Lack of predictive correlation between peripheral arterial tone and colour flow Doppler parameters in men with erectile dysfunction


Correspondence: Daniel A. Shoskes, Department of Urology Cleveland Clinic Foundation, 9500 Euclid Ave Desk Q10-1, Cleveland, OH 44195, USA.




  • To evaluate whether peripheral arterial tone (PAT) can predict the results of penile colour flow Doppler in the evaluation of erectile dysfunction (ED).

Patients and Methods

  • Fifty men presenting to an ED clinic were tested with an Endo-PAT2000 machine (Itamar Medical, Caesarea, Israel), which assessed augmentation index (AI), a measure of arterial stiffness, and reactive hyperaemia index (RHI), a measure of endothelial vasodilatation.
  • Penile haemodynamics were measured and used to identify both arterial insufficiency (abnormal peak systolic velocity [PSV]) and veno-occlusive disease (abnormal end diastolic velocity [EDV]) after pharmacological erection with prostaglandin E1 using colour flow Doppler.
  • Between-group comparisons were carried out using a Wilcoxon rank-sum test for continuous variables and a chi-squared test for categorical variables. Simple and multivariable logistic regression analyses were used to analyse the correlation between outcome measures.


  • Using Doppler analysis, 58% of patients were found to have arterial insufficiency (abnormal PSV) and 48% had veno-occlussive disease (abnormal EDV).
  • Using the Endo-PAT machine, 44% of patients were found to have increased arterial stiffness (abnormal AI) and 54% had decreased endothelial relaxation (abnormal RHI).
  • Neither AI nor RHI were correlated with PSV or EDV. The closest association was between high AI and low PSV, with a sensitivity of 0.55 and specificity of 0.71.


  • In our patient cohort, PAT did not reliably predict the results of penile Doppler.
  • The two tests appear to measure different although potentially complementary aspects of the local and systemic vasculature.

peripheral arterial tone


erectile dysfunction


augmentation index


reactive hyperaemia index


peak systolic velocity


end diastolic velocity


cardiovascular disease


International Index of Erectile Function


It is accepted that arteriogenic erectile dysfunction (ED) is frequently a sign of systemic cardiovascular disease (CVD). The association between widespread atherosclerosis and ED is well documented, with The Massachusetts Male Aging Study estimating the age-adjusted risk of ED per 1000 man years in men with treated heart disease to be 58.3 cases and with treated hypertension to be 42.5 cases [1]. Given the prevalence of both ED and CVD, characterizing the pathophysiology of and relationship between these two entities has the potential to have a significant impact on a large number of patients. Risk factor profiles for CVD and ED are similar, and ED is believed to predict future cardiovascular events [1].

Vasculogenic ED is commonly related to impairment of smooth muscle relaxation (either endothelial-dependent or independent) or occlusion of penile arteries [2]. Vascular evaluation of ED with colour penile Doppler is aimed at identifying contributing arterial and venous abnormalities. Ultrasonography during erection allows measurement of peak systolic velocity (PSV), which has been correlated with the presence of arteriogenic causes of ED [3, 4] as well as end diastolic velocity (EDV), which can be used to assess veno-occlusive disease [5]. The limitations of colour Penile Doppler include operator-dependent measurements, a dependence on erection, and patient unease at the requirement for intracavernosal injection.

Impaired endothelial function is predictive of poor cardiovascular health [6, 7]. Peripheral arterial tone (PAT), as measured by the Endo-Pat2000 (Itamar Medical, Caesarea, Israel) is a non-invasive test of endothelial response to ischaemia. Using calculations derived from PAT, the reactive hyperaemic index (RHI) assesses change in pulse volume in response to local ischaemia and reperfusion, while the augmentation index (AI) is used to estimate arterial stiffness. The test is simple and not operator-dependent, requiring two finger probes, a blood pressure cuff and a computer to collect the data.

Given the above discussion, if PAT could accurately predict the findings of penile Doppler, it might serve as a less invasive and less expensive replacement or screen. The aim of the present study was to compare the results of PAT and colour penile Doppler in the evaluation and risk stratification of men with ED.

Materials and Methods

Fifty men who presented to an ED clinic and requested aetiological evaluation were tested with an Endo-PAT2000 machine according to the manufacturer's instructions. Probes were fitted on both fingers to allow measurement of arterial tone. A blood pressure cuff was placed on the left arm to occlude arterial blood flow for 5 min before being released. Software accompanying the machine was subsequently used to calculate the AI (normal <3%) and RHI (normal >1.8).

Usually at a second clinic visit, patients underwent evaluation of penile haemodynamics by duplex ultrasonography after intracavernosal injection with prostaglandin E1 (PGE1), according to standard protocols [8]. PSV was used to define arterial insufficiency (normal >30 cm/s). EDV was used to define venous insufficiency (normal <3 cm/s). Information regarding patient comorbidites was collected at each patient's initial visit and their degree of ED was assessed using the International Index of Erectile Function (IIEF) [9].

The Doppler measures PSV and EDV and the Endo-PAT measures RHI and AI were categorized based on the above-defined values of normal. Abnormalities of PSV and EDV as binary outcomes were used to assess the predictive power of Endo-PAT measures after adjusting for clinical covariates when appropriate. Between-group comparisons were carried out using a Wilcoxon rank-sum test for continuous variables and a chi-squared test was used for categorical variables. Simple and multivariable logistic regression analyses were used for analysis of both Doppler measures. Statistical analyses were performed using package R version 2.14.1 (R Development Core Team, and the rms package.


The mean (range) patient age was 51.1 (21–74) years and their mean IIEF score was 28.6. The mean (range) total testosterone level was 413 (176–890) ng/dL. In all, 16 patients (32%) had known comorbid hypertension, 10 patients (20%) had hyperlipidaemia, nine patients (18%) had coronary artery disease and five patients (10%) had diabetes. On penile Doppler analysis, 58% of patients had decreased arterial inflow as defined by PSV ≤ 30 cm/s and 48% had venous insufficiency defined by EDV ≥ 3 cm/s. Using the Endo-PAT machine, 54% had decreased endothelial relaxation (RHI ≤ 1.8) and 44% had increased arterial stiffness (AI ≥ 3). Patient demographics are shown in Table 1.

Table 1. Demographic characteristics of patients stratified by abnormal penile Doppler findings.
CharacteristicPSV > 30 cm/s, N = 21PSV ≤ 30 cm/s, N = 29P
Mean age, years50.551.60.852
Mean IIEF score29.6280.924
Mean total testosterone level, ng/dL4164100.729
Mean RHI score1.851.640.066
Mean AI score−1.445.790.122
Mean EDV, cm/s6.142.240.043
Hypertension, n (%)6 (28.6)10 (34.5)0.893
Hyperlipidaemia, n (%)3 (14.3)7 (24.1)0.616
Coronary artery disease, n (%)4 (19)5 (17.2)0.835
Diabetes, n (%)1 (4.8)4 (13.8)0.567
CharacteristicEDV < 3 cm/s, N = 26EDV ≥ 3 cm/s, N = 24P
Mean age, years45.257.90.001
Mean IIEF score30.726.30.364
Mean total testosterone level, ng/dL4074180.521
Mean RHI score1.781.670.232
Mean AI score−0.7816.580.345
Mean PSV, cm/s25.5310.171
Hypertension, n (%)6 (23.1)10 (41.7)0.269
Hyperlipidaemia, n (%)4 (15.4)6 (25)0.62
Coronary artery disease, n (%)3 (11.5)6 (25)0.385
Diabetes, n (%)2 (7.7)3 (12.5)0.925

The PAT and Doppler parameters correlated poorly (Fig. 1). Neither RHI nor AI was predictive of EDV or PSV in either univariate (Table 2) or multivariate models. AI and PSV represented the closest association, with AI ≥ 3 predicting PSV ≤ 30 cm/s with a sensitivity of 0.55 and specificity of 0.71. Sensitivity and specificity results of RHI in terms of predicting low PSV were 0.62 and 0.57, respectively. The ability of RHI ≤ 1.8 to predict EDV ≥ 3 cm/s had a sensitivity of 0.63 and a specificity of 0.54, and for AI ≤ 3% these values were 0.49 and 0.58, respectively (data not shown).

Figure 1.

Correlation between PAT and Doppler outcomes. The largest filled circles represent abnormal values for both PAT and Doppler measurements.

Table 2. Univariate analysis of Doppler outcomes. Association of listed variables with abnormal Doppler measurements of PSV and EDV
 Odds ratio (95% CI)P
Predicting PSV ≤ 30 cm/s  
RHI score0.57 (0.27–1.2)0.14
AI score1.6 (0.8–3.2)0.18
Age1.2 (0.44–3.0)0.78
IIEF score0.82 (0.3–2.2)0.69
Total testosterone level0.96 (0.43–2.2)0.92
Hypertension1.3 (0.39–4.5)0.66
Hyperlipidaemia1.9 (0.43–8.5)0.39
Coronary artery disease0.89 (0.21–4.0)0.87
Diabetes3.2 (0.33–31)0.32
EDV0.36 (0.17–0.78)0.0089
Predicting EDV ≥ 3 cm/s  
RHI score0.73 (0.36–1.5)0.39
AI score1.6 (0.82–3.0)0.17
Age6.8 (2.0–23.)0.0025
IIEF score0.57 (0.21–1.6)0.29
Total testosterone level1.1 (0.5–2.4)0.81
Hypertension2.4 (0.7–8.1)0.16
Hyperlipidemia1.8 (0.45–7.5)0.4
Coronary artery disease2.6 (0.56–11.)0.23
Diabetes1.7 (0.26–11.)0.57
PSV2.0 (0.8–4.8)0.14

In exploring other relationships between our measured outcomes, only EDV was associated with PSV. Not surprisingly, increasing age correlated with abnormal EDV (odds ratio 6.8, 95% CI 2.0–23, P = 0.003).


In our ED patient cohort, PAT did not reliably predict arterial or venous findings on penile colour flow Doppler; therefore, in the patient with ED who desires further evaluation to determine aetiology, Endo-PAT is not an adequate screening test or replacement for penile Doppler. Nevertheless, PAT findings may give complementary information regarding cardiovascular risk, independently of the aetiology of ED.

There are several potential confounding factors when considering the lack of correlation between Doppler and Endo-PAT demonstrated by the present results. First, the association may be weak and therefore not detectable with our sample size. Second, any association may be most evident at the extremes of cardiovascular health. Only severity of ED has been shown to be an independent predictor of risk of coronary artery disease [10]. One study of men in a health screening project found only moderate and severe ED were associated with an increase in relative risk for the development of CVD as compared with those without ED [11]. With a mean (range) IIEF score of 28.6 (5–61), our group of patients represented the spectrum of ED and, therefore, our patients' systemic CVD may have been less clinically significant than in a purely severe ED cohort. Last, ED is known to precede coronary artery disease (both symptomatic and asymptomatic) by several years [12]. Given the young mean age of our study participants, we must also consider that these results are capturing a timepoint after the onset of ED and before the onset of systemic disease. Nevertheless, the majority of patients who present with ED do not represent the extremes of disease severity or age, and as such it is in the intermediate groups that the usefulness of these diagnostic tests is most relevant.

Endo-PAT is a non-invasive measure of peripheral vasodilatory response which has been associated with both endothelial dysfunction as well as adverse cardiac events, as defined by myocardial infarction, need for revascularization, cardiac hospitalization and cardiac death [6, 13]. More specifically, abnormalities in response to hyperaemic environments measured by Endo-PAT have been found to correlate with coronary microvascular endothelial dysfunction, as measured by coronary angiography [14]. The usefulness of Endo-PAT as a non-invasive tool for evaluating current and future vascular risk is well documented in the cardiac literature; therefore, it is a clinically relevant diagnostic test when making treatment decisions for patients at risk of CVD.

Vasculogenic ED is an early marker of systemic vascular disease, therefore, the relationship between ED and CVD is important, not only because of the large number of patients affected by these two conditions, but also because of the potential for early identification of higher cardiovascular risk. Our results suggest that PAT, as measured by Endo-PAT is not a useful surrogate of penile arterial or venous insufficiency during evaluation for ED; however, the relationship between these two tests deserves continued consideration as they measure potentially complementary aspects of the local and systemic vasculature and, together, may provide a more complete picture of the aetiology of disease as well as of systemic risks that may influence treatment algorithms with regard to both urological and cardiovascular health.

In conclusion, neither RHI nor AI as measured by Endo-PAT was predictive of EDV or PSV as measured by duplex ultrasonography; therefore, in men desiring further evaluation for ED, Endo-PAT cannot replace Doppler testing. Nevertheless, given the established relationship between abnormal PAT and systemic CVD, both tests may give complementary and valuable information for the patient.

Conflict of Interest

Daniel A. Shoskes is a Consultant for Farr Labs and Eli Lilly.