This preview shows page 1. Sign up to view the full content.
Unformatted text preview: Nearly all men can Every gun that is made, Every gun that is made, withstand adversity; if every warship you want to test a man's launched, every character, give him rocket fired, signifies power. … a theft from those Abraham Lincoln
who hunger and are source unknown
not fed, from those who are cold and are Never do anything not clothed.
against conscience even if the state Dwight D. Eisenhower
demands it. April 16, 1953
Albert Einstein quoted in obituary Of Dimers and Spirals
Of Dimers and Spirals
Joe Lex, MD, FAAEM
Temple University Hospital
July 29, 2003 The Trouble with PE
The Trouble with PE
Most common preventable cause of hospital death Most common undiagnosed cause of hospital death
Most cases recognized by the true expert – the pathologist
(70 – 80% of PE is diagnosed postmortem) The Trouble with PE
The Trouble with PE
650,000 cases / year in USA
200,000 deaths
Incidence unchanged for 3 decades
Frequency in ED: ? ? ?
21% of ED patients with pleuritic chest pain have PE by angiography The Trouble with PE
The Trouble with PE What We’ve Tried Before
What We’ve Tried Before
ABG
Aa gradient
Chest xray
EKG
FSP / FDP
Leg ultrasound
Venous plethysmography V/Q scan
Chest US
TTE
TEE
MRI
MRA
Alveolar dead space www.emedhome.com/archivesdata.cfm?ID=news032502&Type=news High Probability V/Q Scan
High Probability V/Q Scan This means PE, right?? High Probability V/Q Scan
High Probability V/Q Scan
NOT unless pretest probability is moderate or high
Using angiography as Gold Standard Low pretest probability + high probability V/Q scan = 45 – 66% false positive
Wells P et al. Clin Chest Med. 2003 (Mar);24(1):13 Pulmonary Angiography
Pulmonary Angiography
“Gold Standard” not used
Fear of mortality – Recent study with n=1400 showed 0 deaths, 4 major complications (0.3%) Limited after hours availability
Special expertise required Pulmonary Angiography
Pulmonary Angiography
“Gold Standard” not used
Costs a lot of money
Side effects: arrhythmias, hypotension, contrast reaction
Not 100% sensitive: 1.6% VTE rate at one year followup (CI 0.3 – 6.0%) Current Expert Opinion
Current Expert Opinion
Treat for PE when probability of PE >80%
Safely NOT treat for PE when probability ≤ 2%
MUST use Bayesian thinking (more on this later) The New Kids
The New Kids
Ddimer Latex agglutination
Standard ELISA
Rapid ELISA
Erythrocyte agglutination
– Turbidimetric
– Immunofiltration
–
–
–
– Helical (Spiral) Computerized Tomographic Pulmonary Angiography (HCTPA)
–
–
– Singlebeam
Multiple beam
Electron beam DDimers
DDimers
Dimer = 2 identical monomers
Fibrin breaks into D, E, X, Y (“fibrin split products”)
Ddimer = 2 D fragments joined by covalent bond
T½ 4 to 6 hours
Measurable ~1 week
Part of MOCHAJAVA MOCHA – JAVA
MOCHA – JAVA
Markers of Cascade and Hemostasis Activation, Juxtaposed Against Vascular Activity Bottom RightHand Corner
Bottom RightHand Corner
Coagulation Pathway Lysis Pathway Fibrin Splits
Fibrin Splits Some Essential Statistics
Some SpPin: with high Specificity, a Positive result tends to rule in
SnNout: with high Sensitivity, a Negative result tends to rule out Bayesian Analysis
Bayesian Analysis
An objective mathematical process in which the estimated pretest probability of a disease or the strength of a clinician's suspicion of disease is altered by the result of a diagnostic test to create a posttest probability for that particular disease Rt. Rev. Bayes
Rt. Rev. Bayes
Example
Sn=95%
Sp=50% PTP=20% Some Essential Statistics
Some Pretest Probability (PTP): from clinical model or gestalt
Pretest odds = PTP ÷ (1 – PTP) Positive LR = SN ÷ (1 – SP)
Negative LR = (1 – SN) ÷ SP
LR x pretest odds = posttest odds
Posttest odds ÷ (1 + posttest odds) = posttest probability Some Essential Statistics
Some Example:
SN=95% SP=50% PTP=20%
LR+ = 0.95 ÷ (1 – 0.5) = 1.9
LR = (1 – 0.95) ÷ 0.5 = 0.1
Pretest odds = 0.2 ÷ (1 – 0.2) = 0.25 = 25% SN=95% SP=50% PTP=20%
SN=95% SP=50% PTP=20%
LR+=1.9 LR=0.1 PTO=25%
Post test odds with positive test: 0.25 x 1.9 = 0.475 Posttest probability with positive test: 0.475 ÷ 1.475 = 32%
Thus we’ve gone from a pretest probability of 20% to a posttest probability of 32%
This doesn’t help very much SN=95% SP=50% PTP=20%
SN=95% SP=50% PTP=20%
LR+=1.9 LR=0.1 PTO=25%
Post test odds with negative test: 0.25 x 0.1 = 0.025
Post test probability with negative test: 0.025 ÷ 1.025 = 0.024
So a negative test in a patient with a 20% pretest probability leaves a 2.4% posttest probability (i.e. ~1 in 40 negatives is false) Or…
Or…
Pull out your PDA
Find MedCalc
Scroll down to “Posttest probability (sens/spec)” and you’re in business
If you don’t have it but you’ve got your PDA, see me after the talk and I’ll beam it to you In an ideal world…
In an ideal world…
…we would have PTP for all patients with suspected disease
…we would have LRs for each test
…low PTP plus a negative test would safely rule out
…high PTP plus a positive test would accurately rule in Summary
Summary
If the probability of disease is low, a negative test is likely to rule out the disease BUT a positive test does not rule it in
If the probability of disease is high, a negative test does not rule out the disease BUT a positive test is likely to rule it in Decision Rules
Decision Rules
Miniati et al (a.k.a. PISAPED): ‘low’ group, prevalence 10%
Am J Respir Crit Care Med 1996;154:13871393 Wicki et al (a.k.a. Geneva Rules): lots of rules; hard to memorize; external validation?
Arch Intern Med 2001;161:9297 Decision Rules
Decision Rules
Wells et al.
3 points if PE most likely
3 points if signs/symptoms DVT
1.5 points for immobilization or surgery, HR >100, prior VTE
1 point for cancer, hemoptysis
0 – 1: low prevalence
Clin Chest Med;24(1):March, 2003 Wells Criteria
Wells Criteria
Advantage
No need for ABG
Demonstrated reliability and accuracy
Tested in ED patients Disadvantage
Hard to memorize
Must consider alternative diagnosis and decide whether more likely than PE Clin Chest Med;24(1):March, 2003 Wicki vs. Wells
Wicki vs. Wells
When Wells model is compared to Wicki model using a fixed set of patients, there is a 30% chance that a patient labeled "low probability" by one model would be labeled "moderate probability" by the other model (and vice versa)
Chagnon I. Am J Med. 2002; 113(4): 26975 Kline’s Decision Rules
Kline’s Decision Rules
Kline et al.
If age ≤ 50 years
If heart rate ≤ systolic BP
Eligible for Ddimer testing Kline et al. Ann Emerg Med 2002;39:144152 Kline’s Decision Rules
Kline’s Decision Rules
If age >50 or SBP > HR, ask:
1. Hypoxemia? (SaO2 <95%) 2. Unilateral leg swelling?
3. General anesthesia within 4 weeks?
4. Hemoptysis?
If “No” to all, then PTP = 13.3% and negative Ddimer <1% New Five ‘n Dimers
New Five ‘n Dimers
Latex agglutination
ELISA / Rapid ELISA
Erythrocyte agglutination
Turbidimetric
Immunofiltration (“silk screen”) Latex Agglutination
Latex Agglutination
INACCURATE TEST: DON’T USE IT TO RULE IN OR OUT
15/30 patients with PE proven by angiography had a negative (Accuclot®) study Kutinsky et al: Arch Int Med 1999;159:15691572 Kline et al. Ann Emerg Med 2000;35:168180 ELISA
ELISA Enzyme Linked ImmunoSorbent Assay
*Metaanalysis n=2,069
Sn=97% Sp=42% LR = 0.07
If Pretest probability= 10%
Posttest probability= 0.8%
BUT it takes at least 4 hours ELISA by Bayes
ELISA by Bayes
Sn=97%
Sp=42% PTP=10% Screening: use high sensitivity… …but you’ll do a lot more imaging Kline et al. Ann Emerg Med 2000;35:168180 Rapid ELISA
Rapid ELISA Vidas DDimer, Instant IA
Sn=93% Sp=30% LR = 0.23
If pretest probability = 10%
Posttest probability = 2.3%
Missing one in 40 Rapid ELISA by Bayes
Rapid ELISA by Bayes
Sn=93%
Sp=30% PTP=10% Erythrocyte Agglutination
Erythrocyte Agglutination
SimpliRED® (n = 2,297)
Sn=95% Sp=35% LR = 0.14
If pretest probability = 10%
Posttest probability = 1.4%
Miss ~1 in 70 (i.e. <2%) Turbidimetric DDimer
Turbidimetric DDimer
Turbiquant, Liatest (n = 654)
Sn=98% Sp=43% LR = 0.046
If pretest probability = 10%
Posttest probability = 0.4%
About as good as we can get, but you’ll do a lot of imaging based on falsenegative studies Highest Sensitivity
Highest Sensitivity
Disadvantage: fewer negatives (i.e. more false positives) more imaging
If possible, use high specificity combined with clinical suspicion
Ddimers NOT specific by nature 1000 Suspected PE Patients
1000 Suspected PE Patients
Prevalence of PE = 10%
TP
FP
TN
FN Rapid ELISA
Sn = 97%
Sp = 42%
97
522 ! !
378
3 !!! BUT… 1000 Suspected PE Patients
1000 Suspected PE Patients
Each imaging study also has limited sensitivity, specificity, predictive values, etc.
At each step you will miss some patients with true VTE / PE Conclusion on DDimers
Conclusion on DDimers
IF your patient has low pretest probability for venous thromboembolic disease, and…
IF you use an ELISA, rapid ELISA, turbidimetric, or erythrocyte agglutination Ddimer test… Conclusion on DDimers
Conclusion on DDimers
…THEN you can drive your false negative rate to below 2% and safely rule out pulmonary embolism
PTP + DDimer with small LR Conclusion on DDimers
Conclusion on DDimers
IF pretest probability is high, then NO Ddimer can safely rule out VTE
Ddimer is NOT a “screening test.” It is a diagnostic test to “Rule out” in appropriate patients What About HCTPA?
What About HCTPA?
Helical (Spiral) Computerized Tomographic Pulmonary Angiography “Slip ring” technology vs. “step and shoot”
3rd, 4th, and 5th generations
Single vs. multiple beams HCTPA
HCTPA Technique
Technique
Contrast injection ≥ 3 ml/sec
Start at diaphragm, work up
15 – 30 seconds total
1 or 2 breath holds for patient
Single beam slices 2 – 3 mm
Multibeam scanners: 1.25 mm Primary Findings
Primary Findings
MOST IMPORTANT FINDING: intraluminal filling defect (“doughnut” sign) cross
section Primary Findings
Primary Findings
MOST IMPORTANT FINDING: intraluminal filling defect (“railroad track” sign) longitudinal Secondary Findings
Secondary Findings Normal enhanced lobar artery Secondary Findings
Secondary Findings
Next inferior slice: same artery completely unopacified Secondary Findings
Secondary Findings Eccentric filling defect Secondary Findings
Secondary Findings
Wedgeshaped opacities representing pulmonary infarcts Secondary Findings
Secondary Findings
Oligemia: decreased attenuation in periphery, small vessels Accuracy
Accuracy
“100% sensitive and 100% specific” for large central clots
Overall sensitivity ~90% (when compared to angiography)
Lower sensitivity for smaller (subsegmental) clots Accuracy
Accuracy
In several studies, HCTPA done after indeterminate V/Q preselected to miss small clots?
Up to 2/3 of HCTPA can lead to or support an alternative diagnosis The “Big 4” Questions
The “Big 4” Questions
1. What is the frequency of isolated subsegmental clots?
2. How well does HCTPA find them?
3. Are they clinically significant?
4. Can a patient be managed safely based on negative HCTPA alone? Subsegmental Frequency?
Subsegmental Frequency?
PIOPED showed 6%
Other studies: up to 30%
ALL PRESELECTED – had earlier normal studies
In unselected patients: low Angiography Accuracy?
Angiography Accuracy?
Pigs: injected w/simulated clot
Pulmonary angiography vs. 1mm HCPTA vs. 3mm HCTPA
No difference in sensitivity
No difference in positive predictive value Small Clots Significant?
Small Clots Significant?
627 patients with nondiagnostic V/Q and negative leg ultrasound (preselected for small clots)
No treatment
Longterm followup: 6 patients developed significant PE (0.9%) Small Clots Significant?
Small Clots Significant?
PIOPED: 20 patients with known PE not treated
No difference in outcome when grouped according to size of vessel involved Safe If HCTPA Negative?
Safe If HCTPA Negative?
Small studies using singlebeam scanners show NPV ≥ 97%
Multibeam scanners now becoming more common
PIOPED II ongoing answer in a few years Safe If HCTPA Negative?
Safe If HCTPA Negative?
760 patients
358 low PTP, negative Ddimer STOP! Normal at 3 months
402 had HCTPA & leg US
332 had normal studies, all normal at followup
Prevalence: 70/760
Swensen SJ et al. Mayo Clin Proc. 2002 Feb;77(2):1308 www.homestead.com/emguidemaps/files/PErationalthinking.html British Thoracic Society
British Thoracic Society
Clinical
Assess and document clinical probability on everyone
Consider alternative clinical explanations at presentation and when PE excluded
No recommendation as to which decision rule for probability
Thorax 2003;58:470484 Ddimer
Ddimer
Low probability patient: negative SimpliRED®, Vidas, or MDA reliably excludes PE
Intermediate probability patient: negative Vidas or MDA reliably excludes PE
No imaging needed in either group
Thorax 2003;58:470484 Ddimer
Ddimer
Should not be performed on patients with high clinical probability
Each hospital should provide information on sensitivity and specificity of its Ddimer test Thorax 2003;58:470484 Imaging
Imaging
HCTPA is recommended initial lung imaging modality
Patient with good quality HCTPA does not require further investigation or treatment for pulmonary embolism Thorax 2003;58:470484 Isotope lung scan if…
Isotope lung scan 1.
2.
3. Facility is on site
Chest xray is normal
No significant concurrent cardiopulmonary disease
4. Use standardized reporting
5. Nondiagnostic study followed by HCTPA
Thorax 2003;58:470484 Cost Effectiveness
Cost Effectiveness
No study has yet performed all the relevant comparisons
HCTPA replacing V/Q scan at 2/3 the cost
Ddimers cost $20 $50 each
Algorithms will save money safely Algorithms Work!!
Algorithms Work!!
Prealgorithm: 55% of patients with “abnormal” perfusion scans got heparin before final diagnosis
With algorithm: 13% Berghout et al. QJM. 2000;93:33540 Conclusion
Conclusion
Clinical probability, a high sensitivity Ddimer, and HCTPA decrease the cost of diagnosis, improve safety, and increase convenience for both physicians and patients ...
View Full
Document
 Fall '11
 Dr.Aslam

Click to edit the document details