The Effect of Storage on Blood Used for Prothrombin Estimations By

Suzman (1956) and Suzman, Ruskin, and Goldberg (1955) have confirmed that the incidence of repeated attacks of coronary thrombosis is less in patients undergoing prolonged anticoagulant therapy than in untreated cases. The main objection to anticoagulant therapy, however, is the necessity for strict laboratory control to avoid therapeutically induced haemorrhage. In an editorial in the Lancet (1956) it is stated that in " good risk " cases " the relatively small advantage to be obtained from anticoagulants must be set against the risk of transport to hospital; if the assessment is ' poor risk,' then the patient should be got to hospital for anticoagulant treatment as soon as possible." It appears from this statement that adequate control of anticoagulant therapy necessitates admission to hospital. We cannot agree that this is necessary or, in fact, advisable, our view being based on an experience extending over several years during which over 1,000 patients have been under laboratory control either ambulatory or domiciliary. Haemorrhage in adequately controlled cases has been rare. It has, however, hitherto apparently not been customary in England to undertake laboratory control in domiciliary practice. The major objections have been (a) the fear of haemorrhage, which certainly occurred more frequently (Biggs and Macfarlane, 1949; James, 1949) when prothrombin estimations were performed with Russell's viper venom as the source of thromboplastin (Fullerton, 1940), and (b) the necessity of using freshly drawn blood for Quick, Stanley-Brown, and Bancroft's (1935) one-stage prothrombin test. The latter has apparently been based on Quick's (1951) statement that for research studies blood should be used within 30 minutes of collection, whereas up to four hours' storage was permissible for routine clinical purposes. Since specimens submitted by distant medical practitioners could not reach our laboratories within four hours after collection, an investigation was undertaken in 1953 similar to that reported by Watson (1956), but no report was made at the time. However, in view of the increasing use of anticoagulants and of favourable recent reports (Suzman et al., 1955; Suzman, 1956; Wright, Marple, and Beck, 1948; Gilchrist and Tulloch, 1956), it was thought of interest to record a confirmation of Watson's findings. The main purpose of the study was to determine the effect of a 24hour delay on the estimation of the prothrombin index and to assess the significance of the possible loss of labile factor (Factor V).