Central venous catheter tip misplacement: A multicentre cohort study of 8556 thoracocervical central venous catheterisations

There is a paucity of data on the incidence of central venous catheter tip misplacements after the implementation of ultrasound guidance during insertion. The aims of the present study were to determine the incidence of tip misplacements and to identify independent variables associated with tip misplacement.


Editorial Comment
With routine central venous catheter placement, determination of the final location of the catheter tip is usually performed after the catheter placement procedure has been completed.In this analysis, follow-up for central venous catheter tip position showed that suboptimal catheter tip position still occurred in a small minority of cases in the cohort that was assessed, and was associated with several factors.

| INTRODUCTION
Central venous catheters (CVCs) are essential in modern healthcare.
3][4][5][6][7] The definition of tip misplacement varies in different studies.The correct tip position is somewhere within the lower superior vena cava (SVC), at the cavo-atrial-junction (CAJ) or in the cranial part of the right atrium (Figure 1). 1,2,8,9ere are no prospective multicentre studies on the incidence of catheter tip misplacements after the implementation of ultrasound guidance for central venous access.1][12] There is also a paucity of large observational studies investigating associations between independent variables and risk of CVC tip misplacements.It has been suggested that the risk of tip misplacement increases with left-sided insertions in the external and internal jugular veins and that the subclavian vein is associated with more tip misplacements compared to the internal jugular vein. 5,6,12,13

| Aim
The aims of this prospective multicentre cohort study were to determine the incidence of CVC tip misplacements and to identify patient-, operator-and catheter-related variables associated with tip misplacement.

| METHODS
This exploratory study is a sub-study of the CVC-MECH trial 14 which was approved by the Swedish Ethical Review Authority (Dnr 2018/295).As the study did not include any interventions the Authority waived the requirement for informed consent.The study was registered at ClinicalTrials.gov(NCT03782324) and the manuscript was written according to the STROBE guidelines for observational studies (Supplementary Appendix, p. 14). 15l CVCs inserted in the jugular and subclavian veins from 2 March hospitals: insertion in the subclavian vein, expected long-term use (≥7 days), anticipated administration of chemotherapy, hemodynamic measurements, dialysis, and symptoms of mechanical complications.

| Data collection
Previously unpublished data from the CVC-MECH database were used.All data were prospectively collected as described in detail in the Methods section of the main study. 14The follow-up time was 24 h after the insertion.In reasonable time after documentation of the CVC insertion, a careful review of the patients' medical records including all relevant CVC insertion information and radiologic evaluations of chest radiographies, was performed by dedicated collaborators (researchers, research nurses or medical secretaries), thereby enabling operators to correct inadequate or missing information.In cases where the CVC tip position was not mentioned in the radiologic evaluation, one or more of the authors (EÄ, GB, MA or TK) reviewed the chest radiography.

| Outcomes
The outcomes were the incidences of 'any tip misplacement', 'minor tip misplacement' and 'major tip misplacement'.9][20] Figure 1 shows the classification of different tip misplacements.Tip positions within the superior vena cava, at the cavo-atrial junction or within the cranial part of the right atrium were considered correct. 1,2,9For 10-cm-long CVCs, a position in the brachiocephalic vein was also considered correct as this was the desired position.Further details regarding the classification and grading of different tip placements are shown in Table 1.Radiographic landmarks used to define tip positions are described in detail in Table 2.The cranial limitation of the superior vena cava was defined as 3 cm above the carina, 14 whereas the caudal limitation of the superior vena cava was defined as the cavo-atrial junction (4 cm below the carina). 12Tip misplacements classified as 'deep in the right atrium' (too close to the tricuspid valve) were defined as >7 cm below the carina.Tip misplacements in the azygos vein were not included as lateral view chest radiographies were uncommon in the database.

| Independent variables
The following independent variables with possible associations with the outcomes were predefined based on previous studies: patient age, gender, height and body mass index, vascular insertion site, operator experience, insertion technique (ultrasound guidance before insertion T A B L E 1 Classification and grading of tip positioning.

| Statistics
The sample size was calculated in the main study. 14The present sub-study included all available jugular and subclavian CVC insertions with post-procedural chest radiographies in the CVC-MECH database.
A statistical analysis plan was defined before assessing the data (Supplementary Appendix, pp.2-6).The statistical analyses were performed using SPSS, version 29 (IBM Corporation, Armonk, NY, USA).
Multivariable logistic regression analyses were performed to determine the association between the independent variables and the outcomes.Separate multivariable logistic regression analyses were performed for minor tip misplacements and major tip misplacements.
6][27] Hence, BMI was divided into three groups (BMI > 20 kg/m 2 , 20-30 kg/m 2 , >30 kg/m 2 ).BMI < 20 kg/ m 2 and BMI > 30 kg/m 2 were compared to BMI 20-30 kg/m 2 .Leftsided insertions were compared with right-sided insertions and the subclavian and external jugular vein insertion sites were compared to the standard insertion site at the studied departments, that is, the internal jugular vein. 2,28The following variables were dichotomised: catheter length (≤15 vs. >15 cm), operator experience level (<100 vs. ≥100 insertions in the chosen vein prior to study start) and insertion technique (US guidance in real-time vs. no US guidance in real-time).
The number of events per outcome determined how many of the independent variables that were included in each multivariable logistic regression model.An independent variable was included if there were at least eight outcome events for that variable.With fewer events, variables were excluded starting with the least important one (as determined by univariable logistic regression analyses, possible correlation, clinical experience, and results from prior studies).
A bivariate correlation analysis was performed to rule out any collinearity among the independent variables.One of the variables was excluded if the correlation coefficient was >0.8 or <À0.8.

Goodness of fit for the multivariable logistic regression models
was tested with Hosmer-Lemeshow test and a p-value >.05 represented an acceptable goodness of fit.
Normally distributed continuous variables are presented as the mean and standard deviation and others are presented as median and interquartile range.Categorical and binary variables are presented using counts and percentages.The incidences of tip misplacements are presented using counts and percentages, with a 95% confidence interval.The results from the multivariable logistic regression analyses are presented as odds ratios (OR) with 95% confidence intervals.A pvalue ≤.05 was considered statistically significant.

| RESULTS
A flow chart of the selection of CVC-insertions is presented in   10 This high number may be explained by a wide definition of tip misplacement ('anywhere else than in the superior vena cava').
In our study, CVC insertion in the subclavian or external jugular vein was associated with minor tip misplacement.Only the subclavian vein was associated with major tip misplacement.The incidence of any tip misplacement for the subclavian vein was 7.6 (6.2-9.1)%, of which 4.3 (3.3-5.5)% were major tip misplacements.These results are in line with the results from several other studies, showing a higher risk of CVC misplacement when the subclavian vein is used compared to the internal jugular vein. 5,6,12,13 a prospective, observational study from our group we demonstrated that misplaced guidewires during right-sided subclavian catheterisation can be corrected during the insertion procedure in most cases by using a right supraclavicular fossa ultrasound-guided technique. 29Although this technique has not yet been adopted by most colleagues within the region where the current study was performed, it is reasonable to assume that it has contributed to a lower incidence of tip misplacements after CVC insertions in the right subclavian vein.
Previous studies have shown that the left subclavian vein is associated with less tip misplacements compared to the right subclavian vein. 5,12,30,31However, our findings could not confirm this (Supplementary Appendix, p. 13).One possible explanation may be that an unknown number of operators used the right supraclavicular fossa ultrasound view for correction of the guidewire position during catheterisation procedure, 29 which would even out potential differences between the sides.
The lowest incidence of tip misplacements in the current study was seen when the right internal jugular vein was catheterised.This reinforces the right internal jugular vein as the insertion site of choice for short-term CVCs.However, for longer anticipated need of CVCs, the subclavian vein probably allows for better patient comfort and fewer CVC-related infections and thrombosis. 4,24,32,33veral studies, both landmark and real-time ultrasoundguided insertions, have reported a higher number of cannulation attempts to be an important risk factor for mechanical complications in general. 5,11,14,25,34,35In agreement with that, an increasing number of skin punctures was associated with major tip misplacement in the present study.One explanation for this association may be that insertion problems due to patient factors such thrombosis or anatomical variations may lead to more puncture attempts and more tip misplacements.
Real-time ultrasound guidance was not associated with a lower risk for tip misplacements in the present study.7][38] However, if ultrasound is not used to confirm the guidewire tip position prior to inserting the CVC, the use of real-time ultrasound guidance does not seem to decrease the risk of tip misplacement. 39male patient gender was associated with major tip misplacement.This is a new finding that, to the best of our knowledge, has not been described before and needs to be confirmed and further investigated before any conclusions can be drawn.Although the bivariate correlation analysis did not show a significant correlation between patient gender and height, it cannot be ruled out that these two variables co-variate to some degree, as both low height (with Youden index 168.5 cm) and female patient gender were associated with major tip misplacement.Interestingly, female patient gender was not associated with minor tip misplacement.
Outcomes were divided into minor and major tip misplacements based on the potential consequences of tip misplacement.This classification can be useful when evaluating the need to correct tip misplacements.If the CVC is to be used for chemotherapy, it is mandatory that the tip position is optimal (i.e., neither minor nor major tip misplacement should be present).However, for parenteral nutrition the demands on the tip position are less precise (i.e., a minor tip misplacement does not need correction before use).In line with this reasoning, Ablordeppey and colleagues found that the majority of the mispositioned CVCs in their study remained in clinical use without adjustment with similar complication rates as the ones adjusted. 10her studies have demonstrated that short-term CVC tips misplaced in the right atrium or brachiocephalic vein may not need to be repositioned prior to use. 12,40

| Limitations
This study has several limitations.First, it was an observational study and the variables found to be associated with catheter tip misplacement need to be further investigated in future studies.Given the  observational character of the study, any causality cannot be proven.Second, the chest radiographies used to define the tip position were only single plane, thus it was not possible to determine misplacements in the azygos vein.Third, ECG-guidance for CVC-tip positioning was not used.Fourth, we have no data whether CVC insertions were performed under emergency conditions.Thus, the incidence of nightly CVC insertions was used as a surrogate marker.Fifth, although this is the largest study on the subject this far, the power of the multivariable logistic regression analyses was limited so the results from those should be interpreted with some caution.Sixth, although the proportion of missing values was low, it cannot completely be ruled out that missing values affected the results.

| Conclusion
In this large prospective multicentre cohort study, performed in the ultrasound-guided era, we demonstrated the incidence of tip misplacements to be 3.

2019 to 31
December 2020 in patients ≥16 years with a postprocedural bedside anterior-posterior chest radiography were included in the study.With a few exceptions, all the insertions were performed by anaesthetists.Four hospitals in the Scania Region in southern Sweden participated in the study: one university hospital with approximately 1300 beds and three county hospitals with approximately 200-300 beds each.All four hospitals used the same electronic health record (EHR) system (Melior™, Cerner Corporation, North Kansas City, Missouri, USA), in which each CVC insertion was recorded in an insertion form.The indications for performing a postprocedural chest radiography were the same for all participating F I G U R E 1 AP chest radiography illustrating different tip placements.Green stars: correct tip position.Orange stars: minor tip misplacements.Red stars: major tip misplacements.IVC, inferior vena cava; LAXV, left axillary vein; LBCV, left brachiocephalic vein; LIJV, left internal jugular vein; LSCV, left subclavian vein; RA, right atrium; RAXV, right axillary vein; RBCV, right brachiocephalic vein; RIJV, right internal jugular vein; RSCV, right subclavian vein; RV, right ventricle; SVC, superior vena cava.

2
Radiographic landmarks used to define different tip positions.the lateral border of costa I to the lateral border of the clavicular-sternal junction Right brachiocephalic vein From the right clavicular-sternal junction to the right first costo-sternal junction (app.2.5 cm) Left brachiocephalic vein From the left clavicular-sternal junction to the right first costo-sternal junction (app.6 cm) Internal jugular vein Cranial to the superior margin of the clavicular-sternal junction 21 CVC from the left side with the tip indenting the wall of the superior vena cava CVC from the left side with the tip indenting the lateral wall of superior vena cava.The CVC is not aligned with the superior vena cava.Cranial superior vena cava ≤3 cm above carina Persistent left-sided superior vena cava Suspected position in persistent left-sided superior vena cava Cavo-atrial junction 4 cm below carina 9Deep in right atrium >7 cm below carinaRight ventricleThe tip reaches the midline at >7 cm below carinaInferior vena cava Tip below the diaphragm only, real-time in-plane ultrasound guidance, real-time out-of-plane ultrasound guidance, anatomic landmark method, change over guidewire), catheter length and bore size and number of skin punctures.4,5,[22][23][24][25]

Figure 2 .
Figure 2. A total of 14,071 CVC insertions were assessed for eligibility in the study.Eight thousand five hundred and fifty-six insertions remained after the exclusion criteria had been applied.Seven insertions resulting in failed catheterisations were recognised in separate journal entries in the EHR and thus added afterwards.Patient baseline data and CVC insertion characteristics are presented in Table3.Of the 5587 patients included, 38% were

a
Compared to male patient gender.b Youden index for height = 168.5 cm.c Compared to BMI 20-30 kg/m 2 .d Youden index for high BMI = 31.5.e Compared to right-sided insertion.f Compared to internal jugular vein.g Compared to CVC bore size <9 Fr. h Compared to operator experience (>100 insertions).i Compared to CVC length >15 cm.j Compared to no ultrasound guidance in real-time.
7 (95% CI: 3.3-4.1)%.Right internal jugular vein catheterisation had the lowest incidence of both minor and major tip misplacement.Causal mechanisms will require further evaluation.AUTHOR CONTRIBUTIONS Study conception: EÄ, GB, MA, and TK.Literature search: EÄ, MA, and TK.Protocol and study design: EÄ, GB, MA, OB, and TK.Collection/ monitoring of data: MA, and TK.Preparation of databases: MA, EÄ, and TK.Verification of databases: MA, and TK.Design of statistical analyses: MA, and TK.Statistical analyses: EÄ.Verification of statistical analyses: TK.Writing of draft article: EÄ.

Table 3
. Of the 5587 patients included, 38% were female.The mean (SD) BMI was 27 (6.0)kg/m 2 .Of the 8556 All primary outcomes measures are presented per vascular insertion site in Table 4. Any tip misplacement occurred in 316 (3.7 (3.3-4.1)%) of the CVC insertions.Of these, 179 (2.1 (1.8-2.4)%) were classified as major tip misplacements.Catheterisation of the right internal jugular vein resulted in 2.2 (1.9-2.6)%tip misplacements compared to 7.8 (6.3-9.6)% after catheterisation of the right subclavian vein.The most common tip misplacement was deep in the right atrium, which occurred in 80 (0.9 (0.7-1.2)%) of the insertions, followed by ipsilateral brachiocephalic vein placement in 65 (0.8 (0.6-1.0)%) of the insertions.No catheter tips were misplaced into the contralateral axillary vein.The incidences of both minor and major mechanical complications are presented in

Table 5 .
Patient baseline data and CVC characteristics.Tip misplacements per vascular insertion site.In another retrospective study by Ablordeppey et al., tip misplacements occurred in 13.4% of 2025 CVC insertions.
12e results from the multivariable logistic regression analyses for minor tip misplacements and major tip misplacements are reported in Table6.No variables were excluded due to collinearity, and the goodness of fit test was valid with p > .05forbothregressionmodels.Only the variables shown to have a significant associ-(p.13)alongwiththecorrespondingunivariateanalyses (pp.11-  12).There were no associations between independent variables and any tip misplacement in the multivariable logistic regression analysis.Deviations from the statistical analysis plan are presented in the Supplementary Appendix, p. 7.T A B L E 3 T A B L E 4The 2.1 (1.8-2.4)%incidence of major tip misplacements is lower than in a previous prospective, single-centre study on a similar cohort by Pikwer et al.where tip misplacements (defined as extrathoracic or ventricular tip positions) occurred in 3.3% of the 1619 CVC insertions included.12 T A B L E 5 Mechanical complications.