SEARCH

SEARCH BY CITATION

Keywords:

  • comparison;
  • emergency department;
  • Europe;
  • pulmonary embolism;
  • suspicion;
  • United States

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Addendum
  9. Acknowledgement
  10. Disclosure of Conflict of Interests
  11. References

Summary.  Background: If the prevalence of pulmonary embolism (PE) differs significantly between the US and Europe, this observation could reduce the generalizability of diagnostic protocols for PE derived in either location.Objective: To determine possible causes and potential clinical consequences of these PE prevalence differences.Methods: Secondary analysis of three prospectively collected multicenter samples (two French and one from the US) including 3174 European and 7940 American PE-suspected patients in Emergency departments (ED) (117 for Europe and 12 for US). Comparison of clinical features, resource use and outcomes of European- and US-suspected PE populations in ED.Results: European patients evaluated for PE were significantly older and had a higher clinical pretest probability (CPP) for PE. The final PE prevalence was significantly higher in Europe, in the overall sample (26.5% vs. 7.6%) and in each level of CPP. Suspected European patients categorized as low CPP had a higher posttest probability than US low CPP patients. Suspected US patients categorized as high CPP had a much lower posttest probability of PE than in Europe. The mean number of tests performed for one PE diagnosis was lower in Europe (7.4 vs. 21.6). Among patients diagnosed with PE, European patients had a higher mean severity of illness score and a higher PE-mortality rate (3.4% vs. 0.7%).Conclusions: Among patients suspected of a PE and those ultimately diagnosed with a PE, European patients had higher acuity, a higher pretest probability and worse outcome than US patients. The present study underscores the importance of disease prevalence for pretest probability scoring approaches and for significance interpretation of imaging tests.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Addendum
  9. Acknowledgement
  10. Disclosure of Conflict of Interests
  11. References

Pulmonary embolism (PE) remains a diagnostic problem because of the frequency of occurrence and its presentation with non-specific clinical signs and symptoms that overlap with other common cardiopulmonary disorders. On one hand, the fear of missing a potentially mortal diagnosis drives the desire to test, but on the other hand, the wish to avoid unnecessary examinations and their harmful consequences drives the desire to not test. Some authors speculate that as many as one-half of emergency patients with a PE go unrecognized and untreated leading to a mortality rate as high as 30% [1,2].

The fear of misdiagnosis as well as the availability of several non-invasive diagnostic procedures have fueled an increase in diagnostic testing for PE, especially in emergency departments (ED). The introduction of computed tomographic pulmonary angiography (CTPA) has seen a slight increase in the diagnosis of PE in the US but the number of patients tested without a PE has increased even more [2–4]. In recent diagnostic studies in North America, PE prevalence was as low as 5% to 10%, and approximately one-third of patients were found to have undergone repeated CTPA scanning that was negative for PE [5–7]. Conversely, in European diagnostic studies, PE prevalence remains around 20% to 30% [8,9]. As predicted by Bayes’ theorem, such differences in suspected PE populations may have important consequences for medical practices because the predictive value of diagnostic test results varies with disease prevalence. To our knowledge, no previous study has directly compared the clinical features and outcomes of European and U.S. patients with a suspected PE. We hypothesized that comparison of two large, multicenter samples from Western Europe and the U.S. would reveal significant and clinically important differences in pretest probability, acuity and outcomes.

Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Addendum
  9. Acknowledgement
  10. Disclosure of Conflict of Interests
  11. References

We analyzed three prospective collected databases from patients suspected of a PE.

The compilation of the first two samples represents the ‘European-suspected population’, and the third sample represents the ‘US-suspected population’. The first European sample was a prospective cohort designed to measure the appropriateness of diagnostic criteria used in routine practice to rule in or rule out a PE in 117 EDs from France and Belgium (n = 1529) [10]. The second European sample was the intervention phase of a cluster-randomized trial measuring the effectiveness of a hand-held clinical decision support system to improve the diagnostic work-up for PE suspicion in 20 French EDs (n = 1645) [11]. The US sample was a prospective observational study of patients undergoing testing for a possible PE in 12 EDs from the US (n = 7940) [12]. In all three studies, a standardized form was prospectively completed reporting patient characteristics. Before any diagnostic testing, physicians were invited to give their gestalt assessment of the pretest probability of a PE (as low, intermediate or high for European studies and as <15%, 15%–40% and >40% for the US study, considered low, intermediate and high, respectively). All three studies had a follow-up period (3 months in the European studies and 45 days in the US study). The patients, relatives or general practitioner were interviewed at the end of a follow-up period about the possible occurrence of a venous thromboembolic event or bleeding complication. Moreover in the US study, in absence of telephone follow-up, the patient’s medical record and social security death index were searched. The diagnosis of a thromboembolic event was confirmed according to predefined criteria that included definitive findings on imaging followed by a clinical plan to treat [10–12]. Sudden deaths with no obvious cause were adjudicated as possibly related to PE. In the present study, to standardize reporting, only events occurring in the first 45 days of follow-up were used for outcome designation. Patients were excluded in the three studies if the diagnosis of thromboembolic disease was documented before admission. In European studies, patients were also excluded if (i) a PE was suspected during a hospital stay of more than 2 days duration; or (ii) diagnostic testing was cancelled for ethical reasons, because of rapid death, or because the patient decided to leave the hospital against medical advice or declined testing. In the US study, patients were also excluded before enrollment if (i) the patient indicated that the enrollment hospital was not his or her hospital system of choice for follow-up or (ii) any circumstance suggested that the patient would be lost to follow-up. We considered as a final diagnosis of PE: (i) a PE or a deep vein thrombosis (DVT) diagnosis ruled in at the end of the initial diagnostic work-up; (ii) a thromboembolic event (PE or DVT) occurring during the follow-up period, among patients in whom the diagnosis of PE was initially ruled out or (iii) death adjudicated as related or possibly related to PE. We collected the clinical gestalt assessment prospectively documented and retrospectively calculated the Wells’ score. The subjective criterion about the likelihood of an alternative diagnosis was prospectively collected in the standardized form. As the criterion ‘unilateral lower limb pain’ was not collected in the US database, we calculated the Revised Geneva score (RGS) assuming this criterion was absent. The primary variables related to pretest probability assessment included the proportion of patients categorized in each clinical probability group (low, moderate and high) and the accuracy of categorization compared with observed outcome rate of a PE. The primary patient outcome variables included the overall rate of PE diagnosis, and the Pulmonary Embolism Severity Index (PESI) [13], overall mortality, PE-related mortality and bleeding complications. As previously it has been shown that the majority of PE-related deaths occurred within 2 weeks of the initial work-up [14], we also calculated PE-related mortality in the first 2 weeks of follow-up. As the ‘altered mental status’ was not available in the overall sample, we calculated the PESI assuming this criterion was absent.

We calculated the mean number of tests performed in all European and US patients for each new PE diagnosis as being the ratio of the total number of tests done in all suspected patients divided by total number of new PE diagnoses.

All statistical analyzes were performed using SPSS 15.0 (SPSS Inc., Chicago, IL, USA). A chi-square test (for categorical variables) or Mann–Whitney U-test (for continuous variables) was used to compare characteristics between European and US samples. A multivariate logistic regression was performed in order to determine the relationship between mortality and patient’s origin, D-dimer use, CTPA use or severity (PESI).

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Addendum
  9. Acknowledgement
  10. Disclosure of Conflict of Interests
  11. References

Table 1 reveals multiple differences in baseline characteristics of European (3174 patients) compared with US (7940 patients) PE-suspected populations. PE-suspected European patients were older, had a higher mean respiratory rate, lower oxygen saturation, higher frequency of syncope and hemoptysis, but a lower frequency of chest pain or dyspnea. Compared with US-suspected PE patients, European patients had a higher frequency of a personal history of VTE, congestive heart failure and active cancer. Conversely, patients in the US sample had more chronic respiratory disease, recent surgery and a higher frequency of pregnancy or postpartum status. At the time of diagnostic evaluation, the treating clinicians in Europe considered a PE as the most likely diagnosis significantly more often (34.5% vs. 16.8%P < 0.001). Regardless of the method used, clinical probability assessment categorized a significantly higher proportion of European patients as having a moderate or high pretest probability of PE. Clinicians in Europe ordered significantly more D-dimer tests, ventilation/perfusion (V/Q) scan and leg ultrasonography but fewer CTPA. In Europe, when a D-dimer test was performed, the assay format was more likely to be a high sensitivity quantitative test (98% vs. 73.9%; P < 0.001). The percentage of positive exams for each investigation was higher among the European samples. During the follow-up period, 298 patients were lost in Europe, none in the US.

Table 1.   Characteristics of suspected pulmonary embolism (PE) populations
Suspected populationsEuropean population n = 3174US population n = 7940 P-value
Data (missing/collected)Mean or n (SD or %)Data (missing/collected)Mean or n (SD or %)
  1. VTE, venous thromboembolism; SBP, systolic blood pressure; RGS, Revised Geneva score; CTPA. computed tomographic pulmonary angiography; V/Q, ventilation/perfusion.

  2. *Positive V/Q scan means exam with a high probability of a PE.

Demographic characteristics
 Mean age, years0/317462.4 (0.3)6/793449.0 (0.2)<0.001
 Gender, female0/31741867 (58.8)0/79405328 (67.1)<0.001
Clinical characteristics
 Mean SBP, mmHg46/3128141.2 (0.5)26/7914131.0 (0.3)<0.001
 Mean heart rate, bpm4/317089.9 (0.4)22/791892.2 (0.2)<0.001
 Mean respiratory rate, rpm760/241421.7 (0.1)62/787820.8 (0.1)<0.001
 Mean O2 saturation186/298894.5 (0.1)24/791696.5 (0.0)<0.001
Sign or symptom
 Chest pain12/31621867 (59.0)0/79405767 (72.6)<0.001
 Dyspnea17/31572061 (65.3)0/79405587 (70.4)<0.001
 Syncope or dizziness22/3152666 (21.1)0/7940479 (6.0)<0.001
 Hemoptysis0/3174128 (4.0)0/7940228 (2.9)0.002
 Palpation pain and lower limb oedema2/3172426 (13.4)0/7940710 (8.9)<0.001
 Personal history VTE0/3174605 (19.1)0/7940858 (10.8)<0.001
 Known congestive heart failure15/3159439 (13.9)0/7940581 (7.3)<0.001
 Chronic respiratory disease1/3173401 (12.6)0/79401711 (21.5)<0.001
 Stroke9/3165139 (4.4)0/7940300 (3.8)0.134
 Cancer0/3174242 (7.6)0/7940489 (6.2)0.005
 Past surgery <1 month0/3174155 (4.9)0/7940520 (6.5)0.001
 Fracture0/317467 (2.1)0/7940149 (1.9)0.419
 Current pregnancy6/316825 (0.8)0/7940147 (1.9)<0.001
 Postpartum <4 weeks7/316710 (0.3)0/7940141 (1.8)<0.001
 Current anticoagulant treatment10/3164211 (6.7)0/7940520 (6.5)0.818
 PE is the most likely diagnosis4/31701095 (34.5)0/79401336 (16.8)<0.001
Clinical probability classification
 Gestalt assessment
  Low1142/2032766 (37.7)8/79325357 (67.5)<0.001
  Moderate848 (41.7)2087 (26.3)
  High418 (20.6)488 (6.2)
 Wells score
  Low2/31721740 (54.9)0/79405482 (69.0)<0.001
  Moderate1222 (38.5)2201 (27.7)
  High210 (6.6)257 (3.2)
 RGS
  Low5/31691027 (32.4)28/79123292 (41.6)<0.001
  Moderate2011 (63.5)4421 (55.9)
  High131 (4.1)199 (2.5)
 Examinations performed
  D-dimer test0/31742838 (89.4)0/79405907 (74.4)
  Quantitative D-dimer test0/28382783 (98.0)0/59074363 (73.9)
  CTPA0/31741161 (36.6)0/79404237 (53.4)
  V/Q scan3/3171484 (15.3)0/7940468 (5.9)
  Leg Ultrasonography0/31741137 (35.8)0/7940987 (12.4)
  Angiography0/31743 (0.09)0/7940499 (6.3)
 Positive Exams
  D-dimer test0/28381825 (64.3)0/59072242 (38.0)
  Quantitative D-dimer test0/27831775 (63.8)0/43631908 (43.7)
  CTPA0/1161415 (35.7)0/4237450 (10.6)
  V/Q scan*0/484157 (32.4)0/46839 (8.3)
  Leg Ultrasonography0/1137371 (32.6)0/987155 (15.7)
  Angiography0/32 (66.7)0/49922 (4.4)
 Diagnosis
  VTE during emergency work-up0/3174733 (23.1)0/7940544 (6.9)<0.001
  VTE during FU298/287622 (0.7)0/794016 (0.2)0.002
  All VTE = final PE diagnosis298/2876763 (26.5)0/7940561 (7.1)<0.001
Final PE prevalence according to pretest probability
 Gestalt
  Low110/65651 (7.8)0/5357167 (3.1)<0.001
  Moderate75/773201 (26.0)0/2087223 (10.7)
  High14/404263 (65.1)0/488171 (35.0)
  Total199/1833515 (28.1)0/7932561 (7.1)
 Wells score
  Low217/1523151 (9.9)0/5482172 (3.1)<0.001
  Moderate74/1148470 (40.9)0/2201294 (13.4)
  High7/203134 (66.0)0/25795 (37.0)
  Total298/2874755 (26.3)0/7940561 (7.1)
 RGS
  Low133/89491 (10.2)0/329293 (2.8)<0.001
  Moderate160/1851585 (31.6)0/4421401 (9.1)
  High5/12678 (61.9)0/19963 (31.7)
  Total298/2871754 (26.3)0/7912557 (7.0)

The overall final PE prevalence was significantly higher in Europe (26.5% final PE diagnosed in Europe vs. 7.1% final PE diagnosed in the US, 95% confidence interval [CI] for the difference of 19% = 17.8% to 21.2%). At each strata of the pretest probability, the PE prevalence was approximately two-fold higher in Europe vs. the US, regardless of the method of assessment used (gestalt, Well’s score or RGS) (Table 1). The mean number of tests performed in all patients for each new PE diagnosis was 7.4 in Europe and 21.0 in the US (P < 0.001).

The severity of PE assessed by PESI was significantly higher in Europe 55.8% vs. 44.6% patients with a PESI > 85 (P < 0.001) (Table 2). Significantly more patients had a PE diagnosed using CTPA scanning in the US compared with Europe. PE-related mortality was significantly higher in the European population (3.4% vs. 0.7%, 95% CI for difference = 1.2% to 4.2%) whereas overall mortality and bleeding complications were similar. In the first 2 weeks, the results of PE-related mortality were similar (Table 2).

Table 2.   Characteristics of patients with a final pulmonary embolism (PE) diagnosis
Patients with final VTE diagnosisEuropean population n = 763US population n = 561 P-value
Data (missing/collected)Mean or n (SD or %)Data (missing/collected)Mean or n (SD or %)
  1. VTE, venous thromboembolism; PESI, Pulmonary Embolism Severity Index [13]; CTPA, computed tomography pulmonary angiography.

Demographic characteristics
 Mean age, years0/76369.0 (0.6)1/56055.6 (0.8)<0.001
 Gender, female0/763462 (60.6)0/561305 (54.4)0.024
Severity assessment – PESI
 ≤65 (Class I)179/58498 (16.8)7/554198 (35.7)<0.001
 66–85 (Class II)160 (27.4)109 (19.7)
 86–105 (Class III)142 (24.3)117 (21.1)
 106–125 (Class IV)99 (16.9)63 (11.4)
 >125 (Class V)85 (14.6)67 (12.1)
Exams
 CT performed0/763482 (62.3)0/561479 (85.4)<0.001
Complications
 Overall death0/76345 (6.7)0/56132 (5.7)0.62
 PE-related death0/76326 (3.3)0/5614 (0.7)0.003
 PE-related mortality in the first 2 weeks0/76318 (2.4)0/5612 (0.4)0.006
 Bleeding0/76337 (4.5)0/56125 (4.5)0.840

We used a multivariate logistic regression analysis to test if the independent variables country of origin, CTPA (performed or not), PESI (classes I and II vs. III, IV and V) and D-dimer (performed or not) were significant predictors of mortality. This analysis found that only the continent of origin was a significant predictor of mortality (Table 3).

Table 3.   Mortality multivariate logistic regression analysis in patients with a pulmonary embolism (PE)
Co-variablesDependent variable: mortality
P-valueOR (95% CI)
  1. CTPA, computed tomography pulmonary angiography. Equation: Mortality = −2.65 − 0.972 Origin + 0.83 CTPA + 0.958 PESI II − 0.773 D-dimer. Pearson’s chi-square goodness of fit: 5.368. Hosmer–Lemeshow P-value: 0.615.

Continent origin (US vs. Europe)0.0040.39 (0.20–0.73)
Quantitative D-dimer performed0.1060.47 (0.19–1.17)
CTPA performed0.8881.07 (0.42–2.74)
PESI (classes I and II vs. III, IV and V)0.0732.56 (0.92–7.14)

Table 4 shows the hypothetical posttest probabilities that could be expected in each country, using published estimates of likelihood ratios for CTPA [15,16]. The primary findings of importance are the relatively high posttest probability for PE (6.9%–25.1%) after a negative CTPA in high pretest probability patients in Europe and the relatively low posttest probability (40.8%–45.8%) after a positive CTPA in low pretest probability patients in the US. Posttest probabilities were similar when calculated using the RGS and the original Well’s score (data not shown). We also calculated the posttest probabilities using the second-level Well’s score categorizing patients as unlikely or likely as having a PE (Table 4).

Table 4.   Calculation of posttest probability for computed tomography pulmonary angiography (CTPA) results
Pretest probabilityPE prevalence in EuropePE prevalence in the USPosttest probability in Europe if negative*Posttest probability in the US if negative*Posttest probability in Europe if positive†Posttest probability in the US if positive†
  1. *Assuming a negative likelihood ratio of 0.04 or 0.18 [15,16].

  2. †Assuming a positive likelihood ratio of 19.6 or 24.1 [15,16].

  3. PE, pulmonary embolism.

Gestalt      
 Low7.83.40.3–1.50.1–0.662.4–67.140.8–45.8
 Moderate26.011.41.4–5.90.5–2.387.3–89.471.6–75.5
 High65.136.06.9–25.12.2–9.297.3–97.891.7–93.2
Wells      
 Likely (≤4)16.33.90.8–3.50.1–0.677.2–80.644.3–49.4
 Likely (>4)52.723.84.3–16.70.5–2.395.2–96.186–88.3

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Addendum
  9. Acknowledgement
  10. Disclosure of Conflict of Interests
  11. References

The present study confirms the presence of major differences in PE prevalence in patients tested, clinical characteristics and the diagnostic management between European- and US-suspected PE populations. Additionally, the data show significant differences in the severity of PE after diagnosis. Our results suggest some causes and consequences of these discrepancies.

Differences in clinical patients’ characteristics suggest that at the time of diagnostic evaluation in the ED setting, European patients had a higher acuity than US patients: they were older, had a higher mean respiratory rate, a lower mean O2 saturation and more frequently had syncope and a serious comorbidity such as cancer and heart failure. Even if all differences were statistically significant, some of them were clinically moderate. Nevertheless, an overall analysis suggests this higher acuity at ED admission of European patients. Dyspnea and chest pain were found more frequently in the US-suspected population, suggesting that these general classic symptoms were more often the reason that diagnostic testing was initiated in the US. Because the prevalence of PE was very low in the US, as a consequence, the number of tests ordered in all patients needed to achieve one PE diagnosis was near three-fold more in the US than in Europe.

Differences were also observed in clinical probability assessment. Regardless of the method used, the proportion of patients who were classified as low pre-test probability was much lower in Europe than in the US. Moreover, PE prevalence was around two-fold higher in each level of clinical probability in Europe. These findings have important clinical consequences in terms of estimating posttest probability of diagnostic strategies using a Bayesian approach. The risk of inappropriately ruling out PE on a negative CTPA scan in high clinical probability patients may be as high as 25% in Europe. Conversely, the risk of over-diagnosis of PE with a positive CT scan may reach 64% in US low clinical probability patients suspected of a PE (Table 4). This underscores the classic teaching that posttest probability, what clinicians and more importantly patients are most concerned about, is highly dependant on the baseline prevalence of disease among all those who are tested. Indeed, the risk is the highest when the clinical probability evaluation and the test result are discordant, for example low clinical probability and positive CTPA or high clinical probability and negative CTPA. As a consequence of very low prevalence in the US low clinical probability and the relatively high prevalence in the Europe high clinical probability, the problem is crucial in US low clinical probability with positive CTPA, as it is in Europe in high clinical probability with negative CTPA. In our sample, 135 American patients had low pretest probability and positive CTPA; on the other hand, 93 European patients had high pretest probability and negative CTPA. These figures mean that in 30% (135/450) of all positive CTPA in the US there is a potential risk of false-positive CTPA; and that in 13% (93/709) of all negative CTPA there is a potential risk of a false-negative, suggesting adaptation of our practices. There are no evidence-based recommendations, but first we suggest considering a second reading of the examination focusing on the quality of the test and the reliability of the result; a second evaluation of the CTPA performed by an expert CTPA radiologist would probably be effective to decrease the CTPA evaluation error [17–19]. In the case of suboptimal CTPA (and/or discordance between radiologists), consider performing another CTPA optimizing parameters or a V/Q scan in young patients without previous a PE or chest disease. In the case of a confirmed negative CTPA and high clinical probability in Europe, consider lower limb deep vein analysis [15,16,20]. A proximal lower limb or pelvic deep venous thrombosis means a false-negative CT and justified anticoagulant treatment. Conversely, if both are negative, clinically relevant venous thrombosis disease can be ruled out and anticoagulant can be withheld with confidence. In the case of a confirmed positive CTPA and low clinical probability, especially in the US, consider the level of pulmonary arteries involved. In the case of subsegmental PE, performing an ultrasonography or CT venography and withholding anticoagulant treatment in the absence of thrombosis seems appropriate [16,21,22].

These data show important differences between European and US patients diagnosed with PE in the ED. Severity and PE-related mortality were significantly lower in US PE patients. PESI classes I and II have been demonstrated to identify patients with a very low risk mortality which could benefit from out of hospital treatment [23]. Our results infer the possibility that a larger proportion of patients in the US could be considered for outpatient treatment compared with Europe.

Several explanations could be hypothesized to explain these differences in PE patients. The wide availability of D-dimer tests is associated with an increase of tested patients [2,24]. Multi-detector CT may allow detection of smaller emboli than other diagnostic strategies such as V/Q scanning, and may increase the number of diagnosed PE without impacting mortality [4,17,25]. However, more D-dimer testing was performed in European patients and the multivariate regression analysis shows that the rate of PE-related mortality was significantly higher in European PE patients unrelated to CTPA use.

We speculate that European PE patients are more severely ill and have a higher PE mortality because European clinicians suspect PE in patients with more serious symptoms and/or a worse health condition than American physicians: PE patients are more serious because suspected PE patients are more serious. This may reflect a lower suspicion threshold in the US than in Europe. European practice seems to be not to suspect PE and not to order further tests in patients with mild symptoms. That may increase the risk of missed diagnosis by under-investigation. Conversely in the US, clinicians seem to suspect PE and initiate a diagnostic process at a very low level of suspicion (clinical probability) in patients with fewer serious symptoms and a better health condition. This may lead to over investigations and to possible over-diagnosis in the US. This lower suspicion threshold could be a consequence of medical teaching and practices more based on recommendations and standardized protocols in the US than in Europe, probably owing to the fear of legal procedures. Conversely, the medical and economic consequences of possibly unnecessary exams such as CTPA seem to be emphasized in Europe and, combined with difficulties in radiologic test access, may lead to under-investigation. Another explanation would be differences in populations evaluated in EDs and health care organization in the US vs. Europe. In Europe, as access to primary health care is easy; patients in emergency departments are often referred there by their general practitioner. In contrast, a larger proportion of patients in the US may have limited or no access to primary care physicians, and therefore must rely on the ED as their only portal of access to health care. Thus, it can be inferred that more young and healthy persons with chest pain or dyspnea may first visit a general practitioner in Europe, whereas these patients must to go to an ED in the US. In order to confirm these hypotheses, a prospective study with patient’s inclusion based on classic PE symptoms (dyspnea or chest pain) and not on clinician’s suspicion would be necessary. Such a study, including American and European patients, could allow us to derive a first step rule able to standardize the PE suspicion threshold.

The present study has some limitations. It was a secondary analysis of three prospective studies and the period of the European and the US studies were not exactly the same. Secondly, while no patient was lost during follow-up probably because of the US enrollment criteria, some European patients were lost during follow-up and could have had thromboembolic events or PE-related death. However, European lost to follow-up patient populations were similar than the analyzed population (data not shown), decreasing the likelihood of such a bias. Two extreme situations could have been possible: (i) no thromboembolic event was diagnosed among these patients, the overall PE prevalence would be in this case of 24.0% which do not change our results; or (ii) all lost of follow-up patients had a thromboembolic event leading the overall PE prevalence to 33.4% which only would further increase the PE prevalence difference between the two continents. Finally, similar to most existing studies of PE diagnosis, our investigation concerned only PE suspected patients in the ED, not allowing our results to be extrapolated to inpatients and general medicine practice. However, our results were obtained from unselected PE suspected patients from a large number of EDs (117 in Europe and 12 in the US.) and patients were managed as in daily practice. Therefore, our results are highly likely to provide an accurate sample of European and American prevalence and practice-pattern differences between the two continents.

Conclusion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Addendum
  9. Acknowledgement
  10. Disclosure of Conflict of Interests
  11. References

Suspected PE populations in ED in Europe and the US have several clinically important differences in acuity and outcome. Patients in Europe who were judged to be low risk by clinicians had a higher posttest probability of disease than US low-risk patients. Patients tested in the US who were considered to be high risk had a much lower post-test probability of PE than in Europe. The present study underscores the importance of considering prevalence of disease when applying pre-test probability scoring approaches and when interpreting the significance of imaging tests.

Addendum

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Addendum
  9. Acknowledgement
  10. Disclosure of Conflict of Interests
  11. References

A. Penaloza and P.-M. Roy had full access to all data study and take responsibility for the integrity of the data and the accuracy of the data analysis. Study concept and design: A. Penaloza, J. Kline, F. Verschuren, P.-M. Roy. Analysis and interpretation of data: A. Penaloza, J. Kline, F. Verschuren, D.M. Courtney, F. Zech, B. Derrien, B. Vielle, F. Thys, P.-M. Roy. Drafting of the manuscript: A. Penaloza, J. Kline, F. Verschuren, D.M. Courtney, P.-M. Roy. Critical revision of the manuscript for important intellectual content: F. Zech, B. Vielle, B. Derrien, A. Armand-Perroux, F. Thys. Statistical analysis: A. Penaloza, F. Zech, B. Vielle. Study supervision: A. Penaloza, J. Kline, F. Verschuren, D.M. Courtney, B. Derrien, A. Armand-Perroux, F. Thys, P.-M. Roy.

Acknowledgement

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Addendum
  9. Acknowledgement
  10. Disclosure of Conflict of Interests
  11. References

We gratefully acknowledge the ‘Fondation Saint-Luc’ for providing a research grant to A. Penaloza.

Disclosure of Conflict of Interests

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Addendum
  9. Acknowledgement
  10. Disclosure of Conflict of Interests
  11. References

The authors state that they have no conflict of interest.

References

  1. Top of page
  2. Abstract
  3. Introduction
  4. Methods
  5. Results
  6. Discussion
  7. Conclusion
  8. Addendum
  9. Acknowledgement
  10. Disclosure of Conflict of Interests
  11. References
  • 1
    Benotti JR, Ockene IS, Alpert JS, Dalen JE. The clinical profile of unresolved pulmonary embolism. Chest 1983; 84: 66978.
  • 2
    Goldstein NM, Kollef MH, Ward S, Gage BF. The impact of the introduction of a rapid D-dimer assay on the diagnostic evaluation of suspected pulmonary embolism. Arch Intern Med 2001; 161: 56771.
  • 3
    Le Gal G, Bounameaux H. Diagnosing pulmonary embolism: running after the decreasing prevalence of cases among suspected patients. J Thromb Haemost 2004; 2: 12446
  • 4
    Wiener RS, Schwartz LM, Woloshin S. Time trends in pulmonary embolism in the United States: evidence of overdiagnosis. Arch Intern Med 2011; 171: 8317.
  • 5
    Kline JA, Mitchell AM, Kabrhel C, Richman PB, Courtney DM. Clinical criteria to prevent unnecessary diagnostic testing in emergency department patients with suspected pulmonary embolism. J Thromb Haemost 2004; 2: 124755.
  • 6
    Runyon MS, Webb WB, Jones AE, Kline JA. Comparison of the unstructured clinician estimate of pretest probability for pulmonary embolism to the Canadian score and the Charlotte rule: a prospective observational study. Acad Emerg Med 2005; 12: 58793.
  • 7
    Kline JA, Courtney DM, Beam DM, King MC, Steuerwald M. Incidence and predictors of repeated computed tomographic pulmonary angiography in emergency department patients. Ann Emerg Med 2009; 54: 418.
  • 8
    Righini M, Le Gal G, Aujesky D, Roy PM, Sanchez O, Verschuren F, Rutschmann O, Nonent M, Cornuz J, Thys F, Le Manach CP, Revel MP, Poletti PA, Meyer G, Mottier D, Perneger T, Bounameaux H, Perrier A. Diagnosis of pulmonary embolism by multidetector CT alone or combined with venous ultrasonography of the leg: a randomised non-inferiority trial. Lancet 2008; 371: 134352.
  • 9
    Perrier A, Roy PM, Aujesky D, Chagnon I, Howarth N, Gourdier AL, Leftheriotis G, Barghouth G, Cornuz J, Hayoz D, Bounameaux H. Diagnosing pulmonary embolism in outpatients with clinical assessment, D-dimer measurement, venous ultrasound, and helical computed tomography: a multicenter management study. Am J Med 2004; 116: 2919.
  • 10
    Roy PM, Meyer G, Vielle B, Legall C, Verschuren F, Carpentier F, Leveau P, Furber A. Appropriateness of diagnostic management and outcomes of suspected pulmonary embolism. Ann Intern Med 2006; 144: 15764.
  • 11
    Roy PM, Durieux P, Gillaizeau F, Legall C, Armand-Perroux A, Martino L, Hachelaf M, Dubart AE, Schmidt J, Cristiano M, Chretien JM, Perrier A, Meyer G. A computerized handheld decision-support system to improve pulmonary embolism diagnosis: a randomized trial. Ann Intern Med 2009; 151: 67786.
  • 12
    Kline JA, Courtney DM, Kabrhel C, Moore CL, Smithline HA, Plewa MC, Richman PB, O’Neil BJ, Nordenholz K. Prospective multicenter evaluation of the pulmonary embolism rule-out criteria. J Thromb Haemost 2008; 6: 77280.
  • 13
    Aujesky D, Perrier A, Roy PM, Stone RA, Cornuz J, Meyer G, Obrosky DS, Fine MJ. Validation of a clinical prognostic model to identify low-risk patients with pulmonary embolism. J Intern Med 2007; 261: 597604.
  • 14
    Carson JL, Kelley MA, Duff A, Weg JG, Fulkerson WJ, Palevsky HI, Schwartz JS, Thompson BT, Popovich J Jr, Hobbins TE, Spera MA, Alavi A, Terrin ML. The clinical course of pulmonary embolism. N Engl J Med 1992; 326: 12405.
  • 15
    Roy PM, Colombet I, Durieux P, Chatellier G, Sors H, Meyer G. Systematic review and meta-analysis of strategies for the diagnosis of suspected pulmonary embolism. BMJ 2005; 331: 259.
  • 16
    Stein PD, Fowler SE, Goodman LR, Gottschalk A, Hales CA, Hull RD, Leeper KV Jr, Popovich J Jr, Quinn DA, Sos TA, Sostman HD, Tapson VF, Wakefield TW, Weg JG, Woodard PK. Multidetector computed tomography for acute pulmonary embolism. N Engl J Med 2006; 354: 231727.
  • 17
    Courtney DM, Miller C, Smithline H, Klekowski N, Hogg M, Kline JA. Prospective multicenter assessment of interobserver agreement for radiologist interpretation of multidetector computerized tomographic angiography for pulmonary embolism. J Thromb Haemost 2010; 8: 5339.
  • 18
    Blackmon KN, Florin C, Bogoni L, McCain JW, Koonce JD, Lee H, Bastarrika G, Thilo C, Costello P, Salganicoff M, Joseph Schoepf U. Computer-aided detection of pulmonary embolism at CT pulmonary angiography: can it improve performance of inexperienced readers? Eur Radiol 2011; 21: 121423.
  • 19
    Costantino G, Norsa AH, Amadori R, Ippolito S, Resta F, Bianco R, Casazza G, Biagiotti S, Rusconi AM, Montano N. Interobserver agreement in the interpretation of computed tomography in acute pulmonary embolism. Am J Emerg Med 2009; 27: 110911.
  • 20
    Mos IC, Douma RA, Erkens PMG, Nizet TAC, Durian MF, Hovens MMM, van Houten AA, Hofstee HM, Kooiman J, Klok FA, ten Cate H, Ulmann MV, Buller HR, Kamphuisen PW, Huisman MV. Diagnostic safety of a stuctured algorithm with use of clinical decision rule, D-dimer and CT scan for clinically suspected recurrent pulmonary embolism. J Thromb Haemost 2011; 9: 304.
  • 21
    Perrier A, Desmarais S, Miron MJ, de Moerloose P, Lepage R, Slosman D, Didier D, Unger PF, Patenaude JV, Bounameaux H. Non-invasive diagnosis of venous thromboembolism in outpatients. Lancet 1999; 353: 1905.
  • 22
    Torbicki A, Perrier A, Konstantinides S, Agnelli G, Galie N, Pruszczyk P, Bengel F, Brady AJ, Ferreira D, Janssens U, Klepetko W, Mayer E, Remy-Jardin M, Bassand JP. Guidelines on the diagnosis and management of acute pulmonary embolism: the Task Force for the Diagnosis and Management of Acute Pulmonary Embolism of the European Society of Cardiology (ESC). Eur Heart J 2008; 29: 2276315.
  • 23
    Aujesky D, Roy PM, Verschuren F, Righini M, Osterwalder J, Egloff M, Renaud B, Verhamme P, Stone RA, Legall C, Sanchez O, Pugh NA, N’Gako A, Cornuz J, Hugli O, Beer HJ, Perrier A, Fine MJ, Yealy DM. Outpatient versus inpatient treatment for patients with acute pulmonary embolism: an international, open-label, randomised, non-inferiority trial. Lancet 2011; 378 (9785): 418
  • 24
    Kabrhel C, Matts C, McNamara M, Katz J, Ptak T. A highly sensitive ELISA D-dimer increases testing but not diagnosis of pulmonary embolism. Acad Emerg Med 2006; 13: 51924.
  • 25
    Anderson DR, Kahn SR, Rodger MA, Kovacs MJ, Morris T, Hirsch A, Lang E, Stiell I, Kovacs G, Dreyer J, Dennie C, Cartier Y, Barnes D, Burton E, Pleasance S, Skedgel C, O’Rouke K, Wells PS. Computed tomographic pulmonary angiography vs ventilation-perfusion lung scanning in patients with suspected pulmonary embolism: a randomized controlled trial. JAMA 2007; 298: 274353.