Dietary patterns of patients with chronic coronary heart disease: a cross-sectional study

Este trabajo fue recibido el 05 de junio de 2019. Aceptado con modificaciones: 24 de septiembre de 2019. Aceptado para ser publicado: 20 de noviembre de 2019. ABSTRACT The objective of this study was to identify dietary patterns in an outpatient cohort of coronary heart disease (CHD) patients, to compare these patterns with dietary recommendations of the international cardiology guidelines, and verify associations with cardiovascular risk factors. Dietary intake was assessed through a food frequency questionnaire. Dietary patterns were identified by cluster analysis. The total energy intake, fiber, cholesterol, macro and micronutrients were calculated. Two dietary patterns were identified in 123 patients. Pattern I was characterized by a greater consumption of whole carbohydrates, beans, meats, vegetables, and fruits. Pattern II was rich in refined carbohydrates, fried foods, and sweets. Participants in pattern I had lower values of diastolic blood pressure (DBP) with 77.1 ± 9.9 mmHg (p= 0.002) and glycated hemoglobin (HbA1c) of 7.83 ± 1.76 % (p= 0.029) compared to pattern II with 84.1 ± 14.3 mmHg and 9.02 ± 2.29 %, respectively. Therefore, pattern I had a healthier nutritional composition, however, nutritional adequacy was still lacking. Despite this, participants in pattern I had significantly lower values of DBP and HbA1c, in addition to being associated with a better control of DBP.


INTRODUCTION
Coronary heart disease (CHD) is a chronic inflammatory disease of multifactorial etiology which occurs in response to endothelial aggression 1 . Lifestyle modifications are important to assist in the control of cardiovascular risk factors, including adequate management of comorbidities, such as diabetes mellitus, hypertension, and dyslipidemia. A healthy lifestyle involves body weight control, smoking cessation and adherence to a healthy eating pattern 2 .
Consistent data have shown that healthy dietary patterns characterized by consumption of vegetables, fruits and whole grains are associated with a lower risk of CHD and incidence of major cardiovascular events 3,4,5,6,7,8 . Similarly, inadequacies in the consumption of diet components, such as carbohydrates, fats, and their subtypes are related to a worse control of risk factors and cardiovascular outcomes 9,10 . In recent decades, the role of dietary factors in the development of CHD has been studied in epidemiological research. Many studies have been conducted to identify dietary patterns that characterize the combination of the usual intake of food groups, thus providing a more comprehensive understanding of how diet affects the etiology of diseases 11,12 . This evaluation can be analyzed in two ways: a priori, eating patterns are defined based on a scoring system or a posteriori, when data from dietary surveys are aggregated through specific statistical analysis 13 .
However, the relationship between healthy dietary patterns and their diet components with cardiovascular risk factors should be better established. Despite research in recent years, there are still few studies on the dietary pattern of patients with ischemic disease in Brazil. Therefore, the objective of this study was to identify and describe dietary patterns defined a posteriori, in order to compare them with the dietary recommendations from literature, and to verify associations with the control of cardiovascular risk factors in a group of patients with stable CHD.

METHODS
This cross-sectional observational study was conducted in an outpatient tertiary care cardiovascular clinic at a university hospital in the south of Brazil. The study was approved by the hospital research and ethics committee under protocol number 16-0362. All procedures involved in this study was conducted in accordance with the ethical principles of the Declaration of Helsinki of 1975, updated in 2013. All participants signed an informed consent document before entering the study. This study was conducted and described according to STROBE guidance (Strengthening the Reporting of Observational Studies in Epidemiology) 14 .

Study Sample
This study was conducted in patients with stable CHD who had been receiving cardiovascular care for at least 3 months. CHD was defined as a clinical history of at least one of the following documented episodes: 1) a documented episode of acute myocardial infarction (AMI); 2) percutaneous or surgical revascularization; 3) coronary angiography with evidence of lesions ≥ 50% of the left coronary artery, or ≥ 70% in the diagonal, circumflex or right coronary arteries; 4) positive non-invasive testing of ischemia. In addition, patients should have preserved cognitive capacity to respond the data collection questionnaires. The patients were required to have also at least six teeth in the mouth, because this present study is related to a line of study on oral health and nutrition. Exclusion criteria were: no stable CHD diagnosis, outpatient follow-up <3 months, presence of documented neurological sequelae and <6 teeth present in the mouth.
The protocol of cardiovascular care in this outpatient clinic included statins and antiplatelet agents for all patients. Oral hypoglycemic drugs, insulin, b-blockers or angiotensin-converting enzyme inhibitors and other antihypertensive medications were also prescribed when necessary. Non-pharmacological counseling included general guidance about lifestyle modification in relation to diet and physical activity.
The size of the group of subjects to measure the dietary pattern was adequately planned. Ten individuals were included for each food group studied (12 food groups were studied along with 123 subjects) 15 .
The medical records of 541 patients were consecutively reviewed from November 2016 to December 2017, in accordance with the inclusion and exclusion criteria of the study. Fifty-three subjects (9.8%) were not diagnosed with stable CHD, 9 (1.6%) had less than 3 months of follow-up at the outpatient clinic, and 6 (1.1%) had documented neurological sequelae. Of the remaining 473 patients, 236 were invited for convenience, depending on the patient's availability to participate in the study on the day of outpatient medical appointment. Forty-three patients (18.2%) refused to participate in the study, 68 (28.8%) did not meet the criteria to perform the oral examination, and 2 (0.8%) were excluded because they had incomplete questionnaires. Thus, the final group of subjects was 123 individuals. Figure 1 illustrates the study flow diagram.

Dietary intake
Dietary intake was assessed by a quantitative food frequency questionnaire (FFQ) previously prepared and validated with individuals from Southern Brazil 16 . A trained nutritionist carried out its application in all participants. The FFQ included 135 food items and the retrospective period of the questionnaire comprised the past 12 months of food intake. In addition, a food photo album was used to assist participants in responding to the size of the portion consumed 17 . The application of the questionnaire took on average one hour and thirty minutes per participant.
The intake report obtained by the FFQ was converted into daily consumption to estimate the nutritional composition based on the Programa de Apoio à Nutrição (NutWin) 18 . The monounsaturated, polyunsaturated and trans fatty acids types were not included in the original validation of the FFQ and their nutritional composition was also estimated based on the NutWin, in addition to Nutribase Nutrition Software. Nutrients intake was compared with the nutritional recommendations of the guidelines of the Brazilian and International Societies of Cardiology 2,11,19 .

Cardiovascular risk factors and therapeutic targets
Biochemical exams of cardiovascular risk factors were reviewed in the medical records and if there were no results referring to the month prior to the date of the interview, a request was made for the participant to have blood tests done. Triglycerides (TG), total cholesterol (TC), and highdensity lipoprotein cholesterol (HDL-C) were measured by automated enzymatic calorimetry methods. Low-density lipoprotein cholesterol (LDL-C) was calculated using the Friedewald formula [LDL-C= total cholesterol -(HDL-C + TG/5)] for individuals with TG <400 mg/dL 19 . Glycated hemoglobin (HbA1C) was collected only from diabetic patients and its measurement was carried out by the highprecision chromatography method.

Statistical analysis
FFQ foods were aggregated into 12 groups and the amount consumed from each food group was converted into a percentage of total daily caloric intake. We performed a cluster analysis based on food groups to derive two non-overlapping groups (dietary patterns) using the K-means method. Median and interquartile range were calculated for each of the 12 food groups and two clusters were selected because they were best interpretable because of the number of individuals in each group.
The assumption of normality was examined for all evaluated variables by Shapiro-Wilk test. The Student's t test, Mann-Whitney test, and chi-square test for independent samples were used to test differences across the dietary patterns in comparisons of parametric, nonparametric, and categorical variables, respectively. Energy and nutrient intake data were adjusted before analyses for energy intake according with the residual method 21 . To investigate the associations between dietary patterns and achieve therapeutic targets, the Poisson regression with robust variance analysis was carried out. Thus, in the analysis the effect of dietary patterns on each therapeutic target (dependent variable) was estimated.
Analyses were performed using the SPSS version 22.0 and type I error rate was fixed at P<0.050 (2-tailed).

RESULTS
Our sample consisted of 123 patients with CHD. Two dietary patterns were identified by cluster analysis, as shown in table 1. The first cluster, defined as Pattern I was characterized by a high intake of whole carbohydrates, beans/oilseeds, white and red meats, vegetables and fruits. The second cluster identified was defined as Pattern II and showed a higher intake of refined carbohydrates, fried foods, and sweets/sugars. Clinical characteristics of the patients according to the dietary pattern are shown in table 2. Most clinical characteristics did not differ between groups, but the participants of Pattern I were older and had a longer outpatient follow-up time. Considering smoking, a significantly amount of individuals from pattern II were current smokers (or had stopped <1 year). Use of medication was similar between both groups. Individuals who had previously had a nutritional counseling, specifically with a nutritionist, were significantly associated with pattern I. Nutrient intake reflected food groups' intake. Pattern I presented significantly higher amounts of fiber, protein, cholesterol, iron, zinc, folate, magnesium, potassium, and vitamins A and C. Pattern II had a significantly higher intake of calories, trans fatty acids, and sodium. Table 3 shows the differences in energy, macro and micronutrients between dietary patterns. Table 4 shows the consumption of the total sample and of the individuals categorized in the two dietary patterns, which take under consideration nutritional recommendations of the guidelines for CHD and prevention of atherosclerosis 2,11,19 . Carbohydrate intake was within the recommended percentage range of total energy intake (TEI), however, fiber intake of the subjects in both dietary patterns was below the minimum recommended amount of 25 g/day. Among all subjects, only 18.7% (23 individuals) reached the daily recommended intake of fiber.
Considering total lipid intake, the amount consumed was close to the maximum value recommended of 35% of TEI. Regarding their subtypes, the intake of monounsaturated fatty acids was below the minimum value recommended of 15% of TEI. Only 13% (16 individuals) reached this recommendation. Consumption of polyunsaturated fatty acids was found to be similar for the entire sample and the two patterns identified, with its consumption being within the recommended range. Nevertheless, saturated fatty acids had an intake greater than 7% of TEI, with only 9 subjects (7.3%) reaching this target.
Data are expressed as median (interquartile range). TEI: total energy intake.  Data are expressed as means±standard deviation or median (interquartile range). *Student's t-test for independent samples; †Data adjusted for energy intake according to the residuals method; ‡Mann-Whitney U test.

DISCUSSION
Cluster analysis was used to identify two dietary patterns in a sample of individuals with stable CHD. Pattern I was characterized by greater consumption of whole carbohydrates, beans/oilseeds, white and red meats, vegetables and fruits. On the other hand, pattern II had a greater intake of refined carbohydrates, fried foods and sweets. These dietary patterns differed significantly in the amount of fiber, protein, trans fatty acids, cholesterol, and most micronutrients. In both patterns, there was a consumption below the recommended target for fiber, monounsaturated fats and vegetables; however, the intake of sodium and saturated fats was above recommendations. Even so, individuals in dietary pattern I had better control of DBP and HbA1.
Literature shows that healthy dietary patterns are related to the reduction of CHD risk. Therefore, these are oriented in the clinical practice of patients with ischemic heart disease 3,22,23 . In a meta-analysis of prospective studies two dietary patterns were identified, the first was called "Healthy" because it contains fruits, vegetables, whole grains, white meats, and low-fat dairy products; which presented an inverse association with CHD risk. The second is denominated "Western" for having the presence of processed/red meats, refined carbohydrates, sweets, and fatty foods 3 . When compared to our results, the characteristics of the "Healthy" pattern resemble dietary pattern I. Similarly, the "Western" pattern has a composition equivalent to dietary pattern II.
There are several studies on the association between dietary pattern and CHD in international literature, however, there are few Brazilian studies in patients with heart disease. A Brazilian population-based study showed that the consumption amount of healthful foods, such as fruits and vegetables was below ideal 10 . Data that are similar to our results. The patients maintained an insufficient consumption of vegetables in the total sample and in both groups of dietary patterns. Fruit consumption reached the recommended value only in the total sample and in pattern I.
Intake of whole carbohydrates was higher in dietary pattern I. Both dietary patterns had insufficient fiber intake, despite the significant difference found. Fiber consumption is important for adequate glycemic control in diabetic individuals 24,25 . In the current study, patients following dietary pattern I presented significantly lower levels of HbA1 than the ones from pattern II, even though for both groups the glycated hemoglobin value was outside the target (<7%). This can be explained by the fact that the first pattern has greater amounts of whole carbohydrates, fruits, and vegetables.
There are consolidated dietary patterns in the literature which are used in cardiovascular prevention research 12 . The Mediterranean diet has olive oil as the main source of fat, therefore it is a diet rich in monounsaturated fat 4 . The consumption of total lipids by our sample of individuals was close to the upper limit of the recommendation. When the quality of the ingested fat was analyzed, the intake of the monounsaturated fatty acids was observed to be below the recommended by the guidelines and the Mediterranean pattern. On the other hand, the intake of saturated fatty acids is above these recommendations.
The results of the recent Prospective Urban Rural Epidemiology (PURE) cross-sectional study showed that  9 . In this study, blood lipids were within recommended targets, except for triglycerides. This is justified in part by the use of statins in the pharmacological treatment of these patients. However, if there was an improvement in lipid intake, with a decrease in the intake of saturated fat from food groups such as red meat, and an increase in fiber-rich foods, such as whole carbohydrates and vegetable, this could improve lipid profile -including triglycerides-in these patients. Another protective dietary pattern is the "Dietary approaches to stop hypertension" (DASH), emphasizing a low content of total lipids, saturated fats, cholesterol and sodium. There is a high consumption of fiber, white meat, potassium, calcium, and magnesium 7 . When the DASH diet is compared to the dietary patterns identified in the present study, a higher consumption of total lipids, saturated fatty acids, cholesterol, and sodium is observed in our sample. There is also a decreased intake of fiber and micronutrients. Another study "Optimal macronutrient intake to prevent heart disease (Omniheart)" showed that regardless of the macronutrient predominant in a diet, if a healthy dietary pattern is followed, such as the DASH diet, improvements in lipid profile can be obtained as well as a decrease in CHD risk 26 .
Regarding the clinical and non-dietary characteristics of the participants, there was a significant difference in age and follow-up time in the outpatient clinic, as well as in smoking and nutritional counseling. Individuals in dietary pattern II, which food composition is less healthy, had a higher prevalence of smoking (22.6%), thus suggesting a lifestyle issue. Similar results were found in a previous study in which a higher proportion of smokers was observed in unhealthy dietary patterns 23 .
In the analysis of the proportion of individuals who reached therapeutic goals according to the dietary pattern, we had an association between pattern I and DBP control. This may be influenced by the clinical situation of the patient, since in SBP there was no significant result. However, subjects in dietary pattern I also had significantly lower DBP values. The presence of healthy food groups, such as fruits and vegetables, higher amounts of fiber and lower sodium values present in this pattern may have helped to better control this risk factor.
In this study, few individuals achieved the therapeutic targets for most of the cardiovascular risk factors. The sample size may have impacted our results. On the other hand, it should also be considered that the recommendations used in clinical practice are either not adequate for this population or that there was low adherence in this sample. In the Brazilian population, a CHD secondary prevention study found a patient profile similar to the one in this study and after a 3-month nutritional intervention there were significant reductions in weight, BMI, AC, and HbA1 27 .
In our study, some methodological precautions were also taken into account. We used a FFQ previously constructed and validated in patients from Southern Brazil, and macro and micronutrient data were adjusted for energy using the residual method 21 . However, limitations in this study should be considered. Although FFQ is a well established method for dietary pattern evaluation, one of its limitations is the fact that it relies on the memory of the individual interviewed, which may result in some errors of actual measurements of food consumption 21,28 . Moreover, the adopted cross-sectional design hinders any causal inferences. Another limitation of our study was that physical activity level was not evaluated, and this is a lifestyle characteristic that may influence the control of cardiovascular risk factors.
To our knowledge, our study was the first to use cluster analysis and to characterize the dietary pattern of Brazilian patients with CHD. Cluster analysis findings are easier to interpret because an individual is in one cluster only, outcomes are specific to individuals within each cluster, and each cluster has a specific food and nutrient composition 29,30 .

CONCLUSIONS
In conclusion, two dietary patterns were identified. Both differed significantly in food groups and nutrients, pattern I presented a healthier nutritional composition than pattern II, but still needs adjustments according to guidelines of the Societies of Cardiology. Diabetic patients with dietary pattern I had a higher consumption of fiber-rich foods and lower value of HbA1c compared with those on dietary pattern II. In addition, pattern I individuals also had lower values and better control of DBP. Finally, patients who had previously had nutritional counseling, with a nutritionist, were more likely to be in pattern I, showing the importance of a multi-professional team in the non-pharmacological treatment of CHD.