Highlights of the Study
CT pulmonary angiography often proves a suspected pulmonary embolism in emergency patients.
A pulmonary embolism is quite common even in low-risk Wells score or in young patients <40 years.
Differential diagnosis of chest pain or dyspnea should include pulmonary embolism.
Introduction
Pulmonary embolism (PE) is still one of the main causes of death in hospitalized patients [1]. Studies have shown that PE can be identified in >26% of autopsies, with 9% relating patient death to fatal PE. Being a potentially life-threatening condition, unrecognized PE may have fatal consequences [2]. In the past, mortality of missed PE has been estimated to be in a range from 5 up to 30%. Therefore, a fast and precise diagnostic workup is indispensable. However, the unspecific signs and symptoms associated with PE make a rapid diagnostic evaluation challenging for the physician in an emergency department. Various guidelines recommend a graduated clinical approach using algorithms such as Wells score and revised Geneva score, categorizing patients into low, intermediate, and high clinical probability [3-5]. The reference method for diagnosing or excluding PE is computed tomography pulmonary angiography (CTPA). CTPA represents the most sensitive and precise diagnostic test especially in high-probability cases [6]. Determination of D-dimer levels is recommended in case of low and intermediate clinical probability [5]. Carrying a relatively low specificity for PE, the approach of using positive D-dimer as a standard prerequisite for performing CTPA would lead to a higher number of potential consequences such as contrast nephropathy, allergic reactions to contrast media, radiation exposure, and higher financial burden for health care facilities [7]. In general, a combined strategy of clinical probability and D-dimer testing represents an opportunity for fast clarification. Pathologic values make a CTPA mandatory. This challenging situation in emergency departments has led to various clinical retrospective studies, evaluating the probability of PE in the general patient population [8, 9]. However, little is known about PE in young patient groups in which this diagnosis is usually not expected. Based on this knowledge, we evaluated the role of CTPA in diagnosing PE in an emergency department, with an emphasis on analysis in different age-groups.
Material and Methods
Patients
This retrospective analysis included data from medical records of our emergency department of Internal Medicine (University Hospital Erlangen, Friedrich-Alexander-University Erlangen-Nurnberg, Bavaria, Germany). The study complied with the ethical principles of the World Medical Association Declaration of Helsinki and was approved by the local Ethics Committee. About 10,500 patients per year are admitted to the emergency department. All patients were included who underwent CTPA for suspicion of PE between January 1, 2016, and December 31, 2017. CTPA – available 24 h – was ordered by the physician based on the overall clinical evaluation in the emergency department or when clinical suspicion of PE was high. Alternatively, CTPA was ordered in unclear patient presentation with elevated D-dimers.
Clinical Features
Data were obtained from medical records filled in by nurses and physicians in the emergency department on admission as well as from physician letters and other medical results documented in Soarian Health Archives (Soarian Clinicals, Medical Data Software of the medical faculty, Siemens, Erlangen). Apart from basic data (sex, age, date of CTPA, and reason for admission), records were reviewed for vital signs, blood gas analysis, symptoms and reasons for admission, clinical scores (Wells and revised Geneva scores), laboratory values, as well as results from electrocardiogram (ECG), echocardiography, deep vein ultrasound (within 24 h after admission), and CTPA. Vital signs, reasons for admission and presence of unilateral lower leg edema, and/or pain were evaluated (specific items shown in Table 1). ECG results were evaluated for signs of S1Q3 and/or right bundle branch block. The Wells score was taken from emergency department records or retrospectively calculated, along with the (revised) Geneva score. Core diagnostic parameters (patient age, D-dimer level, presence of dyspnea, chest pain, unilateral leg edema and/or leg pain) were analyzed according to the presence or absence of PE. D-dimer cutoff was 500 ng/mL for patients <50 years old. For patients >50 years, the age was multiplied by the factor 10 µg/L, and the result served as an age-adjusted cutoff value [3, 10, 11]. Sensitivity of D-dimer testing (ELISA) for excluding thromboembolism is ≥95% in our laboratory. Troponin I (normal <0.5 ng/mL) and Cr (normal <0.95 mg/dL) values were analyzed. Hereafter, patients were classified according to Wells criteria for PE in its original version [12, 13]. Earlier deep vein thrombosis or PE, heart rate ≥100/min, surgery, or immobilization during the last 4 weeks represents 1.5 score points each. The presence of hemoptysis and/or active neoplasia represents 1 point each. Three points are given in case of clinical sign of deep vein thrombosis or if an alternative diagnosis is unlikely. Hence, patients were classified into low (0–1 points), intermediate (2–6 points), or high risk (≥7 points) for PE. The Geneva score in its simplified version counted 1 point each for the following parameters: earlier deep vein thrombosis or PE, surgery or fracture during the last 3 months, hemoptysis, active neoplasia, unilateral leg pain, pain on deep palpation of lower extremity + unilateral edema, age >65 years, and heart rate 75–94/min 2 points were added if the heart rate was ≥95 [4, 14].
Table 1.
Overall characteristics of 323 patients with suspicion of PE and performance of CTPA in an emergency department
CT Tomography Pulmonary Angiography
The CTPA of each patient was performed with either Somatom Definition AS CT or Somatom Force CT (Siemens Medical System, Erlangen, Germany). The scan volume included the entire thorax from lung apex to its base in the craniocaudal direction. The slice thickness was 0.75 mm. Monitoring started 10 s after administration of contrast media (in most cases Imeron 350, 50 mL). Flow rate was at least 3 mL/s with Somatom Definition AS CT or at least 4 mL/s with Somatom Force CT and flushed with 50–60 mL NaCl. CTPA interpretation was done by a radiologist highly experienced in diagnosing pulmonary conditions like PE. Results of CTPA were categorized into “PE” or “no PE” as well as the extent of blood vessel occlusion (Table 3) [15]. Extent to peripheral branches was considered a definite PE diagnosis (or positive CTPA), as was PE of the other, higher 3 categories.
Statistical Analysis
For statistical evaluation of collected data, values were presented as frequency, mean, and standard deviation together with the range or proportion (%). To compare the means, the t test was used. A statistically significant result was indicated by p < 0.05. The Statistical Package for the Social Sciences (version 24.0.0.2, IBM Corp., Armonk, NY, USA) supported the explorative data analysis.
Results
Among the 323 patients (mean age 62 years), 158 were female, and 165 were male. In 123/323 (38%) patients, PE was detected (Table 1). Dyspnea, chest pain, or both symptoms were found to be the reason for admission in overall 76% of all patients, presenting the same proportions in patient groups with PE and without. Respiratory infection and syncope were shown in 15.8% of admissions. Tachycardia and tachypnea were present in approximately 30% in the groups with PE and without PE, respectively. Leg pain and/or edema were observed in 23% of cases without PE and in 43.9% of cases with PE, along with proven deep vein thrombosis in up to 54.5% cases. A documented Wells score was found in 8/323 (2.5%) patients. The following data were missing (or examinations not performed): 71 D-dimers, 30 troponin I, 3 Cr, and 115 echocardiographies. In patients with PE, the mean Wells score was 3.0, and in patients without PE 1.5.
Overall, the calculated Wells score showed mostly intermediate risk of PE, then low risk, and rarely high risk (Table 2). The presence of proven PE accounted for 84% of high-risk and 25% of low-risk patients (Fig. 1). Patients <40 years with PE showed low risk according to the Wells score in 1 case. D-dimers were pathologic (age-adjusted) in 98% of PE patients and 95% in the absence of PE. In 15/121 (12%) patients with a low risk, determination of D-dimer levels was not performed or missing. In the group of <40 years, 43% of all CTPA were positive for PE (Table 2), compared to 33% in the group 40–60 years and 39.5% in the group >60 years (with the highest absolute number of proven PE [n = 79]). 60% of patients <40 years had PE within peripheral branches or segmental arteries, whereas the extent to the main pulmonary artery or its branch showed a higher percentage in patients aged ≥40 and >60 years (Table 3). In our study, the extent of PE did not correlate with patient age (r = −0.03; p > 0.05). Looking at the characteristics of 123 patients with PE according to different age-groups (Table 4), 20/29 patients with PE in the group 40–60 years were male. Tachypnea, tachycardia, and chest pain showed the highest percentage in patients <40 years (Table 4). Ultrasound of leg veins was performed in 66.6% of patients <40 years with proven PE, the highest percentage of all age-groups. The mean D-dimer was lower in patients <40 years with PE than that in the higher aged groups with PE. In 1 patient <40 years with peripheral PE, D-dimers were not pathologic.
Table 2.
Calculated risk by the Wells score, pathologic D-dimers, and presence of PE according to different age-groups
Fig. 1.
Positive rate of CTPA in an emergency department according to clinical risk assessment (Wells) and age-groups. CTPA, computed tomography pulmonary angiography.
Discussion
Our results show that 38% of patients who underwent CTPA had the final diagnosis of PE. In particular, in case of chest pain and dyspnea, PE should be part of the differential diagnosis, also in younger patients. A systematic calculation and documentation of the risk of PE (Wells score) seems to be mandatory in accordance with the corresponding guidelines [3-5]. In a similar retrospective analysis evaluating the role of CTPA in an emergency department in the USA, 295 patients in the year 2015 were analyzed [8]. The results showed a PE prevalence of 5.4% in that study population. This result is similar to comparable studies from the USA in which positive PE rates of 2% were shown [9, 16]. Our results seem to be rather close to the rate of 30% in European countries [17-22].
This divergence results from the fact that physicians in many European hospitals seem to adhere to guidelines more often than their American colleagues, generally being more restrictive in terms of ordering CTPA. CTPA is known to be more accessible to physicians in the USA [8]. Another explanation for the comparatively high rate of PE observed in our cohort might be the high proportion of peripheral and subsegmental extent of PE diagnosed in CTPA with both >40% of all PE. Overdiagnosis of PE, including common diagnosis of peripheral and subsegmental PE, is reported in the literature due to an increase in imaging quality of CTPA [23, 24]. Care of subsegmental PE is still debatable. Surveillance of untreated subsegmental PE without deep vein thrombosis might be a therapeutic option [25, 26] In our study, all patients were symptomatic in a certain way, so anticoagulation therapy was undertaken also in subsegmental and peripheral PE (unless contraindicated) in our hospital. Proportion of peripheral and subsegmental PE was higher in younger ages of patients with PE.
PE was found in CTPA in 84% of our patients with high clinical risk (according to Wells score). 12% of the patients with low risk had no determination of D-dimers, probably due to underlying diseases such as chronic renal failure, cancer, or inflammation/infection that might interfere with the interpretation of pathologic levels of D-dimers. In contrast, in 68% of high-risk patients, the D-dimer level was ascertained although guidelines do not recommend this approach. In daily practice, taking a blood sample is performed before or concurrently to anamnesis and physical examination to speed up the diagnostic process, especially when many patients had to be evaluated at the same time. In the intermediate-risk situation, a balanced diagnostic workup involving more clinical and laboratory parameters became more significant: 41% of our patients in this group had a PE. In the literature, the prevalence of 28–40.7% for the detection of PE in intermediate-risk patients is found [27]. Because 25% of our patients classified as low risk were still diagnosed with PE, the standard algorithm of ordering D-dimers and proceeding with CTPA in case of pathologic D-dimers should be followed to avoid the potentially fatal consequences of missed PE. Nevertheless, D-dimers were also regularly positive for patients without PE and showing diagnoses other than PE. This low specificity does not make isolated D-dimer testing without considering clinical parameters reliable enough.
Various risk factors listed in clinical scores seem to apply more often to the older population; however, PE as a probable diagnosis in younger patients should not be underestimated. In 43% of our patients <40 years, CTPA proved PE. There are few data looking at the clinical presentation of young patients with PE. A retrospective analysis performed in the 1990s compared clinical presentation in PE of 40 younger patients with 40 older patients [28]. Clinical signs in younger patients were more subtle apart chest pain, whereas tachycardia and tachypnea were more common in the older group. In contrast, in our emergency department, patients <40 years were more symptomatic: tachycardia, tachypnea, cough, hemoptysis, and especially chest pain were found more often than in patients ≥40 years. In a different study in 2015, symptoms such as chest pain and hemoptysis were more common in the younger population [29].
It seems that in young patients presenting with any kind of these symptoms (or in combination) ordering CTPA should be prioritized or at least PE should be part of the differential diagnosis from the start. In general, PE as a differential diagnosis in young patients still seems to be undervalued because patients <40 years showed a relatively higher probability of PE than the other 2 patient age-groups of our analysis. Moreover, central PE was present in 27% of the 15 patients <40 years with PE. The extent of PE has been generally found to correlate with patient age [30].
More than half of patients (54.5%) diagnosed with PE were found to have deep vein thrombosis (while only 1% of patients with an alternative diagnosis had 1). In 141 patients with PE in a South Korean study, the prevalence of deep vein thrombosis was 45.4% [31]. If thrombosis is to be expected in half of the patients with PE, screening for deep vein thrombosis by performing compression ultrasound in Patients with confirmed PE in the emergency department seems advisable.
Certain weaknesses might impair the interpretation of our results because data were based on hospital records and retrospectively analyzed. The Wells score was mostly calculated retrospectively. Because the Wells score involves the parameter “an alternative diagnosis is less likely than PE,” the physician would have calculated differently ad hoc, which might have classified the patient into a higher or lower risk category (considering the 3 points related to this factor). But even in the acute situation, this issue might be somewhat arbitrary. Another drawback is that only analysis of patients with CTPA was performed, so we cannot conclude for all patients (including those without CTPA), in particular for all young patients or all patients that would have had a low clinical risk and presented in the emergency department.
Conclusion
CTPA frequently helps to confirm PE in Patients with suspicion of PE in an emergency department of Internal Medicine. Even in low-risk calculations (Wells score) or in young patients <40 years (with chest pain or dyspnea), the presence of PE is quite common. This should make us rethink PE in our diagnostic strategies by considering these aspects in future clinical scores and adapt them accordingly.
Statement of Ethics
This retrospective study complied with the ethics of the World Medical Association Declaration of Helsinki and was approved by the local Ethics Committee (No. 340_17Bc from august 20, 2018).
Conflict of Interest Statement
The authors have no conflicts of interest to declare.
Funding Sources
This research did not receive any specific funding from agencies in the public, commercial, or not-for-profit sectors.
Get Permission
PubMed ID
