If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
Diabetes is a serious and increasing global health burden and estimates of prevalence are essential for appropriate allocation of resources and monitoring of trends.
Methods
We conducted a literature search of studies reporting the age-specific prevalence for diabetes and used the Analytic Hierarchy Process to systematically select studies to generate estimates for 219 countries and territories. Estimates for countries without available source data were modelled from pooled estimates of countries that were similar in regard to geography, ethnicity, and economic development. Logistic regression was applied to generate smoothed age-specific prevalence estimates for adults 20–79 years which were then applied to population estimates for 2013 and 2035.
Results
A total of 744 data sources were considered and 174 included, representing 130 countries. In 2013, 382 million people had diabetes; this number is expected to rise to 592 million by 2035. Most people with diabetes live in low- and middle-income countries and these will experience the greatest increase in cases of diabetes over the next 22 years.
Conclusion
The new estimates of diabetes in adults confirm the large burden of diabetes, especially in developing countries. Estimates will be updated annually including the most recent, high-quality data available.
]. Urbanisation has driven dramatic changes in lifestyle and in particular in developing countries. With these rapid transitions come accompanying increases in risk factors for noncommunicable diseases like type 2 diabetes. Estimates of the current and future burden of diabetes are important to appropriately allocate resources, drive health-promoting policies, and encourage action to prevent diabetes in future generations.
The International Diabetes Federation (IDF) has produced estimates of diabetes prevalence since the year 2000 [
]. Previous estimates of the prevalence of diabetes have demonstrated a large and increasing burden, with significant regional variability. The estimates in this paper provide the latest figures based on the most recent and highest quality data on diabetes prevalence for 219 countries and territories.
2. Methods
The methodology applied to generate the estimates is based largely on that used for previous estimates and is described in detail in Guariguata et al. [
]. Briefly, a literature search of PubMed, Medline, and Google Scholar for data sources reporting the age-specific prevalence of diabetes conducted from January 1980 through April 2013 using the search terms: ‘diabetes’ or ‘impaired glucose tolerance’ and ‘prevalence’ and ‘country name’ or ‘region/continent’; ‘cardiovascular risk factors’ and ‘country name or region/continent’. In addition, data sources were gathered from national health surveys conducted by governments, or non-governmental organisations such as the World Health Organization or World Health Surveys. Relevant citations from published literature were also reviewed and investigators within the IDF network were consulted to identify data sources.
Studies were considered for inclusion that reported age-specific prevalence of diabetes for at least three age-groups for adults between 20 and 79 years. Methodological information from studies was abstracted and classified by the following: method of diagnosis; sample size; study type (e.g. population-based, clinic-based, diabetes registry, medical records review); representation (e.g. nationally representative, regionally representative, single city or village, single ethnic group or cohort); age of the data source; type of publication (e.g. peer-reviewed publication, national report).
Using the classification criteria described above, a scoring system was developed using the Analytic Hierarchy Process [
] which allows the comparison of different parameters (e.g. study type versus type of publication) to create a system of weights whereby each criterion for characterisation receives a corresponding score. Experts from the IDF network were asked to complete pairwise comparisons of criteria and a composite score based on the collective opinion was derived. Additional experts were consulted in 2013, producing an updated composite score. The updated scores are available in the online appendix.
The scores for the criteria characterising each element in each study were summed to obtain an overall study score which was used in the selection process. Scores were plotted and produced a bimodal distribution with a smaller cluster at high scores (those scoring 0.5 and above), and a second large cluster ranging from 0.2 to 0.5. The median score was 0.32. Thresholds were selected based on the overall distribution of the scores with one top threshold in the nadir of the first cluster and a bottom threshold at the median. Studies that scored above an upper threshold (0.5) were always selected and those that scored below a lower threshold (0.32) were excluded. Studies scoring between 0.32 and 0.5 were considered for selection if no studies scoring above 0.5 were available from the same country. If considered, the top-scoring study in this middle range and any study scoring within 0.06 of the top-scoring study were selected. In countries where more than one study was selected, the age-specific prevalence of diabetes was calculated using a weighted average taking into account the underlying study scores for that country. Studies which scored most highly were nationally representative, conducted in the last 5 years, based on OGTT, and published in peer-reviewed literature.
When no studies were available for a country, studies from countries matched for geographic region, income group, ethnicity, and IDF Region were aggregated. The groupings used for these extrapolated estimates are available in the online appendix.
Studies were excluded that did not contain sufficient methodological information for characterisation, did not provide enough data on age-specific prevalence of diabetes, were conducted in hospital or clinic-based settings, were based only on treated diabetes, or were conducted before 1980. In addition, duplicates were removed where multiple publications were referencing the same study population or where an update of a survey was conducted using the same sampling frame (e.g. repeated national health surveys), previous ones were excluded. Studies reporting only on type 1 diabetes, or newly diagnosed diabetes were excluded.
2.1 Statistical methods
Statistical programming was done using the R statistical program version 2.15.2 [
]. The age- and sex-specific prevalence of diabetes was calculated for urban and rural settings for each country using logistic regression with diabetes prevalence as the dependent variable and age and a quadratic term for age as the independent variables. Where data were not available stratified by urban and rural setting, a ratio was applied to estimate the proportion of diabetes for each setting derived from aggregated data available within a data region together with the percentage urbanisation by country available from the UN Population Division to estimate the proportion of diabetes in urban and rural settings. Where sex-specific data were not available in the study, the distribution of diabetes in males and females was assumed to be the same. If a study was based only on self-reported diabetes, an estimate of undiagnosed diabetes was derived from proportions reported in population-based studies in the same data region. Full details of the methods and results on estimates of undiagnosed diabetes are available from Beagley et al. [
The calculated age-, sex- and setting-specific estimates were then multiplied by corresponding estimates of population for 2013 and 2035 for each of 219 countries and territories to produce cases of diabetes in adults. Prevalence estimates were standardised using the WHO Standard Population [
The number of people with diabetes for each of the seven IDF Regions (Africa (AFR); Europe (EUR); Middle East and North Africa (MENA); North America and Caribbean (NAC); South and Central America (SACA); South-East Asia (SEA); and the Western Pacific (WP)) and World Bank income group (low-income (LIC); lower middle-income (LMIC); upper middle-income (UMIC); and high-income (HIC)) were calculated by aggregating the numbers of people with diabetes for each country within the respective regions. Global estimates were calculated by aggregating the total cases of diabetes.
3. Results
From the literature search, 744 data sources were identified representing 162 countries. Of these, 174 were selected representing 130 countries. These data produced an estimate of 381.8 million adults in 219 countries and territories with diabetes for 2013; and projected the number to rise to 591.9 million in 2035. The estimates derived from these sources for the 80 most populous countries (with adult populations greater than 6.5 million) are presented in Table 1 and the underlying data sources are presented in Table 2. Sixteen country estimates presented in the table were based on extrapolation. Detailed estimates of prevalence and numbers of people with diabetes on all 219 countries and territories are presented in the online appendix.
Table 1Prevalence of diabetes and estimated number of people with diabetes (20–79 years) for 2013 and 2035: 80 most populous countries.
Country/territory
Prevalence adjusted to the national population (%)
A first national prevalence estimate of diagnosed and undiagnosed diabetes in France in 18- to 74-year-old individuals: the French Nutrition and Health Survey 2006/2007.
Prevalence rate of diabetes mellitus and impaired fasting glycemia in Hungary: cross-sectional study on nationally representative sample of people aged 20–69 years.
Ministry of Health, Directorate of Public Health and Primary Health Care, Ministry of Planning and Development Cooperation, Central Organization for Statistic and Information
Ministry of Health, Directorate of Public Health and Primary Health Care, Ministry of Planning and Development Cooperation, Central Organization for Statistic and Information
Third national Surveillance of Risk Factors of Non-Communicable Diseases (SuRFNCD-2007) in Iran: methods and results on prevalence of diabetes, hypertension, obesity, central obesity, and dyslipidemia.
Ministry of Health, Directorate of Public Health and Primary Health Care, Ministry of Planning and Development Cooperation, Central Organization for Statistic and Information
Ministry of Health, Directorate of Public Health and Primary Health Care, Ministry of Planning and Development Cooperation, Central Organization for Statistic and Information
Public Health Agency of Canada Government of Canada
Report from the National Diabetes Surveillance System: Diabetes in Canada, 2009: Table of Contents – National Diabetes Surveillance System – Public Health Agency of Canada.
Public Health Agency of Canada Government of Canada,
February 2009
Prevalence of diabetes and prediabetes (impaired fasting glucose and/or impaired glucose tolerance) in urban and rural India: phase I results of the Indian Council of Medical Research-INdia DIABetes (ICMR-INDIAB) study.
There were regional differences in the prevalence of diabetes across the seven IDF Regions and by income group (Table 3 and Fig. 1). The highest regional, unadjusted, prevalence was in the North America and Caribbean region (11.0%). However, after age-adjustment the Middle East and North Africa region had the highest prevalence at 10.9%. While the Africa region has the lowest prevalence of adults with diabetes (5.7%), it is projected to have the largest proportional increase in the numbers of adults with diabetes by 2035, with an increase of 109%. All regions are projected to have an increase in the numbers of people with diabetes larger than those projected for growth in the adult population alone. Overall, the numbers of adults with diabetes will increase by 55% by 2035.
Table 3Diabetes prevalence and number of people with diabetes (20–79 years) by IDF Region for 2013 and 2035.
The top ten highest prevalence countries are predominantly Pacific island nations led by Tokelau (Table 4). The other three countries are from the Middle East and North Africa Region: Saudi Arabia, Kuwait, and Qatar. Indeed, many of the countries in the Middle East and North Africa Region have high prevalence rates, well above the global prevalence of 8.3%. In addition, many island nations in the Caribbean, Indian Ocean, and Western Pacific have higher prevalence rates than mainland countries in the same regions.
Table 4Top 10 countries/territories for prevalence and number of people with diabetes (20–79 years), 2013 and 2035.
Countries with large adult populations are also those with high numbers of people with diabetes. China has the highest number of people with diabetes with over 98.4 million adults affected; this is followed closely by India with 65.1 million. Eight out of the ten most populous countries in the world are also found in the top ten countries with the highest number of adults with diabetes. Only Bangladesh and Nigeria are not in the top ten countries with the highest numbers of people with diabetes (replaced by Mexico and Germany) although these countries are ranked 14 and 16, respectively.
Diabetes prevalence and numbers of people with diabetes vary substantially by World Bank income group. The vast majority of people with diabetes live in low- and middle-income countries. When seen across income groups, the greatest increases in people with diabetes over the next 20 years will parallel increases in the adult population. However, for every percentage increase in the adult population, the proportional increase in the numbers of people will be greater and especially for developing countries. The greatest proportional increase in the number of adults with diabetes is expected in low-income countries (108%), followed by lower middle-income countries (60%), upper middle-income countries (51%), and finally high-income countries (28%).
Diabetes prevalence increases with age across all regions and income groups (Fig. 1). The highest age-specific prevalence is in people 60–79 years of age (18.6%), although the largest numbers of people with diabetes are in the 40–59 year age group (184 million). These patterns are expected to persist over the next 20 years. The proportions of adults with diabetes under the age of 50 vary by region and income group with the highest proportion in Africa region (61%) and low-income countries (67%); although there is significant overlap in these figures as most low-income countries are in the Africa region. The greatest proportional increase in the number of people with diabetes by age group is expected to occur in people between 60 and 79 years of age.
4. Discussion
The estimates presented here confirm the large and growing burden of diabetes in the world established by previous estimates. There remains considerable variation in the burden of diabetes across regions and income groups with developing countries disproportionately affected. People with diabetes in high-income countries are predominantly over the age of 50 (74%) while those in low- and middle-income countries are mostly under the age of 50 (59%). Particularly for developing countries, demographic patterns are expected to change substantially over the next generations with increases in life expectancy, decreases in the infectious disease burden, and higher rates of urbanisation. These changes will also drive increases in diabetes prevalence in those countries.
Fig. 1 presents the age-specific prevalence of diabetes when divided by income group and shows a much lower prevalence of diabetes in older adults in low-income countries, compared to higher income countries. This may be a result of a number of factors including: higher mortality in people with diabetes in these countries, or a cohort effect whereby it is younger people developing diabetes. As life expectancy increases, we can expect to see greater numbers of adults with diabetes living longer in these countries which will drive the large proportional increases for low- and middle-income countries over the next generation. In addition, the higher prevalence of diabetes among younger adults in lower and upper middle-income countries will also drive increases as those people age, countries develop, and life expectancy increases. Without concomitant improvements in the health system for early detection and treatment, the rates of complications and deaths due to diabetes can also be expected to increase substantially in these countries. Until prevention and awareness of the disease improves, we cannot expect to see a shift in the curves for low- and middle-income countries towards a decrease in prevalence for younger age groups like that seen for high-income countries.
The IDF approach to estimating diabetes prevalence is simple, reproducible, and intentionally conservative. Prevalence estimates for 2013 are derived from the most recent available data with an effort made to minimise extrapolation and use the available evidence to inform the assumptions applied. The estimates are greatly affected by changes in the underlying data. For example, estimates of diabetes in Turkey have increased from previous estimates due to the inclusion of data from the TURDEP-II study [
] showing a prevalence of 16.5%; an increase from the 11.4% estimate from the previous study. Similarly, a new study on the prevalence of diabetes in China published in the Journal of the America Medical Association [
] after the estimates presented here were calculated showed a prevalence of 11.6%, which, if selected for future estimates, would likely increase the prevalence estimate in that country and countries dependent on the data for extrapolation.
Global estimates of diabetes prevalence have shown increases over the past 15 years. The latest estimates surpass projections made by previous estimates. In 1998, King et al. projected that the number of adults with diabetes would reach 300 million by 2025 [
]. All of these estimates have been surpassed by the estimate in this paper of 382 million in 2013 and we can expect that the projections presented here are likely also to be an underestimate. Each estimate has been based on the latest and highest quality data available. While the methodology has changed and been adapted as more information has become available, the increases seen are likely due to increases in incidence reflected in the availability of newer data reporting increasing prevalence of diabetes. Some increases in prevalence could also be a result of decreases in diabetes mortality where health systems are preventing deaths from complications; however the evidence for this is lacking.
The estimates are also particularly sensitive to the data underlying the assumptions applied to estimate undiagnosed diabetes. As discussed at length in Beagley et al. [
], undiagnosed diabetes is particularly sensitive to the performance of the health system and not to the underlying biology of diabetes. Thus, sampling frames and grouping of countries for extrapolation can have a profound effect on the pooled proportion of undiagnosed diabetes applied to a country. The recent study from China [
] also reported a much higher proportion (70%) of undiagnosed diabetes than previous studies in the same country, which, if included in the future, could have repercussions for estimates in the region.
Changes in the classification of countries used for extrapolation are also likely to have an effect on the way the estimates are calculated. For instance, since the generation of these estimates, the World Bank has reclassified a number of countries into new income group categories [
]. Most notably, the Russian Federation has now been classified as high-income. Because of the way countries are grouped for extrapolation, and the size of the population in the Russian Federation, this shift is likely to have a substantial effect on future estimates.
The modelling approach used here takes into account only changes in the population and urbanisation for the calculation of the 2035 projections and presents a conservative estimate of the prevalence of diabetes. Thus, increases in obesity and other risk factors for diabetes could drive the prevalence even higher over the next 20 years.
There are more high quality, population-based studies of diabetes available than ever before. For the estimates presented here, 744 data sources were considered and 174 included representing 130 countries. Of those studies, 142 were nationally representative, 69 were based on fasting blood glucose values, and 55 were based on OGTT. There is some regional variability in the availability of high quality data and the Africa region had lowest proportion of countries with original data of sufficient quality for inclusion.
Nationally representative studies are particularly important for high population countries within which regional variability may be considerable. The availability of large, nationally representative, recent studies for India and China, and high quality routine data collection in Mexico, Australia, the United States and Canada, translate to more stability and a higher confidence in the estimates for those countries. However, there is a considerable lack of population-based surveys based on fasting blood glucose or OGTT in high-income countries where self-reported surveys are abundant. This has a particular impact on the estimation of undiagnosed diabetes which leads to more uncertainty in the estimates for those countries.
In conclusion, while it is clear that the availability of more high-quality data is essential to improve the certainty around the estimates of diabetes prevalence, the latest evidence shows that diabetes continues to be a large and increasing global health burden and is likely to continue to grow substantially in the next decades. Increases in diabetes prevalence are driven by rapid development and subsequent changes in lifestyle. Coupled with the comparatively slow development of health systems, diabetes and its complications will continue to be an especially high burden in low- and middle-income countries.
Conflict of interest
The authors have no conflicts of interest to disclose.
Acknowledgements
The authors would like to acknowledge the contribution of the IDF Diabetes Atlas committee and in particular Professor Nam Han Cho, chair of the committee, in the generation of the estimates presented here.
The 6th edition of the IDF Diabetes Atlas was supported by the following sponsors: Lilly Diabetes, Merck and Co., Inc., Novo Nordisk A/S supported through an unrestricted grant by the Novo Nordisk Changing Diabetes® initiative, Pfizer, Inc., and Sanofi Diabetes.
Appendix A. Supplementary data
The following are the supplementary data to this article:
A first national prevalence estimate of diagnosed and undiagnosed diabetes in France in 18- to 74-year-old individuals: the French Nutrition and Health Survey 2006/2007.
Prevalence rate of diabetes mellitus and impaired fasting glycemia in Hungary: cross-sectional study on nationally representative sample of people aged 20–69 years.
Condizioni di salute, fattori di rischio ricorso ai servizi sanitari Anno 2005.
Instituto nazionale di statistica,
Roma2007 (Available from: http://www3.istat.it/salastampa/comunicati/non_calendario/20070302_00/testointegrale.pdf [Internet])
Ministry of Health, Directorate of Public Health and Primary Health Care, Ministry of Planning and Development Cooperation, Central Organization for Statistic and Information
Third national Surveillance of Risk Factors of Non-Communicable Diseases (SuRFNCD-2007) in Iran: methods and results on prevalence of diabetes, hypertension, obesity, central obesity, and dyslipidemia.
Public Health Agency of Canada Government of Canada
Report from the National Diabetes Surveillance System: Diabetes in Canada, 2009: Table of Contents – National Diabetes Surveillance System – Public Health Agency of Canada.
Public Health Agency of Canada Government of Canada,
February 2009 (Available from: http://www.phac-aspc.gc.ca/publicat/2009/ndssdic-snsddac-09/index-eng.php [Internet])
The Central America Diabetes Initiative (CAMDI): Survey of Diabetes, Hypertension and Chronic Disease Risk Factors.
Pan American Health Organization,
Belize, San José, San Salvador, Guatemala City, Managua and Tegucigalpa Washington, DC2011 (Available from: http://www.paho.org/hq/index.php?gid=16710&option=com_docman&task=doc_view [Internet])
Prevalence of diabetes and prediabetes (impaired fasting glucose and/or impaired glucose tolerance) in urban and rural India: phase I results of the Indian Council of Medical Research-INdia DIABetes (ICMR-INDIAB) study.
Australian Bureau of Statistics,
2012 (Available from: http://www.abs.gov.au/ausstats/[email protected]/Lookup/4364.0.55.001main+features12011-12 [Internet])
00385-9&id=gr1.jpg){kind=link}