Prevalence and Associated Factors of Insulin Resistance in Adults from Maracaibo City, Venezuela

Background and Aim. Insulin resistance (IR) is a prominent pathophysiologic component in a myriad of metabolic disorders, including obesity, prediabetes, and type 2 diabetes mellitus, which are common in our locality. The objective of this study was to determine the prevalence of IR and factors associated with this condition in an adult population from Maracaibo city, Venezuela. Methodology. A cross-sectional, descriptive study with multistaged randomized sampling was carried out in 2026 adults. IR was defined as HOMA2-IR ≥ 2. A multiple logistic regression model was constructed in order to evaluate factors associated with IR. Results. The prevalence of IR was 46.5% (n = 943), with 46.7% (n = 450) in the general population, 46.4% (n = 493) in females, and 47.90% (n = 970) in males (p = 0.895). IR prevalence tended to increase with age and was significantly greater in subjects aged ≥30 years (χ 2 = 16.726; p = 2.33 × 10−4). Employment, alcohol consumption, obesity, high triacylglycerides, low HDL-C, and dysglycemia were associated with greater odds of IR, whereas a high level of physical activity appeared to be weak protective factor against IR. Conclusions. The prevalence of IR is elevated in our locality. The main determinants of this condition appear to be the presence of obesity, high triacylglycerides, low HDL-C, dysglycemia, and alcohol intake.


Introduction
The term insulin resistance (IR) was coined in 1936 to describe a metabolic disturbance characterized by decreased cellular responsiveness to insulin signaling in insulin-dependent tissues such as skeletal muscle, liver, and adipose tissue [1]. IR has now been established to play a key role in pancreatic cell dysfunction and type 2 diabetes mellitus (DM2) [2]. This disease represents one of the leading causes of mortality in the adult population worldwide [3], with alarming projections for the future, with an estimated prevalence of 33% in adults in the USA by 2050 [4]. Furthermore, IR is a fundamental pathophysiologic component of several other endocrinemetabolic disorders, such as hypertension, dyslipidemias, polycystic ovary syndrome, metabolic syndrome, and cardiovascular disease [5,6].
In light of the great morbidity and mortality rates associated with these disorders, prominent scientific effort has been directed to the refinement of tools for the detection and assessment of IR. In this regard, HOMA2-IR [7] stands as one of the most broadly accepted models in research settings yet remains relatively underutilized in our local clinical scenario. In this context, previous analyses by our research group have estimated specific cutoffs for this variable in our population [8], laying the groundwork for their potential clinical and epidemiologic use.
IR is intimately linked with psychobiologic habits such as smoking, alcohol consumption, hypercaloric diets, and physical inactivity. In turn, these factors can be strongly conditioned by demographic, social, and economic aspects, which shape daily human living. Thus, the objective of this 2 Advances in Preventive Medicine study was to determine the prevalence and determinants of IR in the adult population of Maracaibo city, Venezuela.

Sample Selection.
This report is part of the Maracaibo City Metabolic Syndrome Prevalence Study, a cross-sectional study whose purpose is to identify MS and cardiovascular risk factors in the adult population of Maracaibo, the second largest city of Venezuela. The sample (1,986 individuals) was calculated based on estimations of the city's population by our National Institute of Statistics (1,428,043 inhabitants for the year 2007). A total of 244 subjects (12%) were added for oversampling, in order to increase accuracy of the estimates obtained from smaller subgroups from the overall sample, amounting to a total of 2,230 individuals. Maracaibo city is divided into parishes, each of which was proportionally sampled with a multistaged cluster method: In the first stage, clusters were represented by sectors from each of the 18 parishes, selecting 4 from each parish by simple randomized sampling. In the second phase, clusters were represented by city blocks within each sector, which were selected by simple randomized sampling using a random number generation tool. Further details of the sampling process have been previously published elsewhere [9].

Ethical Considerations.
All participants signed written consent prior to undergoing physical examination and blood sample collection. All procedures were approved by the Ethics Committee of the Endocrine and Metabolic Diseases Research Center of The University of Zulia, Maracaibo, Venezuela.

Subject Evaluation.
Data were collected through completion of a full clinical record carried out by trained personnel, which encompassed interrogation regarding ethnic origin, marital status (single, married, cohabiting, or divorced), occupational status (employed or unemployed), and educational status, under the following definitions: (a) illiterate, subjects with no reading or writing skills; (b) primary education, those who attended or finished primary or middle school; (c) secondary education, those who attended or finished high school; and (d) high education, those who attended or finished college/university [9]. Socioeconomic status was assessed with the Graffar scale modified by Méndez-Castellano for the Venezuelan population [10]. This tool allows stratification of the population according to four variables: occupation of the head of household, educational level of the subject's parents, income source, and housing conditions. Each of these items is rated in a Likert-type scale from 1 to 5 (1 representing the best possible scenarios and 5 the worst). The scores for each item are tallied into a total integrated score ranging from 4 to 20 points, categorized as follows: Stratum I or High Class (4-6 points); Stratum II or Middle-High Class (7-9 points); Stratum III or Middle Class (10-12 points); Stratum IV or Relative Poverty (13-16 points), and Stratum V or Critical Poverty (17-20 points).

Smoking Habit.
Based on information obtained during the clinical interview, subjects were categorized by their smoking habits as follows [11]: (a) nonsmokers, individuals who had never smoked or had smoked <100 cigarettes in their lifetime; (b) current smokers, subjects who had smoked ≥100 cigarettes in their lifetime or reported current habitual smoking at the time of evaluation or had quit smoking less than one year prior to our assessment; and (c) past smokers, individuals who had consumed ≥100 cigarettes in their lifetime and quit over one year prior to our questioning.

Alcohol Consumption.
For the assessment of alcohol intake, subjects were asked to estimate the amount of alcoholic drinks they consumed within a month, with the approximate quantity and frequency of daily intake for each type of drink: beer, spirits, and wine and its derivatives. Accounting for the standard content of alcohol grams in each kind of beverage [12], daily intake of alcohol grams was calculated through the following formula [13]: where 0.8 is a constant representing ethanol density in drinks. Based on this estimation, drinkers were defined as subjects who consumed ≥1 gram of alcohol daily [14].

Physical Activity.
Physical activity (PA) was assessed with the International Physical Activity Questionnaire (IPAQ) [15]. For statistical analysis, PA was evaluated in 4 domains: occupational, household, transport, and leisure. In each of these domains, subjects were categorized as follows: (a) inactive, MET/week = 0, or (b) active, MET/week >0. The latter were then subcategorized by gender-specific MET/week quintiles in each domain (  25-29.9; and (c) ≥30 [17]. Waist circumference (WC) was evaluated with calibrated measuring tapes in accordance with the anatomical landmarks proposed by the USA National Institutes of Health protocol [18].
Advances in Preventive Medicine 3 2.3.6. Laboratory Analysis. Overnight fasting determination of glucose, total cholesterol, triacylglycerides (TAG), and HDL-C was done with an automated analyzer (Human Gesellschaft für Biochemica und Diagnostica mbH, Germany); the intra-assay variation coefficients for total cholesterol, TAG, and HDL-C were 3%, 5%, and 5%, respectively. LDL-C and VLDL-C levels were calculated applying Friedewald's formula [19] but when TAG levels were <400 mg/dL. LDL-C concentrations were directly measured through lipoprotein electrophoresis and densitometry with a BioRad GS-800 optical densitometer. Insulin was quantified using ultrasensitive ELISA double-sandwich methodology (DRG Instruments GmbH, Germany, Inc.).

Statistical Analysis.
Qualitative variables were expressed as absolute and relative frequencies, evaluating association through Pearson's Chi-squared ( 2 ) test, while the test for proportions was used to assess differences between proportions. Quantitative variables were evaluated for distribution normality with Geary's test and were expressed as arithmetic means ± SD. Variables with nonnormal distribution underwent logarithmic transformation; when normalization could not be achieved, these variables were expressed as medians (25th percentile-75th percentile). Student's -test or Mann-Whitney's tests were applied to evaluate differences between means or medians, respectively. A multiple logistic regression model was constructed in order to estimate odds ratios (confidence interval 95%) for the presence of IR, adjusted by gender, age groups, ethnic groups, socioeconomic status, educational status, marital status, occupational status, smoking habits, leisure-domain PA, glycemic status, presence of high TAG, presence of low HDL-C, JNC-8 classification, presence of elevated WC, and BMI classification. Data were analyzed with the Statistical Package for the Social Sciences (SPSS) version 21 for Windows (IBM Inc., Chicago, IL), and the results were considered statistically significant when < 0.05.

Characteristics of the General Population.
The sociodemographic and psychobiologic features of the sample are summarized in Table 2, while metabolic and anthropometric aspects are presented in Table 3

HOMA2-IR and Other Sociodemographic Factors.
When assessing the epidemiologic behavior of IR according to sociodemographic factors (Table 4), an association was found only in regard to marital status ( 2 = 17.293, = 0.001), with a significantly greater proportion of IR subjects among married individuals (51.5% versus 48.5% individuals with HOMA2-IR < 2).

HOMA2-IR and Metabolic
Characteristics. The presence of HOMA2-IR ≥ 2 was found to be associated with various metabolic characteristics (  Table 7 displays the behavior of several biologic and anthropometric characteristics by IR status. Individuals with this disorder had significantly higher age (    29; < 0.01), as well as impaired fasting glucose and DM2, were also associated with increased odds of IR in our population. Furthermore, no relevant differences were revealed after adjustment of a second model with exclusion of diabetic subjects.

Discussion
In recent years, a myriad of biologic, psychological, sociodemographic, and economic factors have driven the epidemiologic transition, which features a shift in the major worldwide causes of mortality from infectious diseases to chronicdegenerative disorders [22]. This group includes cardiovascular disease and DM2, where IR plays a fundamental role in their pathophysiology, in close association with Westernized lifestyles such as physical inactivity and hypercaloric diets [1]. Thus, detection and management of this condition have become a priority in primary care. Nevertheless, its epidemiologic behavior remains incompletely elucidated in many populations, including our locality.
We found the prevalence of IR in Maracaibo to be relatively elevated, in harmony with other reports within our country, 54.8% in Valencia city [23], 50.6% in a community from Tinaquillo [24], and up to 61.2% in a Canarian-Venezuelan population [25]. This may be explained by the growing prevalence of obesity and other cardiovascular risk factors in Venezuela [26], as well as methodological limitations and heterogeneity.
Moreover, IR prevalence is widely variable across demographics: whereas European populations appear to boast lower prevalence of IR, estimated at 17% in a Danish study, for example, [27], it has been reported to be as high as 51% in Iran, West Asia [28]. On the other hand, IR prevalence has been estimated at 39.1% in Hispanic subjects residing in Texas, USA [29]. Indeed, differing genetic, epigenetic, and sociocultural factors are important determinants of insulin sensitivity [30,31]. In addition, utilization of IR-assessing methods different from HOMA2-IR or arbitrary cutoffs may also contribute to these disparities [32].  In our sample, IR prevalence was similar between genders, diverging from findings in a Danish population [27] and the Framingham Heart Study [33], where men showed a higher prevalence, ostensibly attributed to the protective role of estrogens in the development of IR [34]. However, in our study, multivariate analysis revealed women to be at greater risk of IR, echoing a previous report by Rojas et al. [35] on subjects with acute coronary syndrome from Maracaibo. Furthermore, it should be noted that, in this study, the average age in females was 41 years, when a progressive decline in ovarian function tends to begin in most women [36]. In addition, females older than this average age displayed greater prevalence of DM2 and physical inactivity than their male counterparts; these factors favor endocrine-metabolic aging [37] which may explain our findings.
Indeed, aging has been closely linked with IR, chiefly due to the loss of lean mass and gain of visceral adiposity seen in this process [38]. Nonetheless, in multivariate analysis, we found individuals aged ≥60 years to have lower odds of IR. In this regard, certain reports have outlined a protective or innocuous role for adiposity in old age, due to reduced lipolytic activity in this tissue [39]. Likewise, in subjects with chronic-degenerative diseases, a BMI ≥ 25 Kg/m 2 has been reported to be linked with lower mortality, in a phenomenon termed the "obesity paradox" [40,41]. This effect may be responsible for the lower prevalence of IR in our study, as subjects aged ≥60 years exhibited high prevalence of abdominal obesity and comorbidities such as DM2, metabolic syndrome, and hypertension. At any rate, this relationship is very complex with numerous confounding factors, for example, smoking habits, and other comorbidities; therefore, future studies should further explore the interplay between obesity, IR, and old age.
When examining sociodemographic factors, we found marital status to be associated with IR, which was more frequent among married subjects. Notably, a large part of these were male, were inactive, and had elevated BMI and WC, possibly underlying this epidemiologic behavior. This harmonizes with the research by Sobal et al. [42] and Kiadaliri et al. [43] who described married men to have increased risk of overweight and obesity, associated with more frequent and abundant eating and decreased PA. Occupational status also   displayed an interesting relationship with IR in our study, with employed individuals having increased odds for this condition. In this context, unemployed subjects exhibited higher leisure-domain PA, along with lower prevalence of elevated WC than their counterparts, in consonance with findings in an Iranian population where unemployment was a protective factor against overweight and obesity [43].
On the other hand, performance of any degree of leisure-domain PA, even under the recommended threshold, has been described to reduce global, cardiovascular, and cancer-related mortality [44], accompanied by a favorable impact on insulin sensitivity and glucose tolerance [45]. We found leisure-domain PA to be a protective factor against IR in males with >297 MET/week and females with >231 MET/week; thus, the American Diabetes Association recommendations on PA intensity and quantity may be applicable for prevention of DM2 in Maracaibo [46].
Alcohol consumption was found to be another risk factor for IR in our population. In a previous publication [14], we reported 45% of men and 16.7% of women in Maracaibo to be habitual drinkers; of these, 72.4% and 76.2% were shown to drink approximately 1-4 beers daily with a high prevalence of abdominal obesity in this group. In this respect, a metaanalysis by Bendsen et al. [47] encompassing 35 observational studies and 12 experimental studies determined high beer intake (>500 mL/day) to be associated with abdominal obesity, which in turn plays a central role in the development of IR.
We also found each of the separate components of the metabolic syndrome to be independently associated with IR, with the exception of hypertension. The relationship with obesity appears most prominent: In our population, BMI ≥ 30 kg/m 2 grants 4 times greater odds of IR; this link has been extensively demonstrated in several cross-sectional and longitudinal studies [48,49]. Similarly, high TAG has been associated with decreased insulin sensitivity [50], and both this disorder and low HDL-C have been independently linked with IR [51]. However, in our population, the mild association between elevated WC and IR appears to be dependent on the presence of DM2.
As may be expected, we also found high odds of IR in subjects with hyperglycemia; in this state, oxidation of fatty acids is attenuated, favoring their esterification and leading to a subsequent increase in synthesis of adipokines, which disrupt insulin activity and peripheral utilization of glucose [52]. In parallel, hyperglycemia potentiates oxidative stress via activation of the ceramide and diacylglycerol pathways and other intermediaries, which also hinder insulin signaling [53].
Our results show a close relationship between IR and various sociodemographic, psychobiologic, and cardiometabolic factors in our population, highlighting the clinician's role in the early detection of this condition. Indeed, regardless of the methods employed [54], identification of subjects at high cardiometabolic risk is a fundamental preventive strategy for reducing the morbidity and mortality of disorders associated with IR.