*Strengths and Limitations*

Based on our knowledge, this is the first study, including a large sample, which examined healthy ageing within the functional ability framework and lifestyle behaviour in LAC countries. Our results contribute to the advance of healthy ageing knowledge in under-examined areas and especially in LMICs as current research has mostly focused on other countries [2]. The limited attrition between baseline assessment and follow-up together with our further analyses, which indicated that in most of the cases participants included in the study did not differ with those excluded, limit any potential information bias. In addition, the observed heterogeneity (I2) was minimal in most cases, reflecting the limited variability in point estimates among countries.

In our study, we created a healthy ageing index within the functional ability framework, as it is the latest recommendation from the WHO [8]. Previous research has mostly focused on the presence or not of some disease/illness as a recent review indicated [7]. Yet, the functioning framework is considered more useful for effective public-health responses than considering specific diseases [8]. For this reason, we created this healthy ageing index by not taking into account specific conditions (i.e., depression, cardiovascular diseases, etc.). Furthermore, we did not consider other domains of healthy ageing (i.e., psychological well-being and social well-being) [48] since the healthy ageing model provided by WHO also incorporates "resilience" as an ability to maintain or improve the level of functioning during challenging periods. This suggests that older people could use psychological/physiological resources and environmental characteristics (i.e., social relationships) to maintain or improve their functional ability. However, these domains were not included in our index, as our ultimate intention was to build an index focusing on the final observed outcome -functional ability-.

Our findings should be interpreted within the context of this study limitations. Lifestyle behaviours were all assessed via self-reported questionnaires. Therefore, potential measurement error could have occurred. Furthermore, questions were too broad not allowing to holistically assess the impact of different frequencies and intensities in physical activity, or of specific fruits and vegetables that were consumed. In addition, even though interviewers underwent substantial training to ensure consistency in the way surveys were conducted in the various settings, we cannot exclude the bias of cross-cultural differences among countries in the conceptualisation of some questions. Another limitation of our study could be the quite arbitrary way of categorising participants who are in the two highest fifths of the baseline healthy ageing score distribution as healthy agers and all others as normal agers. Nevertheless, supplementary analyses (Supplementary File S5) indicated that our conclusions were not influenced by a different categorisation of participants in the health and normal ageing groups. In addition, we characterised healthy behaviour when participants were "very" or "fairly" physically active and when they consumed fruits and vegetables on a daily basis. These categorisations constitute assumptions of our study and consequently our findings should be interpreted within the limitations caused by these. The limited follow-up time inherited with these data also made it impossible to examine the impact of lifestyle behaviours in a more prolonged time. Finally, our sample included participants 65 years old and over. Hence, we were unable to assess lifestyle behaviour and examine healthy ageing within a life course perspective approach [49].
