Summary
Objective: To determine the reasons why 125I-fibrinogen leg scanning, a screening test which was reported to be very sensitive for the detection of post-operative thrombosis, has shown poor sensitivity in contemporary studies.
Study Identification: English-language reports were identified through a Medline computer search (1965-1991), Current Contents, and an extensive manual search of the bibliographies in identified articles.
Study Selection: Studies in orthopaedic or general surgical patients were reviewed that compared 125I-fibrinogen leg scanning with venography in all patients (accuracy studies) or in patients in whom 125I-fibrinogen leg scanning became positive (positive predictive value studies).
Data Extraction: A systematic appraisal of study design and specific descriptive information concerning the selection of patients.
Results of Data Synthesis: Six of the 15 studies which compared 125I-fibrinogen leg scanning with venography were level 1 studies (potential for bias minimized) and nine were classified as level 2 studies (potential for bias not minimized). In orthopaedic surgical patients, the pooled sensitivity of leg scanning for isolated calf vein thrombosis, for all venous thrombosis, and the pooled specificity were 55%, 45%, and 92% for the level 1 studies, respectively. These indices were 88%, 82%, and 79%, respectively for the level 2 studies (P <0.001). Only two (level 2) studies were found that evaluated the accuracy of leg scanning for venous thrombosis in general surgical patients.
Conclusion: We conclude that leg scanning is an insensitive method for the screening of post-operative venous thrombosis in orthopaedic patients. Our findings call into question the validity of the many studies (including meta-analyses) evaluating prophylactic agents for venous thrombosis which used leg scanning as the only test for the assessment of efficacy.
