A protocol to assess the risk of dementia among patients with coronary artery diseases using CAIDE score

Study design

This cross-sectional study was planned from August 2019 to July 2020. However, it was delayed due to coronavirus 2019. Figure 1 shows the study procedure that started from the literature review through to the end of dissemination. Different study objectives are described below.

Figure 1. Timeline of the study.

Here month and duration-wise description of overall study procedure is illustrated.

Study objectives

Primary objective 1: Dementia risk prediction— to assess the level of dementia risk of the study subjects as per CAIDE risk score and categorize as low, moderate and high.

Primary objective 2: Burden related to dementia

• Assess the burden of sociodemographic risk factors (age, sex, marital status, education level, occupation, economic background) among the study subjects

• Evaluate the distribution of behavioral risk factors (tobacco use in the form of smoke and oral consumption, inadequate fruits and vegetable intake, physical inactivity) among the study subjects

• Assess the personal and family history of chronic diseases (hypertension, diabetes, CAD, stroke, chronic kidney disease, dementia or Alzheimer's disease) among the study subjects

• Determine the burden of anthropometric risk factors (overweight and obesity) among the CAD patients

• Assess how these risk factors distribute in different risk categories of CAIDE score

• Assess the biochemical and other investigation profile (Echocardiogram, angiogram, electrocardiogram etc. also included) of the study population

Secondary objective: Identify the influencing factors

• Evaluate the association of cardiac biomarkers (Brain Natriuretic Peptide (BNP) and N-terminal prob-type Natriuretic Peptide (NT-pro-BNP)) with the risk of dementia development

• Identify the relationship between the number of coronary vessels involved (Single vessel disease (SVD), double vessel disease (DVD) and triple vessel diseases (TVD)) and the level of dementia risk of the study subjects

Study setting

The study was conducted in Ibrahim Cardiac Hospital and Research Institute (ICHRI), a tertiary level cardiac hospital in Dhaka, Bangladesh. It is one of the eight affiliated institutes of the Diabetic Association of Bangladesh (BADAS) that has the objective to improve both preventive and curative cardiac care with quality services at an affordable cost. At the initial stage, it was conceived as an extension of the cardiac outdoor facility of another affiliated hospital, Bangladesh Institute of Research and Rehabilitation in Diabetes, Endocrine and Metabolic Disorders (BIRDEM), which was the World Health Organization Collaborating Centre on Diabetes, Endocrine and Metabolic Disorders, the only one of its kind in Asia. However, thereafter it was agreed that the project would best achieve its objectives from a stand-alone specialized cardiac hospital supported by the multidisciplinary hospital of BIRDEM for treatment of cardiac patients concurrently suffering from other medical conditions. It is a 150-bedded specialized cardiac hospital that has a cardiac emergency, inpatient, and outpatient services.

Study population

Data collection of this study involved patients with stable CAD who were admitted to ICHRI for revascularization at the inpatient department. A total of 285 patients (age range 40 to 59 years) were selected using a convenient sampling technique due to economic and logistical issues. As this was Master’s thesis that has a pre-defined period, this non-probability sampling method will enable us to achieve the desired sample size in a relatively fast and cost-effective way. All the registered and eligible patients under the cardiac surgery department were invited to take part in the study. For this, the researchers counseled both the patient and their caregivers about the potential benefit of screening the future risk of dementia and the importance of their participation. The researchers also informed that the participation is voluntary and they have the right to withdraw themselves or refuse to answer any question. Before enrollment, patients were screened as having “no cognitive impairment” based on the Bangla adaptation of mini-mental state examination (BAMMSE), a self-reported statement (If they have been diagnosed by any neurologist as having signs and symptomps of cognitive impairment or dementia), and documented clinical history (If any document or prescription from authorized physician stated him/ her as having cognitive impairment or dementia and it is written at the hospital history sheet). The sample size was determined using the prevalence of vascular dementia among the Indian population (11.4%) extracted from published literature¹⁶, with a precision of 4% at a 95% confidence interval. The final sample size was adjusted considering 15% non-response rate. Thus, a total of 285 patients were selected for the study. We consider this prevalence for the following reasons-(a) there is a lack of dementia prevalence study at Bangladesh (b) available prevalence study is not representative to such group of population (c) our study population is more exposed to develop vascular dementia as they already have such risk factors (d.)We selected the urban area of our neighboring country as the food habit, lifestyles are almost similar. Exclusion criteria included patients with a history of cognitive impairment, congenital heart diseases, clinically unstable, pregnant, and refusal to participate in the study.

Questionnaire development

A semi-structured questionnaire was being used to collect data¹⁷. The questionnaire was prepared based on the study objectives and the conceptual framework (Figure 2). The questionnaire had five components: (1) sociodemographic information (age, gender, marital status, educational status, occupation, monthly household income); (2) behavioral information (tobacco use-smoking and oral consumption, dietary servings of fruits and vegetables and work-related physical activity); (3) personal history of chronic diseases and medication (hypertension, CAD, diabetes, chronic kidney disease); (4) family history of chronic diseases (Alzheimer's disease, hypertension, CAD, stroke, chronic kidney diseases); (5) measurement (anthropometric - height, weight and body mass index [BMI]; blood pressure - systolic and diastolic; and investigation parameters - blood glucose, creatinine, troponin-I, HbA1C, lipid profile, angiogram, echocardiogram, and electrocardiogram). The behavioral part comprised of relevant questions adapted from the STEPwise approach to Surveillance (STEPs) of Noncommunicable diseases risk factors questionnaire (version 3.2) of the World Health Organization (WHO) with appropriate modification¹⁸. The English questionnaire was translated into the Bengali language to maintain cultural sensitivity and understandability by the participants. Then the translated questionnaire was pre-tested to detect any inconsistency, unclear wording, or unusually long time taken to administer. To pre-test, ten men and ten women patients with stable CAD within the pre-selected age group were recruited randomly and interviewed by the trained data collectors (one-on-one, in-person, and face-to-face) in the cardiothoracic department of ICHRI. Before this, the objective and importance of pre-testing were explained to the participants and the invitation followed the similar approach mentioned in the previous section. To maintain consistency among the interviewers and the pretesting sessions, we used a pretest answer sheet to note verbal and nonverbal responses to the translated questionnaire. The collected responses were analyzed and interpreted based on the following parameters: trends in responses; fundamental flaws with the design or format (whether the data quality match with the study design or not?); attractiveness (do the questionnaire materials attract the audience?) ; comprehension (are the materials properly understood by the target participants?); acceptance (is there anything in the questionnaire which is not sensitive or unfitting?); and relevance (are the facts ever faced by the participants we are talking about).

Figure 2. Conceptual framework of the study.

In this framework, interaction of different factors on the risk of dementia is illustrated. Here, risk of dementia is outcome variable and the factors are independent variables. CAD, cardiovascular disease; S. creatinine, serum creatinine; LVEF, left ventricular ejection fraction; HbA1c, haemoglobin A1C; FBG, fasting blood glucose; RBG, random blood glucose.

Data collection procedures

The data were collected by the research student with the help of a male and female assistant in the inpatient department of ICHRI. Data collection had three steps covering the five components of the questionnaire. In the first step, a face-to-face interview was organized to gather information on sociodemographic background, behavioral risk factors, and personal and family history of chronic diseases. In the second step, physical measurements were conducted with adequate privacy. Anthropometric measurements (height and weight) were carried out following a standard method described in “Noncommunicable disease risk factors survey Bangladesh 2010”¹⁹ and values were recorded in the checklist. Generalized obesity is determined by BMI according to the international guidelines of BMI²⁰ and calculated as weight in kg/height in m². Blood pressure was measured using an aneroid sphygmomanometer on the right arm in a sitting position with their hand in resting on the handle of the chair or some objects. Following resting for at least 15 minutes the first reading is taken, followed by a second reading after three minutes of resting interval. Systolic and diastolic measurements were taken in mmHg. The mean of the two measurements was used to determine the final value of blood pressure. In the third step, all the investigation (biochemical, angiogram, echocardiogram, and electrocardiogram) data were collected from the patient’s record file. The coronary angiogram report was used to determine the number of vessels involved in the disease process. The echocardiogram report was used to assess the left ventricular ejection fraction.

Dementia risk prediction using CAIDE score

CAIDE risk score, a mobile application-based risk prediction tool, was used to assess the long term risk (20 years later) of dementia using all the collected data²¹. The Dementia Risk Tool uses the CAIDE dementia risk score model to predict the risk of developing dementia in the future among middle-aged adults²¹ and among older adults. This is the first evidence-based tool for dementia risk prediction. This study used the Dementia Risk Tool mobile application, which uses the CAIDE Dementia Risk Score to assess The tool uses the risk score model to predict the risk of developing dementia in the future among middle-aged and elderly people²¹. It was developed by the European Institute of Innovation and Technology (EIT-Health) within the project Multimodal strategies to promote a healthy brain in aging: Innovative evidence-based tools (MULTI-MODE). The project partners were Karolinska Institute (Sweden), Research Institutes of Sweden (Sweden), Imperial College London (UK), “Parc Sanitari Sant Joan de Déu (Spain), and Erasmus University Medical Centre Rotterdam (the Netherlands). This risk score has been externally validated to predict mid-life dementia risk²² and was incorporated into the inclusion criteria of the Finnish Geriatric Intervention Study to Prevent Cognitive Impairment and Disability (FINGER) 8220;It was also shown that the use of the CAIDE dementia risk score through a mobile application can motivate participants to reduce their risk factors²³.

The application includes a series of questions, concerning their lifestyle and medical condition, and then calculates a dementia risk score based on the answers. There are a total of 15 points against all of the variables and there is the flexibility of skipping cholesterol level and blood pressure measurements wherever appropriate. The tool then generates the percentage of risk and categorizes risk as low, moderateand high.

The CAIDE risk score based mobile application tool has never previously been applied within the Bangladeshi population and hence we will pre-test it in a sample of CAD patients before the final application. The variables of the tool were incorporated in the semi-structured questionnaire and pre-tested among the patients described earlier in the ‘Questionnaire development’ sub-section. Tool-specific informations were inputted and analyzed to detect any inconsistency. The research student entered the necessary data using an android (version 7.0) mobile phone where the application (CAIDE risk tool) was downloaded from the Google play store.

Definitions of key variables

Stable CAD patients— defined as admitted patients with myocardial infarction or ischemic cardiomyopathy who are hemodynamically stable, free from acute symptoms, and ready for prescribed revascularization.

Cognitive impairment— defined according to BAMMSE in which any score < 24 points (out of 30) indicates cognitive impairment.

Current tobacco user— those who have smoked or consumed tobacco orally in the past 30 days are considered to be a 'current' user¹⁹.

Inadequate fruit and vegetable intake— consumption of less than five servings of fruits and/or vegetables in a day is considered inadequate¹⁹.

Low physical activity—we define it as per STEPS protocol and for that, we convert all the work-related physical activities in minutes per day (metabolic equivalent of task or MET-minute) as follows:

• One minute in sedentary position (sitting quietly) = 1 MET-minute

• One minute in moderate activities = 4 MET-minutes

• One minute in vigorous activities = 8 MET-minutes

Then all the MET-minutes will be added together to get the cumulative physical activity in MET-minutes. Based on the cumulative MET-minutes participants will be categorized as low, moderate, and high as follows: ≥3000 MET-minutes per week = highly active; 600–3000 MET-minutes per week = moderately active. The respondents who do not fall within the moderately active group will be categorized in the low activity group¹⁹.

Overweight and obesity—we classify overweight and obesity as BMI 25.0 - 29.9 kg/m² and BMI ≥ 30.0 kg/m², respectively²⁰.

Hypertension—is diagnosed based on “Seventh Report of Joint National Committee on Prevention, Detection, Evaluation, and Treatment of High Blood Pressure (JNC7)” criteria when systolic blood pressure ≥140 mmHg and/or diastolic blood pressure ≥90 mmHg or known case of hypertension or on antihypertensive drugs²⁴.

Diabetes mellitus—is will be diagnosed as fasting plasma glucose ≥7.0 mmol/l (126 mg/dl) or 2–h plasma glucose ≥11.1 mmol/L (200 mg/dl) or a known case of diabetes or on anti-diabetic drugs²⁵.

Quality assurance

To maintain quality, a quality control panel is formed by the researchers who are acting in different levels of study. The panel is comprised of a research student who is the principal investigator, three senior public health specialists with expertise in prevention and control of chronic diseases, and a cardiothoracic surgeon. The cardiothoracic surgeon and the student together assessed and recruited stable CAD patients as per inclusion and exclusion criteria. The questionnaire was filled, checked, and entered into the database by the student herself. The groups of three public health specialists are still monitoring the overall patient recruitment and data collection procedure. Besides them, an external monitoring panel of two researchers from other institutions was recruited to evaluate the overall procedures and comment on the performance. To prevent bias, the monitoring panel is being anonymized for the data collectors. The overall performance at the end of the study was categorized as satisfactory based on three parameters: (1) mode of measurement; (2) maintenance of data collection environment; and (3) mode of administration of the questionnaire. In addition, to assure the quality of the study, we also maintained and followed specific protocol: (1) pre-testing of the questionnaire and the tool; (2) use of the standard method of measurement as per the STEPS survey of Bangladesh 2010; (3) use of show cards for a better understanding of dietary servings and intensity of physical activities; (4) maintenance of adequate privacy during physical measurements; (5) ensuring the use of robust equipment for physical and biochemical measurements.

Planned analysis

Data will be analyzed using the Statistical Package for Social Science (SPSS) version 20.0 for Windows (SPSS, Inc. Chicago. IL. USA). All estimates of precision will be presented at a 95% confidence interval (CI) in the tables. In this study, the p-value (two-sided) will be considered statistically significant at the threshold of p<0.05. Analyses are planned based on study objectives.

Descriptive statistics. To address the primary objective 1 and primary objective 2, a descriptive analysis will be conducted using mean, standard deviation (SD), median with interquartile range (IQR), frequencies and percentages where appropriate.

Exploratory data analysis. To address the secondary objectives, an in-depth exploratory analysis will be applied. For normally distributed variables, the correlation between biochemical parameters and dementia risk percentages will be assessed using the Pearson correlation coefficient or the Spearman correlation otherwise. For continuous, normally distributed variables, we will use a t-test while comparing two variables and one-way analysis of variance (ANOVA) while comparing more than two variables. For continuous variables not normally distributed, we will use non-parametric tests: the Mann-Whitney U test to compare two groups, and the Kruskal-Wallis test to compare three or more groups.

To identify the relationship between the numbers of coronary vessels involved (SVD, DVD and TVD) and the risk of dementia (Low, moderate and high) among the study subjects, we will employ the χ2 test or Fisher’s exact test, depending on the number of observations obtained in each considered category. Similar statistical analysis will also be applied initially to assess the association of different risk factors with the risk of developing dementia among CAD patients. Moreover, we will identify the influencing factors that are associated with increasing the long-term risk of dementia development using multivariate analysis. For this purpose, univariate analysis will be used to select variables according to the p-value obtained (p≤0.25) to include in the multivariate analyses^26,27. For the multivariate analyses, we will run multinomial logistic regression considering “dementia risk categories” as the dependent variable and the “low risk” as reference. We will also calculate odds ratios (OR), and 95% CI for each independent variable. In the regression table, predictors with OR >1 will be presented. To ensure the presence of no multicollinearity, we will use the variance inflation factor (VIF) before the regression analysis. The STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) guidelines will be used to report the results of our study.

Ethics and dissemination

The purpose of the study and data safety issues was explained to the participants. All the participants provided their informed written consent for inclusion in the study¹⁷. The study was conducted following the Declaration of Helsinki, and the protocol has been approved by the Ethical Review Committee of Bangladesh University of Health Sciences (identification number: BUHS/BIO/EA/19/208). The permission of data collection has been provided by the appropriate authority of ICHRI. No personally identifiable information was collected to maintain the anonymity of the participants. Patients have not been involved in the setting of the research question or outcome measures. They had no role in designing or implementing this work or interpretation of the results.

This study will bring no direct and immediate benefits to the patients. Indirect benefits include a contribution to understanding the underlying risk of dementia among patients with CAD. This will eventually increase the awareness around the importance of early lifestyle modification and the integration of mental health in the cardiac rehabilitation program. No important risk is foreseen as a direct result of the study.

We will apply a comprehensive strategy to disseminate the findings of the study. The strategy will use various communication platforms to reach out to a diverse range of beneficiaries: CAD patients, caregivers, cardiologists, psychiatrists, researchers, journalists, etc. We will use academic media (i.e., peer-reviewed journal articles, national and international conference presentations), social media (i.e., Facebook), online news portal, other platforms in the internet (i.e., links to study reports on the university website) and electronic mail (i.e., posting of study findings to participants and stakeholders) to disseminate the outcomes from this study. The study data will be made available publicly through an online repository using CC0 license.

Study status

It remains ongoing (permanently closed to additional enrollment but subjects continue to undergo research-related activities like checking for consistency, completeness, coding, analysis, report preparation).

Underlying data

No underlying data are associated with this article.

Extended data

Zenodo: A protocol to assess the risk of dementia among patients with coronary artery diseases using CAIDE score. http://doi.org/10.5281/zenodo.4033006¹⁷.

This project contains the following extended data within the file ‘Extended data file.pdf’:

Data are available under the terms of the Creative Commons Zero "No rights reserved" data waiver (CC0 1.0 Public domain dedication).