Need for the Study

Evidence from developed countries indicates significant fall in stroke burden attributable to secondary prevention strategies and prevention of risk factors (such as smoking) in the general population. Rural Andhra Pradesh data diverges from WHO Premise study which reveals a large urban to rural gap in knowledge of cardiovascular risk factors and use of medication for stroke prevention with less than 25% rural subjects using prescribed medication.⁹ The erstwhile combined state of Andhra Pradesh is now re-carved into states of Telangana and Andhra Pradesh. The proposed study will ascertain the risk factors for stroke in rural Telangana population and provide a guide to health care providers in adopting treatment and prevention strategies.

Study Design

The study design applied is community-based unmatched case-referent study.

Study Setting and Time Frame

The population of Moinabad mandal of Ranga Reddy District, Telangana, India was studied, between July 2016 and December 2017. Moinabad mandal, has a population of approximately 56,000, residing in nearly 13,000 households from 26 villages in the subdistrict (mandal). Among them, approximately 29,000 (52%) are male and approximately 27,000 (48%) are female. Seventy-five percent of the whole population are from general caste, 25% are from schedule caste, and none from schedule tribes.

Recruitment and Enrolment

The patients were recruited by line listing from the village of Moinabad through a door to door survey. Cases were live stroke patients. Controls were live subjects without history of stroke who are from the same neighborhood and from similar socioeconomic status as cases. Controls were not taken from the same family as the cases. For every case, one control was enrolled.

Data Collection

All the houses of Moinabad were approached by a door to door survey. Families with a history of stroke and a live stroke patient were considered for enrolment. The subject was examined by a Medical Officer in the field, who has confirmed the diagnosis clinically. Data collected included the variables—demographic details including age, gender, and educational qualification, history of stroke (previous/family), hospital visits, current medical treatment, medical reports, computed tomography scan, or magnetic resonance imaging reports where available. Physical examination included blood pressure (BP) check, blood sample collection for lipid profile evaluation, and a random blood sugar.

Exclusion Criteria

Patients who are not willing to consent, patients who are not available at the time of study, patients who have expired following stroke, patients with history suggestive of transient ischemic attack were excluded for the study.

Kelsey’s sample size method for unmatched case–control study was used.10 A two-sided confidence level (1α) of 95, power (% chance of detecting) of 80, ratio of controls to cases of 1, least extreme odds ratio (OR) to be detected, 1.99, based on OR for hypertension in stroke and proportion of controls with exposure of 40 in previous studies were used as input parameters.6 11 The minimum sample size was calculated to be 136 cases and 136 controls with a total sample size of 272. Proportion of cases with exposure was derived to be 57.02. Finally, 144 cases and 144 controls participated.

Data Analysis

Data are analyzed at two levels. At primary level, prevalence and association of various risk factors were assed using Chi-square test. To derive age and gender adjusted odds ratios (AORs) of various risk factors, binary logistic regression analysis was performed with age and gender being controlled.

Operational Definitions

We defined stoke to be “a clinical syndrome characterized by rapidly developing clinical symptoms and/or signs, and at times global, loss of cerebral function, with symptoms lasting more than 24 hours or leading to death, with no apparent cause other than a vascular one.”12 13 14 Those who gave a history of hypertension or who had an elevated blood pressure of either systolic above or equal to 140 mm of Hg and or a diastolic above or equal to 90 mm of Hg as a mean of three reading were termed as hypertensives.15 Diabetics were those who had a history of diabetes mellitus or a GRBS (glucometer random blood sugar) ≥200 mg/dL.16 17 Generalized obesity was termed as a body mass index (BMI) above or equal to 25.18

WHO STEP-wise Approach to Stroke Surveillance Manual19 was used as reference, in design and conduct of the study including anthropometry and laboratory procedures.

Demographic Characteristics of the Study Population

Out of the 144 controls, 59 were women and 85 were men with a mean age of 45.69 and 47.40, respectively. Out of 144 cases, 44 were women and 100 were men with a mean age of 60.77 and 62.49.

Estimated crude prevalence of stroke in Moinabad mandal is found to be 257(220–300) per 1, 00,000 population. Prevalence of stroke among men is 344(280–410) per lakh men and among women it is 163(110–210) per lakh women.

A higher prevalence of stroke was seen in higher age groups, men, smokers, alcohol consumers, hypertensives, and diabetics (Table 1). This association was tested for statistical significance by Chi-square which has revealed that age and hypertension were significantly associated with stroke patients.

Binary Logistic Regression Model for Deriving AOR of Various Risk Factors

Comparison of Means of Risk Factors of Cases and Controls

The mean systolic (138.73 ± 19.59 mm Hg vs. 127.43 ± 13.05 mm of Hg, p = 0.000) and diastolic (88.26 ± 13.81 mm of Hg vs. 83.75 mm of Hg, p = 0.002) blood pressures were significantly higher in stoke cases compared with controls. Similarly, the means of, “mean arterial pressures,” abdominal girth (inches) were significantly higher in stroke cases than controls. The means of GRBS and total cholesterol were also higher in cases than controls; however, they are not statistically significant. The mean BMI among stroke patients was a unit less compared with controls (Table 3).

Risk Factors for Stroke

Age and gender trends were similar to earlier stroke studies23 with greater prevalence of stroke, hypertension, and tobacco use in older subjects5 6; and greater number of males with exposure to tobacco and alcohol consumption.

Hypertension was the most frequent risk factor in our study with an OR of 2.234, consistent with data from other rural and urban studies in India,11 24 the United States,25 as well as the international multicenter INTERSTROKE study.26 Reports from AIIMS, New Delhi and SCTIMST, Trivandrum also reveal hypertension as the leading risk factor in “young stroke” (stroke < 45 years).27 28

An estimated 54% of all stroke-related deaths in low income countries2 have been linked to hypertension20 as have 40% of all stroke-related deaths in Gadchiroli district. The Earth Institute report in 2004 stated that cardiovascular disease-related death rates across all age groups and both genders are much higher in India than in Portugal and the United States.29 These observations are indicative of both urban and rural India going through phase 2 of the epidemiological transition.

The prevalence of diabetes in our study is similar to what has been observed in rural Kerala.6 With an OR of 1.6, diabetes has a weaker stroke risk association than hypertension, smoking, and alcohol consumption.

Alcohol consumption was the major stroke risk factor in our population (OR 1.997) and was also linked to 24.6% of male stroke deaths in Gadchiroli.7 Many Indian studies however did not document such a relationship.6 30 In rural Karnataka, although alcohol consumption was prevalent in 30% of the population, it did not significantly correlate with direct cardiovascular risk factors for MI or stroke outcome.8Binge drinking increased stroke risk by an OR of 1.5 in the INTERSTROKE study conducted across 22 countries including India.26

Current or past history of smoking increased stroke risk by the odds of 1.81; high-risk association of smoking and stroke is well established by many Indian urban and rural studies with very high ORs ranging from 3.92 in WB31 to 7.8 in Kerala.6 Tobacco inhalation was a direct cardiovascular risk factor in the rural Karnataka population.8 Nearly 37% of our subjects either smoked or chewed tobacco, a number similar to the Trivandrum Rural Stroke Registry; whereas there were significantly more documented smokers (59%) in a hospital-based case-referent study.28 Interestingly, young strokes in the AIIMS study had low exposure to smoking.27 Whether such differences in behavioral risk factors influence stroke incidence and subtypes is not known.

Obesity as measured by waist circumference rather than BMI was more prevalent in our stroke subjects compared with controls with an OR of 1.033 (female) and 1.33 (male) stroke risks. Abdominal obesity rather than BMI seemed to be a better indicator of stroke risk in other Indian and international stroke studies.32 33 34 Two large population-based cardiovascular risk factor studies in rural AP9 and Karnataka8 utilized high BMI as an indicator of obesity and a direct cardiovascular risk factor for stroke or MI.

Elevated serum, total cholesterol, and a low HDL were significant stroke risk factors in this study with ORs of 2.278 and 1.658, respectively. One-third of our stroke subjects had dyslipidemia as did 26% of those in the Trivandrum Registry.6 Urban and rural differences in risk factor profiles are disappearing in epidemiological transition. It is note-worthy that unlike well-established links between dyslipidemia (lipid factors) and cardiac ischemic events,35 the impact of dyslipidemia on stroke in the Indian context requires further study.