Prevalence of Deep Vein Thrombosis in Hospitalized Patients With Suspected Pulmonary Embolism Ruled Out by Multislice CT Angiography

Abstract

Current evidence suggests that for outpatients with suspected pulmonary embolism (PE), multislice computed tomographic angiography (CTPA) is sufficient to rule out PE. However, the accuracy of CTPA alone has not been established for hospitalized patients. Our goal was to determine the prevalence of deep vein thrombosis (DVT) in hospitalized patients who had PE ruled out by CTPA. We conducted a prospective cohort study of patients who developed symptoms indicative of PE, after being admitted to the hospital for any reason other than PE and were evaluated with multislice CTPA. The main outcome was proximal DVT. Between November 2011 and December 2014, 191 hospitalized patients were screened. A total of 99 patients satisfied our inclusion criteria. The average length of hospitalization for this group was 14 days (range: 2-127 days). While hospitalized, 54 (28%) patients underwent a major surgical procedure and 80 (79%) were receiving thromboprophylaxis. Of the 99 patients included, 7 (7.07%; 95% confidence intervals [CIs]: 3.4-13.8) were diagnosed with a proximal DVT. The likelihood of developing a proximal DVT was higher for those with subtle and nonspontaneously reported symptoms of DVT, odds ratio [OR] was 50.93 (95% CI: 5.35-2572) and for those classified as PE likely OR was 37.54 (95% CI: 4.05-186.1). Given the prevalence of DVT in hospitalized patients with suspected PE ruled out by a negative multislice CTPA, our study suggests that compression ultrasonography would, in fact, be justified for patients with these characteristics.

Keywords

deep vein thrombosis vein thrombosis pulmonary embolism suspect pulmonary embolism prevalence venous thromboembolism epidemiology

Introduction

The utility of compression ultrasonography (CUS) in ruling out proximal deep vein thrombosis (DVT) associated with suspected pulmonary embolism (PE) remains a matter of debate.¹ Current evidence shows that for outpatients presenting to emergency departments, suspected PE can be safely ruled out by multislice computed tomographic angiography (CTPA) with no need for leg CUS.^2
–4 However, to date there is a paucity of evidence to show if the same holds true for hospitalized patients with suspected PE^2,5 and that a normal multidetector CTPA result alone, may not safely exclude PE in patients with a high clinical pretest probability.⁶ Based on the limited amount of available data, we estimated the prevalence of DVT in hospitalized patients with suspected PE ruled out by multislice CTPA to be approximately 6% (95% confidence intervals [95% CI]: 1.6-16.1). This estimate suggests that CUS is needed to rule out DVT in hospitalized patients with suspected PE.⁷ The purpose of this study was to provide a more accurate estimate of the rate of proximal DVT in hospitalized patients with suspected PE ruled out with multislice CTPA.

Methods

Ethics

The study was approved by the ethics board of Hospital Italiano de Buenos Aires. The ethics board of Hospital Italiano de Buenos Aires approved the use of verbal consent to enroll patients in the study. Current legislation accepts verbal consent for observational studies. Verbal consent was obtained for all participants and recorded on a standardized form by the research coordinator. Medical records missing documentation of verbal consent were not reviewed. All records and information were rendered anonymous prior to analyzes.

Design

We conducted a prospective cohort study approved by the ethics board of the Hospital Italiano de Buenos Aires. Hospital Italiano de Buenos Aires is an academic hospital with 650 beds. In a prior analysis of 48 patients with suspected PE ruled out by multislice CTPA, 3 had a proximal DVT confirmed by CUS (6.25%; 95% CI: 1.6-16.1). We calculated that an additional 52 patients were needed to determine if the rate of proximal DVT in hospitalized patients with suspected PE was lower than 2.5% when PE was ruled out by multislice CTPA (assuming that the real prevalence of DVT was >6.5%). We selected 2.5% as the lower bound of the CI as it is a commonly accepted measure for studies evaluating diagnostic strategies for venous thromboembolism (VTE).^1

–4

Study Population and Intervention

Patients were eligible to participate in the study if (1) they were able to provide consent, (2) they presented with symptoms suggestive of PE (defined as acute onset of new or worsening shortness of breath, chest pain, hemoptysis, presyncope, or syncope), (3) symptoms developed >48 hours after admission in medically ill patients or >24 hours after admission following a major surgical procedure (defined as an anesthesia time >90 minutes), and (4) PE was ruled out after multislice CTPA. All scans were done with a 64-multislice CT scanner (Aquilion; Toshiba, Japan). Patients were excluded if they had symptoms compatible with PE prior to admission, had symptoms of DVT without symptoms of PE, refused to provide informed consent, refused bilateral leg CUS, or had a prior history of VTE. Within 24 hours of providing verbal consent, all patients received a bilateral whole-leg CUS.

Outcomes

The main outcome was proximal DVT (defined as noncompressible segment above the trifurcation of the popliteal vein). The secondary outcome included the prevalence of proximal DVT in patients with symptoms of DVT or in those classified as high probability for PE (ie, those with a Wells score greater than 4 points). Pulmonary embolism was considered ruled out when the multislice CTPA showed no filling defects in a segmental or larger pulmonary artery (single subsegmental embolus was considered as a PE only if the CUS was positive for a proximal DVT). Patients diagnosed with PE based on CTPA were excluded from the study.

Analysis and Statistics

Proportions, 95% CI, χ² test with Yates correction, Mid-P exact test, and Student t tests were used. Odds ratios and their respective 95% CI were calculated. All analyzes were performed using Stats Direct.

Results

Between November 2011 and December 2014, 191 hospitalized patients were screened. Excluded from the study were 66 patients with confirmed PE, 14 patients with a history of prior VTE, and 12 who did not complete the assigned CUS. A total of 99 patients satisfied our inclusion criteria and were enrolled.

The cohort consisted of 38 males and 61 females. The number of those with cancer was 41. The mean age was 65.5 (standard deviation 18). The average length of hospitalization was 14 days (range: 2-127 days). During their admission, 54 patients had a major surgery and 80 were receiving thromboprophylaxis (70 were on low-molecular-weight heparin, 7 unfractionated, 2 on vitamin K antagonists, and 1 on mechanical thromboprophylaxis). The initial symptoms triggering the investigation for PE were dyspnea in 77.7%, followed by pleuritic chest pain in 15.2%. The mean duration of immobilization (more than 90% of the time in bed) was 12.1 days (range: 0-90). Fourteen (14.3%) patients presented subtle symptoms suggestive of DVT, which were not reported spontaneously, but to the directed and exhaustive questioning. The average Wells score in patients without symptoms of DVT was 2.6 versus 5.1 in those with DVT symptoms (P value <.001). Baseline characteristics are presented in Table 1.

Table 1.

Baseline Characteristics of the Study Population.

	%
Age (SD)	65.5 (18.1)
Female	69
Active cancer	41
Surgery during admission	54
Cancer surgery during admission	18
General surgery (not associated with cancer)	26
Major orthopedic surgery	10
Infections	18
COPD exacerbation	10
Congestive heart failure	10
Stroke	7
Classified as PE likely (>4 points)	18
Classified as PE unlikely (≤4)	79
DVT prophylaxis at time of evaluation	81
Days fully immobilized (SD)	12 (18)
Most proximal location of the DVT
Iliac	0
Common femoral segment	57
Superficial femoral segment	14
Popliteal segment	28

Abbreviations: COPD, chronic obstructive pulmonary disease; DVT, deep vein thrombosis; PE, pulmonary embolism; SD, standard deviation.

Of the 99 patients included, 7 (7.1%; 95% CI: 3.4-13.8) were diagnosed with a proximal DVT. Of these, 1 had an extensive DVT extending from the common femoral segment to the popliteal vein and 2 had an isolated popliteal DVT.

The prevalence of proximal DVT in patients without signs or symptoms of DVT was 1/84 (1.2%; 95% CI: 0.0-6.9). In contrast, the prevalence of DVT in patients with subtle symptoms was 6/14 (42.86%; 95% CI: 21.3-67.4) and OR was 50.93 (5.3-2572). The incidence of proximal DVT in patients classified as PE likely was 6/18 (33.3%; 95% CI: 15.7-56.9), whereas, in those classified as PE unlikely, the incidence was 1/79 (1.3%; 95% CI: 0.06-6.2) and OR was 37.54 (95% CI: 4.1-186.1). No other variables were associated with the presence of DVT (see Table 2).

Table 2.

Variables Associated With the Diagnosis of DVT.

	OR (95% CI)	P Value
Age	0.98 (0.9-1.2)	.38
Cancer	1.06 (0.2-5.0)	.76
Female	1.53 (0.2-8.3)	.56
Surgery during admission	1.12 (0.2-5.2)	.88
Use of thromboprophylaxis	0.60 (0.1-3.3)	.55
Days of hospitalization	0.99 (0.9-1.0)	.79
Signs or symptoms of DVT	55.33 (5.9-2572)	.0004
Wells score greater than 4 points	37.54 (4.1-186.1)	.001

Abbreviations: CI, confidence intervals; DVT, deep vein thrombosis; OR, odds ratio.

Discussion

Our study suggests that in contrast to the outpatient population, the prevalence of proximal DVT in hospitalized patients with suspected PE ruled out by negative multislice CTPA is high. Therefore, bilateral leg CUS seems appropriate for this population (especially for patients with subtle or not spontaneously reported symptoms of DVT or patients classified as having a high clinical probability of PE using the Wells score). Our study was based on the assumption that patients with suspected PE ruled out by multislice CTPA do not require a bilateral leg CUS even if symptoms of DVT are present. To date, there are 4 large studies (3 prospective cohort studies and 1 randomized controlled study) supporting this approach for outpatients. Righini et al compared a strategy using d-dimer/CUS/CTPA versus d-dimer/CTPA to rule out PE (10% with symptoms of DVT) and found no difference in the rate of recurrent VTE in both arms—suggesting that CUS is not required for the outpatient population.³ A prospective study by Huisman et al evaluated 3306 (including some hospitalized) patients with suspected PE using d-dimer/CTPA and found that omitting bilateral CUS was a safe approach (5.6% had symptoms of DVT).² Perrier et al found that CUS was of no benefit to outpatients with suspected PE managed with d-dimer/CTPA (18% had symptoms of DVT).⁴ Douma et al¹³ evaluated outpatients who had suspected PE ruled out by d-dimer/CTPA according to clinical prediction rules. They found that omitting CUS was not associated with an increased risk of recurrent VTE (6% had symptoms of DVT).

Unlike prior studies, our study looks specifically at hospitalized patients. Omitting bilateral leg CUS in patients with PE ruled out by CTPA has not been adequately validated in patients who develop symptoms of PE during hospitalization. The high incidence of DVT in our cohort could be explained by the high number of patients who had cancer or recent surgery. The incidence of cancer or recent surgery for patients in our study was double that of prior studies.^{2
–4,8

–14} Van der Hulle et al¹⁴ recently reported outcomes for hospitalized patients from 2 prospective cohorts who had suspected PE ruled out by clinical prediction rules, d-dimer, and multislice CTPA. They concluded that leg CUS was unnecessary. In contrast, our study employed a stricter definition of “hospitalized.” That is, we excluded patients exhibiting PE symptoms prior to admission and we included more patients with comorbidities such as cancer, recent surgery, or immobilization. Similar to previous reports by Rosen et al¹⁵ and Wells et al,¹⁶ our study suggests that when PE is ruled out for patients who do not have risk factors for VTE or symptoms of DVT, the prevalence of proximal DVT is low. Also consistent with their findings is a high rate of proximal DVT among patients exhibiting symptoms of DVT—further supporting the utility of bilateral leg CUS in hospitalized patients after PE has been ruled with a multislice CTPA.^14,15,16

Geographical variation is another possible explanation for the high rate of DVT in our cohort. Current information regarding the efficacy and safety of diagnostic strategies for suspected PE in developing countries is limited.¹⁷ As previously suggested by Penaloza et al, current diagnostic approaches used to rule out PE have been validated in North America and therefore, may not apply in other regions around the world (eg, European countries) where the prevalence and severity of the disease are higher.¹⁸ Our study highlights the need for adequate validation of clinical prediction rules across different settings (both global and local) to account for variation in the patient population.

Our findings have limitations. First, as this was a single-center study, there exists a potential for practice bias regarding diagnostic approach. Second, the interpretation of CUS reports lacked rigorous blinding. Nevertheless, all reports were read by radiologists specializing in CUS. Third, patients classified as PE unlikely did not receive a d-dimer test before evaluation by multislice CTPA. Fourth, our study did not answer if a CUS of the symptomatic leg alone would be more effective than screening both legs. Fifth, DVT was very uncommon in patients with negative CTPA who had no symptoms of DVT and PE could be safely rule out. Finally, the presence of DVT in a patient with suspected PE confirms the diagnosis of VTE, and the patient requires at least 3 months of anticoagulant therapy.

In conclusion, given the high prevalence of DVT in hospitalized patients for whom PE is ruled out by a negative multislice CTPA (especially in those with symptoms of DVT or a high clinical probability of PE), a proximal bilateral leg CUS could be required to safely exclude DVT. This study does not prove the clinical importance of DVT found and that is why we strongly believe that this should be carefully assessed with a randomized controlled trail.

Footnotes

Authors’ Note

The results of this study were presented at the XXV Congress of the International Society of Thrombosis and Haemostasis in Toronto, Canada, 2015.

Acknowledgments

The authors would like to thank Maria Laura Martin and Alejandra Romero, who collaborated in collecting the data.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

References

Le Gal

Righini

. Controversies in the diagnosis of venous thromboembolism. J Thromb Haemost. 2015;(13 suppl 1): S259–S265.

van Belle

Büller

Huisman

; Writing Group for the Christopher Study Investigators. Effectiveness of managing suspected pulmonary embolism using an algorithm combining clinical probability, d-dimer testing, and computed tomography. JAMA. 2006;295(2):172–179.

Righini

Le Gal

Aujesky

. Diagnosis of pulmonary embolism by multidetector CT alone or combined with venous ultrasonography of the leg: a randomised non-inferiority trial. Lancet. 2006;371(9621):1343–1352.

Perrier

Roy

Sanchez

. Multidetector-row computed tomography in suspected pulmonary embolism. N Engl J Med. 2005;352(17):1760–1768.

Le Gal

Righini

Sanchez

. A positive compression ultrasonography of the lower limb veins is highly predictive of pulmonary embolism on computed tomography in suspected patients. Thromb Haemost. 2006;95(6):963–966.

Moores

Kline

Portillo

. Multidetector computed tomographic pulmonary angiography in patients with a high clinical probability of pulmonary embolism. J Thromb Haemost. 2016;14(1):114–120.

Vázquez

Posadas-Martínez

Sevilla

Giunta

Gándara

. The role of compression ultrasound in hospitalized patients with suspected pulmonary embolism. Thromb Res. 2014;134(5):1162–1163.

Wells

Anderson

Rodger

. Derivation of a simple clinical model to categorize patients probability of pulmonary embolism: increasing the models utility with the SimpliRED D-dimer. Thromb Haemost. 2000;83(3):416–420.

Stein

Fowler

Goodman

; PIOPED II Investigators. Multidetector computed tomography for acute pulmonary embolism. N Engl J Med. 2006;354(22):2317–2327.

10.

Righini

Le Gal

Aujesky

. Complete venous ultrasound in outpatients with suspected pulmonary embolism. J Thromb Haemost. 2009;7(3):406–412.

11.

Elias

Colombier

Victor

. Diagnostic performance of complete lower limb venous ultrasound in patients with clinically suspected acute pulmonary embolism. Thromb Haemost. 2004;91(1):187–195.

12.

Anderson

Kovacs

Dennie

. Use of spiral computed tomography contrast angiography and ultrasonography to exclude the diagnosis of pulmonary embolism in the emergency department. J Emerg Med. 29(4):399–404.

13.

Douma

Mos

Erkens

; Prometheus Study Group. Performance of 4 clinical decision rules in the diagnostic management of acute pulmonary embolism: a prospective cohort study. Ann Intern Med. 2011;154(11):709–718.

14.

van der Hulle

den Exter

Mos

. Optimization of the diagnostic management of clinically suspected pulmonary embolism in hospitalized patients. Br J Haematol. 2014;167(5):681–686.

15.

Rosen

Sheiman

Weintraub

McArdle

. Compression sonography in patients with indeterminate or low-probability lung scans: lack of usefulness in the absence of both symptoms of deep-vein thrombosis and thromboembolic risk factors. AJR Am J Roentgenol. 1996;166(2):285–289.

16.

Wells

Ginsberg

Anderson

. Utility of ultrasound imaging of the lower extremities in the diagnostic approach in patients with suspected pulmonary embolism. J Intern Med. 2001;250(3):262–264.

17.

Carrier

Righini

Wells

. Subsegmental pulmonary embolism diagnosed by computed tomography: incidence and clinical implications. A systematic review and meta-analysis of the management outcome studies. J Thromb Haemost. 2010;8(8):1716–1722.

18.

Penaloza

Kline

Verschuren

. European and American suspected and confirmed pulmonary embolism populations: comparison and analysis. J Thromb Haemost. 2012;10(3):375–381.