What are the optimum components in a care bundle aimed at reducing post-operative pulmonary complications in high-risk patients?

Post-operative pulmonary complications (POPC) are defined as ‘a pulmonary abnormality that produces identifiable disease or dysfunction that is clinically significant and adversely affects the clinical course’ (Restrepo and Braverman 2015). They are common, with a reported incidence ranging from 2–40%, and adverse outcomes include mortality, with studies reporting that 1 in 5 patients with POPC will die within 30 days of surgery, and the average hospital stay is lengthened by 8 days (Restrepo and Braverman 2015). A recent international prospective study (LAS VEGAS) of 9864 patients suggested that 28% of patients are at “high risk” of developing POPC and in this cohort, the incidence of POPC is of 19.2%, with respiratory failure affecting 3.4% (Rogerson et al. 2017). Unplanned supplemental use of oxygen is the commonest POPC, and in another contemporary study, even this ‘mild’ POPC was associated with increased risk of mortality and morbidity (Fernandez-Bustamante et al. 2017).

Alongside other perioperative complications, POPC affect long-term outcomes, with patients having lower long-term survival rates following POPC independent of preoperative risk (Khuri et al. 2005). Furthermore, the economic burden of POPC is substantial, with a 55% increase in hospital costs in non-cardiac patients (Shander et al. 2011). Reduction in the rate of POPC starts with the recognition of moderate- and high-risk patients through the use of validated tools such as ARISCAT (Assess Respiratory rIsk in Surgical patients in CATalonia) (Canet et al. 2010; Canet et al. 2015).

Enhanced recovery (ER) pathways are advocated as ‘standard of care’ for major surgery, and there are pathways specific to each surgical specialty (Ljungqvist et al. 2017). Several aspects of the pathways are either established to be associated with reduced POPC (minimally invasive surgery (Lee et al. 2015)) or strongly suspected of being associated with reduced POPC (early mobilisation (Pasquina and Walder 2003)). Adherence to ER components is highly variable, and improved adherence has been associated with improved outcomes (Pecorelli et al. 2017).

Several perioperative interventions aiming to reduce the incidence of POPC have been studied sometimes in combination. A care bundle (CB) is a group of interventions that when used together have greater clinical effect than when used individually. A CB for POPC used in the ‘I COUGH’ study demonstrated a trend towards reduction in POPC, and CBs have significantly improved outcomes in the critical care environment to prevent ventilator-associated pneumonia (Rello et al. 2013).

The Delphi consensus method was first introduced by the RAND corporation in Santa Monica in the 1950s. The method involves gathering expert opinions through rounds of questions and bringing these opinions together to establish a consensus. Each round is followed by anonymous feedback to all participants, and the results of the previous round are then used to create the subsequent round. This means that over the rounds, a consensus can be derived which is as reliable as possible (Dalkey and Helmer 1963).

Our primary objective was to recruit a team of experts to participate in an electronic Delphi consensus, leading to the formulation of a perioperative care bundle ‘POPC-CB’ intended to reduce the incidence of POPC. The POPC-CB was designed to be applied in addition to, rather than replacing, enhanced recovery pathways in patients at moderate to high risk of POPC.

We performed a survey of members of the European Society of Intensive Care Medicine (ESICM) Perioperative Intensive Care (POIC) section, both to identify ‘expert volunteers’ for the Delphi and to gain an appreciation of clinical opinion regarding the potential components of a POPC-CB. From a survey sent to the 6623 members of the POIC group of ESICM, there were 362 responses which helped establish the potential of POPC risk assessments and possible interventions in a POPC-CB for patients screened at moderate to high risk. One hundred thirty-four experts were identified during this process, and they were invited to collaborate with the project. Additionally, 41 authors from key literature on the subject were invited to participate and 38 experts agreed to form the consensus group.

Three rounds of questions were anonymously circulated to the Delphi group via email with a 2-week deadline for responses to each round. The questions comprised a series of statements which the experts were asked to respond to using a 5-point Likert scale alongside free text spaces for comments and references to relevant publications. Some rounds included more open-ended questions and a ranking system for a preferred option. If an expert wanted to skip a question, the response box would be left blank. After each round, the project coordinators collated the results of the round and feedback of themes was given to the group. By the end of the process, the final feedback was sent to the group, this time not anonymised, with the consensus and a suggested care bundle.

We used a Delphi process to design a care bundle (POPC-CB), for patients at moderate to high risk of post-operative pulmonary complications which may have the potential to improve clinical outcomes. The proposed interventions are implemented throughout the perioperative journey: prior to the operation, supervised exercise programme and inspiratory muscle training, and during the operation, low tidal volume ventilation with individualised PEEP, use of routine monitoring to avoid hyperoxia and efforts made to limit neuromuscular blockade, and also post-operatively, deep breathing exercises and mandatory elevation of the head of the bed.

The quality and extent of clinical evidence to support each of the proposed components of the POPC-CB is variable. The Delphi process did not involve a formal evaluation of the strength of evidence for the interventions included in the care bundle, nor for those that were rejected. Some of the individual interventions have been evaluated by others using GRADE or Jadad methodologies (Guyatt et al. 2011; Jadad et al. 1996). A Cochrane review of preoperative inspiratory muscle training demonstrated efficacy at reducing POPC (Katsura et al. 2015). Using GRADE criteria, the Cochrane reviewers suggested the strength of evidence for pneumonia reduction was moderate whilst the strength of evidence for atelectasis and adverse event reduction was low and recommended caution due to the potential for overestimation of treatment effect. Although our Delphi process supports the inclusion of supervised preoperative exercise programme, the evidence for improved postoperative clinical outcomes is limited (O’Doherty et al. 2013). A meta-analysis of nine prehabilitation trials in 435 abdominal surgery patients suggests that supervised exercise can reduce overall postoperative complications, and five of these studies specifically demonstrated a reduction in postoperative pulmonary complications (Moran et al. 2016). A very low quality rating was observed using GRADE criteria due to the small sample sizes and significant risk of bias. Furthermore, the precise components of such a supervised programme have not been fully evaluated in relation to POPC reduction.

Protective lung ventilation strategies incorporating tidal volumes 5–8 mL/kg, higher PEEP and control of plateau pressure are beneficial in critically ill patients with acute lung injury or ARDS (Villar et al. 2006; Petrucci and Iacovelli 2007). In ICU patients without ARDS, a protective lung strategy reduced pulmonary and systemic inflammation and reduced the development of lung injury (Determann et al. 2010). A meta-analysis of 20 studies, using GRADE criteria, suggested the evidential strength for protective lung strategy in patients without ARDS to be moderate and high in reducing lung injury and mortality, respectively, but low in preventing pneumonia (Neto et al. 2012). These findings from critical care patients led to perioperative research exploring whether protective lung ventilation has a role during major surgery. The IMPROVE trial compared a ‘protective’ strategy of tidal volume 6–8 mL/kg ideal body weight, PEEP 6–8 cm H₂O, recruitment manoeuvers every 30 min with ‘conventional’ management of tidal volume 10–12 mL/kg, and PEEP and recruitment manoeuvers at clinicians’ discretion. The protective strategy was associated with a reduction in a composite of pulmonary and septic complications (Futier et al. 2013a). In another trial, PROVHILO, all patients received tidal volumes 8 mL/kg and were randomly assigned to either rather high PEEP of 12 and regular recruitment manoeuvers or to low PEEP 0–2. The result shows no difference in post-operative pulmonary complications but shows increased adverse events in the high PEEP group (Prove Network Investigators for the Clinical Trial Network of the European Society of Anaesthesiology et al. 2014). Even in critically ill ventilated ARDS patients, high versus very high PEEP levels was not shown to influence outcome (Brower et al. 2004). A meta-analysis of 15 studies investigating protective lung ventilation involving 2127 surgical patients concluded that intraoperative low tidal volume is associated with less pulmonary complications but no difference in mortality and length of hospital stay (Neto et al. 2015). The authors used the Jadad criteria to rate the quality of this evidence with seven studies scoring 4/5, seven 3/5 and one 2/5. Our Delphi consensus suggests low tidal volumes for all patients with individualised PEEP and recruitment manoeuvres at the discretion of the attending anaesthetist.

Inadequate reversal of neuromuscular blockade is associated with POPC. Post-operative residual curarization (PORC) is the incomplete recovery of muscle function following perioperative use of neuromuscular blocking agents. The prevalence of PORC is anything between 4 and 45%, when defined as train-of-four < 0.9 and is associated with adverse outcomes including POPC (Berg et al. 1997). Whilst the Delphi consensus recommendation is to limit neuromuscular blockade by careful monitoring of neuromuscular function and effective reversal, we acknowledge that ongoing research assessing the impact of neuromuscular blockade on POPC is awaiting publication and no specific drug or technique can be recommended. The potential for hyperoxia to cause harm is contentious. The association between hyperoxia and atelectasis is well established, although the magnitude of this effect may be overestimated (Pryor et al. 2004). A meta-analysis of 17 randomised controlled trials with 8093 mixed surgical patients used the Jadad score to assess quality and found six studies scored 5/5, four 4/5 and 7 less than 4, with sub-group analysis suggesting that hyperoxia can reduce surgical site infection in colorectal surgery patients (Yang et al. 2016).

Post-operative care from a specialist physiotherapist is a commonplace; however, the interventions delivered by physiotherapists vary significantly, and there is a scarcity of good-quality clinical evidence to inform practice with many small, underpowered studies using inadequate randomisation and unclear interventions (ÖRman and Westerdahl 2010; Pasquina et al. 2006). The incremental benefit of the proposed deep breathing exercises is uncertain, particularly when other aspects of post-operative physiotherapy, such as early mobilisation, are provided (Mackay et al. 2005). A Cochrane review of post-operative incentive spirometry following upper abdominal surgery described a small number of poor-quality studies which were unable to demonstrate any effect on postoperative complications (do Nascimento Junior et al. 2014). Elevation of the head of the bed, allowing gravity to help minimise pulmonary aspiration of oropharyngeal secretions and perhaps improve effectiveness of coughing, has been shown to reduce ventilator-associated pneumonia in critically ill patients although no evidence is available in POPC reduction (Wang et al. 2016).

I COUGH, a post-operative pulmonary care bundle, demonstrated a trend towards reduction in POPC (Cassidy et al. 2013). I COUGH incorporates Incentive spirometry, Coughing and deep breathing, Oral hygiene, Understanding (patient education), Get up (mobilisation) and Head of bed elevation. I COUGH was applied to all patients, not just those at high risk of POPC, and it was evaluated in a single centre and was evaluated prior to widespread adoption of enhanced recovery pathways. Our POPC-CB is designed to supplement established guidelines for enhanced recovery, not replace them (Lassen et al. 2009; ERAS society guidelines n.d.). Although this protocol is designed for all patients undergoing major surgery, elements of ERAS pathways are relevant to reducing the incidence of POPC. These include optimal analgesia, avoidance of nasogastric tubes and bowel preparation, minimisation of ‘nil by mouth’, avoidance of sodium and water overload, routine early post-operative mobilisation, multi-modal analgesia and prevention of nausea and vomiting. A quality improvement programme incorporating ICOUGH into ERAS pathways has reduced POPC over time (Moore et al. 2017).

Patients at moderate to high risk of POPC, undergoing major elective surgery within an enhanced recovery programme, can be identified using screening tools such as the ARISCAT score and may benefit from a care bundle aimed at reducing these complications such as that described during our Delphi process. We believe a multi-centre international evaluation of the efficacy of such a care bundle in reducing the incidence of POPC, length of stay and long-term outcomes should be a priority for research in perioperative care.