See also "Prophylaxis for PE" and "Deep Vein Thrombosis"

I. Risk factors for PE

  1. Recent surgery
  2. Recent fx of a lower extremity, especially with immobilization
  3. Immobilization, particularly complete bedrest or LE paralysis
  4. Previous DVT or PE
  5. Family history of DVT or PE
  6. Cancer
  7. Postpartum period
  8. Male gender
  9. Estrogen use

II. Risk factors for PE among DVT pts:

  1. See section on Deep Vein Thrombosis for risk factors
  2. The following are not independent risk factors for PE among DVT pts: Age, sex, location of thrombus, being post-op, varicose vv., having Ca, acute MI, acute CVA, estrogen use

III. Typical clinical features (ones with asterisks were sig. predictors of eventual diagnosis of venous thromboembolism in a prospective cohort study-Ann. Emerg. Med. 55:307, 2010-JW):

  1. Dyspnea*
  2. Tachypnea*
  3. Tachycardia
  4. Pleuritic chest pain*
  5. Chest pain that is not substernal in location*
  6. Arterial oxygen saturation < 92% on room air
  7. Low-grade fever (occasionally)
  8. Hemoptysis
  9. Pleural rub
  10. Clinical evidence of DVT

III. Clinical features of severe PE:

  1. Hypotension (from reduced left-heart venous return)
  2. Right heart failure
  3. Hypoxemia
  4. Chronic Pulmonary Hypertension is an occasional late complication of large PE's

IV. Diagnosis

  1. Pulmonary angiography is the gold standard but has 3-4% risk of cardiopulmonary complications and may not be more accurate than CT (see below)
  2. CXR may show evidence of pulmonary infarct
  3. Electrocardiography may yield some clues
  4. V-Q lung scan (limited specificity)
  5. Single-photon emission computed tomography ("SPECT") V/Q scintigraphy
    1. Uses less radiation than CT and doesn't require use of intravenous contrast material
    2. Similar accuracy to CT pulmonary angiography (Chest 136:1546, 2009-AFP)
  6. Lower extremity venous duplex (DVT requires same tx as PE)--of course, a negative study does not exclude PE!
  7. In a retrospective study of 62 pts  suspected of having PE, with  low-probability V-Q scans and  negative LE u/s studies, all of whom  underwent pulmonary angiography,  8% had PE on pulmonary angiography  (Chest 115:980, 1999--JW)
  8. Plasma D-dimer--See also D-Dimer for Diagnosis of DVT
    1. Using 500ug/l assessed w/enzyme immunoassay as cutoff, 99.5% sensitivity, 41.4% specificity (Am. J. Resp. Crit. Care Med. 156:492, 1997-JW)
    2. 100% sensitive and 35% specific using same cutoff in another study of 386 pts with suspected PE (Am. J. Resp. Crit. Care. Med. 158:65, 1998--JW)
    3. Sensitivity 85%, specificity 68% in another study of 1177 pts with suspected PE (though only 3% of pts with normal D-dimer and nondiagnostic VQ scans were eventually shown to have PE) (Ann. Int. Med. 129:1006, 1998--JW)
    4. In a series of 930 pts presenting to an ED with suspected PE, combination of low pretest probability (based on clinical criteria) and negative D-dimer was ass'd with a 3/437 incidence of ultimate diagnosis of PE (e.g. based on V-Q scan) (Ann. Int. Med. 135:98, 2001--JW)
    5. False-negatives much more likely in pts with subsegmental PE (sensitivity only 50% in one series--Am. J. Resp. Crit. Care. Med. 165:345, 2002--JW)
    6. False-positives may be ass'd with recent surgery, malignancy, and hepatic disease (Am. J. Resp. Crit. Care Med. 159:1445, 1999--JW)
    7. Combination of plasma D-dimer pus alveolar dead space measurement (airway dead space - physiologic dead space)
      1. In a prospective trial of 380 pts > 18yo with suspected PE, checking these measurements together and taking both being normal as a "negative" result was 98.4% sensitive and 51.6% specific for PE (determined by standardized testing protocol including VQ scanning or CT plus venous duplex and/or pulmonary angiography if indicated); neg. predictive value was 99.25% (JAMA 285:761, 2001--abst)
  9. Helical (a.k.a "spiral") CT
    1. Diagnostic accuracy--Published studies differ in their conclusions!
      1. Sensitivity/Specificity c/w pulmonary angiography of 94%/93%, sig. greater than V-Q scanning, in a series of 216 pts suspected of having PE (Am. J. Roent. 174:1041, 2000--AFP)
      2. In a series of 103 pts with suspected PE and nondiagnostic V-Q scans, negative predictive value of spiral CT was 93% (Am. J. Med. 110:16, 2001--JW)
      3. CT (w/contrast) of lungs and legs to r/o PE (the leg part to look for DVT)--In a study of 650 pts with suspected PE, this approach had 97% sensitivity and 100% specificity for LE DVT c/w doppler ultrasound; there was no clinical follow-up in this study (Radiol. 219:498, 2001--JW)
      4. In a series of 299 pts with suspected PE and abnormal D-Dimer assay, 39% of whom were ultimately dx'd with PE based on angiography, V-Q scan, or lower-limb u/s, Helical CT had sensitivity of only 70% for PE; specificity was 91% (Ann. Int. Med. 135:88, 2001--JW)
      5. In a retrospective study of 239 pts who had negative helical CT for suspected pulmonary embolus and who didn't receive anticoagulation, over the ensuing 3 months, incidence of venous thromboembolic events was 1.7%.  This is similar to incidence after negative pulmonary angiography or low-probability V-Q scan (Arch. Int. Med. 163:2033, 2003--abst)
      6. In a meta-analysis of 15 studies examining the diagnostic accuracy of CT for PE with at least 3mo f/u surveillance for venous thromboembolic events, the negative predictive value of CT for VTE over 3mos was 99.1% (no sig. diff. based on CT modality used, i.e. single-slice vs. multidetector-row).  This is similar to reported rates for pulmonary angiography (JAMA 293:2012, 2005--abst)
    2. It may be that helical CT is highly sensitive for central pulmonary emboli compared with angiography but may be less sensitive for distal emboli, but that the clinical significance of such emboli is negligible.
      1. In a series of 198 pts with suspected PE but negative helical CT scans, followed for 3mos w/o anticoagulation, 2 were subsequently diagnosed with PE; no deaths attributable to PE occurred (Radiol. 215:535, 2000--JW)
      2. In a clinical trial using the strategy of witholding anticoagulation from a series of about 383 pts with suspected PE but normal LE venous duplex and helical CT failing to confirm PE, over 3mo f/u, only 2 had persistence of sx that was confirmed on repeat imaging to be a DVT or PE (1 of each) (Ann. Int. Med. 138:307, 2003--JW)
  10. Multi-Row Detector CT
    1. In a study in which 756 patients with high clinical probabilities of PE by Geneva Score (Arch. Int. Med. 161:92, 2001) or (low-to-intermediate clinical probabilities of PE + positive plasma D-Dimer) underwent chest CT with 4-slice multidetector-row CT and leg ultrasonography. Sensitivity for PE with CT (compared with CT + leg ultrasound) was 96.4% (NEJM 352:1760, 2005--JW)
    2. Multidetector CT-Angiography + venous-phase multidetector CT venography had slightly higher sensitivity and equivalent specifity to multidetector CT angiography alone in a prospective study of 824 pts with suspected pulmonary embolus (NEJM 354:2317, 2006--JW)
  1. In a meta-analysis of 15 studies of sensitivity of contrast-enhanced CT (including single detector row helical CT, multiple detector row helical CT, and electron-beam CT) for predicting incidence of venous thromboembolic events if anticoagulation therapy was witheld for 3mos, a normal CT was associated with a negative likelihood ratio of 0.07 (JAMA 293:2012, 2005--AFP)
  2. MR angiography--Ass'd with only 77% sensitivity c/w angiography in one study of 118 pts with suspected PE undergoing both provedures (Lancet 359>1643, 2002--JW)
  3. ABG's
  1. Traditionally used with assumption that elevated A-a gradient will result from PE
  2. Study from "PIOPED" group (Prospective Investigation of Pulmonary Embolism Diagnosis) looked at this
  1. 12% of pts with angiographically confirmed PE had nl A-a gradient
  2. In series of 768 pts with suspected PE who underwent angiography to confirm (Chest 109:78, 1996-JW)
  1. Looked at combination of: nl A-a gradient (<20), nl PaO2 (>80), and nl PaCO2 (>35)
  2. 7% of pts with PE had all these; 17% of pts without PE had them
  3. So, 29% of all pts with all 3 nl still had PE!
  1. Decision Rules
    1. Wells Clinical Prediction Rule for PE (click link to see)

V. Treatment

  1. Anticoagulation
  1. General approach:
    1. Heparin (either traditional, adjusted-dose continuous IV infusion or intermittently-dosed SQ low molecular-weight heparin) + immediate initiation of oral warfarin, with heparin discontinued when the INR is therapeutic (2.0-3.0)
    2. Continuation of the warfarin for 3-6mos
  2. Duration of oral anticoagulation
    1. In a 3y randomized trial of 326 pts with h/o PE who had received oral warfarin for 3mos without recurrence of bleeding, continued treatment with oral warfarin for an extra 3mos (or 9mos, for those who had PE w/o identifiable risk factors) vs. placebo) was ass'd with RR 0.81 (3.1%/pt-yr vs. 4.1%/pt-yr; nonsig.) for recurrent symptomatic venous thromboembolism (Ann. Int. Med. 139:19, 2003--abst)
  3. Low molecular weight heparin 
    1. 612 pts w/non-massive PE randomized to tinzaparin SQ vs. unfractionated heparin IV followed by PO anticoagulation; f/u x 3mos showed no sig. diff. in rates of death, recurrent venous thromboembolism, or major bleeding (NEJM 336:663, 1997-JW) 
    2. 1021 pts w/venous thromboembolism (73% w/DVT only; 27% w/PE) randomized to revirapin vs. unfractionated heparin; f/u at 3mos showed no sig. diff. in rates of death, recurrent venous thromboembolism, and major bleeding (NEJM 337:657, 1997-JW)
    3. In a meta-analysis of 12 randomized trials comparing fixed-dose, subcutaneous LMWH with dose-adjusted, intravenous unfractionated heparin as initial treatment for PE, there was no sig. diff in 3mo recurrence of venous thromboembolism, major bleeding, or all-cause mortality (Ann. Int. Med. 140:175, 2003--JW)
    4. In a study in 505 adults with DVT or PE randomized to dalteparin vs. tinzaparin followed by 3mos of oral warfarin, there was no sig. diff. in incidence of recurrent thromboembolism, major hemorrhage, or both (Arch. Int. Med. 165:733, 2005--JW)
    1. Non-warfarin anticoagulants
      1. In a study in about 5,000 pts with acute symptomatic pulmonary embolism randomized to (rivaroxaban 15mg PO BID x 3wks then 20mg PO QD) vs. (enoxaparin + (warfarin or acenocoumarol)) x 3-12mos (duration at discretion of treating physician), there was no sig. diff. in incidence of symptomatic recurrent VTE or minor bleeding; ribaroxaban pts had sig. lower incidence of major bleeding (1.1% vs. 2.2%).  Note-Exclusion criteria included: criteria CrCl <30 mL/min,  significant hepatic disease, severe HTN, or need for thrombolysis (NEJM 3/26/2012, e-publication ahead of printing at: http://dx.doi.org/10.1056/NEJMoa1113572).
    2. Synthetic Heparin Analogues
      1. 2213 pts with acute PE randomized to fondaparinux (5-10mg SQ QD adjusted by weight) vs. adjusted-dose IV unfractionated heparin (continued x 5d or until warfarin resulted in INR > 2.0, whichever came later).  3mo incidence of recurent venous thromboembolic events (PE or DVT) was nonsig. lower in fondiparinux group (3.8% vs. 5%).  No sig. diff. in 3mo incidence of major bleeding or all-cause mortality (NEJM 349:1695, 2003--abst)
  1. Thrombolytics
    1.  Generally used for "massive" PE, i.e. patient hypotensive or thrombi present in R heart or main PA's
    2. Use in "submassive" PE
      1. In a randomized trial of 256 pts with PE and RV dysfunction or pulmonary hypertension (on echo or CVG) but no systemic hypotension randomized to Alteplase vs. placebo (all pts also received unfractionated IV heparin); incidence of (in-hospital death or clinical deterioration requiring escalation of treatment, e.g. "rescue" thrombolysis) was sig. lower in Alteplase group (11% vs. 25%); no sig. diffs in death, recurrent PE, or major bleeding (NEJM 347:1143, 2002--JW
    3. In a meta-analysis of 11 randomized trials of thrombolytics vs. heparin for pulmonary embolus, thrombolysis was not ass'd with sig. reduction of (recurrent PE or death), though a sig. reduction was seen in the subset of trials looking only at treatment of major hemodynamically unstable PE (9.4% vs. 19.0) (Circ. 110:744, 2004--AFP)
  1. "Rheolytic thrombectomy catheter" is a Water-Pik like device that has been used to break up thrombi in the pulmonary aa. (Circ. 96:2498, 1997--JW)
  2. Home vs. in-hospital care
    1. Low-risk patients with PE may be treated at home
    2. "Prognosis in Pulmonary Embolism" (PREP)-Score 0-7 indicates 3% 30-day mortality and 4% 90-day mortality with 99% negative predictive value
      1. Altered mental status (disorientation, stupor, coma): 10 points
      2. Cardiogenic shock (SBP < 90 mm Hg): 6 points
      3. Cancer: 6 points
(Sources include Core Content Review of Family Medicine, 2012)