Exercise Training and Insulin Resistance: A Current Review

There is a general perception that increased physical activity will improve glucose homeostasis in all individuals. While this is an attractive concept, this conclusion may be overly simplistic and even misleading. The topic was reviewed extensively over 30 years ago and it was concluded that acute exercise enhances glucose uptake. However, in some cases the chronic influence of interventions utilizing exercise may have little effect on glucose metabolism. Moreover, insulin resistance often returns to near baseline levels within a couple of days following cessation of the exercise bout; leaving the overall effectiveness of the intervention in question. Since improving glucose homeostasis should be the focal endpoint of any intervention designed to mitigate the overwhelming degree of insulin resistance in individuals at risk for metabolic disease, it is essential to evaluate the key components of a successful approach.
 

Weight loss; Inflammation; Metabolism
 

Insulin resistance is commonly presented in older adults, but has become increasingly prevalent in all ages, including middle-aged individuals that are overweight and sedentary [1-8]. An excessive deposition of adipose tissue, defined as overweight (BMI 25-30 kg/m²) and obese (BMI ≥ 30 kg/m²), is a risk factor for the development of type 2 diabetes [7], and is associated with a reduced uptake of glucose [9]. However, the overall etiology of insulin resistance is quite complex. It is known that a disproportionate accumulation of subcutaneous and abdominal fat contributes to the desensitization of insulin receptors that is characterized by an inhibited uptake of glucose within skeletal muscle and an impaired ability to suppress endogenous glucose production [10-13]. From a simplistic standpoint, these physiological alterations may lead not only hyperglycemia, but also to exacerbations in fat deposition [10,11], promoting a chronic state of inflammation characterized by the release of pro-inflammatory cytokines [9,11,14]. As insulin resistance worsens, ß-cells continually stimulate insulin production in an attempt to reduce overwhelming hyperglycemia, but remain ineffective and may influence an additional state of hyperinsulinemia [15].

To combat these deleterious consequences, the management of obesity has expanded. For example, the popularity of physical activity as a common intervention strategy may be due to its long-term effectiveness against insulin resistance when using the Oral Glucose Tolerance Test (OGTT) as a benchmark for improved glucose homeostasis [15]. The acute and chronic influence of physical activity has a multifaceted influence on glucose metabolism, including acute changes in contraction mediated glucose uptake and chronic changes in insulin-stimulated glucose uptake that may be mediated by a variety of mechanisms.

Since the translation of scientific investigation to clinical care and community engagement represents a tremendous responsibility (NCEP ATP III), recommendations for physical activity designed to mitigate the detrimental influence of insulin resistance must be precise in their application. From our perspective, the distinction between acute exercise, and chronic exercise training (with or without weight loss), or in conjunction with caloric restriction has not been clearly delineated. Therefore, the purpose of this review is to outline the key variables of physical activity that may have a different influence on the mitigation of insulin resistance, and how they might serve to ameliorate the clinical condition altogether.
 

The PubMed database was utilized to conduct a search in November of 2013-April 2015 that consisted of the following keywords: insulin resistance or insulin sensitivity each combined with exercise. Upon retrieval of over 5,000 publications, a year filtration was applied to identify studies between the years 2008 and 2013. An additional filter for “Free Full Text available” was applied, which yielded a total of 978 publications.

Inclusion criteria/rationale

There seems to be a direct connection between obesity and insulin resistance, but is highly dependent upon regional fat deposition [9,12,15]. For example, in a study by Jenkins et al. 2011, mean baseline BMI between normoglycemic and prediabetic individuals was 27.9 ± 0.4 kg/m² and 28.9 ± 0.6 kg/m² respectively with no significant difference (p=0.18) between groups. However, the prediabetic group had a significantly (p=0.005) higher mean percentage of trunk fat (as determined by DXA scan) of 39.9 ± 1.2% vs. 35.7 ± 0.8% of the normoglycemic group. This translated into a mean baseline homeostasis model assessment for insulin resistance (HOMA-IR) score of 3.8 ± 0.2 for the prediabetic group versus a significantly more favorable HOMA-IR score of 2.8 ± 0.1 for the normoglycemic group. Therefore, the evaluation of regional fat distribution differentiation is important to evaluate since these areas of fat deposition may influence insulin resistance differently and exercise training may have a variable influence on specific fat depots [17].
 

Exercise alone without weight loss

Influence of interval exercise training

Since exercise without weight loss is capable of enhancing insulin-stimulated glucose disposal especially when accompanied by a reduction in VAT, it seems intuitively obvious that insulin sensitivity would improve with exercise-induced weight loss. In many cases, weight reduction via exercise is accompanied by reductions in visceral adipose tissue, which seem to drive improvements in insulin resistance. For example, a 12 week aerobic exercise intervention of progressive vigorous intensity (60-65% to 80-85% of HR_max) most days of the week (5 sessions/week) for 50 to 60 minutes a session induced approximately 8% weight reduction with significant reductions in VAT amongst all groups with various types of glucose tolerance. In particular, there was a significant reduction in group lost a significant amount of VAT in individuals with Impaired Glucose Tolerance (IGT) (ie., -75 cm²), with Normoglycemic (NGT) Group (ie., -46.9 cm²),individuals with Impaired Fasting Glucose (IFG) (ie., -53.3 cm²), individuals with Combined Glucose Intolerance (CGI) (-15.5 cm²). These reductions were all accompanied by improvements in glucose tolerance demonstrated by reduced plasma glucose at 120 minutes of an OGTT in all groups. Additionally, insulin sensitivity and ß-cell function, assessed by a hyperinsulinemic-euglycemic clamp procedure, was enhanced following the intervention within all groups highlighting the multifaceted benefits of vigorous exercise that promotes significant weight loss [29].

Interestingly, slight reductions, but still significant, in body weight or body mass index may mitigate insulin resistance. For example, an aerobic intervention lasting 16 weeks of vigorous intensity (75% of VO_2peak), performed three to five sessions per week for 45 minutes each session promoted a 10% reduction in fat mass that correlated with an improvement in the glucose infusion rate (r=-0.635) used to assess insulin sensitivity. More specifically, insulin sensitivity was assessed 36-48 hours following the last exercise session via the hyperinsulinemic-euglycemic clamp procedure and revealed an ~23% increase in the glucose infusion rate during steady state period at the end of the procedure (Dube et al. 2012). A similar aerobic exercise intervention lasting 16 weeks of vigorous intensity (60-80% of APMHR) just two to three sessions per week for 60 minutes each session reported reductions in body mass index throughout the intervention from 24.6 ± 3.4 kg/m² to 24.0 ± 3.4 kg/m² at week 8 and 23.9±3.4 kg/m² at week 16. While BMI is a nonspecific measure of adiposity and was only slightly reduced in these studies, the reduction was significant and accompanied by increases in adiponectin from 4.9 ± 2.4 to 7.5 ± 3.9 µg/mL at week 16. Additionally, these changes significantly improved lipid profiles and reduced insulin resistance, as assessed by HOMA-IR, from 2.2 ± 1.1 to 1.8 ± 1.1 at week 16 [30].
 

Exercise intensity and weight loss: Impact on insulin sensitivity

It seems evident from our previous studies that the volume of exercise required to elicit a reduction in adipose tissue or induce weight loss with improvements in insulin sensitivity is quite cumbersome. For example, in our studies we recognized that for a middle-aged 82 kg female with a VO_2peak, exercise-induced weight loss of 0.5 kg/week necessitated 90 minutes cycle ergometer 5 days a week. As a result of this demanding volume of exercise training, dietary modification or caloric restriction are commonly implemented and utilized to reduce insulin resistance.

While caloric restriction-induced weight loss may have a large influence on insulin sensitivity, the subsequent loss of fat free mass may compromise glucose uptake, given that skeletal muscle is responsible for a larger portion of glucose uptake. This is observed in a study that implemented caloric restriction alone (-400 kcal/day or 2800 kcal/week) as well as with moderate (45-50% of HRR) and vigorous (70-75% of HRR) intensity exercise of progressing durations (moderate intensity=20-25 to 55 minutes/session, vigorous intensity=10-15 to 30 minutes/session) three days a week for 20 weeks. Although average weight loss was 13.4 ± 4.6% and included an approximate 25% reduction in visceral adipose tissue among all groups, plasma glucose at 120 minutes of an OGTT was unchanged. More importantly, fat and lean mass was reduced in each group but relative lean mass was reduced to a greater extent in the caloric restriction group (-36.2 ± 17.1%) compared to the moderate intensity exercise with caloric restriction group (-27.7 ± 14.8%) and the vigorous intensity exercise with caloric restriction group (-26.7 ± 12.1%). However, relative aerobic capacity increased within each group but significantly greater within the vigorous intensity exercise with caloric restriction (+24.2 ± 27.6%) group [36].

When fat free mass was preserved following a 12 week aerobic exercise intervention with or without a caloric deficit (500 kcals/day), insulin-stimulated glucose disposal was significantly improved (20-28% increase in relative). The exercise included progressing vigorous intensity (60-65 to 80-85% HR_max) for 50-60 minutes most days of the week (5 days/week) but the exercise alone group was eucaloric for the duration of the study. Weight loss occurred in both the aerobic exercise (3.8%) and aerobic exercise with caloric restriction (7.4%) group that was accompanied by similar reductions in abdominal adiposity and improvements in insulin sensitivity (~30%) [37].

Since weight loss typically enhances insulin sensitivity especially when skeletal muscle is retained, it seems logical that resistance training with weight loss will improve in the same manner. In an intervention consisting of 6 exercises with progressing intensity (2-3 sets/15 repetitions progressed to 3-4 sets/10-12 repetitions) three days a week for 6 months, resistance training was combined with caloric restriction (-624 ± 133 kcals/day) and was compared to caloric restriction alone (-621 ± 128 kcals/day). Although significant reductions occurred in body mass visceral adipose tissue, abdominal subcutaneous adipose tissue occurred, no significant changes were observed for relative glucose disposal. Additionally, the intervention actually resulted in significant reductions in mean total lean body mass within the caloric restriction and resistance training with caloric restriction groups [37].

Interestingly, another resistance training intervention lasting 12 weeks implemented 4 exercises at a low intensity just one day a week for 40 minutes while a 1600 kcals/day diet was provided. Notably, there was no significant weight loss but reductions in total cholesterol (-14.5 g/L) and triglycerides (-11 g/L) occurred. Although reductions of inflammatory markers were observed, insulin resistance remained, as HOMA-IR scores were not affected [38]. Unfortunately, it seems resistance training, much like aerobic exercise, requires sufficient stimulus to drive reductions in body weight or fat mass and affect insulin resistance. However, the physiological responses from resistance training versus aerobic exercise or both together on insulin resistance remain unclear.

A comparative study of aerobic exercise, resistance training, and caloric restriction has also been investigated. Aerobic exercise included a range of vigorous intensity (67-80% of HR_max) with progressing duration (20 to 40 minutes/session) three days per week. Whereas the resistance training included 10 exercises performed at 80% of 1RM (1-2 sets/10 repetitions) three days per week. The exercise groups also took part in the 800 kcals/day meals that were also provided to the caloric restriction group. The goal of the exercise and/or caloric restriction was to reach a body mass index less than 25 kg/m². Following weight loss, subjects were instructed to maintain a balanced diet and the exercise groups continued activity until 1 year follow-up measurements. Although significant weight loss occurred, weight regain was observed within each group. All groups experienced significant reductions in intra-abdominal and subcutaneous adipose tissue following weight loss with increases after a year. Although lean mass was reduced in both aerobic and caloric restriction groups, insulin sensitivity was significantly improved within all groups after weight loss, but only continued to improve within the aerobic exercise group after weight loss and the one year evaluation [38].

Interestingly, it seems that a 10% reduction in body weight following an intervention has a dramatic influence on insulin sensitivity. For instance, in a study that elicited a 10% reduction in body weight from caloric restriction (500-1000 kcals/day) or combined exercise with caloric restriction, significant increases in ISI-M were observed. The progressive aerobic exercise intensity (70% to 85% of HR_peak) for 30 minutes and resistance training (65% to 80% of 1RM, 1-2 sets/8-12 repetitions to 2-3 sets/6-8 repetitions) three days per week with caloric deficit reduced body weight (-10±2%) and intrahepatic fat (-45 ± 8%) similar to the caloric restriction group. These reductions were accompanied by a 66 ± 25% and 68 ± 28% increase in ISI-M for the caloric restriction group and combined group, respectively [39].

Similarly, a study by Schenk and colleagues compared the effects of caloric restriction alone (500-800 kcals/day) or with vigorous intensity (85% of HRmax) aerobic exercise (3-4 sessions per week for 45 minutes each) after a 12% reduction in weight was elicited [40]. The aerobic exercise with caloric restriction group reached 12% weight loss quicker (20 ± 2 weeks) than the caloric restriction group (30 ± 3 weeks), and was followed by a weight maintenance period. Although the aerobic exercise with caloric restriction group demonstrated increased levels of resting whole-body fatty acid oxidation more than 20%, insulin sensitivity increased similarly by 60-70% within both groups [40]. Additionally, skeletal muscle pro-inflammatory c-Jun N-Terminal Kinase (JNK) and fatty acid mobilization and uptake were 40% and 30% lower, respectively [40]. These reductions were likely due to a significant loss of fat mass altogether.

The importance of significant weight loss was also supported by a study from Mason and colleagues that compared the amount of weight loss relative to changes in HOMA-IR scores [41]. The intervention included aerobic exercise alone (70-85% of THR 5 days/week for 45 minutes each) or with caloric restriction, compared to caloric restriction alone to induce 10% weight reduction followed by 6 months of maintenance. Although weight loss was significant within the aerobic exercise group (-2.4%), the aerobic exercise with caloric restriction and caloric restriction groups demonstrated further weight loss (-10.8% and -8.5%, respectively) and significant HOMA-IR score reductions (-26% and -24%, respectively). The percent weight loss correlated with HOMA-IR score reductions with an r value of -0.23 for <5%, -0.69 for >5-10%, and -1.10 for >10% weight loss [41].

However, the intervention responsible for the weight loss may elicit different physiological responses. For instance, weight lost solely from dietary intake reduction may affect fat depots but weight lost from a combination of both dietary intake reduction and increased energy expenditure may improve muscle metabolism while reducing adiposity. This concept is demonstrated in a study by Toledo and colleagues that compared caloric restriction (reduce intake by 25%) to that with aerobic exercise (60-70% of HR_max for 30-40 minutes 3-5 days/week) to induce at least a 7% reduction in weight over a 16 week period [42]. Fortunately, each group experienced ~10% weight loss accompanied by ~19% reduction in fat mass, but a ~17% of visceral adipose tissue in the aerobic exercise with caloric restriction group as opposed to the ~25% reduction in the caloric restriction group. However, the aerobic exercise with caloric restriction group demonstrated a 49 ± 16% increase in mitochondrial density supporting an improved mitochondrial oxidative capacity, versus the 17 ± 4% reduction in mitochondrial size observed in the caloric restriction group [42].

Comparatively, in studies by Ryan et al. 2012 and Strasnicky et al. 2009, both interventions of similar structure comparing vigorous intensity aerobic exercise with caloric restriction to caloric restriction itself induced ~8% weight loss [43,44]. However, Strasnicky and colleagues grouped the responses together, given that the weight loss was similar between the groups. Both groups experienced significant losses in fat mass (-6.4 ± 0.6 kg) and waist circumferences (-8.6 ± 0.8 cm) that was accompanied by a significant improvement in ISI-M of ~45% [44]. Additionally sympathetic responsiveness was assessed by measuring norepinephrine spill over rate during an OGTT using a radioisotope dilution method. Weight loss promoted resting norepinephrine to decrease, but only the aerobic exercise with caloric restriction group experienced significant sympathetic responsiveness to glucose at 90 minutes of OGTT. This suggests weight loss may also reverse the blunted sympathetic responsiveness to glucose ingestion commonly seen in glucose intolerant individuals [43]. Whereas in the study by Ryan and colleagues, significant reductions in fat mass (-14%), visceral fat area (-13%), and subcutaneous abdominal fat area (-12%) were observed within each group [44]. This supported the overall improvement in glucose utilization and non oxidative glucose disposal during the hyperinsulinemic-euglycemic clamp procedure, 14% and 24% respectively. However, insulin-stimulated glycogen synthase activity was significantly higher in the aerobic exercise with caloric restriction group of impaired glucose tolerant individuals only. This suggests that for individuals with glucose intolerance, weight loss from caloric restriction combined with aerobic exercise improves fitness that may have an effect on insulin to increase glycogen synthase activity greater than weight loss from caloric restriction alone [44].

Under rare circumstances, weight loss from a combined effort of dietary modification with exercise may not result in an enhanced insulin sensitivity. For instance, a study by Oh and colleagues examined various forms of exercise (Tae-Bo, Yoga, Walking) for 6 months that consisted of various intensities performed 2 to 3 days per week for 40 minutes while maintaining a diet less than 1500 kcals/day [45]. Although significant weight loss occurred at 6 months (-5 kg) and at a one year re-evaluation (-4.3 kg), insulin resistance, as assessed by HOMA-IR, remained unchanged [45]. Although this intervention was new to a group of community residing individuals, the prescribed amount of exercise seemed inadequate and uncontrolled.

Differences in weight loss

Significant improvements in insulin sensitivity seem largely driven by weight reduction. Achieving weight loss through caloric restriction or dietary modification may be less cumbersome than energy expenditure, but the adaptations are not necessarily equivalent. When adaptations following weight loss are compared between caloric restriction and exercise, improvements in insulin stimulated glucose disposal occur similarly with greater adaptations from exercise-induced weight loss. Additionally, exercise-induced weight loss stimulates mitochondrial oxidative capacity and impacts endogenous glucose production by significantly suppressing unnecessary gluconeogenesis. The efficacy of sustained improvements in glucose metabolism may be influenced by exercise intensity as it relates to changes in body composition. Dietary weight loss effectively lowers adipose tissue considerably, but does not influence muscle metabolism similar to exercise training. However, both interventions may be used together as a powerful weight reduction technique as well as a weight management tool and a mechanism proven effective to improve glucose metabolism.
 

The ability to detect changes in insulin sensitivity is only as good as the assessment tool used. Specificity is a vital training principle necessary for exercise training and this concept is similar to detecting differences in insulin responsiveness. The hyperinsulinemic-euglycemic clamp is the gold standard for assessing insulin action, but its difficulty and cost commonly influences the use of other assessments to examine insulin action at baseline and following an intervention. Other assessments using fasting or an Oral Glucose Tolerance Test (OGTT) are typically inexpensive assessments and are commonly used in a clinic setting or epidemiological studies that identify glucose intolerance. Unfortunately, the OGTT was not designed to assess intervention-based changes in glucose metabolism The Insulin Sensitivity Index (ISI) calculations correlate quite well with the clamp but still may not be sufficient assessors of insulin action compared to the clamp [17]. Additionally, the Homeostasis Model Assessment (HOMA) equation fails to correlate with the clamp assessment, but is commonly used to assess insulin action in many investigations.

Although it may not be realistic for each investigator to use the clamp as their assessment tool, the assessment tool used during an intervention must be considered when evaluating the impact on clinical care. Therefore, it is important to consider interventions that did not induce changes in insulin action due to the limitations of the tool utilized to assess the endpoint.
 

Funding was provided by National Institutes of Health, National Institute of Diabetes, Digestive and Kidney Diseases K01 DK 64716-01, National Institute on Aging P30 GM 103325-01, and American Heart Association Grant 0335172N.