Introduction
Gaining vascular access in patients in the prehospital environment often is crucial and challenging. In the circumstance in which an intravenous (IV) access is delayed or not obtainable, an alternative site is required, and often may be achieved using an intraosseous (IO) device. While IO infusions traditionally have been used in pediatric patients, the frequency of use in adults is growing.Reference Fowler, Gallagher and Isaacs 1 Since Drinker et al first proposed the technique 90 years ago,Reference Drinker, Drinker and Lund 2 interest has been overshadowed by the development of the IV route.Reference Fowler, Gallagher and Isaacs 1 , Reference Engle 3 The Australian Resuscitation Council recognizes that while IV access is the first line choice, it should be attempted for no longer than 90 seconds during the management of a victim of cardiac arrest.Reference Tibballs 4 If unsuccessful, an IO insertion is recommended as both a safe and necessary substitute.Reference Tibballs 4 Similar approaches to difficult vascular access have been suggested by Gazin et al,Reference Gazin, Auger and Jabre 5 who proposed that only two peripheral venous attempts be made before attempting IO insertion in cardiac arrest victims.
Currently in Australia, only intensive care paramedics have authorization to insert an IO needle. The majority of these paramedics use the manual intraosseous infusion technique (MAN-IO) using the Cook needle (Cook Med. Inc., Bloomington, Indiana USA). Other manual IO needles that are available include the Jamshidi (Cardinal Health, McGaw Park, Illinois USA) and the Sur-Fast (Cook Critical Care, Bloomington, Indiana USA) needles. The more recently evolved semiautomatic devices, such as the FAST-1 (Pyng Medical Corporation, Richmond, British Columbia, Canada); the EZ-IO (Vidacare, Shavano Park, Texas USA); and the Bone Injection Gun, BIG (Waismed Ltd., West Hempstead, New York USA), are used by some flight paramedics. Current literature suggests that there is a move toward the use of semiautomatic devices in other prehospital settings throughout the world.Reference Byars, Tsuchitani and Erwin 6 , Reference Findlay, Johnson and Macnab 7 While this paper focuses on the Australian paramedic system, the current trends suggest that this review of the IO literature has application and relevancy for paramedic systems beyond Australia.
The availability of a variety of IO devices requires that the Australian paramedic sector respond to the rapidly changing technologies based on sound evidence-based research. Moreover, the findings from empirical research should form the platform for a cost-benefit analysis, and assist in determining whether the currently used manual technique should be replaced with the use of the more contemporary semiautomatic intraosseous devices. While a body of knowledge exists on user satisfaction with the various IO devices, less literature is available on the direct clinical benefits associated with the use of IO infusions. To the authors’ knowledge, no prehospital-based research has been undertaken examining IO insertion success rates or clinical outcomes in Australia. The purpose of this paper is to review the literature on IO insertion in the prehospital setting.
Methods
A comprehensive search strategy was conducted to include both peer-reviewed and non-peer-reviewed literature. A prehospital search filter initially developed by the Cochrane prehospital field Reference Smith, McDonald, Wasiak, Jennings, MacPherson and Archer 8 (Table 1) was used to search the following electronic databases for articles published during the dates indicated: Medline (US National Library of Medicine, Bethesda, Maryland USA), 1950 to the end of May 2010; Embase (Elsevier B.V., Amsterdam, The Netherlands), 1974 to the end of May 2010; and CINAHL (EBSCO Publishing, Ipswich, Massachusetts USA), 1986 to the end of May 2010. The following medical subject heading (MeSH) terms in the US National Library of Medicine index were used: “infusion,” “intraosseous,” “IO,” “vascular access,” “bone injection,” and “devices.” In addition, a manual search of relevant grey-literature was performed.
Table 1 Filter Used in Search Strategy
The inclusion criteria consisted of any published studies in the prehospital environment or relating to paramedics reporting intraosseous insertion and/or infusion in both adult and pediatric patients. Studies involving animals were included if they related the findings to in-field practice on humans. Non-English papers were excluded.
Results
A total of 2,100 articles were identified. Following the elimination of duplicates and a review and critique of each article, a total of 20 papers met the inclusion criteria. Table 2 describes the study types, sample sizes, findings, and study limitations.
Table 2 Studies Referenced in Analyses
Abbreviations: BIG, bone injection gun; EZ-IO, semiautomatic intraosseous infusion system; FAST-1, semiautomatic intraosseous infusion system; IQR, interquartile range; MAN-IO, manual intraosseous device; peds, pediatrics; s, second(s); y, year(s).
aOut of hospital.
bSaphenous vein.
Semiautomatic Intraosseous Devices
EZ-IO—The device was independently examined in three separate studiesReference Frascone, Jensen, Wewerka and Salzman 9 - Reference Levitan, Bortle and Snyder 11 involving a combined total of 114 insertions in patients,Reference Frascone, Jensen, Wewerka and Salzman 9 , Reference Horton, Beamer, Horton and Beamer 10 and 297 insertions in cadavers.Reference Levitan, Bortle and Snyder 11 Each of these studies reported high insertion success rates (range: 94.0 to 97.3%).Reference Frascone, Jensen, Wewerka and Salzman 9 - Reference Levitan, Bortle and Snyder 11 Complication rates associated with the use of the EZ-IO device included infiltrations, slow flow rates, and needle dislodgment.Reference Frascone, Jensen, Wewerka and Salzman 9 Horton et al reported a 6% rate of insertion failures and a patient complication rate of 4% with use of the EZ-IO device.Reference Horton, Beamer, Horton and Beamer 10
User satisfaction with the EZ-IO was high, with the majority of users reporting that they felt comfortable or very comfortable with the device.Reference Frascone, Jensen, Wewerka and Salzman 9 Reported insertion times varied among the three studies. Frascone et alReference Frascone, Jensen, Wewerka and Salzman 9 reported that the time required for insertion of the EZ-IO was <60 seconds in 72% of the insertions, although these data were obtained retrospectively via telephone interviews. Horton et alReference Horton, Beamer, Horton and Beamer 10 reported insertion times of <10 seconds in 77.2% of the insertions, although the time interval measured only the time from skin contact to proper insertion of the needle. Levitan et alReference Levitan, Bortle and Snyder 11 reported an average insertion time of six seconds as measured from the time of skin contact until the stylet was removed.
FAST-1—This intraosseous device was investigated in three studiesReference Findlay, Johnson and Macnab 7 , Reference Macnab, Christenson and Findlay 12 , Reference Miller, Guimond and Hostler 13 involving nine in-field insertions and 39 scenario-based insertions. Miller et alReference Miller, Guimond and Hostler 13 reported a 55% success rate of first insertion attempts, while Macnab et alReference Macnab, Christenson and Findlay 12 reported an 84% insertion success rate among first-time users, and a 95% success rate among experienced practitioners. The data regarding complication rates in each of these studies was not clearly articulated, or not examined.
Bone Injection Gun—Three studies have evaluated the use of the bone injection gun (BIG) with 229 insertions in living humans, and 13 insertions in cadavers.Reference Gerritse, Scheffer and Draaisma 14 - Reference Hubble and Trigg 16 Overall, the insertion success rate in adults ranged from 73.0 to 92.3%.Reference Gerritse, Scheffer and Draaisma 14 - Reference Hubble and Trigg 16 Insertion of the BIG device in children was successful in 73% of reported attempts.Reference Gerritse, Scheffer and Draaisma 14 The complication rates associated with insertion using this device ranged from 7% to 27% and consisted of issues with the device and operator failures. The issues with the device were needle not firing, needle not securely placed, no flow achieved, and bone but not marrow entered. The operator issues identified were failure to remove the trochar needle after successful insertion, misidentifying bony landmarks, and missing the bone altogether. Additionally studies reported extravasation as a complication.Reference Gerritse, Scheffer and Draaisma 14 - Reference Hubble and Trigg 16 In terms of user satisfaction, Gerritse et alReference Gerritse, Scheffer and Draaisma 14 reported that 80% of respondents deemed the use of the BIG device to be “very satisfactory.” Inexperienced paramedic students gave the use of the BIG a procedure score of 96.2 out of a possible 100, although it is unclear if the rating scale demonstrated appropriate psychometric properties.Reference Hubble and Trigg 16
The average insertion time of the BIG device was 3.9 minutes, which was the time to fluid flow in the cadaver model study.Reference Hubble and Trigg 16
Manual Intraosseous Devices
Eight independent studies evaluated the use of manual intraosseous devices (MAN-IO), mainly investigating the Cook or the Jamshidi. One studyReference Fiorito, Mizra and Doran 18 compared the Baxter IO needle (Baxter Healthcare Corporation, Deerfield, Illinois USA) and the Monoject IO needle (Sherwood Medical Company, St. Louis, Missouri USA) in a total of 400 insertions.Reference Anderson, Arthur and Kleinman 17 - Reference Smith, Keseg, Manley and Standeford 24 The reported insertion success rates in these studies ranged from 76 to 94%.Reference Anderson, Arthur and Kleinman 17 - Reference Smith, Keseg, Manley and Standeford 24 Glaeser et alReference Glaeser, Hellmich and Szewczuga 19 reported the insertion success rate to be higher in pediatric patients <3 years of age compared with older children. Infiltration and local edema were the most commonly reported complications occurring in up to 13% of the manual insertions.Reference Anderson, Arthur and Kleinman 17 - Reference Miner, Corneli and Bolte 20 , Reference Seigler 22 , Reference Smith, Keseg, Manley and Standeford 24 Other reported complications can be classified as device- or operator-specific. Issues with the needle were dislodgement and bending, while operator issues again pertained to wrong location and failure to adhere to landmarks. Additionally, hematoma and fluid leak were reported.
In terms of user satisfaction, Smith et alReference Smith, Keseg, Manley and Standeford 24 reported that 10 out of 11 operators scored the Jamshidi manual intraosseous device as “excellent” regarding its ease of use; the other studies did not assess this variable. The measurement of insertion times of the manual intraosseous devices was not reported in any of the studies, with the exception of the study by Seigler.Reference Seigler 22 He reports the insertion time to be less than one minute in between 47% and 76% of attempts.Reference Seigler 22 , Reference Seigler, Tecklenburg and Shealy 23 Conversely, between 17% and 23% took more than three minutes or required initiation of another bone marrow site. Smith et alReference Smith, Keseg, Manley and Standeford 24 estimated all IO insertions to be less than 30 seconds.
Comparisons of Intraosseous Devices
There were three papers comparing different devices within the same study.Reference Hartholt, van Lieshout and Thies 25 - Reference Brenner, Bernhard and Helm 27 Hartholt et alReference Hartholt, van Lieshout and Thies 25 compared the Jamshidi 45G, the BIG 15G and the FAST-1. The Jamshidi was successful in 91.7% and the FAST-1 in 89.5% of attempts. This differed significantly (P = .010) from the low success with the BIG 15G (59.1%). Regarding the pediatric population, only the Jamshidi 15G and the BIG 18G were compared. The Jamshidi was successful in 100% of its 12 attempts, while the BIG 18G was successful in seven out of 10 patients. This difference was not statistically significant. They also reported that insertion of the semiautomatic FAST-1 and the Jamshidi IO needle was successful in 90% of attempts, compared with a 59% insertion success rate of the semiautomatic BIG device (P = .010). Regarding pediatric patients in particular, no statistically significant difference was detected, although there was a trend toward the manual Jamshidi being superior.
Frascone et alReference Frascone, Jensen, Kaye and Salzman 26 compared the use of the semiautomatic EZ-IO device with the semiautomatic FAST-1 IO device and found the successful insertion rates to be 72% and 87%, respectively (P = .009). In a study comparing the semiautomatic EZ-IO device with the MAN-IO technique, Brenner et alReference Brenner, Bernhard and Helm 27 reported first attempt success rates of 97.8% and 79.5%, respectively, (P < .01), and failure to obtain access rates of 0% and 12.8%, respectively, (P < .02). The same study also found complication rates to be more frequent with the use of the MAN-IO technique compared with the semiautomatic EZ-IO device (15.4% vs. 0%, respectively; P < .01).
Hartholt et alReference Hartholt, van Lieshout and Thies 25 and Frascone et alReference Frascone, Jensen, Kaye and Salzman 26 concluded that the differences in user satisfaction between the MAN-IO and semiautomatic EZ-IO device were insignificant statistically (P = .52 and .13 respectively) while Brenner et alReference Brenner, Bernhard and Helm 27 reported the EZ-IO to be preferable to the MAN-IO method (1.9 vs. 1.2 on a scale from 1 to 6; P < .01) in their randomized, prospective trial. The average insertion times for both devices were identical.Reference Brenner, Bernhard and Helm 27
In a comparison of the MAN-IO and the FAST-1, accurate placement was achieved more quickly with the use of the Jamshidi needle, 37 (P25- P75 = 30-49) vs. 62 (P25- P75 = 50-131) seconds, respectively (P = .002).Reference Hartholt, van Lieshout and Thies 25
Discussion
Due to the nature of situations requiring an IO insertion, prospective, in-field research is ethically and clinically challenging. Consequently, investigations have been conducted in animalsReference Halvorsen, Bay and Perron 28 - Reference Halm and Yamamoto 30 and classroom-based experiments,Reference Fuchs, LaCovey and Paris 31 , Reference Shavit, Hoffmann, Galbraith and Waisman 32 which limit the ability to generalize the findings of the studies reviewed. Another confounding factor is related to the variations in sample sizes, design, and definitions of outcome measures. There are some limitations across all of the 20 papers included in this study. Several of the studies did not include unpacking, set-up, and preparation time when calculating device insertion times. This may have contributed to some of the reported contrasting findings regarding insertion times. It would seem logical that measuring variables, such as insertion times, should reflect real-time practice.
Several of the studies of IO use in real-life situations relied on retrospective questionnaires to assess outcomes, thereby diminishing the accuracy and external validity of the results. Moreover, due to the infrequent utilization of IO device insertions, most sample sizes are restricted, and subsequently lead to statistically insignificant findings. A further confounder is that some IO devices are utilized for the administration of fluids only, while others are utilized for the administration of drugs. Although some studies included other interesting components related to IO device insertion (i.e., ease of teaching, and skill acquisition and decay), the data provided are too sparse to draw conclusions. However, they do provide a good platform for future research projects, as evidenced in the recent work by Byars et al.Reference Byars, Tsuchitani and Erwin 6
Overall, the paucity of research relating to the FAST-1 device, plus the biasing role some authors hold as both researchers and company shareholders, emphasize the difficulty in generalizing the findings more broadly; a greater examination of its utility is needed.
Interestingly, the study by Gerritse et alReference Gerritse, Scheffer and Draaisma 14 was undertaken in the Netherlands, where insertions are performed by physicians based in the prehospital setting. Whether similar results are found in non-physician-based prehospital care systems in other countries, such as Australia, the United States, or the United Kingdom, remains to be seen. In terms of the user satisfaction ratings in that study, the potential for bias must be considered as representatives from the manufacturer of the device facilitated the instructions on its usage.
Hubble and Trigg,Reference Macnab, Christenson and Findlay 12 who studied IO insertions by inexperienced paramedic students, argue that the teaching and learning process involved in successful IO insertion is relatively straightforward, suggesting that the procedure might be extended to use by intermediate-level paramedics.
While the majority of articles reviewed noted that intraosseous access (regardless of technique) offers a safe and simple method for gaining access to the patients’ circulation, none of the studies clarify which device may be most appropriate for Australian paramedics to use in the prehospital setting. Overall, one IO device does not clearly appear to be better clinically. This suggests the need for cost-benefit analysis and financial feasibility studies.
Conclusions
In most Australian states, manual insertion of an intraosseous needle is the technique used by intensive care paramedics; however, the literature suggests that semiautomatic devices may be more effective. Further research, including cost-benefit analysis, is required at a national level to examine skill acquisition and decay, and whether other devices offer clinically relevant differences.
Abbreviations
- BIG:
-
bone injection gun
- EZ-IO:
-
semiautomatic intraosseous infusion system
- FAST-1:
-
semiautomatic intraosseous infusion system
- IO:
-
intraosseous
- IV:
-
intravenous
- MAN-IO:
-
manual intraosseous device
