Asymptomatic deep vein thrombosis (DVT) and pulmonary embolism are leading causes of morbidity following the hospitalization of elderly people. The diagnosis of DVT is supported by the D-dimer laboratory assay. The concentration of D-dimer increases in patients with DVT, but may be high in other conditions too (i.e. cancer, infections and inflammation). Old age coincides with a physiological increase in D-dimer values, and that is why D-dimer assay in the elderly is characteristically highly sensitive but scarcely specific. The aim of our study was to explore the reliability of different D-dimer cutoffs for the diagnosis of asymptomatic DVT in a population of bedridden hospitalized elderly patients. We studied 199 patients who were a mean 86.3 ± 6.7 years old. All participants underwent lower limb Doppler ultrasound (DUS) and D-dimer venous blood sampling on admission. In our cohort, the usual cutoff proved highly sensitive (100%), but its specificity was very poor (20.1%). To find a higher cutoff that could improve the method's specificity, we analyzed our data using a receiver operating characteristic curve analysis. The resulting D-dimer cutoff of 492 μg/l enabled us to retain the same sensitivity while improving the test's specificity to 39.1%, with a consequent improvement in its positive predictive value and accuracy. In addition to improving the method's reliability, this result may be helpful in clinical practice, in both medical wards and nursing homes. By adopting a cutoff of 492 μg/l, clinicians could significantly increase the proportion of older patients in whom DVT can be safely ruled out, reducing referrals for DUS and administration of heparin, with consequent clinical, practical and economic advantages.