## Abstract

Only 12 to 25% of patients suspected of clinical Venous Thromboembolism (VTE) are confirmed by an image study. This low efficiency in diagnosis adds a high cost to each VTE evidenced Academically it is recommended, before performing any test, to proceed to a clinical stratification between likely or unlikely VTE following explicit criteria. In this way the above prevalence is changed, increasing the test prediction power. However in general clinical practice, outside of clinical trials, these explicit rules are rarely employed. We have recently showed, by a management survey, that a normal D-Dimer value could reject with safety 22.34 % of suspicions of VTE, independently of the pretest clinical probability, whenever a high sensitivity assay (HSA) is used for D-Dimer measurement (VIDAS Exclusion, Bio-Merieux, France). So we have found, between December 2007 to May 2008, only one patient with Deep Vein Thrombosis (DVT) and a normal D-Dimer value, from 410 normal D-Dimer detected, in 1835 orders realized (Patients in anticoagulant treatment were excluded). D-Dimer values were normal in 20.2 % of patients with a suspected DVT and in 23.5 % suspicions of Pulmonary Embolism (PE) (fractions (F) of 0.202 and 0.235 respectively). The systematically use of D-Dimer measurement, as first step in the diagnosis of VTE, could be expensive, with a vast cost in laboratory determination which could be superior to the cost in saving radiological explorations. To explore this problem we have computed the cost-benefit utility for this approach. There are internationally wide cost differences between explorations used to detect an episode of VTE: D-Dimer or Radiological exploration (CT, Ultrasounds, V/Q Lung Scan, Angiography). Therefore to compute the cost-benefit, we need to define a ratio (R=D/I) between the Hospital Service purveyor price of the D-Dimer assay (D) and imaging exploration cost (I) (If two different radiological exploration are indistinctly used, we use the sum of weighted cost for each exploration). If R is less than F the use of D-Dimer measurement to all VTE suspicions is worthwhile. These computed ratios (R) assuming only one exploration by patient, from our hospital are compared with another European hospital, two USA hospitals and another one from Canada, the results are showed in the table

UltrasoundsV/Q Lung ScanCTD-Dimer
Cost ($)RCost ($)RCost ($)RCost ($)
Barcelona (Spain) 161 0.052 203 0.041 407 0.021 8.4
Geneva (Switzerland) 90 0.366 301 0.109 600 0.055 33
USA # 1 95.19 0.103 309.06 0.032 9.86
USA # 2 69 0.174 683 0.017 135 0.088 12
Montreal(Canada) 36 0.333 69 0.174 138 0.087 12
UltrasoundsV/Q Lung ScanCTD-Dimer
Cost ($)RCost ($)RCost ($)RCost ($)
Barcelona (Spain) 161 0.052 203 0.041 407 0.021 8.4
Geneva (Switzerland) 90 0.366 301 0.109 600 0.055 33
USA # 1 95.19 0.103 309.06 0.032 9.86
USA # 2 69 0.174 683 0.017 135 0.088 12
Montreal(Canada) 36 0.333 69 0.174 138 0.087 12

Therefore, as well in Barcelona as in the USA, the proposed strategy would be cost-benefit useful for DVT diagnosis, but not in Geneva or Monreal; however it is worthwhile to exclude PE in all hospitals studied. The saving magnitude in dollars, for each VTE detected, could be computed from the formula (FxI−D)/P; where P=VTE prevalence/100 (P=0.248 for DVT and P=0.129 for PE in our study), So, in our Hospital, the saving cost is 97.3$for each DVT diagnosed and it is 304.7$ for each PE confirmed, if V/Q Lung Scan is used or it is 676.3\$ if CT is employed. Old age and cancer are two conditions believed to have increased levels of D-Dimer. In our study, only patients over age 90 had a low probability (3.1 %) of normal values. On the other hand, a normal D-Dimer was evidenced in 138 out of 538 patients with cancer. Conclusion: D-Dimer determination by a HSA to all patients with suspicion of VTE, independent of a pretest clinical evaluation, could reject one out of every four or five patients, with safety for the patient and an important cost reduction in image explorations, provided that the cost ratio was inferior to the prevalence of normal D-Dimer existing in the tested population.

Disclosures: No relevant conflicts of interest to declare.

## Author notes

Corresponding author