Reducing GHG emissions while improving diet quality: exploring the potential of reduced meat, cheese and alcoholic and soft drinks consumption at specific moments during the day

The typical Western diet has a high environmental impact and is associated with obesity and other diet-related diseases, and therefore it can currently not be called sustainable. In the Netherlands, 50% of the adult population is overweight or obese [1]. The resulting chronic lifestyle diseases, such as cardiovascular disease and type 2 diabetes, cause major declines in quality of life and are primary causes of death [2, 3]. Like in many other countries, in the Netherlands the intake of fruits and vegetables is too low while the intake of saturated fatty acids (SFA) and sodium is too high [4, 5].

Food production and consumption systems are important drivers of greenhouse gas (GHG) emissions, with up to 30% of the worldwide emission of greenhouse gases being related to the production and consumption of food [6, 7]. Increasing the efficiency of food production systems is not enough to reduce GHG emissions to acceptable levels: our consumption patterns will have to change as well [6, 8]. Shifting to healthier diets will only result in reduced GHG emissions if consumers choose healthy foods with a relatively low impact on GHG emissions [9, 10]. At a comparable level of dietary energy intake, GHG emissions of diets may vary substantially. For example, diets containing 2000 kcal per day may vary from ~ 2 to > 12 kg CO₂-equivalents (CO₂-eq) per day [11].

Animal-based foods are generally the most important contributors to resource use and dietary GHG emissions [12]. This is also true for Dutch food consumption patterns [13], where almost half of the GHG emissions of food consumption can be attributed to meat and dairy consumption. The total consumption of all beverages also contributes considerably to dietary GHG emissions [13, 14]. Previous research in the Netherlands has already quantified the potential effects of replacing meat and dairy with plant-based foods on GHG emissions and nutrient content of the diet, for the general adult population [15]. Especially the diets of those with high dietary GHG emissions hold a large potential for improvement. To our knowledge, little research has been done to estimate the effects of dietary changes to lower GHG emissions in this specific group, on GHG emissions as well as intake of selected nutrients. Since people with relatively high dietary GHG emissions also have significantly higher energy intakes [13], reducing the intake of high GHG emission foods such as meat might be preferable over replacing these foods. If there is overconsumption in this subgroup of the population, reducing food (and therefore energy) intake could be beneficial for health as well as the reduction of GHG emissions.

This scenario study aims to provide suggestions to reduce the GHG emission of diets with high associated GHG emissions, while simultaneously improving the nutrient content of these diets. The reference scenario is the diet of people in the highest tertile of dietary GHG emissions in the Netherlands. For specific moments of consumption during the day, the intake of foods with the highest associated GHG emissions is either reduced or replaced with healthier alternatives with lower GHG emissions, by means of several scenarios. The results of this study can provide valuable knowledge for the design of policies and interventions directed at stimulating healthy dietary patterns with reduced GHG emissions, for people with relatively high dietary GHG emissions.

In the highest tertile of dietary GHG emission for women, age was slightly higher compared to the intermediate and low tertile of dietary GHG emission (Table 2). There were more non-Dutch participants in the lowest tertile of dietary GHG emission for women. The proportion of participants with overweight (including obesity) was not significantly different between the tertiles of dietary GHG emission, for men as well as women. There were no significant differences in BMR between groups. Men in the lowest tertile of dietary GHG emission were less physically active (based on the MET score) than participants in the other tertiles, but the MET scores of men in the intermediate and high tertile were not significantly different. There were no significant differences in MET scores for women in the three tertiles of dietary GHG emission.

This study aimed to provide suggestions for dietary change to reduce GHG emissions and at the same time improve nutrient quality of diets associated with high GHG emissions among Dutch adults. Several dietary changes at specific moments of food consumption can reduce GHG emission of these diets by up to one third, while lowering energy, SFA and sugar intakes, but also iron intakes. Considering that 54% of men and women in our study population were overweight, reduced energy intake will be beneficial for most, but for part of the population the scenarios could lead to energy and iron intakes below requirements. In those cases, increased consumption of low GHG emission and nutritious replacement foods will be needed.

In the Dutch diet, the GHG emission varies per consumption occasion. In the highest tertile of dietary GHG emission, dinner is associated with the highest GHG emission, and consumptions in between meals are associated with the second-to-highest emissions. Meat consumption contributes most to the GHG emission of dinner, while beverages contribute most to the GHG emission of consumptions in between meals. Dairy (including cheese) is an important contributor to total daily dietary GHG emission throughout the day.

The most effective single dietary change to reduce GHG emissions in this study was to reduce the consumption of meat, with a 75% reduction in red and/or processed meat during dinner resulting in a 24% reduction in GHG emissions for men and a 22% reduction for women. Average daily meat consumption in the highest tertile of dietary GHG emission of our study population was 185 g for men and 119 g for women, of which almost 90% was red and/or processed meat. This is well above the recommended maximum intake of 500 g meat per week (or 71 g per day), of which a maximum of 300 g (or 43 g per day) may be red meat (60%), according to the current Dutch food based dietary guidelines [26]. Reducing the consumption of meat is therefore in line with the food based dietary guidelines. In the ‘meat75’ and ‘combimax’ scenarios, remaining weekly meat consumption is 700 g (of which 567 g red and/or processed meat) for men and 420 g (of which 308 g red and/or processed meat) for women. Meat consumption of women is slightly lower than the recommended maximum in the food based dietary guidelines, but protein intake remained adequate. The ‘meat75’ and ‘combimax’ scenarios resulted in decreased saturated fatty acid intakes. Reducing the consumption of red and/or processed meat may also lower the risk for chronic diseases, as research shows that every 50 g portion of processed meat eaten daily increases the risk of colorectal cancer by about 18% [27]. In the scenario that reduced meat consumption during dinner with 75%, the average consumption of red and/or processed meat was reduced with 85 g/day for men and 59 g/day for women.

Replacing the consumption of all alcoholic drinks and soft drinks (including mineral water and fruit and vegetable juices) with tap water was also successful in reducing dietary GHG emission. At the same time, the intake of energy and mono- and disaccharides (as proxy for sugar) was reduced significantly and for women there was also a significant and relevant reduction in iron intakes. This can be explained by the fact that women drink relatively more iron-containing drinks. For example, 61% of all alcoholic drinks consumed by women is wine, while for men 87% of alcoholic drinks consumed is beer. According to the food composition table used for this study, beer contains no iron while wine contains 0.6 mg per 100 g. Women also drink relatively more fruit juices (containing some iron) while men drink more other soft drinks. However, because the consumption of both alcoholic drinks and fruit juices is discouraged in dietary guidelines, these foods are not the most appropriate dietary iron sources. Dietary iron could be obtained from types of foods with a more healthy nutritional profile, containing other essential nutrients while the content of nutrients that are considered unhealthy is low. When drinks were partly (50%) replaced, iron intakes were not compromised and there was a significant decrease in mono- and disaccharides intakes for men and women, while energy intake was reduced for men only. This can be explained by the fact that men have a higher consumption of these drinks in the reference scenario. The ‘water50’ scenario did however not lead to significant GHG emission reductions. The consumption of beverages is highest in between meals. Consuming water instead of sugar-containing or acidic beverages in between meals has the additional benefit of a less frequent exposure to foods that can damage tooth enamel (when tap water is consumed without food) [28].

Our scenarios show that reducing the consumption of soft and alcoholic drinks and of red and processed meat will lead to lower dietary GHG emissions, even if meat intake is only reduced at dinner. These dietary changes are also beneficial for health because they bring food consumption more in line with food based dietary guidelines [26] and will reduce SFA and sugar intakes. However, these changes will also lead to lower energy and iron intakes. Considering that 54% of men and women in the highest tertile of dietary GHG emissions were overweight or obese (see Table 2), reduced energy intake is beneficial for the majority of the population under study. Estimated average energy requirements as previously calculated [29] for the Dutch food based dietary guidelines [26] range from 2430 kcal/day to 2790 kcal/day for men aged 19–69 years, and from 1790 kcal/day to 2020 kcal/day for women aged 19–69 years. The mean energy intake in most of our scenarios is still higher than these average requirements. However, for part of the population the scenarios could lead to energy intakes below requirements. We were unable to assess energy adequacy and thus the proportion of the population with an intake below the requirement, since we could not determine individual energy requirements sufficiently accurate from the DNFCS 2007–2010 data available. Individual energy requirements may vary widely depending on BMR and level of physical activity. Iron adequacy could also not be assessed: the EAR cut-point method cannot be used because the requirement distribution for iron is not symmetric [25]. Because the iron requirement distribution is unknown for our study population, it was also not possible to use the probability method to assess adequacy [25]. The Nordic Council [30] has established an EAR for iron of 7 mg/day for men aged 19–69, 10 mg/day for women aged 19–50 and 6 mg for women aged 51–69 (postmenopausal women). Mean iron intake of men in our scenarios is well above the EAR, and for women mean intakes are close to the EAR for menstruating women. Since mostly women of childbearing age are at risk of having inadequate iron intakes [4], iron intake reductions as observed in our scenarios are undesirable for this subgroup. However, because iron inadequacy already occurs in this subgroup [4], the scenarios in this study might increase the proportion of women with inadequate iron intakes but do not introduce a new problem. Effort is needed to ensure adequate iron intakes of women of childbearing age, regardless of whether meat and alcoholic and soft drink consumption are maintained or reduced. For people with energy and/or iron intakes below requirements as a result of the dietary changes in our scenarios, the reduced meat and soft and alcoholic drinks intake should be compensated for by increased intake of other low GHG emission, healthy foods. Previous research in the general Dutch population has shown that replacing meat and dairy with plant-based alternatives can significantly reduce GHG emissions while increasing estimated iron intakes [15]. However, the iron from plant-based sources is of lower bioavailability and future research should take this into account in order to determine if iron intake from plant-based foods replacing meat is adequate to meet requirements [31].

Combining the meat reduction and drinks replacement scenarios into the ‘combimax’ scenario leads to the highest reductions in dietary GHG emission. The reduction in mean habitual daily dietary GHG emission is 34% for men and 31% for women in the ‘combimax’ scenario. These combination scenarios also include the replacement of all cheese consumed in between meals by plant-based alternatives. This individual change led to a small but non-significant decrease in dietary GHG emission. The GHG emission reductions in this scenario are therefore mainly attributable to the reduction in meat consumption and the replacement of alcoholic drinks and soft drinks. These results are in line with previous research. Green et al. have shown that for a UK population, the consumption of meat (especially red meat) and soft drinks should decrease in order to achieve GHG emission reductions [14]. Similarly, Horgan et al. showed that for sustainable diets, the consumption of soft drinks and red and/or processed meat should be reduced in the diets of most individuals in their UK study population [9]. Previous research for the Dutch situation has shown that vegetarian and vegan diets are most effective in reducing GHG emissions [32, 33].

What makes our approach novel is that we have taken consumption occasion into account. Of all meat consumed, the vast majority is consumed during dinner. Important reductions in dietary GHG emission as a result from reduced meat consumption can be achieved by focussing only on that consumption occasion. Previous research in the Dutch population has shown that willingness to change the diet is generally low, and that people are most willing to adopt changes that do not cost money and that are relatively effortless [34]. Both reducing meat consumption during dinner and replacing alcoholic drinks and soft drinks by tap water in between meals will save money. Both changes are also less drastic than eliminating meat from the diet entirely or replacing all drinks (including coffee and tea) by tap water, which might make it relatively easy to implement these changes. Indeed, Sijtsema et al. have showed that willingness to limit meat consumption is higher than willingness to change to a vegetarian diet [34] The first results from the Dutch National Food Consumption Survey 2012–2016 show that there has been a decrease in meat consumption in the Netherlands [35], so possibly further reductions in meat consumption would be acceptable to the population. Consumption of alcoholic drinks has decreased as well, while intake of non-alcoholic drinks increased. However, the latter was caused by increased consumption of tea, coffee and water [35]. These changes are in the direction also proposed in the scenarios in this study. However, it is unknown if the reductions in meat and soft and alcoholic drinks consumption in the scenarios are indeed acceptable to the subgroup of the population included in this study. Taste and food preferences were not assessed. Also, if reduced meat intake during dinner would be compensated for by increased intakes of meat during other moments of food consumption, this will affect the GHG emission reductions. Similarly, if a reduced consumption of meat during dinner would be compensated for by increased consumption of other foods during dinner, it depends on the GHG emissions associated with the replacement foods what the final effect on dietary GHG emissions will be.

The strengths of this study include the focus on GHG emissions as well as nutrient content of the diet, the fact that the scenarios are based on current food consumption using data from the most recent DNFCS (based on a large and representative group of Dutch adults), and the fact that the suggested dietary changes are based on the highest potential GHG emission reductions at specific moments of food consumption.

A limitation of our study could be that we relied on GHG emission data only. While this indicator is highly correlated to some other indicators of environmental impact [36], it does not mean that a reduction in GHG emissions leads to similar reductions in all other indicators of environmental impact. Additional research would be needed to quantify the effect of the scenarios included in this study on other indicators of environmental impact, such as land or water use. Future research would also be necessary to determine how the subgroup of the population included in this study could be moved to adopt the dietary changes proposed, as it is very difficult to achieve dietary change. A long-term approach is needed to change consumer values, and a positive attitude towards sustainable food consumption may not be enough to change dietary behaviour [37]. Health-related arguments may be more effective than environmental motives to promote sustainable consumption [38], but coupling these arguments in communication strategies or in dietary guidelines may be worthwhile since both arguments may appeal to different segments of the population. As this study shows some dietary changes will be beneficial for both health and the environment, at least for people in the highest tertile of dietary GHG emissions. Therefore, an effort should be made to further integrate environmental sustainability into dietary guidelines. Even more than individual approaches, there is a need for social and institutional changes that facilitate environmentally friendly food consumption [37].

The scenarios in this study were focussed on people in the highest tertile of dietary GHG emission, since that subgroup of the population holds the largest potential for dietary GHG emission reductions. Previous research has shown that for the entire Dutch population, meat and cheese combined are the largest contributor to dietary GHG emission, followed by drinks [13]. Out of all types of drinks, soft drinks are an important contributor to GHG emission for the whole population, and alcoholic drinks are an important contributor to GHG emission of adult men specifically. It is, therefore, expected that the dietary changes proposed in this study will also lead to reductions in the GHG emission of people in the lowest or intermediate tertile of dietary GHG emission. However, the effect of the scenarios included in this study on nutrient intakes of the general population are unknown, and therefore the results of this study cannot be extrapolated to other groups. Additional research will be necessary to quantify how the scenarios in this study affect nutrient intakes of people with lower dietary GHG emissions.

Reducing the consumption of red and processed meat during dinner and of soft and alcoholic drinks throughout the day leads to significantly lower dietary GHG emissions of people in the Netherlands in the highest tertile of dietary GHG emissions, while also having health benefits. For subgroups of the population not meeting energy or iron requirements as a result of these dietary changes, low GHG emission and nutritious replacement foods might be needed in order to meet energy and iron requirements.