Biomarkers in Acute High-risk AbdoMinAl Surgery

Acute high-risk abdominal surgery (AHA) is performed in hospitals worldwide. Ethiologies are heterogeneous, but overall emergency surgery carries a very high mortality rate (1)(2). In particular, emergency laparotomies performed on elderly people has a high mortality rate(3)(4). Different quality improvement programmes for that particular patient group have been suggested, but the quality of care and thereby mortality varies considerably between hospitals(5)(6). The use of postoperative intensive care and monitoring seem to be inadequate for this high risk population in a variety of hospitals(1)(7)(8). It is of paramount importance to identify the frailest and most acutely deranged patients, who are in risk of poor outcome, in order to allocate resources for improved monitoring and optimisation postoperatively. It has been shown that failure to rescue patients after having developed postoperative complications and inability to escalate care intensity affects the outcome. Organisation, teamwork and culture is important in the postoperative phase to be able to escalate care especially in standard care wards(9)(10). However, it is found difficult to predict which patients will experience complications in a standard care ward.

Different risk assessment tools have been proposed for patients undergoing AHA(11)(12). The APACHE-II score, even though developed for critical care, seems to be the one with best prediction of outcome. Risk assessment tools are important to support clinical decision making by the perioperative team as subjective clinical assessment often underestimates the risk for the patients in highest risk of complications and death(13). Good clinical decision-making is likely to improve the clinical outcome by allocating appropriate resources. Prognostic tools are also useful to inform patients about what to expect in the immediate postoperative phase and of long-term outcome. Especially in the elder population with increased risk of long hospitalisation and loss of function or independency this can be useful to give informed consent to treatment strategy. Furthermore, good risk assessment is important to optimize palliative care after end-of-life decisions, which is often ignored in research, but highly relevant in clinical work.

During recent years use of prognostic biomarkers in other high mortality populations have received much attention for risk stratification(14). An ideal biomarker should be readily available upon decision-making, easy to measure, reliable and biologically stable. Furthermore it has to be able to accurately differentiate prognosis for patients to have value in the clinical decision-making and guide the treatment of the patient. It should be linked to the clinical outcomes.

The investigators aim to identify serological AHA biomarkers that are prognostic or predictive for postoperative morbidity, mortality and length of hospitalisation.

Data collection Data on mortality was obtained using the Danish Civil Registration System(30). This register is maintained by the Danish government and assigns a unique personal identification number to all Danish citizens. It contains data on address, immigration emigration, gender, date of birth and exact day of death. It is updated within days of any change in information.

Other data were collected in the medical records of patients. Complications were defined using the Clavien-Dindo Classification for Surgical Complications(31).

Outcome measures Primary outcomes are short and long-term mortality (30 days and 180 days respectively) and major complications during index hospitalization (Clavien-Dindo >2). Secondary outcomes are length of hospitalisation, length of intensive care admission as well as perioperative need for inotropic or vasopressor support and high volume fluid resuscitation.

Description of biological material Blood samples were collected at the time of induction of anaesthesia. For AHA-patients a pre-operative optimisation protocol was introduced in the hospital throughout the data collection period, which stated that the patient was optimized with goal directed fluid therapy before induction of anaesthesia. The blood samples were usually taken from the arterial line used for invasive blood pressure monitoring but some were obtained from venous puncture with standard preoperative tests.

The blood samples were collected in tubes with EDTA as anti-coagulant for plasma and in tubes with coagulation enhancer for serum. After centrifugation for 10 mins at room temperature the plasma and serum, respectively were transferred to freezing tubes for subsequent long-term storage at -80C.

Limitations of the study Data collection during clinical work including blood samples requires resources and is susceptible to selection and information bias. It is possible that resources for collecting blood samples correctly where reduced for the most unstable patients at the time of induction of anaesthesia. Patients who were considered too frail to undergo surgery, where a palliative or conservative approach was chosen, were not included in this study. Postoperatively, patients in the AHA-cohort were assigned to postoperative intensive monitoring in the post-operative care unit if they had American Society of Anaesthesiologist physical performance score(34) above two. This stratification is susceptible to information-bias because of inter-observer variation(35).

The data is collected in a single medical facility with and optimized care protocol for AHA and can therefore not uncritically be applied to other populations.

Economical support and budget:

The clinical study as well as establishment of the biobank has been supported by a grant from the Capital Region of Denmark (Region H). The cost of biochemical analysis will be covered by the Department of Clinical Biochemistry, Hvidovre Hospital. Other expenses will be covered by the AHA research group, Department of Anaesthesiology and Intensive Care and Department of Surgical Gastroenterology, Hvidovre Hospital.

Publication of results:

Results from the studies will be published in medical journals indexed in the PubMed/EMBASE database.