How Have Researchers Estimated the Impact of Excess Weight on Mortality? A Systematic Review

The aim of this review is to describe and reflect on the methodology of the studies that estimate the attributable mortality (AM) to excess weight. The purpose was to provide a methodological framework and serve as a starting point to improve the estimation of AM to excess weight. This will lead to more comparable estimates over time and across countries.

A systematic review was conducted considering MEDLINE (Ovid), Embase and Web of Science databases. Studies that had estimated AM to excess weight published until December 2024 were included. Data extraction was performed in an ad-hoc table based on the STREAMS-P tool, and a descriptive analysis of the methodology employed was performed.

The search strategy yielded 2,163 results. Twenty-three studies estimating AM in four continents were included. The majority estimated AM to all-cause mortality (n = 9) and for cardiovascular diseases (n = 9), followed by cancer without specifying which type of cancer (n = 7). Variation was observed in the body mass index (BMI) cut-off points used to define overweight, obesity and excess weight. Thirteen studies applied the population attributable fraction (PAF) formula denoted by PAF = [P (RR-1)] / [1 + P (RR-1)]. Most of the relative risks (RR) derived from meta-analysis (n = 14), and only a minority studies (n = 6) used RR from national studies based on the population for which AM was estimated.

This review emphasizes methodological constraints on AM estimation to excess weight. Causes of death causally associated with excess weight should be prioritized rather than relying solely on all-cause mortality. The ideal scenario would be to consider lag times; RR derived from the population under study and to use WHO BMI cut-offs and measured data rather than self-reported weight and height. The absence of this data should not prevent AM estimation if limitations are acknowledged. Expanding available estimates worldwide is essential to improve understanding of the mortality burden associated with excess weight.