Lifestyle modifications are the crux of atherosclerotic disease management. The goal of this study was to determine the effectiveness of diet and exercise in decreasing coronary and carotid atherosclerotic burden. Randomized controlled trials examining the effects of intensive lifestyle measures on atherosclerotic progression in coronary and carotid arteries as measured by baseline and follow-up quantitative coronary angiogram and ultrasonographic carotid intimal-medial thickness (CIMT), respectively, were included. Studies were excluded if the intervention additionally included a medication. MEDLINE, EMBASE, CINAHL, Cochrane Controlled Trials Registers, reports, and abstracts from major cardiology meetings were searched by 2 researchers independently and verified by the primary investigator. Standardized mean difference (SMD) with 95% confidence intervals (CIs) was calculated using random-effects model. Publication bias and heterogeneity were assessed. Fourteen trials were included. Seven used quantitative coronary angiogram, and 7 used CIMT; 1,343 lesions in 340 patients in the coronary group and 919 patients in the carotid group were analyzed. Overall, lifestyle modifications were associated with a decrease in coronary atherosclerotic burden in percent stenosis by −0.34 (95% CI −0.48 to −0.21) SMD, with no significant publication bias and heterogeneity (p = 0.21, I 2 = 28.25). Similarly, in the carotids, there was a decrease in the CIMT, in millimeter, by −0.21 (95% CI −0.36 to −0.05) SMD and by −0.13 (95% CI −0.25 to −0.02) SMD, before and after accounting for publication bias and heterogeneity (p = 0.13, I 2 = 39.91; p = 0.54, I 2 = 0), respectively. In conclusion, these results suggest that intensive lifestyle modifications are associated with a decrease in coronary and carotid atherosclerotic burden.
Diet and exercise are the cornerstones in atherosclerotic disease management. Guidelines for secondary prevention of atherosclerotic disease from committees like the American Heart Association/American College of Cardiology Foundation and the European Society of Cardiology include dietary modifications and regular exercise. However, patient adherence to this is spotty at best. Several studies have shown that dietary modifications and regular exercise reduce cardiovascular events, mortality, and morbidity. Demonstration that these lifestyle changes, in addition to reducing clinical events, reduce the likelihood of the progression of atherosclerotic disease should strengthen the argument for adherence. However, to date, randomized controlled trials (RCTs) designed to establish the beneficial effects of these measures on atherosclerotic disease progression have not been convincing. Small sample sizes in these trials appear to have led conflicting results. As an attempt to overcome these shortcomings, this study was undertaken to critically examine the effectiveness of intensive lifestyle modifications on atherosclerotic disease progression by a systematic review and quantify them using meta-analytical tools. We decided to focus our study on the coronary and carotid arteries as they are known to hold important prognostic value for cardiac cerebrovascular events and overall morbidity and mortality.
Methods
Only RCTs comparing the efficacy of lifestyle measures, that is, diet and/or exercise versus usual care, were identified by a systematic search. Two independent researchers searched MEDLINE, EMBASE, CINAHL, and Cochrane Controlled Trials Register for eligible studies. Studies from 1946 (MEDLINE) or 1947 (EMBASE) or earliest time frame available (CINAHL, Cochrane Controlled Trials Register) to third week of March 2014 were included. Search terms included: “diet,” “exercise,” “lifestyle,” “risk factors,” “health behavior,” “food habits,” “exercise therapy,” “lifestyle change,” “behavior modification,” “risk factor reduction,” “risk factor modification,” “arteriosclerosis,” “carotid artery diseases,” “carotid artery thrombosis,” “coronary disease,” “angiography,” “carotid intima thickness,” “plaque thickness,” “gensini,” “stenosis,” “coronary artery,” “carotid artery,” and “atherosclerotic burden.” Abstracts and unpublished studies from major cardiology meetings were reviewed. Lastly, references of relevant peer-reviewed publications were evaluated. PRISMA guidelines were used. Institutional review board approval was not required because of the nature of the study.
We included only studies that (1) were prospective controlled trials, in which patients were randomized to diet and/or exercise alterations or usual care; (2) in the methods, clearly reported subject selection process, interventions, and follow-up; (3) included baseline and follow-up data on quantitative coronary angiograms (QCAs) or carotid intimal-medial thickness (CIMT) or both. Usual care in the trials entailed patients being left at their own discretion or their physician’s advice and not being enrolled into an active diet and/or exercise program that the trial offered. Studies were excluded if they were not RCTs or if subjects received a medication along with diet and/or exercise as their intervention. Investigators were contacted to obtain additional data or verify data. Studies meeting all inclusion and no exclusion criteria were reviewed. Two individual investigators (SJ and DP) independently used the Jadad’s system to assess for the quality of the studies. Discrepancies were resolved by consensus.
Relevant data from the RCTs including year of study, characteristics of subjects, details of intervention, and imaging procedures were collected by 1 investigator and verified by another. Total number of patients and coronary lesions in intervention and control groups were reported. Data regarding QCA and CIMT were collected at baseline and follow-up. When available, clinical outcomes were recorded to complement the review; although if unavailable, they were not actively sought as this was not the theme of this study.
Studies were grouped based on the imaging tool used, that is, QCA and CIMT. Outcome of interest was change in QCA and CIMT. Follow-up imaging was conducted after the intervention ended. Compliance to the intervention was recorded as reported in the individual studies. In our meta-analysis, we report overall changes by QCA and CIMT in percent stenosis and millimeter, respectively, with reference to the standardized mean difference (SMD).
Comprehensive Meta-analysis software (version 2, 2005; Biostat, Englewood, New Jersey) was used for analysis. To assess publication bias, funnel plots were drawn and assessed for asymmetry. Statistical heterogeneity was also calculated, and p <0.1 was considered as statistically significant heterogeneity. Fixed- and random-effect models were used to determine overall effect of lifestyle modification.
Results
Our initial search provided us with 178 citations (EMBASE: 16, MEDLINE: 138; CINAHL: 24, Cochrane: 0). After removing duplicate and irrelevant citations, there were 14 studies that met the inclusion and exclusion criteria ( Figure 1 ): 7 used QCA, whereas 7 used CIMT. Table 1 outlines their characteristics. Of the QCA studies, 5 used diet and exercise as the intervention and 2 diet only. Meanwhile, in the CIMT studies, 3 examined both diet and exercise as the intervention, 3 included only exercise, and 1 diet only. In studies using diet and exercise as their intervention, an equal emphasis was laid on diet and exercise. The interval between baseline and follow-up imaging ranged from 6 months to 6.5 years, with a mean of 2.5 and 3 years for QCA and CIMT studies, respectively. Duration of intervention was similar to the interval between baseline and follow-up imaging in all studies expect 1, in which the participants had lifestyle modifications for 2.7 years before the baseline CIMT study. Thus, duration of intervention ranged from 3 months to 9 years, with a mean of 2.5 and 3.4 years for QCA and CIMT studies, respectively. A total of 1,343 lesions in 340 patients were followed by QCA and 919 patients, with CIMT. Table 2 compares baseline and follow-up QCA and CIMT measurements in both groups.
| Study Type | Study | Follow up | Mean Age | Male | Dietary Interventions | Exercise Interventions | Control | Adherence | Jadad score |
|---|---|---|---|---|---|---|---|---|---|
| QCA | Ornish (1990) ∗ | 1 year | 58 years | 88% | LF VD (10% fat (PU/S >1), 15-20% Pr, 70-75% CCa, Ch ≤5mg/day), no Cf, Al ≤2 units/day | Moderate Ar (walking, ≥3 hrs/wk, ≥30 min/Ses) | LATOD | 122% ‡ | 2 † /5 |
| Watts | 3 years | 51 years | 100% | 27% F, 8-10% SF, 8% ù-6 & ù-3 PU, Ch 100mg/1000kcal, plant derived soluble fiber equivalent of polygalacturonate 3.6 gm/1000 kcal, 1000-1200 kcal/day if BMI >25 | N/A | Ed & LATOD | Most people | 3 † /5 | |
| Schuler | 1 year | 53.5 years | 100% | LF, LCh AHA 3 (<20% F (PU/S >1), 15% P, 65% Ca, Ch <200 mg) | Cycle ergometer (daily, ≥30 min/Ses) & group Ex (2 sessions/wk, 60 min/Ses) | Ed & LATOD | 68% | 1 † /5 | |
| Huh | 1 year | 59.5 years | 86% | LF, LCh (<10% F (PU/S >1), Ch <50 mg/day, calorie restriction: 1255 kJ/day), no smoking, no Al | N/A | ANTMC | Excellent | 3/5 | |
| Niebauer | 6 years | 53.5 years | 100% | LF, LCh AHA 3 (<20% F (PU/S >1), 15% P, 65% Ca, Ch <200 mg) | Cycle ergometer (daily, ≥30 min/Ses) & group Ex (2 Ses/wk, 60 min/Ses) | Ed & LATOD | 37% | 2 † /5 | |
| Ornish (1998) ∗ | 5 years | 59.6 years | 94% | LF VD (10% F (PU/S >1), 15-20% P, 70-75% CCa, Ch ≤5mg/day), no Cf, Al ≤2 units/day | Moderate Ar (walking, ≥3hr/week, ≥30 min/Ses) | LATOD | 106% ‡ | 2 † /5 | |
| Manchanda ∗ | 1 year | 51.5 years | 100% | LF, LCh AHA 1 (15% F (mainly PU), 65% Ca (mostly complex), Ch <50 mg/day, fiber ≥1gm/day), 15 gm psyllium husk | Moderate Ar (UK) & yogic Ex (90 min/day) | Ed & LATOD | 79.5 ± 14.5% | 1 † /5 | |
| CIMT | Rauramaa | 6 years | 57.2 years | 100% | N/A | Ar (jogging, skiing, swimming & cycling, 5 Ses/wk, 45-60 min/Ses) | Ed & LATOD | UK | 3 † /5 |
| Wildman | 6.5 years | 49.2 years | 0% | 25% F (7% SF), 100 mg Ch, 1300 kcal/day | Increase leisure-time physical activity to 1000-1500 kcal/wk | LATOD | UK | 1/5 | |
| Anderssen | 4 years | 57 years | 100% | ↑ fruits, veg, CCa, fish & fish products; ↓ SF & sugar; caloric restriction; based on Oslo diet | Endurance Ex (Ar, fast walking, 2 Ses/wk) | LATOD | Poor | 3 † /5 | |
| Hjerkinn | 3 years | 70 years | 100% | Mediterranean European (≤30% F (≤30% SF), 12-15% P, ≥55% Ca, Ch <300 mg/day; ↑ veg oil, veg, fruits & fish; ↓ meat, calorie restriction if overweight), Al ≤2-3% | N/A | LATOD | UK | 3 † /5 | |
| Kim | 6 Months | 54.4 years | 24% | UK | Moderate Ex (brisk walking, ≥150 min/wk) | Ed & LATOD | 75% | 1 † /5 | |
| Meyer | 6 Months | 13.9 years | 53% | N/A | Mon: Swimming & Ar (60 min), Wed: Sports (90 min), Fri: Walking (60 min) | ANTMC | 66% | 3/5 | |
| Farpour-Lambert | 6 Months | 9 years | 36% | N/A | Ar (30 min), strengthening (20 min), stretching & cooling (10 min)/Ses, 3 Ses/wk | ANTMC | 82% | 2 † /5 |
∗ Study used stress management in the intervention group additionally.
† Patients did more than expected and thus got >100% adherence scores.
‡ Studies in which image acquisition and/or analysis of reading was performed in the blinded fashion.
| Study Type: | First Author (Study) (Year) | Patients (n) | Lesions (n) | Baseline stenosis/thickness | Follow up stenosis/thickness | Change in stenosis/thickness | ||||
|---|---|---|---|---|---|---|---|---|---|---|
| Quantitative Coronary Angiogram | Intervention | Control | Intervention | Control | Intervention | Control | ||||
| Ornish et al (1990) | I: 22, C: 19 | I: 105, C: 95 | 40 ± 16.90 ∗ | 42.7 ± 15.50 ∗ | 37.8 ± 16.50 ∗ | 46.1 ± 18.50 ∗ | NA | NA | ||
| Watts et al (1992) | I: 26, C: 24 | I: 169, C: 157 | 23.3 ± 31.2 | 18.6 ± 23.81 | NA | NA | ↓ by 0.5 ± 52 ∗ | ↑ by 5.6 ± 45.11 ∗ | ||
| Schuler et al (1992) | I: 40, C: 52 | I: 122, C: 158 | 65 ± 24 | 63 ± 29 | 64 ± 23 | 66 ± 28 | ↓ by 1 ± 13 ∗ | ↑ by 3 ± 13 ∗ | ||
| Huh et al (1996) | I: 7, C: 7 | I: 14, C:13 | 63.2 ± 14.59 ∗ | 65.8 ± 16.22 ∗ | 56.8 ± 13.47 ∗ | 66.6 ± 27.76 ∗ | NA | NA | ||
| Niebauer et al (1997) | I: 32, C: 34 | I: 98, C: 106 | 58.9 ± 27.7 | 54.7 ± 34.7 | 62.0 ± 25.9 | 66.6 ± 30.2 | ↑ by 3.1 ± 24.6 ∗ | ↑ by 11.9 ± 28.2 ∗ | ||
| Ornish et al (1998) | I: 20, C: 15 | I: 109, C: 77 | 38.92 ± 19.31 | 42.5 ± 19.31 | NA | NA | ↓ by 3.07 ± 15.1 ∗ | ↑ by 11.77 ± 37.47 ∗ | ||
| Manchanda et al (2000) | I: 21, C: 21 | I: 61, C: 59 | 62.4 ± 14.5 ∗ | 59.7 ± 17.7 ∗ | 60.9 ± 16 ∗ | 68.4 ± 16 ∗ | ↓ by 1.5 ± 9.7 | ↑ by 8.7 ± 12.9 | ||
| Carotid Intimal Medial Thickness | Rauramaa et al (2004) | I: 64, C: 61 | NA | NA | NA | NA | ↑ by 0.12 ± 0.551 ∗ | ↑ by 0.20 ± 0.598 ∗ | ||
| Wildman et al (2004) | I: NA, C: NA, TOTAL: 113 | NA | NA | NA | NA | ↓ by 0.0003/year (p= 0.867) in I Vs C ∗ | ||||
| Anderssen et al (2005) | I: 142, C: 143 | 0.798 ± 0.247 ∗ | 0.804 ± 0.186 ∗ | 0.849 ± 0.196 ∗ | 0.878 ± 0.185 ∗ | ↑ by 0.049 ± 0.165 | ↑ by 0.076 ± 0.101 | |||
| Hjerkinn et al (2006) | I:117, C: 114 | 0.929 ± 0.21 | 0.909 ± 0.19 | 0.967 ± 0.21 | 0.977 ± 0.21 | ↑ by 0.038 ± 0.09 ∗ | ↑ by 0.068 ± 0.10 ∗ | |||
| Kim et al (2006) | I: 32, C: 26 | 0.714 ± 0.138 | 0.678 ± 0.205 | 0.664 ± 0.124 | 0.761 ± 0.147 | ↓ by 0.040 ± 0.136 ∗ | ↑ by 0.083 ± 0.167 ∗ | |||
| Meyer et al (2006) | I: 33, C: 34 | 0.48 ± 0.08 ∗ | 0.47 ± 0.06 ∗ | 0.44 ± 0.08 ∗ | 0.45 ± 0.06 ∗ | NA | NA | |||
| Farpour-Lambert et al (2009) | I: 18, C: 18 | 0.487 ± 0.028 | 0.482 ± 0.031 | NA | NA | ↓ by 0.02 (p = 0.045) in I Vs C ∗ | ||||
Jadad’s scoring was used to evaluate for the quality of the studies ( Table 1 ). Because of the nature of the imaging techniques used, blinding of study participants and physicians was not always possible. For this reason, all studies lost 2 points. Thus, overall Jadad scores were low. It is, however, noteworthy that image acquisition and/or analysis of reading was performed in a blinded fashion in 11 reports. In addition to this, details on study criteria, changes in weight, body mass index, lipids, blood pressure, medications, and smoking patterns were sought and reported in Supplementary Table 1 . All studies apart from 1 QCA and CIMT study were associated with decreases in weight with lifestyle. In addition, all QCA studies were associated with decreases in (1) total cholesterol, apart from 1 ; (2) low-density lipoprotein cholesterol, apart from 2 ; and (3) triglyceride, apart from 3 studies. As for CIMT studies, data were available on limited number of studies. None of the 2 reported CIMT studies showed changes in total cholesterol. Two of the 4 studies were associated with decreases in low-density lipoprotein. These 2 studies in addition to 1 more were associated with triglyceride decreases. No QCA or CIMT study demonstrated changes in high-density lipoprotein. In addition, no QCA study was associated with blood pressure changes. However, 3 of the 6 CIMT studies showed decreases in blood pressure. There were no significant changes in smoking patterns in QCA and CIMT studies reporting them; and none of their participants used lipid-lowering medications, except for 1 QCA study, in which at baseline, none were on statins, but at follow-up, more patients in the control arm were found to be on statins compared with the lifestyle group; however, this was not significant.
Heterogeneity and biases were accessed in depth. Funnel plots were shown in Figure 2 . In the QCA arm, all RCTs included patients with coronary artery disease. The exercise regimen was aerobic in nature lasting from 3 to 4.5 h/wk with some interstudy variability ( Table 1 ). The dietary regimen was a variant of a low calorie, carbohydrate, fat, cholesterol diet with specific variations laid down in Table 1 . There was a big variation (37% to >100%) in the adherence to intervention. Studies by Ornish et al and Niebauer et al were long-term follow-ups of the studies by Ornish et al and Schuler et al, respectively. When all QCA studies were looked at, there was no significant heterogeneity (p = 0.21, I 2 = 28.25; Figure 3 ). There was no significant publication bias when QCA studies were looked ( Figure 2 ). As for the CIMT studies, the study population varied from the healthy to those with risk factors like obesity, hypertension, hyperlipidemia, and diabetes. The exercise regimen included aerobic exercise lasting 2.5 to 3.75 h/wk or increases in leisure activity to 1,000 to 1,500 kcal/wk. The diet was a form of low fat, cholesterol, calorie diet with variations pointed out ( Table 1 ). Thus, these interventions were not very different from the QCA studies. Adherence to intervention was mentioned in 4 of 7 studies and ranged from poor (unknown) to 83%. When all the 7 CIMT studies were looked at, there was no significant heterogeneity (p = 0.13, I 2 = 39.91). However, when funnel plots were examined for publication bias, it was found that the study by Kim et al had highly favorable effects, thus contributing to the asymmetry. When CIMT studies were analyzed without it, the 6 studies displayed no heterogeneity at all (p = 0.54, I 2 = 0), and the bias was not apparent. The analyses for CIMT studies were performed with and without the study by Kim et al.
