Iranian Journal of

---

Background: Diabetes mellitus is the most common cause of renal failure, blindness, non- traumatic amputation and neuropathy. Homocysteine, a sulfurated amino acid, has a close correlation with Methionine and Cysteine. The conversion of Methionine to Homocysteine and Cysteine is required coenzymes like vitamin B6, B12 and Folate. The effect of Metformin on serum Homocysteine level by decreasing vitamin B12 level in patients with type 2 diabetes mellitus was described previously. Methods: This is a prospective clinical trail study among patients with type 2 diabetes mellitus in Shiraz. 76 patients were divided into two groups (38 patients in each group). First group treated with Metformin 500-2000 mg/day and the second group treated with Glibenclamide 5-20 mg/day with follow up period of at least 6 months. Hb and MCV were used in follow up to detect megaloblastic anemia, indicator of B12 and folate deficiency. Fasting plasma Homocysteine level Hb A1C and blood sugar were measured in baseline and at 3 and 6 months follow up periods. Results: There was no significant difference between age, sex, weight, height and BMI and baseline serum profile between the two groups. Homocysteine level increased significantly in Metformin group at 3 and 6 months(P=0.003 and 0.001 respectively). Mean plasma homocysteine level after 6 months were 10.98±0.58 μmol/l in Metformin and 10.0± 0.88 μmol/l in Glibenclamide group, with significant difference between the two groups (P=0.001). Conclusion: Metformin increases the plasma Homocysteine level. Metformin will accumulate highly in gastrointestinal wall and cause malabsorption of vitamin B12, therefore we can conclude that the use of Metformin for 6 months can cause vitamin B12 malabsorption and increase in plasma homocysteine level. Increase in plasma homocysteine level was 7.54% in our study that is higher in comparing to the other studies. It can be explained by longer duration of Metformin therapy in our study. Rising in Homocysteine levels may have detrimental effect on vessels that need further study.

---

Background: Provide a health care service to the patients with diabetes provides useful information that could be used to identify, treatment, following up and prevention of diabetes. Explore and investigation of large volumes of data requires effective and efficient methods for finding hiding patterns in the data. The use of various techniques of data mining in particular Classification and Frequent patterns can be helpful.

Methods: This article is a narrative review. We searched keywords related to application of data mining in the field of diabetes, through related databases, in English language articles published from 2005 to 2015. Also related articles in the selected articles list have been analyzed.

Results: From the 2144 articles obtained in the initial search, 38 articles related to the subject of study, were selected. Several studies shown that classification and clustering algorithms, association rules and artificial intelligence are the most widely used data mining techniques for predict the risk of diabetes has been successfully used.

Conclusion: The important step in control of diabetes, use of the methods that could determine the possibility or lack of diabetes. According to studies conducted in this area seem to use data mining techniques to prevent, treat and discover the connection between diabetes and its risk factors, can lead to significant improvements in the field of diabetes research and provide better health care for this group of patients.