Attempt CPR—language matters inside our hospitals

Cardiopulmonary resuscitation (CPR) was initially defined as chest compressions and rescue breaths to manually empty the heart, offering a degree of cerebral perfusion while the cause of cardiac arrest is reversed.¹ In the United Kingdom (UK), this original definition has been replaced. CPR now includes high-voltage electric shocks, ventilation and injection of drugs,¹ and decisions documented in the ReSPECT (Recommended Summary Plan for Emergency Care and Treatment) process.² Most other international resuscitation councils, including the New Zealand Resuscitation Council, retain the original definition of chest compressions and rescue breaths (Table 1). Despite this, almost all healthcare practitioners in New Zealand include broader resuscitation measures in their understanding of the term CPR when talking with each other and patients.³

This international variance in terminology means that “CPR” and “resuscitation” are used interchangeably and non-specifically.³ This is particularly significant when considering cardiac arrest in hospitals. CPR is not a definitive treatment. It is part of the broader range of measures that constitute resuscitation, and in a patient with an easily reversible cause of cardiac arrest, bridges a patient to those measures.

In the hospital, most cardiac arrests are not sudden or shockable and follow a period of clinical deterioration. CPR, therefore, has questionable utility, particularly after attempting other resuscitative measures during the peri-arrest state. We will argue that CPR has little value here and withholding it (do-not-attempt-CPR) can cause harm in terms of limiting access to other treatments. Returning to the original definition allows patients access to other resuscitative measures to prevent further deterioration into cardiac arrest while protecting them from harmful interventions at the end of life.

View Table 1–4, Figure 1.

Cardiac arrest is different in hospital

Cardiac arrest in hospitals has poor survival in most locations.⁴ Patients are already being treated for medical or surgical conditions, are located in general wards, are mostly unmonitored and have low rates of shockable arrhythmias.⁵ Cardiac arrest incidence in hospitals peaks at 48 hours¹ and follows a period of deterioration heralded by a change in patient vital signs in 60% of cases.⁶ Such conditions are in marked contrast to cardiac arrest that occurs out-of-hospital, or in-hospital in operating theatres and coronary care units; here cardiac arrest is sudden, and CPR is a vital bridge to defibrillation or other treatment for reversible causes.⁴

It is time to be more explicit about the difference between sudden cardiac arrest, when CPR may be an effective first aid intervention, and subacute deterioration. For the latter, cessation of cardiac activity occurs at the point of death despite other attempts at treatment, and often resuscitation.

The latter process is better referred to as natural dying, and in a hospital it can sometimes be a rapid process. Clinicians and patients should appreciate natural dying and its distinction from sudden cardiac arrest. Natural dying occurs when illness or injury exceeds the person’s ability to recover, often despite intensive treatments, leading to the secondary cessation of circulation. In the latter situation, where CPR is deployed as the final resuscitation manoeuvre before attempts are halted, CPR serves no medical purpose. It may also be contrary to patient-defined preferences or treatment goals.

Reframing CPR as chest compressions and rescue breaths will provide transparency to clinicians and patients about what treatments are warranted and wanted if clinical deterioration occurs in the hospital. It also affords clinicians clarity in situations where communication and decision-making are complex and emotionally charged. It offers an individualised plan in the event of deterioration, with nuanced resuscitative measures, like drugs, intravenous fluids, intensive care treatment or defibrillation for monitored arrhythmias. Most importantly, it can incorporate a focus on comfort care without intrusive interventions for dying patients.

CPR benefits only a specific minority of hospitalised patients (mechanism of deterioration)

For out-of-hospital cardiac arrest, the time to initiation of CPR is a key determinant of the outcome, but in hospitals, the relationships are more complex. Favourable outcomes are more strongly linked to the aetiology of cardiac arrest (Table 2) and patient type/location (Figure 1) rather than treatment delay (adjusted odds ratio [OR] for 30-day survival 2.81 for monitored wards [95% CI 2.63–3.01] vs 0.55 for delay in calling [95% CI 0.50–0.61]).⁴

The 1-year survival for in-hospital cardiac arrest for cardiac patients is 39.3% versus 10.7% for non-cardiac patients.⁷ Causes of in-hospital cardiac arrest with good outcomes are those that are highly reversible, such as general anaesthesia, intoxication and hypothermia, with survival rates of around 50%.⁸

Conversely, CPR adds little value when cardiac arrest has occurred due to catastrophic diseases such as aortic dissection/rupture or intracerebral haemorrhage (Table 3),⁸ or when the cause cannot be treated within the timeframe of cardiac arrest, such as exsanguination or aortic stenosis.^8,9

The cost of survival for those who survive

Proponents of the drive towards “catch-all CPR first, think later” algorithmic management of in-hospital cardiac arrest will point to the improvements in survival over the last 20 years for all types of in-hospital cardiac arrest documented in both large United States of America (US)¹² and Swedish registry data⁵ (Table 3). However, cardiac arrest survival is highest in countries with the lowest incidence (Table 3). Given the international standardisation of care, it seems likely that patient selection or possibly attitudes towards resuscitation substantively impact survival. Furthermore, adjusted 30-day survival only improved in younger patients, whereas older patients (>85 years) experienced no improvement.⁵ It is important to emphasise to proponents of CPR that system improvements such as time to treatment—mainly defibrillation and adrenaline—are also driving survival gains more than chest compressions.^13,14

For survivors, recovery is complex. Despite 79.5% of US in-hospital cardiac arrest survivors attaining CPC category 1–2,¹² 40% have life-changing disability after discharge⁵ and only half are discharged home. This harm is not readily available in the statistics.¹⁵ Longer-term survival continues to decline after discharge, particularly for general ward patients compared with those on monitored wards such as coronary care and operating theatres (Figure 1).⁵

Monitored wards include coronary care, intensive care, operating theatres, cardiac catheter laboratories and emergency departments.

CPR can be harmful

Closed chest cardiac massage was the natural evolution from open cardiac massage conducted through an emergency thoracotomy. Although simple to administer, closed chest cardiac massage only produces about 20–40% of the usual cardiac output¹⁸ and causes rib fractures, with older age as a risk factor.¹⁹ It can also cause life-threatening injuries,²⁰ and this could contribute to the high (60%) incidence of death following the initial return of spontaneous circulation (RoSC).^13,19

Survival rates for CPR in subacute decline are low

Deciding which patients are likely to benefit and should opt in for “attempt CPR” requires three separate considerations: patient factors, with frailty being more important than age (Table 4), the aetiology of illness and cardiac arrest as outlined above and whether that arrest has occurred despite optimal treatment. Rates of survival after cardiac arrest in sepsis and all patients in intensive care units are low (Table 2).^8,10 In subacute decline, treatment has often already been optimised.

While RoSC occurs in around 50% of in-hospital cardiac arrests, that reprieve is often brief before death ensues.¹ Although injuries related to CPR and resuscitation may contribute, it is more likely that deterioration occurs in a patient not responding to treatment, and cardiac arrest results as the end-point of natural dying.¹ Over 60% of patients who survive their initial cardiac arrest will subsequently have active treatment withdrawn or limited and die naturally during the same hospital admission.¹³

DNACPR should only limit CPR, not halt other forms of resuscitation in hospital

DNACPR decisions must not compromise the quality of care for any patient. Decisions to prevent CPR should not be misinterpreted as an unofficial stop sign to other appropriate interventions. DNACPR decisions were associated with doubling the risk of death at 30 days for intracranial haemorrhage (OR 2.17 95% CI 1.38–3.41) despite adjustment for disease and age.²¹

Narrowing the definition of CPR to chest compressions and rescue breaths clarifies the practice, enabling clinicians to offer other treatments and resuscitative measures as appropriate.

This is enabled by the New Zealand national Shared Goals of Care form (Appendix), where individual levels of care can be ascribed between four options. Shared Goal of Care A is where the patient is for treatment with curative or restorative intent and includes CPR. Shared Goal of Care B is where the patient is for treatment with curative or restorative intent B but excludes CPR. Option C focusses on improving symptoms with non-burdensome treatments, and D cares for the dying patient. Patients who have been ascribed a Shared Goal of Care B can still access other resuscitative measures in the peri-arrest period aimed at halting their deterioration into cardiac arrest.

Communication with patients and families about CPR in hospital

Despite clear benefits and extensive training, healthcare teams remain hesitant to have comprehensive end-of-life or resuscitation discussions. In a study of patients with advanced cancer recruited to examine the impact of these conversations, only 37% of patients reported prior discussions.²⁴ A review of calls requesting medical emergency advice for hospital inpatients suggests that a third (estimated 30,000 calls per year in New South Wales, Australia) were for end-of-life issues in elderly frail patients who were often unaware that they were dying.²⁵ Initiatives such as ReSPECT in the UK demonstrate similar low engagement with discussions (6–41% of patients admitted to test sites had a ReSPECT form).²

The preferences of patients and their families for cardiac arrest treatment requires clinical support beyond asking a patient if they would like their “heart restarted”. This approach undersells the complexity of treatment and does not represent shared decision-making. Patient selections are often confused and paradoxical. Treatment can be complicated and dynamic. For example, respondents to a treatment preference survey indicated that 74% wished to receive chest compressions, whereas only 61% selected defibrillation and 42% ventilation.²⁶ Over-estimation of surviving CPR is common and imagined to be between 55%.²⁶ This is perhaps unsurprising due to the misrepresentation of outcomes by clinicians, but also in television and other media.²⁷

Initiatives such as ReSPECT and the national Shared Goals of Care programme in New Zealand encourage early discussions. Clinicians ultimately lead the medical decision-making process in these programmes. Although patient engagement has improved, qualitative interviews suggest a culture of informing rather than partnering with patients in these discussions, especially for CPR conversations.²⁸ An example from interview-based analysis, one consultant said:

“I went in with quite clear views of what had to be done and as you say the patient’s son started to suggest that ‘actually he would want to be resuscitated wouldn’t you Dad’… and I gently had to steer him away to explain why I didn’t think that would be a good idea.”²⁸

In the US, stronger tactics with graphic descriptions of resuscitation are common.²⁹ One doctor’s comment in a study of DNAR discussions illustrates how choice of language manipulates patient autonomy:

“This is kind of paternalistic, but if I feel strongly that the patient wouldn’t benefit from resuscitation, I’ll be pretty graphic ... I want you to know we have to press really hard and break ribs.”²⁹

The culture of conversations in New Zealand is likely to be similar and may use graphical illustrations to manipulate patient requests or use medical leading of the conversation. Although these practices are not illegal, the opportunity to bring the person and their goals and values into the conversation is completely lost.

Better CPR conversations

The ideal process in hospital is a discussion centred solidly around the patient’s goals, and frameworks such as the serious illness conversation guide³⁰ have been adopted and regionalised for use in New Zealand. Although these frameworks derive broader treatment goals than just CPR, the information gathered from a structured approach will help the clinician better understand what matters to the individual and their whānau. This enables the clinician to make a clinical recommendation incorporating patient-centred goals alongside factors for successful CPR.

For the patient, articulating their goals should be followed with clear explanations of the specific, medically indicated treatments that will support those goals, in the event of deterioration.

Partnering with patients to improve their understanding of CPR as a bridging measure could shift emphasis from graphic illustrations of resuscitation to a more open and honest discussion about what may lie ahead.

Conclusion

The limited scope of CPR (chest compressions and rescue breaths) in hospitalised patients is poorly recognised by the public (and possibly clinicians); a conversation about this is overdue.

We argue that non-specific language around CPR complicates how well it is understood and discussed. Many hospitalised patients at risk of subacute deterioration towards anticipated dying have risk factors for CPR-related injuries and low survival after ROSC. Their treatment should be focussed on preventing cardiac arrest with other measures while preparing to meet their care wishes during dying. Reframing CPR as its original concept and uncoupling it from resuscitation may improve how clinicians discuss treatment options and enable patients to recognise that beneficial treatments are not being withheld.

View Appendix.