Editor s key points. M. Dulce*, I. G. Steffen, A. Preuss, D. M. Renz, B. Hamm and T. Elgeti

Transcription

1 British Journal of Anaesthesia 112 (2): (214) Advance Access publication 31 October 213. doi:1.193/bja/aet341 Topographic analysis and evaluation of anatomical landmarks for placement of central venous catheters based on conventional chest X-ray and computed tomography M. Dulce*, I. G. Steffen, A. Preuss, D. M. Renz, B. Hamm and T. Elgeti Department of Radiology, Charité-Universitätsmedizin Berlin, Hindenburgdamm 3, 1223 Berlin, Germany * Corresponding author. miriam.dulce@charite.de Editor s key points Placement of central venous catheters (CVCs) outwith the pericardial sac is important to minimize complications. This retrospective analysis compared computed tomography and chest X-ray for anatomical landmarks in CVC placement. Considerable variation was found in the extrapericardial length and position of the superior vena cava. CXR is not a reliable way to confirm extrapericardial placement of a CVC. Further studies are needed to assess the reliability of the proposed measures for CVC placement. Background. Positioning central venous catheters (CVCs) in the proper part of the superior vena cava (SVC) is difficult. The aim of this exploratory study was to analyse topographic relationships of the extrapericardial SVC using chest X-ray (CXR) and computed tomography (CT). This included an appraisal of rules for optimal CVC tip placement. Methods. We retrospectively evaluated 1 patients with CVCs who underwent bedside CXR and CT on the same day. Distances between the sternoclavicular joint (SCJ), tracheal carina, SVC origin, pericardial reflection, and CVC tip were analysed on CT and, if visible, on CXR. These measurements served to locate the extrapericardial SVC in relation to anatomical landmarks. Different strategies for CVC tip positioning were evaluated. Results. The mean (standard deviation) extrapericardial length of the SVC was 26 (12) mm. The average position of the pericardial reflection was 5 mm below the carina (range, 29 mm below to 25 mm above). In our patient population, the bestresults in terms of tip positions in the extrapericardial SVC would have been achieved by using 85% of the SCJ-to-carina distance (in 86%) or by positioning the CVC tip 9 mm above the carina (in 84% of patients). Conclusions. The extrapericardial part of the SVC varies considerably in length and position, and rules of thumb based on anatomical landmarks should be used cautiously. In our series, using 85% of the SCJ-to-carina distance or placing the CVC tip 9 mm above the carina would have resulted in a high percentage of positions in the extrapericardial SVC. Keywords: central venous catheter; chest; computed tomography; sternoclavicular joint; X-ray Accepted for publication: 1 July 213 Central venous catheters (CVCs) are invaluable in intensive care, and CVC placement is one of the most frequent vascular interventions performed in intensive care units. 1 If the tip of the catheter is inadvertently placed inside the pericardial sac, this may cause life-threatening complications such as pericardial tamponade secondary to vessel wall erosion. 2 4 Pericardial tamponade is a rare complication with a reported incidence of up to 1.4% and mortality rates of 47 1%. 35 To avoid this severe complication, it is recommended to place the CVC tip within the superior vena cava (SVC) above the pericardial reflection the duplication of the pericardium at the upper end of the pericardial sac outside the heart, ideally in the extrapericardial part of the SVC. 6 Although correct catheter placement is commonly checked by portable chest X-ray (CXR), 7 neither the pericardium itself nor the pericardial reflection can be identified by projection radiography. Therefore, various studies have been undertaken to elucidate the relationship between the pericardial reflection and anatomical landmarks detectable with projection radiography. These cadaveric studies may not accurately represent the in vivo situation due to tissue shrinkage. 89 Up to now, the carina and the right tracheobronchial angle have been considered the most suitable landmarks for catheter placement In clinical practice, the carina is commonly used for anatomy-based catheter positioning. 15 In addition, the clavicular notch, as a palpable anatomical landmark of the sternoclavicular joint (SCJ), may also be helpful for CVC placement. In previous studies, it has been used to define the confluence of the internal jugular vein and the subclavian vein. 11 The aim of the present study was to conduct an exploratory analysis of topographic relationships of the extrapericardial SVC, tracheal carina, and SCJ using CXR and computed tomography (CT) (standard of reference). Using these anatomical These authors contributed equally. & The Author [213]. Published by Oxford University Press on behalf of the British Journal of Anaesthesia. All rights reserved. For Permissions, please journals.permissions@oup.com

2 BJA Dulce et al. landmarks, we applied and compared different rules for CVC tip placement on CXR and CT. Methods Patients and sample size Our retrospective study was approved by the responsible ethics committee (application No. EA1/278/11), and a waiver for informed consent was granted. All CT and CXR examinations were performed for clinical indications. Patients over 18 with CVCs in place who had undergone CXR and CT on the same day were identified by a database search. Patients with chest wall deformities were excluded. The study was designed as an exploratory analysis with the extrapericardial length of the SVC as the primary target variable. Based on published data, we assumed a mean [standard deviation (SD)] extrapericardial length of the SVC of 3 (15) mm. 916 A 95% confidence interval (CI) of +3 mm (i.e mm) was aimed at since this length is considered to be accurately measurable on vascular CT and resulted in a sample size of Imaging Bedside chest radiographs were obtained using standard portable equipment (Mobilett XP Eco, Siemens Medical Systems, Forchheim, Germany) and storage phosphor plates (Kodak PQ Elite CR direct view, Carestream Health Inc., Rochester, NY, USA) with the following parameters: tube current, 1 kv; tube film distance, 1.2 m; and exposure time product, mas. All CXRs were performed in the anterior posterior projection with the patient s arms besides the chest. Chest CT was performed after i.v. administration of iodinated contrast medium (Iobitridol 35, Guerbet, Sulzbach, Germany, or Ultravist 37, Bayer, Leverkusen, Germany) with vessel contrast varying according to the clinical questions to be answered. All examinations were performed on routine clinical CT scanners (Definition and Somatom 16 slice, Siemens Medical Systems, Forchheim, Germany). A tube current of 12 kv and rotation time of.5 s was used. Slice thickness varied between.75 and 2 mm, depending on the clinical protocol used. For the CT scans, arms were usually placed above the head, except for trauma scans. Data evaluation and measurements Images were evaluated on a standard clinical picture archiving and communication system workstation (Centricity RIS 14i, General Electric Healthcare, Barrington, IL, USA). The readers of CT and CXR images were blinded to patient name and imaging date. On CXR, the standard software instrument measured the distances of the catheter tip to the carina and to the upper margin of the right SCJ (example shown in Fig. 1). On CT, distances between the tip of the CVC and the pericardium and also the carina and the right SCJ were calculated using the table positions on axial images (example shown in Fig. 1). Statistical analysis Means and SD were used for describing the distribution of the landmark measurements. CIs were calculated using normal approximation, as no relevant deviation from normal distribution was detected when checking Q Q plots and histograms. The agreement of distances between CTand CXR was analysed using the Bland Altman plots and 95% limits of agreement (95% LoA). 18 The association of between-method differences and mean differences in the Bland Altman plots was analysed using Pearson s correlation coefficient (r). Statistical analysis was performed using SPSS for Windows (version 2.) and R (version , R Foundation for Statistical Computing, www. R-project.org). Optimization of tip position using landmarks We used the individual patient measurements to assess three different strategies for CVC placement within the extrapericardial SVC while avoiding intrapericardial placement. Fixed distance to the upper margin of the right SCJ. Fixed distance to the carina in CXR and CT. Percentage of the SCJ-to-carina distance in CXR and CT. Results The 1 patients included in the analysis (64 men, 36 women) had a mean age of 53 yr (range, 18 8 yr). All images were interpretable and none of the examinations had to be excluded due to non-visibility of the tip of the CVC or the anatomical landmarks. Anatomical presentation in CT The target segment for CVC placement the extrapericardial SVC had a mean length of 26 (12) mm and ranged from 1 to 52 mm. In 1% of patients, the length of the extrapericardial SVC was,9 mm and in 3%,2 mm. The mean intrapericardial length was 38 (12) mm, ranging from 18 to 86 mm. On average, the pericardial reflection was 5 (1) mm below the carina. In 31% of all patients of our series, the pericardial reflection was located above the carina with a maximum of 25 mm. In 92% of all patients, the pericardium extended no more than 1 mm above the carina. The mean distance from the SCJ to the pericardial reflection was 78 (18) mm with the smallest distance measuring at 43 mm. The mean distance from the SCJ to the origin of SVC was 53 (14) mm, ranging from 13 to 93 mm. The data are summarized in Table 1. In our series, CVC tips were placed intrapericardially in 48%, in the extrapericardial SVC in 34%, and above the confluence of the SVC in 18% of cases. Agreement of distances between CT and CXR CT and CXR measurements are outlined in Table 2. The Bland Altman plots showing the agreement between distances measured in CT and CXR are presented in Figure 2. The mean difference in the SCJ-to-carina distance between CT and CXR 266

3 Anatomical landmarks for CVC placement based on CXR and CT BJA A SCJ B SCJ confluence SVC tip of CVC carina carina pericardial reflection 1 cm Fig 1 Overview of the anatomical landmarks on CXR (A) and CT (B). On CXR, the SCJ and carina can easily be identified. In this example, one additional CVC has been inserted over the left subclavian vein on the CXR. The tip of the right CVC is therefore taken for measurements. Four different axial CTslices illustrate the anatomical landmarks used on CT for measurements: SCJ (upper margin of the right clavicular notch), confluence of the SVC (cephalad origin of the innominate veins), carina (cartilaginous ridge of the tracheal bifurcation), and pericardial reflection (thin line between SVC and ascending aorta or aortic arch). The extrapericardial SVC measures from the cephalad origin to pericardial reflexion, the intrapericardial part of the SVC from the pericardial reflection to the atriocaval junction (not shown). Table 1 Anatomical landmarks in CT. Summary of distance measurements between anatomical landmarks in CT. CI, confidence interval; SD, standard deviation. Negative values indicate that the pericardium ends above the carina Distance Mean 95% CI for mean SD Range SCJ pericardium Carina pericardium to 29 SCJ origin of SVC SVC length Intrapericardial Extrapericardial Table 2 Comparison of CT and CXR measurements and corresponding 95% CIs. CI, confidence interval; SD, standard deviation. Negative values indicate that the CVC tip is above the carina Distance Method Mean 95% CI for mean SD Range SCJ carina CT CXR Carina CVC CT to 79 CXR to 9 SCJ CVC CT CXR to 171 was,2 mm, whereas the carina-to-cvc tip and SCJ-to-CVC were overestimated on CXR by around 1 mm. CVC tip placement using different landmarks Topographic relationships of different anatomical landmarks are presented as bar plots in Figure 3A. Different rules for CVC tip placement were applied and evaluated using varying distances as illustrated in Figure 3B D. Quantitative data on the resulting CVC positions relative to the SVC (above, extrapericardial, and intrapericardial) are summarized in Table 3. Sternoclavicular joint With a fixed insertion length from the upper margin of the SCJ, the highest percentage of tips correctly positioned in the extraperciardial SVC would be 61% (18% intrapericardial SVC, 21% 267

4 268 A B C Between method differences Between method differences Pearson correlation coeff. =.24 ; P<.1 Pearson correlation coeff. =.17 ; P<.9 Pearson correlation coeff. =.37 ; P< Mean of methods Mean of methods Mean of methods Fig 2 Bland Altman plots demonstrating the agreement of measurements in X-ray (CXR) and CT for the following distances: SCJ to carina (A), carina to CVC tip (B), and SCJ to CVC tip (C). The solid green lines represent the mean difference and the dashed blue lines indicate 95% LoA. Linear regression is displayed by the pink dot-dashed lines. CXR overestimates the carina-to-cvc distance on average by 1 mm (B) and the SCJ-to-CVC distance by 11 mm (C). Between method differences BJA Dulce et al.

5 Anatomical landmarks for CVC placement based on CXR and CT BJA A 15 1 SCJ to SVC-orgin carina-ct carina-cxr extrapericardial SVC intrapericardial SVC Distance Patient number B Percentage of CVC position C D above extra intra 1 above extra intra 1 above extra intra Percentage of CVC position Distance from SCJ Distance from carina in CT SCJ-carina distance (%) in CT Percentage of CVC position Fig3 Topographic relationships of the SCJ, SVC, and carina measured in CT (standardof reference) and CXR. Distancesfrom the SCJ to SVCorigin and intra- and extrapericardial SVC sections are presented as bars (A), whereas the positions of the carina (as a possible landmark for CVC placement) are represented by dots. The determination of the optimal position of the CVC in CT is demonstrated by line plots using three different strategies for orientation: a fixed insertion length from the SCJ (B), a fixed distance from the carina (C), and a percentage of the SCJ-to-carina distance (D). CVC placements above the confluence of the SVC are displayed as orange lines, whereas CVC tips in the extrapericardial (extra) SVC are indicated by green lines and placements in the intrapericardial (intra) SVC by pink lines. The optimal distance to exclude intrapericardial placement is marked by a vertical dashed line, and the highest percentage of extrapericardial positions in the SVC is depicted by a vertical dotted line. above SVC), accomplished with the use of an insertion length of 62 mm. An insertion length of 43 mm below the SCJ would exclude intrapericardial positioning in CXR and CT (extrapericardial SVC: 29%, above SVC: 71%; Fig. 3B). Carina Using the carina as the landmark for positioning the CVC tip, the best position of the CVC tip for maximum extrapericardial placement was 11 mm above the carina in CXR (68%, extrapericardial SVC, 12% intrapericardial SVC, 2% above SVC) and 9 mm above the carina in CT (84% extrapericardial SVC, 8% intrapericardial SVC, 8% above SVC). A 1% extrapericardial position of the CVC tip would be achieved by positioning the tip 51 mm in CXR (2% extrapericardial SVC, 98% above SVC; Fig. 3C) and 25 mm above the carina in CT (32% extrapericardial SVC, 68% above SVC). Percentage of the SCJ-to-carina distance Using a percentage length between the two landmarks, SCJ and tracheal carina, a distance of 85% would result in extrapericardial SVC placement in 68% in CXR (12% intrapericardial SVC, 2% above SVC) and in 86% in CT (6% intrapericardial SVC, 8% above SVC). To accomplish extrapericardial placement in all cases, the SCJ-to-carina distance has to be 55% in CXR (14% extrapericardial SVC, 86% above SVC; Fig. 3D) and 56% in CT (8% extrapericardial SVC, 92% above SVC). 269

6 BJA Dulce et al. Table 3 Landmark methods and distances foroptimized CVC placement. Fixed distance from the SCJ was onlyavailable for CT because the distance to the pericardial reflexion was not measurable in CXR Method Distance CVC position Above SVC (%) Extrapericard. SVC (%) Intrapericard. SVC (%) Optimized for maximal percentage of extrapericardial SVC placement SCJ (CT) 62 mm Carina (CXR) 211 mm Carina (CT) 29 mm SCJ carina (CXR) 85% SCJ carina (CT) 85% Optimized for exclusion of intrapericardial placement SCJ (CT) 43 mm Carina (CXR) 251 mm 98 2 Carina (CT) 225 mm SCJ carina (CXR) 55% SCJ carina (CT) 7% 92 8 Discussion Our in vivo CT measurements show that, with a mean length of 26 (12) mm, the extrapericardial segment of the SVC the target site for proper positioning of the CVC tip is very short. The upper end of this segment is 53 (14) mm (range, mm) below the right SCJ. Its lower end, that is, the pericardial reflection, is 78 (18) mm from the SCJ (range, mm). Note that, in our study, the extrapericardial SVC length was,9 mm in 1% of patients and,2 mm in 3% of patients. Thus, the target segment for placement of the CVC tip is very short in a substantial percentage of patients. One strategy of catheter placement is to insert the CVC to a fixed depth below the SCJ. In the present series, the optimum depth was 62 mm below the SCJ, which would have resulted in correct placement in 61% of cases. However, this approach would still have resulted in the tip of the CVC lying intrapericardially in 18% of patients and above the confluence of the SVC in 21% of patients. Already before the advent of CT, Greenall and colleagues 4 used the clavicles as landmarks on CXR for placement of the CVCtip outside the pericardium and recommended not to place the CVC tip more than 2 mm below the lower surface of the clavicle in order to assure that the CVC ends at or just proximal to the origin of the SVC. As our series confirms, this approach should avoid pericardial tamponade, although the CVC tip might still lie in one of the innominate veins or near their confluence, a position that increases the risk of thrombosis or vessel wall erosion. A higher percentage of extrapericardial placements of the CVC would be achieved by using 85% of the SCJ-to-carina distance for orientation (correct placement in 86% of cases in CT and 68% in CXR) or by using a fixed distance of 9 mm above the carina (correct placement in 84% in CT and 68% in CXR). Using this fixed distance above the carina appears to be the most practicable method for reliable CVC placement in clinical routine. Overall, our results confirm the current opinion that placement of the CVC tip in the extrapericardial segment of the SVC in adults remains difficult Even when a CT scan is available, it is not possible to derive a rule that will allow correct CVC placement in all cases. In the present study, the pericardial reflection was a mean of 5 (1) mm below the carina, a finding that is in agreement with results obtained in cadavers. 89 In these cadaver-based studies, the pericardial reflection always ended below the carina. In contrast, 3% of the patients in our study had a pericardium ending above the carina, with a maximum distance of 25 mm. This finding is in keeping with a study of Caruso and colleagues, 22 who also reported the upper limit of the pericardial reflection to be slightly above the level of the carina in some patients. There are two possible explanations for the different results in the above-quoted cadaver studies. On the one hand, Schuster and colleagues and Albrecht and colleagues might have measured distances more laterally, not taking into account that the pericardium might be higher towards the aortic arch. On the other hand, distances are different in formalin-fixed cadavers due to tissue shrinkage. 89 Our CIs of up to 5 7 mm in CT (Table 1) are comparable with results reported by Albrecht and colleagues 9 and reflect the variability in the lengths of the extra- and intrapericardial segments of the SVC. Very similar means and CI for the SCJ carina distance in CT and CXR indicate a very small parallax effect for these two anatomical landmarks. 8 As expected, the CIs for measurement of the CVC tip position in CXR are considerably higher (up to 13 mm; Table 2), reflecting the difficulties of tip localization in CXR. In one patient, the CVC was flipped back into the internal jugular vein, accounting for the measurement of a CVC tip 44 mm above the SCJ, 69 mm above the carina on CXR, and 15 mm above the carina on CT (datapoint in the lower left corner in Fig. 2B and C). The systematic overestimation of distances measured from CXR compared with CT, revealed by the Bland Altman plots (Fig. 2), can be attributed to the magnification effect of CXR. Although our measurements were performed in vivo, our study has some limitations. 27

7 Anatomical landmarks for CVC placement based on CXR and CT BJA We cannot rule out that some outliers are attributable to intentional (iatrogenic repositioning) or unintentional (slipping) changes in the position of the CVC tip between CXR and CT. Furthermore, individual patient positioning (e.g. arms up or down, bending of the spine) might be different for CT and CXR, and the path of X-rays to the detector plate in CXR might not be as orthogonal as in CT. No subanalysis was performed to adjust for possible confounders such as age, weight, height, or gender. Nevertheless, previous studies have shown that there is no correlation between the length of the SVC or the pericardium and these patient characteristic variables Lastly, our conclusions are drawn from measurements in our collective and caution should be exercised in applying them more widely. Conclusion Owing to the short and variable length and position of the extrapericardial, SVC optimal placement of the CVC tip is difficult. While CXR can rule out complications such as pneumothorax or haematothorax, it allows only a crude estimate of the correct localization of the CVC tip. CXR provides no definite proof that the tip lies exactly in the extrapericardial part of the SVC. Physicians should be aware that the pericardial reflection is located above the carina in one-third of cases. Based on our findings, the highest rate of extrapericardial CVC placements would be achieved by positioning the CVC tip at 85% of the SJC-to-carina distance below the upper margin of the right SCJ or 9 mm above the carina. These rules have to be confirmed in further studies. Authors contributions M.D., I.G.S., and T.E. designed the study, wrote and revised the manuscript. M.D., A.P., I.G.S., and T.E. conducted the study and analysed the data. B.H. and D.M.R. helped to write the revised manuscript. Acknowledgements The authors gratefully acknowledge Bettina Herwig for language editing and Prof. Reinhard Meister for statistical advice. Declaration of interest None declared. References 1 Bhutta ST, Culp WC. Evaluation and management of central venous access complications. Tech Vasc Interv Radiol 211; 14: Collier PE, Blocker SH, Graff DM, Doyle P. Cardiac tamponade from central venous catheters. Am J Surg 1998; 176: Aldridge HE, Jay AW. Central venous catheters and heart perforation. Can Med Assoc J 1986; 135: Greenall MJ, Blewitt RW, McMahon MJ. Cardiac tamponade and central venous catheters. Br Med J 1975; 2: Booth SA, Norton B, Mulvey DA. Central venous catheterization and fatal cardiac tamponade. Br J Anaesth 21; 87: Scott WL. Central venous catheters. An overview of Food and Drug Administration activities. Surg Oncol Clin North Am 1995; 4: Godoy M, Leitman B, Groot P, Vlahos I, Naidich D. Chest radiography in the ICU: Part 2, Evaluation of cardiovascular lines and other devices. Am J Roentgenol 212; 198: Schuster M, Nave H, Piepenbrock S, Pabst R, Panning B. The carina as a landmark in central venous catheter placement. Br J Anaesth 2; 85: Albrecht K, Nave H, Breitmeier D, Panning B, Troger HD. Applied anatomy of the superior vena cava the carina as a landmark to guide centralvenous catheterplacement. Br JAnaesth24; 92: Aslamy Z, Dewald CL, Heffner JE. MRI of central venous anatomy: implications for central venous catheter insertion. Chest 1998; 114: Kim MC, Kim KS, Choi YK, et al. An estimation of right- and left-sided central venous catheter insertion depth using measurement of surface landmarks along the course of central veins. Anesth Analg 211; 112: KremserJ, KleemannF, ReinhartK, SchummerW. Optimizedmethod for correct left-sided central venous catheter placement under electrocardiographic guidance. Br J Anaesth 211; 17: Lee JH, Bahk JH, Ryu HG, Jung CW, Jeon Y. Comparison of the bedside central venous catheter placement techniques: landmark vs electrocardiogram guidance. Br J Anaesth 29; 12: Ryu HG, Bahk JH, Kim JT, Lee JH. Bedside prediction of the central venous catheter insertion depth. Br J Anaesth 27; 98: Stonelake PA, Bodenham AR. The carina as a radiological landmark for central venous catheter tip position. Br J Anaesth 26; 96: Kwon TD, Kim KH, Ryu HG, Jung CW, Goo JM, Bahk JH. Intra- and extra-pericardial lengths of the superior vena cava in vivo: implication for the positioning of central venous catheters. Anaesth Intensive Care 25; 33: Elefteriades JA, Farkas EA. Thoracic aortic aneurysm clinically pertinent controversies and uncertainties. J Am Coll Cardiol 21; 55: Bland JM AD. Statistical methods forassessing agreement between two methods of clinical measurement. Lancet 1986; 327: FletcherSJ, BodenhamAR. Safeplacement of central venous catheters: where should the tip of the catheter lie? Br J Anaesth 2; 85: Gravenstein N, Blackshear RH. In vitro evaluation of relative perforating potential of central venous catheters: comparison of materials, selected models, number of lumens, and angles of incidence to simulated membrane. J Clin Monit 1991; 7: Bayer O, Schummer C, Richter K, Frober R, Schummer W. Implication of the anatomy of the pericardial reflection on positioning of central venous catheters. J Cardiothorac Vasc Anesth 26; 2: Caruso LJ, Gravenstein N, Layon AJ, Peters K, Gabrielli A. A better landmark for positioning a central venous catheter. J Clin Monit Comput 22; 17: Pridie RB, Parnell B. The importance of magnification in left ventriculography. Br J Radiol 198; 53: Ravi B, Rampersaud R. Clinical magnification error in lateral spinal digital radiographs. Spine (Phila Pa 1976) 28; 33: E311 6 Handling editor: L. Colvin 271