Bone Marrow Transplantation, (1999) 24, 793–797  1999 Stockton Press All rights reserved 0268–3369/99 $15.00 http://www.stockton-press.co.uk/bmt Vascular access devices used during harvest of peripheral stem cells: high complication rate in patients with a long-term dialysis central venous

E Johansson1, M Hansson2, A Solle´n Nilsson2 and P Engervall1

1Division of Haematology and Infectious Diseases, Department of Medicine, and 2Immunhemotherapy Unit, Division of Clinical Immunology and Transfusion Medicine, Karolinska Hospital, Stockholm, Sweden

Summary: from the leukaphereses machine to the patient. However, many patients need a double lumen central venous dialysis PBSC harvesting requires good quality venous access. catheter (dCVC) to ensure harvesting. There are two differ- The efficacy and complication rate of the venous access ent categories of dCVC, the stiff dCVC for short-term use2 devices used during stem cell harvest in 101 consecutive and the softer silastic dCVC for long-term use.3 patients were examined. Four different categories of The long-term type of dCVC can also be used for routine venous access were used: (1) long-term dialysis central venous access, starting with delivery of chemotherapy for venous catheter (dCVC), (2) short-term dCVC, (3) per- the mobilisation of stem cells through the period of high- ipheral venous cannulae (PVC), and (4) PVC and con- dose chemotherapy until bone marrow recovery. Most ventional . The number of har- experience of long-term dCVCs has been collected in the vest occasions per patient or harvest days per occasion haemodialysis setting. The majority of reported compli- were similar between the various categories of access. cations leading to removal of the long-term dCVCs are Complications during harvest occurred in 13 out of 48 catheter-related infections (9–28%)4–7 and failure due to (27%) occasions using a long-term dCVC compared to poor flow rates or thromboses (5–30%).4,6–10 The published six out of 97 (6%) in the other three categories pooled experience on use of long-term dCVC for stem cell together (P Ͻ 0.01). Forty-two of the 101 patients harvesting is limited.3,11,12 received a long-term dCVC to facilitate the harvest. The In March 1993 we decided to use long-term dCVC for long-term dCVC was planned to stay in place and also stem cell harvesting and high-dose therapy in patients be used as a conventional i.v. line during the following unable to have peripheral venous cannulae (PVC) due to high-dose treatment. Twenty-one (50%) of the long- poor quality of the peripheral , with the intention of term dCVCs were removed due to complication. Thir- reducing the number of CVCs inserted in these patients. teen (31%) of the long-term dCVCs were usable The primary aim of this retrospective study was to esti- throughout the entire treatment period. In conclusion, mate the function and complication rate of long-term we recommend that PBSC harvesting is performed dCVCs used during harvesting and as a conventional intra- through peripheral venous when practically venous line. Secondly, we evaluated efficacy and compli- possible, otherwise via short-term dCVC. cation rate during harvesting in all patients during the study Keywords: PBSC collection; harvest; central venous period, comparing four different categories of venous access; central venous catheter; complication; thrombosis accesses. In summary, half of the long-term dCVCs were removed due to complications and only one-third of the long-term dCVCs were usable throughout the entire treatment period. An increasing number of patients with various malignant disorders are undergoing high-dose therapy followed by autologous stem cells rescue as a part of their treatment.1 Patients and methods PBSC collected by leukapheresis has become the main cell source in recent years. The procedure requires a high blood flow through intravenous (i.v.) lines between patient and Leukaphereses performed between 1993 and 1996 apheresis machine. In some patients it is possible to use Between 1993 and 1996 101 patients were harvested on intravascular devices inserted in the peripheral arm veins, 145 occasions (each occasion consists of a varying number or to insert a femoral dialysis catheter. In other patients of harvests over a number of consecutive days (range 1–5 with a functional central venous catheter (CVC) already in days)), over 363 leukapheresis days. Patient characteristics place, this catheter can be used as the return line for blood are summarised in Table 1. In eight patients with CML, harvests were performed at the time of diagnosis and leu- kapheresed cells were saved as a back up to be used later Correspondence: E Johansson, Division of Haematology and Infectious diseases, Department of Medicine, Karolinska Hospital, S-171 76 Stock- during blast crisis. The remaining 93 patients were planned holm, Sweden for PBSC rescue at the time of harvest. Received 23 December 1998; accepted 13 May 1999 Venous access routes used during harvest were divided Venous catheters for harvest of PBSC E Johansson et al 794 Table 1 Characteristics of 101 patients who underwent leukaphereses Leukapheresis procedure and of 42 patients in whom a long-term dCVC was inserted Collection of peripheral stem cells was performed by apher- 101 patients 42 patients with esis nurses, using a COBE Spectra blood cell separator who underwent a long-term (Cobe, Lakewood, CO, USA) on standard settings of the leukapheresis dCVC automated program (presently software version 5.1) which has been upgraded regularly since the beginning of the Female/Male (n) 52/49 24/18 study. Unless otherwise stated, at least 10 l of blood were Age, median, range (years) 50 (16–66) 48 (17–61) Diagnosis (n) processed during each procedure, aiming at a steady flow Acute leukaemia 26 10 rate of 50–60 ml/min and a collection rate of р1.5 ml/min. Chronic myelocytic leukaemia 19 6 All complications requiring special action during the apher- Multiple myeloma 21 8 eses were documented by the apheresis nurses. Lymphoma 22 12 Solid tumour 11 6 Amyloidosis 2 – Use and care of long-term dCVCs

dCVC = dialysis central venous catheter. dCVC dressings were changed twice a week in accordance with the findings of a previous study.13 The exit site was aseptically cleaned with chlorhexidine (5%) and a trans- parent dressing was used. All parenteral therapy including into four categories: (1) long-term dCVC, (2) short-term medications, fluid therapy, blood products, peripheral stem dCVC, (3) PVC and (4) PVC and conventional CVC. cells and total nutrient admixtures were given via the long- In 66 of the 145 (46%) harvest occasions, patients had term dCVC. Daily blood samples were also taken. Two poor quality peripheral veins and dCVCs were inserted to three-way stopcocks with a 10-cm extension tube were con- facilitate the harvest procedure. In 18 of these 66 harvest nected to the long-term dCVC. The inner stopcock was occasions (16 patients), patients received a short-term changed at the time of the dressing change and the outside dCVC (Mahurkar; Quinton Instruments). In the remaining stopcock was changed once a day, most often directly after 48 of these 66 harvest occasions (37 patients) PBSC were blood sampling. New caps were put on after every event. collected through long-term, tunneled, dual-lumen dCVC The insertion site was inspected daily through the trans- (PermCath; Quinton Instruments). Venous access could, in parent dressing to identify early signs of infection. To the same patient, differ from one harvest occasion to maintain function between use, each lumen was primed another, depending on the quality of the peripheral veins with 1.3 ml or 1.4 ml of heparin (5000 IU/ml). Before each at the time of harvest. use the indwelling heparin was aspirated.

Impaired blood flow/occlusion Patients with a long-term dCVC In long-term dCVCs with signs of occlusion (ie total failure During the study period 42 patients received a long-term of infusing or when it was possible to infuse but not to dCVC (Table 1). Thirty-seven of these 42 patients mobil- aspirate) flushes were administered. Fibrinolytic ther- ised stem cells and were subsequently harvested. The long- apy (t-PA, Actilys; Boehringer Ingelheim) was used if cath- term dCVCs were surgically installed under aseptic con- eter dysfunction persisted after mechanical manipulation. ditions in an operating theatre. Patients with a platelet count Suspicion of catheter-related deep venous thrombosis was Ͻ × 9 50 10 /l received prophylactic platelet transfusions confirmed either by X-ray or Doppler examination. Verifi- before catheter insertion. Catheter placements are shown in cation of a venous dCVC-related thrombosis resulted in Table 2. Thirty-one of the 42 (74%) patients had a history immediate catheter removal. None of the patients received of having had one to four previous CVCs. prophylactic oral anticoagulants.

Infection Table 2 Localisation of 42 long-term dialysis central venous catheters The decision to remove a long-term dCVC due to infection, was made by the physician in charge. During neutropenia No. all patients received prophylactic antifungal and antiviral therapy with fluconazole 50–200 mg once daily and acyclo- Internal jugular vir 200 mg five times daily, respectively. Prophylactic anti- right 7 left 9 biotic therapy was not given. External jugular right 12 left 10 Statistics Subclavia right 3 Catheter survival analysis was calculated using the Kaplan– Saphena magna Meier life table method for estimating survival. Catheter right 1 survival times were recorded in completed days. Nominal data were compared using the Fisher’s exact test. Differ- Venous catheters for harvest of PBSC E Johansson et al 795 ences in group proportions were assessed by the Mann– geons. There were no major operative complications. Minor Whitney U test. All statistical calculations were made with bleeding from the insertion site was observed in 10 patients. StatView 4.51 software. There was a trend towards less complications leading to removal of catheters inserted in the external jugular vein (3/12, 30%) compared to the remainder (18/30, 60%; NS). Results A history of earlier lines did not increase the risk of a cur- rent catheter complication. Antibiotic prophylactic before Efficacy and complication rate in all venous access insertion, blood values at the time of insertion, insertion devices used during leukaphereses performed 1993–1996 time or bleeding after insertion, did not differ between patients in whom the long-term dCVC was later removed One hundred and forty-five harvest occasions were due to complications and those where it was not (data not recorded in 101 patients. The number of harvest occasions shown). Pain, in the catheter site, requiring analgesics dur- per patient or harvest days per occasion were similar ing the first 24 h after insertion was recorded in 30% of between the various types of access (Table 3). Compli- patients. cations during harvest occurred on 13 out of 48 (27%) Twenty-one of 42 (50%) long-term dCVCs were occasions when using a long-term dCVC compared to six removed electively after a median of 119 days (range 27– out of 97 (6%) using other venous access devices pooled Ͻ 308). The remaining 21 catheters were removed after 47 together (P 0.01; Table 2). days (median, range 21–176) due to complications; infec- When all harvest occasions were compared, without tion (n = 16), infection and thrombosis (n = 2), thrombosis regard to the catheter category used, there was a statistically (n = 2) and occlusion (n = 1; Figure 1, P Ͻ 0.01). significant (P Ͻ 0.0001) difference in number of days per In 11 out of the 18 long-term dCVCs removed due to harvest occasion with more days during the first 2 years of clinically suspected infection, positive bacterial culture this study compared to the last 2 years (data not shown). findings were obtained at the time of removal. In five There were more complications during harvest occasions patients the same microorganism was isolated from both in patients with lymphomas and solid tumours compared to blood and the long-term dCVC exit site (Coagulase-nega- the remaining patients (13 out of 43 (37%) vs six out of tive staphylococcus (CNS, n = 3), Staphylococcus aureus 102 (6%); P Ͻ 0.001).

Long-term dCVCs 1 Twenty-five of 42 patients completed the planned high- dose treatment with autologous PBSC rescue. In 17 of these 0.8 25 patients the long-term dCVC was still in place at the time of PBSC rescue. Thirteen of the 17 long-term dCVCs 0.6 were used without complications throughout the treatment period. 0.4 Three of 42 patients underwent allogeneic stem cell 0.2 transplantation; in two of these three patients the initial long-term dCVC were removed electively and in one due 0 to complications. Fourteen of 42 patients did not complete the planned high-dose treatment and PBSC rescue due to Proportion long-term dCVC survival 0 50 100 150 200 250 300 350 poor mobilisation, poor growth of CD34-positive cells Time (days) (n = 6) or progressive disease or relapse (n = 8). In six of Figure 1 Cumulative proportion of remaining long-term dialysis central these 14 patients the long-term dCVC was still functioning venous catheters (dCVC). Kaplan–Meier life table showing catheters well with no signs of infection or occlusion at removal. removed electively (n = 21, ———) and catheters removed due to compli- The long-term dCVCs were inserted by 13 different sur- cation (n = 21, ———).

Table 3 Function and complications during leukaphereses performed 1993–1996

Long-term dCVC Short-term dCVC PVC PVC and CVC Total

Patients (n) 37163717101a Harvest occasions (n) 48185425145 Harvest occasions per patient (mean ± s.d.) 1.30 ± 0.52 1.13 ± 0.34 1.40 ± 0.64 1.41 ± 0.62 1.32 ± 0.56 Harvest days per occasion (mean ± s.d.) 2.65 ± 1.07 2.16 ± 1.06 2.49 ± 1.11 2.37 ± 1.29 2.48 ± 1.12 Complications during harvest Non-seriousb (n) 1110315 Seriousc (n) 2200 4 aIn six patients, two different types of access were used. bHarvest was completed. cHarvest was interrupted before 10 l of processed blood. dCVC = dialysis central venous catheter; PVC = peripheral venous cannulae; CVC = central venous catheter. Venous catheters for harvest of PBSC E Johansson et al 796 (n = 1) and Citrobacter freundii (n = 1)). Three patients had mobilization with G-CSF was not used as a standard and positive cultures in their blood (CNS (n = 1), S. aureus (3) the target level of collected CD34+ cells was higher (n = 1) and Escherichia coli (n = 1)) and three patients had during the early period. There were clearly more compli- positive exit site cultures (CNS (n = 1), S. aureus (n = 1) cations during harvesting in patients with long-term dCVCs and Corynebacter sp. (n = 1)). The remaining seven long- compared to the other categories. A high complication rate term dCVCs removed due to presumed infection had nega- has also recently been reported by Goldberg et al14 where tive cultures. 50% of their patients with a semipermanent dCVC experi- Patients with malignant lymphoma showed a trend enced at least one episode of catheter occlusion during har- towards a higher complication rate compared to the other vesting which required thrombolytic therapy or cancellation patients (9/12 (75%) vs 12/30 (40%); NS). There was no of the procedure. However, in our study only four out of difference between patients with or without long-term 66 (6%) harvest occasions had to be interrupted due to dCVC complications with regard to the proportion of fever occlusion. A completely different approach is to use a tech- days (Ͼ38.0°C) during neutropenia (Ͻ0.5 × 109/l; nique for translumbar catheterisation of the inferior vena mean ± s.d., 54% ± 32 vs 38% ± 32 respectively; NS). cava.11,12 Using that technique, thrombosis-related access However, patients with long-term dCVC complications failure occurred in 24%, but 64% of the dCVCs were still had, when not neutropenic, a mean of 11 ± 18 febrile days in place during PBSC rescue.11 per 100 catheter days. This can be compared to patients In our study half of the long-term dCVCs were removed without complications, when not neutropenic, in whom a due to complications. Most of the long-term dCVCs were mean of 1% ± 2 febrile days (P Ͻ 0.001) was seen. removed before the start of high-dose treatment. This also Earlier venous thrombosis was present in three of four occurred in patients with catheter complications who patients with catheter-related thromboses compared to one experienced an increased number of febrile days while not of the remainder (3/4 (75%) vs 1/38 (3%); P Ͻ 0.001). Dur- neutropenic. We have no clear explanation for the fact that ing conventional use 20 of the 42 long-term dCVCs were most of the catheters were infected when patients were not occluded on 33 occasions, in one or both long-term dCVC immunosuppressed. Several factors are known to influence lumen. Fibrinolytic therapy (Actilyse) was successful on 24 risk of infection associated with the use of ordinary CVCs; of 33 occasions. Seven occlusions were resolved by mech- number of lumens, insertion site, repeated catheterization, anic manipulation and one long-term dCVC required duration of catheterization, type of dressing, experience of removal due to the occlusion. personnel inserting the device and that of special intra- venous therapy personnel.13,15–17 However, the compli- cation rate, in our study, of the remaining long-term dCVCs Discussion during the high-dose treatment was comparable with that of other reports concerning conventional CVCs.18,19 Central venous access is mandatory in the care of patients To our knowledge, there is only one previous study receiving intensive chemotherapy and harvest of peripheral describing the use of long-term dCVCs in adults during stem cells is not possible without adequate access. There harvesting and subsequent high-dose treatment with PBSC are several issues to consider when choosing a venous rescue.3 In that study, Meisenberg et al3 showed that 16% access device for patients planned for high-dose therapy of long-term dCVCs used in this setting were removed with stem cell rescue. prematurely due to complications and 7% of the long-term When the peripheral route of venous access is deemed dCVCs were infected late. The difference in infection rate inadequate, one may choose a short-term dCVC, a short- compared to that seen in our study may in part be explained term femoral dialysis catheter or a long-term dCVC. The by the use of prophylactic low-dose warfarin. An associ- femoral catheter has to be removed immediately after the ation between infection and thrombosis has been described harvest and the risk of local bleeding must be considered. in patients with cancer and CVC.20 There were also more The short-term dCVC may allow repeated harvests during frequent complications in long-term dCVCs inserted by less a consecutive number of days, but should not be in place skilled surgeons in the Meisenberg study. Thirteen surgeons for more than 10–14 days, thus necessitating at least one inserted the long-term dCVCs in this study and results may further catheter for the rest of the treatment period. To have been better if fewer surgeons had been involved. avoid unnecessary repeated trauma, a long-term dCVC can Another difference between our study and the study by be used for the whole period. Meisenberg et al was the management of exit site dressing We found no differences in harvest yield, expressed as changes. In the Meisenberg study, patients were instructed days per harvest occasion, between the different access on how to change their long-term dCVC exit site dressing, devices used in this study. Our presumption was that the but whether or not this had an impact on the risk for infec- more steady and often higher blood flow obtained by using tion is difficult to evaluate. a dCVC, compared to a PVC, would result in fewer harvest Our results also demonstrate that occurrence of long- days per occasion. We are not aware of any trial comparing term dCVC-related thrombosis was related to a history of efficacy between different vascular access devices used earlier deep venous thrombosis. Therefore, it seems reason- for harvest. able that patients with a history of venous thrombosis In the first part of this study (1993–1994) each patient should avoid long-term dCVCs. If this group of patients was harvested more often, regardless of the type of i.v. still have a catheter, a very low dose of warfarin device. There could be several explanations for this: (1) (1 mg/24 h) may protect against thromboses without selection of more heavily pretreated patients, (2) PBSC inducing a haemorrhagic state.3,21,22 Furthermore, if cath- Venous catheters for harvest of PBSC E Johansson et al 797 eter-related thromboses do occur, the catheter may success- 8 Grote J, Lufft V, Nikutta P et al. Transesophageal echocardio- fully be left in place during full-dose anticoagulation graphic assessment of thrombosis in therapy.3 patients with long-term central venous catheters. As a consequence of our high complication rate with Clin Nephrol 1994; 42: 183–188. long-term dCVCs, both during harvest and later, we have 9 Agraharkar M, Isaacson S, Mendelssohn D et al. Percu- taneously inserted silastic jugular hemodialysis catheters sel- changed our policy in patients unable to have a PVC. These dom cause jugular vein thrombosis. ASAIO J 1995; 41: patients now have a short-term dCVC, which is removed 169–172. after harvesting is complete. This approach is also 10 Shaffer D. Lessons from vascular access procedures for hemo- supported by other reports.2,23 dialysis. Surg Oncol Clin N Am 1995; 4: 537–548. In conclusion, on the basis of our experience we rec- 11 Haire WD, Lieberman RP, Lund GB et al. Translumbar ommend that harvesting is performed via a PVC in patients inferior vena cava catheters: experience with 58 catheters in with veins of good quality. If there is a need for a dCVC, peripheral stem cell collection and transplantation. Transfus a short-term catheter used only during the collection of Sci 1990; 11: 195–200. PBSC is preferable. The long-term dCVCs cannot be rec- 12 Haire WD, Lieberman RP, Lund GB, Kessinger A. Translum- ommended due to the high number of complications. bar inferior vena cava catheters. Bone Marrow Transplant 1991; 7: 389–392. Further studies are needed to optimize different 13 Engervall P, Ringertz S, Hagman E et al. Change of central approaches towards venous access, starting from mobilis- venous catheter dressings twice a week is superior to once a ation before stem cell harvest until recovery after high-dose week in patients with haematological malignancies. J Hosp chemotherapy with stem cell rescue. Infect 1995; 29: 275–286. 14 Goldberg SL, Mangan TR, Klumpp TR et al. Complications of peripheral blood stem cell harvesting: review of 554 PBSC Acknowledgements leukaphereses. J Hematother 1995; 4: 85–90. 15 Keung Y-K, Watkins K, Chen S-C et al. Increased incidence This study was supported by grants from the Swedish Cancer of central venous catheter-related infections in bone marrow Society and Vingmed Svenska AB. transplant patients. Am J Clin Oncol 1995; 18: 469–474. 16 Richard-Smith A, Buh S. Reducing central line catheter infec- tions in bone marrow transplant patients. Nurs Clin North Am References 1995; 30: 45–51. 17 Reed CR, Sessler CN, Glauser FL, Phelan BA. Central venous 1 Gratwohl A, Hermans J, Baldomero H. Blood and marrow catheter infections: concepts and controversies. Intensive Care transplantation activity in Europe 1995. Bone Marrow Trans- Med 1995; 21: 177–183. plant 1997; 19: 407–419. 18 Brandt B, DePalma J, Irwin M et al. Comparison of central 2 Hahn U, Goldschmidt H, Salwender H et al. Large-bore cen- venous catheter dressings in bone marrow transplant recipi- tral venous catheters for the collection of peripheral blood ents. Oncol Nurs Forum 1996; 23: 829–836. stem cells. J Clin Apheresis 1995; 10: 12–16. 19 Pearson ML. Hospital Infection Control Practices Advisory 3 Meisenberg BR, Callaghan M, Sloan C et al. Complications Committee. Guideline for prevention of intravascular device associated with central venous catheters used for the collection related infections. Infect Control Hosp Epidemiol 1996; 17: of peripheral stem cells to support high-dose chemotherapy 438–473. and autologous stem-cell rescue. Support Care Cancer 1997; 20 Raad II, Luna M, Sayed-Ahmedh H et al. The relationship 5: 223–227. between the thrombotic and infectious complications of cen- 4 Shusterman NH, Kloss K, Mullen JL. Successful use of tral venous catheters. JAMA 1994; 271: 1014–1016. double-lumen silicone rubber catheters for permanent hemo- 21 Bern M, Lokich J, Wallach S et al. Very low dose of warfarin dialysis access. Kidney Int 1989; 35: 887–890. can prevent thrombosis in central venous catheters. Am Coll 5 Moss AH, Vasilakis C, Holley JL et al. Use of a silicone dual- Physicians 1990; 112: 423–428. lumen catheter with a dacron cuff as a long-term access for 22 Boraks P, Seale J, Price J et al. Prevention of central venous hemodialysis patients. Am J Kidney Dis 1990; 16: 211–215. catheter associated thrombosis using minidose warfarin in 6 Gibson SP, Mosquera D. Five years experience with the Quin- patients with haematological malignancies. Br J Haematol ton PermCath for vascular access. Nephrol Dial Transplant 1998; 101: 483–486. 1991; 6: 269–274. 23 Alegre A, Requena MJ, Ferna´ndez-Villalta MJ et al. Quinton- 7 Shaffer D, Madras PN, Williams ME et al. Use of dacron Mahurkar catheter as short-term central venous access for cuffed silicone catheters as long-term hemodialysis access. PBSC collection: single-center experience of 370 aphereses in ASAIO J 1992; 38: 55–58. 110 patients. Bone Marrow Transplant 1996; 18: 865–869.