key: cord-301623-uza6gu4r
authors: Moscarelli, Alessandra; Iozzo, Pasquale; Ippolito, Mariachiara; Catalisano, Giulia; Gregoretti, Cesare; Giarratano, Antonino; Baldi, Enrico; Cortegiani, Andrea
title: Cardiopulmonary resuscitation in prone position: A scoping review
date: 2020-09-10
journal: Am J Emerg Med
DOI: 10.1016/j.ajem.2020.08.097
sha: 
doc_id: 301623
cord_uid: uza6gu4r

INTRODUCTION: The ongoing pandemic of COVID-19 brought to the fore prone positioning as treatment for patients with acute respiratory failure. With the increasing number of patients in prone position, both spontaneously breathing and mechanically ventilated, cardiac arrest in this position is more likely to occur. This scoping review aimed to summarize the available evidence on cardiopulmonary resuscitation in prone position (‘reverse CPR’) and knowledge or research gaps to be further evaluated. The protocol of this scoping review was prospectively registered on 10th May 2020 in Open Science Framework (https://osf.io/nfuh9). METHODS: We searched PubMed, EMBASE, MEDLINE and pre-print repositories (bioRxiv and medRxiv) for simulation, pre-clinical and clinical studies on reverse CPR until 31st May 2020. RESULTS: We included 1 study on manikins, 31 case reports (29 during surgery requiring prone position) and 2 nonrandomized studies describing reverse CPR. No studies were found regarding reverse CPR in patients with COVID-19. CONCLUSIONS: Even if the algorithms provided by the guidelines on basic and advanced life support remain valid in cardiac arrest in prone position, differences exist in the methods of performing CPR. There is no clear evidence of superiority in terms of effectiveness of reverse compared to supine CPR in patients with cardiac arrest occurring in prone position. The quality of evidence is low and knowledge gaps (e.g. protocols, training of healthcare personnel, devices for skill acquisition) should be fulfilled by further research. Meanwhile, a case-by-case evaluation of patient and setting characteristics should guide the decision on how to start CPR in such cases.

This scoping review aimed to summarize the available evidence on the cardiopulmonary resuscitation in prone position ("reverse CPR") and to highlight possible knowledge or research gaps to be further evaluated.

For the purpose of this review, we searched PubMed, EMBASE, and MEDLINE for preclinical and clinical studies on prone CPR. Our search included the keywords "resuscitation", "CPR", "chest compression", "cardiopulmonary", "resuscitation", "prone position" as exact phrases and a combination of broad subject headings according to databases syntax. Specifically, the EMBASE engine was used with the following query: A search was also conducted on main pre-print repositories (bioRxiv and medRxiv) from inception to 31st May 2020 for relevant studies. No limitations were imposed for specific contexts, with the aim of including surgical, medical and intensive care settings. Articles on animals or on manikins were also eligible. Randomized controlled trials (RCTs), nonrandomized studies (both prospective or retrospective), case series and case reports were included. Abstracts and conference proceedings were excluded. Snowballing search on the references of selected articles was also performed.

After the exclusion of duplicates and abstracts, two authors (AM, PI) independently screened full-text papers to include the most relevant on the topic and independently charted data using an electronic standardized form. In case of case reports or series J o u r n a l P r e -p r o o f describing more than one patient, we collected data only on patients meeting inclusion criteria (i.e. cardiac arrest occurring in prone position and CPR performed).

We collected data regarding the type of study (e.g. design and country), population characteristics at baseline (e.g. age, main disease), setting (e.g. operatory room, ICU), occurred events (e.g. rhythm and cause of the cardiac arrest), intervention (e.g. prone or standard CPR) and outcomes (e.g. mortality, return of spontaneous circulation -ROSC).

Data were then tabulated for presentation, as appropriate.

The protocol of this scoping review was prospectively registered on 10 th May 2020 (https://osf.io/nfuh9).

The scoping review was conducted following PRISMA statement extension for scoping reviews [8] .

The initial search identified 1301 results from EMBASE, PubMed and other sources.

Following screening of titles and abstracts and removing duplicates, we evaluated 82 articles in full text. Among these, we selected and included 34 articles. The search from pre-print repositories (bioRxiv and medRxiv) resulted in 52 records screened, none included. The details on the inclusion/exclusion process are provided in the PRISMA flow diagram (see Fig.1 ).

We found no RCT comparing prone to standard supine CPR. Three of the included studies [9] [10] [11] had a nonrandomized design and 31 were case reports [12, 13, [22] [23] [24] [25] [26] [27] [28] [29] [30] [31] 14, [32] [33] [34] [35] [36] [37] [38] [39] [40] [41] 15, 42, [16] [17] [18] [19] [20] [21] . In this section, the included nonrandomized studies are described; a complete description of all the studies, including case reports, is provided with details in Table 1 .

The first study evaluating the feasibility of reverse CPR was a simulation study using Laerdal "ResusciAnne" manikins, posed in prone position on a standard examination J o u r n a l P r e -p r o o f coach, with a gel pad under the sternum. Thirty-six trained nurses were asked to perform 100 compressions on the manikin with no breaks for respirations. Using a skillmeter, a total of 3376 compressions (91.8% of the 3600 total possible) were registered, but only 1168 (34.6%) were effective (4 -5 cm compression depth), with 1370 (40.6%) partially effective (2 -4 cm compression depth) and 838 (24.6%) ineffective (< 2 cm compression depth) compressions. Only 9/36 nurses were able to perform 70% or more adequate compressions (usually considered as the acceptable threshold). An important insight was provided by the authors, discussing that all the nurses had judged CPR in the prone position as more tiring than in the standard position. This, in addition to the light weight of the manikin and the position of the nurses (left-hand side of the manikin), may explain the low rate of successful compressions and the number of compressions performed off the midline (11.2% to the right of midline, 4.1% too high, 6.1% too low, none to the left of the midline). None of the eligible studies were conducted on animals.

Among the nonrandomized studies on humans, one was conducted on 11 cadavers and 10 healthy volunteers, evaluating indirect outcomes (blood pressure and tidal volume) [9] . Mean arterial pressure (MAP) was invasively measured in cadavers both during standard and consecutive back chest compressions (55 ± 20 / 13 ± 7 vs. 79 ± 20 / 17 ± 10 p=0.028) performed at a rate of 60 per minute. Healthy volunteers" tidal volume was then measured using a mouthpiece connected to a spirometer, while they were receiving back chest compressions (60 min -1 ). Spontaneous breathing was held, and a nose clip applied on the volunteers. Mean registered tidal volume was of 399 ± 110 ml. The authors highlighted that their findings may support prone CPR in non-intubated patients, since airways open spontaneously, and adequate ventilation seems achievable with compressions only. Nonetheless, the study has limited external validity, especially for the finding regarding tidal volume (e.g. possibly not generalizable to patients with a compromised pulmonary function), requiring further investigations.

Another nonrandomized study enrolled six ICU patients in cardiac arrest, after the declared failure of standard CPR [10] . Main results included a systolic blood pressure mean improvement of 23 ± 14 mmHg, a calculated MAP mean improvement of 14 ± 11 mmHg and a diastolic blood pressure mean improvement of 10 ± 12 mmHg from standard to reverse CPR, but no cases of ROSC.

Despite such a limited basis, we retrieved a total of 31 case reports described in literature from 1982 (date of the first retrieved report) to date, for a total of 34 patients.

Among the included case reports, 29 described cardiac arrests occurred during surgery. Surgeries included spinal surgery in 20 patients (e.g. discectomies, scoliosis correction, vertebral metastases), craniotomy in 10 patients (e.g. primitive or metastatic cancer, hemorrhage), and cases of pelvic fracture and dorsal lipoplasty. One occurred in the ICU [33] in a prone positioned mechanically ventilated patient, admitted for community acquired pneumonia and acute respiratory failure. Another article reported a case of cardiac arrest during endoscopic retrograde cholangiopancreatography procedure [22] . No studies were found regarding reverse CPR in patients with COVID-19.

The feasibility of CPR in prone position has been under investigated. Its use has been described in the settings of both operatory rooms (e.g neurosurgery, orthopedics) and ICUs (e.g. mechanically ventilated patients with respiratory failure), mainly on case reports.

The main finding of this scoping review is that there is insufficient evidence on the topic and further evidence is needed, considering that an increasing number of cardiac arrests in prone position are expected during the COVID-19 pandemic.

J o u r n a l P r e -p r o o f Many aspects, both decisional and technical, have been described but remain controversial and need to be investigated with adequately designed studies.

To date, evidence on reverse CPR comes from case reports and small sized nonrandomized studies conducted on manikins, cadavers, healthy volunteers or patients with previously failed standard CPR. Among the included reports, ROSC after reverse CPR occurred in 23 out of 31 cases, demonstrating that the technique may be effective and deserves further investigations. Nevertheless, publication bias may exist, potentially overestimating the rate of favorable outcomes in prone CPR, and potentially limiting our findings. We did not find any RCT on the topic and nonrandomized studies only provide preliminary and indirect data to support the technique. Despite prone CPR may be a feasible and safe option in specific settings, further literature is needed, with many aspects remaining uninvestigated.

From the evaluated articles, it can be argued that a frequent cause of cardiac arrest in prone position is air embolism, a potentially fatal event, often occurring in neurosurgery and closely related to the position of the patient and the presence of a pressure gradient allowing air flow into the vessels. Other causes are linked with the specific patient positioning, which can sometimes lead to vessel occlusion and reduced venous return.

Hypovolemia in combination with reduced venous return can quickly lead to reduced cardiac output.

The rhythm presentation is various, ranging from asystole to pulseless electrical activity (PEA), or ventricular fibrillation (VF) .

In the event of a cardiac arrest in a prone positioned patient, the clinicians have limited time to decide whether to turn the patient into supine position, before starting CPR, or immediately start a reverse CPR. There is no clear evidence of superiority in terms of effectiveness of reverse compared to supine CPR in patients with cardiac arrest occurring in prone position. Thus, clinicians should take into account several factors when deciding how to start CPR in such cases. Turning the patient into supine position before starting CPR could be very difficult for several reasons: i) this maneuver is time consuming, and can potentially increase the no-flow time to the brain; ii) it may need at least 3-4 operators, not always available in surgical or ICU settings; iii) it could be dangerous for the patients, due to the presence of an open wound, protruding metal work, an unstable spine and fixed head to a Mayfield Skull Clamp potentially causing neurologic injuries, or iv) it could cause the dislodgment of the endotracheal tube and the loss of monitoring. Furthermore, the decision to perform supination and proceed with a standard CPR entails other risks, such as a difficult hemorrhage control (e.g. during spinal surgery) [30] . On the other hand, healthcare personnel might be reluctant to perform reverse CPR as a first option, especially due to the lack of specific training and knowledge of the procedure. Other concerns may regard the presence of an open surgical field, the limited surface to perform compressions, the need for a counterforce under the sternum and the risks of spinal damages. Notwithstanding, cardiac massage in this position is less likely to cause rib fractures, injury to the heart and aspiration pneumonia [9] .

Since 2005, the American Heart Association (AHA) Guidelines for CPR and Emergency Cardiovascular Care recommended that CPR in the prone position might be reasonable when the patient cannot be replaced in the supine position without prejudice, particularly in hospitalized patients with an advanced airway in place [1] .

J o u r n a l P r e -p r o o f AHA interim guidance for CPR in patients with COVID-19 recommends to attempt to place in the supine position for continued resuscitation of patients with suspected or confirmed COVID-19 who are in a prone position without an advanced airway, and to avoid turning the patient to the supine position if an advanced airway is secured, unless able to do so without risk of equipment disconnections and aerosolization [7] . Similar recommendations are also supported by the European Resuscitation Council (ERC) in its COVID-19 guidelines [43] , suggesting to turn patient supine in case of ineffective compressions (arterial line and aim for diastolic pressure greater than 25 mmHg), airway problems or unrestored circulation (after minutes) [43] .

A recent joint position statement from Brazilian societies of cardiology, intensive care medicine, anesthesiology and emergency medicine recommends to turn into supine position the patients once safely possible, due to insufficient evidences on reverse CPR and suggests monitoring using partial pressure of end-tidal carbon dioxide partial (EtCO 2 >10mmHg) and diastolic pressure (>20mmHg) [44] .

An early planning of supination is recommended, since it may require additional help [30, 44] . For the specific setting of neurosurgery, UK Resuscitation Council suggests that there is no immediate need to turn the patient to the supine position once cardiac arrest has occurred in the prone position, suggesting to start CPR in the prone position [45] . If the aforementioned techniques are not rapidly feasible or become unsuccessful, thoracotomy and direct internal cardiac massage and/or defibrillation are available options [30] .

Once a cardiac arrest is detected in a prone positioned patient, first of all it is important to make sure that airway is secured and the tube is not kinked, obstructed or displaced, in case of intubated patients. Then, it is needed to ensure that the ventilator is connected to the patient with 100% FiO 2 concentration and that there is no unidentified J o u r n a l P r e -p r o o f loss of blood [12] . If venous air embolism is suspected, the patient"s head should stay down, to increase venous pressure, and the surgical field should be flooded with saline.

After that, chest compressions should be started.

Unfortunately, only general indications are provided by the guidelines on how to perform cardiac massage in prone position [46] . Some are provided in AHA recent interim guidance for patients with COVID-19 [7] , ERC provides some other practical indications [43] , suggesting compressing between the scapulae (shoulder blades) at the usual depth and rate (5 to 6 cm at 2 compressions per second) and placing defibrillator pad in anterior-posterior (front and back) or bi-axillary (both armpits) position (Fig. 2) . Alternatively, defibrillator pads can be applied in posterolateral position [29] , i.e. one pad in the left mid-axillary line and a second one over the right scapula, or over the left scapula (Fig. 2) [19] . The effectiveness of reverse CPR also depends on the greater strength of thoracic costovertebral joints as compared to sternal costochondral junction and on the absence of abdominal contents displaced anteriorly and dissipating the compression force [15] .

compressions. An option involves the use of two hands together at the midline at the midthoracic level (Fig. 3, panel a) . If sternal support is present, rescuer"s hands can be placed on either side of the eventual incision at the midthoracic level, with the palms placed over the patient"s scapulae, thus preserving the sterility of the surgical field (Fig. 3, 

A one-handed technique can be performed without sternal supports, with the flat of one hand (or with the hand clinched into a fist) on the lower third of the sternum used as counter pressure and the second one on the thoracic spine, and could be performed by one or two physicians (Fig. 3, panel c) [33] .

In the pediatric population cardiac massage can be started in the prone position using the fingers of one hand over the thoracic vertebral column at the level of the scapulae [35] with the same rate and force as they were delivered during supine position.

The COVID-19 pandemic represents a new challenge for clinicians in many fields, Meanwhile, the available CPR guidelines should be followed.

Reverse CPR has been performed and described in several settings, but evidence is based on case reports and preliminary small sized nonrandomized studies. The procedure has been described in settings such as neurosurgery and prone position mechanical ventilation, and an increasing number of cardiac arrests in prone positioned patients is expected in the COVID-19 era. The quality of available evidence is low and knowledge gaps should be fulfilled by further adequately designed studies that are urgently needed. Meanwhile, a case-by-case evaluation of patient and setting characteristics should guide the decision on how to start CPR in such cases. 

American Heart Association guidelines update for cardiopulmonary resuscitation and emergency cardiovascular care

European Resuscitation Council Guidelines for Resuscitation

Use of extreme position changes in acute respiratory failure

Helmet continuous positive airway pressure and prone positioning: A proposal for an early management of COVID-19 patients

Feasibility and physiological effects of prone positioning in non-intubated patients with acute respiratory failure due to COVID-19 (PRON-COVID): a prospective cohort study

COVID-19 kills at home: the close relationship between the epidemic and the increase of out-of-hospital cardiac arrests

Interim Guidance for Basic and Advanced Life Support in Adults, Children, and Neonates With Suspected or Confirmed COVID-19

PRISMA extension for scoping reviews (PRISMA-ScR): Checklist and explanation

Cardiopulmonary J o u r n a l P r e -p r o o f Journal Pre-proof resuscitation in prone position: A simplified method for outpatients

Reverse CPR: A pilot study of CPR in the prone position

The efficacy of cardiopulmonary resuscitation in the prone position

Prone cardiopulmonary resuscitation in elderly undergoing posterior spinal fusion with laminectomy

Cardiac arrest in the prone position caused by central venous cannulation-induced cardiac tamponade

Cardiac pulmonary resuscitation in prone position. The best option for posterior fossa neurosurgical patients

CPR in prone position during neurosurgery

Venous air embolism during removal of bony spur in a child of split cord malformation

Air Embolism During Posterior Spinal Fusion in a 10-Year-Old Girl: A Case Report

Cardiac arrest after epidural anesthesia for a esthetic plastic surgery: a case report

Cardiac arrest during craniotomy in prone position

Cardiopulmonary Resuscitation in the Prone Position

Prone CPR for transient asystole during lumbosacral spinal surgery

Fatal air embolism during endoscopic retrograde cholangio-pancreatography (ERCP): A case report

Venous air embolism: An unusual complication of atlantoaxial arthrodesis: Case report

Erfolgreiche kardiopulmonale Reanimation in Bauchlage

Case Studies of Near Misses in Clinical Anesthesia

Unsuccessful cardiopulmonary resuscitation during neurosurgery: Is the supine position always optimal?

Cardiac arrest during desflurane anaesthesia in a patient with Duchenne"s muscular dystrophy

Successful defibrillation in the prone position

Cardiac arrest during surgery and ventilation in the prone position: a case report and systematic review

A case of ventricular fibrillation in the prone position during back stabilisation surgery in a boy with Duchenne"s muscular dystrophy

Two cases of fatal air embolism in children undergoing scoliosis surgery

Cardiopulmonary resuscitation in the prone position: Kouwenhoven revisited

Non-invasive continuous haemodynamic and PETCO 2 monitoring during peroperative cardiac arrest

Cardiac arrest and resuscitation of a 6-month old achondroplastic baby undergoing neurosurgery in the prone position

Intraoperative Cardiopulmonary Resuscitation in the Prone Position

Efficacit du massage cardiaque externe chez une patiente en ddcubitus ventral

Successful cardiopulmonary resuscitation of two patients in the prone position using reversed precordial compression

Venous Air Embolism During Lumbar Laminectomy in the Prone Position: Report of Three Cases

Intraoperative death during lumbar discectomy

Sitting or prone? Another argument for the latter

Fatal Air Embolism in an Adolescent with Duchenne Muscular Dystrophy during Harrington Instrumentation

Education: European Resuscitation Council COVID-19 Guidelines

Posicionamento Posicionamento para Ressuscitação Cardiopulmonar de Pacientes com Diagnóstico ou Suspeita de COVID-19-2020

Management of cardiac arrest during neurosurgery in adults

Prone Cardiopulmonary Resuscitation: A Rapid Scoping and Expanded Grey Literature Review for the covid-19 Pandemic

and 838 (24.6%) ineffective Reid et al. (1998) [31] Case report