The series studied comprised 6197 patients who had died of or who had cancer at death and represents all patients with cancer from 21,530 necropsies performed at this department from 1960-84. Pulmonary embolism was significantly more common among cancer patients than in those with non-neoplastic diseases. Among those palliatively treated, patients with ovarian cancer, cancer of the extrahepatic bile duct system, and cancer of the stomach had the highest prevalence of pulmonary embolism (34.6%, 31.7%, and 15.2%, respectively). Necropsy patients with cancer of the oesophagus and larynx, together with leukaemia, myelomatosis, and malignant lymphoma had the lowest prevalence (0-5.6%). Palliatively treated cancers in organs of the peritoneal cavity had a significantly higher incidence than all other cancers combined. Cancer of the peritoneal cavity may impede venous drainage from the lower limbs and thus be an important factor in the onset of deep calf vein thrombosis and pulmonary embolism. It is concluded that cancer represents an increased risk factor for onset of pulmonary embolism, in particular in patients with ovarian cancer and cancer of the extrahepatic bile duct system.

SUMMARY The incidence ofpulmonary embolism and the number ofclinically missed diagnoses ofit in necropsies carried out between 1960 and 1984 at this department were investigated. Pulmonary embolism primarily affects elderly people with serious underlying disease; in this study it was found more often in women. The incidence of pulmonary embolism (9% of all necropsies) was unchanged during the period studied. In contrast, pulmonary embolism as the "sole" cause of death increased (p < 0O0005). Although most pulmonary emboli were the immediate cause of death, the clinical diagnosis was often missed (in 84% of all cases). Furthermore, such clinically missed diagnoses increased over the years (p < 0-005), especially in patients with heart disease and cancer.Without necropsy there will be considerable underdiagnosis of pulmonary embolism, therefore providing a misleading figure in the death statistics for this often fatal disease.Pulmonary embolism is a common cause of death,' but the clinical diagnosis can be extremely difficult, being recognised in 8%-70% of cases seen at necropsy.24"0 The incidence of pulmonary embolism is only between 5% and 14% in retrospective studies based on routine necropsies in general hospitals.389 In prospective studies the incidence reaches 50%-64%.4 1" Two studies have shown an increase in the incidence of pulmonary embolism in recent years.6 12In this department we have had a necropsy rate of between 75% and 80% for several decades.'3 (Medicolegal cases are not included.) The tentative clinical diagnosis and a copy of the death certificate accompanies the clinicians' request for a necropsy. Our material was considered to be suitable for a retrospective study of possible changes in the incidence of pulmonary embolism and the incidence of its clinical detection because necropsies are carried out in a uniform manner and the patients come from the same geographical area. Material and methodsThe following data were retrieved from the necropsy reports for the years between 1960 and 1984: necropsy number, department, sex, age, duration of last admis-

The post-sclerotherapy fluid production is probably due to an inflammatory reaction. This may explain the success of performing sclerotherapy in one single session.

A review of 8571 autopsies disclosed 2833 patients with malignant tumours from 1975 to 1984 at the Department of Pathology, The Gade Institute. Cardiac metastases were found in 130 cases. An increase of cardiac involvement was shown in the autopsy material from I . 2% in 1975-1979 to I .8% in 1980-1984. The same trend was seen if cardiac metastases were related to malignant tumours. Numerically, lung cancer accounted for most of the metastases seen, but the increase was made up by other tumours than lung cancer, especially malignant melanoma, mesothelioma, breast cancer and sarcomas. These tumours have a high frequency of heart metastases and the increased incidence of these cancers in the material explains the rise of cardiac metastases. Cardiac metastases increased with rising number of distant metastases. This study shows that mesotheliornas have the highest percentage of cardiac spread. The importance of autopsy for detecting metastatic spread in sites that are difficult to detect clinically is emphasized.

In a series of over 20,000 autopsies carried out over a 25‐year period, 700 cancers (11%) were found in patients in whom the diagnosis of cancer had not been considered relevant clinically. In over half of them the unrecognized tumour was considered an incidential finding and thus of no importance to the individual concerned, though of epidemiological concern. As expected from the literature the main organs involved were kidney and prostate, with stomach cancer in third place. In contrast, the unrecognized cancers that caused death were most often from the pancreas or lung, again with the stomach in third place. These patients with stomach or lung cancer were frequently thought to have died of other diseases in the gastrointestinal/respiratory tract, and the disease cited was often present though not the cause of death. These patients tended to be older than those with clinically recognized disease, and had been hospitalized late in the course of their fatal illness, at which time clinical recognition would not have influenced the outcome. Earlier recognition of this type of cancer is thus essential. This cannot be achieved without definition and further analysis of the entities concerned.