Diagnostic strategy in deep-vein thrombosis

Abstract

Philip S Wells and co-workers (Dec 20/27, p 1795)1Wells PS Anderson DR Bormanis J et al.Value of assessment of pretest probability of deep-vein thrombosis in clinical management.Lancet. 1997; 350: 1795-1798Summary Full Text Full Text PDF PubMed Scopus (970) Google Scholar point out the importance of the calculation of pretest probability in the management of deepvein thrombosis. They outline 5·6% and 1·6% of false-negative ultrasound results in the high and moderate pretest probability group, respectively, and 0·6% of false-positive ultrasound results in the low-risk group. But the findings of ultrasonography in the low probability group were normal for 97% of the patients. For patients with a moderate or high pretest probability with a first normal ultrasound, 98% of follow-up ultrasounds and 76% of venographies were normal or inconclusive. To reduce the cost and time necessary, especially in emergency, D-dimer measurement, a non-invasive biological test, should be introduced into the diagnosis algorithm.2Grippa L D'Angelo SV Tomassini L et al.The utility and cost-effectiveness of D-dimer measurements in the diagnosis of deep vein thrombosis.Haematologica. 1997; 82: 446-451PubMed Google Scholar D-dimers are specific degradation products of crosslinked fibrin, and a low concentration is supposed to exclude venous thromboembolism (VTE).3Bounameaux H de Moerloose P Perrier A Reber G Plasma measurement of D-dimer as diagnostic aid in suspected venous thromboembolism: an overview.Thromb Haemostas. 1994; 71: 1-6PubMed Google Scholar To assess D-dimer under routine conditions, we conducted a prospective study of 200 consecutive patients admitted in our emergency department with suspected VTE. D-dimer measurements were by ELISA with a cutoff of 500 mg/L. The test results are available after 1 h but were not revealed until the end of the study. The emergency physician in charge of the patient assessed the pretest probability by marking a tick on a 10 cm horizontal visual analogue scale. Clinical probability was judged to be low if the tick was below 3·3 cm, moderate at 3·4–6·6 cm, and high at 6·7 cm or greater. The judgment criterion was the diagnosis at hospital discharge. We calculated 95% CI values of the negative predictive value of D-dimer according to the pretest probability by extrapolating the prevalence of VTE in each pretest probability group from the overall population. Four patients were excluded (inclusion criteria or D-dimer testing not available). VTE was confirmed in 54 (28%) patients and all had a positive D-dimer test. For the 142 patients without VTE, 38 (27%) had a negative D-dimer test. Sensitivity of D-dimer testing was 100%, specificity 27%, positive predictive value 34%, and negative predictive value 100% (95% CI 91–100). The clinical pretest probability was available for 87 patients and the rate of VTE was 13% (3/23), 24% (9/38), and 50% (13/26) in the low, moderate, and high pretest probability groups, respectively. The CI of the negative predictive value of D-dimer according to the pretest probability was: low probability (92–100%), moderate (91–100%), and high (87–100%). Our results confirm that D-dimer can be included on the clinical pretest probability. In case of low pretest probability, a negative D-dimer test seems sufficient to rule out deep-vein thrombosis. For moderate or high pretest probability when ultrasonography is negative, follow-up ultrasound or venography may be restricted for patients with a positive D-dimer test. This diagnostic strategy with D-dimer could avoid the need for ultrasonography or venography with the same level of safety.