Can we measure the ankle–brachial index using only a stethoscope? A pilot study

Abstract

Background. Ankle-brachial index (ABI) is an excellent method for the diagnosis of peripheral arterial disease (PAD) when it is performed with Doppler. However, this device is not always available for primary care physicians. The ABI measured with stethoscope is an easy alternative approach, but have not been proved to be useful.

Objective. To assess the accuracy of the ABI measured using a stethoscope comparatively to that of the current eligible method for the diagnosis of PAD, the Doppler ABI, and describe the characteristics of this new approach.

Methods. We conducted a diagnostic study of ABI measured with a stethoscope and a Doppler probe and compared the results. Eighty-eight patients were accessed by both methods.

Results. Mean stethoscope ABI, 1.01 ± 0.15, and mean Doppler ABI, 1.03 ± 0.20, (P = 0.047) displayed a good correlation. Measurements of stethoscope ABI diagnostic accuracy in recognizing a Doppler ABI are described. The comparison of this data with the current gold standard method results gave a sensitivity of 71.4% [95% confidence interval (CI), 41.9–91.6] and specificity of 91.0% (95% CI, 81.5–96.6), with predictive positive value of 62.5% (95% CI, 38.6–81.5) and negative predictive value of 93.8% (95% CI, 85.2–97.6). The study accuracy was 87.7%. The area under the ROC curve was 0.895 (95% CI, 0.804–0.986, P < 0.0001).

Conclusions. According to our study, the stethoscope ABI is a useful method to detect PAD and it may be suitable for its screening in the primary care setting.

Ankle-brachial index, peripheral arterial disease, stethoscope

Introduction

Peripheral arterial disease (PAD) is an important public health problem and it is estimated that 27 million people in the US and Europe suffer from this illness.1 Its prevalence in people aged 75 and older rises over 20%, which makes it relevant since life expectancy has risen sharply worldwide.2,3 However, its presence is neglected by a large number of physicians. MacLafferty et al.4 showed that only 37% of the clinicians ask their patients about PAD symptoms, in contrast to 92% who ask about coronary artery disease symptoms. In another study, the detection rate of PAD in patients with known peripheral arterial insufficiency was only 49%.5

Since PAD is a systemic illness due mainly to atherosclerosis, symptomatic or not individuals have a high chance of having angina, acute myocardial infarct, transient ischaemic attack, stroke, heart failure, lower limb ischaemia and amputation and internal carotid stenosis.1,6–11 It is considered an independent mortality predictor even after adjustment for other cardiovascular disease risk factors.12 The more severe the stenosis is, the higher the mortality rate is, reaching 56% in 5 years,9,13 which gives PAD a worse prognosis than those of non-Hodgkin’s lymphoma and breast and colon cancer.1,13 In addition, peripheral muscle ischaemia is well tolerated during years and it may be accommodated for by decreasing physical activity with advancing age. The relatively easy access to peripheral circulation may allow the early identification of asymptomatic individuals, thus allowing taking preventive measures against serious brain and heart clinical atherosclerotic disease.11

The elective ambulatorial method to detect PAD is the ankle–brachial index (ABI) measured with a portable Doppler probe and a blood pressure (BP) cuff. Normally, arterial BP is slightly higher in the legs than in the arms.14 This is used to establish a relationship between the inferior and the superior territories, which is called ABI. Values under 0.9 are strongly associated with PAD with a sensitivity of 95% and specificity of 100%.1,15 However, portable Doppler is not always available in the office setting. An appealing alternative would be measuring ABI using a stethoscope, a widely available tool in clinical practice, instead of a portable Doppler. Nonetheless, this possibility has been studied only once, and the results were not as good as expected.16 Yet, the sensitivity, specificity, predictive positive and negative values and likelihood ratio of either a positive or a negative test were not described.

The present study was designed to assess whether a stethoscope may be used to detect PAD in clinical practice, compared with the current eligible method (Doppler), and describe the characteristics of this new approach according to previous diagnostic test validation guidelines.17–21

Methodology

Patients

They referred to the haemodynamic laboratory of Felício Rocho Hospital for peripheral, renal, coronary and cerebral vascular territory diagnostic and interventional angiographic exams between January 2005 and July, 2006 were selected for this study. Exclusion criteria were as follows: patients under age of 18, those who needed emergency percutaneous interventions and those with haemodynamic instability. The study was approved by the Ethics Committee of Felício Rocho Hospital and all participants provided their signed informed consent after a detailed introduction to the study.

Methods

Patients were put in resting supine position for at least 10 minutes and each test was done by trained physicians or trained medicine students, who were blind to other results. All examiners used the same methodology. The cuff was positioned at the terminal portion of the leg. The stethoscope and the Doppler systolic pressures were recorded when the first Korotkoff sound was heard and when the flow resumed after gradual and slow cuff deflation, respectively.

The systolic pressure was first taken in the brachial territory with an appropriate sphygmomanometer and then the dorsalis pedis and posterior tibial systolic pressures were measured. The higher systolic pressure of the two inferior arteries was divided by the higher systolic pressure of the arms to obtain the ABI. A value ≤0.90 was considered abnormal. No adverse events were noted during the tests.

Statistical analysis

The sample size of 88 patients was estimated according to Zhou et al.22 to reach a sensitivity of 90%, using the prevalence of the disease studied in that population (18%) calculated by internal ABI measurements, which was similar to other published studies that considered older patients1, a type II error (β error) of 20% and a type I fixed in 5%.

Data are presented using descriptive statistics, including mean values and standard deviation (SD) and absolute numbers and percentages. The correlation between the stethoscope and the Doppler ABI was performed using Pearson coefficient and Bland–Altman difference plots (also known as bias plots).23 They were used to compare the two methods visually. Standard diagnostic measurements were calculated to describe the diagnostic accuracy of the stethoscope ABI in comparison to the Doppler ABI, here considered the gold standard. A receiver operating characteristic (ROC) curve24 was plotted considering the Doppler ABI < 0.90 as the disease state and the area under the curve was calculated. Statistical analyses were performed using Analyse-it for Microsoft Excel (version 2.00, Analyse-it Software, Ltd http://www.analyse-it.com/; 2007). The significance level was 0.05.

Results

From 88 patients selected for this study, seven were excluded due to incomplete data, but they did not differ from the rest of the population included in the analysis. Twenty-nine were women and 52 were men. Fifty-eight were older than 70 years, 7 were current smokers, 31 alleged smoking in the past, 49 had hypertension, 40 were dyslipidemic and 18 were diabetic (Table 1). Fifty patients used aspirin, 1 used heparin, 1 used ticlopidin, 1 used clopidogrel and 30 used statin. Thirteen complained of stable angina, 16 presented with unstable angina, 4 with non-ST segment elevation myocardial infarction, 6 with ST segment elevation myocardial infarction on the post-acute phase, 16 with abnormal treadmill test result, 1 with claudication symptoms, 2 with a history of pulmonary oedema, 13 with other causes and 10 were asymptomatic (Table 2).

TABLE 1

Population characteristics

Characteristics	N (%)
Gender
Female	29 (35.8)
Male	52 (64.2)
>70 years old	58 (71.6)
Smoking	7 (8.6)
Ex-smoking	31 (38.3)
Hypertension	49 (60.5)
Dyslipidemia	40 (49.4)
Diabetes	18 (22.2)
Stroke	3 (3.7)
Myocardial infarction	12 (14.8)
Myocardial revascularization	6 (7.4)
Angioplasty	8 (10.0)

TABLE 2

Indication of the haemodynamic study

Cause	N (%)
Stable angina	13 (16.0)
Unstable angina	16 (19.8)
NSTEMI	4 (4.9)
STEMI	6 (7.4)
ATT	16 (19.8)
Claudication	1 (1.2)
Pulmonary oedema	2 (2.4)
Asymptomatic	19 (23.5)
Other	13 (16.0)

ATT, abnormal treadmill test; NSTEMI, non-ST segment elevation myocardial infarction; STEMI, ST segment elevation myocardial infarction.

Mean stethoscope ABI was 1.01 ± 0.15 and Doppler ABI was 1.03 ± 0.20 (paired t test, P = 0.047) and displayed a good correlation (Fig. 1, r = 0.58, P < 0.001). In two cases, there is a large disagreement between the two methods. The Bland–Altman difference or bias plot (Fig. 2) allows the visual comparison of the two methods. Stethoscope ABI underestimated ABI slightly when compared to the standard method [bias: −0.03, 95% confidence interval (CI) −0.07 to −0.00, P = 0.047 with 95% limits of agreement from −0.35 to 0.28].

FIGURE 1

Scatter plot between stethoscope and Doppler ABI (r = 0.58, P < 0.001)

FIGURE 2

Difference plot (bias plot) for the visual comparison of the two methods, stethoscope and Doppler ABI. Observe a small bias (−0.03, 95% CI −0.07 to −0.00, P = 0.047) and a 95% limit of agreement from −0.35 to 0.28

Fourteen patients displayed abnormal Doppler ABI and 16 patients presented with abnormal stethoscope ABI (Table 3). Measurements of stethoscope ABI diagnostic accuracy in recognizing a Doppler ABI <0.90 are shown in Table 4. Comparing these data with those of the current gold standard method, we found a sensitivity of 71.4% and a specificity of 91.0% with a predictive positive value of 62.5% and a negative predictive value of 93.8%. The diagnostic accuracy was 87.7% (Tables 3 and 4). The area under the ROC curve was 0.895 with a CI of 0.804–0.986 and P < 0.0001 (Fig. 3).

TABLE 3

Abnormal results using Doppler and stethoscope

ABI (stethoscope)	ABI (Doppler)
	Abnormal	Normal	Total
Abnormal	10	6	16
Normal	4	61	65
Total	14	67	81

TABLE 4

Diagnostic accuracy of stethoscope ABI in recognizing a Doppler ABI < 0.90

Features	Value	95% CI
Sensitivity	71.4	41.9–91.6%
Specificity	91.0	81.5–96.6%
Positive predictive value	62.5	38.6–81.5%
Negative predictive value	93.8	85.2–97.6%
Accuracy	87.0	Not applicable
Likelihood ratio of a positive test	7.93	3.47–18.33%
Likelihood ratio of a negative test	0.31	0.14–1.96%

FIGURE 3

ROC curve for evaluating the accuracy of stethoscope ABI in the recognition of abnormal Doppler ABI (<0.9). The area under the ROC curve was 0.895 with a CI of 0.804–0.986 and P < 0.0001

Discussion

Our study shows that as ABI measured with a single stethoscope and a sphygmomanometer has a great correlation with the Doppler ABI, it can be useful to screen PAD in the primary care setting. To our knowledge, there is only another article in literature that explored the use of the stethoscope as a diagnostic tool in PAD.16 However, it did not described some important information such as sensitivity, specificity, diagnostic accuracy and likelihood ratio of a positive and negative test. Another study evaluated the accuracy of automated ABI measurement by oscillometry and found that it can be used safely to exclude the possibility of PAD with a negative predictive value of 96% in the left leg and 88% in the right.25 This is a simpler and faster method that can be used in the office by general physicians.

Since anamnesis and clinical examination do not have a good correlation with the presence of inferior limb stenosis1,8,12,26, ABI measured with a stethoscope may be an easy and practical tool for ruling out PAD in the office. Its independence from electronic devices like Doppler ABI makes it available worldwide with no additional costs for the health care provider.

Furthermore, the use of a stethoscope in the diagnosis of PAD may have a great impact on medicine office practice, since the portable Doppler probe is not available to the majority of the primary care physicians and patient referral to a vascular specialist may become less frequent, saving both money and time for the initial treatment. Yet, the earlier the diagnosis is made, the earlier the opportunity for secondary prevention of adverse effects of atherosclerosis in any body territory is, and the lesser the damage secondary to amputations, acute myocardial infarction, stroke and other events. As the method may be widely and easily applied, it should be used in screening for the early detection of PAD. Patients with abnormal stethoscope ABI must be referred to a vascular specialist to perform a conventional Doppler ABI, the current gold standard method.

Doppler ABI can stratify the risk of major cardiovascular events in a patient depending of its value10, as well as mortality for other causes, both in men and women.27 This is possible because the Doppler probe can identify very slow blood flow, which indicates severe obstruction and a higher chance of atherosclerosis in sites other than the limbs. This stratification is not possible with the stethoscope ABI; thus, it should be used only to exclude the disease.

Some limitations of the present study should be acknowledged. The reproducibility of the stethoscope ABI was not performed in our study. However, another study evaluated the reliability of Doppler ABI by three different observer groups and, according to it, there are only small differences and no systematic bias between the three groups, which makes it safe and reliable28, and we think that this results can also be applied to a simpler method such as the auscultatory one. In two cases, a large disagreement between the results of the methods may have negatively affected the coefficient of correlation. We studied a hospital population, which may not resemble actual office care patients. Though, caution must be taken when interpreting the stethoscope ABI in the general population, especially in younger patients, since the prevalence of the PAD in this setting is lower. Due to the similarities with the population studied, the results of this study can be safely applied to older patients and for those with risk factors for PAD such as diabetes, smoking, hypertension, etc.

There are no data on the prognostic value of the stethoscope ABI, therefore, it should be used only as a screening test. The new electro-acoustic devices may give new insights since they may detect such a slow blood flow that a lower ABI may be acquired and the relationship with prognosis obtained. Future researches with these devices are needed.

To conclude, according to our study, the stethoscope ABI is a useful method to exclude PAD and it may be suitable for screening in the primary care setting. However, more studies are necessary for further confirmation and to establish the prognostic value of the stethoscope ABI.

Declaration

Funding: None.

Ethical approval: The study was approved by the Ethics Committee of Felício Rocho Hospital.

Conflicts of interest: None.

Dr Eduardo Key Marquesini Washizu, Dr Daniel Mendes Pinto and Dr Leonardo Ghizoni Bez provided substantial contribution to this study, by selecting patients and collecting data.

Author notes

Carmo GAL, Mandil A, Nascimento BR, Arantes BD, Bittencourt JC, Falqueto EB and Ribeiro AL. Can we measure the ankle–brachial index using only a stethoscope? A pilot study. Family Practice 2009; 26: 22–26.