Product reformulation in non-alcoholic beverages and foods after the implementation of front-of-pack warning labels in Mexico

Background In late March 2020, the Mexican government announced an updated norm to include front-of-pack warning labels for packaged foods and non-alcoholic beverages. Warning labels came into effect in October 2020. To avoid displaying warning labels, producers can reformulate their products by reducing the content of calories or critical nutrients targeted by the policy (added sugars, saturated fat, and sodium) or removing non-caloric sweeteners or added caffeine. The objective of this study is to assess changes in the percentage of products above warning-label cutoffs for calories and critical nutrients and changes in the content of calories and critical nutrients associated with warning labels in Mexico. Methods and findings We used nutritional panel data collected by the Mexican National Institute of Public Health from ≈1,000 top-purchased products, which represented ≥60% of the market share for each of the included food groups according to household purchases in the Nielsen Consumer Panel commercial dataset for Mexico in 2016. Nutritional panel data is available for three periods: 2016−2017, T0 (pre-policy); Jul–Sep 2020, T1 (post-warning-label announcement); and Feb–Apr 2021, T2 (post-warning-label implementation). We assessed changes in T1 versus T0 (potential anticipatory reformulation before the warning-label implementation) and T2 versus T0 (reformulation after the warning-label implementation) by food group using generalized estimating equations for the percentage of products above warning-label cutoffs or containing non-caloric sweeteners or added caffeine, and fixed-effects linear models and quantile regressions for the content of calories and critical nutrients. Included food groups were cereal-based desserts, bread and other cereals, salty snacks, sweetened beverages, solid dairy, liquid dairy, instant food, and candies. At T0, the food group level with the lowest percentage of products with at least one calorie/nutrient content above warning-label cutoffs was instant food (77.8%); at T2, this fell to 52.6%. Based on our statistical models, we found that all food groups showed reductions in at least one type of warning label. The most common reductions in the percentage of products exceeding warning-label cutoffs were for sodium (up to −63.1 percentage points for bread and other cereals; 95% CI: −77.5, −48.6; p-value < 0.001), saturated fat (up to −26.3 percentage points for salty snacks; 95% CI: −35.8, −16.8; p-value < 0.001), and products containing non-caloric sweeteners (up to −29.0 percentage points for solid dairy; 95% CI: −40.7, −17.2; p-value < 0.001). The reductions in products above warning-label cutoffs were coupled with reductions in products’ content of calories and critical nutrients. According to quantile regressions, these reductions mostly occurred at the 50th–75th percentiles. Product reformulation mainly occurred in T2. Conclusion Our findings show product reformulation due to reductions in critical nutrients/calories after the warning-label policy implementation, which entails improving the nutritional profile of the packaged food and beverage supply in Mexico.