Neurocrit Care (2011) 15:454–460 DOI 10.1007/s12028-011-9554-3

ORIGINAL ARTICLE

The Clinical Significance of Peripherally Inserted Central Venous Catheter-Related Deep Vein Thrombosis Jeffrey J. Fletcher • William Stetler Thomas J. Wilson

•

Published online: 4 May 2011 Ó Springer Science+Business Media, LLC 2011

Abstract Background Peripherally inserted central venous catheters (PICCs) are being increasingly utilized in hospitalized patients as alternatives to centrally inserted central venous catheters (CICVCs). However, concern exists over the risk of PICC-related large vein thrombosis (PRLVT). The incidence rate and significance of symptomatic PRLVT in critically ill patients admitted to the neurological intensive care unit (ICU) is not known. Methods Retrospective descriptive study of consecutive PICCs placed in critically ill patients admitted to a tertiary care neurological ICU between March 2008 and February 2010. Symptomatic PRVLT was defined as an event that prompted Duplex ultrasound of the ipsilateral extremity in which an acute, proximal large vein thrombosis was confirmed in association with the PICC or confirmed within 5 days of PICC removal. Incidence rate of PRLVT and catheter-related complications were calculated per ‘‘line’’ (catheter). Descriptive statistics were performed with twosample, and t-tests for age and categorical variables were

J. J. Fletcher
W. Stetler
T. J. Wilson Department of Neurosurgery, University of Michigan, 3552 Taubman Health Care Center, 1500 E. Medical Center Drive, Ann Arbor, MI 48109-5338, USA W. Stetler e-mail: [email protected] T. J. Wilson e-mail: [email protected] J. J. Fletcher (&) Department of Neurosurgery, University of Michigan Hospitals and Health Center, 3552 Taubman Health Care Center, 1500 E. Medical Center Drive, Ann Arbor, MI 48109-5338, USA e-mail: [email protected]

123

assessed by Chi-square test or Fishers exact test as appropriate Results Four hundred and seventy-nine lines were placed during the study period with 39 developing a symptomatic PRLVT (incidence rate = 8.1%). Male gender was associated with development of a thrombosis (P = 0.02), but size (P = 0.21) and location of catheter were not (P = 0.30). Median line dwell time was 12 days (IQR 16) with a dwell time of 8 days (IQR 9) until thrombosis diagnosis. Pulmonary embolus attributed to PRLVT occurred in 1.3% of line placements and 15% of symptomatic PRLVT. The majority of patients had their line removed. In addition, some patients also had anticoagulation initiated or a superior vena cava filter placed. Conclusions Symptomatic PRLVT is not uncommon in critically ill patients admitted to the neurological ICU. Future research should focus on indentifying modifiable risk factors for PRLVT and on comparing major cumulative complication rates between PICCs and CICVCs. Keywords Central venous catheterization
Upper extremity deep venous thrombosis
Secondary upper extremity deep vein thrombosis
Venous thrombosis
Central venous catheter thrombosis

Introduction Peripherally inserted central venous catheters (PICCs) are being increasingly utilized in hospitalized patients as alternatives to centrally inserted central venous catheters (CICVCs) [1, 2]. Though major mechanical complications of placement are avoided, cumulative complication rates may not be decreased with PICCs compared to CICVCs as concern exists over the risk of PICC-related large vein

Neurocrit Care (2011) 15:454–460

thrombosis (PRLVT) [3, 4]. Recent trials of hospitalized patients have defined PRLVT as thrombosis affecting the deep veins or the large superficial veins (basilic or cephalic) in the proximal upper extremity or neck. Rates of PRLVT are documented between 3 and 58% [4–7]. Importantly, the highest rates of PRLVT are reported from patient populations who have required PICC placement in the intensive care unit (ICU) [5, 7]. Conversely, the rate of large vein thrombosis related to CICVCs is reported to be much lower, in the range of 0–10% in ICU patients, after excluding the femoral site [5, 6, 8, 9]. Though catheter-associated large vein thrombosis occurs more frequently with PICCs than CICVCs, the clinical significance of PRLVT in hospitalized patients has been questioned [5, 10]. Some authors have suggested PRLVT, whether superficial or deep, is almost entirely asymptomatic [3, 5, 10, 11]. Conversely, a large portion of the existing literature consists of retrospective or prospective observational studies in which thrombosis was identified on imaging studies obtained due to symptoms, most commonly pain, swelling or erythema in the ipsilateral arm [4, 6, 12, 13]. In addition, a recently halted prospective study using routine ultrasound in an ICU patient population demonstrated 20% of patients developed symptomatic PRLVT [7]. Though reported in a few studies, there remains a paucity of data concerning the occurrence of pulmonary emboli and late complications of PRLVT, especially in ICU patient populations [4, 6, 14]. More data exist on spontaneous upper extremity deep venous thrombosis (UEDVT) which is frequently reported to be symptomatic, including the risk of pulmonary embolus [15, 16]. Conversely, it is rare for UEDVT related to CICVCs to become symptomatic, with one large trial in ICU patients demonstrating all CICVC thrombosis to be asymptomatic by 30 day follow-up [9]. No study has been completed evaluating the incidence, significance, or independent risk factors for PRLVT in critically ill patients. The only study to exclusively evaluate PICC lines in the ICU was terminated early after enrollment of only 50 patients [7]. Accordingly we performed a descriptive study to determine the incidence rate and significance of symptomatic PRLVT in critically ill patients admitted to the Neurological ICU.

Methods Study Design This study was approved by the institutional review board at the University of Michigan. Descriptive data were obtained from an ongoing retrospective analysis to identify independent risk factors for PRLVT in patients admitted to the Neurological ICU. After approval information

455

technology designed and implemented a search paradigm to query the University of Michigan information systems and electronic medical record to identify all patients admitted to the Neurological ICU from March 2008 to February 2010. This patient list was cross referenced with the electronic order entry system to identify all consecutive patients who underwent PICC insertion during their ICU stay. All patients underwent preliminary chart review to ensure a PICC line had been placed and the following baseline demographics were abstracted from the electronic medical chart: age, sex, ethnicity, size and location of line insertion, length of stay, primary diagnosis, number, and indication for ultrasounds of the upper extremity. Patients found to have a PRLVT had the following additional information abstracted from the medical record: tobacco abuse, obesity (BMI > 30 kg/m2), prothrombotic state, coagulopathy, cancer, congestive heart failure, pregnant, history of venous thromboembolism (VTE), history of deep venous thrombosis in the ipsilateral arm, estrogen use, antiplatelet use, statin use, line flush, VTE prophylaxis, infused with vancomycin or hypertonic saline, or mannitol, surgery greater than 1 h during dwell time, placement in a paretic arm, catheter tip location, catheter diameter, attempts at placement and line manipulations, duration of use, dwell time until thrombosis, location of PRLVT, presence of clotted ports requiring thrombolytics, treatment of thrombosis, need for replacement with another central venous catheter if PICC removed, PRLVT-associated pulmonary embolus, and late (sub-acute and chronic) complications on follow-up visit. Definitions A PICC ‘‘line’’ was defined similar to other studies in which any line exchanged, or replaced within 24 h in same arm, was counted as one line of continued duration [7]. The incidence rate of symptomatic PRLVT was made by review of Duplex ultrasound reports. Symptomatic PRVLT was defined as an event that prompted Duplex ultrasound of the ipsilateral extremity in which an acute, proximal large vein thrombosis was confirmed in association with the PICC or confirmed within 5 days of PICC removal. In patients with a past history of ipsilateral UEDVT, new symptoms must have prompted the Duplex ultrasound and acute thromboses must have been confirmed. All venous Duplex ultrasonography were performed with a portable Siemens Sonoline Antares (Siemens Medical Solutions, Inc, Malvern Pa) or Toshiba Xario XG (Toshiba America Medical Systems, Inc, Tustin, CA) Machine and were interpreted by board certified radiologists. Non-compressibility as well as Doppler mode was used during evaluation. Deep venous thrombosis was defined as thrombosis of the brachial, axillary, or subclavian vein, and extensive clots

123

456

were defined as clots in or proximal to the axillary vein. Similar to others, the basilic and cephalic veins above the elbow were considered as large veins involved in symptomatic thrombosis [6, 7]. All vessels with thrombus were documented; hence, one line could be associated with multiple thromboses. A pulmonary embolus attributed to a PRLVT was defined as symptoms triggering a helical computed tomography (CT) angiography scan of the pulmonary vessels, a high probability ventilation perfusion scan of the chest, or a pulmonary angiogram. However, all patients in our cohort suspected of a pulmonary embolus received CT angiography of the pulmonary vessels which is routinely done at our institution along with CT venography of the pelvis and lower extremities. All CT scans were performed on a 64-MDCT (Light Speed VCT, GE Healthcare), and the protocol used 1.25 mm collimation reconstructed at 0.625-mm intervals and 125 ml of intravenous contrast material (Isovue-300, Bracco Diagnostics). To evaluate sub-acute and chronic complications, we reviewed clinic records from follow-up visits and primary care physician visits within 2 years of discharge. Information on persistent pain, swelling, erythema, cramping, claudication, compromised venous access, symptoms of pulmonary embolism, or documented diagnosis of pulmonary embolus reported from outside our health system were abstracted. These were counted as PRLVT associated if determined by clinic physician. Late complications were defined as present after 6 months. PICC Insertion and Maintenance PICCs were inserted at the bedside by the vascular access team using sterile precautions and a comprehensive prevention program or rarely by interventional radiology if bedside insertion was unsuccessful or inappropriate. Five French (f) or six f polyurethane double lumen (18gauge/ 18gauge) power PICC (Bard Access Systems, Salt Lake city, UT) were used. No anti-thrombotic, antibiotic, and antiseptic material are used in these lines. Real time ultrasound guidance and standard maneuvers were used during insertion. A navigator tip locating system was used during some placements. PICCs were inserted via a modified seldinger technique. Following puncture of the vein with a 21-gauge needle, a guidewire (0.018-inch) was inserted into the vein followed by a peel-away sheath. Catheter length was determined by a standard algorithm. Catheters were sutured with nylon suture and a sterile occlusion dressing applied. All patients received a chest X-ray to confirm placement. The PICCs were adjusted or replaced if in suboptimal position (outside the superior vena cava or in the right atrium). Lines are routinely flushed with heparin unless a heparin allergy exists. Dressing

123

Neurocrit Care (2011) 15:454–460

changes occurred every 7 days or if soiled. Clotted lines received local thrombolysis per unit policy. In addition, our ICU uses the following measures as standard of care to reduce catheter-associated thrombosis in all patients with PICCs; (1) a sequential compression device is placed on the forearm on the side of the PICC; (2) range of motion is performed at least every 2 h on the upper extremity with the PICC; (3) pulsed infusion technique and clapping the PICC near the end of the flush instead of after completion of the flush; and (4) only use a 10 cc syringe to flush and not a 3 cc syringe. Statistical Analysis Statistical analysis was performed using commercially available software (SPSS version 18, IBM Corporation, Sumers, NY, USA). Variables were screened for normality using normality plots and the Shapiro–Wilks test. Parametric data were expressed as means + standard deviations (SDs), and non-parametric data were expressed as medians and interquartile ranges (IQRs). Descriptive statistics performed with two-sample t-tests for age and categorical variables were assessed by Chi-square test or Fishers exact test as appropriate

Results During the study period, there were 2,453 admissions to the Neurological ICU, comprising 2,150 patients. Though 501 PICCs were placed, this number was reduced to 479 after adjusting for PICCs that were exchanged or immediately replaced into the same arm after inadvertent removal. With the exception of male gender and number of Duplex ultrasounds ordered, baseline characteristics were well matched between those who developed a PRLVT and those who did not (Table 1). 44% of PICCs required at least two attempts, 47% required at least one manipulation, and 53.2% required either a manipulation or additional attempt to ensure proper placement on chest X-ray. Size of PICC line and location of insertion were not associated with thrombosis. Table 2 shows the descriptive characteristics of the 39 patients with symptomatic PRLVT. All PICC lines were in optimal position on chest X-ray, and 97% of PRLVT occurred in the presence of VTE prophylaxis and the utilization of heparin to flush the line. Median line dwell time was 12 days (IQR 16) with a dwell time of 8 days (IQR 9) until PRLVT diagnosis. Indications for Duplex ultrasound in those with a confirmed PRLVT can be seen in Fig. 1. Only three of the eight Duplex ultrasounds ordered for fever alone demonstrated an undiagnosed clot, and both ultrasounds ordered as routine were for follow-up of a known thrombosis. No patient had chemotherapy or total parenteral nutrition infused through the PICC line.

Neurocrit Care (2011) 15:454–460 Table 1 Descriptive statistics of 479 PICC placements over 24 months in the neurological ICU

457

Characteristic

All (479)

No clot (N = 440)

Symptomatic clot (N = 39)

P value

Age (mean ± SD)

56 (±16)

55 (±16)

57 (±)19

0.45 0.02*

Sex Female (%)

244 (51)

231 (53)

13 (33)

Male (%)

235 (49)

209 (47)

26 (67)

5 (%)

193 (40)

181 (41)

12 (31)

6 (%)

286 (60)

259 (59)

27 (69)

297 (62)

273 (62)

24 (62)

0.95

Right basilic (%)

222 (46)

203 (46)

19 (49)

0.30

Right brachial (%)

64 (13)

60 (13)

4 (10)

Size 0.21

Side Right (vs. left) (%) Location

Right cephalic (%)

17 (4)

16 (4)

1 (3)

Left basilic (%)

130 (27)

119 (27)

11 (28)

Left brachial (%) Left cephalic (%)

40 (8) 6 (1)

38 (9) 4 (1)

2 (5) 2 (5)

Basilic vs. other (%)

352 (74)

321 (73)

30 (77)

Number of lines

154 (33)

115 (26)

39 (100)

with at least one

N = 216

N = 162

N = 54

<0.000*

Duplex ultrasound ordered (%) (total number of * P < 0.05

ultrasounds)

Significance (Table 3) The incident rate for development of a PRLVT was 8.1% with 64% of thrombosis being extensive. The majority of PRLVT was diagnosed due to symptoms of ipsilateral arm pain or swelling with only one being diagnosed second to suspicion of pulmonary embolus. Six pulmonary emboli were attributed to symptomatic PRLVT (1.3% of PICC lines; 15% of symptomatic PRLVT). The majority of patients had their PICC line removed, though some patients also had anticoagulation initiated or a superior vena cave (SVC) filter placed (Table 3).

Discussion The incidence rate of symptomatic PRLVT (8.1%) in our cohort was slightly higher than reported in most other studies involving hospitalized patients. These studies have reported rates between 3 and 7% for symptomatic PRLVT; however, they have not evaluated patients exclusively in the ICU [4, 12, 13]. This limitation makes the rate of symptomatic PRLVT in the ICU difficult to determine, however, it is likely higher in critically ill patients than in other hospitalized patient populations. Indeed, the only

study evaluating patients exclusively in the ICU demonstrated a symptomatic PRLVT rate of 20%, prompting halting of the study [7]. The 6 month all cause mortality rate of 28% in our patients who developed a PRLVT confirms the high disease severity of our patient population. Compared to CICVCs, PICC lines are associated with fewer, easier to manage, insertion-related complications. However, our data do not necessarily suggest they are easier to place or that they are less painful than CICVCs. To the contrary, over half of the PICCs lines placed required at least one additional attempt or manipulation to ensure proper placement. Similarly, 90% of PRLVT was symptomatic due to pain or swelling. As previously mentioned, it is rare for UEDVT related to CICVCs in the ICU to become symptomatic [9]. Lending support to the theory that CICVCs and PICCs have a different risk of symptomatic thrombosis in critically ill patients is a recent study that reviewed all upper extremity Duplex ultrasounds ordered through the vascular lab at an academic medical center. The authors demonstrated that UEDVT was most frequently associated with PICCs followed by spontaneous formation and lastly, associated with CICVCs [13]. When evaluating PRLVT in critically ill patients, it does not seem appropriate to extrapolating data

123

458

Neurocrit Care (2011) 15:454–460

Table 2 Descriptive characteristics of 39 symptomatic PRLVT Characteristic

PRLVT (N = 39) (%)

Ethnicity Caucasian

32 (82)

African American

2 (5)

Asian

2 (5)

Hispanic

3 (8)

Hospital length of stay

28 days (IQR 14) Fig. 1 Indications for ultrasounds and veins involved in 39 PRLVT

Primary diagnosis Cerebrovascular

22 (56)

Tumor

3 (8)

Trauma

6 (15)

Table 3 Significance of 39 symptomatic PRLVT (among 479 line placements)

Spine Infection

1 (3) 5 (13)

Incident rate (per line) (%)

39/479 (8.1)

Other

25 (64)

2 (5)

Extensive PRLVT (%)

Tobacco use

9 (23)

Symptoms (initial)

Obesity

16 (41)

Swelling/edema/pain (%)

Prothrombotic state

2 (5)

Fever alone (%)

3 (7)

Coagulopathy

2 (5)

Pulmonary embolus (%)

1 (3)

Cancer

7 (18)

Pulmonary embolus attributed

Congestive heart failure

7 (18)

Pregnancy

0 (0)

To symptomatic PRLVT Per PICC line (%)

35 (90)

N = 6a 6/479 (1.3)

History VTE

10 (26)

History VTE (same arm)

3 (8)

Estrogen

0 (0)

Aspirin

5 (13)

Plavix

1 (3)

Statin

15 (39)

Flushed with Heparin

Requirement of new CVC (if line removed due to PRLVT) (%)

38 (97)

Mannitol

9 (23)

Treatment Removal (%)

Hypertonic saline

10 (26)

Anticoagulation (%)

13 (33)

18 (46)

SVC filter (%)

4 (11)

Nothing (%)

9 (23)c

Vancomycin

Per symptomatic PRLVT (%) Late complications (%) Death by 6 month follow-up (%) Lost to follow-up (%)

VTE prophylaxis Heparin SQ

35 (90)

LMWH

2 (5)

Other

1 (3)

None

1 (3)

Surgery >1 h (during line dwell time)

17 (44)

Placed in paretic arm

21 (54)

Cather tip in SVC

39 (100)

Attempts

1 (IQR 1) (range 1–3)

Manipulations

0 (IQR 1) (range 0–2)

Clotted lumens (TPA instilled) Dwell time of linea

7 (18)

Dwell time until clot

8 (IQR 9) (range 1–32)

12 (IQR 16) (range 3–52)

a

Expressed as median (interquartile range = IQR)

123

6/39 (15) 4/39 (11)b 11/39 (28) 4 (11) 17/28 (61)

28 (72)

a

Only two were not extensive thrombosis. One pulmonary embolus was attributed to a solitary basilic thrombosis, and a second pulmonary embolus was attributed to a clot involving the right basilic, brachial, and cephalic veins b

Two patients with persistent pain and swelling, one patient with compromised access, and one with a subdural hematoma while on warafarin

c

Only two extensive thrombosis had no treatment

from CICVCs, catheters used in cancer patients, or data from non-critically ill patient populations as different risk factors for VTE likely exists. Spontaneous UEDVT results in asymptomatic pulmonary embolism in up to 33% of cases and symptomatic pulmonary embolus in 9% of cases [15–17]. Limited data on catheter-associated UEDVT suggest similar rates of asymptomatic and symptomatic pulmonary emboli;

Neurocrit Care (2011) 15:454–460

however, this data is derived from few studies, mostly evaluating tunneled catheters used for chemotherapy in cancer patients or smaller (three and four French) catheters used in ambulatory patients to treat chronic infections [13, 18]. Rates of pulmonary emboli following catheter-associated large vessel thrombosis in critically ill or other hospitalized patients are not known. Lobo et al. reported pulmonary embolus in 1% (8/777) of hospitalized patients following PICC placement with an adjusted mortality rate of 25%; however, none could be attributed to PRLVT as no patient underwent upper extremity imaging [4]. Evans et al. documented pulmonary embolus in 11% (6/57) of symptomatic PRLVT (0.3% of all PICCs) by review of ICD-9 Codes in the 90 days following PICC placement [6]. Unfortunately, they were also unable to attribute the pulmonary emboli to catheter-related thrombosis as three patients had concurrent lower extremity deep venous thrombosis, one had heparin-induced thrombocytopenia, and one was diagnosed with anti-phospholipid syndrome [6]. Two additional studies evaluating PRLVT in critically ill patients did not confirm any pulmonary emboli; however, one trial was stopped early after enrolling only 50 patients due to high rates of thrombosis and the other trial performed routine ultrasound allowing early detecting and removal or treatment of thrombosis before clinical symptoms. Of concern we were able to attribute pulmonary emboli to symptomatic PRLVT in 1.3% of PICC line insertions (15% per symptomatic line). Given all patients had CT angiography of the chest and CT venography of the pelvis and lower extremities, this is the most accurate estimate of PICC line-related pulmonary embolus in ICU patients. However, even more concerning is that we cannot comment on the total occurrence of PICC-related pulmonary emboli given we did not perform routine ultrasound. The vast majority of Duplex ultrasounds in which a PRLVT was initially confirmed were ordered for pain or swelling in the ipsilateral extremity. Given a median line dwell time of 8 days until symptoms and a unimodal probability distribution (Fig. 2), it’s possible that active surveillance or routine Duplex ultrasound has the potential to reduce PRLVT-associated pulmonary emboli. However, until prospective data are available for the cost effectiveness of this approach cannot be evaluated. The treatment of PRLVT in our cohort is also concerning. The majority of patient had their PICC line removed, and 60% of these patients required another central venous catheter. Though it is possible that PICC lines were not being removed as soon as an indication no longer existed, a large percentage of patients were being exposed, a second time, to the risk of central venous catheter insertion. In addition, 33% of patients received anticoagulation and 11% a SVC filter. Anticoagulation carries significant risks in Neurological ICU patient populations

459

Fig. 2 Histogram display of the dwell time until PRLVT was detected

(indeed, one of our patients suffered a fatal SDH by followup visit), and the use of upper extremity filters is controversial as little is known about their complications or efficacy [19, 20]. The limitations of our study include that inherent to the descriptive study design as well as lack of much long-term follow-up. However, we believe the information obtained in our study has identified areas in which future research should focus. Major mechanical complications during insertion of a CICVC are dependent on operator experience and reported to be less than 3%, while the risk of catheterrelated blood stream infection is reported to be equivalent to PICC lines in hospitalized patients [8, 21, 22]. Given this we hypothesize cumulative major complications are increased with PICC lines compared to CICVCs. Future research should focus on indentifying modifiable risk factors for PRLVT and on comparing major complication rates between PICC lines and CICVCs exclusively in critically ill patient populations.

Conclusion The incidence rate of symptomatic PRLVT supports further research to identify modifiable risk factors and complications in critically ill patients. Conclusions regarding the risks versus benefits of PICC lines in the

123

460

neurological ICU can only be made after completion of a prospective randomized trial comparing PICCs to CICVCs. Conflict of interest The authors declare no conflict of interest or financial disclosures. No intra or extramural funding sources were used for this study.

References 1. Horattas MC, Trupiano J, Hopkins S, Pasini D, Martino C, Murty A. Changing concepts in long-term central venous access: catheter selection and cost savings. Am J Infect Control. 2001;29: 32–40. 2. Walshe LJ, Malak SF, Eagan J, Sepkowitz KA. Complication rates among cancer patients with peripherally inserted central catheters. J Clin Oncol. 2002;20:3276–81. 3. Abdullah BJ, Mohammad N, Sangkar JV, Abd Aziz YF, Gan GG, Goh KY, Benedict I. Incidence of upper limb venous thrombosis associated with peripherally inserted central catheters (PICC). Br J Radiol. 2005;78:596–600. 4. Lobo BL, Vaidean G, Broyles J, Reaves AB, Shorr RI. Risk of venous thromboembolism in hospitalized patients with peripherally inserted central catheters. J Hosp Med. 2009;4:417–22. 5. Bonizzoli M, Batacchi S, Cianchi G, Zagli G, Lapi F, Tucci V, Martini G, Di Valvasone S, Peris A. Peripherally inserted central venous catheters and central venous catheters related thrombosis in post-critical patients. Intensive Care Med. 2011; 37:284–99. 6. Evans RS, Sharp JH, Linford LH, Lloyd JF, Tripp JS, Jones JP, Woller SC, Stevens SM, Elliott CG, Weaver LK. Risk of symptomatic DVT associated with peripherally inserted central catheters. Chest. 2010;138:803–10. 7. Trerotola SO, Stavropoulos SW, Mondschein JI, Patel AA, Fishman N, Fuchs B, Kolansky DM, Kasner S, Pryor J, Chittams J. Triple-lumen peripherally inserted central catheter in patients in the critical care unit: prospective evaluation. Radiology. 2010;256:312–20. 8. McGee DC, Gould MK. Preventing complications of central venous catheterization. N Engl J Med. 2003;348:1123–33. 9. Timsit JF, Farkas JC, Boyer JM, Martin JB, Misset B, Renaud B, Carlet J. Central vein catheter-related thrombosis in intensive care patients: incidence, risks factors, and relationship with catheter-related sepsis. Chest. 1998;114:207–13.

123

Neurocrit Care (2011) 15:454–460 10. Periard D, Monney P, Waeber G, Zurkinden C, Mazzolai L, Hayoz D, Doenz F, Zanetti G, Wasserfallen JB, Denys A. Randomized controlled trial of peripherally inserted central catheters vs. peripheral catheters for middle duration in-hospital intravenous therapy. J Thromb Haemost. 2008;6:1281–8. 11. Paauw JD, Borders H, Ingalls N, Boomstra S, Lambke S, Fedeson B, Goldsmith A, Davis AT. The incidence of PICC line-associated thrombosis with and without the use of prophylactic anticoagulants. JPEN J Parenter Enteral Nutr. 2008;32:443–7. 12. King MM, Rasnake MS, Rodriguez RG, Riley NJ, Stamm JA. Peripherally inserted central venous catheter-associated thrombosis: retrospective analysis of clinical risk factors in adult patients. South Med J. 2006;99:1073–7. 13. Ong B, Gibbs H, Catchpole I, Hetherington R, Harper J. Peripherally inserted central catheters and upper extremity deep vein thrombosis. Australas Radiol. 2006;50:451–4. 14. Chemaly RF, de Parres JB, Rehm SJ, et al. Venous thrombosis associated with peripherally inserted central catheters: a retrospective analysis of the Cleveland Clinic experience. Clin Infect Dis. 2002;34:1179–83. 15. Moser KM, Fedullo PF, LitteJohn JK, Crawford R. Frequent asymptomatic pulmonary embolism in patients with deep venous thrombosis. JAMA. 1994;271:223–5. 16. Prandoni P, Polistena P, Bernardi E, et al. Upper-extremity deep vein thrombosis. Risk factors, diagnosis, and complications. Arch Intern Med. 1997;157:57–62. 17. Kirkpatrick A, Rathbun S, Whitsett T, Raskob G. Prevention of central venous catheter-associated thrombosis: a meta-analysis. Am J Med. 2007;120:901.e1–13. 18. Verso M, Agnelli G. Venous thromboembolism associated with long-term use of central venous catheters in cancer patients. J Clin Oncol. 2003;21:3665–75. 19. Owens CA, Bui JT, Knuttinen MG, Gaba RC, Carrillo TC. Pulmonary embolism from upper extremity deep vein thrombosis and the role of superior vena cava filters: a review of the literature. J Vasc Interv Radiol. 2010;21:779–87. 20. Kearon C, Kahn Sr, Agnelli G, et al. Antithrombotic therapy for venous thromboembolic disease: American College of Chest Physicians Evidence-Based Clinical Practice Guidelines (8th Edition). Chest. 2008;133:454S–545S. 21. Kusminsky RE. Complications of central venous catheterization. J Am Coll Surg. 2007;204:681–96. 22. Safdar N, Maki DG. Risk of catheter-related bloodstream infection with peripherally inserted central venous catheters used in hospitalized patients. Chest. 2005;128:489–95.