Intraosseous versus intravenous vascular access in out-of-hospital cardiac arrest: a systematic review and meta-analysis of randomized controlled trials

Abstract

Background

Rapid and reliable vascular access is crucial during cardiopulmonary resuscitation for out-of-hospital cardiac arrest (OHCA). While intraosseous (IO) and intravenous (IV) access are used, their comparative effectiveness for patient outcomes remains uncertain.

Methods

We searched PubMed, Embase, and ClinicalTrials.gov for RCTs comparing IO vs. IV access in adults with OHCA. The primary outcome was survival (30 days or until discharge), while secondary outcomes included sustained ROSC, favorable neurological outcome, successful first-attempt vascular access, and time from emergency medical service arrival to access. Pooled odds ratios (OR), mean differences (MD), and 95% confidence intervals (CI) were calculated.

Results

Four RCTs with 9475 patients were included. No significant differences were found between IO and IV groups in survival (6.6% vs. 6.9%, OR 0.99, 95% CI 0.84–1.18) or favorable neurological outcome (4.7% vs. 4.6%, OR 1.07, 95% CI 0.88–1.30). The sustained ROSC rate was numerically, but not significantly, lower in IO vs. IV access (24.6% vs. 27.0%, OR 0.92, 95% CI 0.80–1.06). IO access had a higher first-attempt success rate (92.3% vs. 62.3%; OR 6.18, 95% CI 3.50–10.91) and was 15 s faster than IV for vascular access (IO: 11.03 ± 5.57, IV: 11.35 ± 6.16 min, MD − 0.25, 95% CI − 0.48 to − 0.01).

Conclusions

IO access had a higher first-attempt success rate and faster establishment than IV access, but no significant differences were found in survival or favorable neurological outcomes in adults with OHCA. Sustained ROSC was numerically lower with IO access than IV access, although the difference was not statistically significant.

Graphical abstract

Introduction

Rapid and reliable vascular access is essential during cardiopulmonary resuscitation (CPR) for patients with out-of-hospital cardiac arrest (OHCA), as timely administration of drugs, such as epinephrine, significantly impacts their effectiveness [1]. Intravenous (IV) and intraosseous (IO) access are critical methods used in advanced life support to facilitate drug delivery during resuscitation efforts [2].

The optimum route for drug administration in patients with OHCA remains uncertain [3]. The IO route has recently gained popularity because of its relative ease and speed in securing vascular access [1, 4]. On the other hand, the IV route has been suggested to offer a more favorable pharmacokinetic profile for drugs administered during resuscitation [3]. Contemporary guidelines from the American Heart Association and European Resuscitation Council suggested prioritizing the IV route for administering drugs during OHCA and reserving the IO route in cases of IV access failure [2, 5]. This recommendation was supported mainly by pooled results from retrospective observational studies, which suggested worse clinical outcomes with the IO method than with IV access [6]. However, as acknowledged by the guidelines, there was substantial heterogeneity across the observational studies regarding study design, selection protocols for access routes, adjustments for resuscitation variables, patient characteristics, and IO access sites [7,8,9,10,11,12,13]. In some of these studies, IO access was applied only when the IV route was unsuccessful or when the choice of vascular access was based on paramedics' judgments [8,9,10,11, 14]. Under these circumstances, patients who received IO access were possibly in more critical condition, and the findings from observational studies were subject to resuscitation time bias [15]. Secondary analyses of some randomized controlled trials (RCTs) of patients with OHCA for different resuscitation interventions suggested no significant effect modification by the drug administration route [14, 16, 17]. However, these trials did not randomize the patients based on vascular access and were underpowered to assess differences between IV and IO access [14, 16, 17].

Recently, some RCTs have provided valuable insights into the choice of vascular access routes in adults with OHCA but were not individually powered for clinical outcomes and had conflicting results in terms of some endpoints, such as sustained return of spontaneous circulation (ROSC) between the IO and IV access routes [18,19,20,21]. Additionally, it is still unknown whether certain patient-related resuscitation factors, such as bystander CPR, witnessed status, time to vascular access, initial rhythm type, and vascular access site, play a role in the choice of vascular access. Therefore, we pooled high-quality data from RCTs comparing the effectiveness of IO versus IV access in adult patients with OHCA, aiming to estimate the comparative effectiveness of each vascular access type across the overall study population and within key prognostic subgroups.

Methods

Protocol and registration

This systematic review adhered to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines [22]. The protocol of this study is registered in PROSPERO (registration number CRD4202461536). Institutional review board approvals were not required since the study utilized study-level results.

Data sources and search strategy

Systematic searches were conducted in PubMed and Embase to identify RCTs investigating the effectiveness of initial attempts at IO or IV vascular access in patients who experienced OHCA (Supplemental Table 1; the last search date was November 20, 2024). No restrictions were applied regarding language or publication date. The references of the included RCTs and other relevant studies were manually searched to detect additional studies. Furthermore, we searched ClinicalTrials.gov to identify ongoing RCTs.

Eligibility criteria and study selection

We sought trials that specifically randomized patients with OHCA based on the type of vascular access—either IO or IV—where resuscitation efforts were initiated and immediate vascular access was required. Adult patients (age ≥ 18 years) with OHCA due to non-traumatic or traumatic causes were eligible. We excluded non-original publications (reviews, commentaries, or editorials), non-randomized investigations, and randomized studies based on other resuscitating interventions. Two investigators (S.A. and S.R.) independently screened the titles, abstracts, and full texts of the selected studies against the eligibility criteria for inclusion. Any disagreements were resolved through discussion with each other.

Outcomes

The primary effectiveness outcome was survival, measured at a 30-day follow-up or until hospital discharge. Other effectiveness outcomes included the sustained ROSC, favorable neurological outcome, successful first-attempt vascular access, and time from emergency medical service (EMS) arrival on scene to vascular access. The sustained ROSC was defined in case resuscitation efforts were initially successful, further chest compressions were not required for 20 min, and signs of circulation persisted for at least 20 min [23]. A favorable neurological outcome (at 30 days or until discharge) was defined as a modified Rankin Scale score of 3 or less or a Cerebral Performance Category score of 1 or 2 [24].

The safety outcome of interest was vascular access-specific adverse events, including extravasation, inflammation, necrosis, phlebitis, osteomyelitis, compartment syndrome, traumatic bone fracture, and inability to weight-bearing [25, 26]. Serious adverse events were defined as those requiring either inpatient hospitalization or prolongation of existing hospitalization, resulting in persistent or significant disability or incapacity, being life-threatening, or resulting in death.

Subgroups

Subgroup analyses were prespecified based on IO vascular access sites (humeral and tibial) and several demographic and clinical patient characteristics, including sex (male or female), initial cardiac rhythm (shockable or non-shockable), presence of a witness during the cardiac arrest (either a bystander or EMS involvement versus no witness), and CPR performance (either performed or not performed).

Data extraction and quality assessment

Two independent reviewers (S.A. and S.R.) performed data extraction using a standardized data collection form. Detailed demographic and clinical data, including age, sex, location of cardiac arrest, initial cardiac rhythm, presence of a witness, CPR performance, and key clinical outcomes, were collected. The type of vascular access (IO or IV), the time from EMS arrival to vascular access, and the drugs administered were also documented.

The second version of the Cochrane Risk of Bias tool for randomized trials (RoB-2) was utilized to evaluate the methodological quality of the included RCTs [27]. This tool assesses RCTs across five domains: the randomization process, deviations from intended interventions, missing outcome data, outcome measurement, and selective reporting of results, ultimately providing an overall risk of bias for each trial [27]. Additionally, The Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) system was used to report the robustness of generated estimates across trials for each outcome [28]. Two investigators (S.A. and S.R.) independently conducted the data collection and quality assessment. All discrepancies were resolved by discussion with each other.

Statistical analysis

For categorical outcomes, pooled odds ratios (ORs) with corresponding 95% confidence intervals (CIs) were calculated using the Mantel–Haenszel method. For continuous outcomes, pooled mean differences (MDs) with 95% CIs were estimated using the inverse variance method. Random-effect models were used due to the presumed high heterogeneity between the RCTs. If the mean and standard deviation were not reported, we estimated them based on the sample size, median, and interquartile range values using the methods described previously [29]. Statistical heterogeneity among studies was assessed using Cochran’s Q test and the Higgins’ I² statistic, with I² values less than 25% indicating low heterogeneity, between 25 and 50% indicating moderate heterogeneity, and over 50% indicating high heterogeneity [30, 31].

Sensitivity analyses were conducted by excluding one trial at a time from the meta-analyses to test each trial’s contribution to the pooled estimates. In subgroup analyses, we examined all the prespecified outcomes of interest based on IO vascular access sites (i.e., humeral and tibial). For other prespecified subgroups, we tested for effect modification regarding the primary outcome (survival). The analyses were conducted in Stata (StataCorp LLC, version 17.0) and RevMan (version 5.4), and a P-value of less than 0.05 was considered significant for all the analyses. Before the analyses, we acknowledged that subgroup analyses can be underpowered and conducted for idea generation.

Results

Among the 1240 identified records, 67 were sought for full-text review (Supplemental Fig. 1). Ultimately, four RCTs conducted between 2011 and 2024 in the United States, United Kingdom, Taiwan, and Denmark were included [18,19,20,21].

Fig. 1

Baseline characteristics of the included randomized controlled trials a PARAMEDIC-3 is a pragmatic RCT designed to evaluate interventions' effectiveness in real-world clinical settings, reflecting routine practice conditions. b VICTOR is a pragmatic cluster RCT with clusters of advanced life support ambulance teams randomized biweekly to perform IO or IV vascular access. c The trial did not explicitly randomize between the humeral and tibial sites for IO access. Paramedics selected the anatomical location of vascular access based on personal preference and patient characteristics, informed by the available evidence. d Among the 731 patients in the IO access group, 361 were randomly assigned to undergo humeral vascular access, and 370 were assigned to undergo tibial vascular access. e Vascular access was randomized to one of 3 locations: the proximal tibial IO, proximal humeral IO, or peripheral IV. IO Intraosseous; IV Intravenous; N/A Not available; RCT Randomized controlled trial; UK United Kingdom; US United States

These four RCTs included 9475 patients with OHCA (Fig. 1, Supplemental Table 2). Overall, 4627 (48.9%) patients were randomized to receive IO access and 4,848 (51.1%) to receive IV access. All studies included patients with non-traumatic cardiac arrest, except for one trial, which also included 86 (0.9% of the whole population) patients with traumatic OHCA [19]. The follow-up durations varied across the studies: one study followed the patients for 30 days [19], another extended the follow-up period to 90 days [18], and one followed patients up to hospital discharge [20]. One study did not follow up the patients for clinical outcomes [21].

The mean age of patients across four trials was 67.5 ± 15.3 years, with 33.2% (3,118/9,396) being female. The data for some baseline patient characteristics were not available for the trial by Reades et al. [21], and a small portion of patients from the PARAMEDIC-3 trial [19]. We excluded the missing or unavailable data and calculated the rates accordingly. The most common location of cardiac arrest was home (7,249/9,264; 78.2%) and public area (1,779/9,264; 19.2%), respectively. Cardiac arrest was witnessed in 57.8% (5,328/9,209), and CPR was performed in 71.6% (6,645/9,280) of patients. The initial cardiac rhythm was non-shockable in 77.6% (7,133/9,181) of patients.

Methodological quality of the included RCTs

The included RCTs were open-label, and only the IVIO trial featured blinded event adjudication for its prespecified outcomes [18]. However, as our primary outcome was survival, we did not consider that lack of blinding significantly affected the reported outcomes. Overall, all trials were considered to have a low risk of bias. Nonetheless, for other clinical and procedural outcomes, it should be noted that the trials other than the IVIO trial did not perform blinded event adjudication and are at unclear risk of bias for these outcomes [19,20,21] (Supplemental Fig. 2 and Table 3A, B).

Primary effectiveness outcome

There was no significant difference in terms of survival (up to 30 days or until discharge) between IO and IV access in patients with OHCA (6.6% vs. 6.9%; OR 0.99, 95% CI 0.84–1.18, P = 0.94, I² = 8%) (Fig. 2).

Fig. 2

Pooled results for the clinical outcomes of interest CI Confidence interval; CPR, cardiopulmonary resuscitation; EMS Emergency medical service; OR Odds ratio; ROSC Return of spontaneous circulation

Other effectiveness outcomes

The sustained ROSC rate was numerically lower in the IO group compared with the IV group (24.6% vs. 27.0%). However, this difference did not reach statistical significance (OR 0.92, 95% CI 0.80–1.06, P = 0.25, I² = 44%). No significant difference was found between IO and IV insertion for favorable neurological outcome (4.7% vs. 4.6%, OR 1.07, 95% CI 0.88–1.30, P = 0.49, I² = 0%) (Fig. 2).

Procedural outcomes

In terms of vascular access-related outcomes, we observed that the successful first-attempt vascular access rate was significantly higher in the IO group compared with the IV group (92.3% vs. 62.3%, OR 6.18, 95% CI 3.50–10.91, P < 0.001, I² = 94%) (Fig. 3).

Fig. 3

Pooled results for vascular access-related outcomes CI Confidence interval; EMS Emergency medical service; MD Mean difference; OR Odds ratio

Figure 4 depicts the time intervals from cardiac arrest to hospital arrival in each trial. There was a significant difference regarding time from EMS arrival on scene to vascular access between the two groups (IO: 11.03 ± 5.57, IV: 11.35 ± 6.16 min, MD − 0.25, 95% CI − 0.48 to − 0.01, P = 0.04, I² = 0%) (Fig. 3). Based on these findings, the IO method provides vascular access approximately 15 s faster than the IV method.

Fig. 4

Time intervals of the included trials Time intervals are reported in median (interquartile range) or mean ± standard deviation. EMS, emergency medical service; IO, intraosseous; IV, intravenous

Safety outcomes

In the PARAMEDIC-3 trial, a patient in the IO group reported one specific adverse event (ongoing mild leg pain during certain activities). In the IVIO trial, prespecified adverse events for patients with sustained ROSC were uncommon and were limited to extravasation (< 1% for both the IO and IV groups). The other two trials reported no other complications. None of the four RCTs reported serious adverse events in either the IO or IV arms.

Humeral IO/Tibial IO vs. IV

Subgroup analyses based on the IO access site (humeral and tibial) revealed no statistically significant interactions in survival, sustained ROSC, favorable neurological outcome, and successful first-attempt access rate (Fig. 5A). For the time from EMS arrival to vascular access, contrasting time trends were observed: humeral IO showed a trend toward longer access times than the IV route, while tibial IO showed a trend toward faster access than the IV method, although the differences did not reach statistical significance (Fig. 5B).

Fig. 5

Pooled results for humeral IO vs. tibial IO vs. IV vascular access CI Confidence interval; EMS Emergency medical service; IO Intraosseous; IV Intravenous; MD Mean difference; M–H Mantel–Haenszel; OR Odds ratio

Humeral IO vs. Tibial IO

Additional findings from the IVIO trial comparing IO methods (humeral vs. tibial) revealed no significant differences in survival (10% vs. 13%, OR 0.73, 95% CI 0.46–1.15, P = 0.17) and sustained ROSC (30% vs. 31%, OR 0.97, 95% CI 0.71–1.33, P = 0.85). However, the favorable neurological outcome was significantly lower in the humeral IO group compared with the tibial IO group (7% vs. 11%, OR 0.58, 95% CI 0.35–0.98, P = 0.04) (Fig. 5C). In terms of procedural outcomes, data from the IVIO and Reades et al. trials showed that the successful first-attempt access rate was numerically lower in the humeral IO than in the tibial IO (78.8% vs. 88.0%, OR 0.29, 95% CI 0.05–1.76, P = 0.18), which was not statistically significant. Additionally, there was no significant difference in terms of time between EMS arrival on scene and vascular access between these two groups (humeral IO: 6.99 ± 4.48, tibial IO: 6.42 ± 4.24 min, MD 1.12, 95% CI − 0.68 to 2.93, P = 0.22) (Fig. 5D).

Subgroup analyses

Subgroup analyses for the primary effectiveness outcome (survival) did not indicate signals for effect modification considering sex (male vs. female) (P_interaction = 0.20), the status of witnessed arrest (witnessed vs. unwitnessed) (P_interaction = 0.56), bystander CPR performed or not (P_interaction = 0.82), and initial rhythm (shockable vs. non-shockable) (P_interaction = 0.22) (Fig. 6).

Fig. 6

Subgroup analysis for the primary outcome (survival) a Includes patients witnessed by emergency medical service or bystanders. CI Confidence interval; CPR Cardiopulmonary resuscitation; OR Odds ratio

GRADE certainty ratings

The certainty of the pooled estimates, assessed using the GRADE system, was moderate for survival and favorable neurological outcome, low for sustained ROSC and successful first-attempt access, and high for time from EMS arrival to vascular access (Supplemental Table 4).

Sensitivity analyses

In sensitivity analyses, which were performed by leaving one trial at a time out of the meta-analyses for the outcomes of interest, there was no change in the significance of the pooled estimates except for sustained ROSC, which, after excluding the IVIO trial, IO access was associated with lower rate compared with IV access (OR 0.87, 95% CI 0.78–0.96). Additionally, the findings related to the time from EMS arrival on scene to vascular access were no longer statistically significant after excluding the PARAMEDIC-3 (MD − 0.04, 95% CI − 0.45 to 0.37) and Reades et al. (MD − 0.22, 95% CI − 0.55 to 0.11) trials (Supplemental Table 5).

Discussion

This systematic review and meta-analysis, synthesizing data from four RCTs, provides comprehensive evidence on the comparative effectiveness of IO and IV vascular access in adult patients with OHCA. IO access achieved a higher first-attempt success rate and a slightly shorter time to vascular access compared to IV access. However, the pooled analysis found no significant differences in key clinical outcomes, including survival and favorable neurological outcome, between the two groups. The sustained ROSC rate was numerically lower in the IO group compared with the IV group (24.6% vs. 27.0%), though this difference was not statistically significant (Fig. 7). These findings support IV access as the preferred first-line approach when feasible. However, IO access remains a viable alternative in cases where IV access is difficult or delayed.

Fig. 7

Meta-analysis summary

Previous studies have shown that each minute delay in epinephrine administration was associated with a 4% decrease in survival and ROSC in patients experiencing OHCA [32, 33]. Establishing IV access in cardiac arrest situations with collapsed peripheral veins is challenging. On the other hand, IO access can provide fast, non-collapsible entry to the venous plexus in the bone marrow, with a relatively high success rate [1, 34, 35]. In this meta-analysis, IO access consistently resulted in more than six-fold higher first-attempt success rates than IV access across all trials (OR 6.18, 95% CI 3.50–10.91), translating to a 15-s decrease in the time to vascular access. It is also noteworthy to mention that, regardless of the type of vascular access, there were substantial differences in time intervals for vascular access between the trials (Fig. 4). Specifically, the IVIO and Reades et al. trials reported faster vascular access times than the PARAMEDIC-3 and VICTOR trials. This variation may be attributable to differences in trial design, as PARAMEDIC-3 and VICTOR (in contrast to the IVIO and Reades et al. trials) were pragmatic trials and more closely reflected real-world clinical practice where procedural delays are more common. Additionally, differences in the EMS systems, paramedic training, and regional protocols may have contributed to this variation. These findings highlight the importance of minimizing delays and optimizing vascular access strategies in real-world practice to improve efficiency in OHCA management. Regarding safety outcomes for each access, no serious adverse events (such as tissue necrosis, osteomyelitis, or compartment syndrome) were reported in the trials.

Despite faster and higher vascular success rates with IO compared with the IV method, we did not observe any differences in clinical outcomes. Some pharmacokinetic and biological mechanisms may contribute to these findings. First, the anatomical site of administration (upper vs. lower body) may influence the effective delivery of a drug to the heart during active CPR, regardless of whether the drug is administered IV or IO. Experimental studies have shown that drugs delivered through an IV route reaching the heart via the inferior vena cava during active CPR experience delayed time-to-peak and lower peak concentrations than those delivered via the superior vena cava [36]. Second, in adults, red bone marrow is replaced by less vascular yellow bone marrow, and the density of this yellow bone marrow also decreases with age, potentially affecting drug absorption from the IO route [37]. Third, adipose-rich tibial yellow bone marrow exhibits a depot effect, acting as a reservoir after drug administration and gradually releasing the drug into circulation [38, 39], especially for lipophilic medications such as amiodarone, which are expected to be highly absorbed in fatty marrow [17]. Fourth, owing to diminished blood flow during circulatory shock or in response to vasopressors, fluid boluses and pressurized infusions are sometimes necessary to enhance drug movement into the central circulation, with infusion rates being more affected and decreased in IO access than in IV access [40, 41].

While the current meta-analysis compared IO and IV access in adult patients with OHCA, it is important to recognize that outcomes for IO access may vary depending on the specific site of placement, i.e., humeral or tibial IO access. The pooled analysis suggested a trend that tibial IO insertion may be faster than IV access (MD − 0.53, 95% CI − 1.67 to 0.60). In contrast, the humeral IO showed a trend in the opposite direction (MD 0.39, 95% CI − 0.13 to 0.91). However, the findings did not reach statistical significance, possibly due to insufficient power. The comparative effectiveness of IO access sites (humeral and tibial) requires further investigation [42, 43].

The International Liaison Committee on Resuscitation (ILCOR) recently reviewed this topic [44]. The ILCOR recommended IV access, compared with IO access, as the first attempt for vascular access during adult cardiac arrest (weak recommendation, low certainty evidence) [45]. The ILCOR added that if IV access cannot be rapidly achieved within two attempts, it is reasonable to consider IO access as an alternative route for vascular access during adult cardiac arrest (good practice statement). These recommendations were largely driven by the findings of the PARAMEDIC-3 trial, which demonstrated reduced sustained ROSC with IO compared with IV access (OR 0.85, 95% CI 0.74–0.98) [19]. Our meta-analysis of three RCTs, aligned with these recommendations, showed a reduced sustained ROSC with IO access compared with IV access (24.6% vs. 27.0%, OR 0.92, 95% CI 0.80–1.06), suggesting a potential preference for IV access, though this difference did not reach statistical significance. Moreover, as noted by the ILCOR [44] and current meta-analysis, there were no significant differences between IO access and IV access regarding longer-term outcomes, including 30-day survival and favorable neurological outcomes. Furthermore, paramedics are generally more familiar with IV access [20]; thus, IV access can be considered as the preferred first-line approach for patients with OHCA when feasible. However, IO access remains a viable alternative in cases where IV access is difficult or delayed.

Generalizing the findings of this meta-analysis to the wide spectrum of adult patients with OHCA requires specific considerations. First, the predominant population in this study consisted of patients with non-traumatic cardiac arrest, with fewer than 0.9% experiencing traumatic OHCA. Therefore, these findings may not apply to traumatic OHCA. Second, some of the included trials excluded patients above 80 years of age (due to concerns for high risk for osteoporosis) and pregnant patients [19, 20]. Further research is needed to ascertain the optimal vascular access route in these specific populations.

This study had some limitations. First, we estimated the mean and standard deviation of time from EMS arrival to vascular access based on the median and interquartile range for some trials with unavailable data [29]. The method applied in this meta-analysis has been used in prior meta-analyses of continuous outcomes and is considered the best method of estimation [46, 47]. Second, in the subgroup analyses, we did not observe signals for effect modification based on patient-related resuscitation factors or IO access sites (humeral vs. tibial) between IO and IV access. These findings should be interpreted with caution due to limited statistical power. Furthermore, data limitations prevented us from exploring the effects of age and certain comorbidities (e.g., previous neurological disorders) between the IO and IV access groups. An individual participant data meta-analysis of RCTs could offer valuable insights into these subgroups. Third, there was heterogeneity in the trial protocols regarding vascular access attempts and study design. Fourth, the trials did not report the details of post-resuscitation care, which may impact long-term clinical outcomes. However, it can be assumed that by virtue of randomization, post-resuscitation factors were similarly distributed between the IO and IV arms.

Conclusion

This systematic review and meta-analysis of four RCTs showed that in patients with OHCA, IO access was associated with a higher first-attempt success rate and a modestly faster time (15 s) from EMS arrival on scene to vascular access. However, no significant differences were observed in key clinical outcomes, including survival and favorable neurological outcome, between IO and IV vascular access for adult patients with OHCA. The rate of sustained ROSC was lower with IO access compared with IV access, though the difference was not statistically significant.