Introduction

Crude global estimates suggest that more than 300 million patients undergo surgery each year and this number continues to rise [1]. Mortality following uncomplicated surgery is low and recent international estimates put this figure at less than 1% [2, 3]. However, within this very large population there are high-risk patient groups. Major elective and emergency procedures are offered to an increasingly elderly and comorbid population [4]. Complications following surgery are associated with a five-fold increase in mortality [3] and contribute to long-term morbidity and cost [5]. Perioperative admission to critical care has long been regarded as an important standard of care for certain types of high-risk surgery, and recent data from the UK suggest that 40% of all critical care admissions follow surgery [6]. Unplanned admission to critical care following surgery is also associated with worse outcomes and increased need for organ support [7]. Thus, interventions aimed at identifying those in need of perioperative critical care admission and improving care of those admitted in the perioperative period are of great importance.

We set out to highlight recent advances, describe the key uncertainties and propose research priorities for perioperative critical care in the next 5 years.

Methods

This article was commissioned by the editorial board of Intensive Care Medicine journal. We undertook a priority setting process (PSP) using a modified Delphi technique, similar to previously published research PSPs for critical care and anaesthesia [8, 9]. A panel of experts in the field of perioperative intensive care, representing different geographical areas, were selected. A patient representative member was included in the panel (N.P.). The process consisted of three phases (Fig. 1).

Fig. 1
figure 1

Overview of the research priority setting process

Full size image

Survey responses were collated and categorised by two experienced specialists in perioperative and intensive care medicine (M.A.G. and R.M.P.). Current standards (question 1) reflected usual care in the authors’ regions of practice. Responses to the “recent advances” (question 2) were grouped into themes and organised into preoperative, intraoperative and postoperative sections. Key uncertainties and possible research priorities were sought (questions 3 and 4). Ten important uncertainties and ten potential research questions were identified by consensus. For each identified uncertainty a search for future or on-going large clinical trials was undertaken (www.clinicaltrials.gov; www.isrctn.com). For each research question a study question using the “Participants, Intervention, Comparator, Outcome” (PICO) format was developed.

Results

Current standards of perioperative critical care

A summary of what was considered to be current standards of perioperative critical care in the geographic region of practice of working group members where resources were not a constraint is presented below.

Preoperative assessment

For many patients, the need for major surgery may accompany deterioration in other chronic diseases, so detailed preoperative assessment before high-risk surgery is essential to plan appropriate management. Medical complications following surgery, particularly nosocomial infections or cardiac complications, are important causes of adverse outcome after surgery [5, 10] and so the risk of these must be minimised. It is therefore important (and increasingly necessary) that other specialists beyond anaesthesia and surgery are involved in optimising patients for major surgery.

International guidelines recommend that functional assessment is carried out before high-risk surgery [11]. One method is cardiopulmonary exercise testing (CPET), which allows an objective assessment of exercise capacity providing a means of risk stratification, and triage for postoperative admission to critical care [12]. Plasma biomarkers such as NT pro-BNP may also be helpful in this context, although evidence for their use is less clear [13].

When patients present for emergency surgery, there is often insufficient time for a detailed preoperative assessment. Here, the use of preoperative risk scoring systems may allow an estimation of risk and help triage patients for postoperative admission to critical care. The American Society of Anesthesiologists Physical Status score, Portsmouth Physiological and Operative Severity Score for the Enumeration of Mortality and Morbidity (P-POSSUM) and the Surgical Risk Scale have all been used for this purpose, although none of these scores is ideal [14].

Intraoperative care

For many years, the primary focus of perioperative care has been safety and over recent decades there has been a focus on team and human factors (e.g. WHO safety checklists). Thus anaesthetic-related and perioperative deaths have declined markedly in the last five decades [15]. The major development area for intraoperative care has now become the delivery of reliable high-quality care in order to prevent postoperative complications.

Current evidence suggests the intraoperative management of major elective gastrointestinal surgery should include consideration of cardiac output monitoring for haemodynamic management. Findings from a recent systematic review suggest a reduction in complication rates following this intervention [16]. Appropriate intravenous fluid therapy and avoiding intraoperative hypotension are also recommended to prevent acute kidney injury [17].

For patients undergoing major abdominal surgery who are at risk of pulmonary complications, a protective ventilation strategy is recommended, avoiding ventilation with excessive tidal volume, some level of positive end-expiratory pressure (PEEP) and recruitment manoeuvres [18]. Finally, although definitive evidence is lacking, most commentators agree that effective postoperative analgesia including epidural and other regional analgesic techniques are recommended following major surgery to minimise opiate use and reduce respiratory complications.

Postoperative care

Immediate postoperative care of high-risk patients should take place in a post-anaesthesia care unit or critical care unit; however, organisation and implementation of such care varies widely, particularly in settings where resources are limited. In most hospitals worldwide, low-risk patients are admitted to a post-anaesthetic care unit or recovery ward for up to 4 h for basic observation immediately following surgery. These facilities represent the minimum standard of care and do not offer organ support on a routine basis.

Many hospitals provide extended post-anaesthetic care for high-risk patients, with many features of the traditional critical care unit, including short periods of invasive and non-invasive ventilation, and inotropic therapy. The key benefit of such facilities is the ability to provide early postoperative critical care interventions as routine to a larger number of patients. Later after surgery, the identification of patients at risk of deterioration, with early intervention and organ support remains key to ensuring good postoperative outcomes for high-risk patients. Delivery of such care may be difficult in setting where resources are limited.

The Enhanced Recovery After Surgery (ERAS) approach is recommended following many types of surgery [19]. This is a group or ‘bundle’ of interventions aimed at encouraging early enteral nutrition and mobilisation, improving postoperative comfort, minimising nausea and vomiting, and optimising hydration. This approach may reduce duration of hospital admission, rate of complications and improve patient experience [20]. Regular review by an acute pain team should be mandatory for patients with complex analgesia needs. Acute pain teams also fulfil many other aspects of the perioperative medicine team role and undertake surveillance for complications and physiological deterioration. Elderly patients are at risk of postoperative cognitive dysfunction. Involvement of specialists in the medical care of the elderly patient is now commonplace, e.g. in emergency orthopaedic surgery [21].

What are the recent advances in perioperative critical care medicine?

A summary of what the expert group considered to be recent advances in perioperative medicine is presented in Table 1.

Table 1 Recent advances in perioperative critical care
Full size table

Organisational: perioperative medicine

The needs of high-risk surgical patients are best met through a specific care pathway covering all aspects of care from the moment of contemplation of surgery until full recovery: preoperative assessment, optimisation of co-existing disease, standardisation of intraoperative care, individualised postoperative care and discharge planning. This may be surgery specific, e.g. emergency laparotomy, fractured neck of femur, or colorectal resection, or focussed on specific high-risk patient groups, such as the diabetic patient. In recent years, the organisation of these many and varied care pathways has been brought under the single umbrella term of Perioperative Medicine to promote the delivery of high-quality perioperative care for all patients. In some types of surgery this has led to marked reductions in inpatient stay [22]. The intensive care physician has a key role to play in this development, in particular by improving access to a simplified form of perioperative critical care, and in the early prevention of physiological deterioration in the standard ward environment after surgery and avoidance of “failure to rescue” [23]. In future, remote monitoring of physiological variables using wireless technology may allow earlier detection of the deteriorating patient.

Preoperative assessment and risk reduction

Preoperative care should involve early contact with an anaesthetist or perioperative physician. This will facilitate individualised patient care, e.g. in the treatment of anaemia or cardiac disease. The use of physical training or ‘prehabilitation’ before surgery to improve physical fitness as well as to optimise nutritional and psychological status before surgery is currently being evaluated in various settings. Along with education packages to improve the expectation and understanding of postoperative care, these techniques may greatly improve how patients engage with perioperative care. Sophisticated preoperative assessment algorithms are now being developed which combine CPET data with medical history, clinical risk scores and epidemiological data. This may provide a more accurate risk of death or complications [24]. There may also be a role for plasma biomarkers such as estimated glomerular filtration rate, NT pro-BNP, and troponins for risk assessment in this context [13, 25]. In some circumstances, the risk of major surgery is very high and the planned operation may not be in the patient’s best interests. Shared decision-making with the patient allows consideration of other options such as chemo/radiotherapy or palliation (for cancer surgery) or physical therapy (for joint replacement surgery), which best meet patients’ hopes and expectations.

Intraoperative care

There is an increasing use of algorithms that incorporate physiological goals to direct intraoperative therapies. This is well established in haemodynamic management where invasive monitoring allows the titration of fluid, vasopressors and inotropic therapy to achieve prespecified blood pressure, cardiac output or other goals. Use of these algorithms may reduce complications and duration of hospital stay [16, 26]. A ventilator strategy targeting low tidal volume ventilation (6 ml kg−1), use of PEEP and recruitment manoeuvres may reduce re-intubation rates and postoperative respiratory complications in high-risk patients undergoing abdominal surgery [18]. A recent meta-analysis suggests that driving pressure may be a key determinant of harm from intraoperative mechanical ventilation [27]. Electroencephalogram (EEG)-based monitors which determine depth of sedation and anaesthesia, e.g. bispectral index, are often used to prevent intraoperative awareness in situations where this is problematic using traditional means (e.g. cardiac surgery, during total intravenous anaesthesia). There is now increasing interest in the use of this monitoring approach to prevent excessive dosing of anaesthesia agents. The use of point of care coagulation testing to guide blood component therapy has become routine for some types of major surgery and trauma. Again, the use of algorithm-based coagulation management may reduce blood loss, transfusion and improve patient outcomes [28,29,30]. A recent meta-analysis investigating its use in trauma found insufficient evidence to make recommendations on its routine use in this setting [31].

Postoperative care

Ensuring patients at high risk of deterioration after surgery are monitored in an appropriate setting is an important standard of care. These patients are often managed in a critical care area in the early postoperative period, but remain at risk of developing complications following discharge to a standard ward. There has been an increase in the use of early warning scores in this setting, and also in the provision of intensive care outreach teams or physician-led perioperative medicine teams in an attempt to identify patients who are at risk [23]. The hospital level rate of death from postoperative complications, known as ‘failure to rescue’ is increasingly used as a metric of the safety and effectiveness of postoperative care [10]. Hospitals with a high-risk case-mix but very effective surveillance systems may use failure to rescue rates to demonstrate the quality of postoperative care.

Current beliefs and uncertainties in perioperative critical care

The process of identifying the key uncertainties is outlined in Fig. 2. The literature review process for these uncertainties is accessible in the Supplementary File. The key uncertainties are listed in order of importance as rated by the panel.

Fig. 2
figure 2

Process of identification of key uncertainties in perioperative critical care. PEEP positive end-expiratory pressure, ICU intensive care unit, CPET cardiopulmonary exercise testing, CPAP continuous positive airway pressure, GDT goal-directed therapy

Full size image
  1. 1.

    Fluid therapy

    Intravenous fluid therapy is a central component of perioperative care. In recent years, the traditional theories of fluid mechanics have been revised, and a new model involving the endothelial glycocalyx layer has been proposed [32]. The glycocalyx plays an important role in fluid ultrafiltration and maintenance of plasma oncotic pressure. These new concepts, along with a greater understanding of the potential harms caused by fluids, is driving a re-evaluation of the optimal approach to fluid therapy. There is limited consensus on the most effective type of fluid (e.g. crystalloid vs. colloid, balanced vs. conventional fluids), and dose (volume) [33] and this is reflected in on-going clinical trials (NCT02691676, NCT02721654, NCT01424150).

  2. 2.

    Postoperative critical care admission

    Postoperative admission to a critical care unit is commonly regarded as an important standard for complex major procedures [19]. Disparities exist in patient selection for intensive care; some high-risk patient groups, e.g. surgery for fractured neck of femur or emergency laparotomy, are not routinely admitted to critical care after surgery, whilst others are, e.g. cardiac surgery. This is especially true in settings where resources are limited.

    The use of postoperative critical care for low-risk major surgery may increase financial costs and duration of hospitalisation without improving patient outcomes [34]. Recent studies have failed to demonstrate any mortality benefit associated with postoperative critical care admission [3, 34, 35]. Many of the perceived benefits of critical care, such as high levels of nursing care and observation, could equally be provided in specialist wards. Hence the optimal use of critical care resources in the perioperative period remains unknown.

    Postoperative surveillance to ensure early identification of patients who are at risk of suffering major complications is also considered an area of great importance [23]. Shared decision-making about postoperative care, including critical care (with an advanced directive if appropriate) might better meet patient expectations of survival with acceptable quality of life [36]. Thus high-risk patients being offered major surgery should be fully counselled about the risks of, and the alternatives to, surgery.

  3. 3.

    Cardiac output-guided haemodynamic therapy

    Perioperative heemodynamic therapy [also termed goal-directed therapy (GDT)] has been promoted as a method of reducing postoperative morbidity and mortality for high-risk surgical patients, and also the financial costs of surgery [37]. In early trials, this treatment was used to optimise tissue oxygen delivery for high-risk patients undergoing major surgery, with trial results suggesting that substantial mortality reductions could be achieved using a pulmonary artery catheter-guided treatment algorithm. However, few subsequent trials have demonstrated such impressive results. Differences between trials in terms of monitoring technology, treatment algorithms, patient cohorts, and timing of the intervention have resulted in a confused and controversial evidence base. Many clinicians believe this approach should be used as standard, whilst others are vocal about the risks of harm. Systematic reviews of efficacy trials suggest the treatment is safe and may be beneficial [16, 25, 26], but the findings of the largest trial to date were equivocal with regards to clinical effectiveness. A large trial of GDT in major abdominal surgery is underway (ISRCTN39653756).

  4. 4.

    Postoperative cognitive dysfunction (delirium)

    Delirium is common and often remains undiagnosed in surgical patients. It can be a serious complication, associated with prolonged duration of hospital stay, cognitive decline, risk of other major complications and death [38]. Clinicians caring for surgical patients should be trained to recognise delirium. Strategies which may prevent postoperative delirium include avoidance of drugs known to precipitate delirium (e.g. benzodiazepines), prolonged critical care admission and use of regional analgesic techniques to minimise systemic opioid administration [39]. Recently evidence has suggested that use of alpha-2 agonist drugs in the early postoperative period (e.g. dexmedetomidine) may reduce the incidence of postoperative delirium [40] and a large multinational multicentre trial is investigating the effect of depth of anaesthesia on postoperative cognitive dysfunction and mortality.

  5. 5.

    Perioperative acute kidney injury

    Sepsis and surgery are the two most common precipitants of postoperative acute kidney injury (AKI), which is associated with a considerable increase in mortality. Baseline preoperative renal dysfunction is also an important predictor of death after surgery [41, 42]. Although the pathophysiological mechanisms are incompletely understood, perioperative AKI may be preventable. Avoiding hypotension, hypovolaemia, fluid overload and nephrotoxic drugs is thought to reduce the incidence, severity and duration of postoperative AKI [43, 44] and there is research planned in investigating the impact of a postoperative AKI prevention bundle involving these elements (NCT02583945). Use of clinical risk prediction models [45, 46] and the measurement of biomarkers of kidney injury may enable earlier intervention and monitor the effectiveness of future treatments [47].

  6. 6.

    Prevention of postoperative respiratory complications

    Pulmonary complications are a major cause of morbidity and mortality after surgery, accounting for up to one in four deaths in the first week after surgery [48]. Identifying patients at high risk of postoperative pulmonary complications would allow modification of perioperative care to improve survival. Many advocate use of epidural analgesia for major thoracic and abdominal surgery, although evidence of reduced postoperative pulmonary complications has not been demonstrated.

    Although postoperative incentive spirometry has not been shown to be beneficial in reducing pulmonary complications after thoracic or abdominal surgery, there is some evidence to suggest that lung expansion using CPAP may be beneficial in higher-risk groups [49, 50]. Current evidence does not support routine use of high flow nasal oxygen after surgery [51]. Other potentially iatrogenic causes of postoperative respiratory failure include hyperoxia which may increase the incidence of wound infection and duration of ventilation [52, 53], and excessive maintenance fluid which may increase the incidence of respiratory failure after thoracic surgery [54]. Current areas of active research in this area include CPAP to prevent postoperative respiratory complications (e.g. ISRCTN56012545) and the effect of oxygen therapy on postoperative infectious complications.

  7. 7.

    Blood and coagulation management

    A restrictive transfusion strategy (i.e. haemoglobin target of 70–90 g L−1) is considered appropriate in stable, hospitalised adults including critical care patients [55]. The situation in patients who have undergone major or high-risk surgery is less clear. Some commentators suggest that a restrictive transfusion strategy may not be safe in this context, particularly for patients with cardiac disease [56]. Use of intravenous iron has not been shown to benefit critically ill patients [57], and trials in surgical patients are ongoing (NCT01692418, NCT02972294).

    Point of care coagulation testing using thromboelastography and thromboelastometry are increasingly used to manage blood component therapy in trauma and major surgery patients. Available data suggests that blood component therapy administered according to an algorithm using these tests may reduce blood transfusion rates, although the effect on patient outcomes is unknown [28,29,30]. A recent meta-analysis of use of point of care coagulation testing in adult trauma patients found insufficient evidence to make a firm recommendation on its use in this setting [31].

  8. 8.

    Perioperative ventilator management

    Modifiable ventilator parameters may influence pulmonary outcomes after surgery. Recent study data suggests [18, 58,59,60] that occurrence of pulmonary complications is influenced by three intraoperative ventilator settings: tidal volume, PEEP and the resulting driving pressure [59]. However, uncertainty remains over the role of each component in trial interventions. A recent trial demonstrated no benefit and the possibility of harm from high PEEP during intraoperative ventilation [61]. Thus benefit from the lung protective ventilation strategies might best explained by reductions in tidal volume size, and not the increase in levels of PEEP [59]. Current trials are focused on driving pressure and the optimum PEEP and their role in prevention of postoperative lung injury (e.g. NCT02851238, NCT02963025, NCT02148692).

  9. 9.

    Perioperative immune and inflammatory dysfunction

    Inflammation and activation of the immune system occurs normally following major surgery to promote tissue healing. The subsequent impairment of innate and adaptive immune responses results in immune compromise and a reduced bactericidal response to pathogens [62]. This phenotype of host immune compromise is further exacerbated by the effects of general anaesthesia. The consequence of postoperative immune compromise is increased susceptibility to nosocomial infections [63]. Although perioperative alterations in immune biomarkers such as IL-6, IL-10 and HLA-DR are associated with an enhanced susceptibility to postoperative infection, it remains unclear which biomarkers most accurately predict infection secondary to immune suppression. Whilst this immunological deficit is reversible using immune stimulants such as interferon gamma (IFNγ) and granulocyte–macrophage colony-stimulating factor (GM-CSF) the clinical consequences of these therapies in the perioperative setting remain unclear [64, 65]. A persistent immune suppression and catabolism has been identified following the tissue injury associated with major trauma but not as yet following elective or non-trauma emergency surgery [66].

  10. 10.

    Antibiotic therapy

    Antimicrobial resistance is perhaps the greatest single challenge facing modern healthcare and clinicians must balance need for antibiotic therapy with preserving their effectiveness. Prophylaxis is often recommended to reduce the risk of surgical site infection, and is the most frequent indication for antimicrobial use during the perioperative period. However, the evidence to support the use of antimicrobial prophylaxis is several decades old, preceding modern innovations which reduce surgical site infection e.g. modern antiseptic preparation, minimally invasive surgery, operating theatre airflow systems. There is widespread variation in practice of antimicrobial prophylaxis for surgical patients. When source control is secured with surgical or radiological drainage, antimicrobial therapy may be unnecessary in uncomplicated conditions e.g. appendicitis or cholecystitis [67].

    When antibiotics are given for life-threatening infections, timely administration and assurance of adequate tissue concentrations at the infection site are essential. Underdosing antimicrobials can result in resistance and treatment failure. Tissue concentrations of antibiotics may change during sepsis by hyperdynamic circulation, capillary leak, hypoalbuminaemia and need for large volumes of fluids during resuscitation. These may affect the volume of distribution, resulting in high antibiotic clearance and the possibility of inadequate dosing. These pharmacokinetic changes may result in failure to achieve pharmacodynamic targets, and hence the ability of antimicrobial agents to treat infection. Clinicians should consider these changes when prescribing antimicrobials as well as the interaction between the pathogen and antimicrobial agents at minimal inhibitory concentrations [68]. Changes to the dose and infusion regimen may be needed as well as more frequent plasma measurement of drug levels [69]. These enhanced dosing regimens require further investigation.

Ten priority research questions for perioperative critical care

The process for identifying and ordering the research priorities is presented in Table 2. The final ten research questions ranked in order of priority, with proposed PICO, are outlined in Table 3. Priorities marked with an asterisk were identified as the top five research questions by the patient representative member of the panel.

Table 2 Research priorities for perioperative intensive care medicine
Full size table
Table 3 Top ten most important research questions in perioperative critical care with “Population, Intervention, Comparator, Outcome (PICO)” structure
Full size table
  1. 1.

    What is the most appropriate use of fluid therapy in the surgical patient?

    Intravenous fluid therapy represents the cornerstone of perioperative management but unanswered questions remain. We need to confirm which formulation of fluid is best suited to surgical patients, and whether balanced crystalloid and colloid solutions offer any advantages over traditional fluid therapies such as normal saline. There is debate about the value of further trials to confirm the safety and efficacy of hydroxyethyl starch solutions for surgical patients who are not critically ill. Some argue that evidence from large trials of critically ill patients should influence this practice, whilst some believe that separate trials are needed [70]. What are the effects of fluid dose (volume) at different stages of the perioperative period, and does this vary between different procedures?

  2. 2.

    What is the best use of intensive care resources in the perioperative period? *

    Research is required to define the groups most likely to benefit from postoperative admission to critical care, and to identify patients in non-critical care areas at risk of deterioration. We need to investigate which preoperative biomarkers may best help to identify patients at greatest risk of postoperative morbidity and mortality, and hence need for postoperative critical care? What is the role of alternatives such as overnight or 23-h post-anaesthesia care units or specialist wards in delivering postoperative care following high-risk surgery, and are they more cost effective than traditional critical care? How can we identify patients at risk of developing postoperative complications in a ward setting? How can we improve the safety and quality of postoperative care using metrics such as failure to rescue?

  3. 3.

    How can postoperative cognitive dysfunction (delirium) be prevented and treated? *

    Postoperative cognitive dysfunction is common and associated with increased morbidity and mortality. Studies are needed to evaluate strategies to prevent cognitive dysfunction including optimum mode and depth of anaesthesia and ICU sedation, neuroprotective agents, and specific drug treatments for delirium.

  4. 4.

    Is cardiac output-guided haemodynamic therapy (goal-directed therapy) an effective way to improve postoperative patient outcomes?

    There have been many small efficacy trials of this treatment, utilising different monitoring technologies, different algorithms and different patient cohorts. Most suggest improved postoperative complication rates and hospital length of stay [16, 26], but definitive evidence of clinical effectiveness is still lacking [16]. There remains a need for a large adequately powered randomised trial to confirm whether the benefits of this intervention are sufficient to justify routine use in all patients. Further questions include the incremental benefit of inotropic therapy as a component for such algorithms, and the role of ‘dynamic’ fluid responsiveness parameters such as stroke volume variation.

  5. 5.

    How can we prevent and treat postoperative acute kidney injury (AKI)?

    Future studies should focus on the prognostic significance of mild (i.e. KDIGO stage 1) AKI. Is this a suitable endpoint for studies aimed at reducing perioperative AKI? What is the clinical significance of preoperative chronic kidney disease, and can this be mitigated? Is there a role for biomarkers in the detection of subclinical AKI or risk stratification in surgical patients? Can protocolised fluid and blood pressure management in these patients reduce the incidence of postoperative AKI?

  6. 6.

    How is postoperative respiratory failure best prevented and treated? *

    The prevention and management of postoperative pulmonary complications may have a major impact on patient outcomes and critical care admission. We need to establish whether specific interventions to improve pulmonary function will prevent pulmonary complications. Candidates include prehabilitation, incentive spirometry, inspiratory muscle training, continuous positive airways pressure, and high flow nasal oxygen therapy. There is also a need to confirm the effects of analgesia on postoperative pulmonary complications.

  7. 7.

    How should patient blood management be adapted for the surgical patient?

    Controversy still exists around the ideal transfusion thresholds in the surgical patient, particularly in the setting of cancer surgery, orthopaedic surgery, cardiac surgery, and patients with cardiac disease. Future studies should address the safety of liberal and restrictive transfusion strategies in these patients. The effect of algorithm-based blood component therapy using near-patient tests of coagulation on transfusion requirement, morbidity and mortality also requires further investigation.

  8. 8.

    What is optimal mode of oxygen therapy and invasive ventilation for patients undergoing high-risk surgery? *

    Inspired oxygen therapy is the single most frequently used treatment in the perioperative period, and yet we know very little about the optimal use of this treatment. There are conflicting reports about the benefits of oxygen therapy whilst potential harmful effects are now better understood [71]. Which monitoring targets should we employ for oxygen therapy, and which inspired oxygen concentration provides optimal benefit? What are the harmful sequelae of perioperative hyperoxia? What is the most effective approach to oxygen therapy in the perioperative period? It is now established that different approaches to mechanical ventilation may affect mortality amongst critically ill patients. We need to establish whether similar effects occur in the surgical population where the exposure period is much shorter, and patients are at much lower risk. Postoperative outcomes may be affected by ventilator driving pressure even in patients ventilated for short periods [72]. Future research should explore which components of an intraoperative ventilation intervention offer benefit and the optimal level of PEEP in different surgical groups (e.g. obese, trauma, laparoscopic surgery, etc.). We need to consider the feasibility of trials of low pulmonary driving pressure during mechanical ventilation for surgery.

  9. 9.

    What is the optimal indication and dosing of antibiotic therapy in the surgical patient with sepsis? *

    Infection is a leading cause of morbidity and mortality in surgical patients, but must be balanced against the need for antibiotic stewardship. Future research should establish the need for routine prophylactic antibiotic therapy in the surgical setting and optimum duration of antibiotic therapy, particularly where infection source control is achieved. The impact of perioperative changes in the pharmacokinetics of antimicrobials also warrants further investigation, as does the optimal dosing regimens in life-threatening infections associated with severe physiological disturbance. Availability of rapid near-patient microbiologic testing in the operating theatre may allow improved selection of antibiotic agent and warrant further investigation.

  10. 10.

    Can we modify the inflammatory response to surgery and improve immune function?

    Subclinical immune suppression following major surgery may increase susceptibility to infection, and possibly cancer recurrence. Important avenues of future research include identification of biomarkers that predict postoperative infection secondary to immune suppression, the benefit of perioperative immune stimulation, characterising a perioperative syndrome of persistent inflammation and immunosuppression and its consequences e.g. cancer recurrence and persistent infections.

Conclusion

The population undergoing major surgery is very large and conceals a significant number at much greater risk of complications and death. Improvements in their care could lead to benefit for greater numbers of patients. Although standards of care continue to improve for this patient group, clinicians are now much more likely to offer high-risk patients surgical treatments than 20 years ago.

While the focus of this review is on clinical trials which should be conducted in the next 10 years, it should be remembered that a great deal of research funding is awarded to basic and translational science. The importance of this must also be emphasised as this research sets the direction of the large clinical trials of the future.

It should be noted that this is a research priority setting exercise rather than a formal systematic review for each topic. Moreover, the PSP for this review considered research questions which were a priority in the developed world. Priorities in settings where resources are limited may well be different and many commentators predict a huge expansion in the volume of complex surgery carried out in this setting over the coming decades [73].

It remains a priority for critical care physicians to deliver better care for patients during the perioperative period in order to ensure that surgical treatments are as effective as possible. Many treatment approaches developed in the intensive care unit are increasingly applied to the operating room or perioperative setting, to prevent and treat harm during this period. The findings of this review are that research priorities among clinicians remain focused on delivery of basic perioperative care i.e. fluid and oxygen therapy, ventilation, prevention of complications, whereas patients prioritise prevention of delirium, antibiotic resistance and availability of critical care for those who need it.