Characterizing resolution of catheter-associated upper extremity deep venous thrombosis

Article Outline

• Abstract
• Patients and methods
• Results
  • Upper extremity DVT
  • Catheter-associated UEDVT
  • Catheter type
  • Thrombus location
  • Thrombus resolution
  • Duration of catheter placement
  • Catheter removal
  • Effect of anticoagulation
  • Influence of combined catheter removal and anticoagulation
  • Influence of patient characteristics
  • Outcome of new-site catheter placement for CAUEDVT
  • Pulmonary embolism
• Discussion
• Conclusions
• Author contributions
• References
• Copyright

Objective

Catheter-associated upper extremity deep venous thrombosis (CAUEDVT) is well known; however, resolution rates and factors affecting resolution of CAUEDVT are not well characterized. This study determined resolution rates and factors associated with resolution of CAUEDVT.

Methods

From January 1, 2002, to June 30, 2006, 1761 upper extremity venous duplex ultrasound (DU) studies were performed, and a new UEDVT was found in 253 (14.4%). Of these, 150 patients had routine follow-up and 101 had CAUEDVT. Demographics, follow-up DU results, and risk factors for venous thrombosis were recorded in the patients with follow-up studies and CAUEDVT. Univariate analysis and multivariate logistic regression analysis was performed to determine independent risk factors for complete thrombus resolution.

Results

There were 49 men (49%) and 52 women (51%) with CAUEDVT and follow-up studies. Mean age was 49 years (range, 5 months-80 years). Patients with CAUEDVT had risk factors for venous thrombosis that included malignancy in 34%, recent surgery/trauma in 34%, known hypercoagulable state in 11%, concomitant lower extremity DVT in 21%, and pulmonary embolism in 5%. Complete resolution of DVT on follow-up was documented in 46%. Thrombosis resolved in only 25% (6 of 24) when the catheter was not removed (P ≤ .05). Anticoagulation did not improve the rate of thrombus resolution (P ≤ 1.0) compared with catheter removal alone. Of the patients who had thrombus resolution, 75% resolved by 100 days (range, 1-914 days) after catheter removal ≤48 hours of diagnosis. In multivariate analysis, only catheter removal predicted the likelihood of thrombus resolution (odds ratio, 3.25; 95% confidence interval, 1.16-9.09; P = .025). New-site UEDVT developed in 86% of patients with CAUEDVT who underwent catheter removal and immediate catheter placement in a new site. Pulmonary embolism developed in five patients with CAUEDVT. Of these, three had documented lower extremity DVT as well. No pulmonary emboli were fatal.

Conclusions

More than half of CAUEDVT resolve ≤113 days when the catheter is removed ≤48 hours of diagnosis. New-site catheter placement has a high rate of new associated UEDVT. Anticoagulation does not appear to augment resolution of UEDVT.

Upper extremity deep venous thrombosis (UEDVT) occurs in 0.15% of hospitalized patients.¹ Indwelling central venous catheters have been reported to be a strong independent risk factor for UEDVT.² In one retrospective study of patients with duplex ultrasound (DU)-proven UEDVT, central venous catheters were present in 110 of 170 (65%) with UEDVT.³ This study also documented a 7% risk of pulmonary embolism (PE). Mortality in this patient population, although not directly attributed to the central venous catheters, was high, at 16% at 1 month and 34% at 3 months. Another study by the same authors evaluated 2451 upper extremity venous DU examinations and found 598 were positive for UEDVT, of which 307 were associated with a central venous catheter. PE occurred in 33 (5%), and mortality at 2 months was 29%.⁴

Although central venous catheters clearly are a significant risk factor for UEDVT, little is currently known about the factors affecting resolution of catheter-associated UEDVT (CAUEDVT). Specifically, we do not know the effect of catheter removal or anticoagulation on resolution of CAUEDVT. The current study was undertaken to further characterize the natural history of CAUEDVT and to identify factors that may be associated with thrombus resolution.

Back to Article Outline

Patients and methods 

From January 1, 2002, to June 30, 2006, symptomatic patients undergoing DU imaging of the upper extremities were identified from a database of all vascular laboratory studies performed at Oregon Health and Science University (Portland, Ore). Most symptoms were swelling or pain, or both, in upper extremities. According to the protocol, patients were examined supine with the head turned opposite to the extremity under examination. Except when impeded by the presence of bandages or fixation devices, named superficial and deep upper extremity veins were examined, including the internal jugular, subclavian, axillary, brachial, cephalic, and basilic veins.

DVT was diagnosed by the presence of intraluminal echoes with lack of flow response to augmentation and compression maneuvers in the internal jugular, subclavian, axillary, or brachial veins. Location and extent of thrombus were documented by the vascular laboratory technician on predesigned upper extremity worksheets that included the upper extremity venous anatomy. A distinction between acute and chronic thrombus was not made unless it was obvious (ie, patient had previous diagnosis of UEDVT). The criteria used to describe an acute thrombus were new symptoms and lack of criteria suggesting chronic thrombus (ie, echogenic veins).

Patients with CAUEDVT diagnosed by DU imaging and who had had at least one follow-up DU examination were selected for this study. UEDVT was considered catheter-associated if a central catheter was present in the anatomic location of the UEDVT ≤30 days from the DU examination documenting DVT. Partial resolution of thrombus was defined as recanalization of at least one venous segment on a follow-up DU study. Complete resolution was defined as the absence of venous thrombus on the last follow-up DU examination.

The electronic medical records were reviewed for documentation of the central venous catheter. Central venous catheter placement and the type of catheter were further confirmed by additional review of all available inpatient and outpatient progress notes, procedure notes, and radiologic images and reports. Documented thrombosis risk factors included inherited and acquired hypercoagulable states, recent surgery, trauma, or immobilization ≤30 days, and cancer. The indication for catheter placement, the type of catheter placed, and the site of catheter placement were also recorded. Catheter removal was noted along with when the catheter was removed and the indication for removal.

Treatment with anticoagulation was documented. Patients were not considered to be anticoagulated unless they were given therapeutic doses of heparin or warfarin, or both. The indication, timing, duration, and therapeutic levels (international normalized ratios) of anticoagulation were recorded. Complications of anticoagulation were noted, along with the presence of lower extremity DVT and PE. Lower extremity DVT was diagnosed with compression DU imaging. A diagnosis of PE was made by ventilation/perfusion scan, spiral chest computed tomography angiography, or by pulmonary angiography. Follow-up DU studies were classified by whether they showed complete resolution of the initial CAUEDVT.

Statistical analysis was performed using χ² analysis for continuous variables in a univariate model. Multivariate logistic regression analysis was used to determine independent risk factors for complete thrombus resolution. Odds ratios (OR) are presented with their 95% confidence intervals (CI). Statistical analysis was completed with SPSS 17.0 software (SPSS Inc, Chicago, Ill).

Back to Article Outline

Results 

Upper extremity DVT 

During the period reviewed, 1761 upper extremity venous DU scans performed; of these, 454 (25.7%) were positive for UEDVT, and a new diagnosis of UEDVT was noted in 253 (14.4%). A total of 226 follow-up examinations were available for 150, representing 1.51 follow-up studies per patient (range 1-10 studies). The average number of follow-up studies for the patients with CAUEDVT was 1.76. Follow-up studies were obtained from 1 day to 7.9 years (2579 days) after the initial diagnosis of UEDVT (Fig 1). There were 70 men (47%) and 80 women (53%). The mean age was 47 years (range, 5 months-99 years).

• 
  • View Large Image
  • Download to PowerPoint

• Fig 1. 

  Histogram shows the timing of follow-up studies relative to the initial diagnosis.

Catheter-associated UEDVT 

Of the 150 patients with follow-up studies, CAUEDVT was present in 101 (67%), comprising 49 men (48.5%) and 52 women (51.5%) who were a mean age of 49 years (range, 5 months-80 years). Additional risk factors, aside from the central venous catheter in the patients with CAUEDVT, included malignancy in 34 (34%), recent surgery or trauma in 34 (34%), and a known hypercoagulable state in 11 (11%). Concomitant lower extremity DVT was present in 21 (21%), and PE was documented in five (5%; Table I).

Table I. Factors affecting resolution of catheter-associated upper extremity deep venous thrombosis in 101 patients

Variable	Total	Resolved	Not resolved	P
Patients, No.	101	46	55
Catheter duration, d		123.35	129.44
Sex, No.
Men	49	25	24
Women	52	21	31	.3214
Catheter removed	77	40	37
Catheter removed <48 h	43	24	19	.4962
Catheter not removed	24	6	18	.0334
Anticoagulated	62	24	38	.1021
DVT location
Internal jugular		16	15	.5166
Subclavian		7	3	.1790
Axillary		3	4	>.99
Brachial		4	3	.6990
Multilevel		16	30	.0705
Catheter type
Central venous		13	17	.8292
PICC		21	19	.3083
AICD		2	1	.5910
Portacatheter		2	3	>.99
Hemodialysis		8	15	.3409
Associated conditions
Cancer		14	20	.6728
Hypercoagulable state		2	9	.624
Lower extremity DVT		10	11	>.99
Pulmonary embolism		2	3	>.99

AICD, Automatic implantable cardioverter-defibrillator; DVT, deep venous thrombosis.

Catheter type 

Of the 101 instances of CAUEDVT, there were 30 central venous catheters (average duration, 9.7 days; range, 0-27 days), 40 peripherally inserted central catheters (PICCs; average duration, 26 days; range, 0-188 days), 3 implantable cardioverter-defibrillators (ICDs; average duration, 1292 days; range, 486-2886), 5 portacaths (average duration, 496 days; range, 66-1412), and 23 hemodialysis catheters (average duration, 221.95 days; range, 1-886 days). Placement for 22 of 23 hemodialysis catheters was through the internal jugular vein. Resolution of CAUEDVT occurred in 13 of 30 central venous catheters (43.33%), 21 of 40 PICCs (52.5%), 2 of 3 ICDs (67%), 2 of 5 portacaths (40%), and 8 of 23 hemodialysis catheters (34.78%).

Thrombus location 

Thrombus was isolated to the internal jugular vein in 31 patients, subclavian vein in 10, axillary vein in 7, brachial in 7, and was present in multiple deep sites in 46. Superficial venous thrombosis was present at the time of diagnosis in 63 of 101 cases (63%) with CAUEDVT. By comparison, superficial venous thrombosis was present at the time of diagnosis in only eight of 49 patients (16%) with non-CAUEDVT.

Thrombus resolution 

Of the 101 instances of CAUEDVT, resolution was documented in 46 (46%) by the follow-up DU study, and 33 (72%) of these had resolved by 100 days (range, 1-914 days). Partial resolution was seen in 21 (21%), no change in 22 (22%), and progression seen in 12 cases (12%; Table II). Complete resolution of thrombus occurred in 16 of the 31 cases (51.6%) of isolated internal jugular thrombus, 7 of 10 (70%) with isolated subclavian vein thrombus, 3 of 7 (42.8%) with isolated axillary vein thrombus, and in 4 of 7 (57.1%) with isolated brachial vein thrombus. There were no differences in the percentages of complete resolution of CAUEDV thrombi with respect to thrombus location (P = .1790 to >.99 for all comparisons). Complete resolution may have been affected by the involvement of the original CAUEDVT with multiple deep venous segments. Thrombus resolved completely in only 16 of 46 (34.7%) where thrombus involved two or more anatomic segments (P = .0705). By life-table analysis, resolution rates of CAUEDVT are projected from 18.8% at 1 month and increasing to 73.3% at 24 months (Table III, Fig 2).

Table II. Duplex ultrasound result by status of central venous catheter

Result	Removed, No. (%)	Not removed, No. (%)
	(n = 77)	(n = 24)
Resolved	40(51.9)	6(25)
Partial resolution	18(23.4)	3(12.5)
No change	14(18.2)	8(33.3)
Progression	5(6.5)	7(29.2)

Table III. Rate of deep venous thrombosis resolution

• 
  • View Large Image
  • Download to PowerPoint

• Fig 2. 

  Kaplan-Meier curve depicts the rate of catheter-associated upper extremity deep venous thrombosis (DVT) resolution by follow-up duplex ultrasound imaging.

Duration of catheter placement 

Resolution rates of CAUEDVT were not affected by the length of time the catheter was in place before the diagnosis of CAUEDVT. The average duration of catheter placement was 127 days (range, 0-2886 days) and was placed ≤30 days before diagnosis in 61 of the 101 patients (61%). Thrombus resolved in 30 (49%) and failed to resolve in 31 (51%). In 21 patients with CAUEDVT, catheters were in place 31 to 90 days before the diagnosis of CAUEDVT, with eight (26%) resolving. The rate of resolution for CAUEDVT for catheters in place >90 days before the diagnosis of CAUEDVT was 42% (8 of 19; P = .41 to .82 for all comparisons to catheters in place ≤30 days).

Catheter removal 

Of the 101 catheters, 77 were removed because of symptoms or after discovery of the CAUEDVT, and of these, 17 were discontinued before the initial DU examination but ≤30 days of the DU study. Complete resolution of CAUEDVT occurred in 51.9% of the patients when the catheter was removed. Only six of 24 CAUEDVTs (25%) completely resolved when the catheters were not removed. The catheter was removed ≤48 hours of diagnosis in 43 of the 77, including four of 17 in whom the catheter was removed before the DU examination. Thrombus resolution in these patients was observed in 24 of 43 (55.8%). Of the patients who had thrombus resolution, 75% resolved by 100 days (range, 1-914 days) after catheter removal ≤48 hours of diagnosis. When the catheter was removed >48 hours after diagnosis, the thrombus resolved in 16 of 34 patients (47%) at an average of 94 days (range, 5-438 days) from catheter removal (P = .4962; Fig 3, A and B).

• 
  • View Large Image
  • Download to PowerPoint

• Fig 3. 

  A, Kaplan-Meier curve depicts the rate of catheter-associated upper extremity deep venous thrombosis (DVT) resolution with and without catheter removal. B, Kaplan-Meier curve depicts the effect of early catheter removal on the resolution of catheter-associated upper extremity DVT.

Effect of anticoagulation 

Anticoagulation was used in 62 of the 101 patients (62%) in the study. The ordering physician chose the type of anticoagulation. Most patients received unfractionated heparin, with eventual transition to warfarin. Thrombus resolution was documented in 24 of 62 patients (38%) when the patient was anticoagulated. Thrombus resolved in 22 of 39 patients (56.4%) who were not anticoagulated for CAUEDVT (P = .1021). After risk factor analysis was factored into the anticoagulated vs the non anticoagulated group, we found the two groups were similar in risk, and there were no significant differences between them.

Resolution was seen in 19 of 41 patients (46%) when anticoagulation was instituted ≤48 hours of diagnosis vs resolution in five of 21 (23.8%) when anticoagulation was started >48 hours after diagnosis (P = .1043). Anticoagulation also was not protective against progression of CAUEDVT, which was found in 12 patients on follow-up DU scanning. Eight had been treated with anticoagulation for the CAUEDVT, and five of the 12 patients underwent catheter removal during follow-up. Anticoagulation was stopped for bleeding complications in 16 of the 62 patients (25.8%) who were treated with anticoagulation.

Influence of combined catheter removal and anticoagulation 

Catheters were removed in 77 of 101 cases (77%) of CAUEDVT, anticoagulation was used in 62 (62%), and both catheter removal and anticoagulation were used in 49 (49%). In multivariate analysis, only catheter removal predicted likelihood of thrombus resolution (OR, 3.25; 95% CI, 1.16-9.09; P = .025).

Influence of patient characteristics 

Risk factors in patients with CAUEDVT included malignancy in 34%, recent surgery or trauma in 34%, end-stage renal disease in 33%, congestive heart failure in 11%, known hypercoagulable state in 9%, and diabetes mellitus in 8%. In multivariate analysis, only end-stage renal disease significantly predicted failure to resolve thrombus (95% CI, 1.6-13.1, P = .004).

Outcome of new-site catheter placement for CAUEDVT 

In 11 patients with a CAUEDVT, the catheter was removed, and a new catheter was placed at a new site presumably not involved with thrombus. This sequence occurred twice in three of the patients where the second catheter also became involved with thrombus. Overall therefore, there were 14 instances in which a new catheter was placed in a patient with CAUEDVT at a site presumably not previously involved with thrombus. A new CAUEDVT developed in association with the new catheter in 12 of the 14 (86%) subsequent catheter placements with all but one of these new CAUEDVTs developing ≤10 days of placement of the new catheter. Three of these 12 patients were receiving anticoagulation therapy when the old catheter was removed, and a new catheter was placed in a different anatomic site.

Pulmonary embolism 

Five of the 101 patients were diagnosed with PE, none of which were fatal, and four underwent anticoagulation. The patient who did not undergo anticoagulation also had a negative result on a lower extremity venous DU examination. In three of the five patients with a diagnosed PE, DU imaging also showed a LEDVT. In the remaining patient, the lower extremity veins were not assessed for DVT.

Back to Article Outline

Discussion 

This study has some weaknesses, including the small number of patients, retrospective design, and variability of follow-up observations. There was no accepted protocol to monitor patients with newly diagnosed UEDVTs with DU examinations. Most of the follow-up DU examinations in this study were performed because of changes in symptoms or in planning the placement for a new catheter. There is therefore a selection bias to patients with changes in symptoms or patients requiring continued central venous access. Nevertheless, the results are intriguing in that they confirm but also challenge some traditional thoughts about CAUEDVT and its management while providing some new follow-up imaging data on CAUEDVT.

CAUEDVT is a common problem. DVT is obviously a very common problem as well, and UEDVT accounts for 1% to 4% of all cases of DVT,⁵ with central venous catheters by far the most common risk factor for UEDVT. In four studies, central venous catheters were associated with 65%, 51%, 55%, and 60% of cases of UEDVT.¹, ², ³, ⁴ CAUEDVT has been documented in 33% of 208 consecutive central venous catheters placed in intensive care unit patients.⁶ In our study as well, a high proportion of patients with UEDVT had CAUEDVT.

Although CAUEDVT is frequent, there is remarkably little follow-up imaging information available, and management is not standardized. Only two previous studies have provided any follow-up imaging information at all after the diagnosis of UEDVT, and not all patients in those studies had CAUEDVT. One author evaluated 53 patients with a diagnosis of UEDVT, and only six were CAUEDVTs. All patients in this study were anticoagulated for 3 months. With follow-up DU examinations every 6 months for 5 years, recurrent UEDVT was documented at 1, 2, and 5 years as 2.0%, 4.2%, and 7.7%, respectively.⁷

A study of subclavian vein stenosis in patients undergoing hemodialysis used venography to document a recanalization rate of 45.4% at 3 months after catheter removal.⁸ There was a trend toward increased recanalization when catheters had been in place <29 days. Therefore, although the number of patients and the information provided here is not ideal, it considerably exceeds previously available data for patients with CAUEDVT monitored with imaging studies. In this series of 101 cases of CAUEDVT with available follow-up DU examinations, we observed an absolute resolution rate of 46% when the average was 100 days to the follow-up DU examination (range, 1-914 days). Only removal of the catheter seemed to affect DU-documented complete resolution of CAUEDVT.

Of interest is that anticoagulation did not appear to affect the rate of resolution or decrease progression of CAUEDVT and was associated with a significant incidence (about 25%) of bleeding complications. Progression of CAUEDVT occurred despite anticoagulation in eight of the 12 patients with documented progression of catheter-associated thrombus; however, seven of these patients with progression of catheter-associated thrombus were already at high risk for progression because of comorbidities, including malignancy, coagulopathy, and lack of catheter removal. Anticoagulation was stopped in 16 of the 62 patients because of bleeding complications. This high rate of bleeding complications may be because those who require central venous access are higher-acuity patients and perhaps are more prone to a bleeding diathesis than the usual patient with lower extremity DVT.

In addition, documented PE without associated lower extremity DVT was unusual in this patient group (2%), and none were fatal. Although screening studies for PE were not performed, the low rate of confirmed PE indicates that short-term clinically important PE is unlikely in patients with isolated CAUEDVT. A low rate of PE in patients with UEDVT has also been recently documented in a community hospital setting.⁹ The low rate of PE in patients with isolated CAUEDVT, combined with an apparent lack of benefit of anticoagulation on the thrombus itself and the associated bleeding complications of anticoagulation in the patients with CAUEDVT, must lead one to question, despite recommendations of the American College of Chest Physicians,¹⁰ the utility of anticoagulation in patients with isolated CAUEDVT who do not have an additional indication for anticoagulation.

It did not appear that prompt removal of the catheter resulted in dramatically improved resolution of CAUEDVT. What mattered was removal of the catheter, not when it was removed. In addition in this study, removal of central venous catheters associated with thrombus and subsequent placement of a new central venous access at a new site previously uninvolved with thrombus almost always resulted in new CAUEDVT. The implication is that it may be best in a patient with CAUEDVT to leave the catheter in place as long as it is needed, functions, and is not infected and then remove the catheter once it is no longer required for therapy. Such an algorithm will limit the development of UEDVT and may not adversely affect resolution of the initial CAUEDVT.

Back to Article Outline

Conclusions 

Additional studies are needed to confirm the current observations. Nevertheless, if the observations presented here are confirmed, it may be possible to greatly simplify the management of catheter-associated upper extremity deep venous thrombosis by eliminating the common practices of anticoagulation and new-site catheter placements in patients with catheter-associated UEDVT.

Back to Article Outline

Author contributions 

Back to Article Outline

References 

1. Mustafa S, Stein P, Patel K, Otten TR, Holmes R, Silbergleit A. Upper extremity deep venous thrombosis. Chest. 2003;123:1953–1956
   • View In Article
   • MEDLINE
   • CrossRef
2. Joffe H, Kucher N, Tapson V, Goldhaber SZ Deep Vein Thrombosis (DVT) FREE Steering Committee. Upper-extremity deep vein thrombosis a prospective registry of 592 patients. Circulation. 2004;110:1605–1611
   • View In Article
   • CrossRef
3. Hingorani A, Ascher E, Lorenson E, DePippo P, Salles-Cunha S, Scheinman M, et al. Upper extremity deep venous thrombosis and its impact on morbidity and mortality rates in a hospital-based population. J Vasc Surg. 1997;26:853–860
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (953 KB)
   • CrossRef
4. Hingorani A, Ascher E, Marks N, Schutzer RW, Mutyala M, Yorkovich W, et al. Morbidity and mortality associated with brachial vein thrombosis. Ann Vasc Surg. 2006;20:297–300
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (458 KB)
   • CrossRef
5. Joffe HV, Goldhaber SZ. Upper extremity deep vein thrombosis. Circulation. 2002;106:1874–1880
   • View In Article
   • CrossRef
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, risk factors, and relationship with catheter-related sepsis. Chest. 1998;114:207–213
   • View In Article
   • MEDLINE
   • CrossRef
7. Prandoni P, Bernardi E, Marchiori A, Lensing AW, Prins MH, Villalta S, et al. The long-term clinical course of acute deep vein thrombosis of the arm: prospective cohort study. BMJ. 2004;329:484–485
   • View In Article
   • CrossRef
8. Hernández D, Díaz F, Rufino M, Lorenzo V, Pérez T, Rodríguez A, et al. Subclavian vascular access stenosis in dialysis patients: natural history and risk factors. J Am Soc Nephrol. 1998;9:1507–1510
   • View In Article
   • MEDLINE
9. Spencer FA, Emery C, Lessard D, Goldberg RJ Worcester Venous Thromboembolism Study Investigators. Am J Med. 2007;120:678–684
   • View In Article
   • Abstract
   • Full Text
   • Full-Text PDF (110 KB)
   • CrossRef
10. Büller HR, Agnelli G, Hull RD, Hyers TM, Prins MH, Raskob GE. Antithrombotic therapy for venous thromboembolic disease: the seventh ACCP conference on antithrombotic and thrombolytic therapy. Chest 126:401S-28S.
    • View In Article