A high demand on thermoregulatory processes may challenge homoeostasis, particularly regarding glucose regulation. This has been understudied, although it might concern millions of humans. The objective of this project was to examine the isolated and combined effects of experimental short-term mild heat exposure and metabolic level on glucoregulation. Two experimental randomized crossover studies were conducted. Ten healthy young men participated in study A, which comprises four sessions in a fasting state at two metabolic levels [rest and exercise at 60% of maximal oxygen uptake (O2) for 40 min] in two environmental temperatures (warm: 31°C and control: 22°C). Each session ended with an ad libitum meal, resulting in similar energy intake across sessions. In study B, 12 healthy young men underwent two 3 h oral glucose tolerance tests (OGTTs) in warm and control environmental temperatures. Venous blood was sampled at several time points. In study A, repeated measure ANOVAs revealed higher postprandial serum glucose and insulin levels with heat exposure. Glycaemia following the OGTT was higher in the warm temperature compared with control. The kinetics of the serum glucose response to the glucose load was also affected by the environmental temperature (temperature-by-time interaction, P=0.030), with differences between the warm and control conditions observed up to 90 min after the glucose load (all P<0.033). These studies provide evidence that heat exposure alters short-term glucoregulation. The implication of this environmental factor in the physiopathology of Type 2 diabetes has yet to be investigated.

CLINICAL PERSPECTIVES

  • Millions of humans reside in regions chronically or temporarily exposed to heat, with very high prevalence of Type 2 diabetes found in populations living in warm regions (India, the Caribbean or the Middle-East).

  • We report an exaggerated increase in blood glucose after a meal or a glucose load taken in a warm environmental temperature.

  • Our results indicate the need for standardization or correction of the temperature during glucose tolerance testing. If substantiated, this would suggest that the rates of Type 2 diabetes in warm regions cannot be explained only by a given populations’ genetic background or lifestyles, but that high environmental temperature may be involved.

INTRODUCTION

Glucose regulation is a key physiological function. Any acute or chronic alteration to glucose metabolism has potentially serious health consequences. The main factors that challenge glucose homoeostasis include dietary carbohydrate restriction and ingestion and exercise. Exercise markedly increases glucose uptake, but this is usually compensated by hepatic and muscle glycogenolysis so that blood glucose remains stable or increases only slightly in most cases of exercise [1,2]. Other factors, such as high ambient temperature, also seem to modulate glucose regulation [35].

Insulin is the primary mediator of carbohydrate fuel fluxes under most circumstances. Exercise causes the translocation of glucose transporter 4 (GLUT4) glucose transporters to the cell surface, which then increases insulin sensitivity. The combination of exercise and heat stress poses a particular challenge to carbohydrate metabolism [6], although the consequences for blood glucose and its determinants during the recovery period have never been described.

Thus, our aim was to examine the isolated and combined effects of exercise and moderate heat stress on the subsequent recovery and postprandial glycaemic response. The first study presented in this report (study A) investigated the effects of ambient temperature (warm compared with control), metabolic level (exercise compared with rest) and their interaction on glucose metabolism. The second study (study B) investigated the effects of ambient temperature on glucose tolerance as assessed by a 3 h oral glucose tolerance test (OGTT).

MATERIALS AND METHODS

Study design (Figure 1)

Both studies were performed after an overnight fast, starting from 6:30 hours. The participants were instructed to abstain from high intensity exercise and alcohol intake the day before each session and to have sufficient carbohydrate intake the day before their first session and comparable food intake from one session to another.

Schematic representation of the study design

Figure 1
Schematic representation of the study design
Figure 1
Schematic representation of the study design

Study A

The study consisted of an initial evaluation and familiarization session followed by four experimental sessions using a Latin square-based randomized crossover design, with at least 4 days between sessions. The participants were tested in random order in each of the following conditions: rest in a control ambient temperature (rest–22°C), rest in a warm ambient temperature (rest–31°C), exercise in a control ambient temperature (ex–22°C) and exercise in a warm ambient temperature (ex–31°C). The exercise and corresponding rest period lasted 40 min. They were followed by 35 min of recovery or additional rest and then an ad libitum meal that lasted 25 min. In each experimental session, six blood samples were drawn: 5 min before the exercise/rest period (T0), after 20 min of exercise/rest (T20), at the end of exercise/rest (T40), 15 and 25 min after the end of exercise/rest (T55 and T70) and at the end of the meal (T100).

Study B–OGTT

The aim was to produce recent data on heat-acclimated subjects performing a standard glucose tolerance test, so as to potentially replicate data reported previously by others [7,8]. The participants underwent two OGTTs under two ambient temperatures: control (22°C) and warm (31°C), presented in a random order. They drank a solution containing 75 g glucose (Gluco75; ODILsas) within 5 min. Blood samples were drawn at 0, 30, 60, 90, 120 and 180 min (T0, T30, T60, T90, T120 and T180). The participants remained in a seated position throughout the entire OGTT and engaged in quiet activities.

Participants and ethics

Ten men with normal body mass index (21.3±1.7 kg/m2) participated in study A, three of whom also participated in study B. Twelve men (23.2±2.4 kg/m2) participated in study B (Table 1). Participants were recruited from October 2013 to November 2014. Advertisements posted on the University of the French West Indies, Guadeloupian campus and in surrounding sports centres were used. During the initial contact by phone or email, the eligibility requirements were explained to the potential participants and some of them were immediately discarded (smokers, dieters, individuals with diseases or low or high weight at birth). The others were asked to come for a second screening, during which they completed questionnaires and read the information note. The eligible volunteers were then invited to participate in the study and signed a written informed consent form before the first session. They were financially compensated for completing the study. All procedures used in the present study were in accordance with institutional guidelines, and were in accordance with the Helsinki Declaration of 2013. The study procedures were approved by the Human Subject Review Committee (CPP13-018a/2013-A01037-38). The protocol (APPA) was registered in EudraCT (2013-003206-25) and ClinicalTrials (NCT02157233).

Table 1
Characteristics of the participants
Study A (n=10)Study B (n=12)
MeanS.D.MeanS.D.
Age (years) 20.9 1.7 20.4 2.0 
Height (m) 1.80 0.06 1.79 0.07 
Body mass (kg) 69.2 7.1 74.4 8.1 
Fat mass (%) 7.2 5.3 10.1 6.0 
O2 max (ml/min/kg) 47.1 7.4 – – 
Study A (n=10)Study B (n=12)
MeanS.D.MeanS.D.
Age (years) 20.9 1.7 20.4 2.0 
Height (m) 1.80 0.06 1.79 0.07 
Body mass (kg) 69.2 7.1 74.4 8.1 
Fat mass (%) 7.2 5.3 10.1 6.0 
O2 max (ml/min/kg) 47.1 7.4 – – 

Cardiopulmonary exercise testing (study A)

An incremental exercise test to exhaustion was performed on a cycle ergometer. The increments were set at 15–20 W/min. Complete metabolic data were collected using a breath-by-breath ergospirometry system (MetaLyzer 3B, Cortex).

Physical fitness was assessed by maximal or peak oxygen uptake (O2) and the ventilatory anaerobic threshold, determined as the point at which the ventilatory equivalent for oxygen starts to rise non-linearly whereas the ventilatory equivalent for carbon dioxide remains unchanged and the respiratory gas exchange ratio does not exceed 1.0 [9]. After a minimum of 15 min of rest, the subjects were familiarized with the material and procedures used during the exercise sessions. The ergometer power target to reach 60–65% of maximal O2 was identified based on individual energetic cost and then adjusted according to the ventilatory anaerobic threshold and the metabolic parameters observed at quasi-stable state at 25–26°C for approximatively 10 min.

Exercise/rest sessions (study A)

During the rest sessions, the participants rested by lying comfortably supine on a medical examining table in the laboratory for 40 min at the same time breathing periodically (first 8 min, then the last 3 min of each 8 min bout) through a facemask equipped with a pneumotachograph, for the collection and analysis of their inspired and expired air with the ergospirometry system (MetaLyzer 3B, Cortex). During the corresponding period of the exercise sessions, they cycled on the ergometer used in the initial session and the exercise intensity was set at the power identified during the familiarization rectangular exercise test.

Ad libitum meal (study A)

Thirty-five minutes after each experimental session, the participants moved to an isolated part of the room. They were presented with a tray of 12 pieces of ham and cheese sandwiches of known energy composition (78.2 kJ/100 g, according to the French food composition table [10]), along with a glass of water equal to 4 ml/kg body weight. The meal lasted 25 min. They were instructed to eat until they were ‘comfortably full’ and were given another fresh tray of sandwich portions 10–15 min after starting the meal. Total energy intake from the meal was calculated by weighing all products before and after the breakfast meal in a separated room. The absolute energy intake did not differ significantly across sessions (258.6 ± 63.8 kJ, mean ± S.D.).

Blood sampling and analysis

In each session, an intravenous cannula was inserted into an antecubital vein after the participant's arrival at the laboratory and kept in place until the end of the session. Venous blood was collected into serum separator gel tubes at each sampling time. Blood samples were immediately placed in a refrigerator at 4°C.

At each sampling time, a subsample of blood was also collected for immediate blood lactate assay using the Lactate Pro IITM analyzer (Arkray KDK) and Lactate Pro IITM test strips. Serum glucose, insulin and cortisol were analysed within 2 h on a Cobas 6000 automatic platform (Roche) (<c501> module for glucose and <e601> module for insulin and cortisol). The total area under the curve (AUC) for lactate, glucose, insulin and cortisol was calculated according to the trapezoidal rule. Insulin sensitivity was calculated using the Matsuda index and the HOMA insulin resistance index according to the published formulas [11,12].

Clinical measurements

Before the incremental exercise test for study A, and during the session in control environmental temperature for study B, the participant's height and weight were measured and used to calculate the body mass index. Body composition was measured using an InBody S10 body composition analyzer (Biospace).

During the familiarization period and all experimental sessions, aural canal temperature was measured with a tympanic thermocouple probe (Mon-a-therm Tympanic; Mallinckrodt Medical) held in position and isolated from the external environment with cotton and surgical tape.

Statistical analyses

The required number of participants was a priori calculated using G*power 3.1 for Mac. α error probability threshold and power were 0.05 and 0.90 respectively. The calculations were performed based on the AUC, with an effect size of 0.3 and 0.85 for correspondence among repeated measures.

All results were analysed with the SPSS v.20 software package (SPSS). Data are presented as mean [95% confidence interval] or median [lower-upper quartile] according to the distribution, except in figures where S.E.M. are used.

Three-factor (in study A) and two-factor (in study B) repeated measure ANOVAs were performed to determine the effects of metabolic level (two levels: rest and exercise, in study A only), environmental temperature (two levels: control and warm), time (six levels) and their interactions on the outcome variables measured over the study period. Data were tested for sphericity using Mauchly's test and if the assumption of sphericity was violated, the Greenhouse–Geisser correction was undertaken to adjust the degrees of freedom. As the single effect of time was significant for all the variables, one-factor ANOVA was performed at each time of measurement, followed by Tukey's post hoc tests to identify mean differences among conditions. The AUC were analysed with two-factor (study A: two levels of metabolic activity and two levels of environmental temperature) and one-factor (in study B) ANOVAs followed by Tukey's post hoc. The sensitivity indices (in study B) being non-normally distributed, Wilcoxon’s rank-sum tests were used to assess the effect of environmental temperatures on the Matsuda and HOMA indices.

Correlation analyses were performed using Spearman's non-parametric rank correlation coefficients to test associations between glucose values at different time points.

RESULTS

No significant variation in tympanic temperature was observed throughout the sessions, indicating that the thermoregulatory process were efficient and not seriously challenged by the environmental conditions.

Study A focused on the understanding of exercise performed in the heat on glucose metabolism. The metabolic and cardiovascular load was increased from rest to exercise. The warm environmental temperature resulted in a significantly increased heart rate at exercise (at T40, single effects of metabolic level and environmental temperature: P<0.0001, interaction: P=0.037, Table 2), with no significant difference with the control temperature on O2 (P=0.720). The significant interaction (P=0.021) of environmental temperature and metabolic level on the respiratory exchange ratio also revealed a greater increase at 31°C than at 22°C, as another marker of increased metabolic solicitation in the exercise in warm conditions. Lactataemia (Figure 2A) was higher in the exercise sessions than the rest sessions (2.0 [1.7–2.4] mmol/l compared with 1.2 [1.1–1.4], P<0.0001) and higher in the control (22°C) than at warm (31°C) temperature (1.7 [1.5–2.0] mmol/l compared with 1.5 [1.3–1.7], P=0.006). The interactions of environmental temperature and metabolic level with time were also significant (P=0.009 and P<0.0001 respectively).

Table 2
Indicators of the metabolic and cardiovascular load during the experimental sessions in study A

* and **: P<0.05 and P<0.01 for the difference with the rest condition (at the same environmental temperature); † and ††: P<0.05 and P<0.01 for the difference with the 22°C condition (at the same level of metabolic activity).

rest-22°Crest-31°Cex-22°Cex-31°C
MeanS.D.MeanS.D.MeanS.D.MeanS.D.
Heart rate (bpm) T40 57.1 3.4 62.6 4.8 148.5** 12.2 166.6††** 14.9 
O2(l/min) T40 0.24 0.05 0.25 0.04 2.04** 0.14 1.98** 0.13 
Respiratory exchange ratio T40 0.84 0.08 0.77 0.07 0.98** 0.03 1.00** 0.05 
Energy expenditure (kJ)  10.8 1.5 11.1 1.5 94.8** 5.6 92.0** 6.2 
rest-22°Crest-31°Cex-22°Cex-31°C
MeanS.D.MeanS.D.MeanS.D.MeanS.D.
Heart rate (bpm) T40 57.1 3.4 62.6 4.8 148.5** 12.2 166.6††** 14.9 
O2(l/min) T40 0.24 0.05 0.25 0.04 2.04** 0.14 1.98** 0.13 
Respiratory exchange ratio T40 0.84 0.08 0.77 0.07 0.98** 0.03 1.00** 0.05 
Energy expenditure (kJ)  10.8 1.5 11.1 1.5 94.8** 5.6 92.0** 6.2 
Figure 2

Metabolic responses during the rest session at 22°C ambient temperature (white triangles), rest session at 31°C (black triangles), exercise session at 22°C (white squares) and exercise session at 31°C (black squares). Means in the same column with different letters are different (one-way ANOVA for condition effect, Tukey's post hoc, P<0.05). Non-significant ‘P ’  are not reported; n=10. (A) Blood lactate; time, P=0.001; temperature, P<0.006; metabolic level, P<0.0001; metabolic level-by-time, P<0.0001; temperature-by-time, P=0.009; metabolic level × time × temperature, P=0.017. (B) Serum glucose; time, P<0.0001; temperature, P<0.0001; metabolic level-by-temperature, P=0.024; temperature × time, P=0.012. (C) Serum insulin; time, P<0.0001; temperature, P<0.012; temperature-by-time, P=0.007. (D) Serum cortisol; metabolic level, P=0.016; metabolic level-by-time, P=0.003.

Figure 2

Metabolic responses during the rest session at 22°C ambient temperature (white triangles), rest session at 31°C (black triangles), exercise session at 22°C (white squares) and exercise session at 31°C (black squares). Means in the same column with different letters are different (one-way ANOVA for condition effect, Tukey's post hoc, P<0.05). Non-significant ‘P ’  are not reported; n=10. (A) Blood lactate; time, P=0.001; temperature, P<0.006; metabolic level, P<0.0001; metabolic level-by-time, P<0.0001; temperature-by-time, P=0.009; metabolic level × time × temperature, P=0.017. (B) Serum glucose; time, P<0.0001; temperature, P<0.0001; metabolic level-by-temperature, P=0.024; temperature × time, P=0.012. (C) Serum insulin; time, P<0.0001; temperature, P<0.012; temperature-by-time, P=0.007. (D) Serum cortisol; metabolic level, P=0.016; metabolic level-by-time, P=0.003.

Glycaemia and insulinaemia (Figures 2B and 2C) were affected by the environmental temperature (glycaemia: 4.5 [4.4–4.7] mmol/l compared with 4.9 [4.7–5.1], P<0.0001; insulinaemia: 68.6 [50.0–87.2] pmol/l compared with 98.5 [68.6–128.4], P=0.012, in control and warm environment respectively), as well as by time (temperature-by-time interaction: P=0.012 and P=0.007 respectively). Environmental temperature and the temperature-by-metabolic level interaction also had significant effects on glucose AUC (P<0.0001 and P=0.025 respectively).

The single effect of metabolic level on glycaemia and insulinaemia was not significant, nor was their interaction with time (all P > 0.208). The temperature-by-metabolic level interaction significantly affected glycaemia (P=0.024), but not insulinaemia (P=0.648). Insulin AUC was affected by the environmental temperature (P=0.022). The metabolic level and the temperature-by-metabolic level interaction effects on insulin AUC were not significant (all P > 0.482).

Glucose values at T0 were significantly correlated with other glucose values. The strength of the association diminished with increasing time exposure to the experimental conditions (from rho=0.718 at T20 to rho=0.319 at T100, all P<0.045). On the other hand, postprandial glucose values were not significantly associated with glucose measured from blood sampled immediately before the meal (rho=0.299, P=0.061) and the strongest association was observed with T20 (rho=0.425, P=0.006).

Cortisolaemia was significantly higher in the exercise sessions than the rest sessions (499 [423–576] mmol/l compared with 384 [336–433], P=0.016, Figure 2D). The metabolic level-by-time interaction was significant (P=0.003). No effect involving environmental temperature reached significance (single effect: P=0.686, interactions with time, environmental temperature or both, all P > 0.095).

In study B, glycaemia (Figure 3A) was higher when the OGTT was performed in the warm environmental temperature as compared with the control: 5.7 [5.4–5.9] mmol/l compared with 5.0 [4.5–5.6], P=0.026. ANOVAs revealed that the kinetics of glucose response to the glucose load was also affected by the environmental temperature, with a significant temperature-by-time interaction (P=0.030). Differences between the warm and control conditions were evidenced at 30, 60 and 90 min after the glucose load (all P<0.033), with no difference before the glucose load, 120 or 180 min after glucose ingestion (all P > 0.325).

Figure 3

Metabolic responses during the OGTT performed at 22°C (white triangles) and 31°C (black triangles) ambient temperature; n=12.(A) Serum glucose; time, P<0.0001; temperature, P=0.026; temperature-by-time, P=0.030. (B) Serum insulin; time, P<0.0001. (C) Serum cortisol; time, P<0.0001; temperature, P=0.001; temperature-by-time, P=0.046.

Figure 3

Metabolic responses during the OGTT performed at 22°C (white triangles) and 31°C (black triangles) ambient temperature; n=12.(A) Serum glucose; time, P<0.0001; temperature, P=0.026; temperature-by-time, P=0.030. (B) Serum insulin; time, P<0.0001. (C) Serum cortisol; time, P<0.0001; temperature, P=0.001; temperature-by-time, P=0.046.

Insulinaemia (Figure 3B) was not affected by the environmental temperature (single effect: P=0.701, environmental temperature-by-time interaction: P=0.388). Insulin AUC did not differ with environmental temperature (single effect, P=0.652).

The composite insulin sensitivity indices were not significantly modified by the environmental temperature. The Matsuda index values were 7.5 [5.7–12.7] and 6.6 [5.9–7.5] (P=0.308) at 22°C and 31°C respectively. The HOMA insulin resistance index values were 1.4 [1.0–2.1] and 1.2 [1.1–1.7] (P=0.308) at 22°C and 31°C respectively.

There were single effects of environmental temperature (P<0.0001) and time (P<0.0001) on cortisolaemia (Figure 3C). The temperature-by-time interaction was significant (P=0.046).

DISCUSSION

We hypothesized that metabolic level and environmental temperature would influence glucose regulation, separately or by interaction. Our main findings are that (1) a meal or a glucose load elicits an exaggerated increase in blood glucose in a warm environment and (2) 40 min of preliminary exercise provides no significant short-term improvement in this temperature-related alteration in glucose metabolism.

Increased hepatic glucose production via augmented sympathetic stimulation is a first track for the interpretation of blunted glucose tolerance in the warmth. The liver is indeed critical to blood glucose homoeostasis, especially when exercise and heat stress are combined. Hargreaves et al. [13] demonstrated that hyperglycaemia during exercise at 40°C is caused by an increase in liver glucose output. The association between glucose data observed at T100 and those observed earlier in the session suggests some early drive of the postprandial glucose excursion, as if the meal would accentuate the preliminary glucose variation. It supports the liver hypothesis since such variations in hepatic glucose production occur quite early. The kidneys also contribute to this regulation [14]. It has been demonstrated in dogs that an infusion with cortisol, glucagon and adrenaline (epinephrine) increases renal glucose release [15]. In our study, cortisol rose during the exercise in both environmental temperatures, and decreased during the corresponding period at rest. Thus, the high level of glucose during the exercise and early recovery in the warm environment was very likely due to increased hepatic and renal glycogenolysis and gluconeogenesis driven by sympathetic stimulation reinforced by adrenaline and cortisol. However, the discrepancies between the patterns of glucose and cortisol levels suggest that the increased glucose release from these organs did not make a major contribution to the relative hyperglycaemia observed in the early postprandial state. Moreover, renal and hepatic glucose release is crucial in the fasting state and during exercise, concomitantly with high quantities of the gluconeogenic precursors: glutamine, glycerol and lactate. Intestinal territories are another potential source of glucose output, since splanchnic glucose is increased in animal during exercise and heat stress [16]. Because we observed large environmental temperature-related blood glucose differences in the postprandial state (study A) and after glucose ingestion (study B), it is unlikely that increased intestinal gluconeogenesis was the primary explanation for our results. Another possible gut phenomenon is the absorption rate. Because heat stress shifts blood volume from thoracic and splanchnic regions presumably to aid in heat dissipation [17], these adaptations seem likely to lead to underestimation of glucose tolerance limitation in a hot environment.

A second important trail to consider for the understanding of our results is the glucose uptake by muscle or other tissues. It is widely acknowledged that the environmental temperature influences substrate use at rest and during exercise [18,19], with a shift toward increased carbohydrate metabolism during exercise and heat stress and a concomitant decrease in fat oxidation. Muscle glycogenolysis is increased in the heat through direct mechanisms [20], as well as under catecholaminergic influence [21]. Increased circulating adrenaline levels have frequently been observed during exercise at intensities and environmental temperatures comparable with those of our study [22] and are known to activate glycogenolysis. Adrenaline and cortisol levels are generally analysed together since cortisol has a permissive effect on adrenaline. Accordingly, the lactataemia and cortisolaemia in study A were higher during and after exercise (up to T55) than rest. Lactataemia was higher during the exercise (T20 and T40) performed in a warm environment than in control, although there was no significant temperature-related difference in cortisol. It can be interpreted as evidence that the increased arterialization in the heat and at exercise does not bias these studies. This also suggests that glucose or glycogen utilization was increased during exercise, in particular when exercise was performed in the warm environment. However, the participants in our studies were all acclimated and this probably reduced sympathetic stimulation and the adrenaline release in the warm condition, which would explain the absence of a temperature-related difference in the respiratory exchange ratio (as a reflection of substrate use) during exercise. Cortisol release appeared more related to the metabolic level and was either slightly or not modified by the environmental temperature in study A. In study B, the diurnal rhythm of cortisol was normal, in both the warm and control sessions. A shift toward higher values at 22°C was observed, probably reflecting the relative stress of this environmental temperature for the heat-acclimated participants.

Muscle glucose utilization and adipose tissue glucose storage are also widely influenced by the mechanical translocation of glucose transporters, especially GLUT4 to the cell membrane. Decreased translocation inhibits glucose entry and subsequent utilization or storage. The effect of heat, combined with rest or exercise, on insulin-dependent and -independent GLUT4 translocation is not straightforward. In vivo and in vitro studies provide evidence that an elevated muscle temperature per se stimulates muscle glucose uptake through amplification of the phosphorylation of AMP-activated protein kinase [23] and that this results in enhanced glycogenesis when heat is locally applied [24]. However, glycogen resynthesis after exercise is slowed down with whole-body heat exposure [25], the latter observation fitting very well with ours in the sense that it suggests some form of glucose intolerance, if not insulin resistance. The exact mechanisms involved cannot be specified from this project, but direct effect of heat on the pancreatic function can be discarded based on the results of study B, in particular the absence of significant difference related to the environmental temperature on insulinaemia. There are hypotheses to explain transiently blunted glucose tolerance in the heat. A systemic deleterious effect of heat on insulin signalling or GLUT4 translocation mechanisms is a candidate. A change in the sympathovagal balance might also occur in warm temperatures, with or without the implication of the hypothalamic–pituitary–adrenal axis, which is known to influence glucose tolerance [26], could occur in the heat. If this is the case, an exacerbated response would occur at exercise, characterized by a greater decrease in parasympathetic activity and an increase in sympathetic activity. This would explain why the preliminary 40-min-exercise did not lead to a short-term improvement in the altered temperature-related glucose metabolism, although this effect could be expected [27].

The higher blood glucose values during the sessions in moderate heat could also be related to changes in the peripheral blood flow. The arteriovenous glucose difference is expected to be reduced in the territories of the blood sample in the sessions performed at 31°C due to increased arterialization of the antecubital blood in relation to the opening of the thermoregulatory anastomoses leading to increased peripheral blood flow [4]. These vascular considerations have been highlighted in the very few studies on the question, so that authors consider as ‘apparent’ the impairment of glucose tolerance at 30–35°C as compared with 20–25°C [4,5] on studies at rest. We observed (in study A) and others reported [35,8] an elevation of postprandial (or post glucose load) insulin concentration. This bears out that the vascular adaptations to the environmental temperature would explain the entire phenomenon.

So, despite limitations in sample size and duration of the postprandial observation period, the study A results can be interpreted as demonstrating that environmental temperature has a stronger short-term effect on glucose regulation than metabolic level, although the latter is widely acknowledged as a strong determinant of metabolic health, including the prevention of Type 2 diabetes.

To the best of our knowledge, this is the first randomized crossover study including exercise and rest which is focused on the association between glucose tolerance and warm ambient temperature. In this way, our observations from study A and B are complementary to previous isolated studies reporting increased glycaemia and/or insulinaemia in a warm environment [35,8,28]. They are also original to suggest that glucoregulation appears to be disrupted by high environmental temperature in acclimated individuals and that preliminary exercise does not provide significant short-term improvement in this temperature-related alteration in glucose metabolism.

Whether or not this effect is time-limited and related to peripheral vascular adaptations, it is likely to have strong impact on clinical practices and our findings may contribute to a greater understanding of glucose tolerance and Type 2 diabetes in populations living in tropical and equatorial regions or exposed to warm temperatures. If our observations are confirmed and identified as acute only, this would indicate the need for standardizing or correcting for the temperature during glucose tolerance testing. This point has been raised by others reporting doubled glucose intolerance rates on warmer days in a temperate zone [28], though not yet acted upon. Given that elevated blood glucose induces oxidative stress and inflammation and alters insulin sensitivity, postprandial glycaemia may ultimately have a strong impact on the pathophysiology of metabolic and cardiovascular disease [29]. Long-term consequences can theoretically be expected. In our studies on young healthy trained men acclimated to the heat however, although the AUC for insulin was higher during the sessions performed in a warm environment, the effect was not present at all-time points, and we did not observe significant difference in insulin values or insulin sensitivity during the OGTT. Collection of data on similar protocols with participants with less physical fitness and established metabolic impairments would be of interest.

It is noteworthy that a very high prevalence of Type 2 diabetes is found in populations living in warm regions like India, the Caribbean or the Middle East. If future studies demonstrate chronicity, this would suggest that the prevalence cannot be explained only by a given populations’ genetic background or lifestyles, but that high mean temperature may be involved. The present study clearly indicates the need for further investigation of the mechanistic links between heat exposure and glyco-regulation.

These studies provide evidence that heat exposure alters short-term glucoregulation. Whether environmental temperature has a chronic impact on the pathophysiology of Type 2 diabetes has not been determined. We propose that it should be considered as a potential contributor to the diabetes epidemic.

AUTHOR CONTRIBUTION

Cécile Faure and Sophie Antoine-Jonville wrote manuscript, researched and analysed data, Keyne Charlot researched data and contributed to data analysis, Stéphane Henri researched and analysed data, Olivier Hue contributed to study design and discussion and reviewed/edited manuscript, Marie-dominique Hardy-Dessources contributed to study design and reviewed/edited manuscript. Sophie Antoine-Jonville is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

We thank the participants and the nurse and physician (Véronique Conord, Thibault Philippe) for their great involvement. We thank Cathy Carmeni for her excellent revision and suggestions.

FUNDING

This work was supported by the European Social Fund and Region Guadeloupe [grant number CR/12-116]; the French Ministry of Oversea Territories [grant number 0123-C001-D971/2013]; and the European Regional Development Fund [grant number 1/1.4/-31793].

Abbreviations

     
  • AUC

    area under the curve

  •  
  • GLUT4

    glucose transporter 4

  •  
  • OGTT

    oral glucose tolerance test

  •  
  • O2

    oxygen uptake

References

References
1
Brooks
 
G.A.
 
Mammalian fuel utilization during sustained exercise
Comp. Biochem. Physiol. B Biochem. Mol. Biol.
1998
, vol. 
120
 (pg. 
89
-
107
)
[PubMed]
2
Wahren
 
J.
Felig
 
P.
Ahlborg
 
G.
Jorfeldt
 
L.
 
Glucose metabolism during leg exercise in man
J. Clin. Invest.
1971
, vol. 
50
 (pg. 
2715
-
2725
)
[PubMed]
3
Akanji
 
A.O.
Bruce
 
M.
Frayn
 
K.
Hockaday
 
T.D.
Kaddaha
 
G.M.
 
Oral glucose tolerance and ambient temperature in non-diabetic subjects
Diabetologia
1987
, vol. 
30
 (pg. 
431
-
433
)
[PubMed]
4
Frayn
 
K.N.
Whyte
 
P.L.
Benson
 
H.A.
Earl
 
D.J.
Smith
 
H.A.
 
Changes in forearm blood flow at elevated ambient temperature and their role in the apparent impairment of glucose tolerance
Clin. Sci. (Lond.)
1989
, vol. 
76
 (pg. 
323
-
328
)
[PubMed]
5
Moses
 
R.G.
Patterson
 
M.J.
Regan
 
J.M.
Chaunchaiyakul
 
R.
Taylor
 
N.A.
Jenkins
 
A.B.
 
A non-linear effect of ambient temperature on apparent glucose tolerance
Diabetes Res. Clin. Pract.
1997
, vol. 
36
 (pg. 
35
-
40
)
[PubMed]
6
Febbraio
 
M.A.
 
Alterations in energy metabolism during exercise and heat stress
Sports Med.
2001
, vol. 
31
 (pg. 
47
-
59
)
[PubMed]
7
Dumke
 
C.L.
Slivka
 
D.R.
Cuddy
 
J.S.
Hailes
 
W.S.
Rose
 
S.M.
Ruby
 
B.C.
 
The effect of environmental temperature on glucose and Insulin after an oral glucose tolerance test in healthy young men
Wilderness Environ Med
2015
, vol. 
26
 (pg. 
335
-
342
)
8
Akanji
 
A.O.
Oputa
 
R.N.
 
The effect of ambient temperature on glucose tolerance
Diabet. Med.
1991
, vol. 
8
 (pg. 
946
-
948
)
[PubMed]
9
Wasserman
 
K.
Whipp
 
B.J.
Koyal
 
S.N.
Beaver
 
W.L.
 
Anaerobic threshold and respiratory gas exchange during exercise
J. Appl. Physiol.
1973
, vol. 
35
 (pg. 
236
-
243
)
[PubMed]
10
ANSES
2013
 
Ciqual tables [Internet]. Available from: https://pro.anses.fr/TableCIQUAL/index.htm
11
Matsuda
 
M.
DeFronzo
 
R.A.
 
Insulin sensitivity indices obtained from oral glucose tolerance testing: comparison with the euglycemic insulin clamp
Diabetes Care
1999
, vol. 
22
 (pg. 
1462
-
1470
)
[PubMed]
12
Matthews
 
D.R.
Hosker
 
J.P.
Rudenski
 
A.S.
Naylor
 
B.A.
Treacher
 
D.F.
Turner
 
R.C.
 
Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man
Diabetologia
1985
, vol. 
28
 (pg. 
412
-
419
)
[PubMed]
13
Hargreaves
 
M.
Angus
 
D.
Howlett
 
K.
Conus
 
N.M.
Febbraio
 
M.
 
Effect of heat stress on glucose kinetics during exercise
J. Appl. Physiol.
1996
, vol. 
81
 (pg. 
1594
-
1597
)
[PubMed]
14
Gerich
 
J.E.
Meyer
 
C.
Woerle
 
H.J.
Stumvoll
 
M.
 
Renal gluconeogenesis: its importance in human glucose homeostasis
Diabetes Care
2001
, vol. 
24
 (pg. 
382
-
391
)
[PubMed]
15
McGuinness
 
O.P.
Fugiwara
 
T.
Murrell
 
S.
Bracy
 
D.
Neal
 
D.
O'Connor
 
D.
Cherrington
 
A.D.
 
Impact of chronic stress hormone infusion on hepatic carbohydrate metabolism in the conscious dog
Am. J. Physiol.
1993
, vol. 
265
 
Pt 1
(pg. 
E314
-
E322
)
[PubMed]
16
Rowell
 
L.B.
Brengelmann
 
G.L.
Blackmon
 
J.R.
Twiss
 
R.D.
Kusumi
 
F.
 
Splanchnic blood flow and metabolism in heat-stressed man
J. Appl. Physiol.
1968
, vol. 
24
 (pg. 
475
-
484
)
[PubMed]
17
Crandall
 
C.G.
Wilson
 
T.E.
Marving
 
J.
Vogelsang
 
T.W.
Kjaer
 
A.
Hesse
 
B.
Secher
 
N.H.
 
Effects of passive heating on central blood volume and ventricular dimensions in humans
J. Physiol.
2008
, vol. 
586
 (pg. 
293
-
301
)
[PubMed]
18
Jensen
 
T.E.
Richter
 
E.A.
 
Regulation of glucose and glycogen metabolism during and after exercise
J. Physiol.
2012
, vol. 
590
 
Pt 5
(pg. 
1069
-
1076
)
[PubMed]
19
Jeukendrup
 
A.E.
 
Modulation of carbohydrate and fat utilization by diet, exercise and environment
Biochem. Soc. Trans.
2003
, vol. 
31
 (pg. 
1270
-
1273
)
[PubMed]
20
Starkie
 
R.L.
Hargreaves
 
M.
Lambert
 
D.L.
Proietto
 
J.
Febbraio
 
M.A.
 
Effect of temperature on muscle metabolism during submaximal exercise in humans
Exp. Physiol.
1999
, vol. 
84
 (pg. 
775
-
784
)
[PubMed]
21
Febbraio
 
M.A.
Snow
 
R.J.
Hargreaves
 
M.
Stathis
 
C.G.
Martin
 
I.K.
Carey
 
M.F.
 
Muscle metabolism during exercise and heat stress in trained men: effect of acclimation
J. Appl. Physiol.
1994
, vol. 
76
 (pg. 
589
-
597
)
[PubMed]
22
Febbraio
 
M.A.
Lambert
 
D.L.
Starkie
 
R.L.
Proietto
 
J.
Hargreaves
 
M.
 
Effect of epinephrine on muscle glycogenolysis during exercise in trained men
J. Appl. Physiol.
1998
, vol. 
84
 (pg. 
465
-
470
)
[PubMed]
23
Koshinaka
 
K.
Kawamoto
 
E.
Abe
 
N.
Toshinai
 
K.
Nakazato
 
M.
Kawanaka
 
K.
 
Elevation of muscle temperature stimulates muscle glucose uptake in vivo and in vitro
J. Physiol. Sci.
2013
, vol. 
63
 (pg. 
409
-
418
)
[PubMed]
24
Slivka
 
D.
Tucker
 
T.
Cuddy
 
J.
Hailes
 
W.
Ruby
 
B.
 
Local heat application enhances glycogenesis
Appl. Physiol. Nutr. Metab.
2012
, vol. 
37
 (pg. 
247
-
251
)
[PubMed]
25
Naperalsky
 
M.
Ruby
 
B.
Slivka
 
D.
 
Environmental temperature and glycogen resynthesis
Int. J. Sports. Med.
2010
, vol. 
31
 (pg. 
561
-
566
)
[PubMed]
26
Chrousos
 
G.P.
 
The role of stress and the hypothalamic–pituitary–adrenal axis in the pathogenesis of the metabolic syndrome: neuro-endocrine and target tissue-related causes
Int. J. Obes. Relat. Metab. Disord.
2000
, vol. 
24
 
Suppl 2
(pg. 
S50
-
S55
)
[PubMed]
27
Lund
 
S.
Holman
 
G.D.
Schmitz
 
O.
Pedersen
 
O.
 
Contraction stimulates translocation of glucose transporter GLUT4 in skeletal muscle through a mechanism distinct from that of insulin
Proc. Natl. Acad. Sci. U.S.A.
1995
, vol. 
92
 (pg. 
5817
-
5821
)
[PubMed]
28
Schmidt
 
M.I.
Matos
 
M.C.
Branchtein
 
L.
Reichelt
 
A.J.
Mengue
 
S.S.
Iochida
 
L.C.
Duncan
 
B.B.
 
Variation in glucose tolerance with ambient temperature
Lancet
1994
, vol. 
344
 (pg. 
1054
-
1055
)
[PubMed]
29
Cavalot
 
F.
Pagliarino
 
A.
Valle
 
M.
Di Martino
 
L.
Bonomo
 
K.
Massucco
 
P.
Anfossi
 
G.
Trovati
 
M.
 
Postprandial blood glucose predicts cardiovascular events and all-cause mortality in type 2 diabetes in a 14-year follow-up: lessons from the San Luigi Gonzaga Diabetes Study
Diabetes Care
2011
, vol. 
34
 (pg. 
2237
-
2243
)
[PubMed]

Author notes

Trial registration: EudraCT (2013-003206-25); ClinicalTrials (NCT02157233).