Kehoe et al. evaluated accelerometer measures in children (6 to 10 year-olds) in rural India. There were 8% of small for gestational age (SGA) children in the sample (considered as a birth weight of <2.5kg) and birth weight varied from 1.57 to 4.75kg. There were no significant associations between any of the anthropometric measures at birth and the physical activity variables.36 It is interesting to notice that in this study socio economic status was negatively associated with physical activity. Although the data were not shown, the authors described that separate analysis for SGA and preterm were conducted and that there was a trend for SGA to be less active. It seems to us that the definition of SGA was not adequate and the use of other definitions for SGA (i.e.: smaller than 10th percentile or birth weight ratio <0.85) would result in a larger proportion of low birth weight for a given gestational age in the population described.

The ALSPAC (Avon Longitudinal Study of Parents and Children) cohort included a large sample of school-aged children and there was no relationship between physical activity levels, measured through accelerometer, and any of the neonatal anthropometrics measurements (birth weight, ponderal index, head circumference) or gestational age.43 Birth weight was used as a continuous variable but the study did not categorize individuals as SGA or IUGR.

Another study included four birth cohorts from European countries and Brazil32 and found no significant association between birth weight and objectively measured physical activity. The cohort from Brazil that was included in this study revealed results in the opposite direction, with higher birth weight being associated with lower levels of activity.32

Kaseva et al. used the same methodology with accelerometer to objectively measure physical activity levels in a cohort (Helsinki Study) of very low birth weight (VLBW) (<1500g) young adults.51 Overall there were no differences in physical activity levels or sedentary time between VLBW and control participants.

Hildebrand et al.31 and Ekelund et al.52 used the International Children's Accelerometry Database and found that birth weight was positively associated with sedentary time and with waist circumference.31 There were no data regarding gestational age, so the authors could not differentiate between low birth weight caused by IUGR or by prematurity. In 2014 van Deutekom et al. described that birth weight was not related to either physical activity or sedentary behavior.53

Two other studies were performed in the previously described Helsinki cohort,28,47 but assessed physical activity through questionnaires of LTPA. The study by Kaseva et al. found markedly reduced LTPA in VLBW adults when compared to controls.47 Kajantie et al. described similar findings.28 When asked about the amount of time spent in physical activity during their leisure time, VLBW individuals had significant less time devoted to LTPA when compared to controls (35% versus 25%). It is important to notice that the VLBW sample included 30% SGA, while the control sample had no SGAs. Subgroup analysis comparing VLBW SGA and VLBW born appropriate for gestational age (AGA) revealed no differences in any of the physical activity variables, nor birth weight SD score and gestational age as continuous variables within the VLBW group. In this study, IUGR does not seem to be implicated in the findings and it is likely that the association is attributable to prematurity.28

In a sample of both children and adults born IUGR, teachers reported lower ability in sports in 13 year-olds and lower self-reported LTPA in adulthood.54 A meta-analysis by Andersen et al.26 included a wide range of age groups, from 14 to 74 years old, derived from 13 Nordic cohorts. Both low and high birth weight were associated with lower LTPA levels in women and men.

Some studies explored other possible factors interacting with birth weight to influence sedentary behavior levels, such as education and socioeconomic status. A cohort of young adults by Fernandes et al. found an association between sedentary and LBW conditioned to individuals who have higher educational levels.22 Another Brazilian cohort revealed an inverse association between birth weight and LTPA in women but found that socioeconomic status was also related to exercise levels.30 Hallal et al. had similar conclusions to lower birth weight being associated with sedentary behavior at 10–12 years old. Other risk factors for sedentary behavior were higher maternal education and family income, similarly to the other two studies described. From the different studies, it is possible to apprehend that socio-economic variables (both income and education) deeply affect the association between birth weight and sedentary behavior, either when comparing different social extracts within a specific community or when comparing studies performed in countries from diverse economic profile.

In 2014 Gatford et al. comprised a review of clinical and experimental evidence that exercise can improve metabolic outcomes after IUGR.55 It is clear that IUGR adults benefit from physical activity when it comes to cardiometabolic outcomes but it is still unclear whether these benefits are different when comparing IUGR to normal birth weight individuals. Table 2 summarizes the main evidence about cardiometabolic effects of exercise in IUGR individuals.

Aoyama et al. studied 379 adults (20–64 years), evaluating their homeostasis model assessment of insulin (HOMA-IR) and cardiorespiratory fitness. Lower birth weight and cardiorespiratory fitness were associated with increased HOMA-IR, while BMI and abdominal circumference were positively associated with HOMA-IR. Further analysis using HOMA-IR as an independent variable, adjusted for cardiorespiratory fitness, little modified the association between birth weight and insulin resistance. Lifestyle factors may be more important than birth weight since the major predictor of insulin resistance was cardiorespiratory fitness.56

Ridgway et al. evaluated a younger population (9–15 years) with accelerometer measurements to verify whether physical activity and aerobic fitness could modify the association between birth weight and metabolic outcomes. Birth weight was not associated with fasting insulin, except when the model was adjusted for adolescent waist circumference. The data were little modified even after further adjustment for time spent exercising.57 It is possible that the young age of evaluation may play a role in these findings, considering that changes in metabolism may still not be present at this age.

One study conducted in India included young healthy men (18–22 years), excluding obese and those using medications that could alter the glucose profile, for a bicycle-based intervention over 6 weeks in order to evaluate changes in glucose metabolism. Approximately half of the sample comprised men born with LBW. Baseline data revealed no significant differences in either fasting plasma glucose or insulin levels between LBW and normal birth weight groups. Fasting plasma insulin levels, HOMA-IR and insulin secretion (HOMA-IS) improved equally after the intervention in both groups.58 One limitation is that the study selected only healthy individuals. Since IUGR is a known risk factor for metabolic syndrome2,10,11 and type II diabetes,11 the exclusion of individuals with altered metabolic states (obese and those using medications) may lead to the false conclusion that they would not benefit even more from an exercise intervention.

Redmond et al. evaluated differential outcomes of an exercise intervention in the metabolic profile of adults born small for gestational age. Individuals born with normal birth weight had a better pre-training lipid profile (particularly total and LDL cholesterol). Following training, group differences between total and LDL cholesterol disappeared. The increased risk of chronic diseases in those born with IUGR may be attenuated through lifestyle changes such as implementing an exercise routine.39

Among the studies showing beneficial effects of exercise, Eriksson et al. evaluated older adults and showed that frequent or moderate weekly exercise was associated with better glucose tolerance and this effect was even stronger among those born with small size at birth. Furthermore, men that were born small exercised more in adulthood than their controls, which could be interpreted as an adaptive response of survival of the fittest in this high-risk group.29 Also investigating adults, Laaksonen et al. found that small size at birth was associated with metabolic syndrome features even prior to the development of diabetes or cardiovascular disease. The association was no longer present when regular vigorous LTPA or good cardiorespiratory fitness were taken into consideration.38

Two studies in adolescents were favorable to physical activity as an attenuator of altered metabolic outcomes for those born small.59,60 The first one found that as birth weight decreases, leptin levels increase in girls not meeting the physical activity recommendations (≥60min/day of moderate to vigorous physical activity). There was no association between leptin levels and birth weight in those meeting the physical activity recommendations.59 Serum leptin levels are increased in obese people, who seem to develop a leptin resistance status, and it has already been shown that LBW is associated with higher serum leptin levels in female adolescents.61 The second article showed that higher levels of physical activity played a role in altering the association between LBW and HOMA-IR.60

Although the cohorts and studies are heterogeneous in terms of many variables such as age, country of origin and type of physical activity, it seems that physical training may have beneficial effects especially for individuals that are more vulnerable to metabolic modifications such as those born with LBW.

Similarly to what was described in the clinical studies, experimental evidence also suffers from heterogeneity in the methodology used to assess physical activity. Spontaneous physical activity is measured either by filming the cage, by infrared spectroscopy or by digital counter attached to running wheels. Table 3 summarizes the main characteristics of the outcomes of experimental studies.

In order to identify changes in physical activity levels among female mice that were born IUGR, Baker et al. tested animals born from lean and obese dams, with all animals being fostered to lean dams. IUGR was noted in the offspring of obese dams. Female animals born from obese dams had reduced levels of spontaneous physical activity and energy expenditure, with increased rates of adult obesity.62 Kaur et al. induced IUGR through surgical manipulation of endometrial placental attachment sites of ewes and measured spontaneous physical activity in the offspring. Placental restriction and low birth weight were associated with greater spontaneous ambulatory activity.63

Cunha et al. evaluated spontaneous physical activity in rats born with IUGR (dams receiving a 50% food restricted diet from day 10 of pregnancy until birth) versus rats born from dams receiving an ad libitum diet and rats born from dams receiving high-fat diet. Both extreme groups (50% food restricted and high fat diets during pregnancy) had IUGR pups at birth, and both had different physical activity levels when compared to controls. However, there were opposite effects in different sexes: females from the extreme groups had increased physical activity in comparison to controls, while males from the extreme groups showed decreased activity in relation to controls.23 Considering that alterations in the mesolimbic system could be a possible mechanism to explain the effect observed in physical activity, D2 receptors in the dorsal striatum were assessed, and females from the food restricted group had decreased striatal D2 receptor levels when compared to the control group.23

Bellinger et al. fed pregnant rats with either a control or a low protein diet in different gestational periods (early, mid or late). Offspring locomotor activity was measured with an infrared sensor array. A low protein diet in all gestational periods was significantly associated with lower locomotor activity in female offspring, while in male offspring the effect was observed only when the low protein diet was offered in early pregnancy.25

Vickers et al. compared adult rats born from dams submitted to an undernourished diet during pregnancy versus controls. IUGR animals (born from dams receiving an undernourished diet during pregnancy) were more sedentary when compared to controls, regardless of the postnatal diet (standard or hypercaloric). The effect occurred in both genders but was more pronounced in males.24 In the study by Miles et al., the animals were given the choice between wheel running and pressing a response lever for food. IUGR offspring (born from undernourished dams) revealed a greater preference for exercise when compared to ad libitum offspring.64

Garg et al. conducted a trial of early exercise training to analyze its effect on glucose kinetics. Pregnant rats were divided to receive either a normal or a restricted diet during days 11–21 of gestation. An exercise training intervention in the offspring was able to suppress glucose-stimulated insulin production and hepatic glucose production, leading to an overall improvement on insulin sensitivity.65 A short period of exercise training in pups was associated with significant increases in β-cell mass for all animals and this early exercise restored β-cell mass to control levels in IUGR offspring. However, insulin secretion was not altered by the intervention.66

Miles et al. was able to demonstrate that daily exercise in adult IUGR rats prevented the development of obesity. Although IUGR offspring had increased adiposity compared to controls, insulin resistance was not observed. Daily physical activity prevented the development of obesity and had positive effects on metabolic parameters in IUGR offspring.67 In a different study, physical activity also affected leptin content of the adipose tissue in IUGR rats. Offspring from low-protein diet fed dams had increased leptin in the visceral tissue but this was reversed by exercise.68 Swimming in IUGR rats also prevented glucose intolerance from developing as well as reducing adiposity in another sample.69 Adult IUGR female rats exercised continuously for 60min during pregnancy also had enhanced peripheral insulin sensitivity although in this sample this was associated with IUGR in the offspring.70

Other interesting effects of physical activity in IUGR rats concern the proportion of muscle fibers,71 improvements in Angiotensin II-induced vasoconstriction, alterations related to superoxide levels72 and improvements in endothelium-derived hyperpolarization-mediated vasodilation in muscle arteries.73 In contrast, study by Maux et al. found that moderate physical training did not modify structural changes associated with malnutrition on the common left carotid and horizontal aorta arteries wall.74 Using a different methodology with bilateral uterine vessel ligation or sham surgery in pregnant dams to induce IUGR, Laker et al. did not find any reprogramming effects related to exercise in skeletal muscle markers of mitochondrial biogenesis in adulthood.75

Therefore, from the different experimental studies, it is possible to infer that there is evidence to support the existence of a fetal programming effect on physical activity/sedentary behavior later in life. In addition, exposure to exercise during the life course seems to be able to revert some of the detrimental effects of being born IUGR, especially regarding metabolic outcomes, although less consistent reversion is seen on the damages found on muscle architecture or function.