PurposeObesity unit attrition is frequent and contributes to treatment failure. Many studies evaluating attrition predictors were part of randomized trials, and different terminology and criteria were used in the engagement field. We aimed to investigate the factors potentially implicated in early (< 12 weeks) and late (> 12 weeks) attrition from an obesity unit in a community settingMethodsThis was a retrospective cohort study of 250 patients with obesity who were followed-up at our obesity unit. Our program included at least 6 meetings in 12 months. Sociodemographic and anthropometric data, and psychometric questionnaires were collected from all participants.ResultsOne-hundred thirty-four (53.6%) participants dropped out. Those individuals showed lower BMI, lower overall health status, and increased depression scores. In a multiple regression model, BMI (inversely; OR = 0.90; 95%CI 0.84–0.96) and depression score (directly, OR = 1.05; 1.00–1.10) were associated with attrition risk. Early dropouts (n = 47) had lower weights, smaller waist circumferences and worse mental health scores than late dropouts (n = 87) and more frequently lived alone. When compared to completers, early dropouts had lower weights, BMIs, waist circumferences, overall health and mental status scores, increased depression scores and percentage of individuals living alone. In a multiple regression, lower BMI (OR = 0.83; 0.75–0.92), lower mental status score (OR = 3.17; 1.17–8.59) and living alone (OR = 2.25; 1.02–4.97) were associated with early attrition risk.ConclusionLower BMI and increased depression score were associated with attrition. Early attrition was associated with lower weight, decreased mental well-being, and living alone. Individuals with these characteristics might need tailored approaches to enhance their engagement.Level of evidenceLevel V, retrospective descriptive study.