Postoperative complications determine long-term survival after high-risk surgery. Although high-risk surgical patients account for a minority of surgical procedures they represent a vast majority of post-operative mortality rates [1]. Therefore, a decrease in post-operative complications is important and can be achieved by using goal-directed fluid optimization [2]. In order to limit post-operative complications, adequate monitoring and optimisation of physiological variables during surgery are necessary.

Hemodynamic variables such as heart rate, arterial pressure and central venous pressure are commonly used but have been proven to be poor predictors of outcome. In contrast, the use of flow-related variables such as cardiac index (CI), oxygen delivery (DO₂) and central venous oxygen saturation (ScvO₂) have been associated with improved outcome after high-risk surgery [3]. Improvement of these hemodynamic and oxygen derived variables may affect tissue oxygenation and as a result preserve organ function. While these variables reflect global changes in either the upstream or the downstream part of the circulation and not regional changes as tissue oxygenation, they may be of vital importance since undetected tissue hypoxia may contribute to postoperative organ dysfunction. Remarkably, even though tissue hypoperfusion occurs frequently during high-risk surgery or in high-risk patients undergoing surgery [4, 5] there is a lack of protocols for the perioperative management of high-risk surgical patients [6].

The purpose of the present interventional pilot study is primarily to evaluate the effect of the use of a protocol based on optimisation of tissue oxygenation (StO₂) on the incidence of perioperative complications. As a secondary objective, we assessed the feasibility of such optimisation protocol in clinical practice.

The study complied with the Declaration of Helsinki. The protocol was approved by the Local Ethics Committee (Medisch Ethische Toetsingscommissie UMC Groningen; METc 2010/209) and the study was registered on ClinicalTrials.gov (NCT01342900). Written informed consent was obtained from all participants during the pre-anesthetic consultation.

Study protocol

StO₂ was monitored non-invasively in all patients by near infrared spectroscopy using the InSpectra System [Model 650; Hutchinson Technology, Hutchinson, MN, USA]. The device consists of a monitor and a sensor, which is placed on the thenar eminescence of the patient. In the control group the anesthetist/investigator was blinded for data from the InSpectra monitor. All patients were treated according to the standard care algorithm (Figure 1). No dobutamine was to be administered in the control group. In addition, patients in the intervention group were treated according to the intervention algorithm aiming to keep or raise the StO₂ ≥ 80% (Figure 1). If the patient’s heart rate exceeded 110 beats per minute or if new rhythm disturbances occurred during dobutamine infusion it was left to the anesthesist’s discretion not to give or not to increase dobutamine.

Randomisation

After induction of anesthesia, but prior to the start of the surgery, patients were randomly assigned to the control group or the intervention group using sealed envelopes.

Follow-up

Upfront, all patients were scheduled to be postoperatively monitored either at an intensive care unit (ICU) or at a prolonged stay at the post-anesthesia care unit (PACU). All PACU and ICU teams were blinded for patients’ allocation. Follow-up for postoperative complications continued during the total hospital stay with a maximum of 28 days.

During (high-risk) surgery or during treatment of critically ill patients, the ultimate goal is to maintain or to restore sufficient StO₂ since it is a prerequisite of aerobic metabolism. In recent years the clinical use of near infrared spectroscopy (NIRS) to detect tissue oxygenation has increased, especially in the monitoring of critically ill patients and during cardiac surgery [8–10]. Data on the intraoperative use of StO₂ in non-cardiac surgery however are scarce. Of note, the study designs differ substantially, which hampers proper comparison of the results. A randomised controlled trial excecuted in a critical care unit showed that postoperative application of stroke volume guided therapy and low-dose inotropic therapy (dopexamine) was associated with improved tissue oxygenation [11]. However, due to study design and insufficient power the investigators were unable to demonstrate an intervention-related reduction in complications rates. Other recent observations revealed that minimum StO₂ values in the perioperative setting are inversely associated with outcome [4]. These findings suggest that perioperative optimisation of StO₂ values may result in better outcome. The present results cannot fully confirm the abovementioned studies. Despite the category of patients randomised here (all ASA III, age > 65 yr) who are at elevated risk for lower StO₂ values [12] and perioperative morbidity [13] a significant reduction in the number of complications in the intervention group could not be demonstrated in this pilot study. The relatively small number of patients is probably the main reason for this. Another important factor is the fact that only half of the patients randomized for the intervention group actually required dobutamine. Also, in only 40% of the patients who received dobutamine the predefined StO₂ threshold of 80% was reached permanently. One could argue if dobutamine administration was the most favorable intervention. The increase in myocardial contractility (β₁ adrenoceptors) and peripheral vasodilation (β₂ adrenoceptors) generated a modest increase in StO₂, CI and ScvO₂.

Our observations raise another relevant issue, namely the adequacy of the predefined StO₂ threshold of 80% used in this study. StO₂ values below 80% are considered inadequate [10, 14] and together with recommendations of the manufacturer a threshold of 80% seemed reasonable. However, in patients scheduled for non-cardiac surgery, the time averaged StO₂ values of approximately 85% have been described [4, 12]. These StO₂ values are comparable to values seen in healthy volunteers breathing roomair [14, 15]. Hence, the StO₂ threshold used in the present study may have been insufficient and the use of a higher StO₂ threshold may have led to fewer complications in the intervention group. Abovementioned studies did not describe an optimal StO₂ target and therefore another StO₂ threshold than 80%, i.e. 85%, may be considered for future studies. Of note, flipside to that coin may be increased use of fluids and vasoactive medication.

We observed a trend of an increased number of RBC transfusion and of more patients in whom RBC transfusion were conducted in the control group compared to the intervention group. Although RBC transfusions were not controlled by study protocol other than a transfusion threshold of Hb = 5.0 mmol/L we doubt if the transfusions conducted were of influence on our results. Surely, the effect of RBC transfusions on tissue oxygenation by increasing oxygen delivery is yet unclear [16, 17]. RBC transfusions elevate the hemoglobin content in the microcirculation but not necessarily the microcirculatory perfusion in cardiac surgery patients [18]. However, considerable interindividual variability exists and the effects of RBC transfusion on StO₂ values seem unpredictable [19, 20].

A yet unraveled issue with application of NIRS is the choice of measurement site [21]. For the measurement of peripheral tissue oxygenation multiple measurement sites are available. However, only few sites including the thenar eminence are established [10]. The thenar eminence has a relatively thin fat layer with only minor inter-individual variation. Also, the thenar eminence participates in systemic edema formation to a lesser extent than other skin regions [22] Abovementioned arguments together with the practical perspective from an anesthesists viewpoint explain our choice for the thenar eminence as measurement site. Nevertheless, in general we suspect that StO₂ measurements mostly convey information on microcirculatory integrity. It is possible that only after a persisting and sufficiently large change in systemic oxygen balance is reflected in thenar StO₂. The main research question is whether the goal-directed StO₂ based treatment is correct and well dosed. Based on the present study we are unable to answer that question.

The development of a feasible protocol to determine any significant differences between the study groups remains important. For a study as such where small differences between study groups are observed, it is essential to accurately collect data. The study protocol should be clearly outlined for investigators, as to avoid protocol deviations. We believe that the present study fulfills these criteria. Data were collected continuously, the attending anesthesists were informed comprehensibly and the study protocol was present in the operating room at all times. Studies have shown that protocol adherence becomes one of the main barriers when implementing changes in a clinical setting, with regards to goal directed therapy (GDT) [23–25]. Reasons for low adherence to GDT include time consumption of conducting the protocol, the tendency of monitoring to be invasive and personal disbelief that such a protocol will result in better outcome [25]. Protocol adherence in this study was 94% which is considered acceptable (≥75%) [2]. This can be attributed to the convenient and non-invasive nature and easy interpretation of the obtained data provided by the NIRS equipment [10].

This study has several limitations. First, the observations of the present randomized controlled pilot study are limited to the intraoperative period only. One could argue that the direct postoperative hours are as interesting in efforts to improve outcome after high-risk surgery. Indeed, clinically important fluctuations in the balance between DO₂ and VO₂ do also occur in the first postoperative hours [26, 27] and corresponding hemodynamic interventions seem appropriate. On top of that, postoperative hemodynamic optimisation improves tissue oxygenation [11] and we believe that goal-directed protocols ideally should also cover the early postoperative period. However, because of technical challenges in reliably transporting all the equipment, including the data logging devices, along with the patient to the ICU or PACU, measurements were restricted to the intraoperative period. While our results tend to indicate an advantage of StO₂ based goal-directed therapy, the limited power of this pilot restricts any clear recommendations before further larger studies are conducted.

In conclusion, this is the first randomized controlled study that used StO₂ values as a target for hemodynamic optimization in high-risk abdominal surgery patients. Our pilot results are inconclusive on the perioperative optimisation of StO₂ values in relation to outcome after high-risk abdominal surgery. Nevertheless, the perioperative strategy described deserves emulation and further research is obligatory to define both the optimal StO₂ threshold and intervention to treat or prevent tissue hypoperfusion.