Our study showed a statistically significant difference in mean rSO2 between OHCA patients who achieved ROSC and not, using both the INVOS™ 5100 and ROOT® O3. We did not find a difference in ETCO2 between the groups achieving ROSC or not. Both the INVOS™ 5100 and ROOT® O3 can be used in a pre-hospital setting and during transport with minimal interference of ALS according to EMS personnel.

The INVOS™5100, ROOT® O3 and other cerebral oximeters on the market have their own algorithms, methodologies, and sensor designs for analyzing oxyhemoglobin. They also have different ratios for approximation of the contributed arterial and venous saturation (INVOS™ 25/75% and ROOT O3® 30/70%). To the best of our knowledge, all cerebral oximeters are validated during a state of normal circulation. During a cardiac arrest, there is hypocirculation, and the ratio of venous and arterial blood in the cerebral circulation might be different than the ratio 25/75% and 30/70%. When using a device that is not validated for a specific situation, such as resuscitation, this need to be taken into consideration.

We found a statistically significant difference between ROSC and no ROSC with both cerebral oximeters. When rSO2 values were placed in a boxplot, there appears to be a cutoff in rSO2 between 30 and 35% measured with INVOS™5100 compared to the no ROSC group where all patients, except one outlier, had an rSO2 < 30%. All patients achieving ROSC had an rSO2 > 35%. A similar pattern was seen in the Root® O3 group but with a smaller range in rSO2. In the ROOT®O3 group, patients with no ROSC had an rSO2 < 48% and in the ROSC group rSO2 > 44%, and only one patient with no ROSC had an rSO2 > 44%. The lowest mean rSO2 in the ROOT®O3 group was 34% compared with 15% in the INVOS™ 5100 group. As pictured in Fig. 1, a cutoff between ROSC and no ROSC seems to be different for the two devices.

The difference between different cerebral oximeters has been shown in hypocapnia and normal circulation. Bickler et al. compared five different cerebral oximeters with SaO2 and SvO2 and showed a between-subject and between-instrument difference in normocapnic hypoxia [14]. It has also been found that Fore-Sight (CAS Medical Systems, Branford, CT, USA,), Equanox (Nonin Medical, Inc., Plymouth, MN, USA) and NIRO (Hamamatsu Photonics, Hamamatsu-shi Shizuoka, Japan) give an overestimation of rSO2 when compared to calculated values from SaO2 (oxygen saturation) and SvO2 (mixed venous oxygen saturation). Ferraris et al. showed that the rSO2 measured on the forearm during cardiac surgery was not interchangeable between oximeters using Equanox™ and ROOT® O3 [15]. Both INVOS™ and ROOT® O3 are validated during normocapnic hypoxia. Schober et al. showed increasing bias in rSO2 when hypocapnia occurs together with hypoxia measured by Fore-Sight and INVOS [16].

Our results are in line with previous studies showing a statistically significant difference in rSO2 between patients achieving ROSC or not after OHCA [5,6,7,8,9, 17,18,19,20,21]. To the best of our knowledge, only three other studies have incorporated cerebral oximetry in everyday clinical practice in a pre-hospital setting, but they used INVOS [19] and Equanox [8, 9]. These studies showed a mean rSO2 at 47%, 37%, and 42% in the ROSC group compared with 31%, 32% and 31% in the no ROSC group, and the differences between groups were statistically significant. Our study showed higher rSO2 in the ROSC group both with the INVOS™5100 and ROOT® O3. In these three previous studies, mechanical chest compressions were used in only two patients compared with all patients in our study. This could be one explanation for the higher rSO2 values in our study. A meta-analysis by Westfall et al. compared manual versus mechanical chest compressions in OHCA and concluded that mechanical compressions were superior regarding ROSC [22].

In our study, we found no statistically significant difference between ETCO2 in the ROSC and no ROSC groups. ETCO2 has a role during ALS to confirm the correct placement of an endotracheal tube. Prior studies have shown that ETCO2 < 1.3 kPa after 20 min of ALS had no survivors [23]. Rognås et al. found that four out of 22 patients received ROSC despite an ETCO2 < 1.3 kPa [24]. Kolar et al. found that median ETCO2 of 1.9 kPa during OHCA could distinguish between ROSC and no ROSC [25]. Prior studies have shown a relationship between a rise in ETCO2 and ROSC but no cut off value is agreed upon. ETCO2 is dependent on respiratory minute volume and can change when epinephrine or sodium bicarbonate is administered [26]. Heradstveit et al. showed that the cause of arrest, initial rhythm, bystander CPR, and time from arrest to ETCO2 measurement were confounding factors in interpreting the significance of ETCO2 values [27]. As in our study, Engel et al. found that rSO2 was superior to ETCO2 in predicting ROSC [28].

Feasibility of measuring cerebral oxygenation during cardiac arrest has been studied both in-hospital and in the pre-hospital setting, using different oximeters, with Equanox™ and INVOS™5100 being the most investigated [5,6,7, 10, 29]. Schewe et al. showed that rSO2 recordings were achieved in 89.9% of the intended recording time during OHCA [8]. Weatherall et al. showed that cerebral oximetry can be measured both during road and helicopter transports in healthy volunteers using Fore-Sight with a signal present > 70% of the time with 100% for road transported and 85.7% helicopter transported patients [30]. To our knowledge, no other study has assessed the feasibility of the ROOT® O3 during OHCA and transportation to the hospital.

To assess the usability of the two devices, we had one question in regard to interference with ALS for the RRC personnel to fill out after each study patient. The personnel rated the statement to a median of 2 (IQR 1–6) on a 10-point Numerical Rating Scale. From the answers with higher scores, we could read that there was problem with the battery running low, the device was heavy and difficult to carry along, the sensors fell off, or the application or documentation took focus away from the ALS. With nine people filling out the questionnaires, the interference rating could change after building up experience with the device. The place where the OHCA occurred and the situation could differ widely, meaning using the device could be easy in one case and more troublesome in another. Sensors falling off and connection problems between the sensor and cables can be a constant problem affecting the feasibility when applying sensors and transporting a patient from the place of arrest to the hospital. To be even more useful in a pre-hospital environment, the oximeter could be incorporated into the existing equipment or be small and easy to carry. Long battery capacity with a quick recharge are other advantageous features to avoid taking focus off of ALS. When measuring rSO2 with two different cerebral oximeters, we had expected a small difference in values but found a higher mean rSO2 in the ROOT O3® group with smaller confidence intervals than in the INVOS 5100™ group. We see potential in rSO2 guiding CPR, and it could contribute in part to making the decision of whether to continue resuscitation as well as if a patient could benefit from extracorporeal CPR.

Limitations.

The on-call physician decided whether to include a patient or not in the study, therefore, inclusion bias could have occurred. Since included patients ages ranged from 19 to 92, we believe even if there were an inclusion bias, it was minor. As we did not want our measurements to negatively impact ALS, sensor placements and the rSO2 measurements started at different times during the OHCA, at the place of arrest or during transportation. This means that we cannot compare all patients at the same timepoint, but as in prior studies, we present group-wise means for rSO2. rSO2 and ETCO2 were documented manually, leading to a risk of recall bias, but with the equipment used in the study, ETCO2 was routinely recorded manually by a physician who was familiar with this way of observation and documentation.