This review assessed the accuracy of clinical examination for peripheral arterial disease (PAD). The authors concluded that clinical examination findings are not independently sufficient to diagnose PAD; they must be used in the context of the pre-test probability. The small numbers of studies, lack of quality assessment details, and potential for language and publication bias limit the reliability of these conclusions.

To assess the accuracy and precision of the clinical examination for lower extremity peripheral arterial disease (PAD).

MEDLINE (from January 1966 to March 2005) and the Cochrane Database of Systematic Reviews were searched for articles reported in the English language; the search terms were reported. In addition, the authors checked the bibliographies of identified articles and four physical examination textbooks.

Study designs of evaluations included in the review

The authors did not state any inclusion criteria relating to the study design.

Specific interventions included in the review

Studies that assessed an element of the history or physical examination, in which the physical examination manoeuvres were clearly described and could be accomplished at the bedside, were eligible for inclusion. The signs and symptoms assessed in the included studies were claudication, skin changes, bruits, pulse palpation, and venous or capillary refill time.

Reference standard test against which the new test was compared

The studies had to compare physical examination with the ankle-brachial index (ABI), angiography or duplex sonography to be included in the review.

Participants included in the review

The authors did not state any inclusion criteria relating to the participants. However, they stated that studies in which patients were identified improperly, i.e. all patients were free of PAD or all patients had PAD, were excluded. The studies included patients who sought medical attention for leg complaints (classified as symptomatic) and patients who did not have a specific leg complaint but the clinician conducted a screening for PAD based on the patient's risk factor profile (classified as screening).

Outcomes assessed in the review

The studies had to report sufficient data to construct 2x2 tables to be included in the review.

How were decisions on the relevance of primary studies made?

Two reviewers independently assessed studies for inclusion, with any discrepancies resolved by discussion.

The quality of the included studies was assessed using checklists developed for the Rational Clinical Examination series of JAMA. Two reviewers independently assessed study quality, with any discrepancies resolved by discussion.

The authors did not state how the data were extracted for the review, or how many reviewers performed the data extraction.

The authors calculated the prevalence, sensitivity, specificity, kappa scores, likelihood ratios (LRs) and 95% confidence intervals (CIs) using conventional definitions. These were calculated for any degree of PAD (i.e. ABI levels of 0.97 or less) and for moderate to severe PAD (i.e. ABI levels of less than 0.50).

How were the studies combined?

Summary estimates were calculated using a random-effects model.

How were differences between studies investigated?

Statistical heterogeneity was assessed using the Mantel-Haenszel Q statistic. The authors also stated that they evaluated studies to look for explanations for heterogeneity.

Eleven studies of diagnostic accuracy (n=6,272) and 6 studies of precision (n=248) were included in the review.

Six diagnostic accuracy studies achieved a quality score of 1, one scored 2 and four scored 3. The authors stated that they identified several high-quality studies, however, since no details of the quality assessment tool were provided in the paper, it was not possible to verify this.

In asymptomatic patients, the most useful clinical findings to diagnose PAD were the presence of symptoms of intermittent claudication (4 studies; LR 3.30, 95% CI: 2.30, 4.80), femoral bruit (1 study; LR 4.80, 95% CI: 2.40, 9.50) or any pulse abnormality (3 studies; LR 3.10, 95% CI: 1.40, 6.60). The absence of these clinical examination features were not found to lower the likelihood of PAD, but the absence of symptoms of intermittent claudication (1 study; LR 0.57, 95% CI: 0.43, 0.76) or the presence of normal pulses (1 study; LR 0.44, 95% CI: 0.30, 0.66) were found to lower the likelihood of moderate to severe PAD.

In patients who were symptomatic with leg complaints, the most useful clinical findings to diagnose PAD were the presence of cool skin in the affected leg (1 study; LR 5.90, 95% CI: 4.10, 8.60), discoloured skin (1 study; LR 2.80, 95% CI: 2.40, 3.30), or wounds or sores (1 study; LR 5.90, 95% CI: 2.60, 13.40), the presence of at least one bruit at rest (iliac femoral or popliteal) (3 studies; LR 5.60, 95% CI: 4.70, 6.70), or any palpable pulse abnormality (6 studies; LR 4.70, 95% CI: 2.20, 9.90). The absence of any bruits (iliac, femoral or popliteal) (LR 0.39, 95% CI: 0.34, 0.45) or pulse abnormality (LR 0.38, 95% CI: 0.23, 0.64) reduced the likelihood of PAD.

Combinations of clinical abnormal findings did not increase the likelihood of PAD beyond that of individual abnormal findings. However, when combinations of clinical findings were all normal, the likelihood of PAD was lower than when individual symptoms or signs were present.

A PAD scoring system, assessed by one study, used hand-held Doppler to derive a score based on the number of auscultated arterial components, grade of the peripheral pulse and history of myocardial infarction. This system provided the greatest diagnostic accuracy: patients scoring less than 6 had an increased likelihood of PAD (LR 7.80, 95% CI: 4.80, 12.70), whilst those scoring 6 or more had a decreased likelihood of PAD (LR 0.20, 95% CI: 0.10, 0.40).

Inter-observer agreement for diagnostic precision of pulse palpation (6 studies) was moderate to high (kappa statistic 0.27 to 1.00) when determining whether a pulse was present or absent. Discriminating between reduced and normal pulsations showed only marginal reproducibility (2 studies). Precision appeared to increase with experience (1 study).

Clinical examination findings are not independently sufficient to include or exclude a diagnosis of PAD alone with any certainty; they must be used in the context of the pre-test probability. The PAD screening score, using the hand-held Doppler, yields the greatest diagnostic accuracy.

The review question was clear and the inclusion criteria were stated in relation to the intervention, reference standard and outcomes of interest. No inclusion criteria relating to the participants or study design were specifically stated. The authors searched only two electronic databases, along with handsearches, therefore other relevant studies might have been missed. Only publications in English were sought, thus increasing the potential for language bias. No attempts were made to identify unpublished studies, thereby increasing the potential for publication bias, which was not assessed. The study selection and quality assessment processes were carried out in duplicate, which helps reduce errors and reviewer bias. However, since the method of data extraction was not reported, the potential for reviewer bias and error cannot be assessed for this procedure. This is particularly important for diagnostic accuracy studies where the results are calculated from 2x2 tables. The quality of the included studies was assessed using a checklist developed for the journal series where the review was published. The results of the quality assessment exercise were reported, along with adequate details of the primary diagnostic accuracy studies. Appropriate measures of effect were calculated and statistical heterogeneity was assessed.

The authors' conclusions appear to follow from the evidence presented. However, the results were based on small numbers of studies of unclear quality, and the presence of language and publication bias cannot be excluded. The authors' conclusions should therefore be interpreted with caution.

Practice: The authors stated that for screening patients who require further testing to diagnose PAD, the most useful individual symptoms and signs are asking about the presence of claudication, listening for a femoral bruit, and palpating for a pulse abnormality. For symptomatic patients with leg complaints, the most useful individual findings are the presence of cool skin, the presence of at least one bruit, and any palpable pulse abnormality. These individual findings must be used in context with the pre-test probability, as none of the symptoms or signs is independently sufficient to diagnose or rule out PAD with certainty. Clinicians who measure the ABI should listen and notice the number of auscultated components. Use of the hand-held Doppler to derive a score in a PAD scoring system yields a greater diagnostic accuracy for PAD than any individual or combined clinical findings.

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

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.