Trends in Prevalence, Awareness, Treatment, and Control of Diabetes Mellitus in Mainland China from 1979 to 2012

Diabetes mellitus (DM) is one of the primary causes of premature death and disability worldwide. We performed a systematic review and meta-analysis of the published literature regarding the trends in prevalence, awareness, treatment, and control of diabetes mellitus in mainland China. PUBMED, EMBASE, Chinese Biomedical Database, China National Infrastructure database, Chinese Wan Fang database, and Chongqing VIP database were searched. Fifty-six eligible studies were included. Increasing trends in the prevalence, treatment, and control of diabetes in mainland China from 1979 to 2012 were observed. The pooled prevalence, awareness, treatment, and control of diabetes mellitus were 6.41%, 45.81%, 42.54%, and 20.87%, respectively. A higher prevalence of diabetes mellitus was found in urban (7.48%, 95%CI = 5.45~9.50) than rural (6.53%, 95%CI = 4.30~8.76) areas. Furthermore, an increasing chronological tendency was shown in different subgroups of age with regard to the prevalence of diabetes. A higher awareness of DM was found in urban (44.25%, 95%CI = 32.60~55.90) than rural (34.27%, 95%CI = 21.00~47.54) populations, and no significant differences were found in the treatment, and control of diabetes among the subgroups stratified by gender and location. From 1979 to 2012, the prevalence, treatment, and control of diabetes mellitus increased; nevertheless, there was no obvious improvement in the awareness of diabetes.


Introduction
With the rapid economic development, elevated standard of living, dietary shifts, lifestyle alterations, and aging, diabetes mellitus (DM) has become an important public health problem worldwide [1][2][3], which is estimated to be the third most challenging disease threatening public health after malignant tumors and cardiocerebral vascular diseases [4]. It has been estimated that the global number of individuals with diabetes will double from 171 million in 2000 to 366 million in 2030 among adults aged ≥20 years [5]. Data from European countries have indicated that the health care expenditure for patients with diabetes mellitus was significantly higher than for those who were not diagnosed with this disease [6][7][8]. Also, in the United States, it has been estimated that approximately 17.5 million people were diagnosed with insulin-dependent diabetes mellitus (IDDM) or non-insulindependent diabetes mellitus (NIDDM) in 2007, and the total annual cost in higher medical costs and lost productivity was estimated to be $174.4 billion [9], including $159.5 billion for the 16.5 million people with NIDDM and $14.9 billion for the 1 million people with IDDM [10].
A recent global study indicated that the prevalence of diabetes mellitus was rising rapidly, particularly in developing countries [5]. Moreover, more than 60% of the population with diabetes mellitus all worldwide comes from Asia, as this remains the world's most populous region. The number of persons living with diabetes will increase substantially in each Asian country over the next few decades [5]. China is the world's largest developing economy and the most populous country, with one-fifth of the global population. Due to the rapid westernization of diet and lifestyle, China has one of the largest diabetes mellitus populations around the world [11]. The total number of people with diabetes in China is projected to increase from 20.8 million in 2000 to 42.3 million
Selection Criteria. All included studies were required to meet the following selection criteria: (i) cross-sectional or longitudinal studies that provided the prevalence, awareness, treatment, and control of diabetes in mainland China; (ii) based on population samples rather than volunteers; and (iii) the study population being representative of the provincial or national population. Exclusion criteria were (1) reviews, editorials, letters, commentaries, or reports; (2) articles repeating data from other articles that were already included; (3) selfreporting data; and (4) studies based on special populations, that is, physical examination crowds, industry, or occupational groups, ethnicity, or age groups.

Data
Extraction. Data were extracted independently by two investigators. Any disagreements were resolved by discussion. We contacted the authors of eligible studies to request further or missing information if needed for subsequent analysis. Data regarding the first author, year of publication, study location, survey date, age range, sampling method, location (urban/rural), diagnostic criteria, diagnosis method, total sample size, total case size, gender distribution, age distribution, and prevalence, awareness, treatment, and control of the diabetes were extracted wherever available. An awareness of DM was considered a self-reported previous diagnosis of DM by a physician or other healthcare practitioner among participants with DM. Treatment of DM was determined as self-reported if taking oral hypoglycemic medications, using insulin, or other nonpharmacological treatments for the management of high glucose levels. The management of fasting plasma glucose (FPG) at levels lower than 7.0 rnmol/L (126 mg/dL) in patients of DM in treatment was defined as control of DM.

Statistical
Analysis. STATA software version 11.1 (Stata, College Station, TX, USA) and Review Manager (RevMan) version 5.1 [25] were used to calculate the pooled prevalence, awareness, treatment, and control of diabetes from all of the eligible studies. A summary of the prevalence, awareness, treatment, and control estimates was obtained using randomeffects meta-analysis. Statistical heterogeneity was assessed using the 2 statistic [26]. Subgroup analyses included gender, location, and age distribution of diabetes. Chongqing VIP database ( = 11,492). After excluding duplicates and following examination of titles, abstracts, and full texts, 172 potentially eligible studies were identified. In addition, with a manual search of references, two more studies [27,28] were obtained from the papers. Finally, there were 174 publications included for further analysis. Among these studies, 57 included repetitive data, 54 studies were researched in specific populations, and 7 studies were self-reports of diabetes prevalence. Therefore, we obtained 56 studies in total that were suitable for inclusion in our study [21][22][23]. Among the included studies, one study [77] contained survey data for two periods, with the data for 2002-2003 being duplicated in another study [43]. All of the included trials were published from 1980 to 2012.

Characteristics of the Included Studies.
We summarized the characteristics of the 56 included studies which contained 7 nationwide studies and 49 provincial studies conducted in the general Chinese population. Among the included studies, there were 56 concerning the prevalence of diabetes, 10 regarding the awareness of diabetes, 8 involving the treatment of diabetes, and 8 regarding the control of diabetes. A total population of 2,058,243 individuals was investigated, and 120,992 patients with diabetes were detected. The 56 studies were conducted among all of the provinces of mainland China, and included 22 provinces, 4 municipalities, and 5 autonomous regions. Four studies were conducted among all ages, with the age of participants in 9 studies being greater than 30 years old; 3 studies did not provide information regarding the age range investigated. The sampling methods applied in the studies included random sampling, stratified sampling, cluster sampling, multistage-stratified sampling, or combinations of these and general surveys. In terms of diagnostic criteria for diabetes, 7 studies used the American Diabetes Association criteria (ADA) from 1997 or 2009, 31 studies used the criteria of the World Health Organization (WHO) from 1985, 1988, 1995, or 1999, one study used the International Diabetes Federation criteria (IDF) from 2005, and one study was performed according to the criteria outlined in the Lanzhou conference on diabetes in China; however, 12 studies did not provide clear information on the diagnostic criteria employed. Eighteen studies used fasting plasma glucose (FPG) and oral glucose tolerance tests (OGTT) as the methods of diagnosis, 17 studies used FPG alone, 13 studies used OGTT, and one study used FPG, OGTT, and urine glucose tests. The prevalence of diabetes in 56 studies varied from 0.61% to 20.85%. Levels of awareness, treatment, and control ranged from 28.50 to 62.54%, 17.72 to 92.50%, and 6.86 to 35.87%, respectively (Table 1). Table 2, the overall prevalence of diabetes was 6.41% (95%CI: 5.50-7.33). The prevalence of diabetes by year varied from 0.81% to 15.60%. Table 2 and Figure 1 Forty-three studies reported the prevalence of diabetes by gender. The prevalence of diabetes in males was 6.91% (95%CI = 5.72-8.09), and the prevalence in females was 6.43% (95%CI = 5.12-7.74); there was no significant difference in the prevalence of diabetes between males and females (OR = 1.07, 95%CI = 0.98-1.16). Overall, trends in the prevalence of diabetes between males and females were increased and similar to the overall trend for the prevalence of diabetes; no significant differences could be observed between males and females (Figure 1(b)).

Trends in the Prevalence of Diabetes. As shown in
Fifteen studies provided the prevalence of diabetes by location. The prevalence of diabetes in urban and rural areas was 7.82% (95%CI = 5.93-9.72) and 6.26% (95%CI = 4.25-8.27), respectively. A statistically significant difference could be found in the prevalence of diabetes between urban and rural areas (OR = 1.61, 95%CI = 1.25-2.06). As shown in Figure 1(c), trends in the prevalence of diabetes in both urban and rural areas increased over time, and the prevalence of diabetes in urban areas was consistently higher than that in rural areas.
There were 26 articles which reported the prevalence of diabetes stratified by age. The prevalence of diabetes in the groups aged less than 39, 40-59, and over 60 was 1.98%, 6.96%, and 13.24%, respectively. Table 2 and Figure 1(d) show the information and trends in the prevalence of diabetes by age. An increasing tendency could be observed in all age groups over time. Also, the prevalence of diabetes increased with age ( Figure 1(d)).

Trends in the Awareness of Diabetes.
Ten studies provided information on awareness of diabetes. The pooled estimate for the awareness of diabetes was 45.81% (95%CI = 37.88-53.74) (Table 3). However, no obvious increasing trend could be observed in awareness of diabetes from 1998 to 2012 (Figure 2(a)).
Seven studies reported the awareness of diabetes by gender. Overall, the rate of awareness of diabetes in males was 40.86% and was 41.58% for females. No significant difference could be found in the awareness of diabetes between males and females (OR = 0.99, 95%CI = 0.87-1.13) ( Table 3).

Trend in the Treatment of Diabetes.
The combined result for the treatment rate of diabetes was 42.54% (95%CI = 13.69-71.38) ( Table 3). As shown in Figure 2 Six studies provided the treatment rate of diabetes stratified by gender. The treatment rate of diabetes in males was 38.48%, whereas it was 41.18% for females. Comparing the treatment rate of diabetes for males to that of females, no   significant difference could be observed (OR = 0.9, 95%CI = 0.72-1.14) ( Table 3). Three studies reported the treatment rate by location. The treatment rates in urban and rural areas were 50.18% (95%CI = 31.22-100.39%) and 45.42% (95%CI = −12.7% and −103.6%), respectively (Table 3). Comparing the treatment rate of diabetes between rural and urban areas, there was no significant difference (OR = 1.30, 95%CI = 0.77-2.20).

Trends in the Control of Diabetes.
The pooled control rate was 20.87% (95%CI = 10. 76-30.97) for diabetes from the combined information of eight studies (Table 3). There was  (Figure 2(b)). Five studies reported the control of diabetes stratified by gender. Overall, the control rate of diabetes in males was 19.26% and was 19.03% for females. There was no significant difference between males and females (OR = 1.06, 95%CI = 0.98-1.15) ( Table 3).

Discussion
Due to the rapid development of the economy in China, the dietary habits and lifestyles of individuals have changed remarkably, which might have contributed to the increase in many chronic diseases such as diabetes [11,80]. Overall, the results of our study showed that the prevalence, awareness, treatment, and control of diabetes in mainland China were 6.41%, 45.81%, 42.54%, and 20.87%, respectively. A higher prevalence of diabetes was found in urban areas than in rural locations. As for the awareness of diabetes, this was also higher among residents in urban than in rural areas. There were no significant differences when data were stratified by gender or location in the subgroup analysis with regard to the treatment and control of diabetes. To the best of our knowledge, our study is the first systematic review on the awareness, treatment, and control of the diabetes and covers the longest time period from 1979 to 2012 in mainland China on the level of survey above the provincial.
The diabetes prevalence in our study ranged from 0.61% [27] to 20.85% [69] with an average of 6.41%; however, many authors have shown different findings when studying other countries in Asia. Similar to our results, it was reported that the prevalence of diabetes in Indonesia is 5.7%. In contrast, a survey conducted in Vietnam indicated a prevalence of 1.4% to 2.5%, which is relatively lower than that of our study [81]. On the contrary, epidemiological data from Thailand and Japan show that diabetes prevalence estimates were approximately 11.9% and 10%, which were higher than our study [82,83]. The wide variation in the prevalence could be explained by the heterogeneity of the studies, geographical differences, and the application of different diagnostic and sampling methods. According to the statistics of the IDF, diabetes affects about 100 million people worldwide. By 2007, the number of affected individuals had reached 246 million [84]. From 1979 to 2002, nationwide epidemiological surveys of diabetes were conducted in China and the prevalence showed a 5-fold increase [27,32,[85][86][87]. In accordance with this, an increasing trend in the prevalence of diabetes was found in our study. However, limited studies were included for pooled prevalence estimates; therefore, more studies with high quality and larger sample sizes are required to further confirm our findings.
As to the prevalence by gender, no significant difference could be found in our study between males and females, which is similar to the results of most studies according to the International Diabetes Federation (IDF) [84]. However, the study reported by Yang et al. in 2010 [72] showed a difference in the prevalence of diabetes between males and females. This might partly be explained by the fact that men are more accustomed to unhealthy diets due to their work and also smoke more frequently than women, which are the main risk factors for diabetes. Moreover, men are much more careless with regard to their health. However, further studies are required to confirm our findings.
In the analysis stratified by location, the prevalence of diabetes in urban areas was higher than in rural locations, with an odds ratio of 1.61 for residents in urban and rural areas, which suggests that residents in urban areas were more likely to suffer from diabetes than those in rural regions. In accordance with our findings, many studies worldwide have also reported a higher prevalence of diabetes in urban areas than in rural populations [32,86,88]. With the rapid development of the economy, urbanization has become a current trend in developing countries [89]. Urbanization is related to the change in the food supply and lifestyles which could lead to an unhealthy diet, sedentary habits, and overnutrition [72,90]. The prevalence of overweight and obesity has increased over the past two decades, especially in urban areas and highincome groups, which are the risk factors for diet-related disease and contribute to the development of diabetes [80]. Age is an important risk factor for the development of diabetes [91][92][93]. It was reported that the prevalence of diabetes increased with age [85,86]. A nationwide epidemiology survey showed that prevalence of diabetes in the age group of 20-39 was 1.01%, while it reached 6.31% in the age group over 40; however, the highest prevalence of 11.34% could be found in the age group over 60 [85]. Results of our study further confirm these findings. With the largest population in the world, China has become an ageing country; therefore, the prevalence of diabetes will increase as a result of the aging population in China. The health care for the aged population needs more attention from the government.
Ten studies provided information on the awareness of diabetes. Among the included study subjects, 45.81% were aware of having diabetes. From 1998 to 2011, there was no obvious improvement in the awareness rate of diabetes. However, we report a difference between the urban and rural areas when conducting stratified analysis by location. This difference in awareness may be a result of a lack of access to health care and knowledge in rural areas. Increasing the awareness through health education remains an effective method for the prevention and treatment of diabetes.
Among the participants who had diabetes in our study, 42.54% of them were undergoing treatment. This is high compared with the study which first reported the treatment rate of diabetes (27.2%) in China [94]. The results of our study also showed an increasing trend in the treatment rate of diabetes from 2001 to 2011. However, no gender or location disparity was observed. The treatment of diabetes is not merely related to public awareness of the disease. More importantly, diagnostic methods play an essential role in the treatment of diabetes. Convenient, economic, and effective detection tools could improve the treatment rate of this disease.
It was reported that controlling blood glucose at a normal level can prevent many diabetes-related complications, such as retinopathy, nephropathy, neuropathy, and macroangiopathy [95]. Our results indicated an increasing trend in the control of diabetes. Also, no gender or location differences were found in the subgroup analysis. The control rate of diabetes in our study is 20.87%, which is higher than that reported from a previous study with a control rate of 9.7% [94], but lower than that identified in another study (40.3%) [96]. The control of diabetes can be influenced by many factors, for example, obesity, smoking, insufficient exercise, and genetic susceptibility. Therefore, diversity exists in the control rate of diabetes in different places. Regular, effective, and timely treatment is primary in diabetes control. Moreover, adjusting dietary habits, changing lifestyles, and giving up smoking are essential when taking the risk factors for diabetes into account.
Coory [97] stated that it was very difficult to avoid heterogeneity in a meta-analysis, and this is also true for the metaanalysis of data from epidemiologic studies due to methodological problems. There were several methodological problems that might help to explain the heterogeneity: (1) different degrees of urbanization and socioeconomic conditions existed; (2) the included studies were carried out in different cities at various time points; (3) large differences existed in the age range, sampling methods, sample sizes, and response rates; (4) different diagnostic criteria and diagnosis methods were used in studies, with other studies not mentioning them at all; and (5) not all of the included studies provided sufficient information on gender, location, and age for subgroup analysis.
There are some limitations of the current study which need to be highlighted. Firstly, the present study might have underestimated the true prevalence of diabetes, since some of the studies included in the present review only used fasting plasma glucose (FPG) levels to diagnose new cases of diabetes; also, some studies did not provide any information about the diagnostic methods. Thus, more accurate measurements in the detection of undiagnosed diabetes are necessary in future research, for example, the two-hour glucose tolerance test and the oral glucose tolerance test (OGTT). Secondly, in addition to the subgroups of gender, location, and age, we did not report any positive risk factors associated with the prevalence of diabetes and the awareness, treatment, and control of the diabetes (e.g., overweight, obesity, hypertension, and hyperlipidemia); thus, we were unable to assess the association between them. Thirdly, the age range is quite different among the studies included. The age distribution in the sample population could markedly affect the results, since some locations contain a large number of individuals >40 years old; therefore, the prevalence of diabetes in some studies is likely to be higher than in others. Fourthly, limited studies were included in our meta-analysis for the pooled estimates; for example, only one study was included for analysis in 2009 and 2012 as studies from that time are possibly yet to be published. This could also be seen with regard to the trends in awareness of diabetes in 2001 and in the treatment and control of diabetes in 2011. Therefore, more studies were required for further analysis. A further limitation of the present study is that information in the subgroup analysis of awareness, control, and treatment by gender, location, and age was limited; therefore, we did not perform the relevant trend analyses.
The study showed an increasing trend for the prevalence, treatment, and control of diabetes among the population of mainland China from 1979 to 2012, but not in the awareness. This review will help us understand the gaps in the current research, which is useful for investigators and health care providers regarding aboriginal health. Thus, urgent measures are needed to prevent the high prevalence of diabetes and to improve diabetes awareness, control, and treatment among the Chinese populations.