We performed a prospective study in 86 consecutive patients with central vein catheter-related deep venous thrombosis (DVT) of the upper extremity, to evaluate the prevalence of pulmonary embolism (PE), and to identify clinical variables that would increase the likelihood of developing PE in an individual patient. Since upper-extremity DVT was established, all patients received intravenous heparin therapy. Then, a ventilation-perfusion lung scan was obtained within 24 h of DVT diagnosis, whether respiratory symptoms were present or not. Six points of clinical information were recorded on entering in the study, and then compared with the scintigraphic findings: age, sex, the underlying disease, the catheter material, the character of the infusate, and the duration of cannulation. Thirteen patients were considered to have PE. Sixty-six patients were finally classified as having a normal lung scan, and 7 patients were excluded from the study (because of indeterminate lung scan 6; because of femoropopliteal thrombosis simultaneously present 1). Two out of the 13 patients with PE subsequently died because of recurrent, massive embolism, despite adequate heparin therapy. PE was more commonly present in patients with polyvinyle chloride or polyethylene catheters (10/38, 26%) as compared to patients with either polyurethane or siliconized catheters (3/41, 7%; p < 0.05, Chi-Square test; Odds Ratio = 4.52, 95% CI 1.01-23.07). We conclude that PE is not a rare event in these patients, and it may be life-threatening even despite adequate heparin therapy.(ABSTRACT TRUNCATED AT 250 WORDS)