0
Original Research: PULMONARY EMBOLISM |

The Importance of Clinical Probability Assessment in Interpreting a Normal d-Dimer in Patients With Suspected Pulmonary Embolism FREE TO VIEW

Nadine S. Gibson, PhD; Maaike Sohne, PhD; Victor E. A. Gerdes, PhD; Mathilde Nijkeuter, PhD; Harry R. Buller, PhD
Author and Funding Information

*From the Department of Vascular Medicine (Drs. Gibson, Sohne, Gerdes, and Buller), Academical Medical Center, Amsterdam, the Netherlands; and the Department of General Internal Medicine (Dr. Nijkeuter), Leiden University Medical Center, Leiden, the Netherlands.

Correspondence to: Nadine Gibson, PhD, Department of Vascular Medicine, Academical Medical Center, Meibergdreef 9, 1105 AZ Amsterdam, the Netherlands; e-mail: n.s.gibson@amc.uva.nl


The authors have reported to the ACCP that no significant conflicts of interest exist with any companies/organizations whose products or services may be discussed in this article.

Reproduction of this article is prohibited without written permission from the American College of Chest Physicians (www.chestjournal.org/misc/reprints.shtml).


Chest. 2008;134(4):789-793. doi:10.1378/chest.08-0344
Text Size: A A A
Published online

Background:  The d-dimer test is widely applied in the diagnostic workup of patients with suspected pulmonary embolism (PE). The objective of this study was to investigate how often the d-dimer test fails when clinical probability is not taken into account.

Methods:  We used data collected in 1,722 consecutive patients with clinically suspected PE to analyze the 3-month venous thromboembolism (VTE) rate in all patients with a normal d-dimer concentration and separately for patients who have a normal d-dimer concentration with an unlikely or likely clinical probability for PE, as assessed by the Wells clinical decision rule.

Results:  The 3-month VTE rate in all patients with a normal d-dimer concentration (n = 563) was 2.3% (95% confidence interval [CI], 1.4 to 3.9%). In the patients with an unlikely probability of PE (n = 477), VTE was confirmed in 1.1% of the patients with a normal d-dimer concentration (95% CI, 0.4 to 2.4%). In those patients with a likely clinical probability of PE (n = 86), VTE was confirmed in 9.3% of the patients with a normal d-dimer concentration (95% CI, 4.8 to 17.3%). The difference in VTE incidence between patients with unlikely and likely clinical probabilities of PE was significant (p < 0.001).

Conclusions:  Our findings indicate that it is of utmost importance to first examine the patient and assess the clinical probability, after which the d-dimer concentration can be taken into account, in order to prevent physicians from being influenced by a normal d-dimer test result when they evaluate the clinical probability of PE. Patients with a likely clinical probability should undergo further testing, regardless of the d-dimer test outcome.

In the last decade, d-dimer testing has gained widespread popularity for excluding pulmonary embolism (PE) and deep venous thrombosis (DVT), mainly as a result of its noninvasive character and high negative predictive value. For the appropriate implementation of the d-dimer test in clinical practice, however, it is necessary that the d-dimer test is used in combination with a clinical pretest probability assessment, since both diseases can be safely excluded in the case of an unlikely clinical probability test result together with a normal d-dimer test result.13 In patients with a likely clinical probability, a normal d-dimer test result cannot be used to exclude PE, and additional imaging testing is necessary.3

Although this strategy is recommended in several guidelines, in daily clinical practice physicians are often influenced by a normal d-dimer.46 Due to logistic advantages in a hectic emergency room setting the d-dimer test result is often available before a physician examines the patient. As a consequence, the physician might be influenced and would tend to decide that the clinical probability for venous thromboembolism (VTE) is low. An even bigger risk arises when the d-dimer test is used as a stand-alone test in making management decisions.

Based on the data collected in a large diagnostic management study1 in consecutive patients with suspected PE, we first assessed the negative predictive value of the d-dimer test when it is used as a stand-alone test, and, second, we compared the negative predictive value of the d-dimer test in patients with an unlikely clinical probability for PE to those patients with a likely clinical probability for PE. Finally, we studied the clinical characteristics of those patients with a false-normal d-dimer test result.

Data were obtained from a prospective diagnostic management study1 that included 3,306 patients with clinically suspected PE who were enrolled between November 2002 and August 2004 in 12 hospitals in the Netherlands. The institutional review boards of all participating hospitals approved the study protocol. Exclusion criteria were patients who had received low-molecular-weight heparin (LMWH) for > 24 h, age < 18 years, pregnancy, a known hypersensitivity to iodinated contrast fluid, a life expectancy of < 3 months, a geographic inability for follow-up, or informed consent had not been obtained.

The main results have been published previously.1 Briefly, the primary findings indicated that a diagnostic management strategy using a clinical decision rule (CDR), a d-dimer test, and a spiral CT scan is effective and safe in the workup of patients with clinically suspected PE.

At presentation, all patients underwent clinical probability assessment with the dichotomous CDR according to Wells et al7 and Gibson et al8 (Table 1). PE was considered to be unlikely if the CDR score was ≤ 4 and was likely if the CDR score was > 4. In those patients with an unlikely clinical probability, a d-dimer test was performed (Tinaquant; Roche Diagnostica; Mannheim, Germany; or Vidas D-Dimer; Biomerieux; Marcy l'Etoile, France), and the test result was defined as normal if the concentration was ≤ 0.5 mg/L. It was considered that the combination of PE being unlikely and a normal d-dimer test result ruled out PE, and anticoagulant treatment was withheld. In all other patients in whom PE was considered likely or in patients with an abnormal d-dimer test result, a spiral CT scan was performed. In 7 of the 12 participating hospitals, a d-dimer assay was performed in all patients, and the data obtained from these patients form the basis for the present analysis.

Table Graphic Jump Location
Table 1 Clinical Decision Rule According to the Criteria of Wells et al7 and Gibson et al8*

*The clinical probability of PE unlikely, ≤ 4 points; clinical probability of PE likely, > 4 points.

Follow-up, which consisted of a hospital visit or a telephone interview at 3 months and the instruction to contact the study center or their general practitioner in case of complaints suggestive of VTE, either DVT or PE, was performed in all patients. In case of clinically suspected DVT or PE in the follow-up period, a compression ultrasound for suspected DVT and a ventilation-perfusion scintigraphy scan or spiral CT scan for suspected PE was required to confirm or refute the diagnosis. In case of death, information was obtained from the general practitioner, from the hospital records, or from autopsy. The 3-month VTE failure rate was defined as either PE at baseline, or symptomatic objectively confirmed VTE in the 3-month follow-up period. All outcomes were blindly assessed by an independent adjudication committee.

For this analysis, we compared the 3-month VTE rate when the d-dimer test used as a stand-alone test to rule out PE, for the total study group, as well as separately for patients with an unlikely clinical probability and for those with a likely clinical probability for PE. Finally, we assessed specific clinical characteristics such as age, anticoagulant usage, the d-dimer assay used, the d-dimer level, and the localization of the thrombi in the patients with a normal d-dimer concentration and a likely clinical probability for PE, but with confirmed VTE. We compared these to the characteristics of the total study population.

Patients

Of the 1,825 consecutive patients from the seven hospitals that were eligible for inclusion in the study, 103 patients (5.6%) were excluded because of predefined criteria or due to the inability to obtain informed consent. The clinical probability assessment using the CDR of Wells et al7 was completed in all 1,722 subjects, whereas d-dimer test results were available in 1,632 of these patients (95%). In 477 patients (28%) with an unlikely CDR score and a normal d-dimer concentration, spiral CT scanning could be withheld and PE was considered to be ruled out. The mean age of the study population was 54 years (age range, 20 to 100 years), and 78% of the population were outpatients (Table 2). The prevalence of PE at baseline in these 1,632 patients was 22%.

Table Graphic Jump Location
Table 2 Baseline Characteristics of the Study Population (n = 1,722)*

*Values are given as No. (%), unless otherwise indicated.

†Values are given as the mean (range).

‡Values are given as the median (IQR).

Safety of a Normal d-Dimer Concentration

A normal d-dimer test result was present in 563 of the 1,632 study patients (34%). Of these, 13 patients were objectively diagnosed to have PE, either at baseline or during the 3 months of follow-up. None of these patients had DVT during the 3-month follow-up period. Therefore, the 3-month VTE rate in all patients with a normal d-dimer concentration was 2.3% (95% confidence interval [CI], 1.4 to 3.9%) [Table 3].

Table Graphic Jump Location
Table 3 3-Month VTE Failure Rate Per Strategy To Exclude PE

The frequency of a normal d-dimer test result in those patients with an unlikely clinical probability for PE was 45% (n = 1,060). In the likely clinical probability group, this frequency was 15.2% (n = 566; p < 0.001 [difference in frequency]).

In patients with an unlikely probability of PE, VTE was confirmed in 5 of the 477 patients with a normal d-dimer concentration (1.1%; 95% CI, 0.4 to 2.4%) [Table 3]. Two of these patients, in whom the protocol was violated, had PE diagnosed at baseline (by a spiral CT scan that was requested by the treating physician, outside of the protocol), and in the remaining patients VTE occurred during the follow-up period.

Among the patients with a likely clinical probability, PE was confirmed in 8 of the 86 patients with a normal d-dimer concentration (9.3%; 95% CI, 4.8 to 17.3%) [Table 3]. Of these, seven patients were found at baseline and one was found during the 3-month follow-up period. The difference in the incidence of VTE between patients in the unlikely and likely clinical probability of PE categories was significant (p < 0.001).

False-Negative d-Dimer Test Result

Table 4 details the clinical characteristics of the eight patients with a likely clinical probability of PE and a normal d-dimer concentration who turned out to have PE. The mean age of these patients was 52 years (age range, 28 to 71 years) vs 54 years (age range, 20 to 100 years) in the total study cohort. Three of the 8 patients (38%; 95% CI, 14 to 69 years) had received one or more injections of LMWH vs 237 of 1,722 patients (14%; 95% CI, 12 to 16) in the total study population. False-negative d-dimer test results were observed using both d-dimer assays (ie, Tinaquant [Roche Diagnostica], four patients; or Vidas D-Dimer [Biomerieux], four patients).

Table Graphic Jump Location
Table 4 Main Characteristics of the Failures When d-Dimer Is Used as a Stand-Alone Test in Patients With a Likely Clinical Probability for PE*

*NDA = no data available.

†This patient had PE diagnosed during the first month of follow-up.

The localization of the emboli in six of the eight patients vs the total cohort was 17% (95% CI, 3 to 56) vs 28% (95% CI, 23 to 33) for central emboli, 33% (95% CI, 10 to 70) vs 48% (95% CI, 43 to 54) for segmental emboli, and 50% (95% CI, 19 to 81) vs 24% (95% CI, 19 to 29) for subsegmental emboli, respectively. The median d-dimer concentration for the eight patients was 0.41 mg/L (interquartile range [IQR], 0.32 to 0.45) vs 0.30 mg/L (IQR, 0.20 to 0.40), respectively, for all patients with a normal d-dimer concentration and a likely clinical probability for PE.

Clinical characteristics of the five patients with an unlikely clinical probability for PE and a false-negative d-dimer test result were as follows (Table 3). Two patients had received one or more injections of LMWH, and the median d-dimer concentration was 0.45 mg/L (IQR, 0.16 to 0.49). In the two patients in whom a spiral CT scan was performed at baseline, one had segmental localized emboli and one had a central embolus.

This study convincingly illustrates that relying on d-dimer testing alone carries an unacceptable risk if the clinical probability of PE is not taken into account. The a priori chance of having PE in consecutive patients with signs and symptoms suggestive for this disease is 20 to 25%.1,911 Our data show that in patients with a normal d-dimer concentration independent of the clinical probability of PE, the 3-month VTE risk is 2.3%, with an upper level of the 95% CI of almost 4%; whereas, in patients with a likely clinical probability of PE despite a normal d-dimer concentration, approximately 1 in 10 patients will still have PE. In contrast and in agreement with numerous publications, in patients with an unlikely clinical probability and a normal d-dimer concentration the risk of VTE, if the patient is untreated, is around 1%. Hence, clinicians need to realize that they should ignore a normal d-dimer concentration when the clinical probability is considered to be likely. The findings of this work support current guidelines and recommendations46 on the diagnostic approach to patients with suspected VTE.

It is difficult to identify indicators that could predict which patients are at risk for a false-negative d-dimer test result. In fact, a few previously suggested reasons for false-negative d-dimer test results, such as small emboli, symptoms existing longer than several days, and pretreatment with anticoagulant therapy, were only possible explanations in four of the eight patients.1214 The spiral CT scan showed major embolism in three of the eight patients, of whom one was even a centrally localized embolus. The complaints of all eight patients did not exist > 7 days, and finally it should be noted that five of the eight patients had a d-dimer concentration between 0.40 and 0.50, which indicates that not all failures had high-normal d-dimer test results. Taken together, it is unlikely that one can predict the risk of a false-negative d-dimer test result.

Some methodological aspects of our study require comment. First, we studied consecutive patients with baseline clinical characteristics similar to patients in comparable diagnostic management studies1517; hence, we believe that our observations are relevant for similar settings that investigate patients with suspected PE. Second, although the present study had a reasonable sample size of 1,722 patients, our conclusions about a normal d-dimer concentration and a likely clinical probability only involve 86 patients. Therefore, it should be noted that the observed frequency of the 3-month VTE failure rate in this patient group was 9.3% (95% CI, 4.3 to 17.3%). However, the lower boundary of the CI is generally considered to be clinically unacceptable for a strategy to rule out PE.18 A third aspect of our study is that the d-dimer test results were not available in 5% of the 1,722 patients. However, since the clinical characteristics of these 90 missing patients (data not shown) did not differ from the study population, we believe that this does not constitute a bias. Finally, although it is conceivable that the subjective element “alternative diagnosis is less likely than PE” of the CDR score may be influenced by a normal d-dimer test result, especially when used by relatively inexperienced physicians, this could not be investigated due to the study design.7

In conclusion, our findings indicate that it is of utmost importance to first examine the patient and assess the clinical probability, after which the d-dimer test result can be taken into account, which may prevent physicians from being influenced by a normal d-dimer test result when they evaluate the clinical probability. In patients with a likely clinical pretest probability of disease, d-dimer testing should not be performed. If testing has already been performed, the results should be ignored until further diagnostic testing has been obtained.

CDR

clinical decision rule

CI

confidence interval

DVT

deep venous thrombosis

IQR

interquartile range

LMWH

low-molecular-weight heparin

PE

pulmonary embolism

VTE

venous thromboembolism

van Belle A, Buller HR, Huisman MV, et al. Effectiveness of managing suspected pulmonary embolism using an algorithm combining clinical probability, d-dimer testing, and computed tomography. JAMA. 2006;295:172-179. [PubMed] [CrossRef]
 
Wells PS, Anderson DR, Rodger M, et al. Excluding pulmonary embolism at the bedside without diagnostic imaging: management of patients with suspected pulmonary embolism presenting to the emergency department by using a simple clinical model and d-dimer. Ann Intern Med. 2001;135:98-107. [PubMed]
 
Ten Cate-Hoek AJ, Prins MH. Management studies using a combination of d-dimer test result and clinical probability to rule out venous thromboembolism: a systematic review. J Thromb Haemost. 2005;3:2465-2470. [PubMed]
 
Roy PM, Meyer G, Vielle B, et al. Appropriateness of diagnostic management and outcomes of suspected pulmonary embolism. Ann Intern Med. 2006;144:157-164. [PubMed]
 
Kelly J, Hunt BJ. A clinical probability assessment and d-dimer measurement should be the initial step in the investigation of suspected venous thromboembolism. Chest. 2003;124:1116-1119. [PubMed]
 
Righini M, Aujesky D, Roy PM, et al. Clinical usefulness of d-dimer depending on clinical probability and cutoff value in outpatients with suspected pulmonary embolism. Arch Intern Med. 2004;164:2483-2487. [PubMed]
 
Wells PS, Anderson DR, Rodger M, et al. 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;83:416-420. [PubMed]
 
Gibson NS, Sohne M, Kruip MJ, et al. Further validation and simplification of the Wells clinical decision rule in pulmonary embolism. Thromb Haemost. 2008;99:229-234. [PubMed]
 
Kruip MJ, Slob MJ, Schijen JH, et al. Use of a clinical decision rule in combination with d-dimer concentration in diagnostic workup of patients with suspected pulmonary embolism: a prospective management study. Arch Intern Med. 2002;162:1631-1635. [PubMed]
 
Perrier A, Roy PM, Aujesky D, et al. 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;116:291-299. [PubMed]
 
Wells PS, Ginsberg JS, Anderson DR, et al. Use of a clinical model for safe management of patients with suspected pulmonary embolism. Ann Intern Med. 1998;129:997-1005. [PubMed]
 
Ray P, Bellick B, Birolleau S, et al. Referent d-dimer enzyme-linked immunosorbent assay testing is of limited value in the exclusion of thromboembolic disease: result of a practical study in an ED. Am J Emerg Med. 2006;24:313-318. [PubMed]
 
Kraaijenhagen RA, Wallis J, Koopman MM, et al. Can causes of false-normal d-dimer test [SimpliRED] results be identified? Thromb Res. 2003;111:155-158. [PubMed]
 
Cogo A, Lensing AW, Koopman MM, et al. Compression ultrasonography for diagnostic management of patients with clinically suspected deep vein thrombosis: prospective cohort study. BMJ. 1998;316:17-20. [PubMed]
 
Perrier A, Roy PM, Sanchez O, et al. Multidetector-row computed tomography in suspected pulmonary embolism. N Engl J Med. 2005;352:1760-1768. [PubMed]
 
Stein PD, Fowler SE, Goodman LR, et al. Multidetector computed tomography for acute pulmonary embolism. N Engl J Med. 2006;354:2317-2327. [PubMed]
 
Righini M, Le Gal G, Aujesky D, et al. Diagnosis of pulmonary embolism by multidetector CT alone or combined with venous ultrasonography of the leg: a randomised non-inferiority trial. Lancet. 2008;371:1343-1352. [PubMed]
 
Kruip MJ, Leclercq MG, van der HC, et al. Diagnostic strategies for excluding pulmonary embolism in clinical outcome studies: a systematic review. Ann Intern Med. 2003;138:941-951. [PubMed]
 

Figures

Tables

Table Graphic Jump Location
Table 1 Clinical Decision Rule According to the Criteria of Wells et al7 and Gibson et al8*

*The clinical probability of PE unlikely, ≤ 4 points; clinical probability of PE likely, > 4 points.

Table Graphic Jump Location
Table 2 Baseline Characteristics of the Study Population (n = 1,722)*

*Values are given as No. (%), unless otherwise indicated.

†Values are given as the mean (range).

‡Values are given as the median (IQR).

Table Graphic Jump Location
Table 3 3-Month VTE Failure Rate Per Strategy To Exclude PE
Table Graphic Jump Location
Table 4 Main Characteristics of the Failures When d-Dimer Is Used as a Stand-Alone Test in Patients With a Likely Clinical Probability for PE*

*NDA = no data available.

†This patient had PE diagnosed during the first month of follow-up.

References

van Belle A, Buller HR, Huisman MV, et al. Effectiveness of managing suspected pulmonary embolism using an algorithm combining clinical probability, d-dimer testing, and computed tomography. JAMA. 2006;295:172-179. [PubMed] [CrossRef]
 
Wells PS, Anderson DR, Rodger M, et al. Excluding pulmonary embolism at the bedside without diagnostic imaging: management of patients with suspected pulmonary embolism presenting to the emergency department by using a simple clinical model and d-dimer. Ann Intern Med. 2001;135:98-107. [PubMed]
 
Ten Cate-Hoek AJ, Prins MH. Management studies using a combination of d-dimer test result and clinical probability to rule out venous thromboembolism: a systematic review. J Thromb Haemost. 2005;3:2465-2470. [PubMed]
 
Roy PM, Meyer G, Vielle B, et al. Appropriateness of diagnostic management and outcomes of suspected pulmonary embolism. Ann Intern Med. 2006;144:157-164. [PubMed]
 
Kelly J, Hunt BJ. A clinical probability assessment and d-dimer measurement should be the initial step in the investigation of suspected venous thromboembolism. Chest. 2003;124:1116-1119. [PubMed]
 
Righini M, Aujesky D, Roy PM, et al. Clinical usefulness of d-dimer depending on clinical probability and cutoff value in outpatients with suspected pulmonary embolism. Arch Intern Med. 2004;164:2483-2487. [PubMed]
 
Wells PS, Anderson DR, Rodger M, et al. 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;83:416-420. [PubMed]
 
Gibson NS, Sohne M, Kruip MJ, et al. Further validation and simplification of the Wells clinical decision rule in pulmonary embolism. Thromb Haemost. 2008;99:229-234. [PubMed]
 
Kruip MJ, Slob MJ, Schijen JH, et al. Use of a clinical decision rule in combination with d-dimer concentration in diagnostic workup of patients with suspected pulmonary embolism: a prospective management study. Arch Intern Med. 2002;162:1631-1635. [PubMed]
 
Perrier A, Roy PM, Aujesky D, et al. 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;116:291-299. [PubMed]
 
Wells PS, Ginsberg JS, Anderson DR, et al. Use of a clinical model for safe management of patients with suspected pulmonary embolism. Ann Intern Med. 1998;129:997-1005. [PubMed]
 
Ray P, Bellick B, Birolleau S, et al. Referent d-dimer enzyme-linked immunosorbent assay testing is of limited value in the exclusion of thromboembolic disease: result of a practical study in an ED. Am J Emerg Med. 2006;24:313-318. [PubMed]
 
Kraaijenhagen RA, Wallis J, Koopman MM, et al. Can causes of false-normal d-dimer test [SimpliRED] results be identified? Thromb Res. 2003;111:155-158. [PubMed]
 
Cogo A, Lensing AW, Koopman MM, et al. Compression ultrasonography for diagnostic management of patients with clinically suspected deep vein thrombosis: prospective cohort study. BMJ. 1998;316:17-20. [PubMed]
 
Perrier A, Roy PM, Sanchez O, et al. Multidetector-row computed tomography in suspected pulmonary embolism. N Engl J Med. 2005;352:1760-1768. [PubMed]
 
Stein PD, Fowler SE, Goodman LR, et al. Multidetector computed tomography for acute pulmonary embolism. N Engl J Med. 2006;354:2317-2327. [PubMed]
 
Righini M, Le Gal G, Aujesky D, et al. Diagnosis of pulmonary embolism by multidetector CT alone or combined with venous ultrasonography of the leg: a randomised non-inferiority trial. Lancet. 2008;371:1343-1352. [PubMed]
 
Kruip MJ, Leclercq MG, van der HC, et al. Diagnostic strategies for excluding pulmonary embolism in clinical outcome studies: a systematic review. Ann Intern Med. 2003;138:941-951. [PubMed]
 
NOTE:
Citing articles are presented as examples only. In non-demo SCM6 implementation, integration with CrossRef’s "Cited By" API will populate this tab (http://www.crossref.org/citedby.html).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Find Similar Articles
CHEST Journal Articles
PubMed Articles
  • CHEST Journal
    Print ISSN: 0012-3692
    Online ISSN: 1931-3543