Association between thrombosis and bloodstream infection in neonates with peripherally inserted catheters

Thrombosis Research (2008) 122, 782–785

www.elsevier.com/locate/thromres

BRIEF COMMUNICATION

Association between thrombosis and bloodstream infection in neonates with peripherally inserted catheters Courtney D. Thornburg a,b,⁎, P. Brian Smith a,b , Mary Laura Smithwick a , C. Michael Cotten a , Daniel K. Benjamin Jr. a,b a b

Department of Pediatrics, Duke University, Durham, NC, United States Duke Clinical Research Institute, Durham, NC, United States

Received 16 July 2007; received in revised form 1 October 2007; accepted 3 October 2007 Available online 13 November 2007

KEYWORDS Thrombosis; Catheter; Infection; Infant

Abstract Introduction: Peripherally inserted catheters are essential for infants in the neonatal intensive care nursery for administration of medications, parenteral nutrition and blood transfusions. We hypothesized that there is an association between catheter associated thrombosis and catheter associated blood stream infection. The primary objective of this study was to determine the association between catheter associated blood stream infection (CABSI) and catheter-related thrombosis in the Neonatal Intensive Care Unit. Materials and methods: We performed a cohort study of the 1540 peripherally inserted catheters placed in 882 infants from August 2002 until November 2005 in the Duke University Medical Center Neonatal Intensive Care Unit. Results: 212 peripherally inserted catheters were removed because of thrombosis. We identified 142 cases of CABSI. There was a positive association between thrombosis and infection in infants who did not have lines removed for infection, p b 0.05. Conclusions: Further study is warranted to determine the pathophysiology between the association between thrombosis and infection and to determine if interventions may decrease the risk of these potentially life-threatening complications. © 2007 Elsevier Ltd. All rights reserved.

Abbreviations: PICs, Peripherally inserted catheters; NICU, neonatal intensive care unit; TPN, total parenteral nutrition; VTE, Venous thromboembolism; CVCs, central venous catheters; CABSI, catheter associated thrombosis and catheter associated blood stream infection; DUMC, Duke University Medical Center. ⁎ Corresponding author. PO Box 2916, Duke University Medical Center, Durham, NC 27710, United States. Tel.: +1 919 684 3401; fax: +1 919 681 7950. E-mail address: [email protected] (C.D. Thornburg). 0049-3848/$ - see front matter © 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.thromres.2007.10.001

Association between thrombosis and bloodstream infection in neonates with inserted catheters Peripherally inserted catheters (PICs) have been utilized in the neonatal intensive care unit (NICU) since the 1980′s. They are essential for critically ill premature and full-term infants, and they are becoming part of the admission process for extremely low birth weight infants (infants b 1000 grams birth weight). They provide a means to provide longterm access for antibiotics, blood products, intravenous fluids, and total parenteral nutrition (TPN). They also decrease the number of painful procedures to which infants are exposed [1]. However, PICs are associated with a number of significant complications including mechanical complications, infection, and thrombosis [2]. Venous thromboembolism (VTE) is a common complication of central venous catheters (CVCs) including PICs and catheter associated thrombosis is the most common type of VTE in children [3]. Neonates are at particularly high risk for catheter thrombosis due to the large bore of the catheter relative to the size of the blood vessel which leads to obstruction to blood flow and damage to the vein, physiologic differences in the neonatal hemostatic system and acquired prothrombotic risk factors such as dehydration and polycythemia. Neonates are also at risk for catheter associated blood stream infection (CABSI). Prior studies have suggested that lower birthweight, TPN administration, colonization with Staphylococcus epidermidis and duration of catheter increase the risk of infection [4]. There is some evidence that there is an association between catheter associated thrombosis and CABSI. This is likely a bidirectional relationship. The presence of catheter thrombus serves as a nidus of infection. For example, the plastic CVC surface promotes platelet activation and thrombus formation, and the thrombus and its ligands promote adherence of local bacteria [5]. Alternatively, an infection and the associated inflammation can trigger thrombus formation. We sought to better define this relationship in infants as both infection and thrombosis can be lifethreatening in this age group. If a relationship is determined, screening and prevention strategies can be implemented to reduce these catheter Table 1

Baseline characteristics of the study population Mean

Gestational age at birth (weeks) Gestational age at PIC insertion (weeks) Gestational age at PIC removal (weeks) Birth weight (grams) Weight at PIC Insertion (grams)

Range

31.1 ± 5 32.8 ± 5.9

23–42 24–66

34.4 ± 5.9

24.7–66.4

1749 ± 1033 1623 ± 1014

484–6160 469–5815

Table 2

783

Reasons for PIC removal

Reason for removal

N

%

No longer needed Line malfunction Thrombosis Infection Line broken or in wrong position Death Replaced Other Transferred with line

549 260 212 61 58 45 7 18 218

38.5 18.2 14.9 4.3 4 3.2 0.5 1.3 15.3

complications. This study was undertaken to determine the association between CABSI and catheter associated thrombosis and to describe the epidemiology of PICs in the NICU at Duke University Medical Center (DUMC). Materials and methods We performed a retrospective cohort study of the 1540 PICs placed in 882 infants from August 2002 until November 2005 in the DUMC NICU. A CABSI was defined as a positive blood culture which was documented N24 h after PIC placement or within 72 h of catheter removal. Positive cultures were ignored if the same organism was cultured in the week prior to catheter insertion. Evidence of thrombus included cord, phlebitis, extremity edema, extremity perfusion, and inability to draw or flush the catheter as previously described [6]. A relevant thrombus was defined as a thrombus leading to PIC removal; this was recorded in a PIC log by nurses trained in the care of neonatal PICS under the supervision of a neonatologist. Additional data including reason for PIC removal were collected from an electronic medical record and PIC log in electronic and written formats. The DUMC Institutional Review Board approved this study. All PICs were placed by using sterile technique by a member of the neonatal PIC team under sterile conditions using povidone–iodine, gowns, caps, masks, and sterile gloves. All PICS were 24 or 28 gauge single lumen catheters. Twenty-eight gauge PICs were never heplocked. Infants received continuous infusions of intravenous fluids with 0.5–1 units of heparin /mL or TPN with 0.5–1 units of heparin/mL at a minimum rate of 0.5– 1 cm3/h. Twenty-four gauge PICs that were heplocked were flushed every 4 h with IVF containing 1 unit of heparin/mL. Data were analyzed using STATA 9.0 (College Station, TX). Descriptive statistics were used to describe baseline characteristics. The association between thrombosis and CABSI was determined by logistic regression adjusting for PIC position and gestational age at insertion. If removal of the PIC prevented thrombus formation then the association between CABSI and thrombosis may be obscured. Therefore, the data were analyzed to compare rates of infection in infants whose PICs were removed for thrombosis versus infants whose PICs were removed for a reason other than presumed or documented infection. Rates of CABSI and thrombosis were calculated per 1000 PIC days.

Results We analyzed 1540 PIC insertions in 882 infants. Baseline characteristics are shown in Table 1. The majority of PICs were placed in the arm (N = 572), foot (N = 488) and hand (N = 326). Other locations included leg, scalp, axilla, neck and groin veins.

784

Figure 1 Occurrence of infections and thromboses over PIC life. The figure indicates the number (N) of infections and thromboses categorized by age of PIC (weeks). The overall rate of infection was 8.9/1000 PIC days. The rate of infection for PICs with and without thrombosis was 9.2/1000 PIC days and 8.5/1000 PIC days respectively.

The most common reasons for catheter insertion included general intravenous access, nutrition, and antibiotics. Reasons for PIC removal were available for 1428 PICs (Table 2). Two hundred eighteen PICs were still in place at the time infants were transferred to another institution. Five hundred forty nine PICs were removed electively because they were no longer needed. The remaining PICs were removed due to malfunction (N = 260), evidence of thrombosis (N = 212), and documented or presumed infection (N = 61). Other reasons for line removal included death, problem with line placement or structure, and line replacement. The rate of thrombosis was 13.2/1000 PIC days. There were 142 documented infections and the overall rate of infection was 8.9/ 1000 PIC days. The overall occurrence of both thrombosis and infection decreased over time (Fig. 1). Twenty-eight percent (40/ 142) of lines with CABSI were removed for infection, and the remainder was successfully treated with antibiotic therapy through the affected line. Of the PICs that were removed for thrombosis, 10.8% (23/212) had current or prior CABSI. Twenty-three of the 97 PICs with CABSI were removed for thrombosis compared to 189 of the 1270 PICs removed for a reason other than thrombus or infection (p b 0.05). The mean PIC duration was 12 ± 12 days (median 8 days) for all PICs, 22 ± 20 (median 18 days) for PICs affected by CABSI and thrombosis, and 11± 12 (median 6 days) for PICS affected by thrombosis only. The overall time to thrombosis was 12 ± 13 days (median 7 days). The mean time from CABSI to thrombosis was 16.5 ± 17.2 (median 14 days).

Discussion In this retrospective study we found a thrombosis rate of 13.2/1000 catheter days, and evidence that there is a positive association between catheter thrombosis and infection. Of interest, the occurrence of both infection and thrombosis decreased over time. The parallel decrease also supports the hypothesis of an association between CABSI and thrombosis. The risk of thrombosis was not associated with catheter site. We would expect a higher rate of both infection and thrombosis in younger infants. However, there was still a significant positive association between thrombosis and infection when adjusting for gestational age at insertion. Of interest, there appear to be two types of

C.D. Thornburg et al. catheter thromboses. One set of thromboses occurred early without evidence of CABSI and may be related to mechanical or other rheologic factors. The other set occurred late in the presence of CABSI and may be related to inflammation induced by the infection. These findings are consistent with the multi-factorial nature of thrombosis. Our study is limited by the retrospective nature and somewhat subjective definition of thrombosis. Neither Doppler ultrasonography nor venography is routinely performed in the NICU for infants with PICs unless there is persistent limb edema or discoloration after PIC removal. We could not absolutely distinguish between intraluminal PIC thrombosis and occlusion associated with a fibrin sheath. However, we hypothesize that both would increase the risk of CABSI, and therefore, the data still suggest that there is an association. Association between TPN use and thrombosis was not assessed since the majority of infants were on TPN. Data was not available on the number of insertion attempts. Strengths of the study include the unique population of neonates, the size of the cohort and complete record keeping of PICs over the time period. The link between CVC infection and CVC thrombosis is supported by the findings of several studies. Van Rooden et al. investigated the relationship between infectious complications of CVCs and catheter associated thrombosis [7]. The authors found that the relative risk of thrombosis was 17.6 in patients with infection compared to those without infection. They also found that the risk of thrombosis was associated with the severity of infection with 57.1% of patients with a systemic CABSI having thrombosis compared to 27.3% with a local infection. Raad et al. compared thrombus and infection in catheterized veins with that in the contralateral, uncatheterized vein in an autopsy study of 72 cancer patients [8]. Thrombosis, inflammation and ulceration were significantly more common in catheterized veins, 49% vs. 9%, p b 0.001. In addition, CVC infection, confirmed by catheter tip culture, occurred exclusively in catheterized veins, 23% vs. 0%, respectively, p b 0.01. Bacteria were adherent to fibrin filaments on all 10 catheters by electron microscopic scanning, suggesting that local bacteria adhere to the catheter and promote CVC infection. Also, in a prospective study of 208 CVC placements in an ICU stetting, Timsit et al. found that CVC related sepsis was highly correlated with the presence of CVC thrombus by Doppler ultrasonography, OR 2.97, p = 0.02 [9]. Additional studies have shown that prevention of thrombus formation and enhancement of clot resolution reduce the risk of CVC infection [10–14]. In conclusion, this large retrospective study of PICs in the NICU reinforces that infection and

Association between thrombosis and bloodstream infection in neonates with inserted catheters thrombosis are important complications and are associated. Our work complements prior research in this field. Further studies of the pathophysiology of thrombosis and infection and of interventions to decrease these complications are warranted in our unique population.

[7]

[8]

Acknowledgement The authors wish to thank the nurses in the DUMC NICU who provide care for the infants and to Katie Cheng and Rose Guo who participated in data entry as students at the North Carolina School of Math and Science.

[9]

[10]

References [1] Janes M, Kalyn A, Pinelli J, Paes B. A randomized trial comparing peripherally inserted central venous catheters and peripheral intravenous catheters in infants with very low birth weight. J Pediatr Surg 2000;35(7):1040–4. [2] Pettit J. Assessment of infants with peripherally inserted central catheters: Part 2. Detecting less frequently occurring complications. Adv Neonatal Care 2003;3(1):14–26. [3] Schmidt B, Andrew M. Neonatal thrombosis: report of a prospective Canadian and international registry. Pediatrics 1995;96(5 Pt 1):939–43. [4] Camara D. Minimizing risks associated with peripherally inserted central catheters in the NICU. MCN, Am J Matern Child Nursing 2001;26(1):17–21 quiz 22. [5] Stillman RM, Soliman F, Garcia L, Sawyer PN. Etiology of catheter-associated sepsis. Correlation with thrombogenicity. Arch Surg 1977;112(12):1497–9. [6] Pettit J. Assessment of infants with peripherally inserted central catheters: Part 1. Detecting the most frequently

[11]

[12]

[13]

[14]

785

occurring complications. Adv Neonatal Care 2002;2(6): 304–15. van Rooden CJ, Schippers EF, Barge RM, Rosendaal FR, Guiot HF, van der Meer FJ, et al. Infectious complications of central venous catheters increase the risk of catheterrelated thrombosis in hematology patients: a prospective study. J Clin Oncol 2005;23(12):2655–60. Raad II, Luna M, Khalil SA, Costerton JW, Lam C, Bodey GP. The relationship between the thrombotic and infectious complications of central venous catheters. Jama 1994;271(13): 1014–6. Timsit JF, Farkas JC, Boyer JM, Martin JB, Misset B, Renaud B, et al. Central vein catheter-related thrombosis in intensive care patients: incidence, risks factors, and relationship with catheter-related sepsis. Chest 1998;114(1):207–13. Shah PS, Kalyn A, Satodia P, Dunn MS, Parvez B, Daneman A, et al. A randomized, controlled trial of heparin versus placebo infusion to prolong the usability of peripherally placed percutaneous central venous catheters (PCVCs) in neonates: the HIP (Heparin Infusion for PCVC) Study 10.1542/peds.2006-0529. Pediatrics 2007;119(1):e284–91. Aquino VM, Sandler ES, Mustafa MM, Steele JW, Buchanan GR. A prospective double-blind randomized trial of urokinase flushes to prevent bacteremia resulting from luminal colonization of subcutaneous central venous catheters. J Pediatr Hematol/Oncol 2002;24(9):710–3. Dillon PW, Jones GR, Bagnall-Reeb HA, Buckley JD, Wiener ES, Haase GM. Prophylactic urokinase in the management of long-term venous access devices in children: a Children's Oncology Group study. J Clin Oncol 2004;22(13):2718–23. Ray Jr CE, Shenoy SS, McCarthy PL, Broderick KA, Kaufman JA. Weekly prophylactic urokinase instillation in tunneled central venous access devices. J Vasc Interv Radiol 1999;10 (10):1330–4. Kalmanti M, Germanakis J, Stiakaki E, Syfridaki C, Christidou A, Tsetis D, et al. Prophylaxis with urokinase in pediatric oncology patients with central venous catheters. Pediatr Hematol Oncol 2002;19(3):173–9.