Q J Med 2002; 95: 637-638
© 2002 Association of Physicians
Correspondence |
Diagnosing deep-vein thrombosis
1 Department of Radiology 2 Department of Epidemiology 3 Department of Medicine, Bradford Royal Infirmary, Bradford BD9 6RJ, e-mail: sam.chakraverty@tuht.scot.nhs.uk
Sir,
There is much interest in ward-based screening tests that might avoid the need for imaging in patients presenting with suspected deep venous thrombosis, most of whom prove eventually not to have DVT. To be of use, such a test must have a high negative predictive value. We were interested to read of the experience in Bournemouth with computed strain-gauge plethysmography (CSGP).1 It is now in use in 100 UK hospitals. In Bradford, we had similar results when audited formally against ultrasound, with a negative predictive value of 97%.2 However other centres have reported better and worse experiences with the same equipment, reflecting the importance of assessing the equipment in the environment in which it is to be used.
An additional screening test is only of value if the attending clinician has enough confidence in the reliability and validity of a negative result to allow discharge of the patient. Most screening investigations have a small false negative rate, and clinicians will be naturally cautious about new diagnostic tests. We therefore looked at the impact of CSGP once introduced to the medical admissions ward as a stand-alone screening tool. The intention was to examine the level of confidence of clinicians in CSGP and to determine what they do in practice with patients with a negative CSGP result.
All patients presenting with suspected DVT in our medical admissions unit underwent investigation with CSGP performed by trained nursing and auxiliary staff. Imaging was used to investigate those patients with a positive or equivocal CSGP result. Patients with a negative CSGP result were discharged according to a written protocol and given advice. Returning patients were to be medically assessed and ultrasound performed. No formal pre-test clinical scoring for DVT was used at the initial presentation.
Patients who subsequently underwent further investigation with ultrasound or contrast venography, conventional or CT-pulmonary angiography and lung scintigraphy within 4 months of negative CSGP were identified and reviewed.
There were 354 CSGP examinations in an 8-month period: 209 (59%) were negative, 124 (35%) were positive and 21 (6%) were invalid or equivocal. In the negative CSGP group, 22 (10.5%) patients had ultrasound performed on the same day. Only one patient (0.5%) proved to have DVT identified.
There were a further 39 re-attendances for suspected thrombo-embolic disease within 4 months, of whom 28 had ultrasound performed. Four patients had DVT (1.9% of the negative CSGP group). These patients may either have had the diagnosis missed by the screening equipment or have had propagation of pre-existing calf vein clot above the trifurcation. A further 11 patients had ventilation/perfusion scans performed. Three patients (1.4% of the study group) had evidence of pulmonary emboli despite negative CSGP readings. No patient had clinical evidence of DVT on examination. Despite this, CSGP was performed outside of protocol. It is possible that the emboli originated from elsewhere, possibly the deep pelvic veins, or that no residual above-knee clot was present at the time of the CSGP study, or that these were false-negative CSGP studies. Management was unaffected by the CSGP result. However, based on this limited experience, we would caution against using a negative CSGP test as a surrogate marker to exclude pulmonary embolism.
One aim of this study was to examine whether clinicians in a medical admissions unit setting had enough confidence in a negative CSGP result to avoid requesting subsequent investigation. It appears that this was generally the case, with 89.5% of patients with a negative CSGP test not having further imaging on the same admission. We have no information about any different clinical variables present in the 10.5% of patients that did have further imaging despite a negative CSGP result. However only one patient was shown to have DVT the same day, and the other traceable results were negative. It must be remembered that no screening test is perfect, and it can be appropriate for patients with a negative CSGP test to proceed to imaging. A formal clinical scoring system, such as that devised by Wells3 may help to make such decision-making more explicit, and hence use imaging rather than preliminary screening tests in some patients.
Several assumptions have to be made in drawing any conclusions from our results. This hospital is not geographically isolated and it is possible that patients might attend other hospitals after initial presentation here. However, all patients were advised to return if problems continued, so were unlikely to seek medical advice elsewhere. Secondly, no information is available about deaths subsequent to a negative CSGP study, and it is possible that such deaths might have occurred without further medical investigation. We have assumed that our radiology and nuclear medicine information systems have captured appropriate investigations that were performed.
The maximum traceable false negative rate for thrombo-embolic disease was 8/209 (3.8%), comparable with our previously reported false negative rate, although it is possible that some of these patients did not have above-knee DVT at the time of the study. Our results are supported by a recently presented study using CSGP in an accident and emergency department setting, which reported a maximum traceable false negative rate of <1%.4 An acceptable false negative rate is clearly difficult to state in absolute terms, but should be comparable with standard imaging procedures. The test should be used in conjunction with clear advice to patients about the need to return if symptoms persist.
Clinicians appear to rely on the test in the majority of cases to inform their clinical management. From 349 patients referred to the medical admissions unit for investigation of suspected DVT, at least 187 patients were judged not to require urgent admission and radiological investigation which has clear benefits in reducing unnecessary in-patient admissions and reducing workload for both the ward and supporting radiology services.
References
1. Flanagan DEH, Creasy T, Thomas P, Cavan D, Armitage M. Computer-assisted venous occlusion plethysmography in the diagnosis of acute deep venous thrombosis. Q J Med2000; 93:277–82.
2. Goddard AJP, Chakraverty S, Wright J. Computer assisted strain-gauge plethysmography is a practical method of excluding deep venous thrombosis. Clin Radiol2000; 56:30–4.
3. Wells PS, et al. Value of assessment of pre-test probability of deep vein thrombosis in clinical management. Lancet1997; 350:1795–8.[Web of Science][Medline]
4. Ryan J, Barden S, Tidey B. An evaluation of emergency department lower limb plethysmography in the diagnosis of deep venous thrombosis. British Association of Accident and Emergency Medicine Annual Conference, 2000.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||