New research sheds light on metabolites, exercise, and Type 2 diabetes risk

A recent study shows that increased exercise improves metabolic markers and takes a broad look at the molecules impacted by physical activity


The Study:

Metabolite Signature of Physical Activity and the Risk of Type 2 Diabetes in 7271 Men
Published: Metabolites
Where: University of Eastern Finland

The Takeaway:

Habitual physical activity (60-90+ min/week), or an increase of at least 30 min of physical activity/week, can improve glucose, insulin, and molecules associated with disease.

What It Looked At:

Physical activity (PA) can produce an incredible amount of change in the body. The sheer number of reactions co-occurring is impressive. How all these processes link together to form adaptations can be challenging for researchers to discern.

The first step is to start with what we already know to identify the knowledge gaps. Previous research shows that PA improves glucose and insulin markers. We also know that reduced fasting glucose, improved insulin sensitivity, and increased insulin secretion can reduce the risk of Type 2 diabetes. However, multiple physiologic processes connect PA with glucose and ultimately reduce disease risk. For example, during exercise, muscle contraction stimulates the activation of a cellular signaling pathway (AMPK) which increases glucose uptake (via translocation of GLUT4).  This process improves insulin sensitivity and reduces disease risk.

Some of these intermediate processes involve metabolites, intermediate products of metabolic reactions. Metabolites are small molecules that can help reveal the mechanisms and pathways that lead to disease. This helps us better understand the evolution of the condition and how to prevent it.

Previous research has used this approach to identify multiple diseases and lifestyle factors, along with their associated metabolites: diabetes (phenylalanine, hexose), cancer subtypes (chlorogenic acid and polyphenols), obesity (acylcarnitines, biogenic amines, phosphatidylcholines), and smoking (glutamate, arginine).

This large study included 7,271 Finnish men and measured PA, glucose, insulin, 1,260 metabolites, and incidence of Type 2 diabetes. The researchers investigated the links (metabolites) between each of these factors at baseline and after a follow-up period of 7.8 years. The metabolites identified in this study may provide insight into the mechanisms by which PA can lower the risk of certain diseases (cardiovascular disease, diabetes, hypertension, metabolic syndrome).

What They Did:

The 7,271 participants were all middle-aged and older Finnish men (45-73) without diabetes at the start of the study. Researchers assessed several factors at baseline and again in 7.8 years, including PA, waist circumference, glucose and insulin measures, and 1,260 metabolites. They also divided participants into four categories based on their PA:

  1. Little to no PA
  2. Occasional PA or in the context of other hobbies
  3. 30+ min of PA, 2 times or less per week
  4. 30+ min of PA, 3 times or more per week.

After the 7.8-year follow-up, researchers tracked how many participants increased, decreased, or did not change their PA level from baseline. They used statistical analyses to assess associations between PA and: glucose/insulin, metabolites, and incidence of Type 2 diabetes.

What It Found:

Physical Activity at Baseline

The study found that, after adjusting for confounding variables, participants who exercised 30+ minutes, one or two times per week, had lower two-hour glucose along with increased insulin sensitivity and secretion (compared to people who had no PA or were only active in the context of other hobbies). This agrees with previous research, showing that habitual PA can improve metabolic health. This association is due to the ability of exercise to enhance skeletal muscle glucose uptake via both insulin-dependent and insulin-independent pathways.

Changes in Physical Activity

After 7.8 years, the participants conducted follow-up tests and questionnaires. When compared to their level of activity at baseline, 27% of participants increased their level of PA. These participants reaped the benefits of their increased activity in multiple ways (more details below). At follow-up, 55% of participants did not change their PA level. It should be noted that 50% of the participants at baseline were already in the highest level of PA, so no change indicates that they maintained their healthy exercise habits. The 18% of participants that decreased their PA level after 7.8 years showed worse glucose and insulin markers.

“While only 30 minutes per week may not be enough to reduce the risk of Type 2 diabetes, the study does show that even minor improvements can significantly benefit metabolic health.”

People who increased their level of PA over the course of the study (27% of participants) reduced their fasting glucose and fasting insulin concentration while increasing their insulin sensitivity and secretion. However, these participants did not have a statistically significant reduction in risk of Type 2 diabetes after adjusting for factors such as BMI, smoking, and age. This lack of association is a bit surprising given that previous research has shown PA to decrease the risk of Type 2 diabetes by as much as 50%.

However, that study assessed PA at 150 minutes per week, more than the 90 minutes per week in the current study. Thus, the lack of results from the current research may be due to small increments of PA being insufficient to produce a significant change. For example, advancing from the third to the fourth category of PA only required an increase of 30 min per week of activity. This change may not be enough to reduce the risk of disease. If the study had investigated more significant changes in PA, the results might have shown a considerable change.


One novel aspect of this study is that it found 198 metabolites significantly associated with PA levels. This means that individuals that participated in a high level of PA had significantly higher (polyunsaturated fatty acids and lysophosphatidylcholines) or lower (sphingolipids, glycerolipids, bilirubins, carotenoids, carboxylic acids) concentration of metabolites than their non-active counterparts (compared to baseline measurements). Previous studies have been smaller and found up to 20 metabolites associated with PA.

The metabolite groups that had the greatest difference between high and low PA groups were:

  • glycerophospholipids (28%)
  • amino acids (15%)
  • other metabolites (14%)
  • glycerolipids (8%)
  • polyunsaturated fatty acids (6%)
  • sphingolipids (6%)
  • other organic compounds (6%)
  • acylcarnitines (5%)
  • carbohydrates (4%)
  • bilirubins (3
  • carboxylic acids (3%)
  • carotenoids (2%)

While the details of the contribution of all 198 significant metabolites are beyond the scope of this article, here are a couple of noteworthy metabolites associated with glucose, insulin, and Type 2 diabetes.

Polyunsaturated fatty acids, lysophosphatidylcholines (a subclass of glycerophospholipids), certain amino acids, and carotenoids are associated with a decreased risk of Type 2 diabetes.  Lysophosphatidylcholines and some amino acids are linked to decreased glucose levels. Lysophosphatidylcholines can increase insulin sensitivity.

Why It Matters:

This research is promising for those already physically active as well as those just starting on their fitness journey. For the people already habitually exercising more than 60 minutes per week, this study should be reassuring that their activity is likely benefiting their health. The improvements in two-hour glucose, insulin sensitivity, and insulin secretion produced by this level of PA are all positive for health.

For individuals who are just starting their fitness journey, the results of this study should be encouraging. It is promising that even small changes in PA level (increasing 30 minutes per week) can induce positive changes in fasting glucose, fasting insulin, insulin sensitivity, and insulin secretion. While only 30 minutes per week may not be enough to reduce the risk of Type 2 diabetes, the study does show that even minor improvements can significantly benefit metabolic health.

The metabolite segment of this study will serve as a starting point for future research. The ability to identify a significant difference in metabolites with varying levels of PA is promising because it gives researchers a better idea of which metabolites to focus on in future studies. While this study could not separate the impacts of diet vs. PA, it effectively identified multiple groups of metabolites associated with PA. Future research can use these groups as a starting point and investigate varying levels of PA, diet, and gut microbiota. Eventually, the compilation of all of this data will aid in preventing disease and provide additional insight into the mechanisms behind PA, glucose, and insulin.

What questions do we still have?

This study was a significant undertaking with an impressive volume of participants and variables. However, the authors acknowledged the considerable limitation that their choice of PA measure had on the study’s applicability. Even the highest level of PA in this study (30+ min regularly 3 times/week= 90 min of PA) does not meet the current PA guidelines (150 min/wk). This begs the question of what would the results have been if the study used individuals who met the current guidelines or an even higher PA group like elite athletes? It would make sense to infer that the current guidelines are the threshold for metabolic changes. So, it is possible that the entire metabolic profile would change for individuals who hit the 150 min/week threshold. It would be interesting to see changes in these profiles, and associations with disease risk, stratified by PA level. In addition, PA was assessed via questionnaires at baseline and after 7.8 years.  These questionnaires are not perfect and not as reliable as other measures of PA (e.g., accelerometers).  Thus, how PA was measured and the minor incremental changes in PA were potential sources of error. Using more significant jumps in the amount of PA—say, 0 minutes to 120 to 300—might reveal more insights.

While it was interesting that PA was associated with such a high number of metabolites, we leave the study wondering which markers are associated with a specific type of exercise. Previous research has shown that different exercise types produce divergent adaptations. For example, endurance exercise can increase your VO2max, while resistance training is likely to increase muscle size. However, this study did not collect data on the type or intensity of exercise, so it is unclear which metabolic markers are associated with endurance exercise and which metabolites result from resistance training. Previous studies have shown that the metabolite profiles between resistance and endurance exercise were significantly different.

Another question that we leave this study with is the potentially very significant confounding impacts of diet. Many of the metabolic markers that were found to be significant (choline plasmalogens, lysophosphatidylcholines, polyunsaturated fatty acids, carotenoids, indolelactate, indolepropionate) are heavily dependent on diet and modulated by the microbiota in the gut. Without a dietary analysis of the participants, it is impossible to know whether these markers result from PA or a healthy diet.

This study provides further confirmation that at least 60 min of PA each week (or an increase of at least 30 min per week of PA) can improve markers of metabolic health such as fasting glucose, fasting insulin, insulin sensitivity, and insulin secretion. This research also provides a foundation for building metabolite research and garnering knowledge regarding the metabolite links between PA, glucose, insulin, and disease risk. What can you do while waiting for this additional research? Keep exercising and improving your metabolic health.