Prevalence and Management of Diabetes and Associated Risk Factors by Regions of Thailand: Third National Health Examination Survey 2004

The NHESIII was conducted between January and May 2004 by the Health Systems Research Institute (see acknowledgments). The study was approved by the Ethical Review Committee for Research in Human Subjects, Ministry of Public Health. All participants provided written informed consent.

A multistage sampling frame based on government registers was used. For areas except for Bangkok, three provinces were chosen by probability proportional to size (PPS) for each of the 12 health administration areas. At the second stage, nine electoral units or villages were selected by PPS from both urban and rural areas for each province. At the final stage, 15 individuals were selected by simple random sampling with replacement from four broad groups (male or female sex and 15–59 or ≥60 years of age) for each electoral unit/village. Replacements within a 5-year age range and of the same sex and electoral unit/village were randomly sampled. For Bangkok, nine electoral units were selected PPS from six geographical zones. The final stage sampling was identical to that for the other provinces. The final collected sample, after selection of replacement individuals, was 39,290 individuals of a target sample size of 42,120 (93.3%). The final collected sample as a percentage of the target sample size, after selection of replacement individuals, by geographic region was as follows: Central, 99.1%; Northeast, 96.5%; North, 92.0%; South, 89.2%; and Bangkok, 72.2%. The proportion of the Thai population that is urban by geographic region is as follows: Central, 32.3%; Northeast, 15.0%; North, 20.1%; South, 23.5%; and Bangkok 100.0%.

Demographic characteristics, previous diagnosis of diabetes, high blood pressure, or high cholesterol and corresponding medication use, and smoking habits were determined on the basis of an interview. Three serial measurements of blood pressure, 1 min apart, were made using a mercury sphygmomanometer. Weight, height, and waist circumference were directly measured using standard techniques (9). Participants were asked to fast for 12 h overnight before the venous blood sample was obtained. Sera were frozen and transferred to a regional university laboratory for analysis of plasma glucose using a hexokinase enzyme method. Serum total cholesterol was measured using enzymatic methods. All regional laboratories were standardized by a central laboratory at the Ministry of Public Health.

Diabetes was defined as FPG ≥7.0 mmol/l, use of medication (oral antiglycemic agents or insulin) for the treatment of diabetes during the previous 2 weeks, or a report of a previous diagnosis of diabetes by a medical doctor. Diagnosed diabetes was defined as meeting the criteria for diabetes and having a previous diagnosis of diabetes by a medical doctor. As A1C measurements were not obtained in the survey, control of plasma glucose was defined as FPG <7.8 mmol/l (10). IFG was defined as FPG ≥5.6 and <7.0 mmol/l. High blood pressure was defined as systolic blood pressure ≥130 mmHg or diastolic blood pressure ≥80 mmHg (average of the two measurements with the lowest variability) or use of blood pressure–lowering medication during the previous 2 weeks (11). High cholesterol was defined as total serum cholesterol ≥6.2 mmol/l or cholesterol-lowering drug use during the previous 2 weeks. Central obesity was defined as a waist circumference of ≥90 cm in men and ≥80 cm in women (12). Overweight was defined as BMI of ≥25 kg/m² (12,13).

The analysis was restricted to respondents who had fasted over the previous 12 h before blood collection (94.5% of respondents) with the sample weighted against the registered 2004 population by public health administration area, area of residence (urban/rural or geographical zone for Bangkok), sex, and 5-year age-groups up to 80+ years. All comparisons by sex, area of residence, geographic region, and fasting plasma glucose status were age- and sex-standardized to the national population. Adjusted Wald tests were used to determine statistical significance with P < 0.05 considered statistically significant. Robust methods of variance estimation were used to take into account the complex survey design using Stata 9.2 (StataCorp, College Station, TX).

FPG was 5.2 ± 0.04 mmol/l (mean ± SE), 5.3 ± 0.04 mmol/l in men, and 5.2 ± 0.04 mmol/l in women. Prevalences of diabetes and IFG in Thais aged ≥15 years were 6.7% (95% CI 5.9–7.6%) and 12.5% (11.0–14.2%), respectively. Diabetes prevalence in men was lower than that in women (6.0 vs. 7.4%; P < 0.05 for the age-standardized comparison), whereas prevalence of IFG was higher in men than in women (14.7 vs. 10.4%; P < 0.001 for the age-standardized comparison) These figures equate to an estimated 3.0 million Thais with diabetes (1.3 million men and 1.7 million women), and 5.6 million with IFG (3.2 million men and 2.4 million women). Diabetes prevalence was notably higher among women aged 55–74 years than among men of the same age (Table 1) (P < 0.001), whereas IFG prevalence was higher among men aged 25–54 years than among women of the same age (Table 1) (P < 0.001). Diabetes was more common in urban men than in rural men (Fig. 1) (P < 0.01) but was similar in urban and rural women (P > 0.05). Male prevalence was higher in Bangkok than in the urban North (Fig. 1) (P < 0.01). Bangkok and the urban and rural Northeast had the highest female prevalence of diabetes, which was higher than that in the urban North and rural South (Fig. 1) (P < 0.05 for all comparisons).

Individuals with diabetes have high levels of concomitant cardiovascular risk factors (Table 2), except for smoking, with 72.7% (95% CI 68.7–76.4%) having high blood pressure (≥130/80 mmHg or receiving medication), 33.0% (28.6–37.8%) having high total cholesterol (≥6.2 mmol/l or receiving medication), 48.8% (43.6–54.1%) being overweight (BMI ≥25 kg/m²), and 53.5% (48.3–58.6%) having central obesity (waist circumference ≥90 cm in men and ≥80 cm in women). Those with diagnosed diabetes tended to have a higher prevalence of high blood pressure, high cholesterol, and overweight but a lower prevalence of smoking than those in whom diabetes had not been previously diagnosed (Table 2). Individuals with IFG had more concomitant cardiovascular risk factors, except for smoking, than those with normal FPG (Table 2).

Of those identified as having diabetes in the survey, 53.3% (95% CI 47.8–58.8%) had not previously had the diagnosis of diabetes. Nondiagnosis was higher in men (62.6%) than in women (62.6 vs. 46.2%, respectively; P < 0.001 for the age-standardized comparison). This equates to an estimated 1.6 million (0.8 million men and 0.8 million women) Thais aged ≥15 years with undiagnosed diabetes. Although diagnosis rates increased with age, 48.1% of men and 36.7% of women aged ≥55 years had diabetes that remained undiagnosed. Of those in whom diabetes was previously diagnosed, 90.7% (95% CI 87.7–92.9%) were receiving either oral antiglycemic agents or insulin, and of those receiving treatment, 41.6% (37.8–45.5%) had FPG controlled to <7.8 mmol/l. Glycemic control was higher in women than in men (44.6 vs. 35.8%; P < 0.01). Of those in whom diabetes had been diagnosed, 71.1% had received advice from health personnel to reduce their weight, 90.3% to increase their physical activity levels, and 92.5% to consume a healthier diet. Diagnosis rates tended to be higher in urban than in rural men, but the difference was not significant (42.2 vs. 34.6% diagnosed, respectively; P = 0.057). Rates of diagnosis were similar for urban and rural females (55.0 vs. 52.7%). Rates of glycemic control were higher in rural than in urban women receiving treatment (47.3 vs. 38.3%; P < 0.05). Both Bangkok and the South had higher rates of diagnosis than the Central and Northeast regions (Fig. 2) (P < 0.01 for both compared with Bangkok; P < 0.05 compared with the South). The fraction of treated individuals achieving glycemic control was lower in Bangkok than in the Central and North regions (P < 0.01 for both comparisons) and higher in the North compared with the South and Northeast (P < 0.05 for both).

Of those individuals with diabetes and concomitant high blood pressure, 65.9% were not previously aware that they had high blood pressure. Of those who were aware, 82.2% were taking blood pressure–lowering medication, and of those taking medication, 14.9% had blood pressure controlled to <130/80 mmHg. Rates of awareness were lower in rural than in urban areas for both men (26.3 vs. 37.0%; P < 0.01) and women (34.9 vs. 43.5%; P < 0.05). The Northeast region had the lowest rates of awareness (Fig. 2) (P < 0.01 for all comparisons), whereas Bangkok had the highest (P < 0.05 for all comparisons). The fraction of those receiving blood pressure–lowering medication who had blood pressure <130/80 mmHg was universally low but higher in Bangkok than in other regions (Fig. 2) (P < 0.01 for all comparisons).

Diagnosis, treatment, and control levels of high cholesterol were low. Of individuals with diabetes and high cholesterol, high cholesterol had not been previously diagnosed in 72.0%. Of those in whom it had been diagnosed, 80.7% were receiving cholesterol-lowering medication, and of those receiving medication, 63.1% had total cholesterol controlled to <6.2 mmol/l. Rates of diagnosis were lower in rural than in urban women (21.0 vs. 36.3%; P < 0.001) but similar in rural and urban men (30.2 vs. 32.2%). The Northeast had the lowest rates of diagnosis (Fig. 2) (P < 0.05 compared with the North and South and P < 0.01 compared with Bangkok and Central). The fraction of those taking cholesterol-lowering drugs who achieved total cholesterol <6.2 mmol/l was lower in Bangkok and the South than in other regions (Fig. 2) (P < 0.01 for all comparisons).

This study demonstrates a continued high prevalence of diabetes and IFG in Thailand consistent with a previous survey (7). These rates are similar to those in the U.S. (14) and higher than those of China (15). A higher prevalence of diabetes was found in middle to older aged women compared with men of the same age, whereas the prevalence of IFG was higher among younger to middle aged men compared with women of the same age. This pattern is similar to that of a recent Korean survey (16). Urban men were more likely to have diabetes than rural men, whereas the prevalence of diabetes was similar in rural and urban women. This finding may be related to the shift in obesity from high to low socioeconomic groups occurring earlier in women, as has been noted in other populations (17). Despite these differentials, diabetes prevalence is relatively uniform throughout Thailand.

In two-thirds of individuals with diabetes in the current survey (1.6 million individuals), the disease had not been previously diagnosed. Subsequently, there were low rates of treatment and control of plasma glucose with insulin or oral antiglycemic agents. Of further concern are the high levels of cardiovascular risk factors, such as high blood pressure and high cholesterol in those with diabetes. Diagnosis, treatment, and control rates of these risk factors among those with diabetes are also low. As diabetes itself is an independent risk factor for coronary heart disease, stroke, and renal disease (2,18), control of cholesterol and blood pressure is essential to reduce the risk of these events. The economic burden associated with these outcomes is highlighted by the potentially large financial burden of including renal dialysis on the Thai universal health insurance scheme (19). A crucial finding is that a large proportion of older individuals have undiagnosed diabetes, and these individuals have the greatest risk of macro- and microvascular diseases. When these findings are combined with the presence of an aging population and the 5.7 million individuals with IFG, it is clear that a scaled-up response is needed to better diagnose and treat diabetes and prevent progression to diabetes in those with IFG (20) if the health and economic burden of the condition is to be reduced.

Recognizing the growing burden of diabetes and other noncommunicable diseases and risk factors, the Ministry of Public Health launched a nationwide program, “Healthy Thailand,” in 2004. This program includes components promoting healthier lifestyles (physical activity and diet) and set a target for 60% of Thai people aged ≥40 years to be screened for IFG and diabetes by the end of 2006. Surveys must continue to be routinely conducted to monitor the impact of these programs. Furthermore, the need for subnational assessment is highlighted by the lower rates of diagnosis for diabetes and associated risk factors in the Northeast and in rural compared with urban populations. The distribution of health resources, such as medical doctors, has historically been biased against both these populations (21).

Several limitations should be considered. Although replacements were made based on age, sex, and cluster, information to determine the response rate before replacement was not available. Differentials in the final collection rate may also obscure variation across geographic regions. Oral glucose tolerance tests were not conducted in the NHESIII, preventing an estimation of the exact prevalence of pre-diabetes. Long-term glycemic control could also not be assessed as A1C was not measured. Likewise, cholesterol fractions were not measured. Despite these limitations, a clear strength of the study is the large sample size and the ability to monitor diabetes prevalence and management at a subnational level.

This study demonstrates a high prevalence of diabetes and IFG across all regions of Thailand with low levels of diagnosis and appropriate management of blood glucose, blood pressure, and cholesterol among individuals with diabetes. These findings serve to emphasize to other countries undergoing similar epidemiological transitions the importance of establishing monitoring systems for diabetes built on valid and representative data. In Thailand, without increased efforts to prevent diabetes by promoting and facilitating healthier lifestyles along with improvements in the diagnosis and control of diabetes and associated risk factors, the health burden of diabetes has the potential to overwhelm the health care system.

The NHESIII was supported by the Bureau of Policy and Strategy, Ministry of Public Health. Participating individuals included the following: Suwit Wibulpolprasert, Wiput Phoolcharoen, Siriwat Tiptaradol, Yawarat Porapakkham, Porapan Punyaratabandhu, Yongyuth Chaiyapong, and Kasame Vejsutanonth (coordinating team); Bodi Dhanamun, Narin Hiransuthikul, Thosporn Vimolkej, Somrat Lertmaharit, Pornarong Chotiwan, Wiroj Jiamjarasrangsi, Poranee Laoitthi, Mayuri Chiravisit, and Sarawuth Urith (Chulalongkorn University); Chalermchai Chaikittiporn, Kanda Vathanophas, Chaovayut Phornpimolthape, Rawiwan Sangchai, and Chanya Siengsanor (Mahidol University); Virasakdi Chongsuvivatwong, Mafausis Dueravee, Somsak Vanseng, Arpapak Kiatkittipong, and Siriwan Deawsurintr (Prince of Songkla University); Thanaruk Suwanprapisa, Nongyao Udomvong, Darunee Tayati, Decha Tamdee, and Thanapan Junyasiri (Chiang Mai University); and Amorn Premgamone, Somdej Pinijsoontorn, Manop Kanato, Suchada Paileeklee, Wattana Ditsathaporncharoen, and Piyatat Tatsanavivat (Khon Kaen University). Data analysis was conducted by the Setting Priorities Using Information on Cost-Effectiveness (SPICE) Project which is funded by Wellcome Trust, U.K. (071842/Z/03/Z) and the National Health and Medical Research Council of Australia (301199).