Interventions
Data on the choice of interventions for blood pressure, cholesterol, and bodyweight and their effectiveness are now presented.
Choice and Classification of Interventions
A variety of population-based and personal interventions could potentially be used to address the risks associated with high blood pressure, cholesterol, and bodyweight. Of the personal interventions discussed in this section—lifestyle and dietary, pharmacological, and surgical interventions—two main strategies exist for choosing whom to treat: those above certain threshold values of single risk-factor levels and those above certain values of absolute cardiovascular (or global) risk, which is determined by the levels of multiple factors.
Targeting treatments by levels of a single risk factor (such as hypertension) does not effectively focus on overall risk of developing CVD, which is mainly determined by the net effects of other risk factors. For example, the predicted 10-year CVD risk for someone with an SBP of 140/90 mmHg can vary from 5 to 50 percent depending on the number of concomitant risk factors. The number of people who would need to be treated to prevent an event can therefore vary by an order of magnitude, even if they have the same blood pressure levels. Thus, a treatment strategy based only on individual risk-factor levels is likely to result in high-risk patients being undertreated and many patients at relatively low risk being treated with little absolute benefit, which is unlikely to be the best allocation of scarce health care resources.
The absolute-risk strategy was developed in New Zealand (Jackson and others 1993) and has been adopted extensively elsewhere, for example, by the British Hypertension Society (Ramsay and others 1999) and the Joint Task Force of European and other Societies on Coronary Prevention (Wood and others 1998). The absolute CVD risk is estimated using risk assessments such as the Framingham risk function (Anderson and others 1991) or the Prospective Cardiovascular Munster Study score (Assmann, Cullen, and Schulte 2002) on the basis of the number and severity of CVD risk factors. Targeting treatments at those at high absolute risk rather than those above arbitrary thresholds ensures a favorable ratio of benefits to risks. It can be expected to reduce events in the large proportion of people who are, for example, nonhypertensive but who still have nonoptimal blood pressure (Rose 1981). Combinations of personal interventions targeted at those at high absolute risk also have the potential of being highly cost-effective.
The simplest indicator of high absolute risk is established CVD, principally myocardial infarction, angina, stroke, or transient ischemic attack. For example, without preventive treatment, people who have had a myocardial infarction face an annual risk of death from coronary heart disease of about 5 percent (Law, Watt, and Wald 2002). That risk persists indefinitely—probably for the rest of a person's life—and varies little with age or sex.
However, many individuals with no history of CVD are at similar elevated risk for future CVD as a result of constellations of elevated risks. Thus, the distinction between primary and secondary prevention is somewhat artificial and could lead to undertreatment of many high-risk individuals. While recognizing that the distinction is somewhat arbitrary, this chapter discusses the cost-effectiveness of efforts to manage those without previous CVD, and chapter 33 reviews the management of those with known vascular disease. A unifying system targeting treatments at those at highest risk, either with CVD or multiple risk factors, is likely to be highly cost-effective because more than 75 percent of events occur in the 5 to 10 percent of people with CVD or specific clusters of risk factors (Haq and others 1999; Tosteson and others 1997).
The limitations of the individual-risk-factor approach, together with increasing evidence that the thresholds do not have any biological justification, have motivated the adoption of strategies that take other risk factors into account. Although the most complete way of doing so is using the absolute-risk strategy outlined earlier, one intermediate strategy involves lowering the thresholds of blood pressure or lipid levels at which treatment is initiated if one or more additional CVD risk factors, such as diabetes, are present (Chobanian and others 2003).
Intervention Effectiveness
This section summarizes data on the effectiveness of population-based interventions and personal interventions (lifestyle and dietary interventions and pharmacological and surgical interventions). The studies concerned have mainly been conducted in developed countries.
Population-based Interventions
Investigators have undertaken a variety of population-based community intervention studies, mostly in developed countries in the 1970s and 1980s (for further details see chapter 44). These studies have tended to be multifactorial projects testing whether comprehensive community programs could produce favorable changes in such risk factors as bodyweight, cholesterol, and blood pressure and in CVD morbidity and mortality (Schooler and others 1997). In general, they included a combination of populationwide and individual interventions, including messages disseminated through local associations, sports clubs, the media, and food associations; healthy food options at restaurants and worksite cafeterias; food labeling at supermarkets; face-to-face communication at meetings and distribution of educational materials; smoking restrictions; and competitions to develop healthy food. Except in Finland, the projects had mixed results, although many demonstrated significant effects with respect to individual components of the interventions. The limitations of many of the projects include inability to detect small but potentially important changes in risk factors, short duration of intervention and follow-up, and issues with outcome measures. Some have also suggested that those trials with less favorable results may have lacked adequate community support and public policy initiatives (Feinleib 1996; Mittelmark and others 1993; Schooler and others 1997; Susser 1995).
A number of population-based interventions have also taken place in developing countries, including the following:
-
In China, the Tianjin Project showed a significant reduction in sodium intake in men after three years of intervention, and after five years, the prevalence rates of both hypertension and obesity decreased among 45- to 65-year-olds (Schooler and others 1997).
-
In Chile, the Mirame Project was a three-year intervention program designed to provide and evaluate strategies to promote healthy lifestyles among schoolchildren and their families. Nissinen, Berrios, and Puska (2001) report a significant positive effect on some risk factors for the intervention schools.
-
In Mauritius, government-led initiatives resulted in a change in the composition of cooking oil from mostly palm oil, which is high in saturated fatty acids, to wholly soybean oil, which is high in unsaturated fatty acids. From 1987 to 1992, total cholesterol concentrations fell significantly, and the estimated intake of saturated fatty acids decreased, with much of this finding reportedly resulting from the change in cooking oil (Uusitalo and others 1996).
An effective populationwide intervention draws together different kinds of feasible activities that combined produce a synergistic effect (Nissinen, Berrios, and Puska 2001; Puska 1999). Even though the projects and trials were undertaken in a range of different communities and used a variety of methods and interventions, several common themes emerge. Some of the important elements of a successful program that enables individuals to adopt healthier lifestyles include the following:
-
clear responsibility for coordinating prevention efforts, with credible agencies with good communication methods carrying out long-term education programs
-
intersectoral collaboration, with multiple messages sourced from different organizations, including health sector entities, nonhealth government agencies, schools, workplaces, religious organizations, and voluntary agencies
-
collaboration with the food industry to ensure the availability of reasonably priced healthier food options, with food labeling that presents relevant information in a clear, reliable, and standardized format
-
realistic multiyear time frames.
Lifestyle and Dietary Personal Interventions
Many guidelines have concluded that lifestyle modifications, such as weight loss, healthy diet (such as one rich in potassium and low in sodium), physical activity, and moderate alcohol consumption are effective in reducing blood pressure (see, for example, Chobanian and others 2003). Trials indicate that a reduction of salt intake lowers blood pressure, with larger blood pressure reductions in the elderly and in those with higher initial blood pressure levels (Law, Frost, and Wald 1991; Whelton and others 1998). An increase in daily fruit and vegetable intake may also lower blood pressure, and when combined with an increase in low-fat dairy products and a reduction in saturated and total fat, may lower blood pressure even more (Appel and others 1997). Weight reduction lowers blood pressure in proportion to the amount of weight lost (Whelton and others 1998), and physical activity appears to lower blood pressure in a way that may be independent of weight loss. High levels of alcohol intake are associated with blood pressure elevation, which is reversible by reducing intake (Kaplan 1995).
Dietary approaches to lowering total cholesterol and low-density lipoprotein (LDL) cholesterol typically involve reduced intake of dietary fats, particularly saturated fats. Evidence suggests a dose-response relationship between saturated fatty acid intake and LDL cholesterol levels (NCEP Expert Panel 2002). Plant sterols and stanols have recently been incorporated into foods such as margarine and can reduce LDL cholesterol by about 10 percent; however, this approach is currently relatively expensive (Law 2000). Dietary advice may also suggest increasing the intake of viscous fiber—for instance, in the form of cereal grains, fruits, and vegetables—because these dietary sources may enhance the lowering of LDL cholesterol. Maintaining bodyweight in the desirable range and engaging in moderate physical activity complement these dietary strategies (NCEP Expert Panel 2002).
Increases in obesity have been related to declines in energy expenditure (for example, reductions in physical activity and adoption of a more sedentary lifestyle) and a higher intake of energy-dense but micronutrient-poor foods, such as most processed foods (WHO 2003b). A variety of trials have recorded beneficial health effects, with weight reduction achieved by a combination of interventions (NHLBI Obesity Education Initiative Expert Panel 1998). These interventions include dietary counseling and therapy that involves a decrease in daily caloric intake and a reduction in saturated fats and total fats. An increase in physical activity is an important component of weight-loss therapy. Behavioral strategies revolving around self-monitoring of eating habits, stress management, problem solving, and social support may also complement these approaches. Overall, however, the effects of lifestyle modifications to reduce weight and maintain the weight loss are relatively poor, with many reports finding that weight returns to baseline levels after several years.
Pharmacological and Surgical Personal Interventions
Randomized trials have shown that medications to lower blood pressure effectively reduce the risk of stroke, IHD, and heart failure. Results from meta-analyses of more than 40 different trials published in 2003 included about 210,000 participants and more than 8,000 stroke and 11,000 IHD events (Blood Pressure Lowering Treatment Trialists' Collaboration 2003; Fox and EUROPA Investigators 2003; Law, Wald, and Rudnicka 2003; Lawes and others 2004; Pepine and others 2003). The trials may be broadly classified into three groups: (a) drug versus placebo trials, (b) more intensive regimens to lower blood pressure versus less intensive regimens, and (c) drug versus drug trials.
The drug versus placebo trials achieved the greatest reductions in blood pressure, and a dose-response relationship was apparent between blood pressure reduction and reduced risk of stroke. Overall, the trials indicated that a 10 mmHg reduction in SBP would result in a 32 percent reduction in stroke risk and a 14 percent relative reduction in IHD risk. This finding is consistent with the size of associations observed in cohort studies.
Clear evidence indicates that all the major drug classes have similar effects on the risk of stroke and coronary heart disease per mmHg reduction in blood pressure (Blood Pressure Lowering Treatment Trialists' Collaboration 2003; Lawes and others 2004). The only clear evidence of clinically important, class-specific effects are with agents that block the reninangiotensin system, which reduce diabetes incidence by about one-quarter, and with calcium channel blockers, which reduce heart failure less than other agents (although this result may be partly caused by misclassification, because a known side effect of calcium channel blockers is ankle edema, which is a diagnostic component of heart failure). Because all agents lower blood pressure by about the same modest amount and because their effects on blood pressure are additive (Law and others 2003), the key issue seems to be which combinations of two or more drugs should be provided and how long-term adherence can be maximized.
Over the past three decades, numerous trials have assessed the effect of different cholesterol-lowering interventions (Law, Wald, and Rudnicka 2003; Law, Wald, and Thompson 1994). The placebo-controlled trials can be broadly classified into those testing fibrates, statins, and other interventions (mostly dietary interventions, but also some other interventions such as resins and niacin). The statins are the most effective in lowering total and LDL cholesterol, with reductions of more than 1 mmol/l in most trials. A good correlation has been found between reduction in total cholesterol and relative risk reduction. This finding suggests, as for trials investigating blood pressure lowering, that even though some drugs are more effective in achieving greater reductions in risk factors, their effect on disease outcomes is similar per unit reduction of cholesterol. Overall, a 1 mmol/l reduction in total cholesterol is associated with a 21 percent relative risk reduction in IHD and a 17 percent reduction in risk of stroke. Again, this finding is consistent with the epidemiology, with the proviso that the vast majority of strokes in clinical trials were ischemic.
In clinical trials of statins, the relative risk reduction in cardiovascular events is similar at all levels of baseline cholesterol, extending to levels below 5 mmol/l total cholesterol, and is also consistent among patients who are and are not taking concurrent blood pressure lowering and other medications (Heart Protection Study Collaborative Group 2002). Similar findings are observed with treatments to lower blood pressure (Progress Collaborative Group 2001), indicating that these treatment effects are independent. This finding is plausible, because they act through different mechanisms and because observational studies do not suggest a large interaction (Neaton and Wentworth 1992).
The benefits of lowering blood pressure and cholesterol are achieved surprisingly rapidly: for most outcomes, risk appears to be fully reversed within 6 to 18 months of beginning treatment. For example, individuals with cholesterol lowered in the past two or more years are at approximately the same coronary heart disease risk as otherwise identical individuals whose cholesterol has been at that level for decades (Law, Wald, and Thompson 1994).
Pharmacological agents for weight loss that have been subject to randomized controlled trials include dexfenfluramine, sibutramine, orlistat, and phentermine/fenfluramine (although the last has been withdrawn because of a reported association between the drugs and valvular heart disease). Overall, trials suggest only modest weight-loss effects, with an average net weight loss of 1.5 kg after eight weeks and 2 to 3 kg after one year (NHLBI Obesity Education Initiative Expert Panel 1998). A systematic review of orlistat trials indicated a pooled net weight loss of 1.2 kg at 12 weeks, 2.9 to 3.4 kg at one year, and 2.5 to 2.4 kg at two years (O'Meara and others 2001). Results of a systematic review of trials assessing sibutramine were similar (O'Meara and others 2002), with fewer data available on long-term sustained weight loss.
Investigators have also undertaken several randomized controlled trials to assess the effects of different surgical interventions, generally in individuals with a BMI equal to or greater than 35 or 40 kg/m2. Weight loss resulting from gastric bypass varied from 50 to 100 kg six months to a year following surgery (NHLBI Obesity Education Initiative Expert Panel 1998). Overall, several trials suggest that surgery resulted in about 23 to 37 kg more weight loss than conventional treatment and that this loss was maintained for eight years (Clegg and others 2002). Furthermore, gastric bypass surgery appears to be more beneficial than gastroplasty or jejunoileal bypass.
In relation to compliance and adherence with pharmacological therapy, population surveys have demonstrated that, even in industrial countries, high blood pressure is either untreated or inadequately controlled in about 70 to 75 percent of patients and that adherence to medications among patients suffering from chronic disease is only about 50 percent (WHO 2003a). The extent of poor adherence is likely to be even greater in developing countries given the relative lack of health services and inequities in access. Pharmacotherapy faces a variety of potential barriers, including the symptomless nature of the conditions, a lack of knowledge or denial of risk, the complicated nature of drug regimens, the risk of side effects (real and perceived), and the costs of treatment.
Health providers may use multiple strategies to increase compliance and adherence. Patient-centered interventions include involving individuals in the decision-making process; providing individualized patient education and disease counseling and adapting treatment to patients' lifestyles; simplifying dosing schedules; providing drug information leaflets, medication charts, and special reminder packaging for medications; holding group sessions for education and family-oriented disease management therapies; and implementing automated telephone assessment and self-care education calls with nurse follow-up (Haynes and others 2003).
Strategies may also aim to increase physician adherence, and interventions may include the use of guidelines, peer review and audit, and prompts to remind physicians to review risks and medications (Ebrahim 1998; NCEP Expert Panel 2002). These strategies obviously do not address issues pertaining to resources and access in poor countries.
Several trials and overviews have attempted to assess the value of different interventions to improve compliance and adherence; however, issues have arisen in connection with the generalizability of the interventions, the low statistical power in many trials, the lack of description of all parts of interventions, and the assessment of complex interventions without assessment of the separate effects of the intervention components. Haynes and others' (2003) systematic review concludes that, overall, no single approach to improving adherence can be recommended. Simpler treatment regimens can sometimes improve adherence and treatment outcomes for both short-and long-term treatments. Several complex strategies, including combinations of more thorough patient instructions and counseling, easier access to care, reminders, close follow-up, supervised self-monitoring, family therapy, and rewards for success can improve adherence and treatment outcomes in some patients. However, even the most effective interventions did not lead to large improvements in adherence or treatment outcomes and were relatively resource intensive. By contrast, Connor, Rafter, and Rodgers's (2004) systematic review indicates improved adherence and clinical outcomes with fixed-dose combination treatment or unit-of-use packaging.
Few good, evidence-based strategies to improve obesity management are currently available, although reminder systems, brief training interventions, shared care, inpatient care, and dietitian-led treatments may all be worth further investigation (Harvey and others 2003). Thus, a clear need for innovations still exists to help people follow medication prescriptions as well as dietary and lifestyle advice.
