Foetal growth is ultimately restricted by uterine capacity. Babies may be born small because they are constrained in this way or because they lacked the nutrients in utero necessary for optimum growth (Harding 2001). Pregnant women exposed to famine give birth to small babies who go onto develop glucose intolerance and obesity in adulthood, for example (Ravelli et al. 1976, 1998). The proportion of protein and carbohydrate in a women’s diet may be a key factor that affects the foetus, with the combination of high carbohydrate and low protein intake being particularly deleterious (Campbell et al. 1996; Godfrey et al. 1996). In addition to maternal intake per se, clinical exposures such as maternal hypertension leading to reduced uterine blood flow can severely affect the supply line of nutrients from the mother to the foetus (Harding 2001).

Other well-known determinants of size at birth include sex, gestational age, ethnic origin, parity, maternal and paternal size, gestational weight gain, general morbidity and episodic illness, malaria, cigarette smoking, and alcohol consumption (Kramer 1987). Maternal gestational diabetes is associated with higher birth weight and increased risk of subsequent obesity (Lawlor et al. 2011), but Gillman et al. (2003) have questioned the causal role of altered maternal-foetal glucose metabolism because adjusting the gestational diabetes-childhood obesity association for birth weight only marginally attenuated the estimate in their study. This might, however, be expected given that even the most exposed foetuses are limited to how much weight they can gain in utero; the greatest anatomical response to altered glucose metabolism will inevitably occur after birth, when there is no upper limit on weight gain. Indeed, the most deleterious profile for childhood obesity comprises rapid infant weight gain in addition to pre-pregnancy obesity and macroscomia (Weng et al. 2012).

Determinants of childhood obesity include maternal smoking, no or short duration of breastfeeding, obesity in infancy, short sleep duration, television viewing, low daily physical activity, and consumption of sugar-sweetened beverages (Monasta et al. 2010). Many of these risk factors continue to operate in adolescence (Morandi et al. 2012), at ages when there is increasing independence from the family and the establishment of more individual as compared to familial risk factors. Most research has focused on shifts in diet and physical activity as the key drivers of the obesity epidemic, particularly during adulthood (Swinburn et al. 2011; McAllister et al. 2009). There is, however increasing evidence for multiple other factors including microorganisms and epigenetics early in life, sleep debt, and endocrine disruptors (McAllister et al. 2009).

Naturally, the exposures responsible for obesity are very different to those for malnutrition. Evidence from low to middle income countries shows that the key determinants of stunting, wasting, and underweight include growth restriction in utero, poor condition of the mother, poverty, chronic dietary insufficiency, marked seasonality, poor levels of sanitation, and infection (Martorell and Young 2012; Frongillo et al. 1997). As these countries transition, it has been argued that the people are likely to face both sets of exposures and a dual burden of stunting and overweight (Varela-Silva et al. 2012). Others have, however, shown that the co-existence of stunting and overweight, at least at the family level, is a statistical artefact with a prevalence that matches what one would expect based on the separate stunting and overweight rates (Dieffenbach and Stein 2012).

Biological pathways are ultimately responsible for the links between environmental exposures, body size, and cardio-metabolic health. The first main type of pathway involves anatomical changes. Anatomical formation of the kidneys occurs exclusively in utero, and nutritional constraint of a foetus can permanently reduce the number of nephrons that are laid down, for example (Barker et al. 2006). Although physiological capability of the kidney develops over the entire life course, small kidney size in prenatally undernourished individuals increases risk for hypertension and renal failure in adulthood (Lampl et al. 2002; Luyckx and Brenner 2005). Similarly, a larger number of adipocytes in response to nutritional excess and a sedentary lifestyle in adulthood can increase CHD risk because these cells secrete inflammatory proteins that speed up the atherosclerotic process (Berg and Scherer 2005).

The second main type of pathway involves the physiological setting or alteration of hormonal and metabolic axes. For example, the foetus responds to reduced nutritional supply by reducing plasma concentrations of hormones, such as insulin and insulin-like growth factor, which in turn limits the transportation of glucose to the muscles and impairs lean tissue growth (Phillips 1996). This adaptation occurs so that glucose is readily available in the bloodstream to maintain growth of high priority organs, such as the brain (Gluckman 1995). The tendency to maintain high blood glucose levels can, however, lead to a progressive decline in glycaemic control and type two diabetes (Phillips 1996). As a different example, dysregulation of the hypothalamic-pituitary-adrenal axis in response to chronic stress in adulthood might lead through the effects of cortisol to central obesity and cardio-metabolic disease (Rosmond and Bjorntorp 2000).

One increasingly well-studied set of mechanisms focuses on the way in which genes governing body size impact on cardio-metabolic health. We know that environmental exposures can modify the expression of genes. A recent meta-analysis reported that the odds of obesity associated with the risk allele in the FTO gene was attenuated in active adults compared to inactive adults by 27 %, for example (Kilpelainen et al. 2011). Chemical modifications occur that alter gene expression in a specific tissue or organ without changing the nucleotide sequence of the DNA (e.g., methylation and histone modification) (Holliday 1994). While an environmental exposure may influence body size in the short term, these “epigenetic” changes can have long lasting effects on disease risk. Maternal protein restriction might lead to smaller offspring and reduce methylation (and therefore enhance expression) of the angiotensinogen receptor gene in the offspring adrenal gland, thereby leading to high blood pressure, for example (Bogdarina et al. 2007; Woodall et al. 1996). Such epigenetic changes can occur from conception onward but, like the other anatomical and physiological pathways, are most sensitive to the environment during specific stages of the life course.

Genes are also being used increasingly in Mendelian randomisation studies, where a genetic variant or set of variants are used as an instrumental variable for some phenotype or exposure, to establish whether or not that phenotype has a causal effect on some outcome (Lawlor et al. 2008). This technique has recently been used, for example, to demonstrate that BMI does have causal effects on various cardio-metabolic traits, including blood pressure and fasting glucose and cholesterol (Holmes et al. 2014). These studies work, subject to certain assumptions, on the premise that alleles are transmitted from parents to offspring randomly at gamete formation, such that the genotype or instrumental variable relationship with the outcome is not confounded by environmental factors or prone to reverse causation (Lawlor et al. 2008).

A critical period refers to an age window in which intrinsic changes in body structure and function are occurring rapidly and may be most easily programmed in a favourable or unfavourable direction (Scott 1986). The focus in the natural sciences was originally on the requirement of specific environmental stimuli to elicit the normal development and functioning of some body part or system (Cameron and Demerath 2002). In contrast, epidemiology focuses on the environmental exposures that result in anatomical and physiological adaptations that have long-term implications for cardio-metabolic health (Kuh et al. 2003). Perhaps the most often used example of a critical period is the teratogenic effect of maternal exposure to thalidomide in the first trimester of pregnancy on limb development in the foetus when the limbs are most rapidly developing, while exposure to thalidomide after birth is harmless (Newman 1986). In this instance, the critical period is truly critical. In other examples, however, the exposure association (with the outcome) is not constrained to such a narrow age window, but may be present across part or whole of the life course. The strongest associations are observed at ages when the individual is most sensitive to the environment, thus these are called sensitive periods of development.

Research has now identified potentially sensitive periods spanning the entire life course. Early childhood and adolescence in particular are emerging as important periods for the establishment of obesity and programming of cardio-metabolic risk. It has recently been shown in the MRC National Survey of Health and Development that socioeconomic disadvantage in early childhood, more so than at any other age, is associated with cardio-metabolic risk factors at age 53 years, for example (Murray et al. 2011). In the same cohort, exposure to obesity emerges during adolescence as being associated with greater carotid-intima media thickness at 60–64 years of age in men (Johnson et al. 2014). The associations reported in both of these papers were independent of birth weight and appeared to be mediated by higher adulthood BMI in those individuals who were disadvantaged in early childhood or obese in adolescence.

Adolescence is often viewed as critical from physical and behavioural perspectives, but individuals experience this stage in the life course at different ages and progress through secondary sexual characteristic maturation at different paces (Ellison and Reiches 2012). A critical or sensitive period may be thought of as a point of transition from one state to another. During puberty, for example, the adolescent transitions from an immature state to a mature state. Various markers of the pace of pubertal development have been reported to be associated with various cardio-metabolic health outcomes in both sexes (see section “Adolescence”). Interestingly, an early age at menopause, which marks the cessation of a women’s reproductive ability, is also associated with increased risk for cardio-metabolic disease (Ebong et al. 2014). Lifelong environmental conditions may influence the timing of menopause (Lawlor et al. 2003), but it is unknown whether or not the biological changes that occur during menopause and impact on subsequent health are particularly sensitive to concurrent exposures. Similarly, behavioral transitions throughout the life course can lead to increased disease risk, but these do not necessarily occur at ages when biological changes are occurring rapidly and are most sensitive to the environment. Transition into a romantic relationship or marriage can cause behavioral changes that lead to the development of obesity (The and Gordon-Larsen 2009; Gordon-Larsen et al. 2004), but there is nothing biologically critical about the exact age when this occurs, for example. The same might be true for starting university (Gropper et al. 2012), leaving the army (Littman et al. 2013), and retiring (Morris et al. 1992). The timing of rapid change in a known biological structure/function is critical for the experience of specific environmental stimuli to cause permanent alterations and predict long-term outcomes (Cameron and Demerath 2002). While a critical or sensitive period may be thought of as a transition, a transition does not automatically meet the requirements to be a critical of sensitive period.

Recent research in human and animal models suggests that biological adaptations to environmental exposures during critical or sensitive periods of development may be transmitted to subsequent generations (Benyshek 2013), such that a trait like insulin resistance in the exposed generation might be passed to successive unexposed generations in diminishing order of magnitude (Benyshek et al. 2006). This transgenerational transmission goes beyond something that might be explained by genetic heritability or by family members of different generations experiencing similar lifelong environments. Maternal exposure to famine, for example, has been shown to be associated with increased risk of giving birth to small offspring with dysregulated lipid profiles (Lumey et al. 2009), who themselves go onto have relatively small offspring with large amounts of adiposity and high risk for cardio-metabolic disease (Painter et al. 2008). In this scenario, direct and simultaneous exposure of three generations (i.e., the pregnant mother, her fetus, and the fetuses’ primordial germ cells) might be responsible for the increased disease risk seen in each generation. Alternatively, and if more than three generations are affected, research is revealing how inheritance of epigenetic modifications to the genome may be responsible (Hackett et al. 2013).