Development of site-specific predictive equations for epicutaneo-caval catheter insertion depth by body weight in neonates

Introduction

Background

Epicutaneo-cava catheters (ECCs) are peripherally inserted central venous catheters, which are essential for providing intravenous nutrition and medication administration for critically ill newborns (1,2). However, the length of the vena cava in newborns, particularly in extremely low birth weight (ELBW) infants, is significantly shorter than that in adults. The superior vena cava may measure only 2–3 cm, and the inferior vena cava is comparably short (3). This anatomical characteristic makes catheter tip malposition a life-threatening risk factor (4-6). The accurate preoperative prediction of ECC placement length combined with individualized catheterization pathway planning can significantly improve catheter tip positioning accuracy, thereby reducing the risk of fatal complications such as tip malposition, thrombosis, and infection.

Rationale and knowledge gap

Previous studies have indicated that the insertion depth of ECCs is correlated with anthropometric measurements. Armbruster et al. found that neonatal weight and length were associated with ECC insertion depth (7). Zhang et al. conducted a comprehensive study on ELBW infants, revealing a significant positive correlation between ECC insertion depth and body weight (8). Another study identified body weight as a reliable indicator for ECC insertion depth in lower limb placements (9). Notably, these studies highlighted the necessity of implementing site-specific measurement protocols, as the vascular pathway distance from upper extremity access points (left or right arm) to the superior vena cava exhibits substantial interindividual variability due to marked anatomical variations. Body weight is a precise and easily obtainable neonatal metric, unaffected by the infant’s condition or position, and can be accurately measured with specialized scales. However, research exploring the correlations between ECC insertion depth and anthropometric measures is still relatively scarce. The limited studies available have offered valuable foundational insights, but their sample sizes may not sufficiently represent the diverse neonatal population, potentially introducing biases and limiting statistical power.

Objective

Consequently, there is a pressing need to develop predictive models based on large sample sizes to enhance the precision in evaluating the relationships between anthropometric measurements and ECC insertion depth. To address this issue, we collected data from a substantial cohort of neonates who underwent ECC procedures between January 2023 and June 2024. By employing linear regression, we aimed to develop site-specific predictive equations for ECC insertion depths based on body weight. The findings have the potential to enhance the accuracy and safety of ECC procedures in neonates, thereby minimizing associated complications and adverse events. We present this article in accordance with the TRIPOD reporting checklist (available at https://tp.amegroups.com/article/view/10.21037/tp-2025-519/rc).

Methods

Study population

The present study is a retrospective study. We extracted clinical data from the electronic medical records of 673 infants admitted to the Children’s Hospital, Zhejiang University School of Medicine from January 2023 to June 2024. Those were included if they met the following criteria: (I) underwent an ECC placement procedure; and (II) had a postnatal age of 28 days or less at the time of insertion. Infants who received ECCs after 28 days of birth were excluded (n=108). To establish a robust baseline model, those who had weight growth curves under the 25th and above the 75th percentile curves (n=78) were excluded, representing a cohort with well-proportioned growth as per the Fenton growth curves. This approach was chosen to minimize potential confounding from atypical body compositions in this initial study. If the number of cases who received ECCs at a specific site was less than 20, this site was excluded from data analysis (n=59). Thus, 428 infants were included for the analysis. Figure S1 shows the flow chart of selecting study population. Among 428 infants, 279 (65.2%) were diagnosed as very low birth weight (VLBW, <1,500 g), while within this group 118 infants (42.3%) were classified as ELBW (<1,000 g).

The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments (10). The study was approved by the Ethics Committee of the Children’s Hospital, Zhejiang University School of Medicine (No. 2024-IRB-0266). Informed consent was waived by the ethics committee of Children’s Hospital, Zhejiang University School of Medicine on the basis that this study constituted a secondary analysis of anonymized data with no identifiable privacy risks.

Data collection

We extracted data from the electronic medical records including date of birth, gender, body weight, length, head circumference, gestational age at birth, insertion site, ECC placement date, total catheter length, actual ECC placement length. The infant’s weight is measured daily in the morning by a nurse using an electronic infant scale, with the measurement verified by a second nurse. The weighing is performed once per day. For preterm infants, if the weight difference compared to the previous day exceeds 50 grams (either gain or loss), a re-measurement is required. The infant’s length and head circumference are measured weekly on Saturday by a nurse. If the length increases by more than 2 cm or the head circumference increases by more than 1 cm compared to the previous week, a re-measurement by two individuals is required for verification. To ensure measurement accuracy and minimize variability, a standardized positioning protocol was strictly followed. The actual ECC placement length is measured by the nurse before puncture, after identifying the insertion site, using a ruler for accuracy. For the upper limb, the measurement is taken with the arm abducted to 90 degrees, from the insertion site to the right sternoclavicular joint, and then downward to the third rib. For the lower limb, with the leg fully extended, the measurement is taken from the insertion site to the inguinal fold (groin), and then upward to the xiphoid process.

Insertion regions

Figure 1 illustrates the insertion regions. The main veins for venipuncture included the basilic vein, axillary vein, great saphenous vein, and small saphenous vein. The antecubital puncture region was defined as within 1 cm of the antecubital fossa in the basilic vein (Figure 1A). The axillary region was defined as within 1 cm of the axilla (Figure 1B). ECCs placed into the right or left greater saphenous veins (ankle or knee regions) were included. The knee site was defined as the segment within 1 cm of the popliteal fossa along the course of the greater saphenous vein (Figure 1C). The ankle region was defined as the segment of the great saphenous vein located 1 cm anterior to the medial malleolus (Figure 1C). The small saphenous vein is located above the space between the Achilles tendon and the posterior border of the lateral malleolus, within 1 cm of the lower 1/3 segment of the calf (see Figure 1D). According to the Infusion Nurses Society (INS) guidelines, the ECC tip inserted via the upper limb should be placed in the lower third of the superior vena cava or at the cavoatrial junction (CAJ) position, while the catheter tip inserted via the lower limb should be placed in the inferior vena cava above the diaphragm level (11,12).

Figure 1 Definition of puncture sites. (A) Antecubital puncture site; (B) axillary site; (C) knee and ankle sites; (D) small saphenous vein site.

Statistical analysis

First, we applied the descriptive analysis. Regarding the continuous variables, normality was tested. Means and standard deviations were calculated for the normally-distributed data, while medians and interquartile ranges (IQRs) for non-normally-distributed data. Numbers and percentages were calculated for categorical data. Second, we presented the ECC inserted depths in each region and applied the one-way analysis of variance (ANOVA) for comparison of differences in depths across various inserted regions. Third, correlations among all anthropometric indicators and ECC insertion depths were analyzed by Pearson correlation tests (see Figure S2). Given the existence of collinearity among corrected gestational age, length, weight, head circumference, only one of the anthropometric indicators with the largest correlation with ECC inserted depths entered the final linear regression, that is, body weight. Lastly, scatter plots and linear regression analysis were applied to assess the association between body weight and insertion depth at different sites, based on which we established the equations. To evaluate the performance of the equations, adjusted R² were calculated. Moreover, residuals of each linear regression model were tested. All statistical analyses were performed using R version 4.2. The statistical significance was indicated by P value <0.05.

Results

Table 1 presents the general characteristics of the study population. A total of 428 infants were included in the analysis. The median corrected gestational age at enrollment was 31.1 (IQR, 28.9, 33.7) weeks. Two hundred and twenty-eight (53.3%) were males. At the time of ECC placement, the median weight was 1.3 (IQR, 1.0, 1.8) kg; the median length was 38.0 (IQR, 34.0, 41.0) cm; and the median head circumference was 26.2 (IQR, 24.0, 29.0) cm. The distribution of ECC insertion sites was as follows: 56 (13.1%) cases were in the ankle region of the greater saphenous vein; 210 (49.1%) in the knee region of the greater saphenous vein; 62 (14.5%) in the small saphenous vein; 20 (4.7%) in the right basilic vein; 24 (5.6%) in the left basilic vein; 25 (5.8%) in the right axillary vein and 31 (7.2%) in the left axillary vein.

Table 2 presents the inserted depth of ECC in each region. The median inserted length of ECC was the largest in the greater saphenous veins, i.e., 22.2 cm, while the shortest in the left and right axillary veins, i.e., 7.0 cm. The difference in length at each site was statistically significant (P<0.001).

Figure 2 shows the scatter plots of body weight and ECC insertion length at each site. Table 3 presents coefficients and model fitting of linear regression models. Figure 3 visualizes the equations. For lower limb ECC placements, the equation for ECC length in the ankle region of the greater saphenous vein was “Length (cm) = 15.81+3.92 × Weight (kg)” (R²=0.850), whereas for the knee region, it was “Length (cm) = 11.11+3.62 × Weight (kg)” (R²=0.838). The equation for ECC placement length in the small saphenous vein was “Length (cm) = 14.02+4.31 × Weight (kg)” (R²=0.709). For upper limb placements, the equation for ECC placement length in the right basilic vein was “Length (cm) = 7.88+2.35 × Weight (kg)” (R²=0.920), whereas for the left basilic vein, it was “Length (cm) = 8.18+2.49 × Weight (kg)” (R²=0.902). The equation for ECC placement length in the right axillary vein was “Length (cm) = 4.40+1.46 × Weight (kg)” (R²=0.900), whereas for the left axillary vein, it was “Length (cm) = 4.89+1.63 × Weight (kg)” (R²=0.785).

Figure 2 Scatter plot of body weight and the inserted depth of ECCs at each site. ECC, epicutaneo-cava catheter.

Figure 3 Linear regression models for ECC insertion length at different sites. (A) All regions’ inserted depth of ECC; (B) greater saphenous veins (ankle regions); (C) greater saphenous veins (knee regions); (D) small saphenous veins; (E) right basilic veins; (F) left basilic veins; (G) right axillary veins; (H) left axillary veins. ECC, epicutaneo-cava catheter.

Discussion

The placement of ECCs is a common procedure in neonates; however, only a limited number of studies have explored the optimal insertion length based on anthropometric measurements (13,14). Postoperative imaging techniques, such as X-rays, electrocardiography, and ultrasound, play a critical role in confirming correct ECC positioning (15,16). Nevertheless, accurately predicting the insertion length prior to the procedure can significantly reduce the need for subsequent corrective actions, such as dressing changes, thereby minimizing the risk of contamination and infection at the sterile puncture site (17,18). Despite the importance of this issue, research on predictive models for ECC insertion length remains scarce. Therefore, our study provides novel firsthand data and robust evidence for predicting ECC insertion length using body weight in neonates. We developed site-specific equations for each ECC insertion site based on body weight, demonstrating strong predictive performance, with body weight explaining 70% to 92% of the variance in insertion length across different sites.

Our study revealed a significant association between the insertion length of ECCs at different anatomical sites and the body weight of neonates. These findings offer a valuable alternative for determining the optimal insertion length during ECC placement in neonates. Beyond traditional surface measurements, these predictive equations provide clinicians with a straightforward and accurate tool that is easy to implement in clinical practice. Unlike prior studies, our research stands out due to the development of site-specific equations tailored to each insertion site. These equations empower clinicians to select the most suitable site and corresponding insertion length based on an infant’s body weight. Consequently, this approach enhances the success rate of catheter placement and reduces the incidence of associated complications.

The predictive equations developed in this study demonstrated strong performance, as evidenced by the adjusted R² values. In the regression models, body weight accounted for 70.9% to 92.0% of the variance in ECC insertion length. Notably, the goodness of fit was highest (adjusted R²=0.920) for ECC placement in the right basilic vein, indicating that body weight could predict 92.0% of the variance in insertion length at this site. This aligns with the INS guidelines, which recommend the basilic vein as the preferred site for ECC placement. The R² values for the knee and ankle regions of the greater saphenous vein ranged from 0.838 to 0.850, significantly outperforming those reported in the Luister model for the greater saphenous vein (R²=0.609–0.666). This finding is also consistent with Armbruster et al., who reported a stronger correlation between body weight and insertion length in the ankle region (R²=0.971) (7). However, the R² value for the small saphenous vein (0.709) was slightly lower compared to other sites. The small saphenous vein offers a larger visible surface area on the lower limbs, with variable and flexible insertion points, which may introduce measurement bias. Furthermore, the relationship between anthropometric parameters and ECC insertion length at this site remains understudied, limiting the availability of comparative references.

This study employed a joint segmentation approach for the lower extremities. While most existing studies typically categorize insertion sites based on specific veins (19), this method may pose challenges in clinical practice. For instance, when inserting a catheter into the great saphenous vein, puncture sites can vary between the knee and ankle joints, leading to differences in the inserted catheter length depending on the chosen location. Consequently, accurately estimating variations in catheter length due to different puncture sites remains difficult. Similarly, Armbruster et al. adopted a joint segmentation approach in their research (7), highlighting the relevance of this methodology in addressing site-specific variations.

Strengths and limitations

This study offers several notable strengths. First, it includes a substantial sample size of 428 neonates, which is among the largest in published research on this topic to date. The study was conducted at a national medical center for child health in China, with one of the oldest and largest Neonatal Intensive Care Units (NICUs) in the country. Patients from across the nation seek care at this center, ensuring a diverse and representative study population. Moreover, our institution benefits from a dedicated neonatal intravenous therapy team that adheres to standardized protocols. The clinicians on this team possess over 5 years of specialized experience in neonatal ECC insertion, managing an annual caseload of 400–500 procedures. Additionally, 100–200 nurses from across the country receive training at our center each year, further enhancing the standardization and expertise reflected in our practices. Second, the predictive equations developed in this study demonstrated exceptional performance, with adjusted R² values ranging from 0.709 to 0.920—higher than those reported in prior studies—and residuals conforming to a normal distribution (see Figures S3-S10). This indicates that body weight serves as a robust predictor of ECC insertion length across various anatomical sites. Compared to other anthropometric measurements, such as height or head circumference, body weight is simpler and more convenient to measure in clinical settings, particularly for newborns. Consequently, our equations provide a feasible, validated, and user-friendly tool for healthcare professionals to optimize ECC placement.

There are several limitations in this study. A key limitation of our study is that its findings are primarily applicable to neonates with weights between the 25th and 75th percentiles. As a single-center, exploratory investigation, our aim was to first build a highly reliable model for this common population. Future multi-center studies with larger sample sizes are necessary to validate and extend these predictive equations for use across the full spectrum of neonatal weights, particularly for infants who are small or large for gestational age. Second, the sample was predominantly composed of infants with VLBW, who constituted 65.2% of the cohort (279 infants). Within this VLBW group, a remarkable 42.3% (118 infants) were classified as ELBW. However, we contend that this demographic accurately mirrors the clinical reality of a tertiary NICU, as these are precisely the infants who most frequently require ECC placement. Therefore, the focus on this specific population enhances the direct clinical applicability and relevance of our findings for the practitioners who serve these most vulnerable patients. As a result, the study does not fully represent the entire newborn population nor include all possible catheterization sites. This underscores the need for future research with a larger and more diverse sample to confirm the reliability and universality of the equations. Third, with the increasing availability of imaging technologies, such as ultrasound, for guiding ECC insertions, the necessity of predictive models like ours may be questioned. While imaging guidance offers precision, it is not universally accessible, particularly in resource-limited settings or smaller healthcare facilities where such technology may be unavailable or cost-prohibitive. In these contexts, our predictive equations, based on a simple and widely measurable parameter like body weight, provide a practical and cost-effective alternative to enhance the accuracy of ECC placement, improve success rates, and reduce complications. Finally, variations in clinical practices, such as differences in insertion techniques or equipment used, may influence the applicability of our equations in other institutions, further highlighting the importance of broader validation studies.

Conclusions

In this study, we developed seven site-specific predictive equations to estimate the insertion length of ECCs in neonates, using body weight as the primary predictor. These equations demonstrated strong predictive performance, with body weight accounting for 70.9% to 92.0% of the variance in ECC insertion length across different anatomical sites. By providing a reliable and accessible method to determine optimal insertion depths, these equations have the potential to significantly enhance the success rate of ECC placements and reduce the incidence of associated complications. Our findings offer a practical tool for clinicians, particularly in settings where advanced imaging may not be readily available, and pave the way for further validation and refinement in diverse neonatal populations.

Acknowledgments

We thank the methodological support from the Pediatric Evidence-based Medical and Clinical Research Laboratory, which is an internal institution of the National Clinical Research Center for Child Health in the Children’s Hospital, Zhejiang University School of Medicine. The authors extend their gratitude to Ms. Mengjia Chen for the exquisite 3D illustration (Figure 1) provided for this article.

Footnote

Reporting Checklist: The authors have completed the TRIPOD reporting checklist. Available at https://tp.amegroups.com/article/view/10.21037/tp-2025-519/rc

Data Sharing Statement: Available at https://tp.amegroups.com/article/view/10.21037/tp-2025-519/dss

Peer Review File: Available at https://tp.amegroups.com/article/view/10.21037/tp-2025-519/prf

Funding: This research was funded by Zhejiang Provincial Department of Education (grant No. Y202454288).

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tp.amegroups.com/article/view/10.21037/tp-2025-519/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved. The study was conducted in accordance with the Declaration of Helsinki and its subsequent amendments. The study was approved by the Ethics Committee of the Children’s Hospital, Zhejiang University School of Medicine (No. 2024-IRB-0266). Informed consent was waived by the ethics committee of Children’s Hospital, Zhejiang University School of Medicine on the basis that this study constituted a secondary analysis of anonymized data with no identifiable privacy risks.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Barone G, Pittiruti M. Epicutaneo-caval catheters in neonates: New insights and new suggestions from the recent literature. J Vasc Access 2020;21:805-9.
2. Biasucci DG, Disma NM, Pittiruti M. editors. Vascular Access in Neonates and Children. Springer Cham; 2022.
3. Racadio JM, Doellman DA, Johnson ND, et al. Pediatric peripherally inserted central catheters: complication rates related to catheter tip location. Pediatrics 2001;107:E28. [Crossref] [PubMed]
4. Elmekkawi A, Maulidi H, Mak W, et al. Outcomes of upper extremity versus lower extremity placed peripherally inserted central catheters in a medical-surgical neonatal intensive care unit1. J Neonatal Perinatal Med 2019;12:57-63. [Crossref] [PubMed]
5. Bashir RA, Swarnam K, Vayalthrikkovil S, et al. Association between Peripherally Inserted Central Venous Catheter Insertion Site and Complication Rates in Preterm Infants. Am J Perinatol 2016;33:945-50. [Crossref] [PubMed]
6. Zareef R, Anka M, Hatab T, et al. Tamponade and massive pleural effusions secondary to peripherally inserted central catheter in neonates-A complication to be aware of. Front Cardiovasc Med 2023;10:1092814. [Crossref] [PubMed]
7. Armbruster D, Slaughter J, Stenger M, et al. Neonatal Anthropometric Measures and Peripherally Inserted Central Catheter Depth. Adv Neonatal Care 2021;21:314-21. [Crossref] [PubMed]
8. Zhang M, Yang W, Yan L, et al. The correlation between weight gain and PICC tip shift in the very low birth weight infants. Transl Pediatr 2020;9:596-602. [Crossref] [PubMed]
9. Chen IL, Ou-Yang MC, Chen FS, et al. The equations of the inserted length of percutaneous central venous catheters on neonates in NICU. Pediatr Neonatol 2019;60:305-10. [Crossref] [PubMed]
10. World Medical Association Declaration of Helsinki. ethical principles for medical research involving human subjects. JAMA 2013;310:2191-4. [Crossref] [PubMed]
11. Gorski LA. The 2016 Infusion Therapy Standards of Practice. Home Healthc Now 2017;35:10-8. [Crossref] [PubMed]
12. Albrecht K, Breitmeier D, Panning B, et al. The carina as a landmark for central venous catheter placement in small children. Eur J Pediatr 2006;165:264-6. [Crossref] [PubMed]
13. Luister A, Khostwal N, Deindl P, et al. Recommendations for Peripherally Inserted Central Catheter Insertion Depths in Neonates. Neonatology 2023;120:263-7. [Crossref] [PubMed]
14. Wu J, Li J, Li H, et al. Comparison of formula-based PICC catheterisation versus common method for the treatment of newborns. Int J Clin Pract 2021;75:e14210. [Crossref] [PubMed]
15. Yi M, Zhang X. Multiple factors analysis for abnormal removal of peripherally inserted central catheter in low birth weight infants. Unrsing Practice and Research 2011;8:5-7.
16. Zhu LB, Liu L, Zhang TS, et al. A clinical study on the tip localization of peripherally inserted central catheter (PICC) guided by intracavitary electrocardiography in newborns: a randomised trial. Transl Pediatr 2021;10:2409-17. [Crossref] [PubMed]
17. Sharpe EL. Tiny patients, tiny dressings: a guide to the neonatal PICC dressing change. Adv Neonatal Care 2008;8:150-62; quiz 163-4. [Crossref] [PubMed]
18. Hill ML, Baldwin L, Slaughter JC, et al. A silver-alginate-coated dressing to reduce peripherally inserted central catheter (PICC) infections in NICU patients: a pilot randomized controlled trial. J Perinatol 2010;30:469-73. [Crossref] [PubMed]
19. Wen J, Yu Q, Chen H, et al. Peripherally inserted central venous catheter-associated complications exert negative effects on body weight gain in neonatal intensive care units. Asia Pac J Clin Nutr 2017;26:1-5. [Crossref] [PubMed]