidate the ideal amount and type of fluid replacement and determine how to guide fluid therapy.

Introduction

Volume and fluid deficits are extremely common in surgical patients. Preoperative depletion, intraoperative fluid loss and bleeding may cause absolute fluid and volume deficits, while vasodilatation mediated by vasodilating substances (e.g. anaesthetics) or during rewarming it can produce relative volume deficits. Hypovolaemia may also develop in the absence of obvious fluid loss secondary to generalized impairment of the endothelial barrier for example, during inflammation resulting in diffuse capillary leakage and shifting of fluid from the intravascular to the interstitial compartment.

Aside from the debate on the ideal type of fluid replacement, a new controversy has arisen: Is fluid and volume restriction of advantage in the management of patients undergoing major abdominal surgery? After decades of infusing considerable amounts of fluids in these patients [^1,2], new restrictive fluid concepts have been developed to keep the patient ‘dry’ which are thought to reduce the rate of postoperative complications or even improve outcome [^3,4,5]. Statements such as ‘there are many reasons that fluid overloading has become the norm’ and ‘the abdominal compartment syndrome is the price for this aggressive fluid replacement’ [⁶] resulted in recommendations to restrict fluid input in these patients. The study results on restrictive fluid management regimens in abdominal surgery patients, however, are not uniform. In the studies focusing on the ‘dry vs. wet’ concept, the volume of infused fluids, the type of infused fluids and the endpoints for assessing the benefit of a certain fluid management strategy differ widely. Thus, this review was focused on the principles of volume and fluid replacement and on the literature dealing with ‘dry vs. wet’ fluid management in patients undergoing abdominal surgery.

Consequences of surgery and hypovolaemia

Adequate volume therapy appears to be a cornerstone of managing the surgical patient. In a prospective review of 111 consecutive patients who died in hospital after admission for treatment of injuries, the most common deficits in patient management were related to inadequate fluid resuscitation [⁷] indicating that an adequate volume replacement therapy may help to improve organ function and reduce patient morbidity or even mortality.

The pathophysiological consequences of major surgery have been well described [⁸]. Major surgery is associated with a stress response of combined endocrine and inflammatory origin. Hypovolaemia may be associated with flow alterations that are inadequate to fulfill the nutritive role of the circulation [⁹]. Many of the manifestations of organ dysfunction result from peripheral, microcirculatory derangements [^10,11]. Stable systemic haemodynamics do not guarantee that organ and tissue perfusion are maintained as well. During low output the organism attempts to compensate perfusion deficits by redistribution of flow to vital organs (e.g. heart, brain) resulting in an underperfusion of other organs (e.g. splanchnic bed, kidneys). Activation of the sympathetic nervous system (SNS) or the renin–angiotensin system (RAS) are compensatory mechanisms to maintain peripheral perfusion. This compensatory neurohormonal activation is beneficial at first, but may become harmful and be involved in the poor outcome of the hypovolaemic surgical patient even though initial volume resuscitation was successful. Two aspects may contribute to microcirculatory deterioration in this situation [^12,13,14]: interaction between endothelium and cellular elements of the blood and endothelial swelling. Swelling of the endothelium of the capillaries may be caused by increased permeability (‘capillary leakage’). Poor capillary perfusion increases the risk of impaired oxidative killing in the wound and additional liberation/activation of mediators that subsequently promote cell adhesion and vasoconstriction [¹⁵]. Microcirculatory impairment initiates a vicious cycle of progressive tissue damage that may subsequently lead to development of organ dysfunction [^16,17].

Principles of fluid replacement

Administered fluids may remain in the intravascular compartment or may equilibrate with the interstitial/intracellular fluid compartments. Blood pressure, intravascular and interstitial volumes, and the composition of each body compartment are controlled by variety of mechanisms (e.g. the atrial-natriuretic peptide (ANP) system, the renin–aldosterone–angiotensin system (RAAS), SNS, the vasopressin system). The principal action of these neurohormonal systems is to retain water and sodium to correct intravascular volume losses and to increase perfusion pressure through vasoconstriction. The activity of these systems is enhanced in stress situations (e.g. during major surgery and during hypovolaemia). Administration of a restricted amount of crystalloids possibly replace water deficits, but replacement of a intravascular volume deficit would require much more volume to inhibit the secretory stimulus of the hormones. One can thus expect that volume replacement with crystalloids alone will not inhibit the normal response of ANP and RAA, whereas administration of a combination of crystalloids and colloids may achieve this goal.

The primary goal of fluid administration is to guarantee stable systemic haemodynamics by rapidly restoring circulating plasma volume while avoiding excessive fluid accumulation in the interstitial space. Crystalloids are more likely to leave the intravascular compartment than colloids, and greater volumes of crystalloids are therefore necessary to replenish volume deficits (approximately 3 to 4 times the volume of colloids) [^18–21]. Because most of the infused crystalloid solution is distributed into the interstitial space, exclusively administering crystalloids is associated with the risk of oedema formation [^20,21]. Moreover, infusing large amounts of saline solution leads to hyperchloremic acidosis [²²] with considerable negative effects on patients comfort (e.g. postoperative nausea and vomiting) [²³].

What does the literature tell us?

PubMed was searched for original articles published in English during the past 15 years (1990–2004) on perioperative fluid replacement in patients undergoing abdominal surgery. The search string was ([volume/fluid replacement or therapy] [fluid/volume restriction] [abdominal surgery] [intra-/postoperative period] [fluid/volume overload] [plasma substitutes] [colloids] [crystalloids]). The articles were not assessed with regard to quality (‘power’) of the study and all studies meeting all the criteria are evaluated. Due to the very limited number of papers on this topic, no evidence-based approach was used and no meta-analysis was performed.

I. Fluid minimization is beneficial (Table 1)

Table 1a: Fluid restriction is beneficial.

Table 1b: Fluid restriction is not beneficial.

• In one of the landmark studies, the effects of a restricted intravenous fluid regimen vs. a standard regimen on complications after colorectal resection were assessed [³]. In this randomized, observer-blinded, multicentre trial in 141 patients undergoing abdominal surgery, a ‘standard group’ received preloading with a hydroxethylstarch (HES) preparation, extra HES infusions when necessary, and saline solution to replace third space losses, whereas in a ‘restricted group’ these replacements were omitted. The ‘restricted group’ received a median of HES 6% 2740 mL and glucose 5% (range: 1100–8050 mL) on the day of operation vs. HES 6% 5388 mL and normal saline (range: 2700– 11083 mL) in the ‘standard group’. Looking at the range of administered volume, it is obvious that the groups overlapped: 15% of the ‘restricted group’ received more volume, while 24% of the ‘standard group’ received less fluid than prescribed by the study protocol. The total number of complications (including pneumothorax, cystitis, headache) was significantly different between the two groups, favouring the ‘restricted group’. The main differences in major complications consisted of non-surgical, medical complications. The major problems with the study are that a fixed volume of fluids was given, no goal-directed volume therapy was performed and the type of fluid differed between the two groups (HES, glucose, normal saline). No haemodynamic data, such as central venous pressure (CVP) or cardiac output (CO) were shown. Some of the patients appear to be simply fluid overloaded because a fixed volume replacement protocol instead of a monitored (‘patient-adapted’) volume replacement strategy was used.
• In another study in patients scheduled for elective colon resection surgery, a restricted postoperative volume infusion strategy (n = 10; less than 2 L of water and 77 mmol of sodium per day) resulted in more rapid return of gastrointestinal function (first passage of flatus and faeces) than patients in whom a more liberal fluid regime was employed (n = 10; more than 3 L of water and 154 mmol sodium per day) [⁴]. All fluids were given in fixed doses. A positive salt and water balance sufficient to cause a 3 kg weight gain after surgery even prolonged hospital stay by an average of 3 days in these patients. Only crystalloid solutions were infused in this study, and colloid osmotic pressure (COP) was expected to be reduced to a greater degree in the ‘liberal fluid regime’ group resulting in an overload of the interstitial compartment. The total number of patients in this study (n = 10 in each group) was definitely too small to draw conclusions regarding outcome.
• The impact of two intraoperative fluid regimens on postoperative outcome was prospectively evaluated in 152 ASA I–III patients undergoing elective intraabdominal surgery [⁵]. Either a liberal (n = 75; bolus of 10 mL kg⁻¹ lactated Ringer's solution (RL) followed by 12 mL kg⁻¹ h⁻¹ of RL) or a restrictive intraoperative fluid strategy (n = 77; 4 mL kg⁻¹ h⁻¹ RL) was implemented. One-third of the patients in the restrictive group required an additional fluid bolus of 1500 mL to stabilize haemodynamics; significantly more than in the ‘liberal’ group. The patients in the liberal group received an average total volume of 3670 mL (range 1880–8800 mL), while those in the restrictive group received 1230 mL (range 490–7810 mL). This suggests that large volumes of fluids were also administered to the patients in the restrictive group. The number of patients with complications of any type was significantly lower in the restrictive volume group without, however, showing significant differences in the individual complications (e.g. wound dehiscence/infection n = 11 in the liberal vs. n = 7 in the restrictive group). The patients of the liberal volume group passed flatus and faeces significantly later and their postoperative hospital stay was significantly longer (mean: 9 days) than in the restrictive group (mean: 8 days).
• A retrospective study in patients undergoing transthoracic oesophagectomy for carcinoma (n = 56) showed that fluid replacement strategy has changed over the years [²⁴]. With a more restrictive fluid approach that was initiated between 1998 and 2000, the patients had a postoperative fluid balance of 749 ± 697 mL, whereas the patients in the period from 1997 to 1998, in whom a more liberal fluid replacement strategy with crystalloids, colloids and blood was allowed, had a postoperative fluid balance of 2386 ± 1307 mL. There were fewer pulmonary complications and the hospital stay was shorter in the restrictive fluid group. However, the value of such retrospective analyses using mixed volume replacement strategies and changing conditions must be questioned.

II. Fluid minimization is not necessarily beneficial (Table 1)

• In patients undergoing major abdominal surgery involving intestinal resection, it became apparent that the administration of 10 mL kg⁻¹ h⁻¹ of crystalloids (including glucose) was not sufficient to maintain cardiovascular stability and urine output, whereas 15 mL kg⁻¹ h⁻¹ was adequate [²⁵].
• In patients undergoing colon resection surgery, the hypothesis was tested whether supplemental fluid administration during and after elective colon resection increased tissue perfusion and tissue oxygen tension [²⁶]. ‘Conservative fluid management’ consisted of 8–10 mL kg⁻¹ h⁻¹ of unspecified crystalloid solutions, an ‘aggressive fluid management’ consisted of a bolus of 10 mL kg⁻¹ of unspecified crystalloid fluids given prior to surgery and 16–18 mL kg⁻¹ h⁻¹ given intraoperatively and for the first postoperative hour. Subcutaneous oxygen tension was measured in the patient's upper arm using small implantable polarographic tissue oxygen sensors. Supplemental perioperative fluid administration significantly increased tissue perfusion and tissue oxygen partial pressure. Improved tissue perfusion and tissue oxygenation seem to be of benefit for the patient, especially in those who are at increased risk of reduced tissue perfusion (e.g. in patients with diabetes mellitus) [^19,27,28]. The authors concluded that even moderate hypovolaemia, which is occasionally difficult to recognize clinically, may be associated with impaired blood flow at the microcirculatory level.

III. Is the choice of plasma substitute of importance in the ‘dry vs. wet’ debate? (Table 2)

Table 2: The kind of fluid for managing the abdominal surgery patient is important.

Fluid overload has been reported to lead to oedema formation [^1,29] and is thus suspected to be associated with complications. This statement per se may lead to the wrong consequences because the type of fluid appears to play a fundamental role in this situation. In surgical patients, four litres of a crystalloid solution given to stabilize systemic haemodynamics can potentially overload the interstitial compartment leading to interstitial oedema, whereas one litre of a colloidal solution is unlikely to do so.

• In patients undergoing major abdominal surgery, the exclusive intraoperative use of crystalloids (Ringer lactate, mean volume 3850 mL) resulted in intestinal interstitial fluid accumulation that was absent in patients, in whom a colloid (e.g. HES 10% 200/0.5, mean volume: 1358 mL) had been administered [³⁰].
• Intravascular volume replacement with a third-generation medium molecular weight HES preparation (HES 6% 130/0.4, n = 21) improved tissue oxygenation in patients undergoing major abdominal surgery (esophagectomy, complex colon surgery) [³¹]. Tissue oxygenation (ptO₂) was measured over 24 h in the patients' forearm muscle using small implantable polarographic tissue oxygen sensors. ptO₂ increased significantly in HES-treated patients, while the administration of RL (n = 21) was associated with a significant decrease in tissue oxygen tensions despite systemic haemodynamics (mean arterial pressure, HR, CVP) being similar in the two groups throughout the study period. Whether muscle ptO₂ is representative for tissue oxygenation in the gut (e.g. at the anastomosis) remains to be elucidated.
• In middle-aged [³²] as well as elderly patients [³³] undergoing major abdominal surgery, markers of inflammation (e.g. IL-6 and IL-8 plasma levels) and of endothelial injury or activation (plasma levels of adhesion molecules) were prospectively studied in two randomized groups: one group exclusively received RL (n = 21), whereas the other group received, in addition, a modern HES preparation (HES 6% 130/0.4, n = 21) to maintain CVP between 8 and 12 mmHg for 24 h. The degree of inflammation and endothelial activation were significantly higher in the crystalloid group indicating that the use of HES attenuated the inflammatory response.
• In a prospective, blinded study in patients undergoing major, elective, non-cardiac surgery including abdominal surgery, the effects of either a crystalloid or a colloid-based volume replacement strategy was assessed with regard to nausea and vomiting as well as the postoperative patient recovery profile [³⁴]. The patients received either RL (n = 30, 5946 ± 1909 mL) or a HES preparation (HES 6% 450/0.7, n = 30; 1301 ± 1079 mL). The patients in the colloid group had a significantly lower incidence of nausea and vomiting, less frequent use of rescue antiemetics, less severe pain, less periorbital oedema, and less double vision. Compared with a crystalloid-based volume replacement regimen, intraoperative fluid resuscitation using a colloid was judged to improve the quality of postoperative recovery.

How is fluid therapy guided? (Table 3)

Table 3: Monitored directed therapy is important.

One fundamental problem in this issue is how to define adequacy of fluid replacement therapy (i.e. what is inadequate, adequate or excessive). Early estimates for replacement with balanced salt solution ranged from 0 to 67 mL kg⁻¹ h⁻¹ [²]. More recent recommendations for intraoperative fluid administration ranged from 5 to 15 mL kg⁻¹ h⁻¹ without specifying the type of fluid [^1,35,36]. Most institutions have defined their individual protocols for fluid therapy and definitions such as ‘standard’, ‘aggressive’, ‘restrictive’ or ‘liberal’ fluid regimen are far from being clearly defined or generally accepted.

Hypovolaemia is sometimes difficult to recognize clinically or by conventional monitoring techniques (e.g. central venous or arterial pressure). Monitoring protocols were used only in few studies. Some studies adapted fluid replacement according to haemodynamics, for example, to maintain filling pressures (CVP, pulmonary capillary wedge pressure (PCWP)) or to keep systolic blood pressure over 100 mmHg. Other studies used fixed doses without adapting the volume load to systemic haemodynamics or other surrogates of the patient's volume status. This concept is associated with the risk of over- or underloading the patient, because the patient's individual needs and history (e.g. co-morbidities, prolonged starvation) are not taken into account.

Guiding fluid replacement is still an unsolved problem as most of the monitoring methods (CVP, PCWP) are of limited value in assessing the actual volume status. Monitoring blood flow may probably be more helpful than measuring filling pressures in this situation. Hypovolaemia-related alterations in tissue perfusion and oxygenation may be associated with inadequate wound healing, are likely to contribute to the development of organ failure and may increase morbidity or mortality [^9,13,37]. Increasing tissue oxygen tension may have a clinical impact in terms of improved wound healing and fewer infectious complications [^26,37,38], and assessing the adequacy of oxygen supply to organs and tissues thus appears to be essential. Monitoring of tissue oxygenation and organ function in the clinical setting is largely based on measuring traditional variables of resuscitation, such as global (systemic) haemodynamics, pulse oximetry, capillary refilling, urine output or indirect biochemical markers. These parameters are insensitive indicators of hypoxia and are considered to be poor surrogates for oxygen availability at the tissue level, since tissue oxygenation is determined by the net balance between cellular oxygen supply and oxygen demand. The fact that regional tissue hypoxia can persist despite the presence of an apparently adequate systemic blood flow, pressure and arterial oxygen content highlights the need for more specific indices of oxygenation at tissue level to assess the ideal amount and type of fluid replacement.

The value of a monitored fluid replacement strategy (‘goal-directed’ therapy) for the ‘dry vs. wet’ debate has only been shown in few studies in abdominal surgery patients:

• In 100 patients scheduled for general, urological or gynecological abdominal surgery with an anticipated blood loss greater than 500 mL, a control group (n = 50) receiving standard intraoperative care was compared with a protocol group (n = 50), in whom additional volume replacement with HES 6% 450/0.7 was guided by oesophageal Doppler monitoring to maintain maximal stroke volume [³⁹]. The group with ‘goal-directed’ intraoperative fluid administration received more volume (847 ± 373 mL of HES and 4405 + 2650 mL of RL) than the control group (282 ± 470 mL of HES and 4375 ± 2452 mL of RL). However, they had an earlier return of bowel function, a lower incidence of postoperative nausea and vomiting, and a shorter postoperative hospital stay. This suggests that a larger fluid replacement volume is not necessarily associated with a detrimental outcome, as long as the fluid replacement is indicated to meet the patients' needs.
• During and after elective colon resection, supplemental fluid administration was guided by monitoring ptO₂ [²⁶]. Although systemic haemodynamics were almost similar, muscle tissue oxygenation (measured in the patient's upper arm) was significantly better in a group receiving ‘aggressive fluid management’ than in a group of patients in whom a ‘conventional fluid management’ was performed. Whether tissue oxygenation in the ‘aggressive fluid management’ group was higher also in the gut or at the intestinal anastomosis cannot be concluded from this study.
• In patients undergoing major bowel surgery, 57 patients were randomly assigned to Doppler-guided volume therapy or a control group without additional monitoring [⁴⁰]. The Doppler-guided patients received significantly more intraoperative colloids than the control patients (mean 28 ± 16 mL kg⁻¹ vs. 19.4 ± 14.7 mL kg⁻¹). CO increased significantly in the Doppler-guided group while that of the controls remained unchanged. Five control patients required postoperative critical care admission. The authors concluded that fluid titration using oesophageal Doppler during bowel surgery can not only improve systemic haemodynamics but may also reduce critical care admissions postoperatively.
• In a retrospective study in patients undergoing orthotopic liver transplantation, two different regimens of intraoperative fluid management were studied [⁴¹]: one centre used normal CVP values as a surrogate for normovolaemia (n = 78), while the other centre aimed for a low CVP (<5 mmHg) and minimized fluid administration (n = 73). The ‘normal CVP’ group had a higher rate of transfusions, whereas a significantly increased incidence of postoperative renal failure (elevated creatinine levels and more frequent need for haemodialysis) was found in the ‘low CVP’ group.

Conclusions

The philosophy of restricting fluid administration in the abdominal surgical patient has aroused enormous interest. Intravenous fluid overload has been shown to be associated with negative sequelae including delayed recovery of gastrointestinal function or even poor outcome. By contrast, it has also been demonstrated that the underresuscitated patient undergoing major abdominal surgery may suffer from malperfused organs and poor tissue oxygenation resulting in detrimental consequences for organ function and possibly also for the patients' outcome. Reviewing the current literature, this debate remains enormously controversial:

• One fundamental problem with the studies dealing with the ideal fluid management in abdominal surgery is that different endpoints were used. Interest was focused on gastrointestinal function, hospital stay, total number of all complications (also including pneumothorax), wound infection and other problems. This makes comparison of the studies difficult.
• Most studies assessing the value of a restrictive volume approach used institutionally fixed limits of the administered volume and compared it with a fixed ‘standard (high)’ volume infusion regimen, instead of adapting fluid therapy to the patients' needs.
• A core problem in this area is the question of which monitoring is the best surrogate for early identification of volume overload or inadequate volume therapy resulting in poor tissue oxygenation or perfusion deficits. The microcirculation represents the common final part of the circulatory system. During major surgery, microcirculatory flow is altered and often become inadequate [³⁸]: hypoperfused regions suffer from low oxygen tension that does not support adequate wound healing. Inadequate perfusion may occur in spite of ‘normal’ systemic (macro-) circulation, and it is significantly aggravated when cellular dysfunction increases the hydraulic resistance of the capillaries. Unfortunately, none of the available techniques for assessing tissue perfusion and oxygenation can be recommended at present for routine use in the abdominal surgery patient. We are still awaiting a simple method that is easy to use and gives sensitive data on tissue perfusion and oxygenation in important organs (e.g. the intestine at the anastomosis). At present, combined information on haemodynamics (BP, pulse pressure variations, CO), filling pressures (e.g. CVP), urine output, and arterial and central venous blood gases (e.g. base–acid status) may help to assess the patient's volume status. Fixed doses of fluid replacement should be avoided because this approach risks over- or underloading the patient.
• Overloading the patient should definitely be avoided – underloading, however, appears to be dangerous as well. No definite conclusion on the value of a restrictive fluid replacement strategy compared to a more liberal strategy can be drawn. In a review of randomized, controlled clinical trials from 1966 to 2001, outcome was evaluated in patients in whom fluid vs. no fluid was administered preoperatively to correct preoperative fluid deficits [⁴²]. In nine studies, fluid administration was less than one litre, and in eight studies it was equal to or more than one litre. Whether a moderate or a more aggressive fluid replacement therapy prior to surgery influenced patients' comfort and morbidity could not be decided definitely by this analysis.
• The type of administered volume has been mostly neglected in the ‘dry vs. wet’ controversy. Patients in whom several litres of crystalloids are infused according to a fixed fluid replacement strategy are at risk of developing interstitial oedema with negative consequences for tissue perfusion, oxygenation, and organ function. This can be easily avoided by using much smaller amounts of a colloid to guarantee sufficient systemic haemodynamics and tissue perfusion.

Summarizing the available meagre information on the ‘dry vs. wet’ debate in patients undergoing major abdominal surgery, an overload with crystalloid solutions should definitely be avoided. A general recommendation for the ‘dry side of life’, however, is far too early. Merits and demerits of keeping the patient more ‘dry’ have to be carefully analysed. In a postal questionnaire survey that was sent to 1091 Fellows of the Association of Surgeons of Great Britain and Ireland, consultant surgeons felt that the present practice of perioperative fluid management was unsatisfactory [⁴³]. Only 30% felt that patients were receiving appropriate amounts of water, sodium and potassium. This shows the need for further research in this area. Future studies are urgently needed, including a ‘goal-directed’ fluid replacement approach and enlarging the ‘dry vs. wet’ debate to a ‘what kind of volume is best’ for the patient undergoing major abdominal surgery.