The safety and efficacy of bedside removal of tunneled hemodialysis catheters by nephrology trainees

Abstract

Background: Some nephrologists remove tunneled hemodialysis catheters (TDC) at the bedside, but this practice has never been formally studied. Our hypothesis was that bedside removal of TDC is a safe and effective procedure affording prompt removal, including in cases of suspected infection. Methods: We reviewed our consecutive 3-year experience (2007–2009) with bedside TDC removal at the University of Mississippi Renal Fellowship Program. Data were collected on multiple patients and procedure-related variables, success and complication rates. Association between clinical characteristics and biomarkers of inflammation and myocardial damage was examined using correlation coefficients. Results: Of 55 inpatient TDC removals (90.9% from internal jugular location), 50 (90.9%) were completed without hands-on assistance from faculty. Indications at the time of removal included bacteremia, fever or clinical sepsis with hemodynamic instability or respiratory failure. All procedures were successful, with no cuff retention noted; one patient experienced prolonged bleeding which was controlled with local pressure. Peak C-reactive protein (available in 63.6% of cohort) was 12.9 ± 8.4 mg/dL (reference range: <0.49) and median troponin-I (34% available) was 0.534 ng/mL (IQR 0.03–0.9) (reference range: <0.034) and they did not correlate with each other. Abnormal troponin-I was associated with proven bacteremia (p < 0.05) but not with systolic and diastolic BP or clinical sepsis. Conclusion: Our results suggest that bedside removal of TDC remains a safe and effective procedure regardless of site or indications. Accordingly, TDC removal should be an integral part of competent Nephrology training.

• Bacteremia
• catheter-related bacteremia
• C-reactive protein
• hemodialysis
• tunneled dialysis catheter
• troponin-I

Introduction

Tunneled dialysis catheters (TDC) are commonly utilized in the care of end-stage renal disease (ESRD) patients receiving hemodialysis. These “semi-permanent” catheters have the advantage of relative ease of placement with immediate usability, providing a bridge while a more permanent access (arterio-venous fistula, peritoneal dialysis catheter) is placed and fully matured. Furthermore, these catheters can be utilized as a final (destination) dialysis access for those patients who have exhausted all other options of permanent dialysis access.

The use of indwelling vascular access for dialysis is not, however, free of complications. Approximately, two-thirds of them would eventually need removal due to malfunction or infectious complications.1,2 Removal of TDCs could take place in a variety of settings, including in a dedicated procedure suite by surgeons or interventional radiologists, or by a nephrologist at the bedside. Any potential delay resulting from a prolonged wait for a dedicated procedure suite may, however, harm these patients. Bedside removal of TDC may be an ideal approach to address these concerns in suitable patients, especially when infection is a concern and urgent removal of the hardware is in focus. Although considered common practice, there is a paucity of published literature on TDC removal by nephrologists. Moreover, there is a lack of data on the safety and complication rates for this procedure was performed by relatively inexperienced trainees in supervised clinical training settings. Many of these patients with clinical symptoms (e.g., chest pains, shortness of breath or fever) will also receive, as part of their usual care, measurements of commonly utilized biomarkers looking for myocardial damage and inflammation (troponin-I and C-reactive protein, respectively). Our primary theory was that bedside removal of TDC is a safe and effective procedure, affording prompt removal of infected hardware. Our secondary objective was to examine the association between these biomarkers and clinical indications for TDC removals in our cohort.

Material and methods

Study population

We performed a retrospective cohort review of our consecutive 3-year bedside TDC removal experience (1 January 2007 to 31 December 2009) at the University of Mississippi Nephrology Fellowship Program. Patients had been referred to the procedure team by nephrology consulting teams, whenever TDC removal became medically necessary in the team’s clinical judgment. The decision to remove the TDC was made solely by the Nephrology consult team and may have been the result of a variety of indications, including proven bacteremia, fever or clinical septic state, catheter malfunction or recovery of renal function. The basis of data recovery was the procedure teaching log of the first author, which included all inpatient TDC removals by the Nephrology Division during the index period.

Measurements

We reviewed and collected data on multiple patient-related variables: age, race, gender and highest blood urea nitrogen, creatinine and blood coagulation tests within 24 hours of the procedure. Age was the number of full calendar years completed since birth. Gender was self-reported as either male or female. Ethnicity was either Caucasian or African American. Additionally, we collected data on certain other peri-procedure parameters, up to three days before and after the procedure: peak and nadir white blood cell count (WBC); nadir hemoglobin, nadir platelet count and vital signs (temperature, heart rate, blood pressure). Two additional biochemical parameters associated with inflammation and myocardial stress, C-reactive protein (CRP) and troponin-I were searched for and recovered from medical records, if available within 48 hours of TDC removal. Procedure-related variables, which included indication for the procedure, the site and location of removal and any complication or difficulties during the procedure were recorded from the teaching log. Contraindications for the procedure included abnormal coagulation results, including prothrombin time INR >1.5 and markedly decreased platelet count (<60,000/mm³), when otherwise correctable. No further selection criteria were applied beyond the above-mentioned exclusion criteria. All patients provided written consent before the procedure.

Procedure description

Once the decision was made to remove the TDC, the procedure took place within 1–6 hours and always on the same calendar day. Preparation for the procedure included obtaining informed consent from the patient or relative, and gathering the limited supplies needed: suture removal kit, syringes, needles, 1% lidocaine, face mask, sterile gown, sterile gloves, sterile 4 × 4 dressing, Hemostat, chlorhexidine swab and silk tape. Most of the TDCs were removed by nephrology fellows with no attending involvement beyond supervision of the procedure. Depending on the operator’s experience, one or two fellows participated in the procedure with the senior fellow assisting the junior level trainee. At first, sutures, if present were removed and the catheter exit area, the surrounding skin, the catheter proximal to the hub (including the buried, but mobile portion distal to the Dacron cuff) were carefully cleaned with either iodine or chlorhexidine-based cleaning solution. Afterwards, the cleaned area was draped to create a sterile field for dissection. Local anesthesia was achieved with injection of 1% lidocaine without epinephrine solution, 10–15 mL total. Subsequently, utilizing a sterile hemostat clamp, the cuff was bluntly dissected through the exit site from the adjacent soft tissue in such a manner that the dissection usually proceeded up to 1–2 cm proximal to the Dacron cuff. Once the cuff was loosened, the TDC could be pulled smoothly with a controlled amount of vigorous force. Hemostasis was obtained by applying direct pressure on the tract and the exit site with sterile gauze for ≥5 minutes until no bleeding was detectable. Finally, 4 × 4 dressings were secured with silk tape. The total procedure time was usually between 15 and 20 minutes. All procedures were performed under the supervision of a single procedure-oriented attending physician (TF).

Statistical methods

This study was reviewed and approved by the University of Mississippi Human Research Office (Protocol 2010/289). Upon review of both electronic and paper-based medical records, predefined information as approved by the Human Research Office was collected in Microsoft Excel data sheet. Data were analyzed with SPSS Statistics 19 (IBM Corporation, Armonk, NY) and reported with means ± SD or medians 25–75% IQR for descriptive data; Pearson’s correlation and chi-square as well as independent-samples t-test were utilized for statistical comparisons.

Results

Our study population consisted of 55 hospitalized patients. All TDCs were removed at the bedside with 50 cases (90.9%) completed by nephrology fellows under the attending physician’s supervision. The rest of the catheters were removed by the attending physician with medical resident(s) observing the procedure. General cohort characteristics are shown in Table 1. The majority of TDC removals took place in general wards (63.6%), with the rest of the removals done either in emergency department (12.7%) or intensive care unit (23.6%) with a median time of 3 days (IQR 1–13) elapsing from admission with TDC in place or TDC placement. Of these, 36 (65.5%) TDCs were removed from the right internal jugular (IJ) position, 14 (25.5%) from the left IJ position and 5 (9.1%) from femoral veins. Most cases had urgent indication for TDC removal with potential for harm with delays. These included proven (culture-positive) bacteremia in 36.4% of the cases, fever in 41.8% of the cases or clinical signs of sepsis with hemodynamic instability or respiratory failure in 20% of the cases. Only three TDCs were removed in patients recovering renal function, on the basis of “no longer needed”. At the time of TDC removal, 4 (7.2%) patients were hypothermic, 33 (60%) were febrile or subfebrile (T > 37 °C); 7 (12.7%) were on vasoactive pressors (norepinephrine or high dose dopamine). All removals were technically successful without any retention of Dacron cuffs or catheter portions. One patient had prolonged local bleeding, that was controlled with local pressure. No case required Interventional Radiology or General Surgery consultation or assistance. Peak C-reactive protein (available in 63.6% of the cohort) was 12.9 ± 8.4 mg/dL (reference range: <0.49), median troponin-I (34% available) was 0.127 ng/mL [IQR 0.03–0.9] (reference range: <0.034) and they did not correlate with each other (p = 0.848). The associations of CRP and troponin-I with clinical indications for TDC removal and selected clinical parameters are shown in Table 2. Interestingly, we could not find any association between CRP and clinical indications for TDC removal. Additionally, clinical sepsis (as indication for TDC removal) correlated with systolic BP nadir (p < 0.0001), temperature (p = 0.002) and lowest platelet count (p = 0.016). Troponin-I had no association with systolic and diastolic BP or clinical sepsis (as indication for TDC removal). However, troponin-I, as a continuous variable showed a trend with confirmed bacteremia (p = 0.075); furthermore, the association of troponin-I as a bivariate variable (abnormal/normal) with bacteremia was statistically significant (Pearson’s chi-square = 0.456, p = 0.049) (Table 2; 3rd column).

Table 1. Baseline cohort characteristics and indications for tunneled dialysis catheter removal.

Age (years)	53.9 (15.4)
Gender (%, female)	54.5
Race (%, African American)	90.9
Heart rate (bpm/min)^a	95.6 (20.6)
Systolic BP nadir (mmHg)^a	116.5 (20.7)
Diastolic BP nadir (mmHg)^a	64.9 (12.3)
Temperature (°C)^b	37.6 (1.8)
Blood urea nitrogen (mg/dL)^c	46.8 (23.6)
Creatinine (mg/dL)^c	6.96 (3.37)
Platelet count, nadir (×10^3/mm^3)^a	189.6 (114.8)
WBC count, peak (×10^3/mm^3)^a	13.6 (8.7)
WBC count, nadir (×10^3/mm^3)^a	7.9 (4.6)
Hemoglobin, nadir (gm/dL^a	9.4 (1.6)
Biochemical markers
CRP (mg/dL)^b	12.9 (8.4)
Troponin-I (ng/mL)^b	0.534 (0.708)
Indications for TDC removal
Bacteremia	36.4%
Fever (temperature >38 °C)	41.8%
Clinical sepsis	20%
Recovered renal function	5.5%

Abbreviations: BP, blood pressure; CRP, C-reactive protein; TDC, tunneled dialysis catheter; WBC, white blood cell.

Obtained within 72 hours^a, 48 hours^b or 24 hours^c of the procedure.

CRP values available for 63.6% of cohort; troponin-I for 34%.

Table 2. Associations of C-reactive protein and troponin-I with clinical indications for TDC removal and selected clinical parameters.

	CRP (mg/dL)^# p-value	Troponin-I (ng/mL)^# p-value	Troponin-I (Bivariate: abnormal/normal) p-value
Indications for TDC Removal
Bacteremia	0.806	0.075	0.049
Fever (temperature >38 °C)	0.005	0.235	0.018
Clinical sepsis	0.977	0.446	0.659
Clinical parameters
Systolic BP nadir (mmHg)^a	0.690	0.561	0.254
Diastolic BP nadir (mmHg)^a	0.526	0.288	0.945
Temperature (°C)^b	0.031	0.941	0.256
Platelet count, nadir  (×10^3/mm^3)^a	0.292	0.745	0.912
WBC count, peak  (×10^3/mm^3)^a	0.225	0.383	0.594

Abbreviations: BP, blood pressure; CRP, C-reactive protein, TDC, tunneled dialysis catheter; WBC, white blood cell.

Obtained within 48 hours^b or 72 hours^a of the procedure; ^#considered as a continuous variable.

CRP values available for 63.6% of cohort; troponin-I for 34%.

Discussion

The use of TDC for hemodialysis has been very common since its introduction. These catheters, originally developed as a short-term bridge to permanent vascular access, constitute an increasing percentage of vascular accesses in newly declared ESRD patients. According to the United States Renal Data System (USDS) 2012 report, in 2010 ∼80% of new dialysis patients utilized catheters at initiation of renal replacement therapy and ∼50% still use them three months later.3 These sobering numbers have deteriorated since the mid-nineties, at which point less than 20% of new hemodialysis patients utilized TDCs at 60 days after initiation of renal dialysis.4 Although not an access of primary choice for hemodialysis, a sizeable minority of patients is also chronically dialyzing with tunneled dialysis catheters. Nonetheless, the use of TDCs to deliver hemodialysis treatments has been associated with major problems. Catheter malfunction, secondary to thrombosis or malposition, and catheter-related infections are common occurrences5,6 and place a burden on health care providers too.2 Published experience suggests that about one-third of TDCs are removed either due to infections, poor function or maturation of an alternate site of dialysis access.1,2 Pisoni et al., using Dialysis Outcomes and Practice Patterns Study (DOPPS) II data, showed that every 20% increase in facility catheter use was associated with a 16% higher mortality risk.7 A later study confirmed that less catheter and graft use is associated with improved patient survival.8 Additionally, the focus on “Fistula First” initiative may have had the unintended consequence, paradoxically, of contributing to a persistent high utilization of TDCs.9

While much attention in the published literature has been focused on the circumstances of catheter placement10,11 and strategies related to maintaining patency and infection-free state of the catheters,12–16 so far little has been published on TDC removals. The procedure can be accomplished either at the bedside or in a vascular suite, and it can be performed by nephrologists, interventional radiologists or surgeons. Although a study comparing the outcomes of hemodialysis catheters placed by interventional radiologists with those placed by surgeons has been done,17 no study has been conducted comparing the safety of TDC removal at the bedside to those performed in a vascular suite. Waiting to remove TDC in a more controlled environment, for example, a vascular suite may potentially delay the procedure and be associated with increased morbidity or mortality. In our series, there was a likely selection bias against referrals solely for TDC malfunction (i.e., malfunctioning catheters were likely referred to catheter exchange in the outpatient setting). Accordingly, this referral pattern of sick inpatients may have contributed to the lack of complications with catheter tears and cuff separations. Furthermore, it may well be the case that the clinical infection in many of our patients contributed to tissue breakdown around the catheter and the relative ease of pulling them. While we have not encountered separation of Dacron cuffs in our series, nothing in the published literature suggests that in the absence of obvious local infection these retained cuffs could not be left in place—a scenario analogous to clotted arterio-venous grafts. Retained catheters tethered to the environment beyond the Dacron cuff, however, present a distinct challenge and require a second surgical incision, endoluminal dilatation18 or the use of a transcatheter extractor device19 to remove them. In the case of right atrial adhesion and retention, laser sheath liberation is an another emerging technique.20

It should be specifically emphasized that our renal fellowship does not offer full-scale formal interventional nephrology training; neither have we had access to a dedicated procedure room. Nevertheless, we have been able to offer training on this one unique procedural aspect of nephrology and with little investment of time renal trainees have gained an additional skill which was viewed uniformly as a positive “add-on value” to them. While a small number of academic programs offer full-scale, formal interventional nephrology training, this relatively easy-to-perform procedure could represent both a valuable addition to a generic Nephrology Fellowship training program and a positive contribution to the timely care of patients. In our experience, it took approximately 5–8 supervised procedures for our first-year renal trainees to master the “learning curve” for the procedure, at which point they were able to assist other trainees. Low albumin, anemia or elevated CRP are known risk factors associated with adverse outcomes during TDC exchange,21,22 which is an alternative to TDC removal in select patients. Our cohort was relatively ill with generally high CRP values. Herewith, the measurement of CRP was unlikely to contribute to clinical decision-making. Of note, CRP values in “healthy” dialysis patients from a similar cohort23 as well as those measured in an elderly cohort of community-dwelling hypertensive patients from a similar environment24 were markedly lower than in our series.

Finally, subtle troponin-I elevation has been associated with worse cardiovascular (CV) outcomes in ESRD and occasionally has been ascribed to the ESRD status alone. Nonetheless, in usual clinical practice, troponin-I levels are obtained typically for reasons such as CV symptoms. While the concept of “normal” reference range for troponin-I can be debated in ESRD patients, our results seem to strongly suggest that some of the mild elevation of troponin-I observed in hospitalized ESRD patients could potentially be attributed to TDC-induced bacteremia. With regard to cardiac enzyme “leaks”, elevated cardiac troponins were noted to be associated with bacteremia in other studies.25 To state it differently, these results are suggesting that minor troponin-I elevation in the proper context (ill dialysis patient with indwelling TDC) may be a potential clue toward the presence of catheter-related bacteremia. Additional, larger studies are needed to confirm this hypothesis.

Limitations

Limitations of our study include the retrospective, single-center design with a small number of participants and the unknown length of TDC placement prior to hospital presentation. The small sample size makes it difficult to discern weaker associations and precludes the application of multivariable analyses toward assessing the role of confounders. We acknowledge that the referral pattern and indications for the procedure had an impact on some of the patient and population characteristics. Our data collection did not contain any outpatient TDC removal procedures, including those performed by interventional radiology, vascular surgery or nephrology (outpatient) advance care nurse practitioners; therefore, our results and conclusions may be applicable to inpatient settings only.

Conclusion

In our series, bedside removals of TDCs were highly successful and no major complications were encountered. Major indications for the procedure in our study were either proven or suspected infections and we observed a significant association between catheter-related bacteremia and elevated troponin levels. Our results argue that TDC removal should be an integral part of general Nephrology training.

Declaration of interest

The authors have no potential conflict of interest to disclose. The authors alone are responsible for the content and writing of the paper.

Acknowledgements

Drs. Naseem A. Qureshi and Vikram R. Beemidi are former Nephrology Fellows of the University of Mississippi Medical Center, Years 2010–2012. Dr. Beemidi currently affiliated with Northwest Mississippi Regional Medical Center in Clarksdale, MS and Dr. Qureshi currently in private practice in Jackson, MS. Dr Kamel A. Gharaibeh currently a third-year Internal Medicine Resident at the University of Mississippi Medical Center. Dr. Mihaly Tapolyai is currently employed at WJB Dorn VA Medical Center, University of South Carolina, Columbia, SC.

We sincerely appreciate the dedicated clinical care of our Nephrology trainees form the years past, making this paper possible: Frederick Lee, Brandon Bean, Son G. Lam, Minesh B. Pathak, Adrian Cosmin, Mohit Ahuja, Gurvinder S. Suri, David E. Pruett, R Sellors Meador, Michael Shoemaker-Moyle, Justin H. Bain, and Derrick Tesseneer. We appreciated the help of Ms. Kathryn L. Roberson, R.N. to review the manuscript for English and grammar. We are also thankful for the statistical advice from Zsolt Lengvarszky, M.S., Ph.D. from Louisiana State University Shreveport, Shreveport, LA.

Preliminary data from this study has been presented in an abstract format at ASN Kidney Week 2011, Philadelphia, PA (PO1960); J Am Soc Nephrol. 22, 2011: 568 A and at the SSCI Meeting, New Orleans, LA (P253); J Investigative Med. 2013 (Feb); 61(2): 444.