Cerebral oximetry levels during CPR are associatedwith return of spontaneous circulation followingcardiac arrest: an observational studyAdam J Singer, Anna Ahn, Loren A Inigo-Santiago, Henry C Thode Jr, Mark C Henry,Sam Parnia

Resuscitation Research Group,Departments of EmergencyMedicine and Medicine, StonyBrook University, Stony Brook,New York, USA

Correspondence toDr Adam J Singer, Departmentof Emergency Medicine, HSC-L4-080, 8350 SUNY, StonyBrook, NY 11794-8350, USA;adam.singer@stonybrook.edu

Presented at the AnnualMeeting of the Society forAcademic Emergency Medicine,Atlanta, GA, May 2013.

Received 2 December 2013Revised 24 January 2014Accepted 1 February 2014

To cite: Singer AJ, Ahn A,Inigo-Santiago LA, et al.Emerg Med J PublishedOnline First: [please includeDay Month Year]doi:10.1136/emermed-2013-203467

ABSTRACTObjectives Cerebral oximetry using near-infraredspectroscopy measures regional cerebral oxygensaturation (rSO2) non-invasively and may provideinformation regarding the quality of cerebral oxygenperfusion. We determined whether the level of rSO2

obtained during cardiopulmonary resuscitation isassociated with return of spontaneous circulation (ROSC)and survival in Emergency Department (ED) patientspresenting with cardiac arrest.Methods We conducted a retrospective, observationalstudy of adult ED patients presenting at an academicmedical centre with cardiac arrest in whom continuouscerebral oximetry was performed. Demographic andclinical data including age, gender, presenting rhythmand mean rSO2 readings were abstracted. Cerebraloxygenation was measured with a commercially availableoximeter.Results A convenience study sample included 59patients ages 18–102 years (mean age 68.7 ± 14.9years); 50 (84.7%) were men. Presenting rhythmsincluded pulseless electrical activity (21), asystole (20)and ventricular fibrillation/tachycardia (17). 24 patients(40.6%) had ROSC and only 1 (1.7%) survived tohospital discharge. Patients with and without ROSC weresimilar in age and presenting cardiac rhythms. The meanof mean rSO2 levels was higher in patients with ROSC,43.8 (95% CI 40.1 to 47.6) compared with thosewithout ROSC, 34.2 (95% CI 30.6 to 37.8); p=0.001.91.7% of patients with ROSC had a rSO2 of 30% orgreater compared with 62.9% in those without ROSC(p=0.01). The area under the curve for mean rSO2 as apredictor of ROSC was 0.76 (95% CI 0.64 to 0.89).Conclusions In ED patients with cardiac arrest highercerebral oxygen saturations are associated with higherrates of ROSC.

INTRODUCTIONEach year there are over 350 000 deaths fromsudden cardiac arrest (SCA) in the USA alone,which is more than the total number of deathsfrom Alzheimer’s’ disease, assault with firearms,breast cancer, cervical cancer, colorectal cancer, dia-betes, HIV, house fires, motor vehicle accidents,prostate cancer and suicides combined.1 Despitemajor advances in the understanding of SCA mor-tality remains high with less than 1 of 10 victims ofout-of-hospital SCA surviving to hospital dis-charge.2 Even among survivors, many are left withsignificant neurological and cognitive deficits.

In order to maximise the effectiveness of cardio-pulmonary resuscitation and increase the likelihoodthat victims of SCA will survive without majorneurological deficits, it is important to be able tomonitor the balance between cerebral oxygendemand and delivery in order to optimise cerebralperfusion during and immediately after cardiacarrest. Cerebral oximetry by near-infrared spectros-copy is a non-invasive optical monitoring methodthat measures regional cerebral oxygen saturation(rSO2) providing a real time indicator of thebalance between cerebral oxygen demand andsupply.3 A recent study at our centre focusing onpatients who arrested on an inpatient floor afterbeing admitted to the hospital has demonstratedthat cerebral oximetry is feasible during inhospitalcardiac arrest and does not interfere with resuscita-tive efforts.4 Patients with return of spontaneouscirculation (ROSC) as well as survivors had signifi-cantly higher overall mean rSO2. However, thefeasibility and role of cerebral oximetry in theEmergency Department (ED) has not yet beendetermined. In the current study we describe ourexperience with and the feasibility of real time non-invasive cerebral oximetry in ED patients without-of-hospital SCA and explore the associationbetween rSO2 and ROSC and survival in this studypopulation.

Key messages

What is already known on this subject?Survival from cardiac arrest is poor and dependenton adequate perfusion of the heart and brain.Currently there are few objective guides formonitoring and optimizing cerebral perfusion incardiac arrest.

What this study adds?This study demonstrates that it is feasible tocontinuously monitor cerebral oxygenation andperfusion during resuscitation. It alsodemonstrates an association betweenimproved cerebral oxygenation and return ofspontaneous circulation in emergencydepartment patients presenting after out-of-hospital cardiac arrest. Optimization ofresuscitation and improved outcomes may beguided by cerebral oximetry.

Singer AJ, et al. Emerg Med J 2014;0:1–4. doi:10.1136/emermed-2013-203467 1

Original article EMJ Online First, published on March 24, 2014 as 10.1136/emermed-2013-203467

Copyright Article author (or their employer) 2014. Produced by BMJ Publishing Group Ltd under licence.

group.bmj.com on November 12, 2014 - Published by http://emj.bmj.com/Downloaded from

METHODSStudy designWe conducted a structured retrospective chart review of EDpatients undergoing cardiopulmonary resuscitation (CPR) in ourED after out-of-hospital SCA in whom continuous cerebral oxim-etry was performed. Given the retrospective nature of our studythe Institutional Review Board determined that the study wasexempt from informed consent. We have previously incorporatedcerebral oximetry into our clinical practice although use of cere-bral oximetry is at the discretion of the attending physician.

Setting and patientsThe study was conducted in an academic, suburban ED with anaffiliated residency programme in emergency medicine and anannual ED census of 90 000. Adult patients arriving in the EDwith SCA in whom continuous cerebral oximetry was recordedduring resuscitation were included in the study. Patients whoarrived after ROSC were not included. In our institutionmanual CPR is initiated on patient arrival. In patients in whomcerebral oximetry readings remain below 50%, mechanical com-pressions are performed using a ‘Thumper’ device (Life Stat,Michigan Instruments, Michigan, USA).

Measures and outcomesDemographic and clinical data were collected using standardiseddata collection forms conforming to Utstein recommendations.5

The primary outcome of the study was ROSC. A secondaryoutcome was survival to hospital discharge.

Cerebral oximetryCerebral oximetry was measured during cardiac arrest using non-invasive infrared spectroscopy (Equanox Nonin, Plymouth,Minnesota, USA). After arrival in the ED the patient was trans-ferred to an ED gurney and CPR was continued by the ED staff.After replacement of cardiac pacing/defibrillation pads a cerebraloximetry sensor pad was placed on the patient’s forehead by anED nurse or physician who was not performing CPR (figure 1)and monitoring was carried out continuously with the brainoximetry monitor until ROSC or CPR was terminated. With theoximetry device, emitted light is scattered by tissues in two para-bolic curves, one corresponding to haemoglobin saturation fromthe skin and skull and the other from the skin, skull and frontalcortex.6 The haemoglobin saturation on the surface of the frontalcortex is calculated using specific computerised algorithms.

For each resuscitation, the complete rSO2 data for eachpatient (sampled every 4 s) was taken from the time the sensorhad been placed on the patient until either ROSC had beenachieved or until CPR had been terminated. Artefact valueswere recognised by either an absent value or values that were atleast three SDs away from the mean rSO2. This was used to cal-culate the mean rSO2 during the entire resuscitation periodafter arrival in the ED.

Data analysisContinuous data were summarised with means and SDs since meanrSO2 was approximately normally distributed. Binary data weresummarised as the percentage frequency of occurrence. Mean rSO2

was compared between those with and without ROSC andbetween survivors and non-survivors using the t test. The receiveroperating characteristic (ROC) characteristics of sensitivity, specifi-city, and positive and negative predictive values and the area underthe curve (AUC) with 95% CI for mean rSO2 as a predictor ofROSC were also calculated. A sample size of 25 patients in each

group had 80% power to detect a 15% difference in the rSO2

between the groups with a statistically significant level of 0.05.

RESULTSBetween January 2012 and July 2013 there were 59 patients whopresented to the ED in cardiac arrest and had continuous cerebraloximetry readings (table 1). Their mean (SD) age was 68.7 (14.9)

Table 1 Summary of patient characteristics.

Mean age (SD), years 68.7 (14.9)Men, No. (%) 50 (84.7)Location of cardiac arrest: outside of hospital, No. (%) 59 (100)Initial cardiac rhythm, No. (%)Pulseless electrical activity 21 (35.6)Asystole 20 (33.9)Ventricular fibrillation/tachycardia 17 (28.8)

ROSC, No. (%) 24 (40.6)Survival to hospital discharge, No. (%) 1 (1.7)

ROSC, return of spontaneous circulation.

Figure 1 Placement of a cerebral oximetry sensing pad on thepatient’s forehead.

2 Singer AJ, et al. Emerg Med J 2014;0:1–4. doi:10.1136/emermed-2013-203467

Original article

group.bmj.com on November 12, 2014 - Published by http://emj.bmj.com/Downloaded from

ranging from 18 years to 102 years; 50 (84.7%) were men.Presenting rhythms included pulseless electrical activity (21),asystole (20) and ventricular fibrillation/tachycardia (17).Twenty-four patients (40.6%) had ROSC and only 1 (1.7%) sur-vived to hospital discharge. Since all patients initially presentedwith cerebral oximetry readings below 50%, mechanical com-pressions were performed in all patients. Patients with andwithout ROSC were similar in age (68.0 years vs 69.2 years), percent male (79.2% vs 88.6%) and presenting cardiac rhythms.The mean of mean rSO2 levels was higher in patients withROSC, 43.8 (95% CI 40.1 to 47.6) compared with thosewithout ROSC, 34.2 (95% CI 30.6 to 37.8); p=0.001 (figure 2).Of patients with ROSC 91.7% had a rSO2 of 30% or greatercompared with 62.9% in those without ROSC (p=0.01). TheROC characteristics of rSO2 using a cut-off of 30% were 91.7%(sensitivity), 37.1% (specificity), 50% (positive predictive value)and 86.7% (negative predictive value). The mean (95% CI)duration of the resuscitation was similar in those with andwithout ROSC (13 (8–19) min vs 19 (14–25) min respectively;

p=0.11). The AUC for mean rSO2 as a predictor of ROSC was0.76 (95% CI 0.64 to 0.89) (figure 3).

A representative monitor recording depicting the relationshipbetween cerebral oximetry and ROSC is presented in figure 4.

DISCUSSIONOur results demonstrate that continuous monitoring of cerebraloximetry using non-invasive infrared spectroscopy during CPRin ED patients presenting with out-of-hospital SCA is feasibleand does not interfere with or interrupt resuscitative efforts.Placement of the brain oximetry sensor pad is similar to place-ment of a pulse oximetry pad. Our study also shows that ROSCis associated with higher mean rSO2 and that ROSC rarely occurswhen rSO2 remains below 30%. Unfortunately, since only one of

Figure 2 Box plot of regional cerebral oxygen saturation (rSO2) byreturn of spontaneous circulation (ROSC).

Figure 3 The area under the curve for mean regional cerebral oxygensaturation as a predictor of return of spontaneous circulation.

Figure 4 Representative monitorrecording of cerebral oxygenation in apatient with cardiac arrest. The suddenrise in regional cerebral oxygensaturation immediately preceded returnof spontaneous circulation (ROSC).

Singer AJ, et al. Emerg Med J 2014;0:1–4. doi:10.1136/emermed-2013-203467 3

Original article

group.bmj.com on November 12, 2014 - Published by http://emj.bmj.com/Downloaded from

the patients in this series survived to hospital discharge we couldnot statistically assess the association between rSO2 and survival.

Our results are in agreement with several prior studies investi-gating the use of cerebral oximetry in cardiac arrest. A smallstudy on patients with out-of-hospital cardiac arrest found thatall survivors had higher cerebral oxygenation readings than non-survivors (63% vs 46%, p=0.003).6 A study of 16 patients without-of-hospital cardiac arrest found that all four subjects withROSC had detectable cerebral oximetry readings while in thosewith no ROSC cerebral oximetry readings were undetectable.7

They concluded that standard CPR rarely provides detectableoxygenation of the brain. In our centre, the updated AmericanHeart Association guidelines emphasising faster and deeperchest compressions, a lower number of ventilations and minimalinterruptions in CPR are followed, likely leading to higheroximetry readings than those reported by Newman et al.7 Inaddition, we also incorporated mechanical compressions (LifeStat, Michigan Instruments, Michigan, USA) when cerebraloximetry readings remained low with notable improvements incerebral oxygenation (not presented). In another study of 33patients with out-of-hospital cardiac arrest all had readingsbelow 15% and none survived.8

However a recent study by our group in 50 patients withinhospital cardiac arrests found that rSO2 may predict ROSC.Compared with those who did not achieve ROSC, patients withROSC had significantly higher levels of mean rSO2 (47.2%±10.7 vs 31.7±12.8%, p<0.0001).9 Our study in the ED sup-ports the findings from this study and suggests that when insti-tuted during the resuscitative process, cerebral oximetry mayhelp predict ROSC as well as gauge the adequacy of cerebralperfusion during CPR allowing adjustments in CPR that maylead to improvements in brain oxygenation with subsequentimprovements in survival and neurological outcomes. Finally, itis unclear why the survival rate in our study was so low. It ispossible that the rate was low since we only included patientswho presented to the ED still in cardiac arrest despite relativelylong durations of attempted CPR en route to the hospital. Also,many of the patients in our series may have had long orunknown downtimes.

Our study has several limitations. First, it only includespatients who presented to the hospital without ROSC. Thisselection bias may have lead to underestimation of the sensitivityof cerebral oximetry for the primary outcome. On the otherhand, exclusion of patients in whom cerebral oximetry was notmeasured may have lead to overestimation of its SEN. Second,the sample size was small and there was only one survivor tohospital discharge and thus we were unable to assess the associ-ation between rSO2 and survival. Third, the device used tomeasure cerebral oxygenation only measures oxygenation in thefrontal cortex. However, since cerebral blood flow and oxygen-ation are globally reduced in cardiac arrest we believe thatoxygen concentrations in the frontal cortex are representative ofthe rest of the brain. Another limitation of this study is that ascerebral oximetry was being used in the hospital but not by theambulance crews, only data from the time of arrival to the EDrather than the true onset of cardiac arrest or CPR was

available. We also could not control for many potential con-founding factors that may have affected cerebral oxygenation,ROSC, and survival. Finally, our study is limited to a singlecentre that may not be representative of other medical settings.

In conclusion, our study demonstrates the feasibility of realtime monitoring of cerebral oxygenation during CPR using non-invasive infrared spectroscopy in the ED. While rSO2 is asso-ciated with ROSC, there was only one survivor in our sample of59 ED patients with SCA. Non-invasive cerebral oximetry mayhave a role in ED resuscitations by helping practitioners gaugethe effectiveness of their interventions and helping to predictROSC and survival. Future multicentre studies of consecutivepatients will help clarify the role of cerebral oximetry.

Contributors The research idea was conceived by AJS, SP. AJS, MCH, SP,participated in study design. Data collection was performed by AA and LAI-S. AJSdrafted the first manuscript. HCT was responsible for statistical analysis. All authorsreviewed the final manuscript and approved it.

Competing interests None.

Ethics approval IRB.

Provenance and peer review Not commissioned; externally peer reviewed.

REFERENCES1 Go AS, Mozaffarian D, Roger VL, et al. Heart disease and stroke statistics–2013

update: A report from the American Heart Association. Circulation2013;127:143–52.

2 Roger VL, Go AS, Lloyd-Jones DM, et al. Heart disease and stroke statistics-2012update: A report from the AHA. Circulation 2012;125:188–97.

3 Pollard V, Prough DS, DeMelo AE, et al. Validation in volunteers of a near-infraredspectroscope for monitoring brain oxygenation in vivo. Anesth Analg1996;82:269–77.

4 Parnia S, Nasir A, Shah C, et al. A feasibility study evaluating the role of cerebraloximetry in predicting return of spontaneous circulation after cardiac arrest.Resuscitation 2012;83:982–5.

5 Peberdy MA, Cretikos M, Abella BS, et al. International Liaison Committee onResuscitation; American Heart Association; Australian Resuscitation Council; EuropeanResuscitation Council; Heart and Stroke Foundation of Canada; InterAmerican HeartFoundation; Resuscitation Council of Southern Africa; New Zealand ResuscitationCouncil; American Heart Association Emergency Cardiovascular Care Committee;American Heart Association Council on Cardiopulmonary, Perioperative, and CriticalCare; Interdisciplinary Working Group on Quality of Care and Outcomes Research.Recommended guidelines for monitoring, reporting, and conducting research onmedical emergency team, outreach, and rapid response systems: an Utstein-stylescientific statement: a scientific statement from the International Liaison Committeeon Resuscitation (American Heart Association, Australian Resuscitation Council,European Resuscitation Council, Heart and Stroke Foundation of Canada,InterAmerican Heart Foundation, Resuscitation Council of Southern Africa, and theNew Zealand Resuscitation Council); the American Heart Association EmergencyCardiovascular Care Committee; the Council on Cardiopulmonary, Perioperative, andCritical Care; and the Interdisciplinary Working Group on Quality of Care andOutcomes Research. Circulation 2007;116:2481–500.

6 Mullner M, Sterz F, Binder M, et al. Near infrared spectroscopy during and aftercardiac arrest-preliminary results. Clin Intensive Care 1995;6:107–11.

7 Newman DH, Callaway CW, Greenwald IB, et al. Cerebral oximetry in out-of-hospitalcardiac arrest: standard CPR rarely provides detectable hemoglobin-oxygen saturationto the frontal cortex. Resuscitation 2004;63:189–94.

8 Ito N, Nanto S, Nagao K, et al. Regional cerebral oxygen saturation predicts poorneurological outcome in patients with out-of-hospital cardiac arrest. Resuscitation2010;81:1736–7.

9 Ahn A, Nasir A, Malik H, et al. A Pilot Study Examining the Role of RegionalCerebral Oxygen Saturation Monitoring as a Marker of Return of SpontaneousCirculation in Shockable (VF/VT) and Non-Shockable (PEA/Asystole) causes of CardiacArrest. Resuscitation 2013;84:1713–16.

4 Singer AJ, et al. Emerg Med J 2014;0:1–4. doi:10.1136/emermed-2013-203467

Original article

group.bmj.com on November 12, 2014 - Published by http://emj.bmj.com/Downloaded from

observational studycirculation following cardiac arrest: anassociated with return of spontaneous Cerebral oximetry levels during CPR are

Mark C Henry and Sam ParniaAdam J Singer, Anna Ahn, Loren A Inigo-Santiago, Henry C Thode, Jr,

published online March 24, 2014Emerg Med J 

http://emj.bmj.com/content/early/2014/03/24/emermed-2013-203467Updated information and services can be found at:

These include:

References

#BIBLhttp://emj.bmj.com/content/early/2014/03/24/emermed-2013-203467This article cites 9 articles, 3 of which you can access for free at:

serviceEmail alerting

box at the top right corner of the online article. Receive free email alerts when new articles cite this article. Sign up in the

CollectionsTopic Articles on similar topics can be found in the following collections

(551)Resuscitation

Notes

http://group.bmj.com/group/rights-licensing/permissionsTo request permissions go to:

http://journals.bmj.com/cgi/reprintformTo order reprints go to:

http://group.bmj.com/subscribe/To subscribe to BMJ go to:

group.bmj.com on November 12, 2014 - Published by http://emj.bmj.com/Downloaded from