Effects of head-up CPR on survival and neurological outcomes: A systematic review

Abstract

Background

The impact on survival and neurological outcomes of head-up cardiopulmonary resuscitation (CPR) remains unclear. This systematic review aimed to examine whether head-up CPR or head-up CPR bundle affects survival and neurological outcomes.

Methods

In this review registered in PROSPERO (CRD42024541714), we searched Medline, EMBASE, Scopus, Emcare and Cochrane library from inception to February 12, 2025. The inclusion criteria were studies: in adults and children in any setting (in-hospital or out-of-hospital) with cardiac arrest, comparing head-up CPR or head-up CPR bundle with standard or compression-only CPR in supine position, and randomized controlled trials (RCT) and observational studies. We excluded non-human studies and used Cochrane’s Risk of Bias-2 and ROBINS-I tools for risk of bias assessment, and GRADE for outcome assessment. We conducted a narrative synthesis of the findings.

Results

We identified four observational studies with 10,099 participants and no RCTs. All outcomes were judged to be very-low certainty of evidence, subject to high-risk of bias. Two pre- and post-intervention studies showed no statistically significant difference in survival with a good neurological outcome. The other two studies, which analyzed overlapping cohorts comparing the head-up CPR bundle with conventional CPR, showed mixed results—one found no statistically significant difference in survival to hospital discharge with favorable neurological status, while the other, focusing on patients with nonshockable rhythms, found that the head-up CPR bundle was associated with increased survival with favorable neurological function.

Conclusions

The available evidence remains limited, highlighted by the absence of RCTs or observational studies with adequate comparisons.

Introduction

Out-of-hospital cardiac arrest (OHCA) is a major global public burden.1, 2 Despite significant worldwide efforts and the advances in science, survival and neurological outcomes of cardiac arrest have remained low, particularly in cases with nonshockable rhythms.

Head-up cardiopulmonary resuscitation (CPR) is a recent resuscitation strategy. First described in 2014, head-up CPR involves the gradual elevation of the head to improve cerebral perfusion, coronary perfusion, and possibly ventilation during CPR.3, 4 Although the intervention may sound simple, previous studies have suggested that it is more complex than initially thought.⁵ Animal studies have indicated that head-up CPR is most effective when used with active compression/decompression (ACD) and an impedance threshold device (ITD). In the absence of these devices, there is inadequate arterial pressure to create upward flow and achieve cerebral perfusion pressure.6, 7, 8 Based on these findings, head-up CPR is often performed as part of a bundled approach, including the use of ACD and ITD devices.⁹

Due to a lack of clinical evidence, the latest International Liaison Committee on Resuscitation (ILCOR) International Consensus on Cardiopulmonary Resuscitation and Emergency Cardiovascular Care Science with Treatment Recommendations (CoSTR) in 2021 suggests “against the routine use of head-up CPR during CPR” (weak recommendation, very low-certainty evidence).¹⁰ However, several new observational studies have since evaluated the impact of head-up CPR on survival and neurological outcomes, warranting a new systematic review on this topic. In this study, we aimed to perform a systematic review of clinical evidence that compared head-up CPR and head-up CPR bundle against standard or compression-only CPR in supine position.

Methods

This systematic review is reported according to the “Preferred Reporting Items for Systematic Reviews and Meta-Analyses” (PRISMA) statement¹¹ and the protocol was registered in PROSPERO (CRD42024541714). We made the following updates to our review after initial PROSPERO registration: as the search strategy was revised this is a new systematic review and not an update of the 2021 review; Emcare and Scopus were added as additional databases; re-ran the search in all and, as a mentee ILCOR member was included in the review team, three reviewers conducted the screening, risk of bias assessment, and data extraction instead of the two originally planned. The ILCOR evidence evaluation process for systematic reviews was followed.¹².

Eligibility criteria

Studies were included if they met the pre-determined inclusion criteria following the PICOST framework for ILCOR systematic reviews. The Population was adults and children with attempted resuscitation after in-hospital cardiac arrest (IHCA) or OHCA. The Intervention was head-up CPR or head-up CPR bundle. The Comparator was standard or compression-only CPR in the supine position. The Outcomes predefined as critical were survival at 30 days or hospital discharge with a favorable neurological outcome (defined as Cerebral Performance Category 1 or 2, or modified Rankin Scale 0–3), and survival at 30 days or hospital discharge. The Outcomes predefined as important was return of spontaneous circulation (ROSC). Eligible Study designs were randomized controlled trials (RCTs) and non-randomized studies (non-randomized controlled trials, interrupted time series, before-and-after studies, cohort studies) in all languages, providing there was an English abstract available. Conference abstracts and trial protocols were excluded. The Timeframe was all years from inception of the searched databases to February 12, 2025.

Search strategy

The search strategy used in the previous systematic review by ILCOR¹⁰ was reviewed and edited by an informational specialist. The version was then peer reviewed by an additional informational specialist and approved by the ILCOR BLS Task Force. As the search was revised, we re-ran the search back to database inception in five databases (Medline, EMBASE, Scopus, Emcare and Cochrane library) to February 12, 2025. The search strategies are provided in the supplementary material. Additionally, we reviewed the reference lists of all included studies to identify any potential additional articles.

Definition of head-up CPR

Head-up CPR was defined as a resuscitation strategy that involves gradual elevation of the head and upper torso during resuscitation. In head-up CPR, the head and upper body are typically elevated at an angle of 30 degrees or more, either using specialized devices or by tilting the stretcher or surface.4, 9, 13 Although the literature describes multiple ways to perform head-up CPR, a mechanical CPR device is typically used because performing manual CPR while the head and upper body are elevated is impractical. Additionally, an ITD is often employed as part of a bundled approach for the reasons outlined above. The systematic review included all papers, regardless of the specific methods employed in their studies.

Study selection

After removal of duplicates, three reviewers (TN, GL, AS) independently reviewed each title and abstract of the studies using the Covidence software (Covidence R). Full texts were obtained for articles of interest and were screened independently by the same three reviewers for inclusion. All disagreements were resolved by discussion and consensus.

Data extraction

Three authors (TN, GL, AS) independently extracted relevant data from all the included studies into an Excel spreadsheet. We resolved disagreement when it occurred by discussions and consensus. Study-level data included author and publication year, country of origin, study design, and intervention category and description. Outcome data included survival at 30 days with a favorable neurological outcome, survival to hospital discharge with a favorable neurological outcome, survival at 30 days, survival to hospital discharge, event survival defined as the proportion of patients surviving to hospital admission/arriving to hospital with a spontaneous circulation, and ROSC.

Risk of bias and certainty of evidence

We employed the GRADE approach to assess the certainty of evidence for each outcome considered critical or important. GRADE evaluates potential limitations in study design, indirectness, imprecision, inconsistency, and publication bias. Where feasible, we planned to create funnel plot to assess publication bias. The same three reviewers independently assessed the risk of bias for each study, and disagreements were resolved by consensus. As no randomized trials were found, we used the ROBINS-I tool for observational studies.¹⁴

Data synthesis with meta-analysis

We planned to assess studies for clinical (participants, interventions, outcomes), methodological (study design, risk of bias), and statistical heterogeneity (forest plots, χ² statistics, I² statistics). In the presence of heterogeneity, a random-effects model would be used for meta-analyses. If heterogeneity—whether clinical, methodological, or statistical—was deemed too substantial, we planned to conduct a narrative synthesis instead of performing meta-analyses.¹⁵ Summaries of individual study effect estimates were collated, and where observational studies reported more than one analysis, adjusted or case matched analyses were used in preference to unadjusted analyses as per Cochrane guidance.¹⁶ We also planned to conduct subgroup analyses for cardiac arrests with shockable and nonshockable rhythms.

Results

The systematic review search identified 419 studies, of which 21 studies were selected for full-text screening (Fig. 1). There were four observational studies17, 18, 19, 20 and no RCTs were included (Table 1). These four observational studies were conducted by the same research group, and no children were included in any of the studies. An additional observational study was excluded because it focused solely on the increase in cerebral blood flow with head-up CPR, rather than on any of our pre-specified primary outcomes and lacked a comparison group.²¹ Pooling of data and meta-analysis was not performed due to overlapping cohorts and substantial heterogeneity in study design and patient characteristics. Therefore, narrative synthesis of the results from included studies was performed.

Fig. 1.

PRISMA diagram for study selection.

Table 1.

Summary of included studies.

Study	Design	Intervention	Comparator	Main findings for ROSC/event survival	Main findings for survival and neurological outcomes
Pepe, USA 201917	Prospective observational study n = 2,322 (1,356 in the intervention group)	Head-up CPR bundle that includes a mechanical CPR device and ITD along with 1) applied oxygen but deferred positive pressure ventilation several minutes, 2) a pit crew approach for rapid placement of the mechanical CPR device, and 3) subsequently placed the patient in a reverse Trendelenburg position (∼20°)	Before intervention (Patients who received only a mechanical CPR device, and ITD)	The rate of event survival increased from 17.9% (144/806) in the pre-intervention period to 27.5% (49/160) during the transition, and to 34.2% (464/1,356) in the post-intervention period.	35–40% intact neurologic status (modified Rankin Score < 3) for both periods.
Moore, USA, 202218	Prospective observational study n = 5,423* (227 in the intervention group)	Head-up CPR bundle that included an automated controlled head and thorax patient positioning device, ACD and ITD.	Conventional CPR with supine position. Data of the individual patients who received conventional CPR were obtained from three large published OHCA randomized controlled trials.	After propensity score matching, there was no statistically significant difference in the rate of ROSC between the head-up CPR and conventional CPR (33% [74/222] versus 33% [282/860], adjusted OR, 1.02, 95% CI, 0.75–1.49).	After propensity matching, overall outcomes between the head-up CPR and conventional CPR group were similar for survival to hospital discharge (9.5% [21/222] versus 6.7% [58/860], adjusted OR, 1.44, 95% CI, 0.86–2.44) and survival to hospital discharge with favorable neurological status (5.9% [13/222] versus 4.1% [35/860], adjusted OR, 1.47, 95% CI, 0.76–2.82).
Bachista, USA, 202419	Prospective observational study n = 2,232 (380 in the intervention group)	Head-up CPR bundle that included an automated controlled head and thorax patient positioning device, ACD and ITD.	Conventional CPR with supine position. Data of the individual patients who received conventional CPR were obtained from two large published OHCA randomized controlled trials.	Propensity matched analysis:ROSC for head-up CPR was 33% (118/353) versus 29% (101/353) for conventional CPR (adjusted OR, 1.25 [95% CI, 0.91–1.72]).	Propensity matched analysis:Survival to discharge for head-up CPR was 7.6% (27/353) versus 2.8% (10/353) for conventional CPR (adjusted OR, 2.84 [95% CI, 1.35–5.96]). Survival to discharge with favorable neurological outcome for head-up CPR was 4.2% (15/353) versus 1.1% (4/353) for conventional CPR (adjusted OR, 3.87 [95% CI, 1.27–11.78]).
Debaty, France, 202420	Prospective observational study n = 122 (63 in the intervention group)	Head-up CPR bundle that includes an automated controlled head and thorax patient positioning device, ACD and ITD.	Before Intervention (Patients who received conventional CPR with supine position)	There was no statistically significant difference in the rate of ROSC between the head-up CPR and conventional CPR (33.3% [21/63] versus 32.2% [19/59], p = 0.89).	Between head-up CPR and conventional CPR, there was no statistically significant difference in 30-day survival (11.1% [7/63] versus 13.6% [8/59], p = 0.68), or in 30-day survival with favorable neurological outcome (11.1% [7/63] versus 13.6% [8/59], p = 0.68).

ACD: Active compression/decompression, CPR: Cardiopulmonary resuscitation, ITD: Impedance threshold device, OHCA: Out-of-hospital cardiac arrest, ROSC: Return of spontaneous circulation, USA: United States of America.

^*While the study reported that the conventional CPR group consisted of 1,179, 2,728, and 1,258 patients from the ROC-PRIMED, ROC-ALPS, and ResQTrial studies, respectively, the supplemental table in their report calculating ROSC rates indicated slightly different numbers: 1,192, 2,825, and 1,335 patients from the same studies, for a total of 5,352.

The studies by Moore et al. and Bachista et al. obtained data for the intervention group of patients from the same registry (Table 2). This registry is referred to as the ACE (Automated Controlled Elevation)-CPR registry (2019–2020) in the Moore et al. paper and as the AHUP (automated head/thorax-up positioning)-CPR registry (2019–2021) in the Bachista et al. paper. Moore et al.’s study examined the time interval from the 9–1–1 emergency call to the initiation of head-up CPR in patients with both shockable and nonshockable rhythms, whereas Bachista et al.’s study focused exclusively on patients with nonshockable rhythms. To obtain their comparator patients, both studies used the large National Institutes of Health (NIH)-funded RCTs conducted approximately 10 years earlier (Table 2).22, 23, 24.

Table 2.

Comparison of two head-up CPR studies.

	Moore Study18	Bachista Study19
The registry from which the head-up CPR group was obtained	ACE (Automated Controlled Elevation)-CPR registry	AHUP (Automated Head/Thorax-Up Positioning)-CPR registry
Registry Timeframe	2019–2020	2019–2021
Comparator Sources	ROC PRIMED study (2007–2009)22 ROC ALPS study (2012–2015)23 ResQTrial (2006–2009)24	ROC PRIMED study (2007–2009)22 ResQTrial (2006–2009)24

CPR: Cardiopulmonary resuscitation

Interventions used across studies

All four studies used a bundle approach, including ACD and ITD. While Pepe et al.¹⁷ utilized a scoop stretcher to elevate the head and torso by placing a hard case toward the top of the stretcher with a mechanical CPR device attached to the scoop stretcher, the other three studies employed an automated head/thorax-up positioning device for head and torso elevation. The interventions used in each study are summarized in Table 1.

Risk of bias

Regarding the risk of bias, all four studies were rated as having serious concerns (Table 3). Specifically, pre-and-post design studies by Pepe et al. and Debaty et al. were assessed as having a serious risk of bias due to several factors, including concerns relating to confounding.17, 20 Studies by Moore et al. and Bachista et al. were also assessed as having a serious risk of bias due to several factors, including concerns related to confounding and selection bias.18, 19

Table 3.

Risk of bias assessment. Non-randomized studies assessed by ROBINS–I.

Study	Confounding	Selection	Classification of interventions	Deviation from intended interventions	Missing data	Measurement of outcomes	Selective reporting	Overall
Pepe, 201917	Serious	Low	Low	Low	Serious	Low	Serious	Serious
Moore, 202218	Serious	Serious	Low	Low	Serious	Low	Low	Serious
Bachista, 202419	Serious	Serious	Low	Low	Serious	Low	Low	Serious
Debaty, 202420	Serious	Low	Low	Low	Low	Low	Low	Serious

Outcomes at hospital discharge

For the critical outcomes of survival to hospital discharge with a good neurological outcome and overall survival to hospital discharge, the systematic review identified very-low-certainty evidence (downgraded due to serious risk of bias, inconsistency and imprecision) from four observational studies (Table 4).17, 18, 19, 20

Table 4.

Certainty of evidence for non-randomized studies in head-up CPR and head-up CPR bundle.

Certainty of evidence for non-randomized trials in head-up CPR and head-up CPR bundle
Outcomes	Studies	Risk of Bias	Inconsistency	Indirectness	Imprecision	Other	Conclusion
Return of spontaneous circulation	3 studies18, 19, 20	Serious	Moderate	Not serious	Serious	None	Very low
Event survival	2 study17, 20	Serious	Moderate	Not serious	Serious	None	Very low
Survival to hospital discharge	4 studies17, 18, 19, 20	Serious	Serious	Not serious	Serious	None	Very low
Survival to hospital discharge with a good neurological outcome	4 studies17, 18, 19, 20	Serious	Serious	Not serious	Serious	None	Very low
Survival at 30 days	1 study20	Serious	NA	Not serious	Serious	None	Very low
Survival to 30 days with good neurological outcome	1 study20	Serious	NA	Not serious	Serious	None	Very low

The observational study by Pepe et al.,¹⁷ including 2322 adult OHCA patients, compared outcomes before, during and after the introduction of the head-up/torso-up chest compression technique. Metrics such as the average emergency medical services (EMS) crew response intervals, relative frequency of ECG presentations, gender, and frequency of bystander-witnessed cases were comparable between the groups. Details on survival with good neurological outcomes was limited, noting that that about 35–40% of those resuscitated achieved “intact neurologic status”, defined as “modified Rankin Score < 3” in both the pre- and post-intervention groups “wherever tracked”. Missing rates in both groups were unreported.

The study by Moore et al.¹⁸ examined 227 adult OHCA patients who received the head-up CPR bundle enrolled in the ACE-CPR registry between 2019 and 2020, and 5,196 adult OHCA patients who underwent conventional CPR with supine positioning enrolled in three RCTs conducted between 2005 and 2015 at high-performing pre-hospital systems in the United States.22, 23, 24 The study found no statistically significant difference for survival to hospital discharge or survival to hospital discharge with favorable neurological status between the head-up CPR group and the conventional CPR group. The adjusted odds ratio of cumulative survival to hospital discharge between head-up and conventional CPR groups, based on the time interval from the 9–1–1 emergency call to head-up CPR start after propensity-score matching, was 1.65 (95% CI 0.93–2.94) for < 20 min and 0.82 (95% CI 0.23–2.97) for 20–38 min and was not statistically significant. Similarly, the adjusted odds ratio for cumulative survival to hospital discharge with favorable neurological function between head-up CPR and conventional CPR groups was 1.85 (95% CI 0.91–3.74) for < 20 min and 0.42 (95% CI 0.05 – 3.39) for 20–38 min; neither was statistically significant (Table 1).

The study by Bachista et al.¹⁹ focused on patients with nonshockable rhythms and included 380 adult OHCA with nonshockable rhythms who received the head-up CPR bundle in the AHUP-CPR registry.¹⁸ As a comparison group, the study included 1852 adult OHCA with nonshockable rhythms who received conventional CPR with supine positioning enrolled in two different RCTs in the United States.22, 24 The study showed significantly greater unadjusted survival to hospital discharge in the head-up CPR group (Table 1), which remained significant after propensity score matching, 7.6% (27/353) in the head-up CPR group versus 2.8% (10/353) in the conventional CPR group (adjusted OR 2.84, 95% CI 1.35–5.96). The head-up CPR bundle was also associated with higher probabilities of survival with favorable neurological function (4.2% [15/353] vs. 1.1% [4/353], adjusted OR 3.87, 95% CI 1.27–11.78).

The study by Debaty et al.²⁰ included 122 adult OHCA patients (59 in the pre-intervention period and 63 in the post-intervention period with head-up CPR bundle) in France. The study found no statistically significant difference between head-up CPR and conventional CPR in survival rates (11.1% [7/63] vs. 13.6% [8/59], p = 0.68) at hospital discharge.

Outcomes at 30 days

For the critical outcomes of survival to 30 days and survival to 30 days with good neurological outcome, we identified very-low-certainty evidence (downgraded due to serious risk of bias and imprecision) from one observational study^20. The study by Debaty et al.²⁰ found that there was no statistically significant difference between head-up CPR and conventional CPR in 30-day survival or 30-day survival with favorable neurological outcome (Table 1).

Event survival

For the important outcome of event survival, we identified very-low-certainty evidence (downgraded due to serious risk of bias, inconsistency and imprecision) from two observational studies17, 20 The observational study by Pepe et al.¹⁷ demonstrated an increased rate of hospital arrival with sustained ROSC 17.9% (n = 806) in the pre-intervention period to 27.5% during the transition (n = 160), and to 34.2% (n = 1356) in the post-intervention period. The study by Debaty²⁰ showed no statistically significant difference in survival on hospital admission between head-up CPR and conventional CPR.

ROSC

For the important outcome of ROSC, we identified very-low-certainty evidence (downgraded due to serious risk of bias, inconsistency and imprecision) from three observational studies.18, 19, 20 All three studies show no difference between groups for ROSC.

Subgroup analysis

We initially considered conducting subgroup analyses for cardiac arrests with shockable and nonshockable rhythms. However, only one of the four included studies¹⁹ reported outcomes based on cardiac rhythms; therefore, we did not conduct the planned subgroup analysis.

Discussion

This systematic review provides an update on the use of the head-up CPR strategy during cardiac arrest since the previous review conducted by ILCOR in 2021.¹⁰ In this systematic review, we identified very low certainty evidence for the head-up CPR strategy (i.e., head-up CPR alone and head-up CPR bundle). We recognized that the currently available evidence is still limited, highlighted by the absence of RCTs or observational studies with contemporary comparisons. As mentioned above, two of the four observational studies identified in the review used a pre-and-post study design. The other two studies obtained their comparator patients from NIH-funded RCTs conducted approximately 10 years earlier, and resuscitation guidelines and practices have evolved since then.

The included studies used different head-up CPR methods. Although the bundle approach that includes head-up position with automated head/thorax-up positioning device, ACD, and ITD has been adopted by certain EMS agencies in the United States and France, the systematic review did not find clinical evidence supporting a particular bundle approach or indicating that the sole use of head-up elevation is superior to other bundles. For example, the aforementioned study by Pepe et al.¹⁷ described a head-up CPR method in which a scoop stretcher was used to elevate the head and torso. This approach differs from the newer head-up CPR bundle, which uses an automated head/thorax-up positioning device rather than a stretcher. Another example is a pilot study conducted by Kim et al. in Korea in 2022.²¹ The study, which lacked a comparison group, described a method that involved using a 15 cm high wedge on the bed to raise the head approximately 15 cm without elevating the chest, while using a mechanical CPR device but no other devices. Future research should define the best approach (e.g., angle, use of other devices).

Timing of the head elevation might also be an important factor. Animal studies suggest that the greatest cerebral perfusion pressure is achieved with a 2-minute priming period in a flat position, followed by gradual elevation of the head and thorax over an additional 2 min when combined with the use of ACD and ITD.25, 26 An observational study conducted by Moore et al. focusing on the impact of time to deployment of the head-up CPR bundle, showed that faster deployment was associated with a higher incidence of ROSC.²⁷ This study, along with previous animal studies, suggests that faster deployment is associated with better neurological outcomes. However, clinical studies on this topic are limited, and further study is needed on this matter.

This present study identified several current knowledge gaps. First, there are no RCTs that evaluated the effect of head-up CPR or head-up CPR bundle. Second, head-up CPR has primarily been assessed as a bundle with mechanical CPR with ACD and the use of an ITD. Finally, if head-up CPR proves to be beneficial, the optimal approach—such as the angle and timing of head elevation still needs to be determined in the future.

Limitations

The present systematic review has several limitations. First, although we used standard and widely accepted tools to assess the risk of bias, the decisions are subjective. Second, the review did not identify any studies involving children or inpatient cardiac arrest, so we were unable to comment on the efficacy of the head-up strategy for those patients. Third, as there was significant heterogeneity and an overlap of study subjects between studies a meta-analysis was not performed.

Conclusion

This systematic review identified very low certainty evidence for the effects of head-up CPR or head-up CPR bundle on survival and neurological outcomes. The substantial heterogeneity and reliance on historical controls in the current evidence pose major challenges to its interpretation. High-quality research is required before this treatment can be considered for clinical practice.

Patient consent for publication

Not required.

CRediT authorship contribution statement

Tatsuya Norii: Writing – original draft, Project administration, Methodology, Investigation, Formal analysis, Conceptualization. George Lukas: Writing – review & editing, Methodology, Investigation, Formal analysis, Conceptualization. Aloka Samantaray: Writing – review & editing, Methodology, Investigation, Formal analysis, Conceptualization. Mio Yabuki: Writing – review & editing. Theresa M. Olasveengen: Writing – review & editing, Supervision, Methodology. Janet E. Bray: Writing – review & editing, Supervision, Methodology.

Ethics approval

Not required.

Funding

This systematic review was funded by the American Heart Association, on behalf of The International Liaison Committee on Resuscitation (ILCOR). None of the following authors received payment from this funding source to complete this systematic review. Janet E Bray is funded by a Heart Foundation of Australia Fellowship (#104751).

Declaration of competing interest

The authors declare the following financial interests/personal relationships which may be considered as potential competing interests: Some of the Task Force collaborators (G Debaty and L Morrison) have published manuscripts related to head-up CPR which are included in this review. However, none of the authors have any financial conflicts of interests and none of the authors have academic conflicts related to ongoing or planned trials. Authors with identified conflicts of interest as per the guidance of the ILCOR Conflict of Interest Committee were not involved in the decision to include/exclude those articles and did not perform the initial data extraction or bias assessment. Janet E Bray is an Editor and Theresa M Olasveengen is an Editorial Board Member for Resuscitation Plus. Janet E Bray and Theresa M Olasveengen are Editorial Board Members for Resuscitation.

Appendix A. Supplementary material

The following are the Supplementary data to this article: