Our patient had bilateral subpleural nodules, normal D-dimer level and, thus, a low clinical prob­ability of PTE that was estimated using the Geneva score [6]. Her PTE occurred shortly after open lung biopsy, most probably because of a combined ef­fect of a congenitally hypercoagulable state, and the surgical procedure due to the presence of a heterozygous FII G20210A mutation.

The radiologic appearance of pulmonary in­farction may vary depending on the underlying cause and the temporal evolution of the lesion [7]. Subpleural nodular opacities on chest radiography are atypical parenchymal infarctions, and surgical lung biopsy may sometimes be required to establish a correct diagnosis. Parambil et al. [3] reported that 65% of initially unrecognized pulmonary infarc­tions, which were later confirmed by histological examination, presented as solitary or multiple lung nodules on chest radiograph and/or CT.

In several studies, the negative predictive value of the combination of low clinical probability and normal D-dimer levels in exclusion of PTE ap­proached 99.5% [8,9]. For that reason, suspicion of PTE in our patient was initially abandoned and a lung biopsy was performed. A normal level of D-dimer, despite proven thromboembolic pulmonary infarctions on histologic examination, represented an intriguing issue. Normal D-dimer levels, when determined by the ELISA technique, have excel­lent negative predictive values of 91-98%, and are considered adequate for exclusion of venous thromboembolism [10,11]. Similarly, normal D­-dmer level determined by the latex agglutination method, has negative predictive value of about 88%, and is not satisfactory for exclusion of recent PTE [12]. Although the rapid immunoturbidomet-ric test that we used exhibited better sensitivity than the qualitative latex agglutination tests, its negative predictive value was lower than that of the ELISA method, which may be responsible for the negative results of D-dimer in our patient [12]. Furthermore, fibrin degradation products have a half-life of from 3-48 hours, and cause the D-dimer level to decrease over time [13]. Therefore, the long period between occurrence of symptoms and drawing of the blood sample may also be responsible for the normal D-dimer levels in our patient.

While the assays now available for D-dimer measurement have different sensitivities and speci­ficities for the diagnosis of venous thromboembo­lism [8,9,12], the D-dimer level is also affected by a patient's characteristics. For example, the value of D-dimer testing is significantly decreased if the renal function is affected, as study of patients with suspected pulmonary embolism and decreased glomerular filtration rate showed [14]. Gibson et al. [15] described cases when PTE was present in spite of a low D-dimer level, suggesting that pa­tients with a likely clinical probability should un­dergo further diagnostic testing. The ventilation/ perfusion scintigraphy is non diagnostic in almost one-half to two-thirds of patients with a clinical suspicion of PTE [16], and its predictive value de­creases if the clinical probability of PTE is low, as in our patient.

The relative risk of venous thromboembolic events is increased 2-4-fold in heterozygous carri­ers of the FII G20210A mutation when compared to non carriers [17]. The risk of venous thrombo­sis was significantly increased in patients with FII G20210A mutation in whom total hip arthroplasty was performed, compared to patients without this mutation [18]. Even a minor surgical procedure in patients with prothrombotic states may represent a risk factor for thrombosis [19].