Transcript Testing for DVT/PE

Testing for DVT/PE
Steve Kizer MD
Why do the strategies for testing for
thromboembolic disease seem so difficult?



Confusion as to the goals of treatment
Poor understanding of the tests
Confusion by proxy
The goals of therapy

By far, most patients presenting with PE
are stable.



For those that are critical, attention is directed to
removing the embolus.
But for nearly all patients, our goal is to prevent a
second PE with little regard to the existing PE.
In some series, 60% of those presenting
with DVT will already have a PE at the time
of presentation.
The Goals of therapy, (cont’d)




If we can show that the deep venous system is
empty, or becomes empty, then PE will not
recur. (Hull & Hirsch, 1997)
Calf vein thrombi (below the popliteal space) are
not a cause of PE and are a risk factor for PE
only if they extend proximally.
Most pelvic disease will also propagate distally.
In a patient suspected of PE, studies of the
proximal veins are often negative because the
clot has moved.
The goals of therapy, (cont’d)

Therefore the goal of therapy is
straightforward:



If the deep venous system is empty in a patient
diagnosed with PE, then we treat to assure that
the venous system remains empty.
If there is clot in the deep venous system, then
we treat to assure that extension and
embolization of existing clot is minimized.
Thus, DVT and PE are the same disease.
Goals of treatment (cont’d)





How certain does one have to be that a
patient has DVT/PE before undertaking
treatment?
90%?
60%?
30%?
10%?
The notion of thresholds

There are four decision nodes for treating any
disease.




What happens if I DON’T treat someone with the
disease?
What happens if I DO treat someone with the
disease?
What happens if I DO treat someone who does not
have the disease?
What happens if I DON’T treat someone who does
not have the disease?
What happens if I don’t treat someone
with the disease?




10% die. Quality adjustment 0, value = 0
30% complication (SOB, post-phlebitic
leg,etc) Quality adj. 0.9, value = 0.27
60% alive and well. QA 1.0, value = 0.6
Total quality adjusted survival 0.87 for this
decision
What happens if I treat someone with
the disease?





Dead 0.5%, Quality adjst 0, value = 0
Complications 12% (bleeds, post-phlebitic
leg, SOB, etc.) QA 0.9, value = 0.11
Alive and well 87.5%, QA 1, value = 0.875
Total survival value 0.986 for this decision.
NET benefit of treatment 0.986 – 0.87 =
0.116
What happens if I treat someone who
does not have the disease?




Dead 0.05%, QA 0, value = 0
Complications 5% (minor, major bleeds,
etc) QA 0.9, value = 0.045
Alive and well, 94.9% QA 1, value 0.949
Total quality adjusted survival value of this
decision = 0.994
What happens if I don’t’ treat someone
and they don’t have the disease?






This is the perfect decision
Dead 0%, QA 0, value = 0
Complications 0%, QA 0.9, value = 0
Alive and well, 100%, QA 1.0, value = 1
Total value of this decision = 1.00
Thus, harm of treating patients who don’t
have the disease 1.00 – 0.994 = 0.006
Finally, the threshold

Threshold defined as ratio of harm to benefit.
Total harm = 0.006
 Total benefit = 0.116
 Harm/benefit (this is an odds) = 0.006/0.116
 Harm/benefit (probability) = 0.006/0.122 =
0.05 or 5%.
 This means if the patient that I am caring for
has a greater than 5% chance of DVT/PE
when all is said and done, then I treat!

Now, understanding the tests

Why do we test?



To increase or diminish the probability of disease
In many cases we test sequentially, but this
demands that the tests perform independently an unproven assumption in most cases.
Nevertheless, we can probaby come close to the
truth with such an assumption in DVT/PE since
the tests are unrelated physiologically.
The tests

Multiple slice CT (corrected for indeterminate
scans)


V/Q scan


High prob LR 17.8, Indeterminate 1.0, Normal or near
normal 0.10
D-dimer (high sens and depends on cut-off)


+LR 7.8 -LR 0.2
+ LR 2.4 -LR 0.10 (only useful in outpatients!!!)
Compression US (symptomatic pts) +LR 24 –LR
0.06
Using the tests



The purpose of testing is to get above or
below the treatment threshold
This requires knowing the test
characteristics as in the two previous slides
and:
Estimating pretest probability of PE/DVT.
Well’s Rule for Pretest probability of PE,
applied when the clinical complaint may be PE.








Previous DVT
1.5 pts
Recent hosp/leg trauma
1.5 pts
Pulse > 100
1.5 pts
Clinical evidence of DVT
3 pts
No other reasonable Dx
3 pts
Malignancy
1 pt
Hemoptysis
1 pt
For a score less <1 Prob <5%, 2-6 30%, 7+ 70%.
Well’s rules for Pretest probability of DVT,
when presenting complaint may be DVT. *May not
be fully valid in primary care.

Recent casting/hemiparesis
Recent surgery/hospitalization
Malignancy
Calf circumference > 3cm diff
Swelling of leg by measurement
Unilateral edema
Past hx of DVT
Unilateral swelling superficial veins
Tenderness along the venous system
Another equally likely dx

0 or less 5% prob of DVT, 1-2 25%, 3+ 50%.









1 pt
1 pt
1 pt
1 pt
1 pt
1 pt
1 pt
1 pt
1 pt
-2 pts
Using the tests.




SJ 32 yo healthy woman presents to clinic
with pleuritic chest pain, no fever, cough.
No trauma. Exam is normal, HR 97. CXR
wnl.
Pretest Prob? (score?)
~30% Prob (no other dx)
What test to order?
Patient SJ – Can we decrease the
Probability?


Pretest Prob 30%, odds 3/7
Neg D-dimer 3/7 x 0.10 = 0.30/7 odds


V/Q scan (normal or near normal) 3/7 x 0.10 =
0.3/7 odds


Post test prob = 0.3/7.3 = 0.04 ~ 4%
Multislice CT 3/7 x 0.2 = 0.6/7 odds


Post test prob = 0.3/7.3 = 0.048 ~5%
Post test prob = 0.6/7.6 = 0.078 ~ 8%
What would you order?
Pt SJ – Can we increase the
probability?

D-dimer 3/7 x 2.4 = 7.2/7odds


V/Q scan (high prob) 3/7 x 17.8 = 53.4/7
odds


Post test prob 7.2/14.2 = 0.5 ~50%
Post test prob 53.4/60.4 = 0.88 = ~ 88%
Multislice CT 3/7 x 7.8 = 23.4/7 odds

Post test prob 23.4/30.4 = 0.77 ~77%
Unanswered questions

Can patient with high probability of PE be
left untreated if deep venous system is, and
remains, empty by US?



This has been verified as a viable strategy for
those at intermediate probabilities.
If so this changes strategy dramatically.
Then approach to patients with PE would
be to verify empty proximal veins.
Patient RJ


56 yo man, recent MVA, frx of left tibia,
casted, also in hospital for 2 wks for
ruptured spleen and kidney. Develops
SOB, fever, HR 115. CXR clear, U/A wnl,
no evidence for abdominal infection. Left
leg swollen. Rt calf 31 cm left 34 cm
Pretest prob? Score 4, 60% probability
Testing RJ



Pretest prob is so high, that our goal is to try to
reduce probability to level not requiring treatment
(5%). If we cannot, then we treat.
D-dimer (not useful in inpatients)
V/Q (normal/near normal) 7/3 x 0.1 = 0.7/3 odds


Multislice CT 7/3 x 0.2 = 1.4/3 odds


Post test prob 0.7/3.7 = 0.19 = ~19%
Post test prob 1.4/4.4 = 0.32 = ~32%
Now what?
Thinking further about RJ


Pretest probability for DVT for RJ, Wells score is
4, ~50% prob of DVT.
Duplex US positive


Duplex US negative


1/1 x 25 = 25/1 odds of DVT or post test prob of 25/26 =
0.97 or 97% prob > Treat
1/1 x 0.04 = 0.04/1 odds of DVT or post test prob of
0.04/1.04 = 0.038 or ~4%. > No treatment.
Repeat US in 1-2 weeks
Further thinking about RJ

This is based on empirical trials of Hull and
Hirsch that show for intermediate
probability PE if the deep venous system is
empty, the risk for PE (even if patient may
have had one) is less than 2%.
What have we learned?



Possibly nothing
Make a reasonable estimate of pretest
probability of disease.
Then based on your knowledge of the 4
available tests, try to obtain a post test
probability less than 5%. If you can, do not
treat. If you cannot, then treat.
One final point

The threshold will change depending on
treatment risks.


For example, as the risks of bleeding or
complications of therapy increase, the threshold
will rise.
This means, if a 90 yr old man with a hx of GI
bleeding 5 yrs ago is considered, the threshold
may go up to 12% or more.
One final, final point


Filters, removable or not are not particularly
effective or durable treatments for
preventing PE.
If the filter cannot be removed, the risk for
long term complications such as DVT of
lower extremities is about 30-35% and after
9mos – 1 year, any protection afforded by
the filter wanes.