Impact of Socioeconomic Status on Health Behaviors, Metabolic Control, and Chronic Complications in Type 2 Diabetes Mellitus

Study population

The study population consisted of patients with T2DM age 30 to 65 years old who participated in the Korean National Diabetes Program (KNDP) cohort study. The KNDP cohort is a prospective, multicenter, observational study performed in Korean patients with T2DM and patients at high risk for diabetes that started enrollment in May 2006, and primary observation of the registered patients ended in March 2014. Study details have been published previously [1516]. The present study population consisted of 3,294 individuals from the KNDP cohort who were diagnosed with T2DM and were 30 to 65 years old at enrollment and had data on SES with concurrent baseline clinical evaluation. The study protocol was approved by the Institutional Review Board of Inha University Hospital (IRB 2006-67) and of each hospital and all participants provided written informed consent before participation.

SES, dietary assessment, and lifestyle variables

SES, dietary assessment, and lifestyle variables were collected according to a standardized questionnaire by trained research personnel. Indicators of SES used in the current study were household income and maximal education attainment. Level of education attainment was divided into three categories: less than high school, high school graduation, and college graduation. Data for monthly household income were obtained by a questionnaire that consisted of multiple-choice questions covering each 1,000,000 Korean won (KRW) as follows: (1) <1,000,000 KRW; (2) 1,000,000 to 2,000,000 KRW; (3) 2,000,000 to 3,000,000 KRW; (4) 3,000,000 to 4,000,000 KRW; (5) 4,000,000 to 5,000,000 KRW; and (6) >5,000,000 KRW. Monthly household income was divided into three categories after merging each two categories as follows: <2,000,000 KRW (1,700 US$), 2,000,000 to 4,000,000 KRW (1,700 to 3,400 US$), and >4,000,000 KRW (3,400 US$).

For dietary intake, patients were interviewed in person by trained registered dietitians using a 24-hour dietary recall questionnaire as previously described [14]. The reported dietary intake was analyzed by the Computed Aided Nutrient Analysis Program version 3.0 (CAN-Pro 3.0; The Korean Nutrition Society, Seoul, Korea) that converted food consumption records to nutrient intake. Intakes of total calories, carbohydrate, protein, total fat, and fiber were assessed in the current analysis. Energy-adjusted nutrient intakes were shown as percent of total energy for protein, fat, and carbohydrates and g/1,000 kcal for dietary fiber.

Leisure time physical activity was evaluated by a questionnaire that was modified from the Minnesota Leisure Time Physical Activity Questionnaire (MLTPAQ) [1417]. The total energy expenditure was calculated into kcal/day as previously described [14]. Smoking habit was classified into two categories: current smoker or not. Alcohol intake was categorized into two categories as current regular drinking or not. Stress level was evaluated by the Korean version of the Brief Encounter Psychosocial Instrument (BEPSI-K). The BEPSI-K score is the average of five closed question items, which is measured using a five-point Likert scale. A higher score indicates a higher stress level [1819].

Anthropometric and laboratory measurements

All anthropometric measurements were standardized for all participating centers and made with participants wearing light clothing without shoes. Weight and height were measured to the nearest 0.1 kg and 0.1 cm. Body mass index (BMI) was calculated as weight (kg) divided by the square of height (m). Obesity was defined as BMI ≥25 kg/m² [20]. Blood pressure was measured after the subject has been at rest for at least 5 minutes in a sitting position. Blood samples were collected after an overnight fast of at least 10 hours and was analyzed as previously described [21]. Glycosylated hemoglobin (HbA1c) levels were determined by a high-performance liquid chromatography method that received National Glycohemoglobin Standardization Program certification. Total cholesterol, triglyceride, high density lipoprotein cholesterol (HDL-C), low density lipoprotein cholesterol (LDL-C), creatinine, and plasma glucose levels were analyzed using an automated analyzer that received a certification of quality control from the Korean Society for Laboratory Medicine.

Diabetic complications and medical history assessment

Diabetic microvascular complications were assessed as previously reported [22]. To identify diabetic nephropathy, albuminuria was assessed by random urinary albumin creatinine ratio (UACR). Diabetic nephropathy with the presence of increased urinary albumin excretion was defined as UACR ≥30 mg/g Cr. Diabetic retinopathy was diagnosed by funduscopic examination.

Statistical analysis

Differences of clinical and laboratory variables between men and women were analyzed by Mann-Whitney U test and chi-square test for continuous and categorical variables, respectively. The association between education level or income level and baseline clinical and laboratory data were assessed separately in men and women. To assess the trend of variables according to education level or income level, spearman correlation was used for continuous variables, and Mantel-Haenszel test of trend was used for categorical variables. Partial spearman correlation was used to assess trend of continuous variables after adjustment for age, duration, or other relevant factors. Logistic regression analyses were performed to evaluate the impact of lower education and income levels on risks of binary variables, including uncontrolled hyperglycemia, obesity, and diabetic retinopathy after adjustment of confounding factors. All analyses were performed with SPSS version 19 (IBM Co., Armonk, NY, USA). Statistical significance was defined as P<0.05.

General characteristics of the population

The study population comprised 1,927 men (58.5%) and 1,367 women (41.5%). Median age was 52 years (interquartile range [IQR], 46 to 58 years), diabetes duration was 4 years (IQR, 1 to 10 years), BMI 25.0 kg/m² (IQR, 23.0 to 27.1 kg/m²), and HbA1c 7.4% (IQR, 6.6% to 8.8%). Men had higher education and higher household income than women. Compared to men, women had older age, lower systolic and diastolic blood pressure, fasting plasma glucose (FPG), triglyceride, and higher HDL-C and LDL-C levels. Women had lower energy, protein and fat intake, and higher carbohydrate and fiber intake than men. Women had lower UACR, lower rates of current drinking, smoking, and albuminuria (Supplementary Table 1).

Characteristics according to education and household income level in men

In men, 424 (22.1%) had an education level of less than high school, 739 of high school graduation (38.5%), and 758 of college graduation (39.5%). Higher education level was associated with younger age, higher income, more current drinking, and lower 2-hour postprandial glucose (PPG), HbA1c, and UACR. Patients with higher education level had a higher stress level, higher fat intake, and lower carbohydrate and fiber intake. There was no significant association between education level and leisure time physical activity, current smoking, or medication use in men. Those with higher education level had lower rates of diabetic retinopathy, albuminuria, and uncontrolled hyperglycemia (HbA1c ≥7.5%). After adjustment for age, diabetes duration, and household income level, higher education level was significantly associated with lower 2-hour PPG and HbA1c level and higher fat intake and lower rate of uncontrolled hyperglycemia (Table 1).

Table 1

Clinical and laboratory characteristics according to education in men

In men, 516 (28.1%) had a household income of <2 million KRW, 679 (37.0%) had an income of 2 million to 4 million KRW, and 642 (34.9%) had a household income of >4 million KRW. Higher income level was associated with younger age, higher education level, more current drinking, higher BMI and diastolic blood pressure, and lower 2-hour PPG, HbA1c, and UACR. Patients with higher household income had higher intake of energy, protein, and fat and a lower intake of carbohydrate and fiber. Patients with higher household income had less use of insulin, while statin use was more prevalent. Higher household income was associated with a higher rate of obesity and lower prevalence of diabetic retinopathy, albuminuria, and uncontrolled hyperglycemia. After adjustment for age, diabetes duration, and education level, higher household income level was significantly associated with higher current drinking rate, higher BMI and diastolic blood pressure, and lower HbA1c and UACR. Associations between income and protein intake, fiber intake, and albuminuria were not significant after adjustment, while associations between income and energy intake, carbohydrate and fat intake, use of insulin and statin, presence of diabetic retinopathy, and uncontrolled hyperglycemia and obesity remained significant (Table 2).

Table 2

Clinical and laboratory characteristics according to monthly household income in men

Characteristics according to education and household income level in women

In women, 719 (52.7%) had an education level of less than high school, 464 of high school graduation (33.9%), and 182 of college graduation (13.3%). Those with higher education level had younger age, shorter diabetes duration, higher household income, lower BMI, and higher FPG level. Higher education level was associated with higher energy, protein, and fat intake and lower carbohydrate intake. Higher education level was associated with lower use of sulfonylurea and renin angiotensin aldosterone system inhibitors (angiotensin converting enzyme inhibitors or angiotensin receptor blocker) and lower prevalence of diabetic retinopathy and obesity. After adjustment for age, diabetes duration, and household income level, higher education level was significantly associated with higher rate of current drinking, lower BMI, higher UACR, lower carbohydrate intake, and higher fat intake. The associations between household income and medication use, diabetic retinopathy, and obesity were attenuated after adjustment (Table 3).

Table 3

Clinical and laboratory characteristics according to education in women

In women, 637 (50.4%) had an income level of <2,000,000 KRW, 418 (33%) had an income level of 2,000,000 to 4,000,000 KRW, and 210 (15.4%) had an income level of >4,000,000 KRW. Higher income level was associated with younger age, shorter diabetes duration, higher education attainment, lower stress level, higher protein and fat intake, lower carbohydrate intake, and less insulin use. After adjustment for age, diabetes duration, and education level, patients with higher household income was significantly associated with lower stress level, higher protein and fat intake, and lower carbohydrate intake (Table 4). The higher stress level in lower-income patients was significant after further adjustment for current drinking, smoking, energy intake, carbohydrate intake, fat intake, and insulin use (ρ=−0.093, P=0.004 by partial spearman correlation analysis).

Table 4

Clinical and laboratory characteristics according to monthly household income in women

SES and uncontrolled hyperglycemia, obesity, and retinopathy

Since the prevalence of uncontrolled hyperglycemia, obesity, and diabetic retinopathy were significantly associated with SES in men, the association was evaluated with further adjustment for confounding factors. Logistic regression analysis, after adjusting for education or income level, age, diabetes duration, smoking, and drinking (Model 1), showed that lower education level conferred a significant risk of uncontrolled hyperglycemia in men. The highest education level (college graduation) was associated with 29% lower odds of having uncontrolled hyperglycemia compared with the lowest education level (less than high school) (95% confidence interval [CI], 0.52 to 0.97; P_trend=0.021) (Table 5). After further adjustment for insulin use, energy intake, and carbohydrate and fat intake (Model 3), lower education level maintained its significance as a risk factor of uncontrolled hyperglycemia in men (odds ratio, 0.63; 95% CI, 0.43 to 0.91) for highest education level compared to lowest (P_trend=0.048) (Table 5). In contrast, there was no significant association between SES and uncontrolled hyperglycemia in women.

Table 5

Odds ratios of education and income levels on uncontrolled hyperglycemia and obesity by gender

Men with higher income level were found to have a greater risk of obesity after adjustment for age, duration, current smoking, current drinking, and education level (Model 1). A household income of >4,000,000 KRW was associated with a 1.51-fold increase in the odds of obesity compared to a household income of <2,000,000 KRW (95% CI, 1.14 to 2.02; P_trend=0.006). This association was significant after further adjustment for insulin use (Model 2) but was attenuated after further adjustment for dietary factors (Model 3) (Table 5). In women, there were no significant association between SES and obesity.

For diabetic retinopathy, men with lower income level showed a higher risk for having diabetic retinopathy after adjustment for age, duration, education level, current smoking, and drinking (Model 1, P_trend<0.001). After further adjustment for insulin use (Model 2), HbA1c, systolic blood pressure, diastolic blood pressure (Model 3), energy intake, and carbohydrate and fat intake (Model 4), this association remained significant. Compared to an income of <2,000,000 KRW, an income of 2,000,000 to 4,000,000 KRW, and >4,000,000 KRW were associated with a 38% and 41%, respectively, lower odds of having diabetic retinopathy after adjustment (Model 4, P_trend=0.048) (Table 6). For women, there was no significant association between SES and diabetic retinopathy.

Table 6

Odd ratios of education and income levels on diabetic retinopathy by gender