Dr Murphy¹ supports a more discriminate approach to thrombophilia testing than I describe in my own practice² and points out that my opinion is at variance with British Committee for Standards in Haematology (BCSH) guidelines. In respect of the latter, these guidelines were not evidence based and were published 10 years ago. Dr Murphy identifies the main indications for thrombophilia testing from the guidelines as: (1) patients with venous thromboembolism before the age of 45 years, (2) recurrent venous thrombosis or thrombophlebitis, (3) thrombosis in an unusual site, and (4) a first venous thromboembolic event with a clear history of venous thrombosis. Applying current criteria for graded recommendations on levels of evidence these indications would be classified as grade C based on level IV evidence.³ The haemostasis and thrombosis task force for the BCSH is currently preparing an updated guideline on the investigation and management of heritable thrombophilia. This guideline will be evidence based. However, given the lack of randomised clinical trials only grade B recommendations at best will be possible. I share Dr Murphy's concern regarding limited health care resources and the cost effectiveness of a practice for which there are no current grade A recommendations.

The important issue that Dr Murphy raises relates to why thrombophilia tests are performed; are they to explain why an individual develops venous thromboembolism or are they to optimise clinical decisions? In other words, are they for the sake of science or medicine. If testing is performed to explain why an individual develops thrombosis then the discriminate approach suggested by the 1990 guidelines quoted by Dr Murphy will increase the proportion of tested patients who are found to have a laboratory abnormality.⁴ However, if testing is performed to optimise clinical decisions then there is no rational basis for such a recommendation given the data that are currently available. There is no evidence that the predictive value of thrombophilia testing is in any way superior in the categories of patients outlined above to that in other patients with venous thromboembolism. For example, a patient with a first event after the age of 45 years in the absence of a family history might still be at risk of recurrence in the future. Therefore, the issue is whether testing patients with venous thromboembolism for laboratory evidence of thrombophilia has predictive value. The presence of a raised titre of anticardiolipin antibodies indicates a higher risk of early recurrence and therefore might be considered an indication for continued anticoagulation (grade B). Because lupus anticoagulant activity is also indicative of antiphospholipid activity it might also be considered an indication for continued anticoagulation.⁵ The predictive value of testing for heritable thrombophilia is also becoming clearer as I indicated in my leader. There is no evidence to support a higher intensity or extended duration of anticoagulation in most patients with laboratory evidence of heritable thrombophilia. Exceptions might be patients with homozygous or combined heterozygous defects (grade B). In my own practice I find testing valuable for assessing risk in family members, particularly females patients who are considering pregnancy or oestrogen/progestagen contraceptive pill use. If a thrombophilic defect has been detected in the symptomatic index family member then that specific defect can be looked for in the relatives who are requesting counselling. This obviates the need for an expensive comprehensive screen in all of the family members and avoids confusion as to whether an abnormal result is relevant or not.

Like all clinicians responsible for the judicious use of scarce health care resources Dr Murphy has to decide whether thrombophilia test results influence his clinical practice. If they don't then there is no necessity to do them.