The landmark 1960 clinical trial that initially validated anticoagulation treatment for pulmonary embolism (PE) included exclusively patients who presented with signs and symptoms of right heart failure or pulmonary infarction.1 Since then, with the advent of noninvasive imaging, the spectrum of disease has radically shifted to include less severe, and even asymptomatic PE. In 1977, Robin2 suggested that clinically unimportant PE was being diagnosed by nuclear scintigraphy, exposing patients to hazards related to unnecessary anticoagulation, incorrect diagnosis of a potentially chronic disease, and unnecessary interventions. This clinical conundrum has grown since computed tomography pulmonary angiography (CTPA) overtook scintigraphy as the dominant PE imaging modality.
Epidemiologists consider overdiagnosis to represent the diagnosis of clinically unimportant disease. PE is particularly prone to overdiagnosis, because some PEs are likely physiological rather than pathologic. Small deep venous thrombosis (DVT) regularly develops in the legs of even young healthy individuals, and one of the functions of the pulmonary capillary bed, which has endogenous thrombolytic activity, is to filter and lyse small silent PEs. Freiman et al3 suggested this in 1965 on the basis of autopsy findings in humans and animals. More recent autopsy studies have identified old or recent asymptomatic PEs in up to 51% to 90% of patients, further suggesting that limited PE is a frequent and benign event.4,5
Two large multicenter prospective clinical trials from the early 21st century showed strong evidence of overdiagnosis of PE on CTPA. PIOPED II evaluated the diagnostic performance of CTPA using a composite reference standard and demonstrated excellent CTPA performance when the results were concordant with clinical probability. Surprisingly, the positive predictive value of CTPA was only 58% (confidence interval, 40-73) for patients with low clinical probability, that is, a large minority of the positive CTPAs in this subset were “false positives.”6 These results cast doubt on the medical community’s ironclad faith in the results of CTPA—a technology that was rapidly embraced and adopted into practice in the 1990s after Remy-Jardin et al’s landmark publication.7 Anderson and colleagues’ large multicenter randomized controlled clinical trial compared CTPA to V/Q scintigraphy for suspected PE. They found that CTPA diagnosed about 50% more PEs than V/Q (17.7% vs. 11.7%) without a difference in false-negative rate (thromboembolism diagnosis within 90 d of a negative result) or mortality on clinical follow-up.8 Editorials accompanying these publications explicitly raised the question of whether the additional PEs diagnosed with CTPA required treatment, or even diagnosis.
Overdiagnosis of PE is also made evident when observing the trends in CTPA utilization, which has corresponded to a significant increase in the number of PEs diagnosed.9,10 However, counter to what may be expected, these additional diagnoses did not translate to a decrease in mortality. In fact, between 1994 and 2004, population data showed that the PE diagnosis rate doubled in New York State, with mortality remaining stable11; this observation was confirmed at the national level.12 This trend, validated using patient-level data for >2000 PE patients from a single center, also showed that patients with PE diagnosed on CTPA were less likely to die (half the odds) than those diagnosed by V/Q.13 We are left to ponder, is CTPA too good an imaging modality for evaluating PE?
The evidence for overdiagnosis of PE at the population level is overwhelmingly supported by the results of multiple studies, with sparse evidence to the contrary. den Exter et al14 refuted the lack of significance of isolated subsegmental PE, demonstrating that the risk profile of patients with subsegmental PE was similar to those with more proximal PE and different from those without PE, and they concluded that patients with subsegmental PE should be considered similar to patients with proximal PE. However, many of the subsegmental PE patients in their series had DVT. Although often conflated, overdiagnosis of PE is not synonymous with subsegmental PE, which occasionally can be accompanied by a substantial DVT burden. More importantly, it is well recognized that even very small PEs can be serious and, at times, fatal in patients with poor cardiopulmonary reserve.
The greatest concern related to PE overdiagnosis is with regard to its overtreatment with anticoagulation, with the attendant complications and costs. The risk of major bleeding is estimated to be 11.5%.15 Overdiagnosis has driven health care cost up, with an increase in total hospital charges of $19,000, between 1998 and 2005.16 There is also an increasing tendency for patients diagnosed with PE to be hospitalized, even though oral anticoagulants are available, and most of these patients could be discharged earlier.17 Treatment with warfarin, which is less costly than newer anticoagulants, requires monitoring with regular bloodwork and, often, dietary adjustment. Labelling a patient with a history of a thromboembolic event can result in unnecessarily intensive future care and repeat imaging, and impact life and disability insurance eligibility, as they are considered at risk of recurrent thrombosis.
Although CTPA is described as a noninvasive imaging study, the term “noninvasive” falsely lulls physicians and patients into a sense of safety and satisfaction. This skewed impression of low risk and high rewards when prescribing a CTPA explains, at least in part, its increased utilization in recent years. However, CTPA comes with its set of risks and complications. Intravenous contrast administration has the potential for allergic-type reactions and, possibly, contrast-induced nephropathy. More importantly, patients are exposed to radiation with attendant stochastic and deterministic risks.
A study reviewing 300 CTPAs performed for emergency room patients determined that CTPA radiation dose corresponded to 65% of patients’ four-year cumulative medical radiation dose.18 Technical advances are reducing the radiation exposure from CTPA, but decreasing population radiation exposure requires bending the imaging utilization curve downward, as the best radiation reduction strategy is to obviate unnecessary imaging. The decreasing positivity rates of CTPA, especially in the United States, is a strong indication of the rise of inappropriate imaging in clinical practice. Poor adherence to appropriateness criteria for CTPA has been widely documented. Clinical tools, including the Wells score, the Geneva score, and PERC (pulmonary embolism rule-out criteria), have been developed to guide management and determine whether additional tests for PE are required. Geeting et al19 studied the impact of requiring assignment of a modified Wells score on CTPA utilization in 96,000 Emergency Department patients and found that only 67% of CTPAs were appropriate. In the United States, nearly half of patients with moderate pretest probability of PE and negative D-dimers undergo imaging studies, twice as many when compared with Canada and European countries (44.4% vs. 19.2%, P<0.01).20 Another study showed that 61% of patients clinically unlikely to have PE did not have a D-dimer test before CTPA.21 This excessive imaging has a snowball effect, with more CTPAs leading to more incidental findings, inevitably leading to increased costs.
Although radiologists are used to making diagnoses on the basis of imaging findings, not every PE that we see represents disease. Understanding overdiagnosis requires us to grasp the concept that PE is a complex process that ranges from normal physiology to a fatal event. As so aptly stated by Jha,22 “We are not passive birdwatchers. We are physicians.” Patient-centered care requires us to shine a bright light on this complexity and lead investigations with the goal of improving outcomes.
1. Barritt DW, Jordan SC. Anticoagulant drugs in the treatment of pulmonary embolism
. A controlled trial. Lancet. 1960;1:1309–1312.
2. Robin ED. Overdiagnosis
and overtreatment of pulmonary embolism
: the emperor may have no clothes. Ann Intern Med. 1977;87:775–781.
3. Freiman DG, Suyemoto J, Wessler S. Frequency of pulmonary thromboembolism in man. N Engl J Med. 1965;272:1278–1280.
4. Goodman LR. Small pulmonary emboli: what do we know? Radiology. 2005;234:654–658.
5. Stein PD, Henry JW. Prevalence of acute pulmonary embolism
among patients in a general hospital and at autopsy. Chest. 1995;108:978–981.
6. Stein PD, Fowler SE, Goodman LR, et al. Multidetector computed tomography
for acute pulmonary embolism
. N Engl J Med. 2006;354:2317–2327.
7. Remy-Jardin M, Remy J, Wattinne L, et al. Central pulmonary thromboembolism: diagnosis with spiral volumetric CT with the single-breath-hold technique—comparison with pulmonary angiography. Radiology. 1992;185:381–387.
8. Anderson DR, Kahn SR, Rodger MA, et al. Computed tomographic pulmonary angiography vs ventilation-perfusion lung scanning in patients with suspected pulmonary embolism
: a randomized controlled trial. JAMA. 2007;298:2743–2753.
9. Prologo JD, Gilkeson RC, Diaz M, et al. CT pulmonary angiography: a comparative analysis of the utilization patterns in emergency department and hospitalized patients between 1998 and 2003. AJR Am J Roentgenol. 2004;183:1093–1096.
10. Wittram C, Meehan MJ, Halpern EF, et al. Trends in thoracic radiology over a decade at a large academic medical center. J Thorac Imaging. 2004;19:164–170.
11. Burge AJ, Freeman KD, Klapper PJ, et al. Increased diagnosis of pulmonary embolism
without a corresponding decline in mortality during the CT era. Clin Radiol. 2008;63:381–386.
12. Wiener RS, Schwartz LM, Woloshin S. Time trends in pulmonary embolism
in the United States: evidence of overdiagnosis
. Arch Intern Med. 2011;171:831–837.
13. Sheh SH, Bellin E, Freeman KD, et al. Pulmonary embolism
diagnosis and mortality with pulmonary CT angiography versus ventilation-perfusion scintigraphy: evidence of overdiagnosis
with CT? AJR Am J Roentgenol. 2012;198:1340–1345.
14. den Exter PL, van Es J, Klok FA, et al. Risk profile and clinical outcome of symptomatic subsegmental acute pulmonary embolism
. Blood. 2013;122:1144–1149; quiz 1329.
15. Meyer G, Vicaut E, Danays T, et al. Fibrinolysis for patients with intermediate-risk pulmonary embolism
. N Engl J Med. 2014;370:1402–1411.
16. Park B, Messina L, Dargon P, et al. Recent trends in clinical outcomes and resource utilization for pulmonary embolism
in the United States: findings from the nationwide inpatient sample. Chest. 2009;136:983–990.
17. Stein PD, Matta F, Hughes PG, et al. Home treatment of pulmonary embolism
in the era of novel oral anticoagulants. Am J Med. 2016;129:974–977.
18. Takahashi EA, Yoon HC. Four-year cumulative radiation exposure in patients undergoing computed tomography
angiography for suspected pulmonary embolism
. Radiol Res Pract. 2013;2013:482403.
19. Geeting GK, Beck M, Bruno MA, et al. Mandatory assignment of modified wells score before CT angiography for pulmonary embolism
fails to improve utilization or percentage of positive cases. AJR Am J Roentgenol. 2016;207:442–449.
20. Pernod G, Caterino J, Maignan M, et al. D-Dimer use and pulmonary embolism
diagnosis in emergency units: why is there such a difference in pulmonary embolism
prevalence between the United States of America and countries outside USA? PLoS One. 2017;12:e0169268.
21. Alhassan S, Sayf AA, Arsene C, et al. Suboptimal implementation of diagnostic algorithms and overuse of computed tomography
-pulmonary angiography in patients with suspected pulmonary embolism
. Ann Thorac Med. 2016;11:254–260.
22. Jha S. Overdiagnosis
versus overtreatment: a false dichotomy. Radiology. 2014;270:628–629.