The importance of good metabolic control in the treatment of diabetes mellitus (DM) is unquestionable. Measurements of the level of control (LC) are available such as glycohemoglobin (GH) and fructosamine (FA). The aims of this study are: to estimate LC in a population based case series of type II DM patients, to assess the effect of clinical, metabolic and sociodemographic variables usually associated with the LC and to estimate the clinical impact of the LC on vascular complications of DM. A low income area of Mexico City was enumerated. All males and non-pregnant females between 35-64 years of age were considered eligible. Home interview was obtained in 2813 and a medical exam including an oral glucose tolerance test was performed in 2282 (65.1%). Three hundred and four subjects met the criteria (WHO) for DM. Three years later, 213 (70.1%) were located for GH and FA measurements. Cases were divided according to the level of GH: Group I, good LC (GH < or = 8.6%, n = 66); and group II poor LC (GH > or = 8.7%, n = 147). A significant difference was observed in the mean value of FA between groups (Group I 255 +/- 52.3 mumol/dl. Group II 306.4 +/- 51.1; p < 0.001). In comparison with subjects with good LC, patients with poor LC had higher values of duration of DM (p < 0.01), waist/hip ratio (p < 0.04), fasting glucose (p < 0.001) and lower values of fasting insulin (p < 0.07), LDL cholesterol (p < 0.006) and VLDL cholesterol (p < 0.09).(ABSTRACT TRUNCATED AT 250 WORDS)