Pulmonary Embolism and Atrial Fibrillation

Motivation: Last November, I admitted a patient with dyspnea.  He had new onset atrial fibrillation, and discussion in the morning predictably revolved around causes of atrial fibrillation.  One of the causes mentioned was pulmonary embolism (the "green book" lists it under pulmonary triggers).  I find PE to be a strange trigger though.  I associate atrial fibrillation with processes that chronically stretch the atria resulting in a dilated atrium, but PE is an acute process.  Searching the literature at that time did not reveal any evidence for or against assoication of atrial fibrillation with PE.  Recently, when revisiting the question, I found an article by a European group published in December of 2011 addressing this issue.

Paper: Gex, G., Righini, M., Gal, G.L., et. al. "Is atrial fibrillation associated with pulmonary embolism?" J. of Thrombosis and Haemostasis (2011) [e-pub ahead of print]

Methods: To analyze association, the authors pooled together data from two large trials (CT-EP3 and CT-EP4) investigating PE diagnostic strategies.  In both studies, patients were included if they presented with clinical suspicion of PE defined largely as acute dyspnea or chest pain without obvious cause.  All patients had EKG at baseline with confirmation of PE diagnosis by CT.

Results:
Cohort: Total of 2,449 patients were analyzed.  The mean age was 59.9 years and 43.7% were males.  Of this group with clinical suspicion of PE, 22.6% had PE confirmed by CT.

Association with AF: Atrial fibrillation was detected at baseline in 133 patients.  Prevalence of PE was 22.8% (n=519) in patients without atrial fibrillation but 18.8% (n=25) in patients with atrial fibrillation (p = 0.28).  To adjust for differences in comorbidities between patients with AF and those without AF, the authors next created an adjusted model accounting for age, sex, CHF, COPD, stroke, renal clearance, and neoplasm.  After adjustment, there was no increased association with AF and PE (OR of 0.68, 0.42-1.11, p = 0.122).   AF failed to be associated with PE even in patient less than age 65 (OR 0.86, 0.35-2.12) or with no heart failure (OR 0.63, 0.37-1.06).  Authors attempted to separate association of AF with presenting symptoms (dyspnea or chest pain), but the numbers were too small to make meaningful conclusions.

Discussion: This paper establishes that in patients presenting with acute dyspnea or chest pain, finding atrial fibrillation on EKG does not meaningfully change the likelihood of PE.  If anything, finding atrial fibrillation had a trend towards reducing the likelihood of PE (adjusted OR of 0.68, 0.42-1.11).  This effect is likely from similar presenting symptoms of atrial fibrillations and PE.  While this paper in part answers my initial question of the utility of using atrial fibrillation to suspect PE, there are some important limitations.  First, the paper only examined patients presenting with acute dyspnea or chest pain.  The paper did not examine patients with new onset AF and ask how many had PE.  Therefore, the results cannot be extended to patients with atrial fibrillation without obvious symptoms though presumably the incidence of PE would be even lower.  Also, given the structure of the paper in which PE had to be a likely diagnosis, patients with AF with RVR and dyspnea were likely excluded.  Incidence of PE in this population is also unclear.  Depsite these limitations, I think that this paper provides important insights into the limited utility of considering AF in the diagnosis of PE.

Does venipuncture cause lymphedema for woman after breast cancer surgery?

Does venipuncture cause lymphedema for

woman after breast cancer surgery?

Motivation: An 85 year old woman is admitted from the ED for IV
antibiotic treatment of community acquired pneumonia. During the
night, the RN pages me because the peripheral IV in the patient’s
right hand infiltrated and attempts to place another IV in that arm
were unsuccessful. A simple solution would be to try on the left hand
but both the RN and the patient are adamantly against this idea
because the patient had “breast cancer surgery” on the left side 20
years ago. They state that inserting a peripheral IV in the left hand
would cause lymphedema which would become debilitating. However, it
seems peculiar to me that having a distant surgery would preclude the
use of an entire extremity therefore I set out to find the evidence.

Studies:
1. Clark B, Sitzia J, Harlow W. Incidence and risk of arm oedema
following treatment for breast cancer: a three-year follow-up study.
QJM. 2005
2. Winge C, Mattiasson AC, Schultz I. After axillary surgery for
breast cancer–is it safe to take blood samples or give intravenous
infusions? J Clin Nurs 2010;19:1270–1274.
3. May;98(5):343-8.Hayes SC, Janda M, Cornish B, Battistutta D, Newman
B. Lymphedema after breast cancer: incidence, risk factors, and effect
on upper body function. J Clin Oncol. 2008 Jul 20;26(21):3536-42.

Methods:
1. Prospective 3 year observational study of 188 woman who had
undergone surgery for breast cancer that involved sanpling, excision
or biopsy of their axillary nodes.
2. Questionnaires were mailed to all 311 patients who had undergone
removal of axilliary nodes in a major Swedish hospitals in 2000-2001.
Study published in 2010 so I assumed 7-8 years of follow up.
3. Observational study of 287 Australian women with invasive breast
cancer were evaluated 6-18 month after treatment.

Results:
1. At follow-up, 39 (20.7%) had developed lymphoedema. Hospital skin
puncture (vs. none) (RR 2.44, 95%CI 1.33-4.47), mastectomy (vs. wide
local excision or lumpectomy) (RR 2.04, 95%CI 1.18-3.54), and BMI > or
= 26 (vs. BMI 19-26) (RR 2.02, 95%CI 1.11-3.68) were the only
significant risk factors.
2. 88 of 311 women reported venipuncture on the affected arm. Only 4
patients developed lymphedema.
3. 33% (n = 62) of the sample were classified as having lymphedema; of
these, 40% had long-term lymphedema. Older age >50, more extensive
surgery (ie, mastectomy), and having a sedentary lifestyle
significantly increased odds (three- to six-fold) of lymphedema.
Removal of 20 or more lymph nodes, experiencing one or more
treatment-related complication or symptom, and being single each
increased odds 2.6- to 5.0-fold, but CIs were wide and included 1.0.
Conversely, having a lower yearly income significantly decreased odds
of lymphedema over the following 12 months five-fold.

Discussion: Lymphedema results when lymphatic drainage systems are
damaged leading to accumulation of interstitial fluid. More than 2
million American women are diagnosed with breast cancer each year with
more than 200,000 developing lymphedema secondary to cancer or its
treatment. Patients with lymphedema suffer chronic pain and swelling
that causes significant physical and mental anguish and leads to
decreased productivity. Treatment for this condition remains
suboptimal therefore attention has focused on prevention.

The above referenced studies all seem to agree that there are certain
risk factors for developing lymphedema including more extensive
surgery and obesity/sedentary life style. Two studies which
specifically looked at venipuncture in the affected arm had
conflicting results. Clark et al. 2005 was a prospective
observational study which provides level 2 evidence linking
venipuncture with increased risk of lymphedema while Winge 2010 was a
retrospective questionnaire study providing only level 3 evidence
against such a linkage. Clearly, the evidence is far from perfect.
Given the importance of the problem, a randomized clinical trial to
study preventive measures for lymphedema should be undertaken.

In the mean time, the results of these studies should motivate us to
attempt to clarify the nature of the patient’s cancer and surgery
before declaring an extremity off limits. Common sense seems to
suggest that we should avoid the affected arm in patients who are
older, obese and have had aggressive disease requiring more extensive
surgery and lymph node dissection. A patient who had DCIS and was
treated with breast conserving therapy without axillary dissection is
less likely to develop lymphedema. In addition, we need to take into
account the clinical context in which this dilemma takes place. For my
patient, a simple external jugular would have probably solved the
issue. On the other hand, in a patient who is having a massive upper
GI bleed and needs every single access that we can provide, the
uncertain risk of future lymphedema is clearly secondary to the
immediate risk of death from hypovolemia.

Lastly, a quick search of the internet will reveal that breast cancer
survivor’s networks and patient advocacy websites regarding lymphedema
are extremely prevalent and vocal about the need to avoid the affect
extremity. Therefore, communication with our patients is essential.
As with any medical procedures, the risks and benefits should be
clearly delineated and the patient included in the decision making
process. In this way, we avoid the anger and frustration that develops
when patients feel their concerns are not addressed, even in the
absence of lymphedema itself.