CRD summary

This publication reviewed methods of diagnosing deep vein thrombosis. The authors concluded that establishing clinical probabilities for patients is useful, and that patients with low clinical probability and negative D-dimer test can be excluded without ultrasound. The reliability of the conclusion is difficult to assess, mainly due to the sparsely documentation of the review.

Authors' objectives

To determine the accuracy of clinical prediction rules, alone or in combination with D-dimer assays, for the diagnosis of deep vein thrombosis (DVT), and to estimate the prevalence of DVT.

Searching

MEDLINE was searched from 1990 to July 2004 for studies published in English or French; the MeSH terms were provided. In addition, articles were identified from the bibliographies of included studies, retrieved reviews and an existing reference library.

Study selection

Assessment of study quality

The authors restricted inclusion in the review to independent blinded comparisons, but did not further assess the validity of the included studies.

Data extraction

Two authors independently abstracted data to determine the prevalence and diagnostic accuracy parameters.

Methods of synthesis

Results of the review

Fourteen studies (8,239 participants) met the inclusion criteria.

The prevalence was 5% (95% CI: 4, 8) in the low clinical probability group, 17% (95% CI: 13, 23) in the moderate clinical probability group, and 53% (95% CI: 44, 61) in the high clinical probability group.

The sensitivity of D-dimer testing in the low probability group was 88% (95% CI: 81, 92) and the specificity was 72% (95% CI: 65,78). The corresponding values in the moderate probability group were 90% (95% CI: 80, 95) and 58% (95% CI: 49, 67), respectively, and in the high probability group, 92% (95% CI: 85, 96) and 45% (95% CI: 37, 52).

The LRs for a normal result among patients with low clinical suspicion were 0.10 (95% CI: 0.03, 0.37) with a highly sensitive D-dimer assay and 0.20 (95% CI: 0.12, 0.31) with a moderately sensitive D-dimer assay. For patients with moderate clinical suspicion the LRs were 0.05 (95% CI: 0.01, 0.21) and 0.23 (95% CI: 0.13, 0.39), respectively, and for patients with high clinical suspicion they were 0.07 (95% CI: 0.03, 0.18) and 0.15 (95% CI: 0.10, 0.38).

Authors' conclusions

The diagnostic accuracy for detecting DVT improves when clinical probability is estimated before applying diagnostic tests. For patients with low clinical probability on the predictive rule and negative D-dimer results, diagnosis of DVT can be excluded without ultrasound; in patients with high clinical suspicion for DVT the D-dimer test results should not affect clinical decisions.

CRD commentary

The publication included a systematic review that was not exhaustively documented. The inclusion criteria were clear and some limited information about the included studies was provided. The search was targeted at only one electronic database and language restrictions were applied. It was unclear whether unpublished studies were sought. Publication and language bias might have been introduced into the review. The included studies were not assessed for validity, but the inclusion criteria restricted the types of eligible studies. Although all of study results were pooled, given the limited reporting of the characteristics of the included participants it was difficult to assess how appropriate this pooling was. The reliability of the conclusions is difficult to evaluate; the review process and results were only sparsely documented but seem to derive from the data overall.

Implications of the review for practice and research

Practice: The authors stated that out-patients with suspected DVT should be initially assessed using a validated clinical prediction rule. The D-dimer test result can be used to rule out DVT without the use of diagnostic imaging for patients with a low clinical probability estimate and a negative D-dimer test; ditto for moderate probability patients with a negative high-sensitivity D-dimer test. For patients with high clinical probability estimates, diagnostic imaging is necessary even with a negative D-dimer test. D-dimer tests should not be used as screening tests.

Research: The authors did not state any implications for further research.

Bibliographic details

Wells P S, Owen C, Doucette S, Fergusson D, Tran H. Does this patient have deep vein thrombosis? JAMA 2005; 295(2): 199-207. [PubMed: 16403932]

Original Paper URL

http://jama.ama-assn.org/

Other publications of related interest

Anand SS, Wells PS, Hunt D, Brill-Edwards P, Cook D, Ginsberg JS. Does this patient have deep vein thrombosis? JAMA 1998;279:1094-9.

These additional published commentaries may also be of interest. Perrier A. Review: the Wells clinical prediction guide and D-dimer testing predict deep vein thrombosis. Evid Based Med 2006;11:119. Perrier A. Review: The Wells clinical prediction guide and D-dimer testing predict deep vein thrombosis. ACP J Club 2006;145:24.

Indexing Status

Subject indexing assigned by NLM

MeSH

Algorithms; Decision Theory; Fibrin Fibrinogen Degradation Products /analysis; Probability; Sensitivity and Specificity; Venous Thrombosis /blood /diagnosis /ultrasonography

AccessionNumber

12006008062

Database entry date

31/05/2006

Record Status

This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.