Objective: To determine the prevalence and clinical characteristics of cardiac autonomic neuropathy in type 1 diabetic patients who were followed up for 4 years to shed further light on the natural progression of cardiac autonomic neuropathy. Methods: It is a prospective cohort observational study, consisted of 57 patients who were originally studied using the standard tests proposed by Ewing and Clarke (1985). At two years follow up, 46 patients were reevaluated, 55 patients from the original study were reevaluated after another 2 years for the 3rd time using the same protocol. The control group comprised 30 age and sex matched healthy volunteers. McNemar test, ANOVA for repeated measurements, paired t test and unpaired t test were used for statistical analyses. Results: The prevalence of established cardiac autonomic neuropathy (CAN) at the beginning was 14%. Q-Tc intervals were found to be significantly higher in patients with abnormal cardiovascular reflex (CVRs) in the 2nd examination (0.4±0.04 vs 0.5±0.05 sec, p=0.006). Eighteen patients showed deterioration of their CVRs test between 1st and 3rd examination. There was deterioration of their glycemic control guided by glycosylated hemoglobin (8.5±1.4 vs 9.9±1.5%, p=0.05*) and albumin/creatinine ratio (4.4±4.0 vs 28.2±28.0 mg/g creatinine, p= 0.04). On the other hand, 12 patients showed regression of their CVRs test. Only their insulin dose showed significant decrease (1.8±1.3 vs 1.1±0.3 dose/kg, p=0.02). Conclusion: The prevalence of established CAN in diabetic patients is high at the beginning of the study. Glycosylated hemoglobin, systolic and diastolic blood pressure were significantly increased in diabetics with deterioration of their CAN. However, the dose of insulin was significantly decreased in diabetics with regression of their CAN.