OBJECTIVE: We aimed to investigate the risk of long-term mortality associated with weight and waist circumference (WC) change among older adults, particularly the overweight and obese ones. PARTICIPANTS: Community-dwelling elderly (n=1138). MEASUREMENTS: Weight and WC were reassessed three years after baseline. Mortality risk associated with a 5% weight/WC loss and gain was compared to that of weight/WC stability by Cox models adjusted for clinical, behavioral and social known risk factors for death (age, gender, BMI, smoking, diabetes, total cholesterol, hypertension, Chagas disease, major electrocardiographic changes, physical activity, B-type natriuretic peptide, C-reactive protein, creatinine, education and household income). RESULTS: Female sex was predominant (718; 63.1%). Mean age was 68 (6.7) years. Weight stability (696; 61.1%) was more common than weight loss (251; 22.1%) or gain (191; 16.8%). WC remained stable in 422 (37.3%), decreased in 418 (37.0%) and increased in 291 (25.7%) participants. There were 334 (29.3%) deaths over a median follow-up time of 8.0 (6.4-8.0) years from weight/WC reassessment. Weight loss (HR 1.69; 95% CI 1.30-2.21) and gain (HR 1.37; 95% CI 1.01-1.85) were associated with increased mortality, except in those who were physically active in which weight gain was associated with decreased mortality. Results were similar for participants who were overweight/obese or with abdominal obesity at baseline (HR 1.41; 95%CI 1.02-1.97 and HR 2.01; 95%CI 1.29-3.12, for weight loss and gain, respectively). WC change was not significantly associated with mortality. CONCLUSION: Although weight loss has been recommended for adults with excessive weight regardless of age, weight change might be detrimental in older adults. Rather than weight loss, clinical interventions should target healthy lifestyle behaviors that contribute to weight stability, particularly physical activity in overweight and obese older adults.