The effect of the head-up position in cardiopulmonary resuscitation – A systematic review and meta-analysis

Cheng-Chieh Huang Changhua Christian Hospital Kuan-Chih Chen Changhua Christian Hospital Zih-Yang Lin Changhua Christian Hospital Yu-Hsuan Chou Changhua Christian Hospital Wen-Liang Chen National Yang Ming Chiao Tung University Tsung-Han Lee Changhua Christian Hospital Kun-Te Lin Changhua Christian Hospital Chu-Chung Chou Changhua Christian Hospital Yan-Ren Lin (  h6213.lac@gmail.com ) Changhua Christian Medical Foundation Changhua Christian Hospital https://orcid.org/0000-0002-9015-2678


Introduction
Cardiac arrest is the most critical challenge faced by every clinical physician because it has varied etiologies and a high mortality rate [1]. High-quality cardiopulmonary resuscitation (CPR) and improvement in emergency medical service (EMS) system has been proven to result in a higher return of spontaneous circulation (ROSC) rate in out-of-hospital cardiac arrest (OHCA) patients, and the survival rate has improved over time since 2006. However, less than 10% of patients survive to hospital discharge, and survival with good neurological outcomes is even lower [2,27,28].
Insu cient brain perfusion is a key factor in poor neurological outcome. Cerebral perfusion pressure (CerPP), which is calculated as the mean arterial pressure (MAP) minus the intracranial pressure (ICP), decreases dramatically during cardiac arrest for the following reasons. First, once cardiac arrest occurs, in ammatory systems are activated to respond to whole-body ischemia, resulting in increased membrane permeability [3]. In addition, the blood-brain barrier (BBB) breaks down because of intracellular acidosis, stopping oxidative phosphorylation and lactate accumulation [4]. Due to these two effects, serum proteins and water pass from blood to brain tissue, leading to neuronal, glial or axonal injuries [5] and increasing ICP. Second, chest compressions are performed to increase intrathoracic pressure, rather than direct compression of the heart, to create blood ow. Increased intrathoracic pressure increases ICP by elevating the pressure in non-valved veins (for example, the paravertebral venous plexus), which reduces the drainage of blood from the brain and thoracic cerebrospinal uid (CSF) [6]. Finally, optimal CPR can provide approximately 20%~30% of prearrest cardiac output [7], and only 30% of it ows to the brain [4]. As a result, the rate of survival with good neurological outcomes is dismal.
In recent years, some studies have revealed that head-up position CPR could decrease ICP and improve CerPP [8], and that it may even improve coronary perfusion pressure (CoPP) [9] and increase the ROSC rate [10] in animal experiments. Compared to conventional CPR, in which the patient lies down at 0 degrees, elevating the head during CPR could accelerate brain venous return and the hydrostatic displacement of CSF from the cerebral ventricles to the spinal cavity [11]. Thus, ICP decreases and CerPP increases. In addition, facilitating venous return may increase CoPP, and a higher CoPP is associated with a higher ROSC rate [12]. However, Park YJ et al. demonstrated that head-up CPR could worsen the survival rate [13]. The outcome of head-up CPR seems inconclusive, so clarifying the effect of head-up CPR is important. Therefore, in this article, a comprehensive systematic review and meta-analysis was performed to evaluate the effect of head-up CPR in animal models and to draw conclusions to establish a new strategy for CPR in the future.

Method
We conducted this study according to the Cochrane Handbook for Systematic Reviews of Interventions guidelines [14], Hooijmans et al [15] and Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statements [16].

Eligibility criteria:
Types of studies All types of studies were eligible except for case reports, reviews, abstract publications and conference presentations because there is no detailed study design to assess quality or data to analyze.

Types of participants
We includes all kinds of animal studies. If different animal species were included, we analyzed and discussed them separately.

Types of interventions
Studies comparing head-up position (HUP) CPR at a xed degree with supine position (SUP) CPR were included. There were no restrictions on the head-up degree, no-ow time, CPR device, CPR protocol, or CPR duration. The head-up degree is de ned as the angle between the head-chest plane and the horizontal plane. A head-up degree of zero means the supine position. No-ow time was de ned as the time from cardiac arrest to the start of CPR. CPR duration was de ned as the time from the start of CPR to the stopping point of CPR regardless of whether ROSC was achieved.

Types of outcome measures
The primary outcome was the CerPP in both groups. The secondary outcomes were the mean ICP, MAP, CoPP and ROSC rates. If the studies presented outcomes such as systolic pressure, diastolic pressure and ICP during the compression/decompression phase, we estimated MAP as (systolic pressure + diastolic pressure)/2 and mean ICP as (ICP during compression + ICP during decompression)/2 because the compression time is approximately equal to the decompression time during CPR.
Search methods for the identi cation of studies Studies were obtained by comprehensively searching 3 databases (PubMed, EMBASE and the Cochrane Library) from database inception to 10 May 2021 without language restriction. The following key words or medical subject heading (MeSH) terms were used: head-up position or head up or torso up or head elevation or tilt AND resuscitation or CPR. We also reviewed the references of the identi ed articles to avoid missing possible articles.
Data extraction and quality assessment Two authors (C.K. Chen and Z.Y. Lin) searched articles from 3 databases and extracted the data independently. We collected the following information from each eligible study: authors, publication year, study design, study group, CPR protocol, CPR device, no-ow time, CPR duration, intervention/control details and outcome data. If the data were presented in a graph instead of as digits, we used GetData Graph Digitizer software, version 2.26 (http://getdata-graph-digitizer.com/download.php), to extract the data.
The Animal Research: Reporting of In Vivo Experiments (ARRIVE) guidelines 2.0 [17] were used to evaluate the quality of the included articles, as assessed by two authors (C.K. Chen and Z.Y. Lin) independently. There were 21 domains that were assessed for each study.
Each domain was rated as "+" if it met the domain recommendation; otherwise, it was rated as "-".
Finally, any disagreement that occurred during the article search, data extraction or quality assessment was resolved by a discussion between a third author (C.C. Huang) and the previously mentioned two authors to reach a consensus.

Statistical analysis
We use Review Manager (RevMan [Computer program], Version 5.4.1, The Cochrane Collaboration, 2020) to analyze all data. Because some of the included studies presented results as the standard error of the mean (SEM), we converted SEM to the standard deviation (SD) by the formula SEM = SD/√(sample size) [18]. We expressed continuous data as standardized mean differences (SMD) and dichotomous data as risk ratios (RR). In addition, 95% con dence intervals (CI) were calculated for all of the results.
We applied a random-effects model for all analyses and chi-square and I 2 tests to evaluate heterogeneity. Subgroup analysis was conducted based on the change in position in CPR and the CPR duration of xed-position CPR. In addition, we performed a visual inspection of the funnel plot to assess publication bias, and a sensitivity analysis was conducted by repeating the analysis after removing one study at a time. Finally, the results were presented in forest plots.
The head-up angle ranged from − 60° to 60°. Three studies [9,13,20] (Table 1). However, we could not analyze the effect of HUP CPR in conventional CPR due to the lack of an experimental group in one study [8]. Quality assessment The quality assessments of our included studies are shown in Table 2. The average score was 16.86 ± 0.83 (mean ± SD), ranging from 15 to 18. A visual inspection of the funnel plot revealed symmetry, suggesting no publication bias (Additional le 1).
In addition, in xed-position CPR studies, we further evaluated whether the duration of CPR affected our primary outcome, and the results still showed a signi cant increase in CerPP in the HUP group regardless of whether the duration of CPR was < 10 min, 10-20 min, or > 20 min (p < 0.01 in all 3 subgroups) (Fig. 2B).

Secondary outcomes ICP and MAP
The overall ICP was signi cantly lower in the HUP group than in the SUP group (SMD, -3.59; 95% CI, -5.

CoPP
There was an increased trend in CoPP in the HUP group, but it did not reach statistical signi cance (SMD, 0.92; 95% CI, -0.24-2.08; p = 0.12; I 2 = 84%). However, the subgroup analysis revealed that CoPP was increased in the HUP group when the position during CPR was changed (SMD, 2.80; 95% CI, 0.15-5.45; p = 0.04; I 2 = 90%) (Fig. 4A). We did not observe a difference between the groups according to the duration of CPR (Additional le 4).

Discussion
The most important result of our study is that the HUP at 30° during CPR can signi cantly increase CerPP, mainly by reducing ICP, compared to the SUP. CerPP, which is calculated by MAP minus ICP, could be increased by higher MAP and/or lower ICP. The reason for the lower ICP is that when elevating the head and body up to 30° during CPR, ICP will decrease by facilitating brain venous return and CSF movement into the spinal subarachnoid space, which is consistent with previous studies, even at different elevation angles ranging from 10° to 50° [9,23]. In addition, the reduction in ICP also decreased the resistance to forward brain blood ow, which is generated by each chest compression. This effect could explain the ndings of 2 of our included studies [9,21], which demonstrate that brain blood ow increased signi cantly in HUP compared with SUP. Furthermore, due to the heterogeneous study protocols in our included studies, we analyzed the CerPP with regard to whether the position during CPR was changed or xed, and in x position CPR studies, we further evaluated the duration of CPR, which showed signi cantly increased CerPP in all HUP groups. Thus, HUP in CPR can reduce ICP, whenever the head is elevated during CPR, and the effect could last the whole CPR duration.
The other important result is that we did not nd a signi cant difference in MAP between the HUP group and the SUP group, regardless of whether the position was changed during CPR and the duration of CPR. Maintaining a su cient blood pressure by pumping upwards to the brain is important in CPR. From a physiological perspective, elevating the head and chest in CPR may reduce MAP because of the gravity effect. Each chest compression will pump more "uphill" than in the supine position. On the other hand, ACD-CPR with ITD could generate sustained aortic pressure, and the HUP may reduce the resistance of blood ow to the brain. Therefore, the net effect of MAP revealed no signi cant difference between the 2 groups in our study. The absolute MAP value was much lower in Putzer et al [22], who did not use ITD in ACD-CPR, and MAP along with CerPP decreased gradually over time. Debaty et al. [9] revealed that MAP showed a signi cant decrease immediately once ITD was removed in HUP CPR. These results could support our inferences. In addition, of our included studies, 2 [13,20] showed decreased MAP in the HUP, while the others revealed no signi cant difference between the 2 groups. Both of them were in the reverse-Trendelenburg position, rather than elevated head and chest only. Because more blood deposits in the lower extremities, we speculate that ACD-CPR with ITD does not overcome the physiological effect of the reverse-Trendelenburg position, resulting in a decrease in MAP. Interestingly, pulmonary edema is a common complication of cardiac arrest [24]. Elevating the chest may have better blood-gas exchange caused by reduced lung congestion and pulmonary vascular resistance because of the gravity effect [19]. This potential bene t should be con rmed by more studies.
Although there were no differences found in CoPP between the 2 groups, there was an increasing trend of CoPP in the HUP, despite the lack of statistical signi cance. CoPP is calculated by diastolic aortic pressure minus right atrial pressure [25]. Theoretically, while the head and chest are elevated, right atrial pressure is also decreased by the gravity effect. As a result, CoPP could be increased under ACD-CPR with ITD to maintain su cient diastolic aortic pressure. Kim et al. [20] revealed that CoPP increased gradually from the headdown position and supine position to the head-up position and reached the highest CoPP at 30°. On the other hand, two of the included studies [9,21] directly measured heart ow and revealed no signi cant difference between the 2 groups. In our study, we did not observe an apparent increase in CoPP in the HUP, perhaps due to different study protocols and small study groups.
Only 3 included studies [13,19,21], including 50 Yorkshire farm pigs, reported the ROSC rate, and the results showed no difference between the two groups. In these 3 studies, Park et al. [13]  To the best of our knowledge, this is the rst meta-analysis comparing HUP to SUP CPR in animal models. The strength of this analysis includes further con rming the effect of HUP CPR. In addition, we performed subgroup analysis according to position changes and the duration of CPR due to heterogeneous study protocols. Moreover, two authors used the ARRIVE guidelines 2.0 to evaluate inclusion study quality. There were also some limitations. First, all of the included studies used healthy animals, and a ventricular brillation model was used to simulate cardiac arrest. However, cardiac arrest is caused by more complex reasons in human beings, and human CPR physiology is more dynamic. Thus, the results of our study may not be totally transferrable to humans. Second, an optimal CPR position is necessary to achieve ideal CerPP and CoPP. We compared only the head-up 30° position to the supine position. Thus, the best CPR position has not yet been found. Third, all of the included studies used calculated CerPP and CoPP. Only 2 studies further measured brain blood ow and heart blood ow directly by using microspheres. Although high perfusion pressure is associated with high blood ow, the evidence is indirect rather than direct. Finally, and most importantly, even though our study demonstrates strong evidence of increasing CerPP by lower ICP in HUP CPR, it is still unclear whether this bene t could equal an increased survival rate with good neurological outcome. Thus, further large-sample and standardized research is essential to con rm the optimal resuscitation position for humans as well as animals.

Conclusion
HUP 30° in ACD-CPR with ITD can increase CerPP by lowering ICP signi cantly and maintain MAP compared to SUP, and the effect is immediate and lasts for the whole CPR duration. In addition, CoPP might also increase compared to that with SUP. This study further con rms the bene t of HUP CPR in animal models, and further large-sample and standardized research is warranted to clarify the optimal resuscitation position for humans as well as animals.