Atherosclerosis 234 (2014) 11e16
Contents lists available at ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Sugar sweetened beverages consumption and risk of coronary heart disease: A meta-analysis of prospective studies Chen Huang, Jianfeng Huang, Yu Tian, Xueli Yang, Dongfeng Gu* State Key Laboratory of Cardiovascular Disease, Fuwai Hospital, Department of Evidence Based Medicine, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 September 2013 Received in revised form 13 January 2014 Accepted 19 January 2014 Available online 15 February 2014
Objective: To summarize the evidence with respect to sugar sweetened beverages (SSBs) consumption and risk of coronary heart disease (CHD) and to recommend field standards for future analysis on this topic. Methods: We searched for articles published up to February 2013 through PubMed, EMbase, and Cochrane Library Database and reviewed reference list of the retrieved articles. Prospective studies with reported relative risks (RRs) with 95% confidence intervals (CIs) of CHD for different categories of SSBs consumption were included. Random-effects models were used to evaluate the associations by comparing the highest and lowest categories of SSBs consumption in relation to risk of CHD. Results: Four prospective studies with 7396 CHD cases among 173,753 participants were included in the meta-analysis. The pooled RR (95% CI) for CHD in the highest category of SSBs consumption in comparison with the lowest category of SSBs was 1.17 (1.07e1.28). Stratified analyses indicated a significant association for men but not for women, with pooled RRs (95%CI) of 1.17 (1.05e1.29) and 1.19 (0.94e1.50), respectively. For studies carried out in America, the pooled RR for CHD was 1.18 (1.07e1.30). Additionally, a one-severing per day increase in SSBs consumption was associated with a 16% increased risk of CHD (RR: 1.16, 95%CI: 1.10e1.23). Conclusion: Our meta-analysis of four studies suggests that consumption of SSBs may increase risk of CHD, especially among men and American populations. However, this finding was based on limited studies; further studies are warranted to critically evaluate the relationship. Ó 2014 Published by Elsevier Ireland Ltd.
Keywords: Sugar sweetened beverages Coronary heart disease Meta-analysis
1. Introduction Sugar sweetened beverages (SSBs) are products that contain added, naturally derived caloric sweeteners such as sucrose (table sugar), high fructose corn syrup, or fruit juice concentrates, all of which have similar metabolic effects [1] and appeal to a preference for a sweet taste. With the westernization of diets, SSBs have become one of the most frequently consumed beverages around the globe [2e5]. Though evidence from United States (US) has indicated that a decrease in sugar consumption between 1999e 2000 and 2007e2008, mainly due to reduction of SSBs consumption, average intakes were still much higher than the recommended limits [6] and SSBs constitute the main source of added sugar in Americans’ diet [7]. Concerning the considerably increased
* Corresponding author. Tel.: þ86 10 68331752; fax: þ86 10 88363812. E-mail address: [email protected] (D. Gu). http://dx.doi.org/10.1016/j.atherosclerosis.2014.01.037 0021-9150/Ó 2014 Published by Elsevier Ireland Ltd.
consumption of SSBs, health effects of SSBs deserve particular attention and cautious evaluation. Habitual SSBs consumption has been associated with several detrimental cardiometabolic effects, including weight gain and obesity [8], hypertension [9e11], diabetes [12], metabolic syndrome [12] and stroke [13]. For example, a 1-serving per day increase in SSB intake was associated with a 0.06 unit increase in BMI in children and 0.22 kg weight gain in adults over 1 year period [8]. Individuals in highest intake of SSBs had a 26% and 20% greater risk of diabetes and metabolic syndrome, respectively [12]. Accumulating evidence from epidemiological studies has indicated that high consumption of SSBs may increase risk of coronary heart disease (CHD) [14,15], but the results remain conflicting for different populations [16,17]. No any overview has been reported regarding to the association between SSBs and CHD at population levels. Therefore, we performed a meta-analysis of prospective studies to assess the association between the consumption of SSBs and the incidence risk of CHD, including a doseeresponse analysis.
12
C. Huang et al. / Atherosclerosis 234 (2014) 11e16
2. Methods
2.4. Statistical analysis
2.1. Literature search and selection
We used the relative risk as the measure of association between consumption of SSBs and CHD, while HRs were directly considered as the estimates of relative risks. Pooled RRs and corresponding 95%CIs were estimated using random-effects models by comparing the highest and lowest categories of SSBs consumption. The Cocharne Q test (significance level of p < 0.10) and I2 statistics (significance level of I2 > 50%) were used to test heterogeneity among studies [20]. I2 was calculated as I2 ¼ 100% (Q df)/Q. The doseeresponse association was estimated on the basis of data for categories of SSBs consumption levels on median dose, number of cases and person-years, and effect estimates with corresponding standard errors using generalized least squares trend estimation method (GLST) developed by Greenland [21] and Orsini [22]. Then the combined RR for a 1 serving per day increment of SSBs was calculated using the inverse of the corresponding variances as weights. The Northern Manhattan Study [16] was not included in doseeresponse analysis by GLST method due to lack of information on person-years. To test the robustness of the dosee response analysis, we included the Northern Manhattan Study [16] using weighted least squares (WLS) regression method [21], which based on the dose categories, RRs and corresponding 95%CI in the absence of information on person-years and cases per strata. A sensitivity analysis was performed to assess the effect of each individual study in which 1 study at a time was removed and the pooled estimate was re-calculated for the remaining studies. Total energy intake and adiposity may be involved in the pathway between SSBs consumption and CHD, and adjustment for these mediators may attenuate the association. Therefore, we compared our results to pooled estimates that were adjusted neither BMI nor energy intake. We further performed stratified analyses to test the robustness of the combined estimates by sex (men vs. women), countries where the studies were carried out (America vs. not America), duration of follow-up (
Contents lists available at ScienceDirect
Atherosclerosis journal homepage: www.elsevier.com/locate/atherosclerosis
Sugar sweetened beverages consumption and risk of coronary heart disease: A meta-analysis of prospective studies Chen Huang, Jianfeng Huang, Yu Tian, Xueli Yang, Dongfeng Gu* State Key Laboratory of Cardiovascular Disease, Fuwai Hospital, Department of Evidence Based Medicine, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China
a r t i c l e i n f o
a b s t r a c t
Article history: Received 23 September 2013 Received in revised form 13 January 2014 Accepted 19 January 2014 Available online 15 February 2014
Objective: To summarize the evidence with respect to sugar sweetened beverages (SSBs) consumption and risk of coronary heart disease (CHD) and to recommend field standards for future analysis on this topic. Methods: We searched for articles published up to February 2013 through PubMed, EMbase, and Cochrane Library Database and reviewed reference list of the retrieved articles. Prospective studies with reported relative risks (RRs) with 95% confidence intervals (CIs) of CHD for different categories of SSBs consumption were included. Random-effects models were used to evaluate the associations by comparing the highest and lowest categories of SSBs consumption in relation to risk of CHD. Results: Four prospective studies with 7396 CHD cases among 173,753 participants were included in the meta-analysis. The pooled RR (95% CI) for CHD in the highest category of SSBs consumption in comparison with the lowest category of SSBs was 1.17 (1.07e1.28). Stratified analyses indicated a significant association for men but not for women, with pooled RRs (95%CI) of 1.17 (1.05e1.29) and 1.19 (0.94e1.50), respectively. For studies carried out in America, the pooled RR for CHD was 1.18 (1.07e1.30). Additionally, a one-severing per day increase in SSBs consumption was associated with a 16% increased risk of CHD (RR: 1.16, 95%CI: 1.10e1.23). Conclusion: Our meta-analysis of four studies suggests that consumption of SSBs may increase risk of CHD, especially among men and American populations. However, this finding was based on limited studies; further studies are warranted to critically evaluate the relationship. Ó 2014 Published by Elsevier Ireland Ltd.
Keywords: Sugar sweetened beverages Coronary heart disease Meta-analysis
1. Introduction Sugar sweetened beverages (SSBs) are products that contain added, naturally derived caloric sweeteners such as sucrose (table sugar), high fructose corn syrup, or fruit juice concentrates, all of which have similar metabolic effects [1] and appeal to a preference for a sweet taste. With the westernization of diets, SSBs have become one of the most frequently consumed beverages around the globe [2e5]. Though evidence from United States (US) has indicated that a decrease in sugar consumption between 1999e 2000 and 2007e2008, mainly due to reduction of SSBs consumption, average intakes were still much higher than the recommended limits [6] and SSBs constitute the main source of added sugar in Americans’ diet [7]. Concerning the considerably increased
* Corresponding author. Tel.: þ86 10 68331752; fax: þ86 10 88363812. E-mail address: [email protected] (D. Gu). http://dx.doi.org/10.1016/j.atherosclerosis.2014.01.037 0021-9150/Ó 2014 Published by Elsevier Ireland Ltd.
consumption of SSBs, health effects of SSBs deserve particular attention and cautious evaluation. Habitual SSBs consumption has been associated with several detrimental cardiometabolic effects, including weight gain and obesity [8], hypertension [9e11], diabetes [12], metabolic syndrome [12] and stroke [13]. For example, a 1-serving per day increase in SSB intake was associated with a 0.06 unit increase in BMI in children and 0.22 kg weight gain in adults over 1 year period [8]. Individuals in highest intake of SSBs had a 26% and 20% greater risk of diabetes and metabolic syndrome, respectively [12]. Accumulating evidence from epidemiological studies has indicated that high consumption of SSBs may increase risk of coronary heart disease (CHD) [14,15], but the results remain conflicting for different populations [16,17]. No any overview has been reported regarding to the association between SSBs and CHD at population levels. Therefore, we performed a meta-analysis of prospective studies to assess the association between the consumption of SSBs and the incidence risk of CHD, including a doseeresponse analysis.
12
C. Huang et al. / Atherosclerosis 234 (2014) 11e16
2. Methods
2.4. Statistical analysis
2.1. Literature search and selection
We used the relative risk as the measure of association between consumption of SSBs and CHD, while HRs were directly considered as the estimates of relative risks. Pooled RRs and corresponding 95%CIs were estimated using random-effects models by comparing the highest and lowest categories of SSBs consumption. The Cocharne Q test (significance level of p < 0.10) and I2 statistics (significance level of I2 > 50%) were used to test heterogeneity among studies [20]. I2 was calculated as I2 ¼ 100% (Q df)/Q. The doseeresponse association was estimated on the basis of data for categories of SSBs consumption levels on median dose, number of cases and person-years, and effect estimates with corresponding standard errors using generalized least squares trend estimation method (GLST) developed by Greenland [21] and Orsini [22]. Then the combined RR for a 1 serving per day increment of SSBs was calculated using the inverse of the corresponding variances as weights. The Northern Manhattan Study [16] was not included in doseeresponse analysis by GLST method due to lack of information on person-years. To test the robustness of the dosee response analysis, we included the Northern Manhattan Study [16] using weighted least squares (WLS) regression method [21], which based on the dose categories, RRs and corresponding 95%CI in the absence of information on person-years and cases per strata. A sensitivity analysis was performed to assess the effect of each individual study in which 1 study at a time was removed and the pooled estimate was re-calculated for the remaining studies. Total energy intake and adiposity may be involved in the pathway between SSBs consumption and CHD, and adjustment for these mediators may attenuate the association. Therefore, we compared our results to pooled estimates that were adjusted neither BMI nor energy intake. We further performed stratified analyses to test the robustness of the combined estimates by sex (men vs. women), countries where the studies were carried out (America vs. not America), duration of follow-up (
