Servicios personalizados
Inglés (pdf)
Articulo en XML
Referencias del artículo
Traducción automática
Enviar articulo por email
Citado por SciELO 
Citado por Google 
Similares en
SciELO 
Similares en Google 
Permalink
INTRODUCTION
Statins and other pharmacological interventions are now widely recommended as primary prevention against cardiovascular disease (CVD) 1 . Yet while statins were initially recommended for high-risk individuals, this threshold has steadily been reduced. For example, guidelines from the National Institute for Care Excellence imply that ten million healthy individuals in the UK, with no previous history of CV events, are now candidates for taking statins for primary prevention 2. This has led to concerns about potential long term risks and over-medicalization 3.
Consequently, attention is shifting to non-pharmacological interventions for primary prevention, which many perceive as more appropriate for healthy individuals, with less risk of harm, and possibly with lower overall cost than statins to the public health system. Undoubtedly, tobacco cessation, reducing dietary fats and increasing physical activities have played a part in reducing rates of CVD 4 5. According to the WHO, appropriate policy and environmental changes could bring about further major reductions. Nevertheless, there is less evidence about which non-pharmacological interventions led by primary care professionals have been shown effective for CVD risk reduction in the general population as an alternative (or complement) to widespread use of statins.
Many candidate non-pharmacological interventions have been studied, including dietary advice, food supplements with vitamins or minerals, physical exercise, mental exercises and general health checks. The objective of this study is to determine the effectiveness of non-pharmacological interventions for prevention of cardiovascular disease and mortality in people with no history of CVD events. We apply an umbrella review methodology. This is a “review-of-reviews”, providing a balanced summary of evidence from several systematic reviews of randomised controlled trials. This enables us to take a broad view of the state of the evidence at a given point in time across a wide range of different treatments. The study is carried out according to methodological recommendations from the Joanna Briggs Institute 6 and Smith et al 7 .
MATERIALS AND METHODS
The research question was to determine the relative effectiveness of non-pharmacological interventions to prevent major CVD events and mortality in adult classified as having no previous history of CVD events or at risk of developing CVD, compared to placebo or no intervention.
The design of the study, an umbrella review of the literature, was carried out. The protocol was registered in PROSPERO (CRD 42015016447 http://www.crd.york.ac.uk).
The search was conducted using PubMed Health, Effective Health Care Program Agency for Healthcare Research and Quality (AHRQ), Health Evidence and the Cochrane Plus databases; and using as free terms: “primary prevention”, “cardiovascular disease”, “review”. In addition, the reference lists of the included articles were examined. Only papers in English or Spanish and published by 30th July 2017 were included.
Inclusion criteria:
systematic reviews of randomized clinical trials (with or without meta-analysis)
adult classified as having no history of CVD events, or at risk of developing CVD events or those of general population
intervention: Non-pharmacological for targeted interventions to improve cardiovascular health in primary prevention. Any duration or time of follow- up periods were considered
comparing group: placebo or no intervention
outcomes intended to be measured: overall mortality, cardiovascular mortality, cardiovascular events, and adverse events related with the interventions
Exclusion criteria:
systematic reviews conducted only in populations with specific conditions (only smokers, only obese people)
reviews which mix primary and secondary prevention populations together
review which mix results of clinical trials with other designs
systematic reviews of quasi-experimental designs
Identification of the reviews was carried out by two reviewers (EMR and AOL) who selected independently all the works and checked both title and abstract against the inclusion and exclusion criteria. Any discrepancies were agreed by consensus or third researcher (DE). Any duplicate articles were subsequently excluded. Articles were managed using Reference Manager Software.
Following the definitive selection of the articles of interest, two researchers carried out an independent evaluation of the methodological quality of each review (EM and AOL). To achieve this, the Assessing the Methodological Quality of Systematic Reviews (AMSTAR) checklist with eleven items was used8. Any discrepancies were discussed and agreed by consensus.
A data extraction form was developed to record bibliographic information. The general characteristics of the study (participants included, description of intervention) and outcome data (including details of outcome assessment and adverse effects) were extracted from each included study. In those articles that reported data for both primary and secondary prevention, we extracted only the primary prevention results.
The authors carried out a qualitative synthesis of results from included reviews, the interventions were classified according to focus of intervention (proposed by the authors), and type of outcomes (all-cause mortality, cardiovascular mortality, both fatal or non-fatal CVD events and adverse events). Heterogeneity between studies was evaluated using Higgin’s I2, where I2 >50% imply a substantial level of heterogeneity 9 . A random-effects model was used to combine study-specific relative risk (RR) and calculate a 95% confidence interval (CI). The analyses were performed with Stata software version 11.
RESULTS
The initial search yielded a total of 643 publications. After excluding by title and/or abstract, a total of 127 full-text works were assessed for eligibility (see Figure 1). One review 10 was excluded in order to avoid duplicity, as it had the same aim and higher risk of bias than another included review 11 . Finally, twenty-four reviews were included in the final synthesis Fig. 1), of which thirteen reported CVD events and/or overall mortality as primary outcome.
Primary prevention
Eight types of non-pharmacological interventions for primary prevention of CVD were reviewed Table 1):
Tabla 1 Characteristics of the reviews included and adverse events
| Author Year | Intervention Comparison | Population | Outcomes | Studies Partici-pants | Follow-up Range | Characteristics of Participants | AE |
|---|---|---|---|---|---|---|---|
| Al-Khudairy (2017) 12 | supplements of vit.C alone, at any dose or frequency vs placebo / no intervention | -healthy adults from GP -adults at moderate / high risk of CVD | -all-cause mortality -CVD mortality-non-fatal end points | 8 15,445 | ≥3 months 3m-8y | 14,641 male physicians aged ≥50 with CVD risk factors (n=1), men and women (n=6), no specified gender (n=1) | reported but no specified (n=2) |
| Bjelakovic (2012) 13 | antioxidant supplements (carotene, vit.A, vit.C, vit.E, and selenium)b at any dose, duration and route of administration vs. placebo / no intervention | -healthy adults -adults from GP | -all-cause mortality | 26 215,900 | NA | NA | NS |
| Bjelakovic(2014) 14 | supplements of vit.D, alone or combined with calcium, at any dose or durationb vs placebo / no interventionc | -healthy adults from GP irrespective of serum vit.D | -all-cause mortality -CVD mortality -AE | 48 94,491 | NA | healthy volunteers (n=4), postmenopausal women (n=9), older people (n=35) | NA |
| Rees (2013) 15 | supplements of selenium alone vs. placebo / no intervention | -adults from GP -adults at high risk of CVD | -all-cause mortality -major CVD events (CVD, MI, non-fatal stroke, revascularisation procedures as CABG or PTCA) -CHD events (fatal CHD, non-fatal MI) -stroke events (fatal and non-fatal stroke, PAD) | 12 19,715 | ≥3 months 1-2w-12y | 17,843 men (n=6), healthy population (n=3), prostate cancer (n=1), at high risk of cancer (n=2) | increased riskd of DM2 (1.06, 0.97-1.15; n=3) or alopecia (1.28, 1.01- .62) and 1-2 grade dermatitis (1.17, 1.0-1.35; n=1) |
| Flowers (2014) 16 | supplements of co-enzyme Q10 alone vs. placebo / no intervention | -adults from GP -adults at high risk of CVD | -all-cause mortality -CVD mortality -non-fatal endpoints** -AE | 6 218 | ≥12 weeks 12-16w | men and women (n=5), no specified gender (n=1), at high risk of CVD (n=6; obese n=1, hypercholesterolemic n=4, hypertensive n=1) | NR |
| Hartley (2015) 17 | supplements of vit.K alone vs. placebo / no intervention | -healthy adults aged ≥18 from GP -adults at moderate / high risk of CVD | -all-cause mortality -CVD mortality -non-fatal endpointsa -AE | 1 60 | ≥3 months | healthy men and women aged 40-6 with BMI=18.5-30 | NR |
| Fortmann (2013) 18 | supplements of vitamins (A, C, D, E) and minerals (calcium, folic acid, carotene, selenium)b,e alone or combined vs placebo / no multivitamin | -adults from GP with no history of CVD | -all-cause mortality -CVD events | 24 NS | ≥6 months 6m-16y (most studies ≥ 10y) | NA | trivial AE related to vit. A (n=1) |
| Hooper (2004) 19 | advice to reduce sodium intake vs placebo / no intervention | -healthy adults aged ≥16 with normal / raised blood pressure | -total mortality -combined CVD events (fatal and non-fatal MI, stroke, angina, heart failure, peripheral vascular events, sudden death, non-scheduled CVD interventions as CABG, PTCA) | 11 3,514 | ≥26 months 6m-7y | 2,326 healthy normotensive adults mostly white, male, mean age 40 (n=3), 387 untreated hypertensive adults aged 16-64 (n=5), 801 treated hypertensive aged 55-67 (n=3) | NS |
| Hartley (2013) 20 | advice to increase consumption of fruit and vegetables (or provision of them) vs no /minimal interventiona | -healthy adults aged ≥18 -adults at high risk of CVDf | -all-cause mortality -CVD mortality -non-fatal endpointsa -AE | 10 1,730 | ≥3 months 3-12m | only women (n=1), healthy post-menopausal women with family breast cancer history (n=1), women with metabolic syndrome (n=2), healthy adults (n=3), adults with LDL=130- 190 mg/dL and triglyceride ≤250 mg/dL (n=1), obese (n=1), with colorectal adenomatous polyps in the last 5y (n=1), pre-metabolic or healthy (n=1) | increased bowel movements, bad breath and body odor (n=3) |
| Rees (2013) 21 | Mediterranean dietary patterng vs no / minimal interventionh | -adults from GP -adults at high risk of CVDf | -all-cause mortality -CVD mortality -non-fatal endpointsa -AE | 11 52,044 | ≥3 months 3-8y | 48,835 women (n=1), 49,185 healthy adults (n=5) | NR |
| Martin (2015) 22 | advice to increase consumption of nuts (or provision of them) vs no / minimal intervention | -adults from GP with no history of CVDe | -all-cause mortality -CVD mortality -non-fatal endpointsa -AE | 5 435 | ≥3 months 3-6m | women, mean age=37.4-54.3 (n=1) | allergic reaction to nuts (n=1), insignificant weight gain (n=1) |
| Hartley (2013) 23 | intake of green / black tea extracts or as a beverage vs no / minimal interventionh / placebo | -healthy adults -adults at high risk of CVDf | -all-cause mortality -CVD mortality -non-fatal endpointsa -AE | 11 821 | ≥6 months 3m-2y | women (n=4), postmenopausal women with osteopenia (n=1), breast cancer survivors (n=1); either DM or two other CVD risk factors (n=1), up to moderate hypercholesterolemia (n=2), healthy participants (n=6) | prostate cancer, appendicitis, hospitalizationfor influenza, and retinal detachment (n=5, unlikely attributable to intervention) |
| Hartley (2016) 24 | advice to increase dietary fibre consumption (or provision of fibre supplements / high-fibre foods) vs. no/minimal intervention | -adults from GP -adults at high risk of CVD | -all-cause mortality -CVD mortality -non-fatal endpointsa -AE | 23 1,513 | ≥3 months NS | healthy participants (n=4), over weighted or obese (n=6), with clear signs of metabolic syndrome (n=3), hypertensive (n=3), women with climacteric symptoms (n=1) | n=14; gastrointestinal events (flatulence and diarrhoea, n=1; flatulence but also constipation, nausea, bloating and diarrhoea, n=7); decrease of antihypertensive medication sideeffects scores from baseline observed only in the intervention group (n=1) |
| Hooper (2004) 25 | advice on diet or provide a diet to increase 3FA by foodstuffs or supplements (oil / capsule)e vs. placebo / no advice / no supplementation / usual diet | -adults at any risk of CVD | -all-cause mortality -fatal and non-fatal CVD events (MI, angina, stroke, heart failure, peripheral vascular disease, sudden death, CABG, PTCA) -cancer -AE | 48 35,140 | ≥6 months 11-48 m | at high (n=21), moderate (n=10) or low risk (n=17) of CVD; male proportion ≥70% (n=24), 31-69% (n=17), 0-30% (n=5), not stated (n=2); mean ages in the 30s (n=6), 40s (n=7), 50s (n=27), 60s (n=4), 80s (n=3), unclear (n=1) | not significant thrombophlebitis (1.59, 0.72-3.51, n=1), urolithiasis (0.80, 0.47-1.36, n=1) and DM (0.87, 0.15- 5.08, n=2)d |
| Al-Khudairy (2015) 26 | advice on diet, supplementation or provide a diet to increase / decrease 6FA vs. placebo / no advice / no supplementation / usual diet | -healthy adults from GP -adults at moderate / high risk of CVD | -all-cause mortality -CVD mortality -non-fatal endpointsa -AE | 4 664 | ≥6 months | over weighted or obese healthy participants aged 35-65 (n=1) and 40-70 (n=1), older men and postmenopausal women aged 45- 70 (n=1), adults aged 23-58 with hypercholesterolemia (n=1) | NR |
| Ebrahim (2011) 11 | health promotion activity (counseling or educational intervention, individually or by group, with or without pharmacological treatment) to achieve a behavioral change altering ≥1 CVD risk factor Parallel-group design | -adults (aged ≥35) from GP -adults from occupational groups -adults with specific risk factorsf | -all-cause mortality -fatal CHD and stroke events -non-fatal CHD (MI, unstable angina, CABG, PTCA, stroke events) | 55 139,256 | ≥6 months. 6m-12y (median=1 y) | mean age=50y; adults ≥60 (n=2), ≥75 (n=53); patients with DM (n=5) and with hypertension (n=16) | NS |
| Uthman (2015) 27 | health promotion intervention (dietary advice, increasing activity, smoking cessation) with or without pharmacological treatment to achieve a behavioral change altering ≥1 CVD risk factor vs. no intervention | -adults from the workforce, including those at high-risk (or not) of CVDf | -CVD mortality -non-fatal endpointsa -AE | 13 NS | ≥6 months 6-30m (mean= 13.3m) | healthy or GP (n=2), high CVD risk groups (preand hypertensive, with metabolic syndrome, obese, with impaired glucose regulation) (n=11) | NR |
| Seron (2014) 28 | aerobic / resistance exercise training vs. no exercise | -people at high CVD risk with a Framingham risk score ≥10% over 10 y | -CVD mortality -all-cause mortality -MI -stroke | 4 823 | ≥8 weeks 16w-6m | mean age=52.6y (n=4) | NR |
| Hartley (2015) 29 | any style of Qigong vs. no / minimal intervention (health or physical activity promotion without faceto- face interaction) | -healthy adults aged ≥18 years from GP -adults at high risk of CVD | -all-cause mortality -CVD mortality -non-fatal endpointsa -AE | 11 1,369 | ≥3 months 3m-30y | men and women (n=5), men (n=4), not specified gender (n=2); healthy participants (n=2), elder with high blood lipids (n=1), hypertensive patients (n=8) | NR |
| Hartley (2014) 30 | any style of Tai chi vs. no / minimal interventionh | -healthy adults -adults at high risk of CVD | -all-cause mortality -CVD mortality -non-fatal endpointsa -AE | 13 1,520 | ≥ 3 months 3-6m | men and women (n=9), only women (n=4); borderline (n=1) and essential (n=2) hypertensive, hypertensive with liver and kidney deficiency syndrome (n=1), hypertensive (n=1), older participants with high risk of falling (n=1), healthy people (903 participants, n=7) | NR |
| Hartley (2014) 31 | any type of yoga with postural exercises, breathing control and meditation vs. no / minimal interventiona | -healthy adults -adults at high risk of CVD | -all-cause mortality -CVD mortality -non-fatal endpointsa -AE | 11 800 | ≥12 weeks 14w-8m | only man (n=1), only women (n=10), including healthy premenopausal (n=1), with metabolic syndrome (n=1), obese postmenopausal (n=1), with menstrual irregularities (n=1); essential hypertensive (n=1), males with high cholesterol and hypertension (n=1), HIVinfected participants (n=1), no yoga exercisers with family history of DM2 and at least one cardio- metabolic risk factor (n=1), healthy participants (n=3) | NR |
| Hartley (2014) 32 | Transcendental meditation vs. no / minimal interventionh | -healthy adults -adults at high risk of CVD | -all-cause mortality -CVD mortality -non-fatal endpointsa -AE | 4 430 | ≥3 months NS | men and women (n=3), only women (n=1); healthy participants (316 adults, n=1), essential hypertensive (n=1), older people (n=1) | NR |
| Widmer (2015) 33 | digital health interventionsb,e (telemedicine, webbased strategies, email, mobile phone, apps, text messaging, monitoring sensors) vs. no active intervention beyond usual care following standard guidelines | -adults at any risk of CVD | -CVD events -hospitalizations -all-cause mortality | 2 1,055 | ≥1 month NS | hypertensive (n=1), over weighted adults (n=1) | NS |
| Krogsboll (2012) 34 | general health checks screening for ≥1 disease or risk factor and in more than one system by doctor, nurse or other health professional vs. control group | -GP or subgroups (employees from a company aged <65, or people with specific risk factors / diseases) | -all-cause mortality -disease-specific mortality -morbidity | 14 182,880 | 4-22 y | no details of interest | NS |
AE: adverse events; vit: vitamin; GP: General Population; CVD: cardiovascular disease; m: months; y: years; NA: not applicable as primary and secondary prevention data were mixed; NS: not studied; CHD: coronay heart disease; w: week; DM2: type 2 diabetes mellitus; NR: not reported; BMI: body mass index; ωFA: omega fatty acids.
a myocardial infarction (MI), coronary artery bypass grafting (CABG), per cutaneous transluminal coronary angioplasty (PTCA), angina, angiographically defined coronary heart disease, stroke, carotid endarterectomy, and peripheral arterial disease (PAD); b only primary prevention information/ results provided; c calcium in the control group was included if equally used in the vitamin D groups of the trial; d relative risk and confidence interval at 5%; e only information about random clinical trials; f studies excluded if ≥25% participants had CVD; g consumption of at least two of the following dietary factors: high monounsaturated/saturated fat ratio (use of olive oil as main cooking ingredient), low to moderate red wine consumption, high consumption of legumes, high consumption of grains and cereals, high consumption of fruits and vegetables, low consumption of meat and meat products and increased consumption of fish, moderate consumption of milk and dairy products; h as leaflets with no person-to-person interaction or reinforcement; i studies excluded if ≥25% participants had diagnosed CVD or DM2.
- supplements: vitamin C 12, antioxidant 13, vitamin D 14, selenium 15, vitamin Co-enzyme Q10 16, vitamin K 17, and vitamin, mineral and multivitamin supplements 18 . Fortmann’s review 22 included different interventions based on single and paired combination of supplementation with vitamins and minerals supplementation, including vitamin C and selenium. The two reviews on vitamin C 12 18 shared elements such as type of population included and intervention, but they differed in both the outcome measures included and some inclusion criteria (minimum follow-up time for required randomized controlled trials (RCT) (three months) 12, trials with good or fair quality, studies carried out in countries whose Human Development Index was considered very high18)
dietary interventions: advice to reduce dietary salt 19, advice to increased consumption of fruit and vegetables 20 and Mediterranean dietary pattern 21
diet supplementation: nuts, green and black tea 23, dietary fibre 24, Omega 3 25 and Omega 6 26 fatty acids
multiple risk factor interventions: counseling or education to modify more than one cardiovascular risk factor 11 27, the last of them in low income countries
physical exercise: exercise training 28, Qigong 29, Tai chi 30, Yoga 31
mental exercise: transcendental meditation 32
digital health intervention 33
general health checks 34
Twenty-two reviews were published by the Cochrane collaboration and two by other groups 18 33, all during the last five years, except two reviews from 2004 20 25 were not published in the last five years (Table 1). The overall number of participants included in these systematic reviews ranged between 60 participants for vitamin K supplements 20 and 215,900 participants for antioxidant supplementation 13. Only four reviews included less than 500 participants 16 17 22 32 . The number of trials per review ranged between one 17 and 55 19. Most studies were carried out in developed countries, except one conducted in low- and middle-income countries 27 . Most reviews defined a minimum follow up time (between three and six months) for RCTs to be included; the maximum follow up time was up to 20-30 years 29.
Low risk of bias for nine of the eleven items included in the AMSTAR checklist (Appendix 1) was showed by all studies, exception made by five studies whose authors did not evaluate the risk of publication bias and/or report the results 18-20 22 28 34 (item 10). Eleven studies did not report outcomes of interest, preventing to combine results and, thus, scoring was not applicable.
Deaths from any cause
Several interventions showed some risk reduction in all-cause mortality 14 15 18 18 27 34 8: reduce dietary salt 19, selenium supplementation 15 18, folic acid + vitamin D + calcium supplementation 18, general health checks 34, vitamin D supplementations 14 and practice of Qigong 29, but only the two last reached statistical significance (Table 2). The antioxidant supplements 13 showed a significant increase of 5% [95% CI 2-8%] for all-cause mortality risk.
Table 2 Effect of the interventions on all-cause mortality
| Intervention | Studies* N | Events N (%) | Effect RR (95%CI) | I2 (%) |
|---|---|---|---|---|
| Antioxidant supplementation13 | 19 | 21,334 (12.0) | 1.03 (0.97-1.08)a, b | 43.0 |
| 19 | 21,334 (12.0) | 1.05 (1.02-1.08)a, c | 46.9 | |
| 7 | 2,479 (6.6) | 0.93 (0.84-1.03)d,b | 0 | |
| Vitamin A supplementation18 | 1 | 115 (5.0) | 1.15 (0.81-1.65) | NA |
| Vitamin C supplementation12 18 | 1 | - | HR=1.07 (0.97-1.18)e | NA |
| 2 | 1,663 (11.2) | 1.06 (0.97-1.16) | NI | |
| Vitamin D supplementation14 | 48 | 11,955 (12.7) | 0.97(0.9-0.99)b | 0 |
| Vitamin D + Calcium supplementation18 | 1 | 1,551 (4.3) | 0.92 (0.83-1.01) | NA |
| Calcium supplementation18 | 3 | 1,827 (23.8) | 1.04 (0.96-1.12) | NI |
| Folic acid supplementation18 | 1 | 29 (2.8) | 0.52 (0.24-1.10) | NA |
| Selenium supplementation15 18 | 2 | 1,336 (7.2) | 0.97 (0.88-1.08)c | 0 |
| Omega 3 fatty acids25 | 44 | 1,995 (5.5) | 0.87 (0.73-1.03) | 0 |
| Advice to reduce dietary salt19 | 4 | 17 (0.7) | 0.90 (0.36-2.24)a | 0 |
| Qigong29 | 2 | 147 (28.8) | 0.54 (0.39-0.90)a,f | 0 |
| General health checks34 | 9 | 11,940 (7.7) | 0.99 (0.95-1.03)c | 0 |
| Multiple risk factor interventions11 | 14 | 2,511 (1.8) | OR=1.00 (0.96-1.05)c | 62 |
*: number of included studies providing information of interest; RR: relative risk; I2: Higgins’ heterogeneity; a: primary prevention trials with low risk of bias; b: random effects model analysis; c: fixed-effect model analysis; d: primary prevention trials with a high risk of bias; e: data from the clinical trial Physicians Health Study II (men aged ≥50); NA: not applicable; HR: hazard ratio; NI: No information; f: this data was calculated by the authors of the present review as it was not provided in the original review; OR: odds ratio. Effects in bold are statistically significant.
Cardiovascular mortality
Both multiple risk factor interventions 11 and Qigong practice 29 showed pooled significant reductions in stroke mortality (risk reduction 25% [95% CI 5-40%] and 45% [95% CI 26-58%], respectively) (Table 3).
Table 3 Effect of the interventions on cardiovascular mortality
| Intervention | Studies* N | Events N (%) | Effect RR (95%CI) | I2(%) |
|---|---|---|---|---|
| Vitamin C supplementation12 | 1 | - | HR=1.02 (0.85-1.22)a | NA |
| Selenium supplementation15 | 2 | 342 (1.9) | 0.97 (0.79-1.20) | 44 |
| Omega 3 fatty acids25 | 14 | 1,418 (4.3) | 0.85 (0.68-1.06) | 0 |
| Advice to reduce dietary salt19 | 2 | 17 (2.3) | 0.82 (0.56-1.21)b | 0 |
| Mediterranean dietary pattern21 | 1 | - | HR=1.01 (0.81-1.27)c,d | NA |
| HR=0.94 (0.65-1.35)d,e | NA | |||
| Qigong29 | 2 | 101 (19.8) | 0.55 (0.42-0.74)b,e,f | 1 |
| General health checks34 | 8 | 4,567 (29.6) | 1.03 (0.91-1.17)g | 64 |
| Multiple risk factor interventions11 | 11 | 901 (2.2) | OR=0.99 (0.92-1.07)c,g | 32 |
| 7 | 342 (0.6) | OR=0.75 (0.60-0.95)e,g | 0 |
*: number of included studies providing information of interest; RR: relative risk; I2: Higgins’ heterogeneity; HR: hazard ratio; a: data from the clinical trial Physicians Health Study II (men aged ≥50); NA: not applicable; b: random effects model analysis; c: coronary heart disease death; d: data from RCT WHI study (postmenopausal women); e: fatal stroke; f: this data was calculated by the authors of the present review as it was not provided in the original review; g: fixed-effect model analysis. Parameters in bold are statistically significant.
Cardiovascular events
The effect of interventions on CVD events, either a combined outcome of fatal and non-fatal events, or only non-fatal events are showed in Table 4. A significant risk reduction of 16% [95% CI 2-23%] was found in multifactorial individual lifestyle interventions 11, despite substantial between-study heterogeneity. Subgroup analysis showed that the combined event relative risk was 0.71 (p<0.05) in patients with either hypertension or diabetes, while there was no significant effect on patients without co-morbidity.
Table 4 Effect of the interventions on all cardiovascular events (fatal and/or non-fatal)
| Intervention | Studies* | Events | Effect | I2 |
|---|---|---|---|---|
| N | N (%) | RR (95%CI) | (%) | |
| Beta-Carotene supplementation18 | 2 | 2,157 (3.5) | 1.01 (0.93-1.09)a | NI |
| Vitamin C supplementation12 18 | 1 | - | HR=1.04 (0.87-1.24)b,c | NA |
| HR=0.96 (0.86-1.07)b,d | NA | |||
| HR=0.93 (0.84-1.03)b,e | NA | |||
| HR=0.89 (0.74-1,07)b,f | NA | |||
| 1 | 1,245 (8.5) | 0.99 (0.89-1.10)a | NA | |
| Vitamin D supplementation18 | 2 | 1,682 (21.1) | 0.94 (0.87-1.02)a | NI |
| Vitamin D + Calcium supplementation18 | 1 | 3,642 (10.1) | 1.01 (0.95-1.07)a | NA |
| Calcium supplementation18 | 1 | 96 (10.3) | 1.09 (0.75-1.60)a | NA |
| Vitamin E supplementation18 | 3 | 4,328 (6.0) | 0.97 (0.92-1.03)a | NI |
| Selenium supplementation15 18 | 2 | 2,036 (11.0) | 0.96 (0.89-1.04)g,h | 0 |
| 1 | 165 (1.0) | 0.79 (0.58-1.07)i | NA | |
| 1 | 22 (0.1) | 0.99 (0.43-2,29)j | NA | |
| 1 | 197 (1.1) | 0.90 (0.62-1.32)k | NA | |
| 2 | 2,329 (12.6) | 1.03 (0.95-1.11)a | 0 | |
| Omega 3 fatty acids25 | 31 | 2,628 (7.5) | 0.95 (0.82-1.12)a,l | 65 |
| 38 | 390 (84.0) | 0.86 (0.60-1.25)c | 48 | |
| 25 | 648 (3.8) | 1.03 (0.70-1.50)m | 51 | |
| 25 | 565 (3.3) | 0.77 (0.59-1.02)e | 66 | |
| 26 | 243 (0.7) | 1.17 (0.91-1.51)f | 0 | |
| 20 | 54 (0.7) | 0.51(0.31-0.85)n | 0 | |
| 17 | 11 (1.0) | 0.26 (0.07-1.06)o | 0 | |
| 37 | 416 (2.2) | 0.85 (0.49-1.48)p | 72 | |
| Advice to reduce dietary salt19 | 2 | 92 (12.3) | 0.82 (0,56-1,21)l,q | 0 |
| Mediterranean dietary pattern21 | 1 | - | HR=0.96 (0.89-1.03)a,r | NA |
| HR=0.91 (0.8-1.04)m,r | NA | |||
| HR=0.91 (0.82-1.01)d,r | NA | |||
| HR=1.04 (0.90-1.19)i,r | NA | |||
| Qigong29 | 1 | 80 (26.1) | 0.56 (0.38-0.83)f | NA |
| Health promotion interventions27 | 1 | 6 (2.6) | 0,57 (0.11-3.07) | NA |
| Digital Health Interventions33 | 2 | 42 (4.0) | 1,21 (0.58-2.58)a | 15 |
| Multiple risk factor interventions11 | 9 | 10,071 (8.3) | 0.84 (0.73-0.98)a,h | 82 |
*: number of included studies providing information of interest; RR: relative risk; I2: Higgins’ heterogeneity; HR: hazard ratio; a: all cardiovascular events (fatal and non-fatal); NI: not informed; NA: not applicable; b: data from the clinical trial Physicians Health Study II (men aged ≥50); c: myocardial infarction; d: revascularization; e: angina; f: stroke; g: non-fatal cardiovascular disease; h: fixed-effect model analysis; i: non-fatal stroke; j: hemorrhagic stroke; k: ischemic stroke; l: random effects model analysis; m: non-fatal myocardial infarction; n: heart failure; o: peripheral vascular events; p: sudden death; q: cardiovascular morbidity; r: data from RCT WHI study (postmenopausal women). Parameters in bold are statistically significant.
Qigong 29 showed a significant reduction of stroke incidence (44% [95% CI 17-62%]).
Increased consumption (dietary or supplemental) of omega 3 fatty acids, showed a significant risk reduction of heart failure of 49% (95% CI 15-69%) 25 .
Adverse events
Eight of the 24 included reviews reported adverse events related with the interventions: vitamin C 12, selenium 15 and mineral and multivitamin 18 supplements, increased consumption of fruits and vegetables 20, and also in nut 22, green and black tea 23 and dietary fibre 24 consumption (Table 1). Al-Khudairy et al 12 reported one RCT describing death, serious adverse events and adverse event without any further detail. In the other reviews, adverse events were a minor problem and/or the difference between the intervention and the control group was not statistically significant. In two reviews 18 21, the data regarding adverse events were aggregated, thus primary was indistinguishable from secondary prevention.
Discussion
This umbrella review provides information about diverse interventions for the primary prevention of CVD events and mortality. Our work constitutes the first comprehensive and systematic summary of diverse non-pharmacological interventions based on the umbrella review methodology. We included twenty-four reviews concerning non-pharmacological interventions; four of the thirteen reviews reporting CVD events or mortality found a significant reduction on them 11 14 25 29 .
Ebrahim et al 11 found that individual lifestyle interventions significantly reduced stroke-related mortality and any CVD event (fatal and non-fatal). These interventions pursued a behavioural change in order to alter more than one cardiovascular risk factor (i.e. diet, reduce blood pressure, smoking, total blood cholesterol or increase physical activity) by means of counselling or educational interventions, with or without pharmacological treatments. They were commonly provided by a variety of health professionals (physicians, nurses, nutritionists, dieticians, personal trainers, cooks, psychotherapists and physiotherapists). Focusing multicomponent lifestyle interventions (dietary advice, smoking cessation, stress reduction, physical exercise, etc.) on individual risk factors -taking account of the characteristics of the patients and, furthermore, targeting at those most likely to benefit (for example, those with hypertension or diabetes)- is very important, as these interventions seem to not benefit patients without the risk factors 11 .
Bjelakovic 14 found that vitamin D supplementation at any dose, for any duration, in monotherapy or in combination with calcium, administrated enteral or parenterally, significantly reduced all-cause mortality.
Hooper 26 found that omega-3 fatty acid intervention (dietary supplementation, a provided diet or advice on diet such that the product orally consumed as oil or capsule or as food stuffs had an omega 3 fat content of at least 10% of the total fat content) significantly reduced the risk of heart failure.
Finally, Hartley 29 found that Qigong significantly reduced stroke incidence, all-cause mortality and stroke-related mortality. Qigong is defined by the National Qigong Association 35 as an ancient Chinese health care system that integrates physical postures, breathing techniques and focused intention.
Systematic reviews included in this umbrella review were generally well conducted. Nonetheless, in most reviews authors called for caution in extrapolating results, as a high risk of bias in some of the RCTs was noted due to participants at varying levels of risk, different intensities interventions, lack of statistical power and insufficient length of follow-up.
Conducting RCTs in primary prevention can be challenging, and perhaps this explains scarcity of interventions and show evidence of benefit on CVD outcomes. The studies are targeted at a general population, which includes people at low risk. The interventions generally have a behavioural element, which requires a certain degree of collaboration from participants to maintain the healthy behaviour over long periods. A long follow-up is required to provide enough statistical power for CVD outcomes. The authors of the Qigong review 29 realised that the included trials were originally designed to study outcomes over one year, but it was unclear from the RCTs whether both Qigong was continuously practiced over the whole follow up and if randomization was preserved in the analysis of the RCTs over the whole follow up. Other reviews have found more positive results on intermediate, short term outcomes such as cholesterol level and high blood pressure 13 16 17 19 22 24 26-28 30 .
The effectiveness of vitamin D in reducing all-cause mortality concur with the results found by Theodoratou et al 36 who conducted an umbrella review across systematic reviews and meta-analyses of observational studies of 25-hydroxyvitamin D or 1,25-dihydroxyvitamin D concentrations in plasma, along with randomised controlled trials where the intervention was vitamin D supplementation for multiple health outcomes. The results of meta-analysis showed some insignificant risk reduction in CVD and mortality outcomes across randomised controlled trials with vitamin D supplementation, while a statistically significant risk reduction was seen in the observational studies about vitamin D concentrations in plasma.
Apart from vitamin D, no other supplements showed a benefit for cardiovascular prevention. The supplements are marketed in several countries, and many people take these pills believing that it will improve their health 37 . Nevertheless the effectiveness and safety of these products ought to be further evaluated. Antioxidant supplements may even increase the risk of overall mortality 13 . Furthermore, other meta-analyses have associated supplements with increased risk of some cancers 37 .
Limitations of this review are: only reviews in English and Spanish were included, and grey literature (e.g. non peer reviewed papers in unconventional information channels) was not searched; thus, we may therefore have missed some relevant studies. The strengths are: the review was conducted over several databases, and the overall risk of bias is likely to be very low as the majority of the included reviews are from Cochrane collaboration.
A limitation of the umbrella review methodology may be that we have assessed the quality of the systematic reviews using the risk of bias tool, not the quality of the individual studies within these reviews. Furthermore, the trials included in each systematic review can be quite heterogeneous in population (the Women’s Health Initiative about diet in postmenopausal women 20, or the Physicians Health Study II about vitamin C supplementation in male physicians aged 50 years or older 12 ), causing that observed results may not be transferable to the general population. More research should be conducted to assess whether the effect of treatment varies according to population characteristics such as gender 38, age or social class 39 . Longer-term monitoring of patients is needed to assess CVD.
Finally, the results of this review should be carefully interpreted as most studies compare non-pharmacological interventions for primary prevention against usual care, which can include recommended pharmacological treatment in higher risk patients (e.g. statins and/or antihypertensive, etc.). This means that these trials are measuring the combined effect of non-pharmaceutical interventions alongside standard pharmacological practice, rather than considering these interventions as an alternative to pharmacology. More research is also recommended to evaluate combinations of different interventions.
These results inform to future guidelines and clinical practice for intervention related to primary prevention of CVD. From the list of putative interventions to improve the cardiovascular health of patients in primary prevention, four of them have been shown to be effective, with minor adverse events if any. Other non-pharmacological interventions did not achieve statistical significance for CVD prevention, but did achieve significant improvements in other intermediate outcomes, such as blood pressure and lipid levels. Hence, further research is required to establish if these interventions, alone or in combination, translate into definite long-term health benefits. Given the limited evidence to date, there is therefore a need for well-conducted RCT with higher-quality design focused on long term outcomes such as CVD, as well as intermediate and surrogate endpoints, involving a long-term monitoring of non-pharmacological interventions on primary prevention to assess their effects on CVD and mortality. The role of the healthcare professional in engaging and communicating with patients also deserves further attention.
