From dearth to excess: the rise of obesity in an ultra-processed food system

Abstract

More people now have obesity than suffer from starvation thanks to our modern food system. Agriculture was transformed over the 20th century by a variety of technological advancements that relied heavily on fossil fuels. In the United States, government policies and economic incentives led to surplus production of cheap inputs to processed food industries that produced a wide variety of heavily marketed, convenient, rewarding, timesaving, and relatively inexpensive ultra-processed foods. The energy available in the food supply increased by much more than the population needs, albeit with large inequities in nutrition security. While most of the rise in per capita food availability during the late 20th and early 21st centuries in the United States resulted in increased food waste, a variety of mechanisms have been proposed by which changes in the increasingly ultra-processed food environment resulted in excess energy intake disproportionately in people genetically susceptible to obesity. As populations continue to grow, substantial investments in coordinated nutrition and agricultural research are needed to transform our current food system to one that relies less on fossil fuels, preserves biodiversity, ensures environmental health, and provides equitable access to affordable, safe and nutritious food that reduces the prevalence of chronic diet-related diseases like obesity.

This article is part of a discussion meeting issue ‘Causes of obesity: theories, conjectures and evidence (Part I)’.

1. Introduction

Producing enough food has challenged humans for millennia. The 18th century English economist Thomas Malthus made dire predictions that population growth will outpace agriculture's ability to provide sufficient food thereby resulting in mass starvation [1]. Echoes of what has become known as the ‘Malthusian trap’ have reverberated for centuries since. For example, in the late 1960s and early 1970s the American biologist Paul Ehrlich predicted that the United States of America (USA) would experience famines killing millions due to mass starvation by the middle of the 1980s [2]—a period that was instead marked by a substantial increase in obesity prevalence [3]. Indeed, more people around the world now have obesity than suffer from hunger [4]. How did this happen? Here, I sketch how the development of the industrialized food system has thus far avoided the Malthusian trap and instead may have caused the increase in obesity prevalence. For purposes of concision, this review will focus on the USA, but food systems are now multinational and parallels exist in other industrialized nations with important implications for emerging markets.

2. The rise of the industrialized food system

Agriculture has relied on hard physical work of humans and draft animals for millennia. Seed drills mechanized planting in the 18th century and steel plows were introduced in the 19th century, but it was not until the twentieth century when agriculture was transformed by machines powered by fossil fuels replacing the vast majority of muscular farm work [5–8]. Mass mechanization of farming dramatically improved efficiency such that more land could be farmed by fewer farm workers and without the need to also produce feed for draft animals.

Fossil fuels also transformed agriculture by indirectly increasing fertility of the soil. Specifically, soil nitrogen content sets a limit on the rate of plant protein synthesis and growth because nitrogen is an essential component of all amino acids. Atmospheric nitrogen is naturally fixed in the soil by bacteria that synthesize ammonia which is readily absorbed by plants. Increasing plant growth could be achieved by supplementing soil nitrogen with application of manure and guano, but the availability of sufficient natural fertilizers to sustain agricultural production was limited. In the early 20th century, German chemists Fritz Haber and Carl Bosch invented a way to artificially synthesize ammonia from atmospheric nitrogen at an industrial scale thereby lifting the nitrogen limitation for agriculture. Unfortunately, the Haber-Bosch process traditionally uses a substantial amount of fossil fuel [7,8].

Twentieth century expansion of agricultural productivity and markets in the USA was also facilitated by public investments in infrastructure, especially irrigation and transportation networks. Agricultural research resulted in the development of hybrid seeds for grains like corn, wheat and soybeans that capitalized on the improved irrigation and fertilization of the soil to dramatically increase crop yields and select for varieties amenable to efficient mechanized harvesting. An industrial chemical revolution also produced new classes of pesticides, fungicides and herbicides to further increase productivity [5–9].

These technological advancements have been collectively called the ‘Green Revolution’ and are credited for avoiding the dire Malthusian predictions. In the past several decades, advancements in molecular biology have resulted in the introduction of genetically modified organisms to agriculture in the hopes of further increasing productivity and improving nutritional quality of certain crops.

The grain commodity crops that have most benefited from the ‘Green Revolution’ provided more than enough protein and calories to feed the growing population without substantial increases in land use. But food is more than protein and calories and the vast majority of commodity crops are not directly consumed by humans. Rather, most are used as feed for cows, pigs and chickens that are subsequently consumed by humans. Animal agriculture became increasingly dependent on grains, antibiotics and confined feedlots to increase the growth efficiency of animals and bring them to market sooner. Furthermore, such practices allowed the animals to be located nearby slaughterhouses and transportation networks facilitated their efficient delivery to market [5,6].

Government policies in the USA incentivized farm specialization, increased commodity crop production, and promoted larger farm sizes. Indeed, the number of farms and people working on farms has decreased substantially over the past several decades, and draft animals are now practically nonexistent [6]. As a result, much of the rural population moved to cities and suburbs whose residents rely on others to produce their food.

Improvements in transportation networks, manufacturing methods, food packaging and preservation technologies enabled the rise of large national and multi-national food brands. These processed food and beverage manufactures relied on the inexpensive agricultural grain inputs and their food scientists invented ingenious formulation and processing methods to ‘add value’ to their products [9]. Products were fortified with vitamins and minerals, flavours were enhanced with additives, textures were engineered to improve ‘mouth feel’, and other technologies were applied to increase the overall appeal of their products that were sold for a considerable profit [10,11].

Processed food companies grew at an astonishing rate in the 20th century. Investments in research and development resulted in an ever-expanding range of new products that were more convenient, time-saving and tasty. Intense competition in the marketplace resulted in a ‘survival of the fittest’ products along with directed evolution of new products that were iteratively more and more successful. When a sufficient market of consumers objected to certain products on offer, industry responded by offering a new line of ‘gluten free’, ‘low fat’, ‘low carb’, or other products thereby capturing additional market share.

While food and beverage industries have a guaranteed market because we all need to eat, these industries are faced with a fundamental limitation when it comes to their growth: Engel's law states that a person's spending on food decreases as a percentage of their increasing income, in part, because food requirements are physiologically limited. Thus, even in a growing economy, profit growth of food and beverage companies is reliant on increasing the size of their market to a greater extent than other industries. Shareholders of individual food and beverage companies demand increased profits which incentivizes aggressive marketing campaigns to replace traditional foods and incentivizes development of a wide variety of products that can be easily consumed above physiological needs. Increasing portion sizes appeal to the consumer's sense of value for money and promotion of snacking behaviour alters social norms of eating and further increases opportunities for consumption [12].

Beyond increasing food and beverage consumption in existing markets, another way for processed food companies to increase profits is to enter new markets and displace traditional foods. For example, in 2010 the world's largest food manufacturer Nestlé issued a press release announcing the launch of a ‘floating supermarket’ on the Amazon River stocked with Nestlé products and touted the project as an ‘Unprecedented Business Model [that] will extend the company's presence in Brazilian households’ [13].

3. Ultra-processed foods

Even before the launch of Nestlé's ‘floating supermarket’, Carlos Monteiro documented an alarming increase in industrially processed food products entering the Brazilian market and displacing more traditional foods [14]. To better document and measure this shift in the food supply, in 2010 Monteiro introduced a new food categorization system called NOVA that classified foods based on the extent and purpose of processing [14]. Originally containing only three categories [14], the NOVA system was later modified to include four categories [15]. Category 1 included unprocessed or minimally processed foods such as whole foods that have been washed, chopped, frozen, dried or fermented; category 2 was processed culinary ingredients such as sugar, butter, cooking oils and salt; category 3 was processed foods that were combinations of category 1 and 2 foods for purposes of preservation or to increase palatability; and category 4 was the so-called ‘ultra-processed foods’ whose definition has been the subject of much debate that has prompted clarifications [16]. The most recent definition of ultra-processed foods is that they are ‘industrial formulations manufactured by deconstructing foods into their component parts, modifying them and recombining them with a myriad of additives and little, if any, whole foods. Importantly, ultra-processed foods are distinct not only in terms of their ingredient composition but also in terms of the purpose for which they are produced. The purpose underlying the manufacture of ultra-processed foods is to create convenient (durable, ready-to-consume), tasteful (often hyper-palatable) and highly profitable (cheap ingredients, value adding) products that are liable to displace all other NOVA food groups’ [17].

Food and nutrition professionals have expressed confusion about the concept of ultra-processed food which is often conflated with ‘processed food’ [18] thereby ignoring the ‘purpose of processing’ criterion as an important part of the NOVA category 4 definition. While manufacturers and producers of all NOVA food categories intend to generate profits, ultra-processed foods have a uniquely wide scope for extensive reformulation to meet market demands and potentially displace consumption of foods from other categories. Thus, some have argued that ultra-processed foods would be better described as ‘ultra-formulated’ foods [19].

The NOVA classification system has been criticized [20–27]. Some critics have argued that the NOVA system is unclearly defined and difficult to implement resulting in substantial discrepancies between different people asked to categorize the same foods according to the NOVA system [28]. However, other studies have found more consistent NOVA ratings between people [29] and standardized approaches to applying the NOVA system are being implemented [30].

Another criticism is that the NOVA system offers no meaningful benefits over traditional nutrient-based classification systems because diets high in ultra-processed foods also tend to have greater amounts of salt, sugar and fat while being lower in protein and fibre [20]. However, the NOVA system largely ignores the nutrient content of foods and rejects the dominant ‘nutritionism’ paradigm that the nutrient composition of foods is the main determinant of their impact on health [31]—an approach with a long history of success in nutrition science [32]. Nevertheless, despite widespread use of ‘nutritionism’-based approaches to communicating about healthy and unhealthy foods or dietary patterns, many consumers appear to understand the basic advice to minimize ultra-processed food consumption more easily than recommendations to reduce intake of precisely defined nutrients such as sodium, saturated fat or sugar.

Official dietary guidelines of health organizations and nations have begun to include statements to limit consumption of ultra-processed foods [33] because diets high in ultra-processed foods have been linked to a variety of poor health outcomes, including obesity [34–36]. The food industry is especially concerned about widespread adoption of the NOVA system because reformulating products out of the NOVA 4 category will be very difficult. Furthermore, the NOVA 4 definition currently represents such a wide variety of products that there is a risk of enacting ineffective and regressive policies without targeting such policies based on a better mechanistic understanding of how ultra-processed foods cause adverse health outcomes [37]. By subcategorizing NOVA category 4 foods based on understanding biological mechanisms, it may be possible to help reformulate ultra-processed foods and provide effective consumer advice on the basis of their individual risk of a disease or health condition such as obesity [37].

4. Mechanisms by which ultra-processed foods promote excess energy intake and obesity

Within a given food environment, individual susceptibility to obesity is largely explained by genetic differences, particularly genes preferentially expressed in the brain and likely involved in appetite control [38]. The changing food environment is thought to explain the increasing prevalence of obesity in the population, and our modern industrialized food system has provided excess calories in the form of increased availability and marketing of inexpensive and convenient ultra-processed foods.

There is much debate about the most important factors in our changing food environment that are responsible for the increase in obesity prevalence. Associations between ultra-processed food intake and obesity observed in epidemiological studies [34–36] do not necessarily imply that ultra-processed foods cause excess intake and weight gain. We recently conducted a controlled randomized crossover trial to investigate this issue and found that a diet high in ultra-processed foods caused an increase in ad libitum energy intake of approximately 500 kcal d⁻¹ and weight gain in adults as compared to an unprocessed diet matched for overall presented calories, carbohydrate, sugar, fat, sodium, energy density, glycemic load and fibre that resulted in spontaneous weight loss [39].

A wide variety of mechanisms, with varying degrees of plausibility, have been proposed to explain why diets high in ultra-processed foods are consumed in caloric excess and may lead to obesity [27,33]. Many potentially important mechanisms involve cost, convenience, preparation time, availability, variety, portion size and marketing that affect food choice and frequency of consumption. However, these mechanisms were not at play in our randomized trial because participants were freely provided with all their meals and snacks in fixed portions and ready to consume. Below, I will focus mainly on the mechanisms that could have played a role in our study.

One highly plausible mechanism by which ultra-processed foods may promote obesity involves their typically high energy density which is a property that is well-known to positively affect ad libitum energy intake [40]. While the overall energy densities of the ultra-processed and minimally processed diets in our study were matched, this was achieved primarily by including more low-calorie beverages with the ultra-processed meals to deliver soluble fibre supplements to match the overall fibre content of the minimally processed diet [39]. Non-beverage energy density differed substantially between the ultra-processed and minimally processed diets, primarily because of decreased water content in the ultra-processed foods. In a secondary analysis of the meal intake data, non-beverage energy density was an important variable positively predicting meal energy intake and mediating a substantial fraction of the effect of ultra-processed versus minimally processed meal energy intake in our trial [41].

Eating rate has been implicated as a mechanism that positively promotes energy intake and is related to the sensory properties and structure of the foods [42–44]. Ultra-processed meals in our study were consumed more quickly than the minimally processed meals when including the masses of the beverages, but non-beverage items in the ultra-processed meals were eaten more slowly than in the minimally processed meals [41]. In a secondary analysis, non-beverage eating rate was a positive predictor of individual meal energy intake in both the ultra-processed and minimally processed diet [41].

While the overall diets were matched each day for sodium, sugar, carbohydrates and fat in our trial, the individual meals differed in the proportion of calories presented as hyper-palatable foods defined as having high amounts of both sodium and fat, sodium and carbohydrates, and sugar and fat [45]. Indeed, neuroimaging studies have suggested that hyper-palatable foods high in sugar and fat are particularly reinforcing [46]. In a secondary analysis of our ultra-processed versus minimally processed diet study, we found that meal energy intake was positively influenced by hyper-palatable foods as a proportion of calories presented [41]. We also found an interesting interaction between non-beverage meal energy density and hyper-palatable foods such that the effect of hyper-palatable foods was greater for meals with lower energy density and vice versa. Interestingly, the participants in our trial did not report significant differences in palatability between the ultra-processed versus minimally processed meals [39]. However, palatability is only a single dimension of the multifaceted concept of food reward and ultra-processed meals may have increased incentive salience, wanting and motivation to eat that may have driven excess energy intake [47].

Some investigators have hypothesized that ultra-processed foods contain highly processed carbohydrates that are quickly absorbed causing rapid increases in circulating glucose thereby enhancing reward-related brain activity via mechanisms similar to addictive substances [48,49]. The subsequent fall in circulating glucose may also result in late-postprandial glucose dips that induce hunger [50]. While the ultra-processed and minimally processed diets used in our trial varied in added sugars, the diets were similar in glycemic load and resulted in no significant differences in mean glucose or glucose variability as measured using continuous glucose monitors [39,51].

Ultra-processed foods tend to have lower protein content [52] and the protein leverage model of obesity suggests that diets with diluted dietary protein result in excess energy intake to achieve a given protein target [53,54]. While our ultra-processed and minimally processed diets differed slightly in the proportion of presented calories as protein (14% versus 15.6%, respectively) we calculated that protein leverage could not fully explain the observed differences in mean energy intake [39,55]. Indeed, in a secondary analysis of our ultra-processed and minimally processed meal data we found that more protein offered with meals significantly increased energy intake [41].

A variety of gut signals are modulated in response to food ingestion and are believed to play important roles in appetite control [56]. The production of ultra-processed foods disrupts the natural food matrix of the whole food ingredients and may result in altered intestinal digestion and absorption of nutrients [57]. For example, if nutrients from ultra-processed foods are absorbed more proximally in the gut they may instigate a different pattern of gut–brain signals as compared to the same nutrients ingested in the form of minimally processed whole foods. These gut–brain signals include vagal afferents, hormones derived from enteroendocrine cells or signals from the intestinal microbiota. Enteroendocrine cells and microbiota may also be negatively influenced by low amounts of insoluble fibre or various emulsifiers or additives in ultra-processed foods. Another ingredient of ultra-processed food involves the materials used in packaging which may alter the endocrine system and act as obesogens [58,59].

Ultra-processed foods are often stripped of bioactive compounds that are found in whole foods while additives, such as artificial flavours, included in their formulations may have unintended biological activity. It has been proposed that the relative concentrations of nutrients, flavours and bioactive compounds that were previously reliable over evolutionary time and were used to guide our food intake to meet physiological demands have become mismatched by the rise in ultra-processed foods and have thereby resulted in aberrant food intake behaviour promoting weight gain [47,60]. For example, the increasing prevalence of ultra-processed foods in the food supply has gone along with an increase in fats derived from processed seed oils which have a high proportion of omega 6 to omega 3 unsaturated fatty acids that have been suggested to promote obesity through various mechanisms, including influencing the endocannabinoid system to increase appetite [61–63].

5. Conclusion

Through a variety of technological advancements, government policies and economic incentives, the current food system in the USA relies heavily on fossil fuels to generate a huge surplus of calories available in a wide variety of heavily marketed foods that are inexpensive, convenient, ubiquitous, energy dense and often hyper-palatable. This flow of matter and energy through the industrialized food system in the USA has increased over the past half century on a per capita basis in correspondence with the rise in obesity prevalence [3] and an even larger increase in food waste [64].

We find ourselves in a unique position in human history where relatively affordable food is produced in excess by relatively small numbers of people. Whereas food purchases represented approximately 60% of disposable income at the beginning of the 20th century in the USA, now food costs less than 10% of disposable income [12]. The benefits of these changes include the fact that more food is prepared outside the home and cooking meals at home is more convenient thereby freeing up time and money for other pursuits.

It is intriguing to speculate that increased obesity prevalence has been an inevitable result of striving to create a food system that avoided Malthus' predictions. In other words, society incentivized surplus agricultural production (relying on cheap fossil fuels) to provide low-cost inputs to food and beverage industries that produced heavily marketed, convenient, rewarding, timesaving and relatively inexpensive ultra-processed foods in great excess of consumption needs of the population, albeit with large inequities in food distribution and nutrition security [4].

We are now faced with a huge challenge and opportunity. As populations continue to grow and agriculture faces the added pressures of tackling climate change and loss of biodiversity, substantial investments in coordinated nutrition and agricultural research are urgently needed to transform the current food system to one that relies less on fossil fuels, ensures environmental health, and provides equitable access to affordable, safe and nutritious food that reduces the prevalence of chronic diet-related diseases like obesity.

Data accessibility

This article has no additional data.

Authors' contributions

K.H.: conceptualization, writing—original draft, writing—review and editing.

Conflict of interest declaration

I declare I have no competing interests.

Funding

This work was supported by the Intramural Research Program of the National Institutes of Health, National Institute of Diabetes and Digestive and Kidney Diseases.