Clinical review: Goal-directed therapy-what is the evidence in surgical patients? The effect on different risk groups

Patients with limited cardiac reserve are less likely to survive and develop more complications following major surgery. By augmenting oxygen delivery index (DO2I) with a combination of intravenous fluids and inotropes (goal directed therapy (GDT)), postoperative mortality and morbidity of high-risk patients may be reduced. However, although most studies suggest that GDT may improve outcome in high-risk surgical patients, it is still not widely practiced. We set out to test the hypothesis that GDT results in greatest benefit in terms of mortality and morbidity in patients with the highest risk of mortality and have undertaken a systematic review of the current literature to see if this is correct. We performed a systematic search of Medline, Embase and CENTRAL databases for randomized controlled trials (RCTs) and reviews of GDT in surgical patients. To minimize heterogeneity we excluded studies involving cardiac, trauma, and paediatric surgery. Extremely high risk, high risk and intermediate risks of mortality were defined as >20%, 5 to 20% and <5% mortality rates in the control arms of the trials, respectively. Meta analyses were performed and Forest plots drawn using RevMan software. Data are presented as odd ratios (OR; 95% confidence intervals (CI), and P-values). A total of 32 RCTs including 2,808 patients were reviewed. All studies reported mortality. Five studies (including 300 patients) were excluded from assessment of complication rates as the number of patients with complications was not reported. The mortality benefit of GDT was confined to the extremely high-risk group (OR = 0.20, 95% CI 0.09 to 0.41; P < 0.0001). Complication rates were reduced in all subgroups (OR = 0.45, 95% CI 0.34 to 0.60; P < 0.00001). The morbidity benefit was greatest amongst patients in the extremely high-risk subgroup (OR = 0.27, 95% CI 0.15 to 0.51; P < 0.0001), followed by the intermediate risk subgroup (OR = 0.43, 95% CI 0.27 to 0.67; P = 0.0002), and the high-risk subgroup (OR 0.56, 95% CI 0.36 to 0.89; P = 0.01). Despite heterogeneity in trial quality and design, we found GDT to be beneficial in all high-risk patients undergoing major surgery. The mortality benefit of GDT was confined to the subgroup of patients at extremely high risk of death. The reduction of complication rates was seen across all subgroups of GDT patients.

practice; and pulmonary artery catheters (PACs) are used in many of the clinical trials but have been largely superseded by less invasive haemodynamic monitors. A recent meta-analysis has demonstrated that although studies prior to 2000 demonstrate a benefi t in mortality, studies con ducted after 2000 demonstrate a signifi cant reduction in complication rates [5]. Furthermore, the reduction in complication rates is signifi cant regardless of the type of haemodynamic monitor used.
We hypothesized that the benefi ts of GDT are greater in patients who are at higher risk of mortality. We defi ned risk by the mortality rate of the study population undergoing major surgery. We conducted this metaanalysis to determine if GDT in high-risk surgical patients undergoing major non-cardiac surgery improves postoperative mortality and morbidity, and if this was aff ected by the mortality risk among the population studied.

Eligibility criteria
We reported only randomized controlled trials, that reported morbidity (complications) and mortality as primary or secondary outcomes. GDT was defi ned as the term encompassing the use of haemodynamic monitoring and therapies aimed at manipulating haemodynamics during the perioperative period to achieve a predetermined haemodynamic endpoint(s). Studies with GDT started pre-emptively in the perioperative period (24 hours before, intraoperative or immediately after surgery) were included. Th e GDT must have an explicit protocol, defi ned as detailed step-by-step instructions for the clinician based on patient-specifi c haemodynamic data obtained from a haemodynamic monitor or surrogates (for example, lactate, oxygen extraction ratio), and predefi ned interventions carried out by the clinician in an attempt to achieve the goal(s). Interventions included fl uid administration alone or fl uids and inotropes together. As the use of inotropic agents was aimed at a specifi c haemodynamic goal(s) and titrated accordingly, fi xed dose studies of inotropes were excluded. Only studies involving adult general surgical populations were included, and studies involving cardiac, trauma and paediatric surgery were excluded.

Information sources
A systematic literature search of MEDLINE (via Ovid), EMBASE (via Ovid) and the Cochrane Controlled Clinical trials register (CENTRAL, issue 4 of 2012) was conducted to identify suitable studies. Only articles written in English were considered. Date restrictions were not applied to the CENTRAL and MEDLINE searches. EMBASE was restricted to the years 2009 to 2012 [6]. Th e last search update was in April 2012.

Search strategy
We included the following search terms: goal-directed therapy, optimization, haemodynamic, goal oriented, goal targeted, cardiac output, cardiac index, oxygen delivery, oxygen consumption, cardiac volume, stroke volume, fl uid therapy, fl uid loading, fl uid administration, optimization, supranormal, lactate and extraction ratio. Search terms were entered into the electronic databases using search strategy methods validated by the Cochrane collaboration (see Box 1 for search strategies used) [7]. In addition to searching electronic databases, previous review articles on the subject were hand-searched for further references.

Methodological quality of included studies
Methodological quality of included studies was assessed using criteria described by Jadad and colleagues [8]. Th e Jadad scale analyzes methods used for random assignment, blinding and fl ow of patients in clinical trials. Th e range of possible scores is 0 (lowest quality) to 5 (highest quality). Studies were not excluded based on Jadad scores.

Analysis of outcomes
Th ree investigators independently screened both the titles and abstracts to exclude non-pertinent studies. Relevant full text articles were then retrieved and analysed for eligibility against the pre-defi ned inclusion criteria. Information from selected studies was extracted using a standardized data collection form. Data were collected independently by three diff erent investigators (GA, NA and CC) and discrepancies resolved by a fourth author (MC).
Hospital mortality was reported in all the included articles and was the primary outcome of our study. Morbidity, expressed as number of patients with complications, was the secondary outcome. Mortality risk groups were based on the defi nition of the high-risk surgical patient by Boyd and Jackson, such that patients whose risk of mortality was 5 to 19% and ≥20% were classifi ed as high-risk and extremely high-risk, respectively [9]. We therefore performed subgroup analyses based on the control group mortality in each study. We created three subgroups based on the mortality rate of the control group. Mortality rates of 0 to 4.9%, 5 to 19.9%, and ≥20% were considered intermediate, high risk, and extremely high risk, respectively. Mortality and complica tions were analyzed according to the above subgroups. Studies were also analyzed according to the type of monitor used, type of interventions, the therapeutic goals, and the use of 'supranormal' physiological goals.

Statistical analysis
Dichotomous data outcomes were analysed using the Mantel-Haenszel random eff ects model and results presented as an odds ratio (OR) with 95% confi dence intervals (CI). Th e meta-analysis was carried out using review manager ('Revman') for MAC (version 5.1, Cochrane collaboration, Oxford, UK). Statistical heterogeneity was assessed using the I 2 methodology. When an I 2 value of >50% was present heterogeneity and inconsistency were considered signifi cant, and when it was >75% these were considered highly signifi cant [10]. All P-values were two-tailed and considered statistically signifi cant if <0.05.

Description of studies
A total of 32 studies were included in the meta-analysis (Table 1)  . Th ese 32 studies included a total of 2,808 patients, 1,438 in the GDT arm and 1,370 in the control treatment arm. Five studies included patients who were considered extremely high risk, 12 included patients who were high risk, and 15 included patients who were intermediate risk. Th e intermediate-risk, highrisk, and extremely high-risk mortality subgroups included 1,569, 924, and 315 patients, respectively. Th ere were similar numbers of patients in the GDT and control arms. Twenty studies initiated GDT at start of surgery, whilst the other studies initiated GDT before or immediately after surgery.

Mortality
Th ree studies did not report any deaths in the control or intervention group. All 32 studies included mortality rates ( Figure 2). Although there was an overall benefi t on mortality (OR 0.52, 95% CI 0.36 to 0.74; P = 0.003), subgroup analyses revealed that mortality benefi t was seen only in studies that included extremely high risk patients (OR 0.20, 95% CI 0.09 to 0.41; P < 0.0001) but not for the intermediate-risk patients (OR 0.83, 95% CI 0.41 to 1.69; P = 0.62). Th ere was a trend towards a reduction in mortality in the high risk group (OR 0.65, 95% CI 0.39 to 1.07; P = 0.09; Figure 2). Further subgroup analyses of mortality as an endpoint revealed that mortality was reduced in the studies using a pulmonary artery catheter (OR 0.3, 95% CI 0.15 to 0.60; P = 0.0007), fl uids and inotropes as opposed to fl uids alone (OR 0.41, 95% CI 0.23 to 0.73; P = 0.002), cardiac index or oxygen delivery index as a goal (OR 0.36, 95% CI 0.21 to 0.36; P = 0.0003), and a supranormal resuscitation target (OR 0.27, 95% CI 0.15 to 0.47; P < 0.00001) ( Table 2).

Morbidity
Twenty-seven studies (including 2,477 patients) reported the number of patients with postoperative complications. Meta-analysis of these studies revealed an overall signifi cant reduction in complication rates (OR 0.45, 95% CI 0.34 to 0.60; P < 0.00001; Figure 3). Consistent with the mortality benefi ts, the reduction in morbidity was greatest in the extremely high-risk group (OR 0.27, 95% CI 0.15 to 0.51; P < 0.0001). However, there was also a signifi cant morbidity benefi t in the intermediate risk group (OR 0.43, 95% CI 0.27 to 0.67; P = 0.0002) and the high-risk groups (OR 0.56, 95% CI 0.36 to 0.89; P = 0.01) ( Figure 3). Th e reduction in the number of patients suff ering postoperative complications was seen across all subgroups, apart from studies that did not use the oxygen delivery index (DO 2 I; ml/minute/m 2 ), the cardiac index (CI; ml/minute/m 2 ), stroke volume (SV; ml), or corrected fl ow time (FTc) as a goal (OR 0.48, 95% CI 0.22 to 1.04; P = 0.06), although this approached statistical signifi cance (Table 3).

Discussion
We believe that GDT in high-risk surgical patients is likely to have the greatest benefi t if carried out early, in the right patient cohort and with a clearly defi ned protocol. We performed this meta-analysis to test the hypothesis that patients with the highest perioperative risk gain the greatest benefi ts from GDT. Studies without clearly defi ned GDT protocols and studies that initiated GDT late in the postoperative course were therefore excluded from our meta-analysis. Studies were stratifi ed into diff erent risk groups based on the mortality rate of the control group in the study. Heterogeneity in the year of study, patient demographics, type and urgency of surgery, and health care facilities among the diff erent studies are likely to account for the diff erence in mortality rates.
A reduction in mortality associated with GDT was seen only in the extremely high-risk group of patients (baseline mortality rate of >20%). A baseline mortality rate of >20% is unusual in current practice [4,96]; in this sense it is interesting to note that two of fi ve studies with a baseline mortality rate of >20% were carried out within the past decade. Neither of these studies demonstrated a survival benefi t with GDT [80,97]. One of these studies demonstrated a reduction in complication rates [97], whilst the other demonstrated a trend towards a reduction in complication rates [80].
Supranormal physiological targets, targeting DO 2 I or CI, the use of inotropes in addition to fl uids, and the use of a PAC were also associated with an improvement in survival. As fi rst demonstrated by Shoemaker and colleagues [19], a supra normal physiological target of global oxygen delivery to ameliorate the oxygen defi cit incurred during major surgery is associated with a survival benefi t. Th is is likely to explain the other associations with an improve ment in morbidity across all risk groups. Th e combination of fl uids and inotropes is more likely to achieve a supranormal physiological target, as opposed to fl uids alone. All eight studies using the oesophageal doppler used fl uids alone, refl ected by the lack of mortality benefi t with the use of FTc or SV as a target. Th e survival benefi t associated with the use of PACs is unlikely to be due to the use of the PACs per se. Th e survival benefi t associated with PAC use may be explained by a number of factors. Th ese include the ability to measure and there fore achieve supranormal DO 2 I, and the use of inotropes in addition to fl uids in all studies using a PAC.
Th e reduction in the number of patients suff ering postoperative complications was seen across all sub groups, apart from studies that did not use DO 2 I, CI, SV, or FTc as a goal. However, there was a trend towards fewer complications among the GDT cohort in these studies. Goals used by these studies included lactate, pulse pressure variation, plethysmographic variability index, Bender et al. [  pulmonary artery occlusion pressure, oxygen extraction ratio, and intrathoracic blood volume [73,74,76,80,87,93,95]. Consistent with the trends seen with mortality, the reduction in complication rates was most profound in the extremely high-risk group of patients, protocols with supranormal physiological targets, targeting DO 2I or CI, and the use of inotropes in addition to fl uids. In contrast to the benefi ts seen in mortality, however, the subgroup using the 'other cardiac output monitors' had a greater reduction in complication rate than the subgroup using the PAC. Th is may relate to the complexity and invasive nature of the PAC in comparison to less invasive cardiac output monitors [98][99][100]. Th ere remains signifi cant heterogeneity in complication rates among postoperative patients in diff erent centres [4,96]. Although diff erences in patient demographics are not modifi able, optimal management of the high-risk surgical patient during the perioperative phase may improve overall outcomes. Despite a requirement for an increase in healthcare resources to off er early GDT to high-risk surgical patients, reductions in immediate postoperative complications translate to overall benefi ts in healthcare costs. Any perceived increase in resource allocation results in a lower patient mortality and morbidity, and therefore a fi nancial saving [101]. Furthermore, reduc tion in immediate postoperative complications has far-reaching eff ects, with a potential benefi cial eff ect on long-term survival [102].
This meta-analysis includes trials from 1988 to 2011. As surgical techniques, perioperative care, and patient selection have been refined over these years, the overall mortality of patients has reduced. As such, the applica bility of historical trials to current day practice may not be valid. This has recently been evaluated in a meta-analysis of 29 perioperative GDT trials carried out between 1995 and 2008 [5]. Th ere was an approximate halving of mortality rates in the control group every decade (29.5%, 13.5%, 7%). Despite a reduction in mortality rate, the morbidity rate remained constant, with approximately a third of patients experiencing post operative complications. Perioperative GDT should there fore off er a reduction in complication rates in current practice.
We acknowledge that there is an element of subjectivity in our decision to include trials in this meta-analysis. Many studies were conducted in single centres with limited patient numbers, and not all studies conducted were of a high quality design. Th is is refl ected by the median Jadad score of 3. Th e eff ect of study quality on outcomes of GDT trials has been analysed in a recent meta-analysis [5]. Most perioperative GDT trials were singe-centre studies, and only a few were conducted in a double-blind manner. In contrast to the lower quality studies, the higher quality studies (defi ned as a Jadad score of at least 3) did not demonstrate any benefi t in mortality reduction. However, the benefi cial eff ect of reduction in perioperative complication rates was evident irrespective of trial quality.
One of the main limitations of this study is the lack of data on the volume and type of fl uids given, and the dose of inotropes used due to variation and inconsistencies in reporting. However, it must be emphasised that the absolute volume of fl uids used per se is not as important as the way in which fl uid is given. Fluid therapy must be titrated against a patient's response to a fl uid challenge, with the use of haemodynamic monitoring [103]. Such 'goal-directed' fl uid therapy must also be given at the right time, as GDT is not benefi cial after complications have already developed [104,105].
One of the other limitations is missing data on the number of patients with complications, due to variations in reporting of complications in the literature, with some studies reporting the number of complications as opposed to the number of patients with complications. Furthermore, we acknowledge that the defi nitions and coding of complications are likely to vary between studies. We have analysed data extracted from studies, rather than data of individual patients. As some of the studies included were carried out several years ago, obtaining data on individual patients would not have been possible. Despite these limitations, the results remain consistent across many subgroups of patients, and are consistent with other recent meta-analyses, supporting our hypothesis [5,106] and the recent EUSOS study which showed a mortality of 4% [107]. Th e benefi t in terms of reduction of complications of GDT in the intermediate risk group may have implications for the majority of the European surgical population.

Conclusion
Despite heterogeneity in trial quality and design, early GDT among high-risk surgical patients has a signifi cant benefi t in reducing rates of complications. Th ere is also an associated reduction in mortality among patients at extremely high risk of perioperative death. GDT is of greatest benefi t in patients with the highest risk of mortality.