Consumption of sugar-sweetened beverages, artificially sweetened beverages, and fruit juice and incidence of type 2 diabetes: a systematic review, meta-analysis, and estimation of population attributable fraction
Bhupathiraju, Shilpa N
MetadataShow full item record
Imamura, F., O’Connor, L., Ye, Z., Mursu, J., Hayashino, Y., Bhupathiraju, S. N., & Forouhi, N. (2015). Consumption of sugar-sweetened beverages, artificially sweetened beverages, and fruit juice and incidence of type 2 diabetes: a systematic review, meta-analysis, and estimation of population attributable fraction. The BMJ, 351 (h3576)https://doi.org/10.1136/bmj.h3576
Background Current evidence is insufficient for prospective associations of consuming sugar-sweetened beverages (SSB), artificially-sweetened beverages (ASB), and fruit juice with type 2 diabetes (T2D) for which adiposity is adequately controlled. The population attributable fraction (PAF) for T2D remains unknown. Objectives To examine these prospective associations before and after adjustment for adiposity; and to estimate PAF for T2D due to SSB consumption in the United States and the United Kingdom. Data sources and eligibility Prospective studies of non-diabetic adults, published until February 2014, were identified by searching PubMed, EMBASE, OVID, and Web-of-Knowledge. The PAF was estimated in national surveys in the US, 2009-2010 (n=4,729 representing 189.1 million non-diabetic adults) and the UK, 2008-2012 (n=1,932 representing 44.7 million). Synthesis methods Random-effects meta-analysis and survey analysis for PAF due to SSB consumption. Results Prespecified information was extracted from seventeen cohorts (38,253 cases/10,126,754 person-years). Higher SSB consumption was associated with higher incidence of T2D by 18% per one serving/day (95% confidence interval=8.8 to 28%, I2 for heterogeneity=89%) and 13% (5.8 to 21%, I2=79%) before and after adjustment for adiposity, respectively; for ASB, 25% (18 to 33%, I2=70%) and 8% (2.1 to 15%, I2=64%); and for fruit juice, 5% (-1.0 to 11%, I2=58%) and 7% (0.8 to 14%, I2=51%). Potential sources of heterogeneity or bias were not evident for SSB. For ASB, publication bias and residual confounding were indicated. For fruit juice, the finding was non-significant in studies ascertaining T2D objectively (Pheterogeneity=0.008). Under specified assumptions for PAF, of 20.9 million events of T2D predicted to occur over 10 years in the US (absolute event rate=11.0%), 1.8 million would be attributable to SSB consumption (PAF=8.7%, 95% confidence interval=3.9 to 12.9%); and of 2.6 million events in the UK (absolute event rate=5.8%), 79 thousand would be attributable to SSB consumption (PAF=3.6%, 1.7 to 5.6%). Limitations Residual confounding and limited evidence of causality. Conclusions Habitual SSB consumption was associated with greater T2D incidence, independently of adiposity. Although ASB and fruit juice also showed positive associations with T2D incidence, the findings were likely to involve bias. Nonetheless, both ASB and fruit juice were unlikely to be healthy alternatives to SSB for the prevention of T2D. Under assumption of causality, SSB consumption over years may be related to a substantial number of cases of new-onset diabetes.
The study was funded by Medical Research Council Epidemiology Unit Core Support [MC_UU_12015/5]. SNB was supported by American Heart Association Postdoctoral Fellowship Grant (13POST14370012).
External DOI: https://doi.org/10.1136/bmj.h3576
This record's URL: https://www.repository.cam.ac.uk/handle/1810/248741
Licence URL: http://creativecommons.org/licenses/by/4.0/