Skeletal muscle oxygenation during cardiopulmonary resuscitation as a predictor of return of spontaneous circulation: a pilot study

Out of hospital cardiac arrest (OHCA) is a major cause of morbidity and mortality around the world [1]. Despite the progress in medicine, new devices and research, there was no great breakthrough in recent years [2]. There are numerous factors that have influence on the outcome of cardiopulmonary resuscitation (CPR) in OHCA. It is usually impossible to predict weather the resuscitation will be successful or not, especially in early stages of cardiac arrest [3].

Near-infrared spectroscopy (NIRS) is a noninvasive optical technique that uses light in near-infrared spectrum of electromagnetic wave (700–1300 nm) [4, 5]. NIRS can be used to assess tissue oxygenation, oxygen consumption and blood flow in various tissues, including brain and skeletal muscle [6‐8]. NIRS has an advantage over pulsatile oximetry, because it can be used in situations, where there is no blood flow [9]. With NIRS we measure oxygen saturation in all vessels that are smaller in diameter than 1 mm (arterioles, capillaries, venules) [4, 10]. Most of the signal comes from capillaries, because they represent majority of vessels in the tissue [6].

During cardiac arrest and during CPR the blood flow through the brain is absent or significantly reduced. Brain injury is major cause for neurologic disability after successful resuscitation [4, 7, 10]. With the placement of NIRS probes on the forehead region, regional tissue oxygen saturation (rSO₂) in superficial areas of frontal brain lobes is measured. Current data confirms that increase of brain rSO₂ is associated with higher probability of return of spontaneous circulation (ROSC) [4, 7, 9‐12]. Different values of basal brain rSO₂ or different values of brain rSO₂ increase were associated with higher probability of ROSC. Genbrugge et al. reported that absolute increase of rSO₂ for 15% or more was associated with ROSC. They also noticed that increase in rSO₂ beyond 1 min after initiation of rSO₂ measurement was associated with more favorable long-term neurologic outcome [13]. Parnia et al. have shown that all patients with ROSC had mean brain rSO₂ of 35% or higher. A rise of brain rSO₂ from baseline was associated with ROSC and values remaining below 30% most of the period of CPR predicted that ROSC will not be achieved [12]. Recent meta-analysis has confirmed prognostic value of brain tissue oxygenation [14].

During cardiac arrest there is no blood flow, the consequence is lower oxygen values in all tissues [15, 16]. Skeletal muscles are not part of vital organs and flow through them is decreased in critical states. In critically ill we currently monitor skeletal muscle rSO₂ in patients with shock or in patients on different types of circulatory mechanical support [8, 17]. We have previously shown that skeletal muscle rSO₂ can predict adequacy of flow (i.e., mixed venous oxygenation) in patients with shock with preserved oxygen extraction; it can also be used to track effects of therapy [8, 18]. Despite new NIRS technologies and design of probes, skeletal muscle rSO₂ monitoring remains technically more reliably compared to brain rSO₂, because skeletal muscle is covered with a thin layer of skin and subcutis compared to brain, which is hidden in the skull and floating in the cerebrospinal fluid [19]. A short paper already reported an illustrative case series of skeletal muscle rSO₂ use in five patients in the emergency department, showing fast response of skeletal muscle rSO₂ value to loss or return of pulse [15].

Aim of our study is to test feasibility to monitor skeletal muscle rSO₂ during CPR after OHCA and to assess changes of skeletal muscle rSO₂ during CPR and after ROSC. In addition, we want to explore the relationship between skeletal muscle and brain rSO₂.

Thirty patients were recruited. Ten patients were excluded due to different technical problems or violation of study protocol (Fig. 2): two patients due to disconnection of NIRS probe and consequent loss of skeletal muscle rSO₂ signal during CPR; in four patients the probes were applied more than 5 min after arrival to the patient; in one patient probes were applied after ROSC; three patients were excluded due to irregular rSO₂ signals, which have not allowed to determine predefined checkpoints.

Twenty patients with skeletal muscle rSO₂ data were finally analyzed. In these patients, one patient had only skeletal muscle and no brain rSO₂ recordings. Both skeletal muscle and brain rSO₂ were recorded in 19 patients.

Basic demographic data, data about arrest and CPR are presented in Table 1. Twenty patients [age: 66.0 years (60.5–79.5), 65% male] with OHCA (50% witnessed, 70% with BLS), time to advance life support (ALS) 13.5 min (11.0–19.0) were finally analyzed. Cumulative dose of adrenaline was higher in no-ROSC group.

Measurement of skeletal muscle rSO₂ and brain rSO₂ are presented in Table 2. There was no statistically significant difference between the basal skeletal muscle and the basal brain rSO₂ (17.5 (12.8–26.0) vs. 31.0 (15.8–41.6), P =  0.1674, respectively). Basal, maximal and end-CPR skeletal muscle rSO₂ were higher in ROSC compared to no-ROSC group (49.0% (39.7–53.7) vs. 15.0% (12.0–25.2), P =  0.006; 76.0% (52.7–80.5) vs. 34.0% (18.0–49.5), P =  0.005; 72.0% (48.7–74.7) vs. 16.0% (12.0–35.0), P =  0.002, respectively) (Fig. 3). Delta rSO₂ for skeletal muscle was not significantly different in patients with ROSC and no-ROSC group.

Basal brain rSO₂ did not differ between ROSC and no-ROSC (38.0% vs. 29.5% (14.5–42.5), P =  0.4). Maximal, delta-rSO₂ and end-CPR brain rSO₂ were higher in ROSC compared to no-ROSC group (77% vs. 42.0% (30.5–53.0), P =  0.01; 27% vs. 10.5% (6.0–15.0), P =  0.007; 77% vs. 39.0% (29.7–52.7), P =  0.01, respectively) (Table 2).

There was non-linear cubic relationship between duration of collapse to establishment of NIRS monitoring and basal skeletal muscle rSO₂ in witnessed OHCA and without BLS (F-ratio = 9.7713, P =  0.0261) (Fig. 4).

There was no correlation between basal rSO₂, delta-rSO₂ and end-CPR rSO₂ of skeletal muscle and brain. There was a correlation between maximal skeletal muscle and brain rSO₂ (n = 18, rho: 0.578, P =  0.0121), which was confirmed also in linear regression model (y = 25.245 + 0.499 x, r = 0.63, P =  0.005) (Fig. 5).

The study confirmed the feasibility of monitoring skeletal muscle rSO₂ in OHCA. Patients with ROSC had higher skeletal muscle rSO₂ at the start of ALS (basal rSO₂) and during CPR (maximal rSO₂). There was a non-linear cubic relationship between basal rSO₂ and time from collapse to start of rSO₂ monitoring. There was also a linear relationship between the maximal values of skeletal muscle and brain rSO₂.

During cardiac arrest there is decrease of rSO₂, the value of decrease depends on the duration of no-flow/low flow and oxygen consumption of the tissue [8]. Basal skeletal muscle rSO₂ is a surrogate for estimating the time of tissue low/no flow. Our study has shown a relationship between duration of witnessed cardiac arrest and basal skeletal muscle rSO₂, which was, however, not significant due to low number of patients, because only patients with witnessed cardiac arrest and without BLS were include in that analysis. Basal skeletal muscle rSO₂ were significantly different between ROSC and No ROSC groups, when all included patients were analyzed.

By rapid cuff inflation it is possible to stop flow through the arm, such as simulation of cardiac arrest, and evaluate skeletal muscle oxygen consumption. We have done this in patients with sepsis/septic shock and controls, aiming skeletal muscle rSO₂ to decrease to 40% [23]. The rate of StO₂ decrease during rapid cuff occlusion test was lower in septic shock patients compared to severe sepsis and controls (− 5 ± 2%/min vs. − 12 ± 2%/min vs. − 37 ± 7%/min, respectively; P < 0.001). In healthy volunteers we could measure rate of skeletal muscle rSO₂ decrease during no-flow for longer period of time without any major risk, to explore skeletal muscle rSO₂ kinetics and construct a normogram for estimating the duration of cardiac arrest for different age and gender groups.

In current study increase of skeletal muscle during resuscitation (delta rSO₂) was not different between ROSC and No ROSC groups, this additionally emphasize the importance to start the resuscitation early as possible, when the basal skeletal muscle rSO₂ (tissue oxygenation) is also still relatively high.

By additional fixation we have improved the position of NIRS probe on the thenar allowing more stable signal monitoring. The problem was big probe size, compared to thenar and higher possibility of detachment while manipulating patient’s hand. This fixation completely removed the possibility of losing contact and consequently signal during rSO₂ monitoring even during different manipulations around and with the patient.

Despite the end of resuscitation (end-CPR rSO₂) was different between ROSC and No-ROSC group, it is a very biased measure, since end-CPR by definition occurs much later in the no-ROSC group. This is a similar problem to the resuscitation-time bias seen in previous observational studies of the use of adrenaline during resuscitation from cardiac arrest [24‐26]. End-CPR rSO2, furthermore, is an impracticable measure that cannot be used in a clinical relevant setting, because of the obvious fact that we cannot predict ROSC.

Brain rSO₂ seems to be physiologically superior compared to other regional tissues. However, other rSO₂, like kidney, are tested to give additional value to brain rSO₂ during resuscitation in pediatric population [27, 28]. Monitoring of skeletal muscle rSO₂ seem to be technically more reliably compared to brain rSO₂, especially due to brain location in the body [19].

Several NIRS devices are available for clinical use. They differ according to numerous aspects, which include the algorithms adopted, the type of light source, the wavelengths of light emitted, the number and distance between the light emitters and detectors [19]. For example, INVOS system (5100C Cerebral/Somatic Oximeter; Medtronic, MN, USA) uses near-infrared light at two wavelengths [29]. Light travels from the light emitting diode of the sensor to either a proximal or distal detector, which allows separate data processing of shallow and deep optical signals. Data from the scalp and the surface tissue are subtracted and suppressed by spatial resolution, which reflects the rSO₂ in deeper tissues [30]. The EQUANOX 7600 and also the SenSmart Model X-100 (Nonin Medical, MI, USA), the model that we have used, uses a dual light emitting and detecting sensor architecture, which has been shown to more effectively target the cerebral cortex and eliminate extracranial contamination from the scalp and skull. The Nonin system uses four wavelengths of near-infrared light. These added third and fourth wavelengths increase the accuracy of reporting the actual percent of oxygenated hemoglobin in the targeted tissues and can compensate for tissue factors that might otherwise reduce the accuracy of the measurements. This also allows the algorithm to reduce inter-subject variability, regardless of age, weight or skin color [30, 31]. We have previously shown high grade diversity of brain rSO₂, with different NIRS devices, in patients with alkaptonuria, who had widespread tissue deposition of black pigment [19, 32, 33]. To guide our resuscitation efforts, we should probably not focus on only one modality, i.e., brain rSO₂. Especially, because there is a report when good neurological outcome was achieved after prolonged CPR despite very low brain rSO₂ [34].

Skeletal muscle rSO₂ could also guide post-resuscitation care [17]. Continuous monitoring skeletal muscle rSO₂ is already used in trauma patients and identifies the severity of shock [35]. Skeletal muscle rSO₂ can track changes of systemic oxygen delivery during and after resuscitation of trauma patients or in patients heart failure patients/cardiogenic shock, who have preserved oxygen extraction [36]. New methods, such as near-infrared spectroscopy, which measures venous oxygen saturation in tissue from the near-infrared spectrum of the amplitude of respiration-induced absorption oscillations, may lead to the design of a non-invasive optical instrument capable of providing simultaneous and real-time measurements of local arterial, tissue and venous oxygen saturation.[37].

There was no statistically significant difference between the basal skeletal muscle and the basal brain rSO₂, despite we would expect lower basal brain rSO₂ due to higher cerebral oxygen consumption in normal human subjects compared to resting skeletal muscle oxygen consumption [38, 39]. There was also a very wide spread of basal rSO₂. In pre-arrest state the patient could have centralization of flow to vital organ, and skeletal muscle rSO₂ would be already low before cardiac arrest, as we previously have shown in patients with cardiogenic shock [8]. In our study, during resuscitation, there was lineal correlation between the maximal brain and skeletal muscle rSO₂.

Repeatability of skeletal muscle rSO₂ with NIRS during vascular occlusion test was confirmed in previous studies [40].

Current study has at least four major limitations. First, the number of recruited patients is low despite long recruiting period. The main cause is the SARS-CoV-2 epidemics, during which the study was temporally stopped to minimize workload of staff in protective clothing. Second, the study was only single center study. Our data should be confirmed in a bigger prospective multicenter study. Third, low number of patients with ROSC, did not allow to study time change of skeletal muscle rSO₂ during resuscitation and prognostic value of skeletal muscle rSO₂ for good neurological outcome. The fourth, study was not designed to study use of skeletal muscle rSO₂ as post-resuscitation therapy guide and non-invasive estimation of adequacy of flow. It should be done in other multicenter study.

Recording of skeletal muscle rSO₂ during CPR in patients with OHCA is feasible. Basal and maximal skeletal muscle rSO₂ were higher in ROSC compared to no-ROSC group. Skeletal muscle rSO₂ during cardiac arrest could provide additional data to brain rSO₂ on duration of arrest and efficiency of resuscitation efforts.