Although increased PA is known to improve IR, the mechanisms underlying this improvement are not completely understood. More specifically, it is difficult to distinguish the metabolic benefits of PA, per se, from the improvement in cardiorespiratory fitness that often accompanies an increase in PA. Presently, we found a strong and significant association between daily MVPA and a measure of IR (HOMA-IR) after adjusting for confounding variables including cardiorespiratory fitness. In contrast, we did not observe the same relationship between cardiorespiratory fitness and log HOMA-IR after correcting for the same confounding variables. Together these results highlight the potential importance of daily MVPA as a target intervention of improving IR independent of changes to cardiorespiratory fitness.
It is clear from previously published longitudinal studies that individuals with greater cardiorespiratory fitness are at lower risk for developing diabetes [18, 19]. Yet improvements in cardiovascular fitness as a result of exercise training are highly variable and can take several months to develop . Also, while the most robust improvements in cardiovascular fitness stem from higher intensity training [21, 22], this kind of activity is also associated with greater participant discomfort and injury contributing to higher dropout rates . Therefore, it is necessary to determine alternative intervention programs that may be more attainable by individuals at risk of developing diabetes. Our prediction analysis shows marked improvements in IR with relatively modest changes to daily MVPA. Indeed the benefits of cardiorespiratory fitness on other health outcomes such as cardiovascular health are undeniable [7, 8], but modest PA may be an appropriate alternative for individuals that may see exercise intensity as a barrier to being more physically active.
Near-optimal fasting plasma glucose and insulin levels yield a HOMA-IR value close to 1, and our model showed that 120 minutes of MVPA per day predicts near optimal HOMA-IR values. This is equivalent to only 7.5 minutes of activity per hour during wakeful hours (accounting for 8 hours of sleep), throughout the day. Some PA guidelines emphasize a bout of at least 10 minutes of PA to improve health . However, our analysis included accumulated minutes of PA throughout the entire day. We did not determine how the MVPA was accumulated, and future research is required to determine whether MVPA accumulated sporadically throughout the day is sufficient to improve IR or if it should be accumulated in bouts to have a positive effect on IR.
Surprisingly, LPA was not significantly associated with IR in our regression models. In contrast, similar methods, using objectively measured PA, identical criteria for defining intensity of PA, and inclusion of confounding variables like adiposity (i.e. waist circumference or BMI), have reported an inverse relationship between LPA and 2-hour plasma glucose concentrations during an oral glucose tolerance test . This discrepancy between our results and previous analysis may be explained by the use of HOMA-IR, which has not been found to be associated with LPA . Furthermore, while the use of accelerometers are currently the ideal method for measuring PA and have a high degree of sensitivity for LPA, HOMA-IR is a less sensitive measure of IR. Therefore LPA could have a positive effect on IR that we were unable to detect in our analysis. The use of more sensitive measures of IR (i.e. intravenous glucose tolerance test and hyperinsulinemic-euglycemic clamp methods) may be required to determine the influence of LPA on IR. Further investigation into the influence of LPA on IR may have important public health implications for individuals incapable or resistant to higher intensity PA.
Obesity is associated with increased risk of IR and type 2 diabetes [26, 27]. In our complete model, we show that body fat percentage was most strongly associated with IR. With two-thirds of the U.S. population classified as overweight or obese, weight loss is an important therapeutic target for metabolic disease prevention [1, 28]. In fact modest weight loss (~5% of initial body weight) can induce clinically relevant improvements in metabolic health . However, clinical trials targeting behavioral changes to diet and exercise as well as pharmacological interventions have shown only modest weight loss and more importantly have proven difficult to sustain . Because sustained weight loss is so difficult for many people , adopting a physically active lifestyle may be a more feasible alternative to weight loss for those at increased risk of developing IR.
Our study has certain limitations. Because our analysis is cross sectional, causation cannot be determined. However, our analysis showed that higher MVPA was indicative of lower IR in healthy adults. Large observational trials confirm the association between PA and HOMA-IR . Controlled intervention trials in which adiposity, cardiorespiratory fitness and PA are manipulated independently in a sample large enough to convincingly prove causal association are likely to be infeasible. However, a more practical approach may be to systematically determine the minimum “dose” (i.e. intensity, duration, caloric expenditure, etc.) of PA, as well as the effect of habitual PA over time on the mechanisms underlying an improvement in IR.
It is also important to note that HOMA-IR is not the most sensitive measure of insulin resistance, and better measures of insulin resistance exist including an intravenous glucose tolerance test and hyperinsulinemic-euglycemic clamp methods. Although our analysis included a relatively insensitive measure of insulin resistance, we were able to detect measurable differences and the HOMA-IR method was ideal for measuring insulin resistance in a population this large and diverse.
The NHANES 2003–2004 survey limited cardiovascular fitness testing to individuals <49 yrs of age because of increased risk for an adverse cardiovascular incident. Examining the independent relationship between physical activity and cardiovascular fitness on insulin resistance in middle and older aged adults is an important question worth pursuing. However, we were unable to make this analysis with the current NHANES data set.
Because only healthy adults were selected for this analysis in order to rule out any unnecessary confounding factors brought about by disease, on average our sample was more active and fit than the average American. While participants averaged nearly 30 minutes of MVPA per day and had an estimated VO2max of 40 ± 9 ml·kg-1·min-1, there was a large range of values and high variability between participants, suggesting a rather heterogeneous subject pool rather than a highly fit population. Although a maximal exercise test was not performed, VO2max was estimated objectively using a submaximal exercise test. Submaximal exercise tests have been shown to be strongly correlated with maximal measures of VO2max . Also, although the participants in the present study were generally younger than the age range where the incidence of newly diagnosed type 2 diabetes is greatest (45–64 years old ), in order to prevent or delay the onset of diabetes it is important to determine how PA behavior may influence IR in younger adults (>45 year old) in order to delay or prevent the onset. Other than requiring an overnight fast, diet and exercise were also not controlled prior to participants’ fasting blood draw or submaximal exercise testing. Although macronutrient meal composition and exercise can acutely alter IR [5, 33] it is not unreasonable to assume that participant’s diets and exercise habits were not altered during this brief testing period.