IN THIS ISSUE of the ARCHIVES, Meignan and associates1 observe that among 622 patients with proximal deep venous thrombosis (DVT) 32% to 45% had silent pulmonary embolism (PE), as assessed by high-probability lung scans. The frequency of silent PE was higher, 40% to 50%, when estimated on the basis of high-, intermediate-, and low-probability lung scans combined with the respective positive predictive values of PE for these categories of interpretation. These observations confirm those of several previous investigations, the first from 1987, quoted by Meignan and associates.2-8 The authors observed a comparable frequency of PE in patients with DVT that was limited to the popliteal veins and in patients with suprapopliteal extension. Regarding the results of treatment, the rate of recurrent PE among those with silent PE was comparable with the rate of new PE among those who only had DVT. This, however, was contrary to the observations of others.9