Diagnosing Pulmonary Embolism in Hospitalized Patients

 

Diagnosing Pulmonary Embolism in Hospitalized Patients: A Comprehensive Approach for the Internist

Dr Neeraj Manikath , claude.ai

Abstract

Pulmonary embolism (PE) remains a diagnostic challenge in hospitalized patients, where traditional risk stratification tools and diagnostic algorithms may not apply as reliably as in ambulatory settings. The atypical presentations, competing diagnoses, and altered pretest probability in the inpatient milieu demand a nuanced approach. This review synthesizes current evidence on PE diagnosis in hospitalized patients, offering practical strategies, clinical pearls, and diagnostic pitfalls to guide internists in this complex clinical scenario.

Introduction

Pulmonary embolism accounts for approximately 100,000 deaths annually in the United States, with many cases occurring in hospitalized patients. The in-hospital incidence of symptomatic PE ranges from 0.4% to 1% despite thromboprophylaxis, rising substantially in high-risk populations such as critically ill patients, those undergoing major surgery, and patients with cancer. The diagnosis becomes particularly challenging when PE develops during hospitalization, as symptoms overlap with the primary reason for admission, prophylactic anticoagulation may create a false sense of security, and the baseline cardiopulmonary status is often abnormal.

The Unique Challenge of Inpatient PE Diagnosis

Why Hospital-Acquired PE is Different

Pearl #1: Hospital-acquired PE has a significantly higher mortality rate (15-30%) compared to community-acquired PE (8-15%), partly due to diagnostic delays and the presence of comorbid conditions.

The Wells Score and Geneva Score, while validated for outpatient and emergency department settings, have reduced specificity in hospitalized patients. A systematic review by Roy et al. demonstrated that these tools may overestimate PE probability in inpatients, where alternative diagnoses are more prevalent and comorbidities confound clinical assessment.

Oyster: A patient with a Wells Score >4 doesn't automatically warrant CT pulmonary angiography (CTPA) in the inpatient setting. Consider the individual components: many hospitalized patients score points for immobilization, prior DVT/PE, malignancy, or tachycardia from unrelated causes.

Clinical Presentation: Beyond the Textbook Triad

The classic triad of dyspnea, chest pain, and hemoptysis occurs in fewer than 20% of hospitalized patients with PE. Recognition requires vigilance for subtle manifestations.

Atypical Presentations in Hospitalized Patients

1. Isolated tachycardia or tachypnea: Often attributed to pain, fever, or anxiety, unexplained persistent tachycardia (>100 bpm) or tachypnea (>20 breaths/min) warrants consideration of PE, particularly when out of proportion to other clinical findings.

2. Hypoxemia without dyspnea: Particularly in elderly or sedated patients who may not report breathlessness. An unexplained drop in SpO2 of ≥3% from baseline should trigger evaluation.

3. Syncope or presyncope: Suggests massive PE with right ventricular compromise. Studies show that 10-15% of PE patients present with syncope.

4. Unexplained delirium: In elderly hospitalized patients, PE can manifest as acute confusion without obvious cardiopulmonary symptoms.

5. Refractory hypotension: When a patient fails to respond to appropriate management of presumed sepsis or cardiogenic shock, consider PE in the differential.

Hack #1: The "Rule of 20s" – Be suspicious when hospitalized patients develop persistent tachycardia >100 bpm AND tachypnea >20/min without clear explanation. This combination has a sensitivity of approximately 90% for PE, though specificity is limited.

Risk Stratification in the Inpatient Context

Modified Approach to Pretest Probability

Pearl #2: The SOME score (Syncope, One-leg swelling, Mental status change, Elevated troponin) may better identify high-risk hospitalized patients than traditional scoring systems, though it requires further validation.

For hospitalized patients, consider a structured three-category assessment:

Low probability (<5%):

• Mobilizing regularly
• Adequate thromboprophylaxis
• No new respiratory symptoms
• Alternative diagnosis clearly explains presentation

Moderate probability (5-30%):

• Mixed risk factors
• Symptoms could represent PE or alternative diagnosis
• Most hospitalized patients fall into this category

High probability (>30%):

• Recent surgery (especially orthopedic, abdominal, or pelvic)
• Active cancer with chemotherapy
• Previous VTE
• Prolonged immobilization
• Known thrombophilia
• Multiple unexplained cardiopulmonary symptoms

Oyster: Don't assume thromboprophylaxis rules out PE. Studies show that 40-50% of hospital-acquired PEs occur despite appropriate prophylaxis. Prophylactic-dose anticoagulation reduces but does not eliminate risk.

D-Dimer: The Double-Edged Sword

In hospitalized patients, D-dimer loses much of its clinical utility due to poor specificity. Elevations occur with infection, inflammation, malignancy, surgery, pregnancy, renal disease, and advanced age—conditions prevalent in inpatients.

When D-Dimer Remains Useful

Hack #2: Use age-adjusted D-dimer cutoffs: Age × 10 μg/L for patients >50 years (e.g., 700 μg/L for a 70-year-old). This improves specificity while maintaining high sensitivity (>95%) in hospitalized elderly patients.

Pearl #3: A normal D-dimer (<500 μg/L or age-adjusted cutoff) combined with low clinical probability can safely exclude PE in most hospitalized patients, potentially avoiding unnecessary imaging and its attendant risks.

When to skip D-dimer entirely:

• High clinical suspicion (proceed directly to imaging)
• Recent surgery (<1 week)
• Active infection or sepsis
• Known disseminated malignancy
• Baseline D-dimer known to be chronically elevated

A meta-analysis by Righini et al. showed that the negative predictive value of D-dimer drops from 99% in outpatients to 85-90% in hospitalized patients, highlighting its limitations.

Imaging Strategies: Getting the Right Test at the Right Time

CT Pulmonary Angiography: The Gold Standard with Caveats

CTPA remains the primary diagnostic modality, with sensitivity of 83-96% and specificity of 89-96% for PE. However, several considerations apply specifically to hospitalized patients.

Pearl #4: Subsegmental PE detected on CTPA in patients without DVT may not require anticoagulation if cardiopulmonary reserve is adequate and risk factors are transient. The ELOPE trial demonstrated that withholding anticoagulation in selected subsegmental PE patients resulted in low recurrence rates (0.7% at 90 days).

Contraindications and concerns:

• Renal insufficiency (eGFR <30 mL/min)
• Contrast allergy
• Hemodynamic instability
• Pregnancy

Hack #3: For patients with borderline renal function (eGFR 30-45), consider sodium bicarbonate infusion (150 mEq/L in D5W at 3 mL/kg/hr for 1 hour before and 6 hours after contrast) plus N-acetylcysteine 1200 mg PO bid × 2 days to minimize contrast-induced nephropathy risk.

Alternative and Complementary Imaging

Ventilation-perfusion (V/Q) scanning:

• Consider when contrast is contraindicated
• Requires relatively normal chest X-ray for optimal interpretation
• PISAPED criteria improve diagnostic accuracy
• Results categorized as normal (rules out PE), low probability, intermediate, or high probability

Compression ultrasonography (CUS):

• Positive proximal DVT in the appropriate clinical context allows PE diagnosis without CTPA
• Useful when CTPA is contraindicated
• Sensitivity only 40-50% for PE (since not all PEs have detectable DVT)
• Negative CUS does not rule out PE

Echocardiography:

• Bedside tool for risk stratification rather than diagnosis
• McConnell's sign (RV free wall hypokinesis with apical sparing) has 77% specificity for acute PE
• RV dilation (RV:LV ratio >0.9) suggests hemodynamically significant PE
• Cannot diagnose or exclude PE but guides management in unstable patients

Hack #4: In hemodynamically unstable patients where transport for CTPA is unsafe, use the "POCUS protocol": Point-of-care cardiac ultrasound showing RV strain + DVT on CUS = presumed PE. Initiate treatment immediately.

Laboratory Adjuncts

Troponin and BNP

Cardiac biomarkers don't diagnose PE but provide prognostic information and identify patients requiring closer monitoring.

Pearl #5: Elevated troponin (any elevation) in PE patients indicates RV strain and is associated with 5-10 fold increased mortality. These patients warrant ICU-level monitoring even if initially hemodynamically stable.

BNP or NT-proBNP elevations similarly indicate RV dysfunction. The 2019 ESC guidelines incorporate troponin and RV function into risk stratification, classifying patients as low-risk, intermediate-low, intermediate-high, or high-risk.

Arterial Blood Gas

While nonspecific, ABG findings can support PE suspicion:

• A-a gradient >20 mmHg (expected gradient = [Age + 10]/4)
• Respiratory alkalosis (early finding)
• Hypoxemia resistant to supplemental oxygen
• Normal PaO2 and A-a gradient make PE less likely but don't exclude it (5-10% of PE patients have normal ABG)

Diagnostic Algorithms for Hospitalized Patients

A Practical Stepwise Approach

Step 1: Clinical assessment

• Document new or worsening cardiopulmonary symptoms
• Review risk factors (immobility, recent procedures, lines/catheters)
• Assess pretest probability

Step 2: Consider D-dimer (selective use)

• Use only if low-to-moderate pretest probability
• Apply age-adjusted cutoff
• Skip if recently post-operative, septic, or disseminated cancer

Step 3: Imaging decision

• Low probability + negative D-dimer → No further testing
• Moderate-high probability or positive D-dimer → CTPA (if no contraindications)
• CTPA contraindicated → V/Q scan or CUS for DVT
• Hemodynamically unstable → Bedside ultrasound or empiric treatment

Step 4: Risk stratification if PE confirmed

• Hemodynamic status
• Troponin, BNP
• Echocardiographic RV function
• Determines treatment intensity and monitoring level

Oyster: Don't order CTPA "just to be safe" in low-probability scenarios. Overutilization leads to detection of clinically insignificant emboli, unnecessary anticoagulation with bleeding risks, contrast nephropathy, radiation exposure, and incidental findings requiring further workup.

Special Populations

Cancer Patients

Hospitalized cancer patients have 4-7 fold increased PE risk. Khorana score helps stratify risk. High D-dimer threshold (>1000 μg/L) may improve specificity without sacrificing sensitivity in this population.

Pearl #6: In cancer patients with PE and no absolute contraindications, direct oral anticoagulants (apixaban, rivaroxaban) or low-molecular-weight heparin are preferred over warfarin. The SELECT-D and ADAM VTE trials demonstrated superiority in preventing recurrence.

Postoperative Patients

Highest risk period is days 1-3 and weeks 2-4 post-operatively. Tachycardia is common postoperatively; look for sustained elevation or other findings.

Hack #5: The "POST-PE rule": Post-operative patients with SpO2 <94% on room air, unexplained tachycardia, and Troponin elevation should undergo CTPA regardless of D-dimer, which is invariably elevated post-surgery.

Critically Ill Patients

PE diagnosis in ICU patients is particularly challenging due to mechanical ventilation, sedation, multiple organ dysfunction, and competing diagnoses.

Pearl #7: In mechanically ventilated patients, sudden increases in dead space ventilation (rising EtCO2 gradient) or worsening oxygenation with increased inspiratory pressures may indicate PE.

Diagnostic Pitfalls and How to Avoid Them

Pitfall #1: Anchoring bias Attributing symptoms solely to the known admission diagnosis. Combat this by maintaining PE on differential for unexplained deterioration.

Pitfall #2: False reassurance from prophylaxis Remember that prophylaxis reduces but doesn't eliminate risk. Up to 50% of hospital-acquired PEs occur despite prophylaxis.

Pitfall #3: Over-relying on "typical" symptoms Many hospitalized PE patients lack chest pain or hemoptysis. Isolated tachycardia/tachypnea may be the only clues.

Pitfall #4: Dismissing subsegmental PE While isolated subsegmental PEs may not always require treatment, they indicate hypercoagulability and increased future VTE risk. Ensure adequate prophylaxis going forward.

Pitfall #5: Delay in unstable patients Don't transport unstable patients for CTPA. Use bedside diagnostics and empiric anticoagulation if suspicion is high.

Treatment Considerations Post-Diagnosis

Once PE is diagnosed, treatment initiation should be swift. Unfractionated heparin provides reversibility in actively bleeding or peri-procedural patients. LMWH or fondaparinux offer convenience with once or twice-daily dosing. DOACs (apixaban, rivaroxaban) can be initiated immediately without heparin bridging and have lower major bleeding rates than warfarin.

Pearl #8: For massive PE (sustained hypotension, shock), systemic thrombolysis reduces mortality from ~30% to ~10%, but bleeding risk is substantial (10% major bleeding, 3% intracranial hemorrhage). Consider catheter-directed therapy in centers with expertise.

Conclusion

Diagnosing PE in hospitalized patients requires synthesizing clinical judgment, selective use of biomarkers, and appropriate imaging while avoiding diagnostic overuse. The key is maintaining appropriate suspicion—not too low (missing deadly PEs) nor too high (overtesting and overtreatment).

Remember the fundamental principle: PE diagnosis in hospitalized patients is as much about thoughtful clinical reasoning as it is about ordering tests. A systematic approach, awareness of atypical presentations, judicious test selection, and recognition of diagnostic pitfalls will serve internists well in navigating this common but complex clinical challenge.

Final Pearl: When in doubt, ask yourself: "Would my clinical decision change if I knew the CTPA result?" If the answer is no (because you'd anticoagulate regardless), proceed with treatment. If yes, proceed with diagnostic workup. This simple question prevents both over- and under-diagnosis.

References

1. Roy PM, Corsi DJ, Carrier M, et al. Net clinical benefit of hospitalization versus outpatient management of patients with acute pulmonary embolism. J Thromb Haemost. 2017;15(4):685-694.

2. Righini M, Van Es J, Den Exter PL, et al. Age-adjusted D-dimer cutoff levels to rule out pulmonary embolism: the ADJUST-PE study. JAMA. 2014;311(11):1117-1124.

3. Konstantinides SV, Meyer G, Becattini C, et al. 2019 ESC Guidelines for the diagnosis and management of acute pulmonary embolism. Eur Heart J. 2020;41(4):543-603.

4. den Exter PL, van Es J, Klok FA, et al. Risk profile and clinical outcome of symptomatic subsegmental acute pulmonary embolism. Blood. 2013;122(7):1144-1149.

5. Carrier M, Righini M, Wells PS, et al. Subsegmental pulmonary embolism diagnosed by computed tomography: incidence and clinical implications (ELOPE). BMJ. 2010;340:c1413.

6. Kucher N, Rossi E, De Rosa M, Goldhaber SZ. Massive pulmonary embolism. Circulation. 2006;113(4):577-582.

7. Khorana AA, Kuderer NM, Culakova E, et al. Development and validation of a predictive model for chemotherapy-associated thrombosis. Blood. 2008;111(10):4902-4907.

8. Young T, Tang H, Hughes R. Vena caval filters for the prevention of pulmonary embolism. Cochrane Database Syst Rev. 2010;(2):CD006212.

9. Raskob GE, van Es N, Segers A, et al. Edoxaban for venous thromboembolism in patients with cancer: results from a non-inferiority trial (SELECT-D). Haematologica. 2020;105(6):e284-e286.

10. McConnell MV, Solomon SD, Rayan ME, et al. Regional right ventricular dysfunction detected by echocardiography in acute pulmonary embolism. Am J Cardiol. 1996;78(4):469-473.