Response: increasing prevalence of type 2 diabetes in a rural bangladeshi population: a population based study for 10 years (diabetes metab j 2013;37:46-53).

We sincerely appreciate the interest and comments regarding our study, "Increasing prevalence of type 2 diabetes in a rural Bangladeshi population: a population based study for 10 years" which was published in Diabetes & Metabolism Journal 2013;37:46-53. Our responses to Nan Hee Kim's comments follow:

We agree with reviewer's comment that it may be more convincing if we included all the mentioned factors (like dietary data, smoking habits, and physical activity, etc.). We should keep in mind that this is a rural agricultural society where manual labour is the only means for farming. Therefore, other than the land lords (few indeed) people work manually for 8 to 10 hours. Therefore questions like "leisure time physical activity" may not be relevant. In addition, variables like dietary data and smoking habits were not included in previous two studies (1999 and 2004). It should be noted that in 2009 study, we did not find any significant association of body mass index (BMI), socioeconomic condition, education level, family history of diabetes, smoking habits, and physical inactivity as risk indicators for diabetes [1]. This is possibly owing to the fact that almost no women in culture smoke and physical activity is high with a low BMI.

Second, we used the same cutoff level for waist-hip ratio (WHR) to maintain the flow of our previous two studies [2]. We have also attempted with 3 cutoff level for WHR but did not get any significant findings possibly due to low statistical power in each stratum. This may also be mentioned that previous cross-sectional and prospective studies in Bangladesh have found WHR a better predictor of type 2 diabetes than waist circumference [1-4].

Third, decreasing prevalence of hypertension is noteworthy. Landmark study like WHO MONICA [5] is in agreement with our findings. It is true that antihypertensive is only used when diagnosed with hypertension and therefore should not affect the prevalence of hypertension in ideal situation. In this context people can buy almost any drugs over the desk in a pharmacy. People often use these antihypertensive drugs with increasing age and obesity without any prescription. Therefore we cannot rule out the possibility for such unexplained observations. However, we cannot ignore the possibility of measurement error but the measurement of blood pressure was conducted by trained physicians in all studies.

We again thank the reviewer for providing very thoughtful comments.