Meat consumption scenarios

To define the baseline scenario (the quantity of beef and processed meat consumed) we gathered the latest available data on average daily consumption by adult meat eaters (aged 18–64) by gender and geographical area from the National Food Consumption Survey (INRAN-SCAI) in 2005–2006 [24]. To define the scenarios (S1 Table), we referred to three different consumption targets. The first scenario, according to the World Cancer Research Fund advice [5], is compatible with the consumption of about 300g per person/week of red meat, which represents 40% of the total meat consumption in Italy. By applying this to the meat categories studied, the first scenario assumed a target of 244gr person/week for beef consumers and 147gr person/week for processed meat consumers. The second and third scenario are based on the Mediterranean Pyramid Model of recommended quantities [27] for red meat (150gr/week which represents 63% of the red meat currently consumed in Italy) and processed meat (not more than 50gr/week which represents 80% of processed meat currently consumed) respectively.

Health benefits

We calculated the health impact due to reducing meat consumption in both annual chronic disease burdens and the long-term impact on overall mortality.

Concerning the first health outcome, for each geographical area and gender, we estimated the changes in cardiovascular disease (CVD) and colorectal cancer mortality that would occur if meat consumption were reduced from the baseline to the hypothetical scenarios. The assessment of avoidable deaths (AD) was ascertained using the following functions [28]:

Which combine the dose-response association with mortality (β) and changes (i.e. difference) in consumption from baseline to suggested scenarios (ΔX) with data from adult meat-eaters (P₀) and cause-specific mortality rate (y₀).

The exposed population (P₀) was estimated using data come from the National Statistics Office Multipurpose survey on families showing frequency of food consumption, by gender, age and geographical area over the period 2002–2012. In particular, referring to the most recent available year (2012), we defined the exposed population through the percentage of adults (aged 20+ years) that regularly at least one a week eats beef and processed meat. Similarly, we used data from the National Statistics Office survey on deaths and causes of death 2012 from cardiovascular diseases (CVD) and colorectal cancer, by gender, age and geographical area to define the baseline mortality rate (y₀). Lastly, the coefficient β is based on a log-linear relationship between relative risk and meat consumption defined by epidemiological studies:

• For red meat: a dose–response relationship with colorectal cancer, with a 17% increased risk (95% CI 1.05–1.31) per 100 grams per day of red meat [15], and a 15% increased risk of mortality from CVD (95%CI 1.05–1.26) [12].
• For processed meat: 18% increase in risk of colorectal cancer (95% CI 1·10–1·28) per 50 grams per day [15], and a 24% risk of mortality from CVD (95%CI 1.09–1.40) [12].

In addition, 95% confidence intervals were calculated to evaluate the range of the variation of attributable deaths.

To estimate the long-term impact of healthier meat consumption habits we used life table methods based on the IOMLIFET model [29]. Separate life tables were constructed for the baseline scenario and the hypotheticals. We computed life tables distinctly for each geographical area and gender by populating them with the population of 2012 with age and sex-specific mortality rates for that year. We used a time-horizon of 106 years to predict gains in the 2012 newborn cohort life expectancy, while we emphasized the impact over the first 30 years (2030) for the 2012 population. The Mediterranean scenario was computed by applying the relative risk of overall mortality and meat consumption based on a meta-analysis of cohort studies:

For red meat consumption: RR = 1.21 (95%CI 1.15–1.26) for men and RR = 1.14 (95%CI 1.00–1.30) for women; For processed meat consumption RR = 1.23 (95%CI 1.10–1.37) for men and RR = 1.34 (95%CI 1.09–1.66) for women [12].

Based on these estimates, we calculated the decrease of age-specific mortality rates (aged 20 and above) corresponding to the difference in meat consumption between baseline and Mediterranean pyramid model scenarios.

Environmental benefits

Owing to the high contribution of livestock ruminants to GHG emissions [7], evaluating the environmental impact focused only on beef consumption by comparing the baseline and the Mediterranean scenario.

We explored the potential beef contribution to GHG emissions attributable to total consumption (grams), consumer’s choices (energy intake) and dietary patterns (low-carbon food substitutions).

All the analyses estimated GHG emissions by coupling the mass and type of food with the specific Global Warming Potential (GWP) coefficient. We used GWP coefficients from life cycle assessment studies (LCAs) [7, 30, 31] since they allow calculating the environmental impact of a product throughout its lifecycle [32].

With respect to the assessment based on total intake, we estimated annual GHG emissions attributable to the total intake of beef in the diet of the adult meat-eating population. We provided two estimates. The first is based on the specific Italian livestock production chains (from cradle to farm gate), assuming that all the consumed meat was raised in Italy [30]. The second is based on a world average GWP coefficient that includes both the post-farm gate processes and the variability of beef production process among LCAs [7].

With respect to the assessment based on consumers’ choices, we evaluated the annual differences in GHG emissions from baseline to the Mediterranean scenario by varying the beef quality from maximum energy intake (low quality) to minimum energy intake (high quality) in the diet of adult meat-eaters [31]. These differences depend mostly on fat contents. The minimum and maximum energy intakes were retrieved from all the beef cuts regularly consumed in Italy [33] and measured as kcal/100 grams of edible portions. An edible portion corresponding to baseline and Mediterranean beef consumption was calculated using FAO correction coefficients [34].

Concerning dietary pattern assessment, we followed the approach of the Australian case study proposed by Clune and colleagues [7] that employed the GWP coefficient to estimate environmental impact for different sustainable diet substitution scenarios. We compared total GHG and food specific contributions for adult weekly per capita consumption at baseline and under the Mediterranean scenario with low-carbon food substitutions for beef. The beef substitute was proposed by the Italian Institute of Nutrition to preserve the overall nutritional balance [35]. We stratified this analysis by geographical area to account for the heterogeneity in the dietary patterns across Italy.

Health co-benefits

This study is the first to provide evidence of health and environmental (in terms of GHG emissions) benefits deriving from reducing red and processed meat consumption by the Italian adult population to the levels recommended by international and national healthy diet guidelines. Our study estimates the health impact of eating less red and processed meat in Italy. It shows that a relevant proportion of colorectal cancer and CVD deaths are attributable to meat intake. It also demonstrates that a change in dietary habits in Italy, coherent with the national guidelines, would bring benefits in term of life expectancy.

We chose the Mediterranean pyramid model as a reference since we assumed it was better suited for a Mediterranean country like Italy [27]. If Italian meat consumers would follow the Mediterranean pyramid model [27], which indicates consuming a maximum of 150 grams of beef per week, it would result in a 1.2% to 6.0% reduction in mortality from colorectal cancer and between 1.2% and 5.2% from CVD. In parallel, eating 50 grams of processed meat per week would correspond to between 3.0% and 7.2% fewer deaths from colorectal cancer and 2.7%-9.3% from CVD. Considering total mortality, the reductions proposed are able to improve life expectancy in a 2012 birth cohort by 4–9 months by 2030.

The Mediterranean targets are coherent with the recommended consumption limits recognized to be safe for people of all ages [37, 38]. Recommendations regarding red meat are guided by the principle of limiting saturated fats, which have a detrimental effect on total and LDL cholesterol by increasing the risk of cardiovascular disease [39].

Comparing these results with those from other countries, especially those from Great Britain [18, 20, 40], where most of the studies have been carried out, is not immediately possible, since reduction scenarios and the health outcomes considered are often different. Nevertheless, the results are consistent with those summarized in a 2013 review, which reported the burden of disease went down by between 1 and 16% [41].

Previous studies provided reductions in total mortality in the range of 0–18% [31], or 6–10% [19, 40] associated with global livestock or red meat reduction consumption. Other studies have evaluated the impact of hypothetical change in specific nutritional factors, such as saturated fats intake, leading to reductions in cardiovascular burden in the range of 6–10% in 9 European countries [42] and of 14% by 2025 in Sweden [43].

It is important to stress that in studies based on food frequency questionnaires, a possible misclassification of food consumption may occur that can result underestimate risk: for example, in the EPIC study the estimated risk for colon cancer associated with 100 grams increases of meat consumption jumps from 1.25 to 1.55 due to systematic and random dietary intake measurement errors [44].

It is therefore possible, given that we used food frequency data surveys to estimate meat consumption, that the results of the present study are underestimated and that the health gains could be even higher. It should be highlighted also that eating red meat is a risk factor for other diseases, such as stomach cancer, diabetes and obesity, diseases that were not taken into consideration in this analysis. It has also been postulated that reducing meat consumption would result also in improved dietary choices, for example an increase in fruit and vegetable intake, allowing further health benefits [19]. In the present study, health impacts were calculated based on the relative risks estimated in meta-analysis studies of meat consumption.

Nevertheless, the life table approach allows to estimate the impact on all-cause mortality and has been acknowledged to be especially suited to express effects of long-term exposures such as diet [45]. This method is beneficial since it is able to provide a clear estimate of the public health significance of proposed reductions in the consumption of red and processed meat.

This analysis considered only adults, because proposed reductions in red and cured meats are applicable only to the healthy adult population in whom we observed significant health benefits such as reducing cardiovascular and cancer risks. Results cannot be transferred to susceptible subgroups, such as children, women of childbearing age, or the elderly who may be more vulnerable to iron or zinc deficiencies [46].

Geographical and gender differences

Few studies have evaluated the impact of geographical differences within the country of reduction scenarios of meat consumption, considering that health and environmental gains differed depending on variations in food production and consumption patterns [18, 19]. Our study showed potential health benefits deriving from the proposed dietary changes, which could be greater in northern areas where people tend to eat more. Such heterogeneity is consistent through the different analytical approaches we used (cause-specific mortality reductions in 2012 according to counterfactual scenarios, life years gained by Italian adults to 2030, improvement in life expectancy in future generations). Across Italian areas, there are slightly different culinary traditions and typical foods, but also differences in demographic and socioeconomic determinants that may affect attitudes towards a specific food. Processed meat consumption is uniform throughout Italy. There is a strong tradition of processed meat production, which includes ham, salami and other varieties. A survey conducted by the Italian Institute of Rural Sociology in 2002 counted more than 650 typical products from all Italian regions [47]. Nonetheless, this could be partly due to the higher costs of beef compared to processed meat, which makes consumption patterns easily influenced by social and economic conditions.

In all regions, there are fewer women who are regular consumers of red and processed meat, and they consume less than men do. Gender differences in specific food attitudes are well documented in western societies and possibly influenced by personal factors such as degree of health consciousness, and weight control that are more common in females than in males. In our study, Italian women not only eat less red meat but they also have a different baseline mortality risk for cardiovascular and colorectal cancer compared with men. These results underline the importance of developing gender-specific nutritional goals for the general population.

GHG emission impact

Current beef consumption by Italian carnivores has an environmental impact in the range of 12,900 to 21,800 Gg CO2eq. per year of GHG emissions. Emission savings according to the Mediterranean pyramid model are between 8000 and 14000 Gg CO2eq. per year. The Mediterranean scenario with low-carbon food substitutions reliable for the Italian population [35] produced an overall per capita reduction of 263 kg CO2eq./year. These potentially avoidable emissions thanks to dietary change are not negligible since overall annual per capita emissions from all sources (not only agriculture) are estimated to be around 7 tons CO2 eq./year [48]. The Italian Institute of Nutrition study produced a net environmental benefit similar to ours (114,6 kg CO2 eq. per capita saved per year) but it did not analyse regional variations [35]. Our study suggests that assessing the environmental impact needs to account for geographical heterogeneity in dietary habits across Italy. The low-carbon food substitute highlights that the greatest GHG benefits can be obtained in the Northwestern and Central regions where baseline beef consumption is higher. Moreover, GHG reductions in the same amount were obtained from different dietary changes.

Other studies have provided similar estimates of environmental benefit like the GHG emissions in the range of 29–70% associated with red meat consumption reductions [19], or have evaluated overall dietary changes, i.e. shifting to a vegetarian diet with reductions between 22–29% [49] or a Mediterranean diet (GHG reductions in the range 10–30%) [50]. Scenarios for reducing beef consumption at Mediterranean targets in Italy seem not only in line with recommended nutritional levels but also address reduction goals of GHG, which the European Union agreed to in Paris in November, 2015 [51]. The Paris agreements anticipate, in fact, a reduction of at least 40% in domestic emissions by 2030 and 80% by 2050. Various studies have stressed that adopting a diet that conforms to the guidelines would contribute to a substantial reduction in GHG emissions associated with food consumption [3]. According to the FAO, meat consumption is expected to increase by about 73% by 2050 due to population and income increases, even in developing countries [23]. To balance this increase, policies to reduce consumption are essential in order to achieve tangible benefits on population health and the environment. We used two GWP coefficients associated with beef production activity, one from an Italian LCA study [30] and another from a published meta-analysis of LCA studies including post-farm emissions which provided comparable food-specific coefficients on which we based the substitution analysis [7]. We also used a third coefficient based on energy units, to account for different cuts of beef consumed [31]. Although we followed a methodology previously suggested by others [7], our estimates are a simplification, since they do not correct the amount of food produced for the quantity actually consumed, stored, cooked, and wasted in the Italian context.

Future research avenues

Our results confirm significant health impacts of reducing red meat consumption to the Mediterranean target (150 grams/week/person) and suggest also important environmental gains. Although using scenarios is a simplification, it opens the way to studies investigating more in-depth alternatives to red meat in the dietary pattern, that can account not only for quantities and nutritional balance but also for costs, acceptability, and regional culinary traditions. Also, different age groups have different dietary needs and substitution patterns, making further study necessary.

Policy implications

The main public health message from this study is that small changes in habits towards red and processed meat, towards higher quality and smaller quantities, can maximize health impacts and minimize the environmental burden in terms of GHG emissions. Due to the heterogeneous health and environmental impact, policy recommendations will need to target specific Italian areas. Overall, the reductions in red meat consumption we propose are compatible with a Mediterranean-type diet, which indicates a variety of proteins (meat and poultry, fish, milk and dairy products, legumes) and does not suggest radical choices, such as the adoption of a vegetarian or vegan diet [27]. For example, diets with fewer animal-sourced foods typically include more legumes, nuts and whole grains which evidence suggests have health benefits and are likely to increase the number of avoided deaths [31, 40, 42, 43].

Possible interventions to reduce emissions from livestock production are based on technologies and practices that improve production efficiency, but mitigation and adaptation options have to include public health policies to reduce meat consumption. In practice, people may be informed and invited to reduce portion size or frequency of consumption of red meat and processed meat. However, how to reach these goals is challenging and should take into account factors influencing food choice, and the external economic and social context in which choices are made. One possible solution is pricing food-related GHG emissions that have been proved to be effective in Australia [19]. Such policies need to address health inequalities by income since poorer people are more prone to make unhealthy choices and are more resistant to possible changes than richer subgroups. The Australian study partly addressed this issue by proposing the use of taxation of GHG emissions from red meat and other foods for health promotion in the poorest population subgroups [19]. Another information action is dietary guidelines to raise awareness of population of environmental impacts of the so-called “discretionary” food, emphasizing that following guidelines is good not only for one’s health but also for the environment [37]. Examples of interventions focusing on achieving a full healthy diet approach rather than on a single objective such as reducing red meat consumption are becoming widespread especially in the US, but also in Europe. These interventions promote healthy eating through educational interactive material (website, app for mobile phones). To address low income families, specific tools are used, such as meetings, support groups, and, more recently, food subsidy programmes (e.g. food vouchers) that have found to improve diet quality in the UK [52].

Dairy and beef fat typically contains around 3–6% TFA (wt. % of total fatty acids), while levels in lamb and mutton can be somewhat higher

The TFA content of fat from pork and poultry are generally below 1% of total fatty acids, but the content may vary, primarily depending on the TFA content of the feed (Aro et al., 1998a).

Among the specific efforts to be set up in this field, there is the need to raise awareness of the environmental impact of different kinds of food, to promote model food consumption which result in the lowest greenhouse gas emissions and sufficient, balanced amounts of protein and micronutrients, and which may be more sustainable.

Experts have highlighted that climate change mitigation policies are difficult to implement in part because their effects are not geographically limited, and also because they will have impact decades from now. Empowering both the general population and policy makers is then a crucial goal [41].