Objectives Obesity is subject to strong family clustering. The relatives of participants in weight-loss interventions may also modify their lifestyle and lose weight. The aim of this study was to examine the presence and magnitude of a halo effect in untreated family members of participants enrolled in a randomized, multi-component, lifestyle intervention. Methods A total of 148 untreated adult family members of participants in an intensive weight-loss lifestyle intervention (the PREDIMED-Plus study) were included. Changes at 1 and 2 years in body weight, physical activity, and adherence to a traditional Mediterranean diet (MedDiet) were measured. Generalized linear mixed models were used to assess whether the change differed between family members of the intervention group compared to the control. Results Untreated family members from the intervention group displayed a greater weight loss than those from the control after 1 and 2 years: adjusted 2-year weight change difference between groups was -3.98 (SE 1.10) kg (p < 0.001). There was a halo effect with regard to adherence to the MedDiet at one year which was sustained at two years: 2-year adjusted difference in MedDiet score change +3.25 (SE 0.46) (p < 0.001). In contrast, no halo effect was observed with regard to physical activity, as the untreated family members did not substantially modify their physical activity levels in either group, and the adjusted difference at two years between the 2 groups was -272 (SE 624) METs.min/week (p = 0.665). Conclusions In the first prospective study to assess the influence on untreated family members of a diet and physical activity weight-loss intervention, we found evidence of a halo effect in relatives on weight loss and improvement in adherence to a MedDiet, but not on physical activity. The expansion of MedDiet changes from individuals involved in a weight-loss intervention to their family members can be a facilitator for obesity prevention.