Clinical assessment of limited value in DVT
diagnosis
Clinical question
Which clinical
features are useful in ruling in or ruling out deep vein thrombosis?
Bottom line
With the exception of
either a previous deep vein thrombosis (DVT) or a previous malignancy,
no other clinical feature effectively increases or decreases the odds of
having a DVT. The Wells Clinical Probability Score, which combines
several clinical features, is much more effective. (LOE = 1b)
Reference
Goodacre S, Sutton AJ,
Sampson FC. Meta-analysis: The value of clinical assessment in the
diagnosis of deep vein thrombosis. Ann Intern Med 2005; 143:129-39.
Study design: Meta-analysis (other)
Funding: Government
Allocation: N/A
Setting: Various (meta-analysis)
Synopsis
The authors of this
meta-analysis sought to determine what clinical assessment is best at
identifying patients with DVT. By searching 8 databases and
bibliographies of the identified papers, they identified 51 studies
enrolling patients with possible DVT. Two reviewers independently
determined the suitability of the study for the analysis, though only
one person extracted the data from the studies. The authors included
studies that evaluated the diagnostic performance of empirical
judgments, clinical findings, or clinical scores in identifying DVT.
They reported their results using a likelihood ratio (LR), which
expresses the odds that a given diagnostic test would be expected to be
positive in a patient with DVT versus in a patient without DVT. For
example, an LR of 1 means that the odds of DVT are no different if the
test result is positive or negative, whereas an LR of 2 means that the
odds are twice as good that the patient has a DVT than the patient's
odds were before the test was done. Individual clinical features, for
the most part, are not useful in ruling in or ruling out DVT; only a
history of DVT and a history of malignancy have an LR of greater than 2
and no clinical features effectively rule out DVT. The Wells Clinical
Probability Score, which scores the probability of DVT on the basis of
the presence or absence of several clinical features, had a positive LR
of 5.2 and a negative LR of 0.25. The Wells Score is more accurate in
younger patients and in patients who have not had a previous
thromboembolism. Other studies have shown that the Wells Score, combined
with D-dimer results, is quite accurate in ruling in DVT, but less
accurate in ruling out DVT (Ann Intern Med. 2005; 143:100-107). In some
of the 8 studies that evaluated it, physician judgment produced similar
LRs as the Wells Score, though the results were heterogeneous among the
studies.
| 