Current issues of ACP Journal Club are published in Annals of Internal Medicine


D-dimer levels detected DVT in patients hospitalized for stroke rehabilitation

ACP J Club. 1997 Nov-Dec;127:82. doi:10.7326/ACPJC-1997-127-3-082

To the Editor

The commentary by Bath on the use of D-dimer levels as a screening test for deep venous thrombosis (DVT) among immobile stroke rehabilitation patients (1) states that compression ultrasonography is insensitive for the detection of DVT in asymptomatic patients. However, the study that he cited included a very different population of patients than those with a neurologic deficit. This is important, given that many stroke patients have prolonged immobilization and may lack the visual or sensory cues that alert a physician to the development of early symptoms of DVT.

By stating that clinicians "should continue to screen for DVT using clinical examination, although it is insensitive," Bath gives mixed messages: Continue to accept a practice with demonstrated poor diagnostic efficacy (2), yet delay adopting a new practice of likely benefit.

The evidence for such prophylactic measures as compression stockings, aspirin, and heparin (the last for "high-risk patients") against DVT among stroke patients is no more compelling than the current D-dimer study, yet Bath recommends all 3. Because the patients included in the D-dimer study were all at high-risk for DVT, the use of low-dose heparin would be warranted; however, its use is associated with such risks as heparin-induced thrombocytopenia (3). A serial screening test, such as the D-dimer assay, seems to be a practical adjunct to this approach.

How many studies does it take to implement an inexpensive, safe, and efficacious technology?

Joel Ray, MD
McMaster University
Hamilton, Ontario, Canada

Commentator's Reply

Ray raises 3 points about my commentary on D-dimer screening for DVT in patients with subacute stroke (4). First, he supports the use of compression ultrasonography as the gold standard for screening for DVT in asymptomatic patients, although this tool was not reliable in a group of asymptomatic postoperative patients (5). Although postoperative patients and those with stroke clearly differ, his comment that patients with stroke may not be able to alert physicians to the development of DVT is irrelevant: We need a screening test that detects DVT in asymptomatic patients.

Second, he challenges the relevance of prophylactic measures. Compression stockings are clearly effective for surgical DVT prophylaxis. Although no definitive evidence proves their effectiveness in patients with stroke, it is unlikely that a randomized controlled trial will be done in this group, so we must extrapolate from existing data. Aspirin also probably reduces the risk for DVT after stroke, as suggested by a reduction in symptomatic pulmonary embolism in the International Stroke Trial (6) and a nonsignificant reduction in the Chinese Aspirin Stroke Trial (7). Although heparin clearly reduces the risk for DVT after stroke (8), it increases the risk for intracranial bleeding and a poor outcome (6). Thus, it should not be used routinely in patients who have had acute ischemic stroke, with the possible exception of morbidly obese patients or those with a history of venous thrombo-embolism. Whether heparin is safe if given later in the recovery period has yet to be tested formally in a randomized controlled trial.

Lastly, Ray questions why D-dimer assays cannot be introduced into clinical practice. Unfortunately, he does not address important practical questions, including assay cost and standardization and the need for hospitals to develop their own diagnostic cut points.

On the basis of current data, I remain unconvinced that D-dimer assay should be used as a routine screening test for DVT in asymptomatic patients with stroke.

Philip M. Bath, MD
King's College School of Medicine & Dentistry
London, England, UK

Author's Response

Plasma D-dimer ELISA has shown promise in several studies as a means to rule out DVT in symptomatic, non-hospitalized patients using cut points ranging between 300 to 540 ng/mL. The pooled sensitivity and specificity of the test are 97% and 35%, respectively (9). We recently showed that plasma D-dimer levels may be similarly used in rehabilitation patients with stroke if a higher cut point is used (10).

Bath expresses concern that we used venous duplex ultrasonography (VDU) rather than contrast venography as our standard for DVT (11). Venography is the gold standard for DVT diagnosis, but it is important to note that VDU in asymptomatic patients often has poor sensitivity while maintaining high specificity (12). Given that VDU is specific and D-dimer levels are sensitive, venography is likely to have shown that the D-dimer assay had a higher positive predictive value than that found in our study without lessening its negative predictive value. Thus, plasma D-dimer levels may be a useful adjunct to clinical examination, which is very unreliable in patients with stroke, as Ray has pointed out.

Despite the prevailing evidence that plasma D-dimer levels are a sensitive indicator of venous thromboembolism, routine clinical use of this test can only be recommended when management trials using decision trees, such as the one proposed in our paper, have been completed. Further, plasma D-dimer screening will only be practical after a rapid assay that maintains the sensitivity of ELISA is developed (9).

I agree that DVT prophylaxis in patients with stroke is very important. Bath recommends that compression stockings be used to prevent DVT in immobilized patients, aspirin for those with ischemic stroke, and heparin for "high-risk" patients. Ray suggests that all immobilized patients with stroke undergoing rehabilitation are at high risk. Currently, no clear measure exists for thromboembolic risk in patients with stroke and hemiplegia, although retrospective data note that DVT is rare in ambulatory patients with stroke (13). To date, no study has supported the efficacy of stockings and aspirin alone for DVT prevention. Low-dose heparin is the more prudent prophylactic choice for all neurologically stable, nonambulatory patients with stroke who are at low risk for bleeding. For patients who cannot tolerate heparin, intermittent application of compression sleeves for the legs is advisable.

Richard R. Harvey, MD
The Rehabilitation Institute of Chicago
Chicago, Illinois


1. Bath PM. Commentary on D-dimer levels detected DVT in patients hospitalized for stroke rehabilitation. ACP J Club. 1997 Mar-Apr; 126:43. Evidence-Based Medicine. 1997 Mar-Apr;2:59.

2. Vaccaro P, Van Aman M, Miller S, Fachman J, Smead WL. Angiology. 1987;38:232-5.

3. Warkentin TE, Levine MN, Hirsh J, et al. N Engl J Med. 1995;332:1330-5.

4. Bath PM. Commentary on D-dimer levels detected DVT in patients hospitalized for stroke rehabilitation. ACP J Club 1997 Mar-Apr; 126:43. Evidence-Based Medicine. 1997 Mar-Apr;2:59.

5. Jongbloets LM, Lensing AW, Koopman MM, Buller HR, ten Cate JW. Lancet. 1994;343:1142-4.

6. International Stroke Trial Collaborative Group. Lancet. 1997;349:1569-81.

7. CAST (Chinese Acute Stroke trial) Collaborative Group. Lancet. 1997;349:1641-9.

8. Counsell C, Sandercock P. Stroke Module of the Cochrane Database of Systematic Reviews, The Cochrane Library. Oxford: Update Software; 1997.

9. Bounameux H, de Moerloose P, Perrier A, Reber G. Thromb Haemost. 1994;71:1-6.

10. Harvey RL, Roth EJ, Yarnold PR, Durham JR, Green D. Stroke. 1996; 27:1516-20.

11. Bath PM. Commentary on D-dimer levels detected DVT in patients hospitalized for stroke rehabilitation. ACP J Club. 1997 Mar-Apr; 126:43. Evidence-Based Medicine. 1997 Mar-Apr;2:59.

12. Agnelli G, Radicchia S, Nenci GG. Haemostasis. 1995;25:40-8.

13. Bromfield EB, Reding MJ. Journal of Neurological Rehabilitation. 1988;2:51-7.