Coronary heart disease (CHD), with myocardial infarction (MI) as a key clinical manifestation, is the leading cause of death in Malaysia and elsewhere in Southeast Asia [1,2]. In recent decades, Malaysia has experienced a sustained increase in the incidence of CHD [3]. According to World Health Organization estimates, the age-adjusted death rates of CHD are approximately twice as high in Malaysia compared with the United States or the United Kingdom [4]. Furthermore, studies in Malaysia and other Southeast Asian countries indicate that CHD events tend to occur at younger ages and are characterized by more severe clinical features than in Western populations [5,6], resulting in substantial losses in productive working years owing to death and disability. Indeed, CHD in Malaysia is characterized by high rates of short-term post-MI mortality [7]. However, there is limited direct evidence regarding the determinants of CHD in Malaysia [2]. Therefore, we have established the Malaysian Acute Vascular Events Risk (MAVERIK) retrospective case-control study of MI to investigate genomic, lipid-related, and other determinants of acute MI in Malaysia. In this paper, we describe the rationale, protocol, and early baseline results of this study.

The MAVERIK study was established in 2017 in collaboration with the British Heart Foundation Cardiovascular Epidemiology Unit at the University of Cambridge, United Kingdom (the study’s international coordinating center); the Institute for Medical Research (IMR), Malaysia (the study’s National Collaborating center); and 17 national collaborating cardiology referral hospitals in Malaysia serving as the study’s recruitment centers (encompassing nearly all states and federal territories in Malaysia; Figure 1).

Complete information on age, sex, and ethnicity was available for 2547 MI cases and 2500 controls (Table 2). Missingness for other variables ranged between 2.2% (history of high blood pressure in controls) and 11.6% (monthly income level in cases). Complete case analysis was used (ie, participants with missing values were excluded from the analysis). The mean (SD) age of MI cases was 50.3 (9.4) years, with 91.2% (2324/2547) of participants being male. In total, 60.7% (1548/2547) of cases self-identified as being of Malay ancestry, 19.1% (487/2547) were of Indian ancestry, 13.5% (346/2547) were of Chinese ancestry, and 5.8% (150/2547) as Sabah and Sarawak Bumiputera (including Iban, Kadazan, Dusun, Bidayuh, Melanau, Bumiputera of Sabah, and other Bumiputera of Sarawak ethnicities). As expected, the age, sex, and ethnicity of the cases and controls were similar owing to the frequency matching (standardized mean difference [SMD] age=0.1443 years; SMD female sex=0.0008; SMD Malay ethnicity=0.0001, and SMD Chinese ethnicity=0.0009).

Approximately 61.55% (1487/2416) of MI cases were current smokers, compared with only 33.51% (817/2438) of controls (P<.001; Table 2). Cigarette consumption in smokers was higher among cases than in controls (median, IQR daily consumption for cases vs controls 20, 10-20 vs 10, 5-20) cigarettes per day; P<.001). However, the prevalence of tobacco consumption by chewing and vaping was low and did not materially differ among cases and controls. Other risk factors that had a higher prevalence in MI cases than controls included high blood pressure (872/2444, 35.68% vs 500/2446, 20.44%; P<.001; Table 2), diabetes mellitus (691/2433, 28.4% vs 308/2446, 12.59%; P<.001), and family history of CHD (70/2376, 11.36% vs 220/2440, 9.02%; P=.007). MI cases also had higher median levels of waist-to-hip ratio than controls (0.96 vs. 0.94; P<.001), but median levels of BMI were not materially different.

Indicators of lower socioeconomic status were more common among MI cases than controls, including higher percentages with a monthly income below MYR1500 (US $357; 846/2251, 37.58% cases vs 700/2315, 30.24% controls; P<.001), and educational attainment to primary school level or lower (421/2284, 18.43% vs 268/2306, 11.62%; P<.001). The prevalence of participants reporting being the offspring of a first cousin marriage (68/2376, 2.86% vs 66/2430, 2.72%) or having a first cousin as a spouse (41/2330, 1.76% vs 34/2385, 1.43%) was not materially higher in cases than in controls. Additional details about the baseline distribution of characteristics, subdivided by age group and sex, are outlined in Multimedia Appendix 1, Tables S2 and S3.

Crude and adjusted (age and sex) logistic regression models for traditional risk factors are summarized in Table 3. Current smoking (OR 4.11, 95% CI 3.56-4.75; P<.001), previous smoking (OR 1.34, 95% CI 1.12-1.60; P=.001), history of high blood pressure (OR 2.13, 95% CI 1.86-2.44; P<.001), history of diabetes mellitus (OR 2.72, 95% CI 2.34-3.17; P<.001), family history of CHD (OR 1.28, 95% CI 1.07-1.55; P=.009), and obesity (BMI >30 kg/m²; OR 1.19, 95% CI 1.05-1.34; P<.01) were associated with MI in age- and sex-adjusted models.

Although Malaysia is experiencing a substantial increase in the burden of CVD, the risk factors for CHD in this population have been relatively little studied. We established the MAVERIK study, an epidemiological bioresource comprising over 2500 confirmed cases of acute MI and 2500 controls, and confirmed the importance of several known risk factors for MI (eg, tobacco consumption, history of diabetes, and hypertension). To our knowledge, this represents the largest case-control study of MI and related traits in Malaysia. Our long-term objective is to enable direct investigation of genetic and other determinants of MI in this multi-ethnic population. In this paper, we report the design and initial results of this study, which suggest several conclusions.

First, our study has demonstrated the ability to use modern and standardized epidemiological methods in a national study in Malaysia encompassing 17 hospitals. For example, the study used electronic data collection, including apps with intelligent study forms, and collected several types of biological samples, including serum, plasma, DNA, and nails.

Second, the mean age at first confirmed MI was only 50 years in Malaysia, suggesting that premature CHD is highly prevalent in this population.

Third, this study’s multi-ethnic population broadly mirrors the overall ethnic composition of the Malaysian national population, which is approximately 61.8% Malay, 21.4% Chinese, 6.4% Indian, and 0.9% others and 9.6% non-Malaysian resident [20]. Malaysia’s complex genetic ancestry is distinct from the commonly studied East Asian populations (eg, Chinese and Japanese) and other Southeast Asian populations (eg, Thai and Filipino) [21]. In contrast, most existing genetic studies of MI and cardiometabolic risk factors involve European or East Asian (eg, Chinese and Japanese) descent populations. The distinctive genetic architecture of the Malaysian population suggests that, with the use of genome sequencing technologies, the MAVERIK study should help identify population-specific genetic risk factors, leveraging transethnic differences. For example, variants in the myosin binding protein C3 (MYBPC3) gene region, which are common (2%-8%) in South Asian ancestry populations but rare (<0.1%) in European ancestry populations, are associated with an increased risk of hypertrophic cardiomyopathies [22].

Fourth, an analysis of initial baseline data supports the validity of the MI outcomes recorded, suggested by the observation of expected associations of MI risk with a panel of conventional risk factors, including high blood pressure, diabetes mellitus, family history of coronary disease, and higher waist-to-hip ratio (a marker of visceral adiposity). Finally, the study’s initial data suggest that in Malaysia, a middle-income country with a population of 32 million people in the midst of economic and epidemiological transition, there are inverse associations of indicators of socioeconomic status (such as income and educational attainment) with MI risk, a pattern similar to that observed in many high-income populations in Western countries.

The strengths and potential limitations of the MAVERIK study merit consideration. Retrospective case-control studies of MI can usefully complement prospective studies because the former involves the ascertainment of exposure information and the blood sampling of people who have already developed MI and a comparable group of controls without MI, enabling the rapid and cost-effective accrual of large numbers of relevant cases. By contrast, prospective studies require many tens of thousands of people to be screened (with all of their blood samples kept in frozen storage) and followed for several years to accrue a sizable number of MI cases. Retrospective studies are also often able to include large numbers of individuals who have developed the disease at younger ages, when associations with risk factors are often stronger, providing particularly sensitive tests of certain hypotheses. Furthermore, owing to the richness of data collected, the MAVERIK study allows for the use of a range of analytical techniques, including machine learning approaches [23,24].

As demonstrated by the Wellcome Trust Case-Control Consortium [11], Myocardial Infarction Genetics Consortium [25], Pakistan Risk of Myocardial Infarction Study [26], and Bangladesh Risk of Acute Cardiovascular Events study [13], case-control studies can powerfully and efficiently facilitate genetic discovery and can use genomic data to quantify and robustly correct for any population structure. Furthermore, the International Study of Infarct Survival (ISIS) (United Kingdom–based) [27] and INTERHEART (conducted across 52 countries) studies [28] have demonstrated that appropriately conducted, large case-control studies can usefully address some nongenetic hypotheses in MI, such as tobacco and alcohol consumption, and serological evidence of infection. However, INTERHEART did not focus extensively on Malaysia; INTERHEART-Malaysia involved approximately only 100 MI cases [29].

Nevertheless, particularly regarding nongenetic hypotheses, retrospective case-control studies may be liable to potential biases, such as selection and recall biases. To minimize such biases, standardized questioning approaches were used for all study participants, and both cases and controls were selected from the same population. Furthermore, there may be less scope for recall bias in a study of acute MI than in a study of chronic stable CHD, as hours rather than months or years may have elapsed since the index event. As noted earlier, the selection of controls in case-control studies invariably involves trade-offs between scientific rigor and feasibility. Following the examples of INTERHEART [28], Pakistan Risk of Myocardial Infarction Study [26], and Bangladesh Risk of Acute Cardiovascular Events study [13], we chose to recruit controls drawn from attendants of people visiting outpatient clinics or (nonblood-related) attendants of cardiac patients. Studies of plasma components may be affected by case-control studies of acute MI because certain circulating markers may be altered by the sampling of blood within 24 hours of MI symptoms. If, as in MAVERIK, the time since the onset of symptoms has been recorded, any material bias can be quantified and at least partially corrected for. Similar considerations apply to fasting status and the recording of the time since the last meal. We seek to evaluate the findings of the MAVERIK study with reference to prospective cohort studies in Malaysia (eg, the Malaysian Cohort project) [30]. Finally, it is important to recognize that we were unable to screen controls for silent MI and that the external generalizability of the MAVERIK study may be limited as the study comprises 91.18% of male participants.

In conclusion, the MAVERIK study is a large, multi-ethnic epidemiological resource for CHD that should help identify and evaluate genetic and other determinants of MI in Malaysia. It should help to hasten the discovery of disease-causing pathways and inform regionally appropriate strategies that optimize public health action.