Estimating Insertion Length of Umbilical Catheters in Newborn Infants

Overview

Insertion Length of Umbilical Catheters Umbilical catheters, inserted through the umbilical artery and vein of newborn babies at birth, are crucial in neonatal care of sick babies. They allow delivery of medication and fluids and to provide access for blood sampling and blood pressure monitoring. Incorrect positioning of the catheter tip due to under- or over-insertion length can lead to significant complications in newborn infants. Currently, several methods are used to estimate insertion length of umbilical catheters based on one of two beliefs; that the insertion length of the umbilical catheter is proportional to either the infant's birth weight or an external length measurement. Several research studies have identified that existing methods often result in incorrect positioning of umbilical catheters, with studies showing a variable range of proportions of umbilical lines being correctly or incorrectly placed. In particular, formulas for predicting umbilical venous catheter (UVC) length have been shown to be particularly unreliable. The investigators propose a new observational study which uses a novel but easy-to-measure external length measurement, the sternal notch (upper end of breast-bone) to umbilicus (upper margin of belly-button) length, along with other clinical information to develop a more reliable formula for estimating the insertion length of umbilical venous and arterial catheters to an appropriate length. Our study population will include newborn babies admitted to the neonatal unit requiring umbilical venous (UVC) and/or arterial catheterisation (UAC) over a two-year period. Demographic information will be recorded for each child and once position has been confirmed, the new external length will be measured. New formulae for estimating required insertion length will be developed using statistical (regression) analysis.

Introduction Umbilical catheters, both arterial (UAC) and venous (UVC), are an integral part of neonatal care to provide access for blood sampling and blood pressure monitoring, and to allow for the delivery of intravenous medication and fluids. It is important to ensure the correct position of the catheter tip on first attempt to minimise additional handling, further radiological exposure and risk of hospital-acquired infection. Other complications of malposition of the catheters include injury of surrounding structures, false passages, bleeding, and injury to other organs including the heart, the liver, and nerves. In one study, the most common complication with umbilical venous catheters was incorrect position. The British Association of Perinatal Medicine (BAPM) advise that the ideal placement of a UVC should be at the thoracic 8-9 (T8-9) vertebrae (backbone) outside of the heart shadow, with those being placed below T10 being at high risk of extravasation injury. The UAC should lie in a high position above the level of the diaphragm in the descending aorta and below the subclavian artery. This correlates to being positioned between the upper border of T6 and the lower border of T10 with T8 being the ideal position. Currently, the recommended method to confirm the position of umbilical catheters is by radiological imaging in the form of antero-posterior chest and abdominal x-rays. Prior to insertion of umbilical catheters, it is routine practice to calculate the estimated length required for suitable placement. Currently, there are several methods for calculating length which are primarily based on one of two measurements; birth weight or an external length measurement. In 1966, Dunn identified in a post-mortem study of 50 newborn babies that the length of the catheter required correlated to external length measurements such as total body length and shoulder to umbilicus length. This study introduced a series of scatter graphs to estimate the length of UACs and UVCs in the clinical setting using these external length measurements. However, the external length proposed by the study, the shoulder-umbilical length, was not based on fixed anatomical landmarks, leading to potential errors. More importantly, the target placement for the UVC was in the heart chamber, which is now an unacceptable position. In addition, these catheters were often performed in an emergency setting with limited access to the graphs. Some of the external measurements were potentially affected by increased flexor tone of newborn infants and variation in methods used leading to inconsistencies, particularly with shoulder to umbilicus length. In view of limitations of external length measurements, Shukla et al identified a formula using body weight to calculate the required length of catheter for insertion. Retrospectively, their study involved 43 neonates with a UAC and 10 neonates with a UVC in suitable positions. They generated mathematical formulae based on birth weight and length of UVC and UAC; UAC length in centimetres was = (3 x birthweight in kg + 9) divided by 2 + 1 and UVC length in centimetres was ½ x UAC insertion length +1cm. Prospectively, they implemented their formulae to calculate the required length of insertion of 26 UACs and 16 UVCs of which all were in acceptable positions. However, there was no explanation in the paper as to how the final formula was derived, as this was different from the original statistical analysis. The authors also correctly cautioned about using this formula in babies who are small- or large-for-dates, as the formula would over- or under-estimate the length in such cases. However, this formula is widely used currently, leading to erroneous results. Evidently, these were both small studies and subsequently there have been multiple studies evaluating these two methods in the clinical setting in larger populations. In 2010, Verheij et al compared both the Dunn and Shukla methods in 153 patients in a single-centre prospective cross-sectional study in the Netherlands. The authors reported that the existing two methods often led to high positions of umbilical catheters, particularly UVCs. The Shukla method has more high-positioned UVCs (75% compared to 57% with the Dunn method, p\<0.05) and the Dunn method has more highly placed UACs (34% compared to 13% in Shukla group, p\<0.05). Furthermore, in both groups, with UVCs there was poor accuracy with correct positioning (24% in Shukla group and 41% in Dunn group, p\<0.05) to reach a correct position which was more accurate with UAC placement (up to 87% correctly placed in Shukla group and 63% in Dunn group, p\<0.05). A randomised control trial by Kieran et al in 2016 also compared the Dunn and Shukla methods in a tertiary neonatal unit in Ireland. The neonates were randomised and stratified according to weight with the outcome assessor blinded. There was no significant difference between correctly sited UVCs between the two methods however the overall rate of correct placement was low (28% in Dunn group compared to 31% in Shukla group, p value 0.826). The Shukla group was more reliable in UAC insertion with 91% in a correct position compared to 50% of the Dunn group (p\< 0.001) which was consistent with the study by Verheij et al. With particular reference to UVC insertion length, the Shukla formula was found to have a higher rate of over-insertion as previously mentioned. Verheij et al proposed a revised Shukla formula by removing the additional 1cm from the original formula; (3 x birthweight BW in kg + 9)/2 cm. This study of 185 neonates compared this revised formula to the original and found higher rates of correct placement their revised formula (43% compared to 26%, p\<0.05). The rate of over-insertion improved with their revised formula (54% versus 73% in Shukla group, p\<0.01) with no difference in under-insertion. However, this study was not randomised and their revised formula still resulted in 56% of UVCs being incorrectly positioned on x-rays. This formula was later compared with the Dunn and original Shukla methods for UVC insertion by Mutlu et al. They found that of 91 successfully inserted UVCs, only 56% were inappropriately positioned and there was no statistical difference between the three groups, highlighting the poor methods which exist for UVC length estimation. Although insertion rates for UACs are often more successful than for UVCs, both Dunn and Shukla only studied low numbers of very low birth weight (VLBW) infants and there has been consistent over-insertion of UACs with the Shukla formula. In view of this, Wright et al proposed a new formula (UAC length in cm = 4 x birth weight in kg + 7) for the insertion of UACs. When compared with the Dunn method in all gestations, 83% UACs were correctly sited with the Wright formula compared to 61% with the Dunn method (p=0.007). This was statistically significant in the term group but not the preterm infant group. 90% were correctly inserted in the preterm group with the Wright formula. However, when Lean et al compared 11 different formulae for UAC insertion length, they found a lower predicted success rates (57.3%) with birth weight related formulae like the Wright formula compared to external length measurement formulae. In this study, the formulae chosen depended on the clinician inserting the catheter and calculations were predominantly theoretical to compare the result with a predefined length required for acceptable placement. All methods proposed so far are based on a fundamental assumption: that internal lengths of vessels are proportional to an external body measurement. Many formulas have used birthweight as an easily accessible measurement to estimate internal length. However, this assumption is inherently flawed, especially when looking at weight and length proportions of newborn infants at various gestations. Centile charts plotting birthweight and length seem to be proportional at lower gestations, but become clearly disproportionate closer to term (when the fetus gains majority of its birthweight). Thus, formulas which were prepared using a mixture of preterm and term infants would result in estimated lengths which are likely to be too long for preterm infants and too short for term infants. In addition, infants who are born small or large for a particular gestation would face similar problems as well. As birthweight does not seem to be proportional to length, it is unlikely to be the ideal external measurement for estimating internal length. On the other hand, an external length measurement is more likely to be closely correlated with internal (vessel) length as it does not suffer from the above problems. Dunn's formula used this principle to create the nomogram. However, the external measurement he proposed was not based on fixed anatomical landmarks. This has probably contributed significantly to the malposition of catheter tips in the above studies. From the current literature, it is evident that umbilical catheters, both arterial and venous, are often incorrectly positioned when using existing methods for calculating required length for correct placement. There is a wide range of different success rates with positioning of these umbilical catheters with lower rates for UVC length estimation. The investigators propose that a new formula, based on an external length measurement between two fixed, easily accessible anatomical landmarks, and adjusted for gestation, is urgently needed to improve accuracy of placement of umbilical catheter. The investigators propose a study which uses a novel external length measurement, the sternal notch to umbilicus length, to develop a new formula for the insertion of UVCs and UACs. This measurement is less affected by the normal newborn flexor tone or weight centile size differences at different gestations. Using this measurement, The investigators propose to use regression analysis to develop novel formulae based on its relationship with length of UACs and UVCs respectively.