Volume 79, Issue 2 p. 186-192
Original Article
Open Access

Peri-operative decisions about cardiopulmonary resuscitation among adults as reported to the 7th National Audit Project of the Royal College of Anaesthetists

J. P. Nolan

Corresponding Author

J. P. Nolan

Professor, Consultant

Warwick Clinical Trials Unit, University of Warwick, UK

Department of Anaesthesia and Intensive Care Medicine, Royal United Hospitals Bath NHS Foundation Trust, Bath, UK

Correspondence to: J. P. Nolan

Email: [email protected]

Search for more papers by this author
J. Soar

J. Soar

Consultant

Department of Anaesthesia and Intensive Care Medicine, Southmead Hospital, Bristol, UK

Search for more papers by this author
A. D. Kane

A. D. Kane

Research Fellow, Consultant

Health Services Research Centre, Royal College of Anaesthetists, Red Lion Square, UK

Department of Anaesthesia, James Cook University Hospital, South Tees NHS Foundation Trust, Middlesbrough, UK

Search for more papers by this author
I. K. Moppett

I. K. Moppett

Director, Professor

Health Services Research Centre, Royal College of Anaesthetists, Red Lion Square, UK

University of Nottingham, Nottingham, UK

Search for more papers by this author
R. A. Armstrong

R. A. Armstrong

Research Fellow, Specialty Registrar

Health Services Research Centre, Royal College of Anaesthetists, Red Lion Square, UK

Severn Deanery, Bristol, UK

Search for more papers by this author
E. Kursumovic

E. Kursumovic

Consultant, Research Fellow

Department of Anaesthesia and Intensive Care Medicine, Royal United Hospitals Bath NHS Foundation Trust, Bath, UK

Health Services Research Centre, Royal College of Anaesthetists, Red Lion Square, UK

Search for more papers by this author
T. M. Cook

T. M. Cook

Consultant, Honorary Professor

Department of Anaesthesia and Intensive Care Medicine, Royal United Hospitals Bath NHS Foundation Trust, Bath, UK

University of Bristol, Bristol, UK

Search for more papers by this author
First published: 22 November 2023

1 Professor, Resuscitation Medicine, Warwick Clinical Trials Unit, University of Warwick, UK

2 Consultant, Department of Anaesthesia and Intensive Care Medicine, Royal United Hospitals Bath NHS Foundation Trust, Bath, UK

3 Consultant, Department of Anaesthesia and Intensive Care Medicine, Southmead Hospital, Bristol, UK

4 Research Fellow, 6 Director, Health Services Research Centre, Royal College of Anaesthetists, London, UK

5 Consultant, Department of Anaesthesia, James Cook University Hospital, South Tees NHS Foundation Trust, Middlesbrough, UK

7 Professor of Anaesthesia and Peri-operative Medicine, University of Nottingham, Nottingham, UK

8 Specialty Registrar, Severn Deanery, Bristol, UK

9 Honorary Professor, University of Bristol, Bristol, UK

Summary

Current guidance recommends that, in most circumstances, cardiopulmonary resuscitation should be attempted when cardiac arrest occurs during anaesthesia, and when a patient has a pre-existing ‘do not attempt cardiopulmonary resuscitation’ recommendation, this should be suspended. How this guidance is translated into everyday clinical practice in the UK is currently unknown. Here, as part of the 7th National Audit Project of the Royal College of Anaesthetists, we have: assessed the rates of pre-operative ‘do not attempt cardiopulmonary resuscitation’ recommendations via an activity survey of all cases undertaken by anaesthetists over four days in each participating site; and analysed our one-year case registry of peri-operative cardiac arrests to understand the rates of cardiac arrest in patients who had ‘do not attempt cardiopulmonary resuscitation’ decisions pre-operatively. In the activity survey, among 20,717 adults (aged > 18 y) undergoing surgery, 595 (3%) had a ‘do not attempt cardiopulmonary resuscitation’ recommendation pre-operatively, of which less than a third (175, 29%) were suspended. Of the 881 peri-operative cardiac arrest reports, 54 (6%) patients had a ‘do not attempt cardiopulmonary resuscitation’ recommendation made pre-operatively and of these 38 (70%) had a clinical frailty scale score ≥ 5. Just under half (25, 46%) of these ‘do not attempt cardiopulmonary resuscitation’ recommendations were formally suspended at the time of anaesthesia and surgery. One in five of these patients with a ‘do not attempt cardiopulmonary resuscitation’ recommendation who had a cardiac arrest survived to leave hospital and of the seven patients with documented modified Rankin Scale scores before and after cardiac arrest, four remained the same and three had worse scores. Very few patients who had a pre-existing ‘do not attempt cardiopulmonary resuscitation’ recommendation had a peri-operative cardiac arrest, and when cardiac arrest did occur, return of spontaneous circulation was achieved in 57%, although > 50% of these patients subsequently died before discharge from hospital.

Introduction

As emergency care and treatment plans are being adopted widely across the UK and many other countries [1, 2], increasing numbers of patients undergoing surgical procedures are likely to have plans in place in the event of their health deteriorating. Decisions relating to cardiopulmonary resuscitation (CPR) are a primary component of such treatment plans. Furthermore, surgical patients have become older and frailer in recent years [3, 4] and it is becoming increasingly important that advanced treatment plans are discussed with those patients who might be at increased risk of peri-operative cardiac arrest. Discussions with patients who are judged high-risk and those with pre-existing emergency care and treatment plans should take place with an anaesthetist well before surgery so that it can be agreed which peri-operative treatments, particularly chest compressions and/or defibrillation, and postoperative critical care, would be appropriate and desired by the patient.

Causes of unexpected peri-operative cardiac arrest may be promptly reversible (e.g. a relative overdose of induction drug, vagotonic response to pneumoperitoneum, sudden arrhythmia), and a witnessed and monitored intra-operative cardiac arrest is associated with better outcomes than out-of-hospital cardiac arrest or in-hospital cardiac arrest in other areas [5, 6]. It is likely that many patients would accept brief resuscitation interventions if the cardiac arrest occurred during anaesthesia was witnessed, monitored and rapidly reversible, and they were unlikely to suffer significant harm consequently. However, CPR can cause injuries. Most commonly, chest compressions can cause rib fractures; after resuscitation from out-of-hospital cardiac arrest, several studies have documented an incidence of rib fractures of > 70% when evaluated by computed tomography and this risk is greater in patients who are older and frailer [7, 8]. Injuries to the viscera including the liver and other intra-abdominal structures may also occur, although less commonly [9]. There is some evidence of poor emergency treatment planning peri-operatively, even in groups of patients known to be at high risk of adverse outcomes, such as patients who are frail and those with hip fracture [10].

The 7th National Audit Project of the Royal College of Anaesthetists (NAP7) studied peri-operative cardiac arrest in the UK. The project provided an opportunity to study current practices in peri-operative ‘do not attempt CPR’ (DNACPR) recommendations.

Methods

The methods of the NAP7 project are described in detail elsewhere [11]. In summary, all NHS hospitals and a subset of independent sector hospitals undertaking anaesthesia care in the UK were invited to take part in the project. A network of local coordinators was established to lead the project in each hospital. A national activity survey examined anaesthetic practices and complications in all cases undertaken in NHS hospitals over a 4-day period and provided the denominator for the NAP7 study of approximately 2.71 million anaesthetic interventions annually in the UK [3, 12]. The activity survey collected data only on survival of the event and not overall hospital survival. A one-year registry collected details of peri-operative cardiac arrests occurring from first anaesthetic contact until 24 h after discharge from immediate anaesthesia care via an anonymous, protected and encrypted online database. Detailed structured case reports omitted date, patient, hospital and clinician details, and each case was reviewed by a panel of clinicians and lay representatives [11]. Included cases formed the numerator for the study and their review enabled quantitative and qualitative analyses. As part of the case review process, cases were classified by specialty and themes of importance, enabling topic-specific analysis (with associated numerator and denominator data). The main results and full project are published separately [6, 13, 14].

Results

The NAP7 activity survey showed that of 20,717 adults (aged > 18 y) undergoing surgery, 595 (3%) had a DNACPR recommendation pre-operatively and, of these, it was suspended in 175 (29%) (Fig. 1). Among the 11,880 adults undergoing elective surgery, 95 (< 1%) had a DNACPR recommendation and this was suspended in 24 (25%). Among 1676 patients aged > 65 y with a clinical frailty scale score ≥ 5 (for whom frailty data were obtained in the activity survey), 418 (25%) had a DNACPR recommendation pre-operatively and it was suspended for 129 (31%) patients (Fig. 1). Of the 418 patients with a clinical frailty scale ≥ 5 and a DNACPR recommendation in place, 252 (60%) were undergoing non-elective surgery.

Details are in the caption following the image
Patients with active ‘do not attempt cardiopulmonary resuscitation’ (DNACPR) recommendations by: (a) age (n = 20,172 patients aged ≥ 19 y) and (b) clinical frailty scale (n = 6466 patients aged ≥ 66 y) in the NAP7 activity survey. Data show the proportion of patients with active () and suspended () DNACPR recommendations during surgery. Data are available in tabular form in online Supporting Information Table S2. CFS, clinical frailty scale.

Of 881 reports to NAP7, 54 (6%) patients had a DNACPR recommendation made pre-operatively and were then reported to NAP7 after a peri-operative cardiac arrest and resuscitation attempt. Of these 54 patients, 38 (70%) had a clinical frailty scale score ≥ 5 (Fig. 2), and 26 (48%) were aged ≥ 85 y. Most of these reported cases (34, 65%) were orthopaedic trauma but included several emergency laparotomies and vascular surgery cases (Table 1).

Details are in the caption following the image
Clinical frailty scale score and cases of peri-operative cardiac arrest in patients with a ‘do not attempt cardiopulmonary resuscitation’ (DNACPR) recommendation who met the criteria for inclusion in NAP7. Two additional individuals aged < 66 y who had a peri-operative cardiac arrest with a DNACPR recommendation have been omitted.
Table 1. Surgical specialty of cases with a pre-operative ‘do not attempt cardiopulmonary resuscitation’ recommendation.
Surgical speciality Number of cases
Orthopaedics – trauma 34
Upper gastrointestinal 3
Vascular 3
Hepatobiliary 2
Neurosurgery 2
Lower gastrointestinal 1
Cardiac surgery 1
Interventional cardiology 1
Electrophysiology 1
Gastroenterology 1
General surgery 1
Thoracic surgery 1
Urology 1
Unknown 2

Relative to the adult surgical population in the activity survey, patients with a DNACPR recommendation were: older (aged > 75 y; 41 (76%) vs. 3080 (18%)) and more frail (clinical frailty scale score ≥ 5; 38 (70%) vs. 1957 (11%)). These patients were also more likely to be undergoing: non-elective surgery (52 (96%) vs. 5378 (31%)); major or complex surgery (33 (61%) vs. 5318 (30%)); and surgery during a weekend (10 (19%) vs. 1494 (9%)) compared with the adult surgical population in the activity survey. Thirty-two patients (59%) had a pre-operative modified Rankin Scale (mRS) score > 1. The only cause of cardiac arrest that was more prevalent in the group of patients with a DNACPR recommendation compared with all adult surgical cases was bone cement implantation syndrome (11 (20%) vs. 21 (2%)).

A total of 25 (46%) DNACPR recommendations were formally suspended, 20 (37%) remained active at the time of cardiac arrest and in 9 (17%) cases the status of the DNACPR recommendation was unknown. Most patients (31, 57%) who received CPR with a DNACPR recommendation survived resuscitation, achieving return of spontaneous circulation for > 20 min, compared with 634 (77%) in patients without DNACPR recommendations. Duration of CPR was > 20 min in 8 (15%) patients with a DNACPR recommendation compared with 168 (19%) in the whole cohort. The duration of CPR was > 1 h in 1 (1%) patient with a DNACPR recommendation compared with 33 (4%) in the whole cohort (Table 2). Of 10 patients surviving to hospital discharge, seven had a functional assessment using the mRS reported on discharge (online Supporting Information Table S1). Overall, mRS increased by 1 in four survivors and was unchanged in three survivors.

Table 2. Outcomes for those patients with pre-operative ‘do not attempt cardiopulmonary resuscitation’ recommendations. Data are number (proportion).
Outcome of event Do not attempt resuscitation order in place All other cases
n = 54 n = 827
Initial
Died 23 (43%) 186 (22%)
Not available 0 7 (1%)
Survived (ROSC for > 20 min) 31 (57%) 634 (77%)
Hospital
Alive 10 (19%) 374 (45%)
Dead 40 (74%) 308 (37%)
Still admitted 4 (7%) 145 (18%)
  • ROSC, return of spontaneous circulation.

Among the patients with a DNACPR recommendation, the NAP7 panel rated overall care as good in 22 (42%), good and poor in 21 (38%) and poor in 1 (2%), with insufficient information to rate care in 13 cases (23%). This is broadly similar to all cases considered by the panel. The panel identified several instances of good practice where pre-operative discussions were held with patients and/or their family members, leading to agreement on temporary suspension or modification of a DNACPR recommendation. When care was rated poor, it mostly reflected a lack of risk assessment, discussion about risks pre-operatively or decision-making as to whether proceeding with surgery was appropriate. In some cases, interventions such as avoidance of general anaesthesia or use of invasive blood pressure monitoring appear not to have been considered.

The panel's opinion was that a DNACPR recommendation should have been considered in a further 34 (4%) cases. This group consisted of mainly older patients (24 (71%) aged > 75 y) with a clinical frailty scale score ≥ 5 (24 (71%)) and most (19 (55%)) were having orthopaedic trauma surgery. Only one patient was having an elective operation. At the time these cases were reported, 23 (68%) patients had died, 4 (12%) had been discharged from hospital and 7 (21%) were still in hospital.

Discussion

The NAP7 activity survey has shown that among all adults undergoing surgery, 3% had a DNACPR recommendation pre-operatively, but among those aged > 65 y and with a clinical frailty scale score ≥ 5 the proportion with a DNACPR recommendation was 25% and this was suspended in approximately one-third of these cases. Of the 881 case reports of treated peri-operative cardiac arrest 54 (6%) had DNACPR recommendations made pre-operatively and 14 (26%) were alive at the time of reporting. Based on an annual surgical case load of approximately 2.7 million [3], we estimate there are 67,000 patients with DNACPR recommendations who present for surgery each year.

A study using the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) reported that in 2012, 566 (0.2%) of 316,997 patients undergoing elective procedures had a DNACPR recommendation and this included those for whom the DNACPR recommendation had been suspended [15]. In the NAP7 activity survey, 0.7% of the 13,743 adults undergoing elective surgery had a DNACPR recommendation, which is four times more common than in the ACS NSQIP study. This difference may reflect the retrospective nature of the ACS NSQIP study as well as the 10-year difference in the periods of study.

In its clinical practice guideline on implementing advance care plans in the peri-operative period, the Association of Anaesthetists recommends that it is usually appropriate to suspend a DNACPR recommendation during the peri-operative period [16]. In contrast to this, we have found that only one-third of DNACPR recommendations are formally suspended pre-operatively. Although DNACPR recommendations are not legally binding (they guide the clinician on what to do in an emergency) [16, 17], and technically do not require explicit cancellation, best practice would be to discuss the implications of suspending or not suspending a DNACPR recommendation with the patient and their families pre-operatively. Such discussions enable shared decision-making based on the patient's values and preferences and support a focus on on patient-centred outcomes.

Most patients are likely to support a suspension of their DNACPR once it is explained that some intra-operative cardiac arrests are rapidly reversible and that any proposed resuscitation interventions would be brief and aimed at restoring a quality of life that would be acceptable to them. The Association of Anaesthetists' working party opines that giving chest compressions to expedite circulation of a drug when cardiac output is low (but when a spontaneous circulation is present) is distinct from CPR [16]. It is also the view of the working party that a peri-operative DNACPR recommendation would not prevent drug administration to treat bradycardia, hypotension or cardiac arrhythmia, or use of defibrillation for a sudden-onset arrhythmia during anaesthesia. However, these potential interventions should be mentioned specifically in the pre-operative discussion with the patient and their relatives so there is a full understanding and agreement about the interventions that will and will not be offered. Such discussions should be documented fully. Failure to engage in discussions about patient preferences and the potential suspension of a pre-existing DNACPR recommendation before surgery may result in the patient receiving treatment that conflicts with their wishes. If an intra-operative cardiac arrest occurs and CPR leads to return of spontaneous circulation, the patient may require a period of organ support in the ICU. Although the patient may not have wanted such interventions, failure to address this possibility pre-operatively, especially if the cardiac arrest is considered iatrogenic, may create pressure to admit the patient to the ICU against their preferences.

Pre-operative discussions with any high-risk patient and their families should include treatment escalation planning, including admission to ICU, invasive ventilation and renal replacement therapy, as well as the potential for CPR. Factors to consider include: patient wishes at that time and in that context; the certainty of death if CPR is not performed; the chances of successful resuscitation; the possibility of harm from the CPR and the impact on organ function subsequently; and the possible need for ICU postoperatively. Defining who belongs to this high-risk group lacks consensus. Recent in-hospital cardiac arrest data indicate poor outcomes for older patients with frailty who undergo CPR [18] and a recent study has established a strong association between higher frailty burden and increased mortality following peri-operative cardiac arrest [19]. Our data and previous studies of peri-operative cardiac arrest [20] show that there is an increased risk of peri-operative cardiac arrest and death in patients who are older with comorbidity undergoing non-elective surgery. Based on the available data, we suggest that the highest risk group of patients would include any patient having surgery with: a clinical frailty scale score ≥ 5; ASA physical status 5; or an objective risk score for early mortality exceeding 5%.

In NAP7, all reported instances of peri-operative cardiac arrest in adult patients with a DNACPR recommendation occurred during non-elective surgery and often out-of-hours. On these occasions, time to speak with the patient, family members, close friends or legal proxies to ascertain patient values and preferences is often limited, and discussions may be hampered by the illness or injury requiring surgery. Even in elective cases, this remains a challenge for anaesthetists who might encounter the patient only shortly before a scheduled procedure. Such conversations are likely to involve intensivists as well as anaesthetists. Decisions to offer surgical treatment are related to, but distinct from, treatment escalation planning and will often be included in these discussions. The option of not proceeding with surgery has been highlighted by the Academy of Medical Royal Colleges in its Choosing Wisely initiative [21] and the Association of Anaesthetists in its human factors guidance for making time-critical decisions [22].

Although NAP7 is a comprehensive national study of peri-operative cardiac arrest over a period of one year, it has several limitations. In some cases, pre-operative discussions with patients and/or their families about their values and preferences may have resulted in a shared decision not to proceed with the surgery. However, we did not capture these data. Apart from those identified in the activity survey, we did not include those cases where a peri-operative cardiac arrest occurred, but CPR was not started. The outcome data are limited to those available at the time case reports were submitted and in many cases, patients were still in hospital and their survival to hospital discharge status is uncertain.

The NAP7 report includes recommendations for improving practice in relation to peri-operative treatment escalation plans (Box 1).

Box 1. Recommendations for improving practice in relation to peri-operative treatment escalation plans.

Institutional
  • Where practical, treatment escalation plans, including but not limited to ‘do not attempt cardiopulmonary resuscitation’ (DNACPR) recommendations, should be discussed and documented before arrival in the theatre complex in any patient having surgery with:

    • Clinical frailty scale ≥ 5
    • ASA physical status 5
    • Objective risk scoring of early mortality > 5%.

  • When appropriate, discussion should include the anaesthetic team.
  • In any patient presenting for surgery who has a clinical frailty scale score ≥ 5, discussions should take place as early as possible pre-operatively with 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.
  • Units should consider development of ‘high-risk patient’ bundles that create a person-centred approach to the management of patients who are peri-arrest and in whom treatment may be withdrawn in the immediate postoperative period.
Individual
  • When discussions take place around treatment planning, the patient's current or previously known wishes should be explored regarding which outcomes they value.
  • It is usually appropriate to suspend a pre-existing DNACPR recommendation in the peri-operative period. These discussions and decisions should be fully documented and discussed at the theatre team briefing.
  • If resuscitation is started, the patient's known wishes should be considered in deciding the extent of interventions undertaken (e.g. a patient may not wish to be in multiple organ failure on intensive care with little chance of surviving or recovering to their previous functional state).

Acknowledgements

The project infrastructure was supported financially and with staffing from the Royal College of Anaesthetists. Other NAP7 panel and team members are: C. Bouch; J. Cordingley; L. Cortes; M. T. Davies; J. Dorey; S. J. Finney; S. Kendall; G. Kunst; J. Lourtie; D. N. Lucas; R. Mouton; G. Nickols; V. J. Pappachan; B. Patel; F. Plaat; K. Samuels; B. R. Scholefield; J. H. Smith; C. Taylor; L. Varney; and E. Wain. We thank all NAP7 local reporters and their teams and all UK anaesthetists who completed surveys or submitted cases. The NAP7 fellows' salaries were supported by: South Tees Hospitals NHS Foundation Trust (AK); Royal United Hospitals Bath NHS Foundation Trust (EK); and NIHR Academic Clinical Fellowship (RA). JS and TC's employers receive backfill for their time on the project (4 hours per week). NAP7 panel members were not paid for their roles. IM is an Editor of Anaesthesia. No other competing interests declared.