**2. Epidemiological Investigations**

Concerning the wide consumption of tea, a number of epidemiological studies have evaluated the e ffects of tea on diabetes mellitus and its complications.

Several cohort studies have estimated the e ffects of tea on diabetes mellitus. A cohort study pointed out that the consumption of tea showed protective e ffects against T2DM (RR = 0.55, 95% CI (0.55, 1.08)) for Vietnamese adults [23]. Another prospective cohort study revealed a negative relationship between the risk of diabetes mellitus and the consumption of tea (HR = 0.77, 95% CI (0.59, 1.00)) for subjects aged less than 60 years old in the United States [24]. It was also reported that tea intake was negatively associated with diabetes mellitus (HR = 0.66, 95% CI (0.61, 1.22) in British subjects [25]. In addition, a prospective cohort study demonstrated an inverse association between green tea intake and the risk of diabetes mellitus (*p*-trend = 0.02) only for women in the Japanese population [26]. Furthermore, the Singapore Chinese health prospective study found that more than one cup of black tea per day could reduce the risk of diabetes mellitus by 14% (RR = 0.86, 95% CI (0.74, 1.00)) [27]. A retrospective cohort study found that green tea was inversely associated with T2DM (OR = 0.67, 95% Cl (0.47, 0.94)), while black tea and oolong tea showed no significant effects for Japanese adults [28]. For black tea, the reasons of inconsistent results may be related to the differences of dose and frequency of consumption as well as research subjects [27,28].

Some case-control studies also have evaluated the effects of tea on diabetes mellitus. For instances, a case-control study found that tea could reduce the risk of T2DM (OR = 0.66, 95% Cl (0.49, 0.89)) for Vietnamese adults [29]. Another case-control study revealed that long-term drinking of green tea had preventive effects on diabetic retinopathy (OR = 0.49, 95% CI (0.26–0.90)), and people who regularly drink green tea had a 50% lower risk of developing diabetic retinopathy than those who don't drink green tea in China [30]. A population-based case-control study in Shantou, China found that long-term consumption of oolong tea could reduce the risk of hypercholesterolemia and triglycerides (OR = 0.10, 95% CI (0.06–0.16)) in southern China [31]. Additionally, one descriptive study revealed that long-term tea intake had negative relation with T2DM in Cyprus [32].

Moreover, several meta-analyses also supported the protective effects of tea consumption against diabetes mellitus. These meta-analyses included cohort studies conducted in different countries, such as America, China, Japan, and South Korea, and involved subjects of different races, genders, and ages. It was found that tea could increase the fasting blood insulin level (1.30 U/L, 95% CI (0.36–2.24)) [6], and tea consumption ≥ four cups per day could reduce the risk of T2DM (RR = 0.8, 95% CI (0.7, 0.93)) [18]. Further, another meta-analysis revealed that three to four cups of tea per day had an approximately 20% lower risk of diabetes mellitus than no tea per day (RR = 0.82, 95% CI (0.73, 0.94)) [33]. In addition, a meta-analysis including sixteen cohorts revealed a significant linear and inverse association between tea consumption and the risk of T2DM (*p* = 0.02) [34], and it was also found that more than three cups of tea per day is beneficial for the prevention of T2DM (RR = 0.84, 95% CI (0.73, 0.97) [35].

However, there are also some inconsistent results in epidemiological investigations. For instance, a prospective cohort study reported that the consumption of tea was not associated with T2DM [19]. In addition, compared with those not consuming oolong tea, multivariable adjusted hazard ratios (HR) for developing diabetes mellitus were 1.64 (95% CI (1.11–2.40)) for those drinking two or more cups of oolong per day [36]. The reasons of negative findings may be due to different sensitivities of different ethnic groups to tea, differences of tea composition in different regions, and bad control of drinking time and dose. Further, green tea was found no effects on T2DM in two meta-analysis [37,38], which may be related to the quality of the included studies, the level of evidence, and the size of the sample.

In summary, many epidemiological evidences have supported the efficiency of tea consumption against diabetes mellitus and its complications (Table 1), although inconsistent findings also exist. In addition, studies are still lacking concerning about the association between tea and diabetic complications, which should be further investigated in the future.



*Antioxidants* **2019**, *8*, 170
