Here, we discuss the link between nutrition, non-communicable chronic diseases and socio-economic standing, with a special focus on the microbiota. We provide a theoretical framework and several lines of evidence from both animal and human studies that support the idea that income inequality is an underlying factor for the maladaptive changes seen in the microbiota in certain populations. We propose that this contributes to the health disparities that are seen between lower-income and higher-income populations in high-income countries.
A prominent paradigm that has developed to explain the complexity of human health involves the mammalian gut microbiota — that is, the intricate community of microorganisms that live within the gastrointestinal tract1. In humans, microbial cells are estimated to roughly equal the number of total human cells (3.8 × 1013 to 3.0 × 1013) or to outnumber nucleated human cells (excluding erythrocytes) by approximately 10:1 (Ref. 2). The gut microbiota affects digestion and absorption by breaking down otherwise indigestible carbohydrates3,4,5, and it protects against pathogens by competitive exclusion6,7. Disruptions in the microbiota, which can occur following antibiotic administration or exposure to persistent genetic or environmental stressors or during disease, can affect the long-term health of both the microbiota and the host8. Moreover, we now appreciate that the intestinal microbiota influences not just the health of the gut itself9 but also health in distal tissues, such as the liver10, the joints11 and the brain12. However, what is less often discussed is how these processes are influenced by the social context in which an individual resides.
The composition of the microbiota can be influenced by host genetics13 and early-life imprinting14,15,16,17 (Box 1), but diet has an inescapably important role in this18,19. The past century has seen a dramatic shift from farm- and home-based eating to industrialized food processing20, an increase in convenience foods and snacks21 and a surge in eating restaurant-prepared foods22. In recent decades, our diets have shifted to higher ratios of processed fats and sugars to fresh foods, and these patterns are influenced heavily by socio-economic status23,24. Individuals with greater socio-economic hardship are often forced to weigh their food selection by a 'calories-per-dollar trade-off', in which the most accessible foods become those that offer more energy and satiation for a lower price. In high-income countries, these trade-offs often necessarily include refined carbohydrates and processed oils and exclude nutrient-dense fresh foods25. In low-income and middle-income countries, the diets of higher-income populations are starting to mimic the processed food intake of higher-income countries, while malnutrition continues to impact the maturity of the microbiota in lower-income populations. These nutritional and epidemiological transitions are a contributing factor that results in a double burden of malnutrition in low-income and middle-income countries, where changes in the food system and increased availability of processed foods has caused these foods to make their way into the diets of the higher-income populations within these countries26,27,28,29.
In this Opinion article, we suggest that socio-economically driven nutritional differences affect the microbiota and, consequently, the immune system and host health30,31,32,33,34 (Fig. 1). We discuss this idea using data obtained from both human and mouse models of chronic metabolic disease, with a focus on high-income countries. With chronic preventable diseases on the rise, especially among those least able to afford health care, the pandemic of chronic metabolic diseases affecting the poor becomes impossible to ignore.
Socio-economic status and chronic disease
Chronic non-communicable diseases have risen in recent decades at a pace too fast to be accounted for by genetics alone35. In 2013, death estimates and disability-adjusted life years suggested a global predominance of non-communicable causes, such as cardiovascular disease, neoplasms and respiratory, metabolic and neurological disease, many of which have a greater incidence since 2000 (Ref. 36). Low socio-economic status is a risk factor for many of these conditions37. Early associations with socio-economic status and health were observed between income level, income instability and mortality, with low income being strongly predictive of mortality independent of educational status or initial health status38. It was reported that raising the minimum wage by a single dollar predicts a 2% decrease in low-birthweight babies and up to a 4% drop in infant death39. Children living in neighbourhoods associated with greater levels of poverty or overcrowding were three times more likely to be hospitalized for influenza than children from wealthier, less crowded neighbourhoods40. These and numerous other reports have shown that societal standing and income status influence susceptibility to many distinct diseases, as we discuss below.
Obesity and metabolic syndrome. In low-income and middle-income countries, obesity is generally more prevalent in individuals of higher incomes41,42,43. However, in high-income settings like Western Europe and the United States, obesity is more prevalent among people with lower incomes44. Of the risk factors for obesity and metabolic syndrome in this setting, food processing and availability appear to be some of the most influential for disease risk41. Low-cost energy-dense foods that are high in fat or sugar content are consistently linked to the prevalence of obesity45, yet these are the most readily available to impoverished populations of wealthier nations. This group further suffers from the loss of protection from nutrient-rich fresh foods, which may be less affordable46. With globalization of the food system, many low-to-medium-income countries, which are increasing their intake of processed and packaged foods47, are also expected to bear a disproportionate burden of metabolic disease incidence in the coming years48. Studies in these regions have shown that dietary and lifestyle changes could significantly prevent onset in these populations48.
Animal studies have highlighted the impact of dietary changes on the microbiota and host health. Germ-free mice are resistant to weight gain in response to a high-fat diet (HFD)49 and have lower metabolic rates than mice maintained under specific-pathogen-free (SPF) conditions3. The introduction of microbiota from SPF mice to germ-free animals induced a 60% increase in body fat, with increased lipogenesis in the liver by facilitating greater monosaccharide absorption from the gut3. In both mice and humans, obesity has been associated with shifts in major phyla of bacteria from a dominance of Bacteroidetes to a dominance of Firmicutes, although this phenomenon is not universally observed, probably owing to regional differences in diet and environment. A cross-sectional study in a German cohort found that obesity was associated with a reduction in Firmicutes50. A study of obese and overweight subjects indicated that at baseline, a lower gut microbial gene count was related to greater insulin resistance, and levels of serum triglycerides and gene diversity increased after a 6-week energy-restricted high-protein dietary intervention with corresponding improvements in insulin resistance but not inflammation in subjects who started with a low gene count51,52. When microbiota from an obese donor is transferred to germ-free recipient mice, the mice gain more weight than recipients of microbiota from lean donors53. Mathematical modelling of the microbiota is quickly expanding to predict diet–microorganism–metabolism contributions to specific metabolites within the gut and is likely to further advance our understanding of the connections between microbiota and metabolism54.
How the microbiota responds or contributes to the obesogenicity of a particular diet is currently under scrutiny. Serino et al.55 suggested that whether or not mice developed diabetes when fed a HFD was related to the nature of their microbiota. Another study indicated that microbial signatures could predispose human subjects to weight-regain after dieting56. Though HFDs can promote changes within the microbiota, the source and nature of the fat are important16,17,30,55,57,58,59 (Table 1). Mice fed a diet based on saturated milk fat developed an expansion of sulfite-reducing bacteria and an exacerbated T helper 1 (TH1) cell immune response, making them susceptible to colitis57. Furthermore, compared with feeding mice a fish oil-based HFD, a lard-based HFD promoted increased weight gain and inflammatory macrophage recruitment to adipose tissue in a Toll-like receptor 4 (TLR4)-dependent manner31. These differential responses to HFDs indicate that the microbiota may determine the distinct metabolic effects that are seen in response to different fat sources.
Bile acids, which are generated in the liver and secreted into the duodenum following a meal, also modify and are modified by the host microbiota and impact host metabolism. Some microbial families are capable of metabolizing bile acids into downstream signalling molecules, whereas other microbial families are susceptible to the detergent-like properties of bile acids. Thus, fat intake can selectively shape the microbiota through effects on the secretion of bile acids, a topic that has been thoughtfully reviewed elsewhere (Ref. 60).
Food additives, which are ubiquitous among shelf-stable processed foods, may also affect metabolism through effects on the microbiota (Box 2). In one study, even low amounts of dietary emulsifiers — which are extremely common among processed foods — were sufficient to induce changes in the mucus layer of the intestine and provoke low-grade inflammation, making mice susceptible to colitis61. The timing of eating also affects the microbiota, which itself oscillates in composition on a diurnal circadian clock62. Feeding mice a HFD suppressed these oscillations and promoted obesity, whereas time-restricted HFD feeding could partially restore circadian oscillations and attenuate the development of obesity62.
Early disruptions in the establishment of the microbiota may also be obesogenic. In both primates and humans, a maternal HFD during pregnancy has a lasting impact on the microbiota of the offspring even if the mother herself is not obese17. However, leptin, insulin and unfavourable ω-6/ω-3 fatty acid ratios within breast milk were found in mothers with higher body mass index (BMI)63, which could have a lasting impact on the seeding microbiota of the infant gut and immunological training. Administration of antibiotics in the first few years of life has also repeatedly been associated with childhood obesity64,65,66, and maternal antibiotic exposure during pregnancy has been associated with low birthweight, which can be a risk factor for obesity67. Importantly, these disruptions may be more common among low-income populations due to the lower prevalence of breastfeeding68, the misuse of antibiotics and poverty-driven practices such as medication sharing, though more epidemiological studies on the misuse of antibiotics are merited69.
Type 2 diabetes. Between 1980 and 2012, the incidence of type 2 diabetes more than doubled in the United States70, with the alarming appearance of type 2 diabetes among younger age groups71. Alongside obesity, diabetes has been identified as potentially one of the gravest public health epidemics of our time71. On a national level, a 1% increase in gross domestic product (GDP) correlates to a 1.07% rise in diabetes prevalence72. Like obesity, however, type 2 diabetes incidence falls harder on deprived populations in higher-income societies73, even though it has been deemed largely preventable by dietary and lifestyle interventions74. The intractability of diabetes prevalence despite its known preventability suggests an attainability gap in these populations that scientists should not be quick to overlook, especially as we uncover the rippling importance of our diets.
It has been estimated that for every additional 150 kilocalories available per person per day, the world diabetes prevalence increases by 0.1%. However, if those kilocalories happen to come from sugar, prevalence increases by 1.1%, or 11-fold more than for generic calorific consumption72. Furthermore, sugar was the only nutrient to produce a significant correlation to the prevalence of diabetes72, which may not be inherently surprising. Still, sugar-free diet options may not be much better. A recent study showed that non-calorific sweeteners, including the most common diet-soda sweetener aspartame, could still promote glucose intolerance by altering metabolism via effects on the microbiota75. Sucralose (Splenda®) was also included in the study and found to have a similar effect75.
Type 2 diabetes frequently follows a progressive developmental path from insensitivity to insulin to low-grade inflammation and glucose intolerance. It is becoming more evident that just as for obesity, the microbiota is also altered in the progression of diabetes, perhaps as part of the same process. While the obesogenic shifts mentioned earlier are also risk factors for type 2 diabetes, there appear to be some shifts that are diabetes-specific. For instance, in patients with type 2 diabetes, the ratios of bacterial taxa appear to correlate more strongly with blood glucose levels than with BMI, with reductions in the butyrate-producing class Clostridia and expansion of Betaproteobacteria76. One metagenomics study in women with type 2 diabetes found that gene expression patterns in the microbiota could predict a diabetic metabolism, although the markers for European women were region-specific and differed from those for a Chinese population77. Patients with type 2 diabetes display a dysbiosis characterized by an enrichment of sulfate-reducing bacteria and some opportunistic pathogens, and multiple studies have reported a reduction in the frequency of butyrate-producing bacteria76,78. The reported lack of butyrate-producing bacteria could be important in the development of diabetes, as butyrate induces regulatory T cells79 (Box 2), which are known to attenuate inflammation within adipose tissue and reduce insulin resistance in ob/ob mice80.
Cardiovascular disease. Cardiovascular disease (CVD) is another example of a disease affected by socio-economic status. A study among office workers showed that the risk of CVD rose sharply the lower the occupational attainment an individual achieved, with the lowest-level workers surpassing the highest-level workers in CVD risk by a factor of three81. A 2012 study showed that living in a neighbourhood with a low socio-economic profile predisposed healthy Americans to a much greater risk of mortality compared with counterparts in neighbourhoods with higher socio-economic status, independent of their individual socio-economic standing82. An economic downturn during a person's birth years was associated with cardiovascular mortality later in life, suggesting the importance of early years in the development of health disparities83. Higher rates of ischaemic heart disease were found among individuals with low socio-economic status in ten different Western European nations84, and mortality from CVD could be predicted by low socio-economic status among individuals with type 2 diabetes85.
The role of socio-economic status in CVD risk is multifactorial; however, dietary factors and inflammation are two major components. A 15-year follow-up study tracking dietary patterns found that higher dietary quality (for example, more fresh food compared with processed food) could reduce cardiovascular mortality86. It has long been observed that obesity, serum cholesterol and blood pressure serve as major risk factors87, whereas the intake of fruit and vegetables, especially dark leafy greens and cruciferous vegetables, is associated with significant protection across multiple studies88,89. One study found a 27% reduced risk of CVD mortality among individuals who consumed at least three servings of fruit or vegetables per day compared with those who consumed less than one90. The same population was protected against stroke, ischaemic heart disease and all-cause mortality as well.
Several studies have demonstrated a role of the microbiota in mediating dietary promotion of atherosclerosis. For example, the gut microbiota is necessary to metabolize dietary choline, phosphatidylcholine or L-carnitine, to trimethylamine (TMA) and then to the atherogenic compound trimethylamine-N-oxide (TMAO) through the action of the host liver enzyme flavin monooxygenase 3 (FMO3)91,92,93,94. These compounds predicted adverse cardiac events and outcomes in humans and led to the development of atherosclerosis in mice, but this effect required the microbiota93. A HFD in humans has since been shown to elevate post-prandial TMAO levels as well95. Although choline is an important mediator in liver and neuromuscular health96, together, these findings underline an additional effect that poverty and diet may have on the microbiota in modulating metabolic disease risk.
Indeed, as is the case in obesity and diabetes, CVD was found to be accompanied by an altered microbiota and a relative dearth of butyrate-producing microorganisms97, perhaps proceeding from the same root causes. It is important to remember that CVD, obesity and type 2 diabetes can all be thought of as inflammatory conditions that emanate from prolonged subclinical inflammation involving circulating inflammatory mediators98, endothelial vessel inflammation99 and adipose tissue inflammation100. As butyrate serves as a primary anti-inflammatory microbial metabolite (Box 2), its deficiency could potentially contribute to the failure to dampen corrosive systemic inflammation. This is an example of one way in which refined, low-fibre diets leave low-income populations vulnerable to disease.
Socio-economic status and stress
It has been hypothesized that one key component connecting low socio-economic status to chronic disease may be the persistence of stressors in the lives of individuals with low socio-economic status101. The term allostatic load is used to describe the cumulative physiological effects of prolonged stress and is an idea that has persisted for more than 20 years102. Prolonged stress can lead to both accumulated visceral adiposity and greater risk of cardiovascular mortality103. A lower social and financial position in society predisposes individuals to a number of interwoven stressors that are not necessarily experienced as frequently by higher socio-economic status individuals, such as financial stress, shift work, low job control, discrimination, noisier and more crowded living conditions, social stress, neighbourhood crime and reduced access to health care, among others. Social status and social stress have also been repeatedly associated with health outcomes. One prospective cohort study found that receiving support from relatives or a spouse could reduce mortality by 19%, and from a network of regular interaction with six or seven friends by 24%104, indicating that investing in strong relationships may be one of the more beneficial health choices an isolated individual could make. A recent study found an increased mortality from low social integration, but not independently from living in a disadvantaged neighbourhood105, indicating that the stresses that mediate the maladaptive effects of allostatic load may be social and socio-physiological in nature. People receiving low social support have more cardiovascular events and higher death rates as a result of cardiovascular causes, as well as higher all-cause mortality106. And indeed, a study in rhesus macaques demonstrated that disruptions in social status could account for direct effects on the innate and adaptive immune system. Specifically, natural killer cell and TH cell subsets were especially sensitive to changes in gene expression in response to changes in social status, and low-status females tended towards greater inflammatory potential through activation of the MYD88 pathway107.
Sympathetic nerve fibres directly connect to primary and secondary immune organs such as the bone marrow, thymus and spleen, which respectively produce, train and act as hubs of signalling for adaptive immune cells108. Stress is associated with differential cytokine expression, and different kinds of stress may affect the immune system differently. Both chronic and traumatic stress can promote elevated levels of circulating IL-6, a pro-inflammatory cytokine with a complex role in health109,110. Circulating IL-6 is a risk factor for numerous chronic diseases, including cardiovascular and metabolic disorders. It was associated with a 38% risk of heart attack for each elevated quartile of circulating cytokine111, and both C-reactive protein (CRP) and IL-6 have been linked to the development of depression112. Il6-knockout mice are resistant to stress-induced depression113, and certain anti-depressants also have the effect of lowering circulating levels of IL-6 (Ref. 114).
Until recently, connections between neurological function and the microbiota had not been thoroughly investigated, but it is now becoming evident that important connections exist. The vagus nerve connects the brain to the intestines and is capable of communicating shifts in the inflammatory milieu of the gut back to the brain115. In one study, stress from social disruption was found to reduce diversity in the gut microbiota, including a reduction in the relative abundance of the butyrate-producing genera Coprococcus and Dorea116. Similarly, male mice that underwent repeated social defeat in an experimental setting also displayed reduced bacterial diversity in the gut microbiota, which is a feature generally associated with disease117. In a model of early-life stress, maternal stress could induce neural stress pathways through the hypothalamic–pituitary–adrenal axis, but the microbiota was required for the onset of anxiety-like and despairing behaviour, suggesting that disruptions in microbial ecology act as a mechanism for the onset of stress-associated behaviour118. In fact, introduction of a SPF microbiota to germ-free mice could reduce symptoms of anxiety and enhance neurotransmitter turnover119. Though disruption of early-life microbiota through the administration of antibiotics in young rats did not influence anxiety behaviour in maze or open-field tests, it could enhance adult hypersensitivity to visceral pain, suggesting a role for the microbiota in pain sensation120. Another study in humans indicated that probiotics could decrease reported feelings of anxiety in a 30 day period121. Together, these studies indicate that the microbiota is intricately involved in stress, pain sensation and neurological well-being, which deepens its connection to the health outcomes of low-income populations.
Socio-economic status and diet
In a globalized food system, the nutrient quality of food depends on decisions made throughout the food chain122,123. In 2012, three-fourths of the calorific energy purchased among U.S. households came from processed foodstuffs, most of which could be categorized as ready-to-eat20. Processed foods are more likely to be populated with large amounts of fat and sugar while lacking the fibre, vitamin and micronutrient benefits of antioxidant-rich fruits and vegetables45. This has implications for the composition of the microbiota and its effect on the immune system (Box 2).
On the other hand, having the resources to modify diet to increase the well-acknowledged benefits of fresh foods and their accompanying micronutrients may protect against these outcomes. Evidence is accumulating for a role for the microbiota in mediating the effects of diet and disease. Ellagic acid, for instance, which is found in berries, nuts and pomegranates, is emerging as a potent antioxidant that has anti-inflammatory and antitumour immunomodulatory effects124 and that may even mitigate the inflammatory effects of diet-induced obesity and metabolic syndrome through involvement of the microbiota125,126.
Flavonoids, which are metabolized by the microbiota, also have beneficial effects on the immune system. Anthocyanins are phenolic antioxidants of the flavonoid family that are found in deep purple fruits and vegetables, such as blueberries, and have potent anti-inflammatory activity. For example, they inhibit the activation of nuclear factor κB (NF-κB) in monocytes127 and were found to attenuate airway inflammation in a mouse model of asthma128 and to inhibit the growth of cancer cells129. Anthocyanins and other flavones were also associated with reduced levels of insulin resistance, as well as lower levels of CRP130. Importantly, anthocyanins and their metabolic derivatives appear to be necessarily modified by the microbiota131.
Quercetin, a flavonol available in fruits and vegetables, is suggested to regulate hepatic lipid metabolism and prevent adipogenesis and the onset of diet-induced obesity132. Indeed, high doses of quercetin were shown to offset metabolic syndrome133. Quercetin appears to specifically attenuate macrophage expression of pro-inflammatory cytokines, such as tumour necrosis factor (TNF), IL-6 and IL-1β, which are an underlying cause of subclinical inflammation134. It has also been shown to dramatically shift the composition of the gut microbiota towards what is thought to be a more favourable metabolic profile135.
Each of the micronutrients discussed above represents just one of a number of beneficial compounds that are abundant in fresh foods. As such, a diet of fresh food confers numerous microbiota-dependent benefits to the host, but such a diet is often cost-prohibitive and consumed by higher-income families46. These studies highlight the interplay between diet, nutrition, health and the microbiota and indicate the potential negative consequences to those who cannot afford to eat fresh foods regularly.
Despite ongoing debate about the health of certain dietary structures, it is nearly universally recognized that the healthiest diets contain lots of fruits and vegetables, and their protection against premature mortality is well documented136. The key trend through all of this is that some of the most healthy micronutrients that affect the microbiota reside in fresh foods — the same foods that are generally less accessible to the financially challenged25,137. Such nutrients are lost in the energy-dense and nutrient-poor diets that are more frequently consumed by the poor.
Low socio-economic status has repeatedly been listed as a risk factor for numerous metabolic conditions, including obesity, type 2 diabetes and CVD. Although the influences of socio-economic status on health are complex and multifactorial, and although metabolic diseases often include environmental and genetic risk, the contribution of dietary factors is not trivial. In this Opinion, we have considered how the dietary trade-offs faced by individuals who do not consume a beneficial ratio of whole, minimally processed foods (for either financial or social reasons) can have a negative impact on the basic metabolism, immune system and general health of the host, in a manner that is at least partly dependent on the microbiota. We emphasize that this is likely a result of both eating too much of harmful macronutrients, such as high fat, high sugar and refined compounds, and a failure to receive the benefits of fibre fermentation and the immunoregulatory benefits of vitamins and micronutrients. Although we still do not understand all of the mechanisms involved, it is clear that the fibre, vitamins and micronutrients that are found in fresh, whole foods can promote shifts in the microbiota that are associated with protection from the chronic low-level inflammation and immune dysregulation that precipitate the development of inflammatory disease.
Of course, anyone may lose these beneficial effects of the microbiota as a result of personal dietary choices, and there are also social and financial factors that affect dietary choices. Nevertheless, the calories-per-dollar challenge faced by socio-economically challenged individuals in higher-income societies, and the compounding influence of greater stress in their lives, leaves them particularly vulnerable. Studies that address these connections more directly are urgently needed (Box 3). We propose that these relationships should be kept in mind as we move forward in creating meaningful research and a food system and food policies that serve all people.
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Financial support: Dorrance Endowed Fellowship in Pediatric Gastroenterology and Nutrition, University of Arizona Department of Pediatrics