The administration of medications or fluids is a critical step in neonatal resuscitation. Although only 0.5% of infants may require intensive resuscitation [1], these infants are typically the most critically ill and at risk of death or severe morbidities. In cases where umbilical venous catheterization is not feasible, IO needle insertion is widely accepted and effective. The NRP 8th edition recommends a single insertion site for needle placement, without specifying any differences between term and preterm infants [3]. However, various factors including gestational age, parental stature, and ethnicity can modify fetal growth. Previous studies showed that gestational age is a key predictor of fetal growth [12]. Tibial bones exhibit high rates of diaphyseal growth in utero [13]. Therefore, it is possible that tibial growth progresses with increasing gestational age, leading to changes in length and cortical thickness, which could potentially affect the depth and insertion site of the tibial bones. It is important to determine the appropriate IO insertion site and depth for each infant to ensure that the needle tip is positioned correctly without causing serious adverse events it is important to determine the appropriate IO insertion site and depth for each infant to ensure that the needle tip is positioned correctly without causing serious adverse events. Additionally, determining the appropriate depth of insertion using the sense of loss of resistance presents a challenge, particularly in infants with small tibial bones. Guidance regarding the insertion depth tailored to the size of each infant is therefore essential.

In our study, the typically recommended position for IO placement, which is 2 cm below the tibial tuberosity and 1–2 cm medially [3], was found to be at 1.46 ± 0.27 cm below the epiphyseal plate. Boon et al. [5] conducted a study involving the dissection of neonatal cadavers and IO needle insertion, 1 cm below the tibial tuberosity. They reported that the needle tip was approximately 10–15 mm away from the epiphyseal growth plate in this position. However, they did not measure the position at 2 cm below the tibial tuberosity due to difficulty in placement. Therefore, they suggested that the position 1 cm below the tibial tuberosity was suitable. We propound that this position may set the needle tip too close to the epiphyseal plate, posing a risk of injury, as evidenced by Schwindt et al. who reported the risk to be as high as 53% [14]. Therefore, we advocate for IO needle insertion at the recommended position of 2 cm below the tibial tuberosity, which is relatively safe and minimizes the risk of injury to the growth plate. Furthermore, the thin cortex observed at this position (mean: 0.16 cm) facilitates easier and faster insertion, requiring less pressure and reducing the likelihood of fracture. Cortical thickness measured in our cohort was similar to the study by Fuchs et al. who reported a thickness of 1.2 mm [11].

We observed an overall success rate of 86.8%, which was 66.7% in VLBW infants. This success rate was higher than the 40% to 61% range reported by Fuchs et al., dependent on the type of IO needles used [11]. During IO insertion, blood aspiration was observed from only 3 EZ-IO® needles across 9.1% of successful insertions, with one of these occurring in a VLBW infant. While blood aspiration was possible, we cannot be certain about the rate of successful blood aspiration since our study involved neonatal cadavers that had been deceased for a median of 8.0 [2.0, 12.0] hours, during which time blood may have clotted. Hence, the absence of aspirated blood does not necessarily indicate IO malposition.

Reported procedural complications from leakage include hematoma or subcutaneous fat necrosis [7]. Among the 6 insertions (15.8%) that exhibited contrast leakage, one EZ-IO® needle had its tip identified within the medullary cavity and displayed contrast filling. In real-world scenarios, instances like these may still permit some medication or infused volumes to flow into the systemic circulation. Therefore, we considered this as a successful placement. On the other hand, the remaining 3 EZ-IO® and 2 Acufirm® needles were not positioned within the medullary cavity and were thus deemed unsuccessful insertions. Overall, the EZ-IO® needles had a higher, non-significant rate of contrast leakage compared to the Acufirm® needles. This difference may be attributed to the tip design of the two needle types as outlined in the materials and methods section. The EZ-IO® needle predisposes the needle to potentially protrude beyond the medullary cavity, which exhibited a mean size of 0.21 ± 0.07 cm across all birth weight groups. Specifically, in extremely low birth weight (ELBW; BW < 1000 g) infants, the mean medullary cavity size was measured at 0.16 ± 0.03 cm, increasing the likelihood of cortical penetration on the opposite side. A higher incidence of contrast leakage occurred in VLBW infants, particularly in the subgroup utilizing EZ-IO® needles. The lack of statistical significance may be attributed to an insufficient sample size to adequately demonstrate a difference. Fuchs et al. [11], also found that IO insertions with a butterfly needle had a 2.4-fold significantly higher odds ratio of appropriate needle placement compared to EZ-IO® needles. While EZ-IO® needles can indeed be manually drilled into the bone, we opted to utilize the DAD, which is widely practiced. Therefore, our observations could potentially stem from either the type of needle tips or the insertion mechanism. However, since the depth of insertion was not different between the needle types (Table 2), we believe that it can be more likely attributed to the type of the needle tip. Additionally, since we did not observe any fractures with either the DAD or manual insertion using a twisted hand motion, we recommend the use of conventional manual needles, such as the Acufirm®, over the EZ-IO® for IO insertion, particularly in ELBW infants.

Suominen et al. investigated the medullary diameter of neonatal tibial bones using x-ray imaging in full-term infants aged 1–28 days and reported a diameter of 7.7 ± 0.4 mm [6], which is considerably wider than the mean medullary cavity diameter of 2.6 mm observed in our neonatal cadavers weighing >2500 g. Several potential reasons may account for the discrepancy between the studies. First, Suominen et al. included 10 full-term infants in their study, whereas the three cadavers in our >2500 g group had gestational ages of 36, 38, and 39 weeks, with postnatal ages ranging from 2 to 7 days. Hence, both studies may have measured neonates with different demographic characteristics. Second, the techniques and position used for measurement differed. Suominen et al. measured the medullary diameter in the antero-posterior and lateral dimensions of x-rays taken 1 cm below the proximal end of the tibia due to technical difficulties in identifying the tibial tuberosity, while we utilized CT scanning images to measure the cross-sectional dimension at 2 cm below the tibial tuberosity. Although the measurement positions by Suominen et al. were also higher than in our study, they are close to our widest medullary cavity diameter of 5.3 mm in >2500 g infants. Last, potential differences may exist due to ethnic factors that contributed to variations in neonatal size. Fuchs et al. investigated formaldehyde-fixed stillbirth cadavers with gestational ages ranging from 26 to 43 weeks, utilizing spectral-CT examination to confirm successful insertion by identifying contrast media in the marrow cavity. The median diameter of the bone marrow cavity at the proximal tibia was reported as 4.0 mm [11], which was wider than our mean medullary diameter of 2.1 mm at the insertion position. However, it is important to note that Fuchs et al. documented the width of the medullary diameter at the proximal tibia, but the exact measurement level was not well defined. Therefore, their findings of the widest medullary cavity measurement of 4.7 mm may closely align with our study.

Based on our measurements, we recommend a minimal insertion depth of 0.62 ± 0.22 cm and a maximum distance of 0.83 ± 0.25 cm. However, the suggested depth should be selected based on the infant’s BW group. There may be concerns about exceeding the depth, especially with the DAD method. The EZ-IO® needle for neonates has a maximum depth of 15 mm, marked by a 1 mm thick black line. The distance from the bottom edge of the black line to the base of the diamond-shaped tip is approximately 10 mm. Additionally, considering the 2.5 mm length of the diamond-shaped tip, inserting the needle until the bottom edge of the black line with the stylet removed would result in a needle tip depth in the marrow of approximately 5–7.5 mm (Fig. 1). Regarding the insertion depth for the Acufirm® needle, it requires estimation based on the tip and total length of the needle. In the clinical setting, we propose convenient insertion depths for each birth weight group based on the estimated success rates presented in Supplementary Table 3. For VLBW infants, the insertion depth should be less than 0.5 cm or not exceeding the bottom edge of the black line on the EZ-IO® needle. Infants weighing between 1500–2499 g should have an insertion depth between 0.5–0.75 cm, approximately at the level of the black line. For infants weighing 2500–3500 g, the insertion depth should be approximately 1 cm or extending a few millimeters beyond the upper edge of the black line to the base of the EZ-IO® needle. For infants with a birth weight ≥3500 g, the needle should be inserted until the base. Hence, to enhance the likelihood of successful IO needle insertion, additional techniques may be necessary. These could include utilizing an angled rather than a perpendicular needle insertion or employing an ultrasound-guided placement [15]. These techniques can improve accuracy and reduce the risk of complications during the procedure [14]. Furthermore, we observed in our study that contrast media injected through IO needles could flow into the abdominal aorta and reach the heart, even in the absence of spontaneous circulation. This confirms that it is possible to flush a medication volume of just 1 mL through an IO needle to reach the heart.

This study aimed to investigate the appropriate insertion site and depth for IO needle placement in Asian infants, recognizing their typical smaller, anthropometric measurements. The positive attributes of our study are the apriori calculation of the sample size and the meticulously planned methodology. The assessment of the proper insertion site was conducted using standardized recommended positions, and needle insertion success was evaluated by confirming the presence of both the needle tip and contrast media in the marrow cavity, to ensure internal validity. Additionally, our study provides depth recommendations for IO needle insertion in Asian infants, categorized by BW groups. Despite the focus on only Asian neonatal cadavers, the cortical thickness and medullary cavity width was similar to previous studies conducted on infants of different ethnicities. Hence, the recommended insertion site and depth can be generalized and extrapolated for use in all infant populations without limitation to Asian neonates.

Several limitations of our study merit consideration. First, we did not exclude infants with hydrops fetalis recognizing that the IO route for the administration of medications and fluids during cardiovascular resuscitation, may be preferentially warranted over the chance of unsuccessful umbilical venous catheter placement in an edematous cord. Two infants with hydrops fetalis may have had thicker skin compared to normal infants, but their measured skin thickness ranged from 0.36 to 0.40 cm and was comparable to the other cadavers. Nevertheless, if IO needles need to be employed in infants with severe edema, deeper needle insertion may be necessary to compensate for the abnormal skin thickness at the insertion site. Second, the number of cadavers were relatively small to explore differences in the success rate between the types of IO needles. The success rate observed with the Acufirm® needle in our study can be extrapolated to other needle brands that share the same bevel tip design. However, all the insertions were performed solely by the principal investigator of the study, which potentially enhanced the success rate beyond what might occur in real-world scenarios, where physicians performing IO insertions may not have extensive expertise. Lastly, we conducted our study on neonatal cadavers, which may undergo post-mortem changes that could potentially affect the measurement of skin thickness. However, the infants in this study were examined at a median duration of 8.0 [2.0, 12.0] hours after death and were stored in a cool environment to minimize post-mortem changes if the examination exceeded 2 h after death. Based on previous histological examination of skin changes after death [16], the epidermal and dermal layers mostly remained intact during this timeframe. Therefore, we feel that any potential effects on skin thickness measurements would be minimal. The clear definitions of insertion sites and depths from this study’s results afford generalizability of our findings and may increase confidence and improve the likelihood of successful insertions.