NAP7 report on perioperative cardiac arrest

Today we have published the report of the College’s 7th National Audit Project (NAP7), which studied perioperative cardiac arrest over a year-long period in almost all UK NHS hospitals and some in the independent sector.  

The report, At the Heart of the Matter, provides the largest, most comprehensive assessment of the incidence, management and outcomes of perioperative cardiac arrest to date. The NAP7 team reviewed 881 cases of cardiac arrest among an annual caseload of 2.71 million anaesthetics: an incidence of 1 in 3,000. 

NAP7 also provides a quantitative snapshot of anaesthetic activity in the UK and data on anaesthetists’ personal experiences of perioperative cardiac arrest.  

NAP7 has been expertly led by Dr Jasmeet Soar and Professor Tim Cook, working with co-authors and Fellows Dr Richard Armstrong, Dr Andrew Kane, Dr Emira Kursumovic and Dr Fiona Oglesby. We are extremely grateful to the almost 11,000 anaesthetists and anaesthesia associates who participated in the baseline survey. We are particularly indebted to our NAP7 Local Coordinators and those who assisted them in data collection and coordination of NAP7 at a local level, and to the Steering Panel and PatientsVoices@RCoA for their advice and guidance.  

Dr Fiona Donald, President of the Royal College of Anaesthetists said:  

“For the first time, we have a clear, comprehensive and detailed picture of perioperative cardiac arrest in all NHS hospitals in the UK. Several million patients undergo surgery each year and this research shows they can have confidence that their risk of a cardiac arrest is very low. For the 1 in 3,000 patients who do experience it, the quality of care is high.   

“NAP7 has huge potential to drive improvements in perioperative care. Alongside the wealth of in-depth data, the report also includes practical and accessible recommendations that will help us improve the prevention and treatment of perioperative cardiac arrest.”   

Headlines from NAP7 

• In the last decade, the population of patients cared for by anaesthetists has increased in age (approximately 2.3 years), comorbidity (the proportion of healthy ASA 1 patients has fallen from 37% to 24%) and in both the prevalence of obesity (overweight or obese patients rising from 49% to 59%) and its extent (a 7.2% increase in severe obesity). These changes significantly increase the risks for patients of undergoing anaesthesia, and impact anaesthesia workload across the entire perioperative pathway. 
• Our survey of overall anaesthetic activity included more than 24,000 patients’ care and identified potentially serious complications in 1 in 18 anaesthetics. Around one-third of these were cardiovascular in nature, and a quarter were related to the airway. Complications occurred disproportionately in urgent or emergency surgery in high risk settings. Complications were more common in the very young and older patients. Complications were also associated with patient comorbidity (ASA grade), male sex, increased frailty, the urgency, duration and complexity of surgery, and out-of-hours procedures. The vast majority of complications were managed effectively by the anaesthesia team. 
• Our registry included 881 reports of perioperative cardiac arrest with an overall incidence of approximately 1 in 3,000 anaesthetics. 
• Three-quarters of patients survived the initial event, and 60% were alive when the case was reported to NAP7, with 44% having been discharged from hospital. These survival rates are notably higher than other in-hospital cardiac arrests – 49.5% surviving the event and 23% leaving the hospital alive. Of those surviving to hospital discharge, 88% had a favourable functional outcome. 
• The most common causes of perioperative cardiac arrest were major haemorrhage (17%), bradyarrhythmia (9.4%) and cardiac ischaemia (7.3%) but varied by surgical specialty. Anaphylaxis was likely overestimated as a cause of cardiac arrest in our survey of anaesthetists and in reported cases, with many of these cases judged by the NAP7 panel to have other causes. 
• The initial rhythm during perioperative cardiac arrest was non-shockable in 82%, and only 17% required defibrillation. Bradycardic cardiac arrest had the highest rate of successful resuscitation and survival to hospital discharge at the time of reporting to NAP7 (86% and 60%), and pulseless electrical activity the lowest (68% and 34%). 
• Patient factors were a key cause in 82% of cases of perioperative cardiac arrest, anaesthesia in 40% and surgery in 35%. In 31% of fatalities, death was judged to be due to an inexorable process. 
• Cardiac arrest was usually well managed. A senior anaesthetist was present at induction of 97% of cases reported to NAP7, including a consultant present at induction in 86% and at the time of cardiac arrest in 73%. Resuscitation was prompt, and help was called for and attended rapidly. Adrenaline was administered in 79% of cases, and other drugs in 38%. Relative overdosing of adrenaline occurred in a small number of cases when lower doses might have been used, and in a small number of cases caused complications. No complications of low dose intravenous adrenaline were seen during anaphylaxis treatment. Calcium (13% of cases) and bicarbonate (7.2%) were administered frequently and sometimes without clear indication, notably in children and postoperatively in critical care. 
• Perioperative cardiac arrest was more common in older. Frailer patients (1 in 5 cardiac arrests with an incidence of 1 in 1,200), and in high-risk and urgent or emergency surgery. 
• Patient groups with better than average outcomes included children (but not those awaiting transfer to a tertiary centre), cardiac surgery patients and cardiac arrest due to suspected anaphylaxis or airway complications. Poorer outcomes occurred in vascular surgery, cardiology, radiology, in frailer and older patients, in cases due to major haemorrhage, in obese patients with a body mass index (BMI) > 40 kg m-2, and in critically ill children awaiting transfer to tertiary care. 
• Perioperative cardiac arrest or death in low-risk patients was very rare. Among patients without significant comorbidity (ASA 1-2) peri-operative cardiac arrest occurred in around 1 in 8,000 cases and fewer than 1 in 100,000 died. 
• The highest prevalence specialties for perioperative cardiac arrest were orthopaedic trauma, lower gastrointestinal, cardiac, vascular surgery and interventional cardiology. The most overrepresented were cardiac surgery, cardiology, vascular and general surgery, with obstetrics underrepresented. 
• Despite many of the patients reported to NAP7 being very high risk patients, only 6.1% had a do-not attempt cardiopulmonary resuscitation (DNACPR) recommendation made preoperatively. DNACPR recommendations were documented in only 24% of cases with documented frailty, with 15% having treatment limitations. Most DNACPR recommendations were not suspended during surgery, and 1 in 5 of those with a DNACPR recommendation who had a cardiac arrest and CPR survived to leave hospital. 
• There were six cases of unrecognised oesophageal intubation in NAP7. Conversely rates of emergency front of neck airway and pulmonary aspiration appeared notably lower than in previous large studies, including NAP4. 
• While care was judged good far more often than poor (76% vs 4.7% of assessments), there are opportunities for improvement, especially in the prevention of cardiac arrest, with elements of poor care before cardiac arrest identified in 32% of cases. 
  • 71% of adult perioperative cardiac arrest cases did not have evidence of pre-operative risk scoring and NAP7 demonstrated that risk tools used to predict short term mortality (eg SORT) have good utility for stratifying risk of perioperative cardiac arrest. 
  • While supervision of trainees by senior anaesthetists was almost universal, access to senior support was occasionally judged inadequate when anaesthesia was delivered in isolated locations. 
  • National guidelines for monitoring during anaesthesia were not followed in a significant number of cases. This reduces the opportunity to recognise early deterioration. Monitoring was notably deficient during transfer of patients to recovery areas and NAP7 included cases where this contributed to cardiac arrest. 
  • Drug choice and/or dosing was judged to have contributed to a substantial proportion of perioperative cardiac arrests. This occurred more commonly in patients who were older and frailer, with higher ASA grade or acute illness and perhaps with propofol and remifentanil based total intravenous anaesthesia (TIVA). Lower doses, slower induction, use of vasopressors and sometimes different drug choices may have prevented some cardiac arrests. 
  • In some patients who were anaesthetised in the anaesthetic room the review panel judged anaesthesia in theatre would have been safer. Also, in cases in which the anaesthetic room was used for induction and cardiac arrest occurred before surgery started, the panel judged anaesthesia to be a key cause of cardiac arrest more commonly and care before cardiac arrest to be poor more often and good less often than in other cases. 
• NAP7 did not receive sufficient engagement and responses from the independent sector (which in addition to externally funded care, provides around one in six NHS funded perioperative care episodes, a proportion which is increasing) and as such has insufficient data to enable us to determine whether perioperative care in that setting is more, equally or less safe than in the NHS. This is a matter of concern. 
• Training of anaesthetists and provision of equipment for managing perioperative cardiac arrest is generally well implemented in NHS hospitals but is incomplete. It is notably less complete for anaesthesia care in children and for anaesthesia delivered in remote locations. 
• Among the over 10,000 anaesthetists responding to our national survey, almost half had been involved in managing at least one perioperative cardiac arrest in the previous two years (7% a child and 4% a pregnant woman) and 85% over the course of their career. Anaesthetists were confident in managing these events but less so in managing the aftermath or communicating with next of kin. 
• There is a potential for unrecognised impact on the staff involved in the management of perioperative cardiac arrest, which may influence future staff wellbeing and patient care. Among around 5,000 anaesthetists who had attended a recent perioperative cardiac arrest 4.5% reported that this had had an impact on their subsequent ability to deliver patient care. This was more common when the cardiac arrest involved a child, an obstetric patient or an unexpected death. In the case registry 3.4% of anaesthetists reported the same and 5.2% declined to answer this question. Formal psychological support for staff after managing cardiac arrests was uncommonly available or accessed. Anaesthetists reported that recent involvement in management of perioperative cardiac arrests most commonly led to negative psychological impacts, while career involvement led most often to positive professional impacts (in one in three anaesthetists) and negative impacts on professional life (in one in four anaesthetists). 
• Comparing these data to previous NAPs – specifically NAP4 which reported on airway complications and NAP6 on anaphylaxis – suggests improvements in the quality of care and patient outcomes over the last decade, despite the increasing challenges of the modern patient population. 

For a summary of findings to accompany these headlines, please see chapter 4 of the NAP7 report. 

NAP7 main recommendations 

NAP7 makes 20 top recommendations following a voting and ranking process by the panel. There are also topic specific recommendations and suggestions for future research at the end of each chapter. 

Organisation of services 

• Resuscitation equipment, that is age appropriate, should be standardised and available in every main and remote site where anaesthesia takes place, including advanced airway management equipment and a defibrillator. 
• Hospital guidelines and individual practice should recognise the following high-risk cardiovascular settings:
  • hypovolaemic and cardiovascularly unstable patients 
  • the frailer and older patient
  • patients presenting for vascular surgery 
  • patients with bradycardia, and those undergoing surgery with vagal stimuli.

In these cases, there should be consideration of the choice, dose and speed of administration of induction drugs. Induction technique may require modification, such as using ketamine instead of propofol or by co-administering vasopressor medication to counteract hypotension. High-dose or rapidly-administered propofol, in combination with remifentanil, should be avoided. Similar considerations apply to the modification of doses of intrathecal drugs. In all high-risk patients, blood pressure should be monitored frequently at induction, whether invasively or non-invasively (eg every 30–60 seconds). 

• All institutions should have protocols and facilities for managing predictable perioperative complications occurring during anaesthesia both in main theatres and remote locations, including:
  • haemorrhage
  • anaphylaxis
  • airway difficulty
  • cardiac arrest.

All clinical staff who deliver anaesthesia autonomously should be trained, skilled and practiced in the management of these emergencies. 

• Each organisation providing anaesthesia and surgery should have a policy for the management of an unexpected death associated with anaesthesia and surgery. Such a policy should include the allocation of a senior individual to oversee care. The policy should include care of the deceased patient, communication with family and provision for staff involved to be relieved from duty and subsequently provided with appropriate support mechanisms. 
• The Independent Healthcare Provider Network (IHPN) and Private Healthcare Information Network (PHIN) should work with commissioners of care, regulators and inspectors to improve engagement with safety-related national audit projects in the independent hospital sector to assess the quality and safety of care delivered. 
• There should be greater clarity in cardiac arrest guidelines for adults and children relating to the closely monitored patient (eg during perioperative care) regarding:
  • when to start chest compressions
  • dosing of adrenaline
  • indications for use of calcium and bicarbonate in cardiac arrest
  • indications for extracorporeal cardiopulmonary resuscitation (eCPR). 

Before 

• Risk scoring, using validated tools, should be a routine part of preoperative assessment and shared decision making. It should be considered both before and after a procedure to ensure patients receive the appropriate level of postoperative care. 
• As part of early preoperative information provision, patients should be provided with a realistic assessment of likely outcomes of their treatment. The information provided should routinely include important risks, including the risk of death during anaesthesia and surgery. 
• Where practical, treatment escalation, including but not limited to do not attempt CPR (DNACPR) recommendations, should be discussed and documented before arrival in the theatre complex in any patient having surgery with any of: 
  • Clinical Frailty Scale score of 5 or above 
  • ASA 5 
  • Objective risk scoring of early mortality greater than 5%. 

Discussions should take place as early as possible preoperatively, with the involvement of an anaesthetist, so that there is a shared understanding of what treatments might be desired and offered in the event of an emergency, including cardiac arrest. 

• Infants and neonates should be recognised as at high risk of airway difficulty during and after surgery and, when critically ill, of cardiovascular collapse soon after induction of anaesthesia. Departments should make provision for senior and expert care of these patient groups at all times of day and night. 

During 

• Regardless of location, anaesthesia should not be performed unless appropriate preoperative observations, investigations, risk assessment and team brief have been performed. 
• Robust supervision processes should be in place for anaesthesia care delivered by those in training or who do not work autonomously. There should be clear processes for contacting appropriate expert assistance during an emergency and both parties should be aware of these processes. This applies particularly when caring for children and when working in remote locations. 
• A standard procedure to effectively call for help, which includes an audible alarm, should be provided across all locations where anaesthesia takes place. 
• Monitoring should be consistent with published guidelines and continuous throughout the perioperative patient journey, including during transfers. Disconnections in patient monitoring should only occur exceptionally. 
• The level of monitoring should match patient risk. The majority of NAP7 reviewers advocated a lower threshold for continuous invasive arterial blood pressure monitoring in theatre and recovery. Research to inform national guidelines would be of value. 
• High-risk or deteriorating patients should be anaesthetised in theatre on the operating table. 
• All clinical staff who deliver anaesthesia care should be trained and competent in the administration of intravenous adrenaline, both as a low-dose bolus and infusion. 
• In monitored patients in early cardiac arrest or a severe low flow state, initially give small doses of intravenous adrenaline (eg 50 μg in adults or 1 μg/kg in children) or an infusion of adrenaline, and if return of spontaneous circulation (ROSC) is not achieved within the first 4 minutes (about two 2-minute cycles of CPR) of cardiac arrest, give further adrenaline boluses using the standard cardiac arrest dose (1 mg in adults or 10 μg/kg in children). 

After 

• Due to the severity of its nature, all cardiac arrests should be reviewed to understand the cause, discover potential learning and support staff. Learning should be shared across the whole perioperative team. 
• All cases of cardiac arrest should be communicated to the patient, next of kin, or parents if the patient is a child, as part of the duty of candour. 

These recommendations are available in chapter 5 of the NAP7 report.