Cardiorespiratory Fitness and Vigorous Leisure-Time Physical Activity Modify the Association of Small Size at Birth With the Metabolic Syndrome

The Kuopio Ischemic Heart Disease Risk Factor Study (KIHD) is a prospective population-based study (22). The original study population comprised a random age-stratified sample of 2,682 men living in eastern Finland who were 42, 48, 54, or 60 years of age at baseline. The University of Kuopio Research Ethics Committee approved the study. All participants gave their written informed consent.

Data on birth weight and length were available for 624 men. Data were missing for many men who were born in the area of Eastern Finland annexed by the Soviet Union at the end of World War II. Many records were destroyed during the war or in fires or were lost for other reasons. Some men were born at home without the assistance of a midwife, nurse, or physician. Of the 624 men, complete data for defining the metabolic syndrome were available for 462. Men for whom birth data were available were younger (50.6 ± 6.0 years) than men with missing birth data (53.9 ± 4.4 years). After adjusting for age, men with complete birth data did not differ significantly from men with missing data in any of the clinical, biochemical, or anthropometric data shown in Table 1.

For the present study, data on birth weight recorded in grams and length recorded in centimeters were collected in 1998–1999 from prenatal and pediatric health clinic and school records obtained from the community health center, hospital, and home. The ponderal index was calculated as birth weight (kg)/length (m³).

Blood pressure was measured with a random-zero mercury sphygmomanometer (Hawksley, U.K.). Three measurements were recorded at 5-min intervals while the patient was supine, one measurement was recorded while standing, and two measurements were recorded while sitting. The mean of all six measurements was used as the systolic and diastolic blood pressure.

BMI was computed as weight divided by height squared. Waist circumference was calculated as the average of two measurements recorded after inspiration and after expiration (mean difference ≅1.5 cm) at the midpoint between the lowest rib and the iliac crest. The waist-to-hip ratio was defined as waist girth divided by hip circumference measured at the trochanter major.

Participants were asked to fast and to refrain from smoking for 12 h and to avoid alcohol intake for 3 days before blood sampling. Blood glucose was measured using a glucose dehydrogenase method. The serum samples for insulin determination were stored at −80°C and later analyzed with a Novo Biolabs radioimmunoassay kit (Novo Nordisk, Bagsvaerd, Denmark). LDL and HDL fractions were separated from fresh serum by ultracentrifugation and precipitation. LDL and HDL cholesterol and triglycerides were measured enzymatically.

The metabolic syndrome for men, according to the WHO definition, was modified for epidemiological studies in part as proposed by the European Group for the Study of Insulin Resistance (23) and defined as insulin resistance in the top 25% of the nondiabetic population, impaired fasting glycemia or diabetes, and presence of at least two of the following: abdominal obesity, dyslipidemia (triglycerides ≥1.70 or HDL <0.9 mmol/l), or hypertension (blood pressure ≥140/90 mmHg or use of blood pressure medication) (19).

Insulin resistance was estimated using a validated insulin sensitivity index (QUICKI) (24). The 25% cutoff was determined from the entire nondiabetic KIHD cohort for whom information on the core components of the metabolic syndrome was available. Impaired fasting glycemia was defined as fasting blood glucose level of 5.6–6.0 mmol/l (19). Diabetes was defined as fasting blood glucose level ≥6.1 mmol/l or a clinical diagnosis of diabetes with either dietary, oral medication, or insulin treatment (19). Men with diabetes were excluded from the study. The original WHO definition of low HDL cholesterol (<0.9 mmol/l) and abdominal obesity (waist-to-hip ratio >0.90 or BMI ≥30 kg/m²) was maintained. As suggested by the European Group for the Study of Insulin Resistance, microalbuminuria was not included in the definition (21). This modified definition has been validated previously (20).

We assessed childhood socioeconomic status with a three-category composite index (low, medium, or high) based on parents’ education and occupation during the subject’s childhood, whether the family lived on a farm, and whether the family was perceived as wealthy (25). Adult socioeconomic status was assessed with a composite index based on education and occupation, income, material wealth, and perception of financial security (25).

The validated KIHD 12-month Leisure-Time Physical Activity Questionnaire was used as described previously (24,25). This is a detailed quantitative questionnaire assessing the duration, frequency, and mean intensity of the most common lifestyle and structured LTPA of middle-aged Finnish men as recalled over the previous 12 months. The cutoff of ≥7.5 metabolic equivalents (METs; 1 MET is the metabolic expenditure at rest, corresponding to an oxygen uptake of 3.5 ml O₂/kg) for high-intensity physical activity included lifestyle and conditioning activities commonly considered vigorous (e.g., chopping wood or shoveling snow, skiing, jogging, ball sports) at the subjective intensities at which they were practiced (6). Low-intensity LTPA was defined as <4.5 METs. Moderate- and high-intensity LTPA was defined as ≥4.5 METs. The durations of LTPA were calculated in min/week.

A graded symptom-limited maximal exercise test was performed on an electrically braked cycle ergometer (model 400L; Medical Fitness Equipment, Mearn, the Netherlands). Vo_2max was measured directly with breath-by-breath respiratory gas exchange analysis (Medical Graphics 2001; Medical Graphics, St. Paul, Minnesota), as described previously (26,27).

Assessment of medical history, medication use, family history of diseases, and smoking (28) has been described previously.

Differences in anthropometric, biochemical, and clinical variables between men with and without the metabolic syndrome were assessed with the Student’s t test or the χ² test as indicated. The age-adjusted associations of birth weight or ponderal index categories with insulin and glucose levels in middle age were assessed with ANCOVA, and the trend was analyzed with linear regression. The association of birth weight or ponderal index categories with the metabolic syndrome was estimated using logistic regression, with forced adjustment for age, and in separate models, adjustment for childhood socioeconomic status, adult socioeconomic status, and adult BMI. Indexes of birth weight and ponderal index were categorized into thirds for both logistic regression and covariate analyses. We also assessed whether the association of the ponderal index with insulin and glucose levels and the metabolic syndrome in middle age was modified by LTPA or cardiorespiratory fitness. For these analyses, we stratified the men by the weekly duration of median LTPA and cardiorespiratory fitness. Because especially Vo_2max is strongly and inversely associated with age, we adjusted Vo_2max and duration of vigorous LTPA for age (using the entire nondiabetic KIHD cohort for whom all variables for calculation of the metabolic syndrome were available) before determining cutoffs for stratification of Vo_2max and vigorous LTPA. For Vo_2max in ml · kg⁻¹ · min⁻¹, we used a lower cutoff (40%) because so few cases (n = 11) of the metabolic syndrome were present in fit men when using median fitness as a cutoff. In stratified analyses, the ponderal index was dichotomized at the lower tertile for the statistical analyses. We also tested for the statistical interaction of physical activity and fitness with thinness at birth in the ANCOVA and logistic regression modes, including age and dichotomized measures of either physical activity or fitness, the ponderal index, and the appropriate product term. Triglyceride and insulin concentrations were corrected for skewing using log transformation but are presented using untransformed values. Data are presented as means ± SD, medians, and interquartile ranges or simple percentages where appropriate in Table 1. Significance was considered to be P < 0.05. All statistical analyses were performed with SPSS 11.0 for Windows (Chicago, IL).

Men with the metabolic syndrome had a higher BMI and waist-to-hip ratio, higher serum triglyceride level and lower serum HDL level, higher fasting blood glucose and serum insulin levels, and higher blood pressure than men without the syndrome (Table 1). Moreover, men with the metabolic syndrome were older and more likely to have CVD or use blood pressure medication. Men without the metabolic syndrome had a much higher Vo_2max and tended to engage in more vigorous LTPA (Table 1).

After adjustment for age, men in the lower third of the ponderal index or birth weight had clearly higher insulin and glucose concentrations than men in the upper third (Table 2). A ponderal index or birth weight in the lower third was also associated with lower insulin sensitivity in middle age.

In age-adjusted analyses, men with a ponderal index or body weight in the lower third at birth were roughly two times more likely to have the metabolic syndrome (Table 3). Adjustment for childhood socioeconomic status or adult socioeconomic status did not affect the association. Adjustment for BMI in middle age seemed to increase the association somewhat. Adjustment for adult height did not affect the association (not shown).

In the 314 men without CVD, men with a low ponderal index were nearly three times more likely to have the metabolic syndrome (odds ratio [OR] 2.91 for the upper versus lower third, 95% CI 1.18–7.15). Similarly, men with a birth weight in the lowest third were also more likely to have the metabolic syndrome (OR 2.31 for the upper versus lower third, 95% CI 0.99–5.37).

Thinness at birth was not associated with cardiorespiratory fitness (for the trend across tertiles of the ponderal index, P = 0.65) or duration of strenuous LTPA (for the trend across tertiles of the ponderal index, P = 0.47) in middle age.

When stratifying the men by median weekly duration of vigorous LTPA, the association of the ponderal index at birth with insulin and glucose levels and the insulin sensitivity index was absent in men engaging in at least 25 min/week of strenuous LTPA (Table 4). In contrast, in more sedentary men, the association of the ponderal index with hyperinsulinemia was enhanced (P = 0.033 for the interaction of vigorous physical activity and the ponderal index at birth), although the effects of stratification by duration of physical activity on the association with glucose levels (P = 0.33 for the interaction) and the insulin sensitivity index (P = 0.50 for the interaction) were less marked.

Similarly, the association of the ponderal index at birth with the metabolic syndrome in middle age was attenuated and no longer significant in men engaging in at least 25 min/week of strenuous leisure-time activity (Fig. 1A). In sedentary men, however, the association of the ponderal index with metabolic syndrome was even clearer. In statistical testing, however, the multiplicative interaction of vigorous physical activity and the ponderal index at birth was not significant (P = 0.39). The duration of total, low-intensity, and moderate-intensity LTPA did not clearly modify the association.

Likewise, the association of the ponderal index with insulin levels (P = 0.009 for the interaction between thinness at birth and Vo_2max), and glucose levels (P = 0.020 for the interaction), and the insulin sensitivity index (P = 0.15 for the interaction) in more fit men was absent, whereas the association was enhanced in less fit men (Table 5). The association of the ponderal index with metabolic syndrome was also absent in more fit men (Fig. 1B and C). In contrast, the likelihood of having the metabolic syndrome was more than twofold greater in unfit men who had a ponderal index in the lower third at birth. In statistical testing, however, the multiplicative interaction of cardiorespiratory fitness and the ponderal index at birth was not significant (P = 0.27 for the interaction between thinness at birth and Vo_2max in ml · kg⁻¹ · min⁻¹; P = 0.31 for the interaction between thinness at birth and Vo_2max in l/min).

Nondiabetic middle-aged men who were thin at birth or who had a low birth weight had higher fasting insulin and glucose concentrations and were at least two times more likely to have the metabolic syndrome. Moreover, this association was present also in men without CVD, suggesting that thinness at birth also predisposes the metabolic syndrome in its earlier stages, even before development of diabetes or CVD. Perhaps the most compelling new finding of our study was that regular vigorous LTPA or good cardiorespiratory fitness seemed to attenuate or abolish the association of thinness at birth with insulin resistance and the metabolic syndrome.

From a clinical and public health standpoint, it is important to focus on the metabolic syndrome itself rather than its components, because of the commonness of the metabolic syndrome (20,29) and its association with subsequent development of diabetes and CVD (1,3,20,21). Several studies have suggested an association of small size at birth with the metabolic syndrome itself (9,14,15), but these studies have used widely varying and unvalidated definitions. Using a modified and validated WHO definition (20), we demonstrate that men who had a ponderal index or birth weight in the lower third were at least twofold more likely than men in the upper third to have the metabolic syndrome. Adjustment for childhood or adult socioeconomic status had no effect, and adjustment for current BMI, if anything, strengthened the association.

According to the thrifty phenotype hypothesis, in utero undernourishment results in redistribution of blood flow and changes the production of fetal and placental hormones that control growth and metabolism (2,16). These changes may be reflected as reduced fetal growth and decreased cell mass, especially in muscle, liver, and β-cell islets, and may result in metabolic and endocrine “programming,” with permanent changes in, for instance, metabolism and action of insulin and IGFs, cortisol levels, sex hormones, and growth hormone (for a review on the human and experimental evidence, see ref. 17). Thinness at birth is believed to be largely a consequence of the fetal environment, but genetic factors also play a role (18).

The association of thinness at birth with hyperinsulinemia and the metabolic syndrome seemed to be enhanced in sedentary men but absent in men engaging in at least 25 min/week of strenuous lifestyle and sporting physical activity. The study may be underpowered, especially for showing a statistically significant multiplicative interaction. These findings are nonetheless strongly suggestive of a relevant biological interaction and imply that thinness at birth is a modifiable risk factor for hyperinsulinemia and the metabolic syndrome. In intervention studies, physical exercise has, to variable degrees and at least in the short term, decreased weight and visceral fat accumulation (30,31), increased HDL cholesterol and decreased triglyceride levels (31), decreased blood pressure (31), and improved insulin sensitivity (30). For most metabolic and disease end points, vigorous physical activity seems to provide additional benefits beyond low- and moderate-intensity exercise (31). We have recently extended these findings, showing that middle-aged men engaging in at least 1 h of vigorous leisure-time activity weekly were nearly three times less likely to develop the metabolic syndrome during follow-up than sedentary men (6). In that same study, men with a high Vo_2max were also much less likely to develop the metabolic syndrome, even after adjustment for age and BMI.

Cardiorespiratory fitness also seemed to modify the association of thinness at birth with hyperinsulinemia, elevated fasting glucose levels, and the metabolic syndrome in middle age. Again, the association was enhanced in unfit men but absent in fit men. The statistical interaction of thinness at birth with hyperinsulinemia and elevated fasting glucose levels was clear. Despite the striking appearance otherwise in Fig. 1B and C, the multiplicative interaction was not statistically significant. These overall findings are again strongly suggestive of a potentially important biological interaction. Cardiorespiratory fitness has been used as a proxy for or as an objective measure of LTPA. Indeed, vigorous LTPA and Vo_2max are quite strongly correlated (r = 0.40) in the KIHD cohort (6) despite the inherent problems in quantifying physical activity. A strong genetic component is nonetheless present (32). Furthermore, especially cardiorespiratory fitness is closely related to the metabolic syndrome (33). It seems clear, however, that the modifying effect of vigorous physical activity and cardiorespiratory fitness is not due to confounding with thinness at birth, because thinness at birth was not related to either the duration of vigorous LTPA or Vo_2max during middle age.

Our study has several important limitations. Records on birth size were available for only part of the KIHD cohort. After adjusting for age, however, there was no difference in the main clinical and biochemical measures between men with and without birth records. Information on gestational age was not available. Nonetheless, in previous studies adjustment for gestational age has not attenuated the association of small size at birth with insulin resistance, glucose intolerance, diabetes, or CVD (8,13,34,35). Study size limited more detailed assessment of the modifying role of physical activity and fitness. Noteworthy strengths include a population-based design, detailed assessment of lifestyle and sporting LTPA with a validated questionnaire (26,27), direct measurement of Vo_2max during a maximal exercise test, assessment of childhood socioeconomic status, and use of an accepted and validated definition of the metabolic syndrome.

Small size at birth was associated with the metabolic syndrome in middle-aged men already in the early stages of the syndrome, before development of diabetes or CVD. Vigorous LTPA and cardiorespiratory fitness seemed to modify the association of thinness at birth with the metabolic syndrome such that the association was absent in active or more fit men but enhanced in sedentary and unfit men. These findings support previous studies showing that thinness at birth may carry with it lifelong metabolic consequences, but suggest that regular strenuous physical activity and maintenance of cardiorespiratory fitness may alleviate or eliminate those consequences.

The KIHD study was supported by grants from the Academy of Finland (41471, 1041086, and 2041022), the Ministry of Education of Finland (167/722/96, 157/722/97, and 156/722/98), and the National Heart, Lung and Blood Institute (HL44199). J.L. is supported by Grant 1 P50 HD38986-01 from the National Institute of Child Health and Development.

We thank the staff of the Research Institute of Public Health, University of Kuopio and the Kuopio Research Institute of Exercise Medicine for data collection in the KIHD study. We also thank Eeva Kumpulainen for the historical data collection and entry.