The relationship between changes in health behaviour and initiation of lipid-lowering and antihypertensive medications in individuals at high risk of ischaemic heart disease

Non-pharmacological treatment, in terms of physical activity, dietary habits and smoking cessation, has been shown to have a beneficial effect on biological cardiovascular disease (CVD) risk factors. Physical activity is known to modify well-known CVD risk factors, such as, hypertension, dyslipidemia, abdominal and general obesity [1–5]. Also, individual dietary interventions in primary prevention have shown moderate effects in relation to cardiovascular risk factors, such as, blood pressure, total cholesterol, LDL cholesterol stemming from dietary changes in terms of lower salt and fat, higher fibre and fruit and vegetable consumption [6–9]. Furthermore, studies have shown that quitting smoking reduces the risk of acute myocardial infarction [10, 11].

Regarding pharmacological treatment, there have been well-reasoned and balanced recommendations for the use of statins for individuals, who have an elevated risk of cardiovascular events [12–14]. Furthermore, blood-pressure lowering drugs reduce the risk of coronary heart disease (CHD) events and stroke, also in individuals with no previous history of cardiovascular disease [15]. There is also evidence that users of medications for hypertension and hypercholesterolemia may benefit from health behaviour changes such as improved diet, quitting smoking and exercise, particularly in the setting of primary prevention [16]. Danish national guidelines, from 2007, recommended that the general practitioner should motivate people at high risk of CVD (10-year absolute risk of CVD death ≥ 5%), and with no pre-existing CVD, to pursue health behaviour changes within 3 months, before considering pharmacological treatment [17]. Only in terms of a very high risk (10-year absolute risk of CVD death ≥ 10%), the general practitioner should not hesitate to initiate pharmacological treatment.

Several observational studies have reported that individuals who initiate and adhere to preventive therapy such as statin use are likely to be more health conscious, the so-called healthy adherer effect [18–21]. ‘Healthy adherer patients’ are more likely to see their doctors on a regular basis, exercise, have healthier diets, stop smoking, adhere to treatment and avoid risky behaviours.

In the present study, we investigated a possible healthy adherer effect among individuals with an elevated risk of ischaemic heart disease (IHD) in terms of high blood pressure and high total cholesterol level. We hypothesised that individuals who were able to change their health behaviours, also were more likely to initiate treatment with lipid-lowering medications and blood pressure-lowering medications, rather than substituting medications with behavioural change. To explore this, we examined whether long-term changes in dietary habits and physical activity were associated with initiation of preventive medications. We conducted the study by using two subsamples from the population-based cohort from the Inter99 study.

The main finding of this study was that positive changes in dietary habits were associated with increased initiation of lipid-lowering and antihypertensive medications. Furthermore, participants, who reported an increase in terms of physical activity were more likely to initiate treatment with lipid-lowering medications; however, in terms of antihypertensives, no association was observed. Some associations may be non-significant due to type 2 errors resulting from the relatively small sample size. Our hypothesis that people who were able to change their health behaviour also would tend to take better care of themselves, by greater adherence to preventive medications, was therefore supported to some extent. The purpose of this study was not to assess whether patients with increased blood pressure or total cholesterol ought to initiate treatment with preventive medications and/or change in health behaviour, but to shed light on a possible health conscious behaviour and thereby also lack of health consciousness in other individuals.

A study by Brookhart et al. reported that patients who adhered to statin therapy underwent a variety of cancer screening tests, received influenza and pneumococcal vaccinations, compared with those who were less adherent [26]. In another study by Majumdar et al., focusing on patients admitted to hospital with community acquired pneumonia, statin users were more likely to be former smokers and have up-to-date immunizations for pneumococcus and influenza [18]. Furthermore, Dormuth et al. reported that patients with good adherence to therapy with statins were less likely to be involved in motor vehicle or work place accidents requiring medical attention than less adherent patients [20]. Although the studies differ in terms of design, outcome measure and definition of adherence, they all suggest a healthy adherer or healthy user effect.

Medical treatment with preventive medications is likely to be determined by system-, provider, as well as patient-level factors, and the factors also interact [27]. In this study, it might be speculated that it was the most health conscious participants who consulted their general practitioner regarding their increased risk of IHD. The most health conscious individuals may be more open-minded regarding preventive medications and may even demand such treatment. Nevertheless, health consciousness in itself does not lead to the prescribing of medications. In Denmark, primary prevention of cardiovascular disease is often handled by the general practitioners. Almost all participants in our study visited their general practitioner within one year after baseline (Table1), but it remains unknown, how many actually consulted their general practitioners due to their increased risk of IHD and/or motivation for health behaviour change, as the Health Insurance Registry does not provide information about diagnoses related to contacts.

On the other hand, participation in a non-pharmacological intervention (as with the Inter99 study) might have affected the general practitioners’ decisions regarding treatment, in terms of a hesitation towards initiating treatment. Generally, the prescribing of preventive medications has been much debated, also among general practitioners. A recent Cochrane review, reported a stronger effect of statins on overall mortality than observed in previous meta-analyses [14], but at the same time the authors did not support the widespread use of statins in individuals without a previous history of heart disease, due to high costs and risks associated with side effects [28]. Another fact is that Danish national guidelines from 1998 and 2002 were less likely to recommend pharmacological treatment as primary treatment for IHD, as was the case in the 2007 guideline. Also, general practitioners were very reluctant to initiate treatment with statins, as the price was very high and regulations on subsidies were very complicated in the first years of the study. This is likely to explain the relatively low prevalence of users.

Generally, participants with unchanged/negative changes in health behaviour were less likely to initiate treatment with preventive medications, but this was not the case for antihypertensives and physical activity, where no healthy adherer affect was observed. One explanation for this finding might be that physicians have a tendency to prescribe medications to patients, who are most likely to benefit from them [27], for example as a result of unhealthy behaviours or poorer health status. However, we tried to take this into consideration, by excluding those who reported a decrease in physical activity due to illness/handicap and by adjusting for blood pressure/cholesterol and risk score. Differences in health status did appear to have some effect, as some associations were strengthened by the inclusion of the above mentioned variables.

The generalisability of our study should be made carefully, as our study population is relatively young, received lifestyle intervention and were referred to their general practitioners. The strengths of this study are the follow-up design and the fact that we were able to link data to the Danish National Prescription Registry of high validity [29]. Our study also has limitations. The relatively low participation rate at baseline might cause some selection bias, because group A had lower BMI than the general population (group C), but poorer lifestyle [22]. Nevertheless, the intervention groups have been found to be representative for the general population (group c) with regard to former admissions for all causes, IHD, CVD and diabetes [22]. Another thing to keep in mind is the selection in drop-out at five-year follow-up. It is likely that it was the most health conscious individuals who returned at follow-up. However, we consider this less relevant, as we investigate associations for participants with complete data. Furthermore, we cannot exclude the risk of residual confounding due to broad categories of dietary habits and physical activity. A weakness is also that health behaviour data are self-report, which involves a risk of bias and misclassification. It cannot be ruled out that some participants may have reported too positive changes in health behaviour; however, this would lead to underestimated associations between changes in health behaviour and initiation of medications. Previous cardiovascular events might also cause a selection and therefore affect the findings; but, due to the relatively young population, and the fact that we excluded those with existing use of medications and diabetes, we consider this less relevant. We might have underestimated the associations, as participants who were the healthiest at baseline may have difficulties in increasing their scores; the so-called ceiling effect. Finally, we did not consider subgroups of medications, combinations of different medications or the fact that comorbidity, such as with migraines, might require antihypertensives.

The study partially supports the hypothesis of health conscious behaviour in terms of increased initiation of preventive medications among those who experienced positive changes in health behaviour. This fact implies that studies that examine the effect of lipid-lowering and antihypertensive medications might consider the healthy adherer effect in terms of a potential bias. From a public health perspective, our study suggest that more attention should be given to identifying individuals and groups, who are less health conscious and thereby less likely to engage in either preventive medication or changes in health behaviour. Further research is needed to understand the complex interactions between healthy behaviour and preventive medication and the role of individual perceptions of health, health behaviour and medication.

NFH conceived the study, conducted the analysis of the data and had primary responsibility for drafting the manuscript. All authors contributed to the design of the study and participated in the interpretation of results and critically reviewed the manuscript. All authors read and approved the final manuscript.