EBM

Erik Zeger, M.D.

August 10, 2001

 

CAT     Low dose warfarin preventing thrombosis in central venous catheters

 

Clinical bottom line:  Low dose warfarin (1mg/day) does lower the risk of thrombosis in patients with central venous catheters.

 

Our patient:  53yo female with stage IV breast cancer with a central line (Port-a-cath) and noted to be taking warfarin 1mg/day.

 

Citation:  Very low doses of warfarin can prevent thrombosis in central venous catheters. A randomized prospective trial. Annals of Internal Medicine. 1990;112:423-428.

 

Clinical question:  In patients with a central venous catheter, does low dose warfarin decrease the risk of thrombosis?

 

Search terms:  “warfarin” and “central venous catheter” in MEDLINE 1966 to present

 

Design:  Prospective randomized trial of patients at risk for thrombosis associated with chronic indwelling central venous catheters were assigned to receive or not receive 1mg of warfarin per day.  Warfarin was initiated 3 days prior to catheter insertion and continued for 90 days.  Prothrombin times were measured weekly for 1 month then monthly unless clinically necessary.  Subclavian, innominate, and superior vena cava venograms were done at the onset of thrombosis symptoms or after 90 days.  Venograms were assessed by radiologists unaware of the patient’s status in the trial.

 

The patients:  121 consecutive oncology patients requiring central venous catheters and with projected survival times greater than 3 months were enrolled.  Approx. 70% of patients had adenocarcinomas.  See table 2 for baseline group characteristics.  Sclerosing chemotherapeutic drugs are a suspected risk factor for thrombosis.  Based on this assumption, randomization was stratified for this subset of patients. 

 

Exclusion criteria:  baseline platelet counts under 125K, acquired or congenital coagulopathies, previous subclavian vein catheters, obstructing mediastinal tumors, prior DVT, anatomic lesions that bleed (ulcers),  creatinine greater than 1.6mg/dL, platelet suppressing medications.

 

Results:  42 of 60 patients in the warfarin group (70%) and  40 of 61 patients in the control group (66%) completed the study.  See table 3 for causes for withdrawal.  Of the 26 patients who died during the study,  no deaths were attributed to their catheters, warfarin therapy, nor to thrombosis. 

 

 

 

 

In the warfarin group:  (n= 60; 42 analyzed)  4 of 42 patients had venogram-documented thrombosis (9.5%). 

In the control group:  (n= 61; 40 analyzed)  15 of 40 patients had venogram-documented thrombosis (37.5%).

 

  Occurrence of thrombosis at 90 days in patients with central venous catheters ± low dose warfarin:

No warfarin

Control event rate (CER)

Warfarin

Experimental event rate (EER)

Relative risk reduction

 (RRR)

Absolute risk reduction

(ARR)

Number needed to treat

(NNT)

          37.5%

          9.5%

          75%

          28%

          3.6

 

 

95% confidence interval      >>>>>

±18% = 10-46%

        2 to 10

 

 

 

Comments:

Are the results of the therapeutic trial valid? 

  1. While the study was randomized the prospective,  neither the patients nor their physicians were blinded. 
  2. Unclear whether a 90 day period is adequate to assess the efficacy of the treatment.
  3. There were a significant number of patients who did not complete study.  All patients remained in the randomized groups, but all were not analyzed.
  4. Were the groups treated equally?  Overall yes, but since the groups were not blinded many more coagulation tests were drawn on the experimental group.  While this may not be significant, were there other unequal treatments given?
  5. It does appear that both groups were similar at the start of the trial.

 

Are the valid results of the trial important?

 

  1. Both the magnitude and the precision of the treatment effect appear to be significant if looking at the smaller (42 and 40) patient groups.  Would these results hold if we analyzed with the original group numbers?

 

Are these valid, important results applicable to our patient?

  1. Our patient is sufficiently similar to the study group in age and underlying problem (malignancy) that the results are applicable. 
  2. The treatment is inexpensive and easily performed. 

 

Other:

  1. Is this pt population applicable to all pt with central venous catheters?
  2. Is our thrombosis endpoint adequate?  Should we be looking at mortality?
  3. No clear assessment of harm.
  4. Will you give your pts with central venous lines low dose warfarin?