Spatio-temporal patterns of the associations between type 2 diabetes and its risk factors in mainland China: a systematic review and meta-analysis
Ting Zhou,Xiang Liu,Yuanyuan Liu,Xiaosong Li
DOI: https://doi.org/10.1016/S0140-6736(18)32661-8
IF: 202.731
2018-01-01
The Lancet
Abstract:Abstract Background There is a lack of accurate information about the associations between risk factors for type 2 diabetes and the incidence of this disease in China. We conducted a systematic review and meta-analysis to examine these associations and their spatio-temporal patterns. Methods We searched PubMed, MEDLINE, Embase, EBSCO, Science Direct, and the Chinese databases CNKI, VIP, CBM, and Wanfang, for relevant studies from January 1997 to December 2017. Observational studies of the associations between risk factors and type 2 diabetes in adults (aged 18 and over) in mainland China were retrieved. Overall analysis and stratification analysis—including secular trends and temporal distributions—were conducted, and random-effects modelling was applied. Findings 35 studies were included in the final analysis. A family history of type 2 diabetes (pooled odds ratio [OR] 2·89, 95% CI 2·38–3·49), hypertension (2·73, 2·25–3·36), central obesity (2·28, 1·94–2·68), dyslipidaemia (2·23, 1·70–2·91), hypertriglyceridemia (2·18, 1·64–2·92), general obesity (1·90, 1·66–2·18), hypercholesterolemia (1·65, 1·32–2·06), smoking (1·26, 1·13–1·40), and drinking (1·20, 1·05–1·36) were associated with type 2 diabetes, while female gender (0·87, 0·78–0·97) was negatively related to type 2 diabetes. In terms of temporal stratification, the estimated effects of general obesity increased gradually during the periods 1992–2005, 2006–2010, and 2011–2017, whereas the estimated effects of a family history of type 2 diabetes decreased. In terms of regional stratification, the magnitude of the pooled effect for hypertension, dyslipidaemia, and hypercholesterolemia in northern areas (Beijing, Gansu, Hebei, Henan, Liaoning, Ningxia, Shandong, Shanxi, and Tianjin) was greater than in southern areas (Anhui, Chongqing, Fujian, Guangdong, Guangxi, Hubei, Jiangsu, Jiangxi, Shanghai, Sichuan, and Zhejiang), yet the magnitude of the pooled effect for a family history of type 2 diabetes in southern areas was larger than in northern areas. Lastly, in terms of provincial stratification, statistical maps showed that the estimated effect of a family history of type 2 diabetes was consistent with the overall results in all of the provinces included, but for the other nine factors, the effects were not consistent with the overall results and the effects also differed among provinces. Interpretation As this is the first meta-analysis to focus on the spatio-temporal distributions of the associations between type 2 diabetes and its risk factors, our evidence may help to guide prevention and control of type 2 diabetes in different regions and populations of China. Funding This work was supported by the National Key Research and Development Program (number 2017YFC0907300) and the China Medical Board (number 12-106).