Background: Abnormal circadian blood pressure (BP) patterns and reduced heart rate variability (HRV) are established risk factors for cardiovascular events in diabetic patients. However, studies assessing all-cause mortality associated with altered BP patterns and HRV in diabetes are limited by follow-up periods of less than 10 years. Methods: We examined a cohort of 349 patients with type 2 diabetes ( n =284) or type 1 diabetes ( n =65) recruited in Pisa (Italy) from 1999 and followed-up for 21 years, all with available 24-hour ambulatory BP and HRV monitoring. Dipping, non-dipping and reverse dipping status were defined as a ≥10% decline, <10% decline, and ≥0.1% increase in average night-time systolic BP (SBP) compared with average daytime SBP, respectively. Results: After 6,251 person-years of follow-up (median follow-up 21.0 [14.0-21.0] years, 183 [52%] women, age 57.1±11.9 y, BMI 29.4±5.9 kg/m 2 , HbA1c 8.6±2.1%), a total of 136 (39%) deaths occurred. Compared with dippers (n=166), non-dippers (n=144) and reverse dippers (n=39) showed progressively higher prevalence of cardiac autonomic neuropathy (11%, 16% and 31%, respectively), low HRV (45% vs 53% vs 62%), 24-hour hypertension (40%, 60% and 67%), isolated nocturnal hypertension (5%, 27% and 49%), postural hypotension (14%, 26% and 43%), and lower prevalence of white-coat hypertension (31%, 17% and 13%). Reverse dippers and non-dippers had progressively lower mean overall survival (OS) compared with dippers (16.1±5.3 years, 17.5±5.3 years and 18.6±4.6 years, respectively). Reverse dippers also showed an increased risk of all-cause mortality after adjustment for age, sex, BMI, office SBP, plasma glucose, and diabetes duration and type (HR 2.3 [1.4-3.8]). Patients with low HRV had reduced mean OS than those with high HRV (16.9±5.5 and 18.8±4.4, respectively) but similar adjusted risk (HR 1.3 [0.9-1.9]). No significant interactions emerged between BP patterns, HRV and diabetes type on OS. Conclusions: Non-dipping and reverse dipping BP patterns are associated with an increased prevalence of cardiac autonomic neuropathy and reduced survival probability in diabetic patients over a 21-year follow-up, with reverse dipping more than doubling the adjusted risk of all-cause mortality.