Article Text

This article has a correction. Please see:

Download PDFPDF
Physical activity throughout life reduces the atherosclerotic wall process in the carotid artery
  1. M Sandrock1,
  2. C Schulze1,
  3. D Schmitz2,
  4. H-H Dickhuth1,
  5. A Schmidt-Trucksaess3
  1. 1
    Freiburg University Hospital, Centre for Internal Medicine, Department for Rehabilitative and Preventative Sports Medicine, Freiburg, Germany
  2. 2
    Heidelberg University Hospital, Centre Mannheim, Department for Anaesthesia, Mannheim, Germany
  3. 3
    Technical University Munich, Department for Sports Medicine, Munich, Germany
  1. Dr M Sandrock, Freiburg University Hospital, Centre for Internal Medicine, Department for Rehabilitative and Preventative Sports Medicine, Freiburg, D-79106, Germany; markus.sandrock{at}


Background: Good cardiorespiratory fitness has been associated with a reduced risk for clinical events of atherosclerotic vascular diseases. It is still unclear how this relates to a slower progression of the early atherosclerosis wall process.

Method: Using a dynamic model, we generated new parameters for describing the pathologic wall process in the carotid artery, based on an automatic layer detection system. In this study, we scrutinised the influence of two ultrasonographic parameters, intima–media thickness (IMT) and roughness, by comparing two groups: a healthy inactive group (PIP) (mean (SD) age 64.37 (5.10) years; n = 50) and a healthy lifelong physically active group (PA) (mean (SD) age 64.48 (3.45) years; n = 51). All subjects underwent a blood test, spiroergometry, echocardiography and carotid ultrasound examination.

Results: There was a significant difference in the well known risk factors for cardiovascular disease (for example, high density lipoprotein cholesterol, triglyceride) between groups. PIP compared to PA had a significantly higher roughness (PIP 0.073 (0.015) vs PA 0.065 (0.0156); p<0.01). No significantly higher IMT was found for PIP (PIP 0.89 (0.18) vs PA 0.90 (0.22); p = 0.63) compared to PA. In this cross sectional study of middle aged men, Vo2max was inversely associated with carotid atherosclerotic parameters.

Conclusion: In this study, good cardiorespiratory fitness was associated with an increase of the proven risk factors and a reduction of atherosclerosis in the common carotid artery. Roughness seems to be significantly more sensitive than IMT for characterising the changes of the arterial wall. We suggest measuring roughness in addition to IMT to gain additional information about the atherosclerotic wall.

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

Increasing physical activity and good cardiorespiratory fitness is associated with a decrease in the incidence of cardiovascular events.1 2 Physical training improves endothelial function, exercise capacity, and collateralisation in patients with coronary artery disease,35 chronic heart failure6 7 and peripheral artery disease.8 Physical activity also is associated with improved mood, body weight, blood pressure, insulin sensitivity, and haemostatic and inflammatory variables.9

Physical activity for itself,2 physical activity combined with a low fat diet10 or comprehensive lifestyle modification,2 together with concomitant improvement in cardiorespiratory fitness, slows the progression of angiographically quantified coronary atherosclerosis in patients with coronary heart disease. Thus, physical activity has a proven direct influence on the atherosclerotic wall process.5 11 12

Atherosclerosis in human arteries develops from an asymptomatic phase to a manifest disease over decades. The occurrence of clinically significant atherosclerotic lesions and consequent symptomatic atherosclerotic vascular diseases increases progressively in middle age. Over the last few years, researchers have tried to quantify the atherosclerotic wall process in humans by measuring the intima–media thickness (IMT). The IMT of the common carotid artery (CCA) can be used to predict major cardiovascular events like myocardial infarction and stroke,1214 and the involvement of other arterial beds with atherosclerosis.15

Previous studies have indicated that an increase in IMT is associated with an increased risk of myocardial infarction and stroke without a history of cardiovascular disease.6 Additional studies have shown that IMT can be viewed as a descriptive general index of individual atherosclerosis.16

Structural changes of the arterial wall, like granulations, might have predictive power with respect to future cardiovascular events.

We recently presented an automatic procedure for measuring IMT,17 18 which can identify the “lumen–intima” and “media–adventitia” boundaries with high accuracy. Based on this procedure we presented additional parameters to quantify the arterial wall process. Roughness quantifies the irregularities in carotid wall thickness.19

The following hypotheses were scrutinised in this study: (1) Is lifetime physical activity associated with a reduction of the atherosclerotic wall process? (2) Does roughness correlate more closely to the Framingham score than IMT?


Study population

The study group consisted of 101 healthy male subjects between the ages of 55–75 years. Fifty were physically inactive persons (PIP; mean (SD) age 64.37 (5.10) years) for at least 30 years, with less then two physical activities per month. Fifty-one were active persons (PA; mean (SD) age 64.48 (3.45) years), with at least three physical activities per week for the last 30 years. Subjects for the physically active group were chosen randomly at sports clubs. Subjects for the control group were recruited on the street. The subjects underwent a detailed oral interview and completed a standardised questionnaire to evaluate their physical activity. Subjects were selected based on this information.

According to their histories, none of the subjects had any of the following afflictions: a proven cardiovascular arterial disease (CAD), stroke, diabetes or myocardial infarction. All subjects reported being non-smokers.

All PIP and PA had a normal electrocardiogram (ECG) (12 leads) and none of the subjects had a resting blood pressure of >160/95 mm Hg or took antihypertensive drugs.

During exercise until subjective exhaustion, either using a treadmill spiroergometer or bike ergometer, no subject had cardiac symptoms or abnormalities in the permanently registered ECG.

None of the older subjects showed signs of regional wall motion disturbances during the resting echocardiography. None of the subjects in either group took antihypertensive drugs, antidiabetic agents or drugs for lipid disorders (table 1).

Table 1 Anthropometric data of the subjects with results for significance of the non-parametric Mann–Whitney test

The hospital ethics committee approved the study protocol, and informed consent was obtained from each patient before the examination.

Risk factor profile

All subjects underwent an ultrasound examination of the heart, a resting ECG, and a blood test for the lipid profile, red blood cells, glucose value and serum factors. Blood samples (15 ml) were collected for all subjects. After overnight fasting, whole blood samples were taken in the supine position at 08:00 h from the antecubital vein with a 20 gauge needle by applying minimal tourniquet force. After blood was taken, all subjects went home for lunch and prepared for the spiroergometry. A detailed medical history with regard to risk factors was recorded for each subject.

Cardiorespiratory fitness

Cardiorespiratory fitness was assessed by using a maximal but symptom limited exercise test on a cycle ergometer (Lode, 91 subjects) or on a treadmill spiroergometer (Excalibur, 10 subjects) until subjective exhaustion. Subjects were allowed to choose between a bicycle or treadmill ergometer, depending on athletic preference (PA: 46 cycle, 5 treadmill; PIP: 45 cycle, 5 treadmill).

The exercise tests were performed by using a standardised testing protocol starting at 100 W, with an increase of 25 W every 3 min. The treadmill test was performed at 6 km/h beginning with 0% incline, and then a 5% incline increase every 3 min with 30 s rest between every increase. All tests were performed at 16:00 h after the ultrasound examination.

Oxygen consumption was measured by using the breath-by-breath method of respiratory gas exchange in oxygen uptake (Jaeger Oxycon Delta, Würzburg, Germany). The Vo2 peak was defined as the highest value for the plateau during spiroergometry, and was indexed by body weight (ml*min−1*kg−1).


Colour Doppler examinations were performed with a commercially available system (Toshiba SA, Japan) using a 2.5 MHz transducer according to the guidelines of the American Society of Echocardiography. All examinations were carried out by the same person for the entire study. The echocardiographic evaluation of heart valve regurgitation followed a structured protocol.

M mode echocardiography was used to determine the left ventricular end-diastolic diameter (LVIDd) and left ventricular (LV) posterior wall thickness (PWTd), as well as septum thickness (IVSTd). Total left ventricular end diastolic diameter (EDD) (TEDD  =  LVEDD + LVIDd + PWTd) was determined on the mitral valve plane (TEDDm) and the papillary muscle plane (TEDDp).

Total LV longitudinal diameter (TLD) was determined in the four chamber view. Parameters were measured three times, and mean values were used to calculate the LV total diastolic volume (TDV (ml)) and absolute heart volume (AHV (ml)) using the following formulae20:

TDV  =  [(TEDDm2 * 0.785) + (TEDDp2 * 0.435)] * TDL/2000 (1)

AHV  =  (TDV * 2.432) + 130 (2)

The relative heart volume was calculated by dividing AHV by the weight of the subject.

Ultrasound examination

After at least 15 min of rest in a supine position, every patient was scanned by the same experienced physician. The examining physician was blinded with regard to which group the subjects were in.

We used an ultrasound scanner (Toshiba SSA-380A, Japan) with a high resolution 10 MHz transducer and an aperture of 32 mm. The carotid tree was scanned and the transducer positioned cross sectional to the CCA. The lumen was maximised and the position changed until the typical double lines of the intima and adventitia boundaries of the arterial wall could be seen. The scanning itself was performed of the lateral and anterolateral view of the CCA. Localised thickening was considered to be carotid plaque, when the echo luminance included high echogenic structures encroaching into the vessel lumen.

A total of eight images were stored for every subject; two lateral and two anterolateral views for the left and right carotid arteries. Images were stored via the cine view function of the Toshiba SSA-380A and digitised, including the simultaneously recorded ECG at the top of the R wave on digital disk (Sony DKR-700, Japan). All examinations were additionally stored on videotape.

Analysis of the images

All offline measurements and analyses were done by the same person blinded to all clinical data. The measurement error was calculated as 0.3%. To analyse the intima–media layer, we used a length of 1 cm from the beginning of the bulb backwards into the CCA.14 We measured the intima–media layers at the far wall of the distal CCA and not at the near wall, because detecting the trailing edges of intima–media layers in the near wall can lead to problems when interpreting the interfaces.14 Results from the left and right carotid arteries have been averaged.


Intima–media thickness (IMT)

The IMT is defined as the mean value of the total intima–media layer in a chosen segment.


Roughness describes the irregularities of the IMT in a defined segment. We used an international standard ISO 4287/1 “surface roughness; terminology, surface and its parameters”, and applied this standard to quantify the IMT profile deviations. Figure 1 shows IMT surface roughness for one subject in the PA group.

Figure 1 Part of the original ultrasound image (A) of the right common carotid artery of a physically active subject with intimal and adventitial tracing lines of the far wall (upper part) and graph of the assessed intima–media roughness (lower part) of the most distal (left side) and adjunctive proximal cm (right side). Arrow indicates the zero value of the ordinate, corresponding to the regression line rotated into the horizontal position, and the red dotted line represents the average value of the intima–media roughness. The blue area is equivalent to the absolute profile deviation of the intimal–media thickness (IMT) from the regression line. The green line represents the profile deviation of the IMT from the regression line.

Assessment of individual’s cardiovascular heart disease risk level

The individual cardiovascular heart disease (CHD) risk level was estimated for each subject based on the Framingham CHD score sheets presented by Wilson et al.21 This scoring system was applied to subjects who were between the ages of 30–79 years. Detailed information on the algorithm for determining the risk score can be found in the original report.


All data were calculated using the software program SPSS (Software Package for the Social Sciences) for Windows, version 12. The arithmetic mean and standard deviations (SD) were used for descriptive statistics. Means of PIP and PA, respectively, were compared using a non-parametric Mann–Whitney test. A value of p<0.05 was considered to be significant, and a value of p<0.01 highly significant.

We used a logistic regression model to rank all measured parameters. This is most useful when you want to model the event probability for a categorical response variable with two outcomes.

The receiver operating characteristic (ROC) curve procedure22 was used to evaluate the difference between both groups in this study. The area under the curve represents the probability that the assay result for a randomly chosen positive case will exceed the result for a randomly chosen negative case.

Linear regression was used to model the value of the Framingham score based on its linear relationship to IMT and roughness. The regression model assumes that there is a relationship between the dependent variable and each predictor.


PIP compared to PA had significantly lower triglyceride and lower high density lipoprotein cholesterol (HDL) values. No significant differences were found for cholesterol or low density lipoprotein cholesterol (LDL). PIP compared to PA had a significantly lower left ventricular mass (ml/kg), left ventricular mass per kg, left ventricular internal diameter, and lower cardiorespiratory fitness. The characteristics showed no significant differences apart from weight and body mass index (BMI). The mean (SD) Framingham risk score in PA (11.4 (1.0)) was significantly higher than in PIP (12.2 (1.4)) (tables 1 and 2).

Table 2 Correlation coefficients for the logistic regression for intima–media thickness (IMT) mean and roughness as independent parameters for the risk factor profile with significance (Sig) <0.05 for a significant correlation (Cor)

In the ultrasound analysis, PIP compared to PA had a significantly higher roughness and higher wall inhomogeneity. No significantly higher IMT was found for PIP compared to PA (fig 2, tables 3 and 4).

Figure 2 Scatterplot of roughness versus Vo2max.
Table 3 Results of intima–media thickness (IMT) mean, maximum, roughness of the two groups with results for significance of the non-parametric Mann–Whitney test
Table 4 Logistic (1 = PA, 2 = PIP) regression analysis for ultrasound parameters

Logistic regression analysis showed that by using the relative heart size, triglyceride, and roughness, 82% of the subjects could be correctly allocated into their established group (table 3). Using the ultrasound parameter roughness, 70% could be allocated into their correct group (table 5).

Table 5 Logistic (1 = PA, 2 = PIP) regression analysis for ultrasound parameters

The area under the ROC curve for roughness was 0.65 (significance 0.01), which was significantly different. For IMT there was a mean of 0.50 (significance 0.96) and IMT max of 0.53 (significance 0.66), respectively, with no statistically significant differences (fig 3).

Figure 3 Receiver operating characteristic (ROC) curve for intima–media thickness (IMT) mean, IMT max and roughness. Area under the curve (AUC) as value with significance <0.05 for a significant difference (IMT mean: 0.5 AUC and significance 0.96; IMT max: 0.53 AUC and significance 0.66; roughness: 0.65 AUC and significance <0.01).

Linear regression analysis for IMT showed no significant correlation to the risk factor profile (39%). Roughness showed a significant correlation to the risk factor profile (roughness: 52%).

The correlations between the Framingham risk score and roughness (Pearson: roughness/Framingham = 0.28; p>0.01) in contrast to IMT (Pearson: IMT/Framingham = 0.08; p = 0.66) were significantly higher.


The main findings of this cross sectional study are: (1) lifetime physical activity correlates with a decrease of the proven risk factors and a reduction of atherosclerosis in the CCA; and (2) roughness seems to be more sensitive than IMT for characterising arterial wall changes.

Only a few population based studies have examined the relationship between cardiorespiratory fitness and early atherosclerosis. In those cross sectional studies in middle aged men, Vo2max was inversely associated with carotid bifurcation IMT. Similarly, in this study, we found an inverse relationship between Vo2max and the indicators of carotid atherosclerosis.

One explanation for the inverse association between Vo2max and the increase in the indicators of carotid atherosclerosis could be that physical activity slows the progression of atherosclerosis by improving cardiorespiratory fitness, as suggested by clinical trials in patients with coronary artery disease.23 Therefore, further prospective studies should be done to confirm this.

In the present study, the positive risk factor profile in physically active subjects (HDL and triglyceride values) was reduced by physical activity. This is supported by the sonomorphological findings of this study.

The logistic regression analysis in our study showed that using ultrasonographic parameters (69%) could correctly predict nearly the same number of subjects as using the “classic” risk factors like weight, BMI, and blood parameters (82%). We used this analysis to show the possibility and potentials of ultrasound analyses. Our results show that ultrasound seems to be a legitimate method of screening.

What is already known on this topic

Increasing physical activity and good cardiorespiratory fitness is associated with a decrease in the incidence of cardiovascular events.2 25 However, little is known about the effects of physical activity on the effector organ of the atherosclerosis—the intima media layer.

What this study adds

This study shows for the first time that the effect of physical activity can be measured with the use of ultrasound examination.

The results of the ROC analysis and of the ultrasonographic parameters in our study showed an advantage for the newly developed parameters compared to the “established” parameters mean, max, and min IMT. In previous studies, we showed that roughness is an independent parameter for examining the atherosclerotic wall process, and is also not dependent on IMT measurement. In his histological work, Stary24 showed that minimal accumulations of lipid filled macrophages (type I and II lesions) do not disorganise or deform the arterial wall, and stabilise at that level. Yet after puberty, an increasing number of people develop lesions in the same susceptible locations with additional features that histologically render them immediate precursors (type III lesion, pre-atheroma) of fully formed atherosclerosis (type IV lesion, atheroma). After the development of atheroma, additional mechanisms and components may begin to accelerate wall thickening and lumen reduction (type V–VIII lesions). Roughness can detect this development more sensitively than IMT. Based upon those facts, the different ultrasonographic parameters have different strengths and weaknesses.

The regression analysis showed that IMT did not correlate to the risk factors in the same way as roughness. This is an indication for the unspecific swelling of IMT during the atherosclerotic wall process described by Stary.24 We showed that the roughness could significantly differentiate between both groups. The correlation of the Framingham score system strengthens the thesis for roughness, but not for IMT. We added this analysis to this study to show the correlation of the new parameter to the proven risk stratifications (such as ARIC, WOSCOPS).

Possible limitations of our study should also be mentioned. As this was a cross sectional study, the causal relationships may not be straight forward and further studies should be done. The number of subjects is large enough to show significance, since the variation of an automatic system is much lower than manual or semi-automatic analyses. The performance the parameters may be influenced by the different degree of exhaustion as to be seen at the level of maximal lactate.

All together, our cross sectional study shows that physical activity over a lifetime significantly reduces the atherosclerotic wall process. Additional information about this process can be quantified through an ultrasonographic examination. The new ultrasonographic parameters seem to characterise the changes of arterial wall process better than IMT, and we suggest adding this parameter to the ultrasound analysis of arterial wall characterisation.



  • Competing interests: None.

  • Patient consent: Obtained.

Linked Articles