Meta-Analysis of the Effect of Different Exercise Mode on Carotid Atherosclerosis

(1) Background: There is increasing evidence showing the health benefits of exercise on carotid atherosclerosis. However, little is known about the different exercise modes for carotid atherosclerosis. This study was designed to perform a meta-analysis of effect of different exercise modes on carotid atherosclerosis so as to provide evidence-based suggestions for the prevention and management of cardiovascular and cerebrovascular diseases. (2) Methods: Six databases were systematically searched to identify randomized trials that compared exercise to a non-exercise intervention in patient with carotid atherosclerosis. We a priori specified changes in cIMT, TC, LDL-C, and HDL-C biomarkers as outcomes. (3) Results: Thirty-four trials met the eligibility criteria, comprising 2420 participants. The main analyses showed pronounced differences on cIMT (MD = −0.06, 95%CI (−0.09, −0.04), p < 0.00001, TC (MD = −0.41, 95%CI (−0.58, −0.23), p < 0.00001), LDL-C (MD = −0.31, 95%CI (−0.43, −0.20), p < 0.00001), and HDL-C (MD = 0.11, 95%CI (0.04, 0.19), p = 0.004), which significantly reduced the risk factors of carotid atherosclerosis disease. In the different exercise modes, the effect was pronounced for aerobic exercise for all outcomes except TC; high-intensity interval exercise also showed significance for all outcomes except TC and HDL-C; aerobic exercise combined with resistance exercise did not affect any outcome except HDL-C; (4) Conclusions: Exercise has a prominent prevention and improvement effect on carotid atherosclerosis. In the perspective of exercise pattern, aerobic exercise and high-intensity intermittent exercise can improve carotid atherosclerosis; however, aerobic exercise has a more comprehensive improvement effect.


Introduction
Arteriosclerosis is a non-inflammatory disease of arteries, which can lead to serious cardiovascular diseases, cerebrovascular diseases, peripheral arterial disease, and type 2 diabetes and greatly threaten human life and health. Most arteriosclerosis is located in the major and medium arteries, including the coronary artery, carotid artery, and brain base ring. In 2020, the global prevalence of carotid plaque in the population aged 30-79 was estimated at 21.1% [1], and about 8.5 million people over 40 years old have peripheral arterial disease in USA every year [2] due to arteriosclerosis. Moreover, ischemic stroke caused by carotid arteriosclerosis accounts for 15% to 20% of cerebrovascular disease, and the more severe the carotid artery stenosis, the higher the risk of stroke and the more severe the disease [3].
Atherosclerosis is the pathological basis of cardiovascular disease, while the formation of atherosclerosis is related to dyslipidemia [4,5]. However, regarding how to prevent and treat atherosclerosis, the state of the reduction of its incidence rate is still in the exploratory Figure 1 shows the flow diagram of the selection process of the literature. In total, 3405 potentially eligible articles were obtained by searching various databases and then were checked manually. A total of 1936 documents remained by removing duplicate records from EndNote x9 software. Next, 1686 irrelevant studies were excluded by preliminary screening of the titles and abstracts of the literature, leaving 223 remaining documents. We excluded 189 of these based on further reading of the full text, with the common reasons for exclusion including having non-RCT design, unrelated outcomes, or no continuous exercise intervention or no control group, etc. Finally, a total of 34 RCT articles were deemed eligible for inclusion and quantitative synthesis for the meta-analysis.

Study Characteristics
The basic characteristics of the included studies are shown in Table 1. All included articles were published between 2004 and 2020. A total of 2420 individuals were involved in the 34 eligible articles were that were included in this meta-analysis. In this meta-analysis, 21 studies were published in English, and 13 studies were published in Chinese. Trial sample sizes ranged from 21 to 160 participants. Regarding exercise intervention mode, 25 involved aerobic exercise intervention, 8 involved a high-intensity interval exercise, and 6 involved aerobic exercise combined with resistance exercise. All the control groups of the included articles did not receive exercise intervention. There were 28 studies deemed to have high quality and 6 studies low quality after assessing the quality of each study by the Jadad scale. continuous exercise intervention or no control group, etc. Finally, a total of 34 RCT articles were deemed eligible for inclusion and quantitative synthesis for the meta-analysis.

Study Characteristics
The basic characteristics of the included studies are shown in Table 1. All included articles were published between 2004 and 2020. A total of 2420 individuals were involved in the 34 eligible articles were that were included in this meta-analysis. In this meta-analysis, 21 studies were published in English, and 13 studies were published in Chinese. Trial sample sizes ranged from 21 to 160 participants. Regarding exercise intervention mode, 25 involved aerobic exercise intervention, 8 involved a high-intensity interval exercise, and 6 involved aerobic exercise combined with resistance exercise. All the control groups of the included articles did not receive exercise intervention. There were 28 studies deemed to have high quality and 6 studies low quality after assessing the quality of each study by the Jadad scale.

Risk of Bias
The risk of bias of all included articles was evaluated by the Cochrane Collaboration's Risk of Bias 2 (RoB2) tool, and these results are summarized in Figures 2 and 3. Overall, 31 of 34 (91.2%) trials described the process of random sequence generation, and they were low-risk in the fields of random sequence generation. Most studies were classified as having an unclear risk in allocation concealment, and only three were low-risk. A high risk of bias was detected in the domain of blinding of participants and personnel; only six were at low risk; a low risk of bias was observed in blinded outcome assessment except five of them that were at high risk of bias. A low risk of incomplete outcome data bias was observed in all of the studies (34, 100%). With regard to selective outcome reporting bias, most studies were determined as low-risk. All studies were graded as unclear risk of other bias. These results are summarized in Figure 2.

Effects of Different Exercise Mode on cIMT
Carotid intima-media thickness (cIMT) is a crucial risk factor for cardiovascular health. A total of 26 articles used cIMT to evaluate the therapeutic effect of exercise on carotid atherosclerosis [12][13][14][15][17][18][19][20]22,23,[26][27][28][29][30][31][33][34][35][37][38][39][40][41]43,44]. The random-effects model was performed to integrate the results. The results showed that, overall, exercise significantly reduced cIMT (MD = −0.06, 95%CI (−0.09, −0.04), p < 0.00001; Figure 4), which was statistically significant compared with the control group. The I-squared (I 2 ) of this results as > 60%, indicating high heterogeneity; therefore, we performed a subgroup analysis based on exercise patterns to discuss the source of heterogeneity. In order to further explore the source of heterogeneity, we further classified and excluded the original literature, and the results showed that I2 of the cIMT effect size could be reduced to 58% if Adams 2017, Kadoglou 2013 (E2), Wang 2018 and Zhang 2012, and Zhang 2020 were excluded. However, we found that there were no differences in the sample size and the intervention of these articles compared to the other literature.

Risk of Bias
The risk of bias of all included articles was evaluated by the Cochrane Collaboration's Risk of Bias 2 (RoB2) tool, and these results are summarized in Figures 2 and 3. Overall, 31 of 34 (91.2%) trials described the process of random sequence generation, and they were low-risk in the fields of random sequence generation. Most studies were classified as having an unclear risk in allocation concealment, and only three were low-risk. A high risk of bias was detected in the domain of blinding of participants and personnel; only six were at low risk; a low risk of bias was observed in blinded outcome assessment except five of them that were at high risk of bias. A low risk of incomplete outcome data bias was observed in all of the studies (34, 100%). With regard to selective outcome reporting bias, most studies were determined as low-risk. All studies were graded as unclear risk of other bias. These results are summarized in Figure 2.  .

Risk of Bias
The risk of bias of all included articles was evaluated by the Cochrane Collaboration's Risk of Bias 2 (RoB2) tool, and these results are summarized in Figures 2 and 3. Overall, 31 of 34 (91.2%) trials described the process of random sequence generation, and they were low-risk in the fields of random sequence generation. Most studies were classified as having an unclear risk in allocation concealment, and only three were low-risk. A high risk of bias was detected in the domain of blinding of participants and personnel; only six were at low risk; a low risk of bias was observed in blinded outcome assessment except five of them that were at high risk of bias. A low risk of incomplete outcome data bias was observed in all of the studies (34, 100%). With regard to selective outcome reporting bias, most studies were determined as low-risk. All studies were graded as unclear risk of other bias. These results are summarized in Figure 2.  .
Adams 2017, Kadoglou 2013 (E2), Wang 2018 and Zhang 2012, and Zhang 2020 were excluded. However, we found that there were no differences in the sample size and the intervention of these articles compared to the other literature. Thirty studies were conducted as a subgroup analysis of exercise intervention patterns, including 20 on aerobic exercise [14,15,20,22,23,[26][27][28][29][30][31]33,[36][37][38][39][40][41]43,44] 6 on high-intensity interval exercise [12,18,20,23,27,34] and 6 on aerobic exercise combined with resistance exercise [13,17,19,26,30,35]. The results showed that cIMT was  Figure  5), but the changes were not statistically significant. The heterogeneity of subgroup analyses was also prominent from the research results. To explore the source of heterogeneity of the subgroup, sensitivity scores were used to analyze the excluded studies one by one and evaluate the cIMT effect size of each study. However, the results showed that there was little difference in heterogeneity among different studies, and the elimination of any article had little influence on the effect size of cIMT, and the results of the meta-analysis were relatively stable.

Effects of Different Exercise Mode on TC
Twenty-two trials used TC to evaluate the clinical effect of exercise on atherosclerosi s [12,13,[16][17][18][19][20][21][23][24][25][26][27][28]32,34,36,[38][39][40]42,43]. A fixed-effect model was used for merge the results. The overall effect of the studies showed that exercise significantly decreased TC compared with the control group (MD = −0.41, 95%CI (−0.58, −0.23), p < 0.00001; Figure 6). The I 2 of analysis was 92%, indicating high heterogeneity. We found that the result of The heterogeneity of subgroup analyses was also prominent from the research results. To explore the source of heterogeneity of the subgroup, sensitivity scores were used to analyze the excluded studies one by one and evaluate the cIMT effect size of each study. However, the results showed that there was little difference in heterogeneity among different studies, and the elimination of any article had little influence on the effect size of cIMT, and the results of the meta-analysis were relatively stable.
Although there was high heterogeneity of the effect of exercise on HDL-C, we found that it also has little influence on HDL-C effect size by eliminating any certain article in sensitivity analysis, so the analysis results of the meta-analysis were relatively stable.

Adverse Events, Sensitivity Analysis, and Publication Bias
Adverse events were not found in all included studies. Hence, this information could not be searched from the RCTs analyzed. The results of this study had high heterogeneity. By removing single studies, the sensitivity analyses showed no obvious changes in the statistical significance of all primary or secondary outcomes. The publication bias of the results in the meta-analysis was evaluated using the Egger's test (Figures 12-15). The Egger's test of cIMT and HDL-C was 0.987 and 0.702, indicating there is no publication bias; however, the Egger's test results of TC and LDL-C are 0.009 and 0.014, which are lower than 0.05, suggesting that there were a certain publication bias. Figure 10.Meta−analyses of the effect of exercise on HDL−C compared with the control group [12,13,[16][17][18][19][21][22][23][24][25][26][27][28][29]31,32,[34][35][36]39,42,43].

Adverse Events, Sensitivity Analysis, and Publication Bias
Adverse events were not found in all included studies. Hence, this information could not be searched from the RCTs analyzed. The results of this study had high heterogeneity. By removing single studies, the sensitivity analyses showed no obvious changes in the statistical significance of all primary or secondary outcomes. The publication bias of the results in the meta-analysis was evaluated using the Egger's test (Figures 12-15). The Egger's test of cIMT and HDL-C was 0.987 and 0.702, indicating there is no publication bias; however, the Egger's test results of TC and LDL-C are 0.009 and 0.014, which are lower than 0.05, suggesting that there were a certain publication bias.  [12,13,[16][17][18][19][21][22][23][24][25][26][27][28][29]31,32,[34][35][36]39,42,43].

Discussion
This study was designed to evaluate the effects of aerobic exercise, resistance exercise, high-intensity interval training and combined exercise on atherosclerosis, focusing on risk reduction for individuals who had not yet progressed to cardiovascular and cerebrovascular diseases.
cIMT was the crucial indicator for evaluating carotid atherosclerosis. Studies have shown that cIMT thickening was the early clinical manifestation of atherosclerosis [46]. Mounting evidence indicates that exercise produces significant physiological and health benefits and prevents or delays the development of atherosclerosis in humans. From the perspective of evidence-based medicine, this meta-analysis showed that exercise can obviously reduce cIMTs, significantly reducing the risk factors of cerebrovascular disease. Seals DR and Che L's article also confirmed that exercise training is an effective non-pharmacological treatment for improving carotid artery stiffness in young and older individuals [47,48]. In addition, the result of the subgroup analysis based on different exercise

Discussion
This study was designed to evaluate the effects of aerobic exercise, resistance exercise, high-intensity interval training and combined exercise on atherosclerosis, focusing on risk reduction for individuals who had not yet progressed to cardiovascular and cerebrovascular diseases.
cIMT was the crucial indicator for evaluating carotid atherosclerosis. Studies have shown that cIMT thickening was the early clinical manifestation of atherosclerosis [46]. Mounting evidence indicates that exercise produces significant physiological and health benefits and prevents or delays the development of atherosclerosis in humans. From the perspective of evidence-based medicine, this meta-analysis showed that exercise can obviously reduce cIMTs, significantly reducing the risk factors of cerebrovascular disease. Seals DR and Che L's article also confirmed that exercise training is an effective non-pharmacological treatment for improving carotid artery stiffness in young and older individuals [47,48]. In addition, the result of the subgroup analysis based on different exercise modes showed that cIMTs were decreased significantly after aerobic exercise and high-intensity intermittent training. No significant effect was observed on cIMT by aerobic exercise combined with resistance exercise. Carpio-Rivera E et al. showed that regular physical activity has potential benefits for arterial elasticity, especially aerobic exercise [49]. Evidence from a recent meta-analysis also suggests that aerobic training is the most effective type of exercise modality to improve blood pressure and arterial stiffness [50]. These were consistent with our research results. Our study also found that high-intensity interval training can improve carotid atherosclerosis as well as aerobic exercise.
The mechanisms by which physical activity counteracts arterial stiffening are not well-known. In order to find the intervention mechanism of aerobic exercise on carotid atherosclerosis, we further explored the effect of exercise on lipidemia metabolism. Because dyslipidemia was considered as a critical risk factor for atherosclerosis, including TC, LDL-C, and HDL-C, the formation of atherosclerosis was related to the deposition of a large amount of TC in blood vessels [51,52]. Reducing the deposition of TC in blood vessels can lessen the formation of atherosclerosis. LDL-C protein particles can carry cholesterol; if LDL-C is excessive, the cholesterol carried by LDL-C will accumulate on the arterial wall, leading to atherosclerosis [26]. In addition, the occurrence of atherosclerosis was negatively correlated with the serum HDL-C in the human body, so this is an important way to inhibit the formation of atherosclerosis by improving HDL-C level. Both clinical drug therapy or exercise intervention mainly focus on regulating blood lipids in the treatment of atherosclerosis [53]. Many studies have shown that exercise intervention can greatly improve blood lipids and lipid metabolism levels, thus further improving the formation of atherosclerosis [54].
The results of this meta-analysis showed that exercise intervention could significantly decrease the content of TC and LDL-C and increase the level of HDL-C to prevent and improve atherosclerosis [55,56]. The results of the subgroup meta-analysis showed that aerobic exercise had the remarkable effect of reducing LDL-C and increasing HDL-C. Highintensity intermittent exercise has a better effect in reducing LDL-C, but it has no obvious effect on TC and HDL-C. Aerobic exercise combined with resistance exercise showed a significant effect on HDL-C but not TC and LDL-C. The Egger's analysis of this study shows that there is publication bias in the improvement effect of exercise intervention on TC and LDL-C, and there is no significant publication bias in the improvement effect of cIMT and HDL-C, and the meta-analysis results were relatively stable. These findings further confirm that aerobic exercise can prevent the formation of atherosclerosis by improving dyslipidemia [12,15,41], and high-intensity intermittent exercise may also play a certain role in regulating dyslipidemia.
According to the above research results, the physiological mechanism of aerobic exercise intervention on carotid atherosclerosis includes the following two aspects: first of all, exercise changes the habits of sedentary and reduces the level of risk factors causing the disease of atherosclerosis [8]; secondly, aerobic exercise can accelerate and improve the activity of the lipoprotein enzyme in the body and the metabolic decomposition of TC and LDL-C, reduce the total blood lipids, and increase the level of high-density lipoprotein [57,58] so as to further prevent and improve atherosclerosis.

Limitations
The study also has some notable limitations. Among the RCTs included, there was great heterogeneity with respect to exercise intervention modes, medicine intake, lifestyle, and low-quality data that may have contributed to unwanted heterogeneity. Moreover, studies with aerobic exercise combined with resistance exercise accounted for less than 18% of included articles, and there were six RCTs with combined exercise data; therefore, medical evidence on the intervention effect of high-intensity intermittent training and combined exercise on atherosclerosis needs to be further explored.

Practical Implications
The results of this study indicate that the intervention effect of aerobic exercise on carotid atherosclerosis is relatively stable, which can be used to guide patients to improve their condition, reduce the risk of cardiovascular disease, and thus improve their quality of life. In order to prevent and improve carotid atherosclerosis more effectively, the prescription of aerobic exercise for atherosclerosis was induced by tracing the original research literature. The minimum standard of aerobic exercise for atherosclerosis was an 8-week intervention period, 60 min of cumulative exercise time per week, and 50-70% HR peak exercise intensity. Secondly, high-intensity intermittent training can also be adopted for atherosclerosis with young patients. However, the intervention effect of aerobic exercise is more stable, and aerobic exercise should be the main intervention.

Conclusions
Exercise can significantly reduce cIMT, TC, and LDL-C and increase HDL-C, which has a good prevention and improvement effect on carotid atherosclerosis. From the perspective of exercise intervention patterns, aerobic exercise and high-intensity intermittent exercise can improve carotid atherosclerosis; however, aerobic exercise has a more comprehensive improvement effect.