A Brief Historical Perspective on Pulmonary Embolism

Abstract

Pulmonary embolism is a major cause of mortality worldwide. In this historical perspective, we aim to provide an overview of the rich medical history surrounding pulmonary embolism. We highlight Virchow's first steps toward understanding the pathophysiology in the 1800s. We see how those insights inspired early attempts at intervention such as surgical pulmonary embolectomy and caval ligation. Those early interventions were refined and ultimately led to the development of inferior vena cava filters, the earliest clinical applications of anticoagulation, and even apparently disparate medical advances such as the successful development of cardiopulmonary bypass. We also see how the diagnosis of pulmonary embolism has evolved from rudimentary monitoring of vitals and symptoms to the development of evermore sophisticated tests such as contrast tomography angiography and echocardiography. Finally, we discuss current approaches to diagnosis, classification, and myriad treatments including anticoagulation, thrombolysis, catheter-directed interventions, surgical embolectomy, and extracorporeal membrane oxygenation guided by Pulmonary Embolism Response Teams.

Pulmonary embolism remains a major cause of morbidity and mortality worldwide. Indeed, they were first described by medical science just 200 years ago, and the first meaningful insights into the pathophysiology were elucidated in the middle of the 19th century by Virchow. ^{1 2} Since then, we have made significant strides in diagnosis and management. Modern diagnosis and management rely on clinical assessment and multiple laboratory, imaging, and treatment modalities. Anticoagulation is the most common treatment. However, more severe cases may require a combination of thrombolytics, various catheter-directed/percutaneous methods, surgery, and mechanical circulatory support. How did our modern armamentarium emerge from Virchow's recognition of pulmonary embolism risk factors and their origins in peripheral veins? Are there lessons from this history that may inform future innovations in the management of pulmonary embolism?

Early Insights into Pulmonary Embolism Pathophysiology

The first description of pulmonary embolism is attributed to René-Théophile-Hyacinthe Laennec, inventor of the stethoscope, in his 1819 treatise De l'auscultation medicale. ¹ The French pathologist Cruveilhier also recognized and reported blood clots in the pulmonary arteries in the early 19th century. ² These early findings were major advances in the field, but the etiology of pulmonary emboli was not understood and they were assumed to arise primarily in the pulmonary arteries.

Virchow provided the first pathophysiologic understanding of pulmonary embolism in the 1850s. ³ Virchow recognized that pulmonary emboli may arise in one location (e.g., the vena cava), break free, and “embolize” to cause “secondary occlusion in remote vessels (i.e., the pulmonary arteries). Perhaps even more importantly, Virchow recognized the key factors predisposing patients to the development of pulmonary embolism as described in “Virchow's triad”: (1) stasis of blood, (2) venous damage, and (3) a hypercoagulable state. It is difficult to overstate the importance of Virchow's contribution to the field. His eponymous triad remains a key conceptual framework in medicine to this day. With this early foundational understanding of the risk factors and pathophysiology of pulmonary embolism, the stage was set for development of interventions.

Surgical Pulmonary Embolectomy Comes Full Circle

Surgical embolectomy was probably the earliest intervention with a pathophysiologic rationale proposed to treat pulmonary embolism. Trendelenburg is credited with the concept. He had studied cases of sudden death from pulmonary embolism as a clinician. After performing experimental studies in calves, he developed the surgical embolectomy procedure in the 1870s. His approach included a thoracotomy followed by exposing the embolus by elevating the pulmonary artery with a rubber tube passed behind the ascending aorta and pulmonary artery in the transverse sinus. He performed this operation in two patients. Neither patient survived. ⁴ In 1924, Kirschner, a student of Trendelenburg, performed the first successful pulmonary embolectomy. ⁵ Over the next few decades, there were hundreds of attempts at the procedure. Survivors of the procedure were rare. ⁶

In 1932, John Gibbon witnessed a desperate attempt at open embolectomy while he was a research assistant at Massachusetts General Hospital. ⁷ In those times, patients suspected of having acute pulmonary embolism were monitored closely until cardiopulmonary collapse was felt to be imminent. They were then taken for surgical embolectomy, “Trendelenburg's operation.” Like many others, the patient Gibbon observed did not survive. However, the experience is thought to have inspired his research that led to the successful development of the heart/lung machine. In 1953, Gibbon performed the first successful operation, atrioseptal defect closure, on cardiopulmonary bypass. ⁸ Later, Sharp performed the first successful pulmonary embolectomy on cardiopulmonary bypass in 1962. ⁹ In 1991, Meyer et al reported a 20-year experience treating 96 patients with pulmonary embolectomy on cardiopulmonary bypass. ¹⁰ Operative mortality was 37.5%. Numerous case series have since been published with a wide range of operative mortalities, ranging from approximately 10 to 50%. In contemporary reports, the operative mortality is approximately 5%. ^{11 12} Surgical pulmonary embolectomy remains an important tool in our armamentarium.

A Century of Anticoagulants and Fibrinolytics

Anticoagulation is the mainstay of pulmonary embolism treatment and prevention. The topic of anticoagulation is broad and beyond the scope of this historical view, but we will attempt to briefly cover a few salient points. Heparin was the earliest anticoagulant routinely used in medicine. It was discovered by Howell at Johns Hopkins University around 1918 to 1922 and first used in a human patient in 1937. ^{13 14} Crafoord, a Swedish thoracic surgeon who had performed pulmonary embolectomy, recognized the potential of anticoagulation for treatment of pulmonary embolism as early as 1929. ¹⁴ However, reports of the use of heparin for treatment of thrombosis were rare until the 1940s. Broad acceptance for the use of heparin in patients suspected of having perioperative pulmonary embolism began in the 1960s after the publication of a randomized trial by Barritt and Jordan. ¹⁵ In the 1970s, heparin began to come into routine use for extracorporeal circuits and for prophylactic perioperative use. ¹⁴ In 1976, low-molecular-weight heparin was discovered and it was first synthesized in 1983. The 1980s also showed development of tissue plasminogen activator (tPA), initially discovered by Astrup in 1952, after cloning of the gene in 1983. ¹⁶ Ultimately, this led to the recombinant production of tPA and rapid approval by the Food and Drug Administration as a thrombolytic. Both heparin and tPA remain essential to contemporary treatment of acute pulmonary embolism.

While early work purifying and understanding the biochemistry of heparin was underway, Paul Link's group discovered Coumadin at the University of Wisconsin-Madison. A local rancher serendipitously presented to Dr. Link's biochemistry laboratory while he was looking for the state veterinarian. The rancher's cattle were dying of a hemorrhagic illness known as sweet clover disease. This led to the discovery and purification of dicoumarol in 1939, followed by the production of numerous analogs, including warfarin/Coumadin in 1940. ¹⁷ In 1954, warfarin/Coumadin was approved for use in humans, and it has been used extensively for prevention and treatment of pulmonary embolism, deep venous thrombosis, and stroke. While warfarin has saved countless lives and remains a key drug in our current pharmacopeia, recent years have seen a proliferation of other anticoagulants including the leech venom hirudin analog bivalrudin and numerous direct oral anticoagulants such as dabigatran. The proliferation of anticoagulants speaks to their important role in multiple facets of modern medicine and pulmonary embolism in particular.

Peripheral Embolectomy, Occlusion, and Filters

Early attempts at central surgical embolectomy were fraught with failure. Recognition that pulmonary emboli result from dislodgment of peripheral clot led to several initial approaches focused on those peripheral sites. Peripheral thrombectomy was proposed by Lawen in 1938. ¹⁸ However, peripheral thrombectomy was not efficacious. Other early concepts included ligation of femoral veins; an area explored by Homans. ¹⁹ Caval ligation was also explored extensively. Ten years after his work on femoral vein ligation, Homans proposed a novel application of caval ligation for the prevention of pulmonary embolism. ²⁰ Caval ligation itself was not entirely novel. Kocher and Billroth had both used caval ligation for trauma in 1883 and 1885, respectively. ²¹ Neither of these first two caval ligation patients survived. However, Bottini successfully performed caval ligation for trauma in 1893. ²¹ Prophylactic caval ligation for pulmonary embolism began in earnest with a small case series in the 1940s. ²² By the mid-1950s, caval ligation seemed to be showing promising results with no recurrent pulmonary emboli in a 60-patient series. ²³ While there was enthusiasm surrounding these early successes, caval ligation was highly morbid. Approximately two-thirds of patients experienced lower extremity edema, pain, and ulceration. Mortality was approximately 20 to 40%. ²⁴ In addition to the significant morbidity and mortality, recurrent pulmonary embolism remained possible from enlarged collateral veins arising from caval occlusion. ²⁵ Operative caval plication was conceived to exploit the benefits of preventing pulmonary embolism while avoiding the morbidity and mortality of complete caval occlusion. One of the earliest successful applications of caval plication, using silk suture to create multiple channels, was reported by Spencer et al in 1962. ²⁶ Numerous other permutations of caval plication followed. These included the use of porcelain beads, sieves, stapling, and clips. ²⁷

While the early caval interventions carried a high morbidity and mortality, they hinted at the promise of blocking peripheral emboli from reaching the pulmonary arteries. Ultimately, this led to the conception of caval filters that could capture emboli in transit while still allowing venous drainage of the lower extremities and pelvis. The first caval filter, Mobin-Uddin caval filter, was introduced in 1967. ²⁸ This early filter design suffered from a high rate of caval occlusion and filter migration. The ubiquitous Greenfield filter was introduced in 1973 and numerous derivations have since been introduced. ^{28 29} While they are indicated in less than 1% of pulmonary embolism patients, namely in patients who fail or have contraindications to anticoagulation, caval filters remain important in contemporary management of pulmonary embolism.

Making the Diagnosis

Diagnosis of pulmonary embolism utilizes a combination of clinical suspicion, history, signs, symptoms, laboratory testing, and imaging. Prior to the 20th century, laboratory and radiologic modalities were not in clinical use for diagnosis of pulmonary embolism. Instead, clinical assessment of history, symptoms, signs, and vitals were utilized. For example, as noted previously, in the 1800s, clinical signs such as jugular venous distension and tachypnea were monitored continuously by bedside medical trainees in patients suspected of having pulmonary embolism. In the 1900s, a rapid proliferation of diagnostic methods relevant to pulmonary embolism occurred. These included advances in laboratory medicine and radiology. ^{19 30}

The early laboratory testing included work by Wilson et al. They reported elevated levels of fibrin split products in patients with pulmonary embolism. ^{31 32} This test showed a high degree of sensitivity, but low specificity. These findings led to the development of the D-dimer test, which improved sensitivity, but still suffers from low specificity. The D-dimer test remains in use to this day and is often used in the emergency department setting. More recently, cardiac enzymes (troponins and beta natriuretic peptide) have been used to help assess ventricular strain in patients with acute pulmonary embolism.

Radiologic methods have been a mainstay of pulmonary embolism diagnosis since the early 20th century. The hallmark study by Hampton and Castleman described many radiographic findings on chest roentgenogram (CXR) in 1940. ³³ These include Hampton's hump, pulmonary consolidation with air bronchogram and/or air lucencies. Their findings were confirmed with postmortem examinations. However, findings on a plain CXR are insensitive and nonspecific. ³⁴ Lung scans were explored by Wagner et al in the 1960s at Johns Hopkins University, but they too suffered from poor sensitivity and specificity. ^{32 35} Robb and Steinberg reported angiographic visualization of pulmonary vessels in humans in 1939, while Aitchison and McKay first reported angiographic diagnosis of pulmonary embolism in humans in 1956. ^{36 37} Pulmonary angiogram became the gold standard for the diagnosis of pulmonary embolism until it was supplanted by contrast computed tomography angiography (CTA). Goodman et al are credited for advancing CTA for the diagnosis of pulmonary embolism. ³⁸ A study in 2006 found this test to be 83% sensitive and 96% specific. ³⁹ Modern CTA provides high-quality images very quickly with reduced radiation goals.

Echocardiography came into clinical use relatively recently but has become an essential radiologic modality in modern management of massive and submassive pulmonary embolism. The physician Inge Edler, working with physicist Hellmuth Hertz, was putatively the first to apply ultrasound to examination of the human heart in 1953. ⁴⁰ Edler's primary interest in ultrasound was its application to rheumatic mitral disease. Around the same time, in the late 1940s, other groups were investigating ultrasound in tissues. ⁴¹ Adoption into clinical use was initially slow but began to gain momentum in the 1960s and 1970s. Technological advances and refinements allowed development of two-dimensional, three-dimensional, and transesophageal echocardiographies in the following decades. ^{40 42} Echocardiography now plays an essential role in assessing right ventricular (RV) strain in acute pulmonary embolism.

Current Diagnosis and Management

The presentation of pulmonary embolism varies in severity from asymptomatic to sudden death. Thus, severity of pulmonary embolism is a key factor in risk stratification and treatment choice. In current practice, pulmonary embolism severity is stratified into three main categories by both the American Heart Association (AHA) and the European Society of Cardiology (ESC). ^{43 44} There are differences in these classification schemes, and the following is a generalized summary of the categories. The massive (AHA) or high-risk (ESC) category is defined by hypotension (for more than 15 minutes) or need for vasopressor. These patients make up ∼5% of hospitalized patients with pulmonary embolism and have a 30-day mortality of ∼15%. ⁴⁵ The submassive (AHA) or intermediate-risk (ESC) group is defined by RV strain/dysfunction without hypotension. RV strain may be assessed by various measures including echocardiography, CTA, and cardiac enzymes. This group represents ∼35 to 55% of pulmonary embolism patients. The mortality of this group is less clear, but ranges from ∼3 to 15% between 30 and 90 days. Finally, the low-risk (AHA and ESC) group is characterized by patients not meeting criteria for either of the preceding categories. This group makes up ∼40 to 60% of hospitalized patients with pulmonary embolism, and they have a 30-day mortality of ∼1%. ⁴⁶ Risk stratification is essential to optimizing individualized treatment plans for pulmonary embolism.

Anticoagulation is the mainstay of pulmonary embolism treatment and prevention of recurrence. Generally, anticoagulation alone is the treatment for patients in the low-risk group. For patients in the submassive/intermediate-risk, and especially the massive/high-risk categories, anticoagulation may be insufficient. A myriad of additional treatment options exists. Systemic thrombolysis, typically with tPA, may be considered first-line therapy for massive/high-risk pulmonary embolism and can be considered in submassive/intermediate-risk patients. Compared with anticoagulation, tPA may allow more rapid improvement in pulmonary artery recanalization and hemodynamics. However, these benefits must be carefully weighed against the higher risk of hemorrhage, including a 2 to 3% risk of intracranial hemorrhage. Catheter-directed fibrinolysis represents an evolution of systemic thrombolysis in which thrombolytics are delivered via catheter localized to the obstructed pulmonary artery. Other catheter-based thrombolytic interventions include the use of ultrasound to facilitate dissociation of fibrin strands. Numerous other percutaneous mechanical techniques (e.g., wire-based mechanical disruption, balloon fragmentation, rheolytic high-pressure jets, and suction/aspiration) have been developed. ⁴⁵ This technological proliferation seems reminiscent of the early attempts at caval ligation and plication. Notably, many of these approaches lack randomized trial data, and further investigation is needed.

Recent years have seen a resurgence in the role of cardiothoracic surgeons in the management of massive/high-risk and submassive/intermediate-risk pulmonary embolism. There are two major areas where cardiothoracic surgeons remain essential to the management of pulmonary embolism: surgical embolectomy with cardiopulmonary bypass and extracorporeal membrane oxygenation (ECMO). Surgical embolectomy, now performed with cardiopulmonary bypass, has re-emerged as a viable treatment option for patients with massive/high-risk pulmonary embolism and failure or contraindications to other methods. In particular, the approach may be useful in patients with emboli in transit due to impending paradoxical thromboembolic stroke. As we noted earlier, operative mortality in the current era is relatively low at approximately 5%. In addition to surgical embolectomy, mechanical circulatory support with venoarterial ECMO (VA ECMO) is increasingly considered as part of the armamentarium for acute massive/high-risk pulmonary embolism. ⁴⁷ Strengths of VA ECMO include supporting perfusion, gas exchange, and RV unloading; all of which are critical issues that can lead to mortality from acute pulmonary embolism. Surgical embolectomy and VA ECMO are important modalities for management of pulmonary embolism. As with other treatment modalities, their application must include careful consideration of the risks and benefits on an individualized basis.

The complexity of the disease and our expanding armamentarium for treatment has given rise to the establishment of Pulmonary Embolism Response Teams (PERTs). These teams comprise a multidisciplinary group of specialists with a goal of providing timely care for patients with acute pulmonary embolism. ⁴⁸ Such teams often include emergency medicine physicians, pulmonologists, critical care physicians, anesthesiologists, pharmacists, interventional cardiologists, perfusionists, and cardiothoracic surgeons. PERTs may help navigate the complexity of patient presentation and treatment options to optimize management of acute pulmonary embolism including low-risk, submassive/intermediate-risk, and massive/high-risk categories. Their role in treatment decisions for submassive/intermediate risk may be particularly important given the nuances of patient presentation, RV dysfunction, comorbidities, and follow-up.

Summary

Pulmonary embolism remains a major clinical problem and its rich clinical history is commensurate with its clinical import. What have we learned during the past two centuries? When we reflect on this history, we note recurrent trends in various aspects of pulmonary embolism diagnosis and management. Examples include imaging modalities, anticoagulation, inferior vena cava filters, surgical embolectomy, and catheter-based technologies. In each case, we can appreciate a general trend of discovery of a pathophysiologic insight or technological advance followed by a proliferation of imaging, laboratory testing, or interventions. These proliferations are often refined by selection of the most accurate or efficacious modality. At times, we return to prior concepts (e.g., surgical embolectomy or partial caval occlusion) when the requisite technology becomes available (e.g., the advent of cardiopulmonary bypass and caval filters). In contemporary practice, a patient suspected of acute pulmonary embolism may undergo rapid diagnosis and risk stratification through imaging with CTA, laboratory work (including D-dimer, troponin, and beta natriuretic peptide), echocardiography, and clinical risk stratification by a PERT that dictates the optimal treatment with a combination of anticoagulation, thrombolysis, catheter-directed therapies, caval filter, surgical embolectomy on cardiopulmonary bypass, or mechanical circulatory support with VA ECMO. These sophisticated diagnostic and treatment modalities were developed over the past 150 years. We believe there is still much to discover regarding pulmonary embolism diagnosis and management.

Funding Statement

Funding None.