Physical Activity, Physical Fitness, and Insulin and Glucose Concentrations in an Isolated Native Canadian Population Experiencing Rapid Lifestyle Change

The methodology of the Sandy Lake Health and Diabetes Project has been described in detail in previous publications (14,15). From July 1993 to March 1995, 728 of 1,018 (72%) eligible residents of Sandy Lake aged 10–79 years volunteered to participate in a cross-sectional survey to determine the prevalence of type 2 diabetes and its associated risk factors. Signed informed consent was obtained from all participants, and the study was approved by the Sandy Lake First Nation Band Council and the University of Toronto Human Subjects Review Committee. The analyses in this report are based on the 530 men and nonpregnant women who were ≥18 years of age at the time of the survey. Sample sizes vary slightly for certain variables due to occasional missing values.

Volunteers provided fasting blood samples for glucose and insulin after an 8- to 12-h overnight fast. A 75-g oral glucose tolerance test (OGTT) was administered, and a second sample for glucose was drawn after 120 min. Individuals were excluded from the OGTT if they had physician-diagnosed diabetes and 1) were currently receiving treatment with insulin or oral hypoglycemic agents, or 2) if they had a fasting blood glucose level >11.1 mmol/l. Diabetes and IGT were diagnosed according to established criteria (16).

Glucose level was determined using standard laboratory procedures. Insulin was measured using a radioimmunoassay (Pharmacia), which has a lower detection limit of 22 pmol/l, and an interassay coefficient of variation of 7.2–8.8%. Although this assay displays low cross-reactivity with C-peptide (<0.18%), the cross-reactivity with proinsulin is very high (100%), and thus reported values refer to concentrations of total immunoreactive insulin.

A modified version of the Modifiable Activity Questionnaire (MAQ) (originally the Pima Indian Questionnaire) (17) was used to assess physical activity levels in this population. The MAQ was designed for easy modification to maximize feasibility and appropriateness of physical activity assessment in a variety of minority populations. The MAQ has been shown to be both reliable and valid in adults as well as adolescents (18). One important feature of this interviewer-administered questionnaire is its comprehensiveness in that it assesses current (past year and, if desired, past week) occupational and leisure activities, as well as extreme levels of inactivity due to disability. The leisure and occupational estimates obtained from the questionnaire can also be combined to give an overall estimate of physical activity levels. The most recent version of the MAQ as well as detailed instructions for administrating the questionnaire and calculating the results can be found in “A Collection of Physical Activity Questionnaires for Health-Related Research” (18).

Another feature of the questionnaire is the ability to weight activities by estimates of their relative intensity. An estimate of the individual’s physical activity level is determined over the past year and expressed as hours per week, or alternatively can be weighted by a crude estimate of the metabolic cost of each activity (known in the exercise physiology literature as METs) and expressed as MET-hours per week. A MET is the ratio of the working metabolic rate of an activity divided by the resting metabolic rate.

With use of extensive qualitative research techniques (including free listing, key informant interviews, and participant observation) (19), a list of leisure and occupational activities that are specific and comprehensive to the Sandy Lake community was established. When completing the leisure section of this questionnaire, the participant was asked to identify all leisure activities from this comprehensive list in which he or she had participated on at least 10 different occasions over the past year. Estimates of frequency and duration were then obtained for each of these activities. The modified occupational section of the MAQ was used to determine, for each job held over the past year, the average number of hours that the individual participated in light, moderate, and hard physical activities during an average workday (17,18).

Anthropometric measurements were performed without shoes and with the volunteer wearing either undergarments and a hospital gown or light athletic clothing. Each measurement was performed twice, and the average was used in the analysis. Height was measured to the nearest 0.1 cm using a wall-mounted stadiometer. Weight was measured to the nearest 0.1 kg using a hospital balance beam scale. BMI was defined as weight/height² (kg/m²). The waist was measured to the nearest 0.5 cm at the point of narrowing between the umbilicus and xiphoid process; the hips were measured to the nearest 0.5 cm at maximum extension of the buttocks. Waist-to-hip ratio was calculated as the ratio of these two circumferences.

Percent body fat was estimated by bioelectrical impedance analysis (BIA) using the Tanita TBF-201 Body Fat Analyzer (Tanita Corp., Tokyo, Japan). We have documented high reproducibility of percent fat estimates using this machine (intraclass correlation coefficient 0.99) in a sample from this population. The instrument has been validated by others against dual-energy X-ray absorptiometry (DEXA) in both Caucasian and Asian populations (20,21), and has recently been validated against DEXA in a sample of individuals with type 2 diabetes who had similar fat distributions to participants in this study (22).

Vo_2max, a measure of cardiovascular fitness, was estimated using a submaximal step test developed and validated by Siconolfi et al. (23). Participants stepped on a 25.4-cm exercise stepper for 3 min per stage for a maximum of three stages. Heart rate was measured using a finger-clip pulse monitor in the last 30 s of each phase. Exclusion criteria included a medical history of cardiovascular, respiratory, or severe muscular-skeletal disease or an unwillingness to perform the test; 72% of men and 61% of women aged ≥18 years completed the step test protocol. In the analysis, Vo_2max was adjusted for lean body mass (estimated from BIA).

All statistical analyses were conducted using SAS version 6.12 (1990) and were performed separately for each sex. Associations between continuous variables were assessed using the Spearman’s rank order correlation coefficient, adjusted for age. This nonparametric technique was used because of the skewed distribution of some of the variables of interest.

Multiple linear regression analysis was used to determine the contribution of physical activity and Vo_2max to variation in fasting insulin and glucose concentrations in nondiabetic subjects. In these models, the natural logarithm of the dependent variable (either fasting insulin or fasting glucose) was used to correct for skewness. For the independent variables, either the square root (total, leisure, and occupational activity) or the natural logarithm (Vo_2max) transformation was used to overcome skewness and improve the model fit. Each model was adjusted for age, total body adiposity, and waist circumference. For models examining the role of activity, BMI was used as an estimate of total body adiposity, whereas percent body fat from BIA was used in models examining the role of Vo_2max. These decisions were made to improve the overall fit of individual models. We tested for statistical interactions between activity and age, waist and BMI, Vo_2max and age, and waist and percent body fat by adding two-way interaction terms to the full models. The importance of these interactions was determined by examining the individual P values of the interaction coefficients, and by assessing the impact of the addition of these terms on the overall fit of the models.

Completed “reliable” physical activity interviews were collected on 530 men and nonpregnant women between the ages of 18 and 79 years from the Sandy Lake Community, which represents a participation rate of ∼72% of the population within the same age range.

The general characteristics of the 231 men and 299 women 18–79 years of age constituting this Sandy Lake cohort are presented in Table 1. The prevalence of both IGT and type 2 diabetes increases with age and appears to be somewhat higher in women than in men. Women also tended to have substantially higher percent body fat than men within the same age-groups. These findings have been previously reported in earlier descriptions of this population (12,14).

Reported leisure physical activity levels during the past year, expressed in median hours per week of activity averaged over the past year, are also presented in Table 1. The commonly reported leisure activities included walking, bicycling, chopping wood, carrying water, and gardening as compared with sporting activities, with the exception of softball and hockey. The contribution of occupational activity to total physical activity levels was greater than that of leisure physical activity for all age-groups in both men and women. Not surprisingly, both leisure and occupational physical activity were inversely associated with age in both sexes. In addition, men were more active than their similarly aged female counterparts with the exception of the reported job activity level for the >35 age-group. However, if the contribution of housework is eliminated (justified by the fact that physical activity demands within housework are difficult to quantify), the median hours per week of occupational activity averaged over the past year is reduced to zero in women of both age-groups.

A comparison of reported total physical activity levels by glucose tolerance status (normal glucose tolerance, IGT, newly diagnosed diabetes, and previously known diabetes) is shown in Fig. 1. A clear trend of decreasing physical activity levels with worsening glucose tolerance status was apparent in men, although this pattern was not noted among women.

As previously mentioned, 72% of men and 61% of women ≥18 years of age completed the submaximal step test protocol to obtain an estimate of Vo_2max. Not surprisingly, those who did not perform the step test were also found to be significantly less physically active and more likely to report being confined to a bed or chair for more than 1 month in their past than those who did complete the step test.

As an indirect assessment of validity, the subjective activity questionnaire results were examined in relation to the objective Vo_2max measures. Specifically, age-adjusted Spearman’s rank order correlations between physical activity and Vo_2max adjusted for lean body mass were determined. Spearman’s rank order correlations were used because of the skewness of the physical activity data. In men, particularly younger men, occupational but not leisure activity was significantly related to Vo_2max (ρ = 0.16 and −0.06, respectively). In contrast, leisure relative to occupational physical activity was more strongly related to Vo_2max in women, although neither was statistically significant (ρ = 0.09 and −0.02, respectively).

Multiple regression analysis was used to determine whether physical activity was independently associated with fasting insulin concentrations in nondiabetic individuals (Table 2). In men, current total physical activity (leisure and occupational), controlled for age, BMI, and waist circumference, was significantly associated with fasting insulin concentrations. In a model in which total physical activity was replaced by physical fitness, physical fitness was also significantly associated with fasting insulin concentrations in men, controlled for age, percent body fat, and waist circumference. The independent association of either physical activity or physical fitness and fasting insulin concentrations remained significant, even when fasting glucose concentration was added to the respective model (data not shown). Finally, if BMI is used in the fitness model instead of percent body fat, or if percent body fat replaced BMI in the activity model, the results are similar (data not shown). In contrast, neither physical activity nor physical fitness was significantly associated with insulin concentrations in women. The combined two-way interaction terms involving physical activity/physical fitness and the other independent variables did not significantly add to the model, so these interaction terms were not included in the full model.

Multiple regression analysis was also used to determine whether physical activity and physical fitness were independently associated with fasting glucose concentrations in nondiabetic individuals as well (data not shown). In general, these relations were not as strong or as consistent as they were when fasting insulin concentration was the dependent variable, although leisure physical activity in men and physical fitness in women were found to be significantly related to fasting glucose concentrations (controlled for age, BMI or percent body fat, and waist circumference).

In this isolated Native Canadian community, both physical activity and physical fitness were independently found to be significantly associated with fasting insulin concentrations after controlling for age, BMI or percent body fat, and waist circumference in men but not women from Sandy Lake. Although these findings are cross-sectional, they suggest a beneficial effect of physical activity/fitness on insulin sensitivity that is separate from any influence of physical activity on body composition.

The fact that both physical activity and physical fitness were found to be significantly associated with fasting insulin concentrations in men is not surprising. Previous epidemiological studies involving a variety of populations, including the current study, have identified a positive effect of physical activity/fitness on insulin sensitivity (estimated using insulin concentrations). Many of the cross-sectional studies in nondiabetic individuals that have examined the association between physical activity and postload insulin concentrations (24) have found significantly higher insulin levels in the less active than in the more active individuals. However, only a few cross-sectional studies have mirrored our findings of a significant relation with activity and fasting insulin after adjusting for body size (25). Enhanced insulin sensitivity due to physical activity has also been demonstrated in a multitude of activity training studies (26). Possible physiological mechanisms responsible for the beneficial influence of physical activity on insulin sensitivity include enhanced insulin receptor and postreceptor function, increased skeletal muscle insulin-sensitive glucose transporter proteins, increased skeletal muscle capillary density, and change in body composition (26).

A key question arising from this current investigation is in regard to why both physical activity and physical fitness were found to be significantly associated with fasting insulin concentrations in men but not women in this population. There is no apparent reason to believe that the physiological mechanisms behind a relation between activity or fitness and insulin sensitivity would not be similar between men and women. A more likely explanation lies in the measurement of activity and fitness in this study and the fact that women in this population appear to be less active than the men and generally have a relatively narrower range of activity/fitness values. In the current effort, physical activity was assessed by questionnaire, which may not be as appropriate an assessment tool in women as it is in men of this population. Women tend to engage in lower-intensity activities, like walking, child care, and housework—all of which are relatively more difficult to assess and less reproducible than higher-intensity activities such as those found with many organized sports and in occupational activities (18,27).

In this current effort, cardiovascular fitness was estimated using a submaximal step test (23). It is possible that the women of this community were less willing or less able than the men to maximize their effort on this test, resulting in a less accurate test result. The fact that the women were less physically active (based on self-reported physical activity levels), more obese, and less likely to perform the fitness testing in the first place (28% of men vs. 39% of women were unable to complete this test) supports this hypothesis.

In summary, a consistent significant relation between insulin concentrations and both physical activity and physical fitness was demonstrated in a population sample of the Algonquian-speaking Ojibwa-Cree inhabitants of the isolated community of Sandy Lake, Ontario. The suggestion that a physically active, physically fit lifestyle would positively influence insulin sensitivity separate from any influence of body composition is certainly consistent with the existing scientific literature (28). What is of concern is the fact that, as this population continues to progress to a lifestyle that is relatively less physically active, with increasing amounts of television watching and other sedentary activities, declines in insulin sensitivity with increases in type 2 diabetes are anticipated. Rather than wait until diabetes has become an epidemic, lifestyle intervention efforts have been implemented in the Sandy Lake Community. Partnership between the investigative team and the study community has resulted in the development and implementation of culturally appropriate intervention strategies, including an elementary school–based diet and exercise curriculum, media and grocery store–based activities, and the construction of walking trails (29). The effectiveness of these interventions is currently being evaluated and will provide valuable public health information.

This work was supported by grants from the National Institutes of Health (91-DK-01 and 1-R21-DK-44597-01) and the Ontario Ministry of Health (#04307). A.M.K. was supported in part by grant RO1-DK-43394-06A1 from the National Institute of Diabetes and Digestive and Kidney Diseases. A.J.G.H. was supported by Health Canada through a National Health Research and Development Program Research Training Award. S.B.H. is a Career Scientist with the Ontario Ministry of Health.

The authors would like to acknowledge the following groups and individuals, whose cooperation was essential in the design and implementation of this project: the chief, the council, and the community members of Sandy Lake First Nation; the Sandy Lake community surveyors (Louisa Kakegamic, Tina Noon, Madeliene Kakegamic, Elda Anishinabie, Annette Rae, and Connie Kakegamic); the Sandy Lake nurses; Mary Mamakeesic; the staff of the University of Toronto Sioux Lookout Program; the Department of Clinical Epidemiology of the Samuel Lunenfeld Research Institute; Dr. Alexander Logan; Dr. Joe Gao; and Annette Barnie.