Journal of Applied Physiology


The relative roles of circulatory glucose, muscle glycogen, and lipids in shivering thermogenesis are unclear. Using a combination of indirect calorimetry and stable isotope methodology ([U-13C]glucose ingestion), we have quantified the oxidation rates of these substrates in men acutely exposed to cold for 2 h (liquid conditioned suit perfused with 10°C water). Cold exposure stimulated heat production by 2.6-fold and increased the oxidation of plasma glucose from 39.4 ± 2.4 to 93.9 ± 5.5 mg/min (+138%), of muscle glycogen from 126.6 ± 7.8 to 264.2 ± 36.9 mg glucosyl units/min (+109%), and of lipids from 46.9 ± 3.2 to 176.5 ± 17.3 mg/min (+376%). Despite the observed increase in plasma glucose oxidation, this fuel only supplied 10% of the energy for heat generation. The major source of carbohydrate was muscle glycogen (75% of all glucose oxidized), and lipids produced as much heat as all other fuels combined. During prolonged, low-intensity shivering, we conclude that total heat production is unequally shared among lipids (50%), muscle glycogen (30%), plasma glucose (10%), and proteins (10%). Therefore, future research should focus on lipids and muscle glycogen that provide most of the energy for heat production.

  • energy metabolism
  • shivering thermogenesis
  • heat loss
  • plasma glucose oxidation
  • stable isotopes

during environmental cold exposure in adult humans, a decrease in core temperature is prevented by increasing heat production ( prod) via shivering thermogenesis. Involuntary muscle contractions during shivering are mainly fueled by carbohydrates (CHO) and lipids, whereas the contribution of protein oxidation remains minor (∼10%) (16, 36, 42). However, the respective importance of CHO and lipid oxidation has not been clearly established. For example, some researchers imply that CHO is the preferred fuel in the cold (∼60% of total prod) (12, 13, 20, 36, 37, 39,41-43), whereas others show a greater reliance on lipids (∼60% of total prod) (21-23, 33, 45). Possible reasons for such discrepancies between studies are differences in shivering intensity (but see discussion), cooling protocol, and/or nutritional state. Over the last decade, most studies of fuel selection during shivering have focused on CHO metabolism as a probable limiting factor for prod. However, two important issues remain unresolved: 1) the relative contributions of hepatic glucose and muscle glycogen to total CHO oxidation have not been quantified, and 2) the potentially important role of triacylglycerol stores (adipose tissue, liver, and muscle) has often been underrated, particularly during prolonged, low-intensity shivering.

Plasma glucose and muscle glycogen have both been shown to play significant roles in prod during cold exposure (16). Vallerand et al. (42, 43) calculated that plasma glucose and muscle glycogen would contribute equally to total CHO oxidation, assuming that 100% of hepatic glucose production (Ra Glu) is oxidized. However, as pointed out by these authors, at such low rates of oxygen consumption (V˙o 2), this assumption is probably not met because nonoxidative glucose disposal could be important. Studies in which measurements of Ra Glu and plasma glucose oxidation were carried out simultaneously indicate that nonoxidative disposal ranges between 25% Ra Glu during submaximal exercise [45% maximal V˙o 2(V˙o 2 max)] and 70% Ra Glu at rest (11).

A series of studies have shown that 90 min of cold-water immersion cause a reduction of glycogen concentration in biopsies from the vastus lateralis (17, 21-24, 47). Unfortunately, estimating total use of muscle glycogen, at the whole organism level, from small biopsy samples is inaccurate at best, because 1) glycogen content varies greatly between and within individual muscles and2) the specific muscles involved in heat generation and their level of recruitment are unknown. Therefore, the exact contributions of plasma glucose and muscle glycogen are presently unclear because their rates of oxidation have never been measured directly during shivering.

The purpose of this study was to quantify the respective contributions of plasma glucose, muscle glycogen, and lipid oxidation to total prod during prolonged, low-intensity shivering by using a combination of stable isotope and indirect calorimetry methods. In human subjects exposed to low-intensity shivering [10°C for 2 h with a liquid conditioned suit (LCS)], we hypothesize that plasma glucose oxidation will play a lesser role than previously suggested (42, 43) and that lipid oxidation will be a major pathway for heat generation because of the small change in metabolic rate observed in the cold.



Six healthy and trained men volunteered for this study, which was approved by the Health Sciences Ethics Committee of the University of Ottawa, and written consent was obtained from the participants. Percent body fat [underwater weighing; Brosek et al. (5)] andV˙o 2 max were measured with a progressive treadmill protocol 5–7 days before the experiments. Physical characteristics of the subjects are presented in Table1.

View this table:
Table 1.

Physical characteristics of subjects

Experimental protocol.

Experiments were conducted between 900 and 1300 after 36 h without heavy physical activity. The last evening meal was standardized (∼988 kcal, ∼52% CHO, ∼18% lipids, ∼30% proteins), and subjects were asked to report to the laboratory the next morning (900) after a 12- to 14-h fast. Ingestion of CHO from plants naturally rich in 13C (C4 photosynthetic cycle) was avoided to maintain low 13C background enrichment in plasma glucose and expired CO2. Care was taken to minimize thermal stimuli between awakening and the start of the experiment (i.e., avoid exposure to hot or cold temperatures, very-low-intensity exercise during transit from home to the laboratory). On their arrival in the laboratory, subjects were instrumented with thermal probes and an indwelling catheter (18-gauge, 32 mm, Medical, Arlington, TX) placed in an antecubital vein (left arm) for blood sampling and were fitted with a LCS (Three Piece Delta Temax, Pembroke, ON). Subjects were then asked to empty their bladder [time (t) = 0 min] and sit quietly for 2 h at 28.1 ± 0.3°C (758 ± 4 mmHg, 20–30% relative humidity). After this habituation period, they were transferred to an environmental chamber (11.1 ± 0.1°C, 760 ± 4 mmHg, 40–57% relative humidity), and a 10°C water perfusion was started through the LCS by using a temperature-controlled circulation bath (Endocal, NESLAB; and model 200–00, Micropump, Vancouver, WA). Thermal response, metabolic rate, and fuel utilization were measured at 28°C and during the subsequent 2-h cold exposure.

Thermal response.

Central body temperature [esophageal temperature (Tes)] was monitored continuously by using a pediatric Tes probe (Mon-a-therm general purpose, Mallinckrodt Medical, St. Louis, MO), which was inserted through the nose to a depth placing the tip of the thermocouple at the level of the left atrium, or one-fourth of the standing height of the subject (25). Heat flux transducers (Concept Engineering, Old Saybrook, CT) were used to estimate skin temperature and nonevaporative heat flux from the forehead, chest, biceps, forearm, abdomen, lower and upper back, front and back calf, quadriceps, hamstrings, and finger. Mean skin temperature (T̄sk) and mean heat flux were calculated with an area-weighted equation (8). Heat flux measurements were used to calculate whole body radiative () and convective heat exchange (). Respiratory evaporative ( resp) and convective heat exchanges ( resp) were determined from ventilation (V˙e) measurements by estimating water loss via the respiratory tract (2,411.3 J heat/g evaporated water) (4). It was assumed that evaporative heat loss ( loss) from the skin was negligible under the LCS. Whole body loss (in watts) was calculated as followsH˙loss=(R˙+C˙)+(E˙resp+C˙resp) Equation 1

Metabolic rate and fuel utilization.

e, V˙o 2, and carbon dioxide production (V˙co 2) were determined by open-circuit spirometry (250 liters, chain-compensated gasometer, Warren Collins, Braintree, MA). All expired gas collections were made at ambient temperature outside the experimental chamber. A mouthpiece, a unidirectional valve (2700 series, Hans Rudolph, Kansas City, MO), and a 44-mm plastic tube were used to direct all expired gases to the collection tank. Expired gases were collected for 5 min every 30 min at 28°C and during cold exposure. V˙e (l/min,btps) was calculated from the displacement of the cylinder and corrected for temperature and pressure. Oxygen and carbon dioxide concentrations in dry expired gases (CaSO4, Drierite, 8 mesh, Fisher Scientific, Ottawa, ON) were determined directly from the spirometer by using calibrated electrochemical gas analyzers (AMETEK model S-3A/1 and CD 3A, Applied Electrochemistry, Pittsburgh, PA).

Total CHO (RGox), lipid (RFox), and protein oxidation rates (RPox) were calculated with the following equations (19)RGox(g/min)=4.59V˙CO2(l/min)3.23V˙O2(l/min) Equation 2 RFox(g/min)=1.70V˙CO2(l/min)+1.70V˙O2(l/min) Equation 3 RPox(g/min)=2.9×Ureaurine(g/min) Equation 4where V˙co 2 andV˙o 2 (Eqs. 2 and 3 ) were corrected for the volumes of O2 and CO2corresponding to protein oxidation (1.010 and 0.843 l/g, respectively), and Ureaurine is urinary urea excretion.

Estimates of RPox (Eq. 4 ) were made by measuring Ureaurine from urine samples collected for a period of 120 min at 28 and 10°C. A correction for urea accumulation in plasma was not required because plasma levels did not change during cold exposure (P = 0.28; pairedt-test) (18). Urinary and plasma urea concentrations were determined on a Synchron Clinical System (CX7, Beckman, Anaheim, CA). Respective contributions of glucose, lipid, and protein oxidation to total prod were calculated by using energy potentials of 16.3, 40.8, and 19.7 kJ/g, respectively (9, 28).

Plasma glucose oxidation.

For the measurement of plasma glucose oxidation, the subjects ingested 10 g of glucose [7 × 1.4 g in 100 ml of water; corn sugar, with a ratio of 13C to C (13C/C) = 0.01098] artificially enriched with 13C ([U-13C]glucose, 13C/C >99%, Isotec, Miamisburg, OH) to obtain a final 13C/C of 0.0476 isotopic composition of exogenous glucose solution (Rexo).

After baseline 13C/12C in plasma and expired CO2 (t = 30 min) were measured, subjects ingested the first dose of [13C]glucose. Subsequent doses were then taken every 30 min until the end of the experiment. Isotopic composition of plasma glucose and expired CO2 were determined in blood and expired gas samples every 30 min before the ingestion of the next dose. On collection, blood samples were put on ice, spun in a refrigerated centrifuge, and separated, and the plasma was kept frozen at −20°C until analysis.

The isotopic composition of plasma glucose (Rglu) was measured as previously described (29). Briefly, plasma samples (1 ml) were deproteinized (BaOH: 1.5 ml, 0.3 N; and ZnSO4: 1.5 ml, 0.3 N) and centrifuged to precipitate the proteins. Double-bed ion exchange chromatography with superimposed columns (resins: AG 50W-X8 H+, 200–400 mesh, and AG 1-X8 chloride, 200–400 mesh) was used to isolate plasma glucose. After evaporation, glucose was combusted (60 min at 400°C) in the presence of CuO, and CO2 was recovered. Measurements of13C/12C in expired CO2 and in CO2 obtained from glucose combustion were determined in a prism mass spectrometer (VG, Manchester, UK). Isotopic composition was expressed as ‰ difference (δ) compared with Pee Dee Belemnitella-1 (PDB-1) Chicago standard with the equation of Craig (6)δ13C={[(13C/12Csample)/(13C/12Cstandard)]1}×103 Equation 5The rate of exogenous glucose oxidation (RGox-exo, g/min) was estimated from isotopic composition of expired CO2 (Rexp) and Rexo as follows (26)RGox­exo(g/min)=V˙CO2×(RexpRref/RexoRref)(1/k1·k2) Equation 6where V˙co 2 is in l/min (stpd), Rref is the isotopic composition of expired CO2 at 28°C before ingestion of the first [13C]glucose dose, k1 (0.7426 l/g) is the volume of CO2 produced from the complete oxidation of glucose (28), and k2 is the fractional recovery at the mouth of CO2 produced in tissues. A fractional recovery of 13CO2 at the mouth (k2) of 0.8 and 1 was used at 28°C and during cold exposure, respectively (46). Because of the large size of the bicarbonate pool, only values in the last 30 min at 28°C and at 10°C were used in the calculation of plasma glucose oxidation rate (RGox-plasma). This delay allows sufficient time for equilibrium of the 13C/12C to be attained in the bicarbonate pool (27).

Plasma glucose oxidation was calculated from13CO2 excretion and the isotopic enrichment of plasma glucose by using the following equation (7, 46)RGox­plasma=V˙CO2(RexpRref/RgluRref)(1/k1·k2) Equation 7Oxidation of glucose released from the liver (RGox-liver) was estimated by subtracting the low rate of RGox-exo from RGox-plasma. Calculation of glucose oxidation derived from glycogen stores (RGox-mus; g/min) in the tissues, either directly or through the lactate shuttle (3), was calculated by subtracting RGox-plasma(Eq. 8 ) from RGox (Eq. 2 )RGox­mus=RGoxRGox­plasma Equation 8

Blood analysis.

Plasma glucose and lactate concentrations were measured spectrophotometrically at 340 nm on a Beckman DU 640 (2), whereas total plasma nonesterified fatty acid (NEFA) concentration was determined by using an analytic assay kit (NEFA C, Wako Chemicals, Osaka, Japan). Insulin concentration was measured by using a radioimmunoassay (#KTSP-11001, Medicorp, Montréal, PQ).

Statistical analyses.

Overall changes in Tes, T̄sk, loss, prod, blood metabolite concentrations, expired CO2 and plasma glucose isotopic enrichments, and gas exchange over time were assessed by using a one-way ANOVA with replication. For each sampling time, a Bonferroni t-test was used to detect potential differences with control values observed at 28°C. Differences in metabolic fuel utilization for CHO (RGox, RGox-exo, RGox-plasma, RGox-mus, RGox-liver), lipids (RFox), and proteins (RPox) over the last 30 min at 28°C and during cold exposure were determined by using two-tailed paired t-tests. Statistical differences were considered significant when P ≤ 0.05. All values given are means ± SE (n = 6).


Thermal response.

Changes in Tes and T̄sk are presented in Fig. 1. Whereas Tes remained constant at 36.4 ± 0.1°C throughout the experiment,T̄sk decreased from 34.0 ± 0.02 to 27.2 ± 0.02°C in the initial 90 min of cold exposure and did not change for the last 30 min. Absolute loss and prod increased by a maximum of 3.3-fold (77.7 ± 0.6 to 258.4 ± 10.6 W) and 2.6-fold (95.3 ± 2.2 to 243.8 ± 4.2 W), respectively (Fig.2). After reaching a maximum 20 min after the onset of cold exposure (t = 140 min), loss decreased by 16% over the next 100 min (238.3 ± 0.6 W). Maximal prodwas reached after 90 min of cold exposure and stayed constant for the remainder of the experiment. Observed shivering activity appeared minimal over the first 60 min of cold exposure but increased progressively in the last hour.

Fig. 1.

Esophageal (Tes) and mean skin temperature (T̄sk) at 28°C and during whole body 10°C cold exposure. Arrows indicate the times at which [13C]glucose solutions were ingested. Values are means ± SE.

Fig. 2.

Absolute heat loss ( loss;Eq. 1 ; A) and heat production ( prod; B) at 28°C and during whole body 10°C cold exposure. Values are means ± SE. * Significantly different from values at 28°C, P ≤ 0.05. † Significantly different from maximal value reached during cold exposure, P ≤ 0.05.

Metabolic response and fuel utilization.

Changes in V˙e, V˙o 2, and respiratory exchange ratio (RER) are shown in Fig.3. V˙e andV˙o 2 increased by 2.4- and 2.6-fold, respectively (Fig. 3, A and B). A small decrease in RER was observed during cold exposure, but it did not reach overall statistical significance (one-way ANOVA with replication;P = 0.075), averaging 0.84 ± 0.01 throughout the experiment (Fig. 3 C; 0.86 ± 0.01 at 28°C and 0.83 ± 0.02 between 150 and 180 min at 10°C).

Fig. 3.

Ventilation (V˙e, in l/min,btps; A), absolute oxygen consumption (V˙o 2, in l O2/min,stpd; B), and respiratory exchange ratio (RER;C) at 28°C and during whole body 10°C cold exposure. Values are means ± SE. * Significantly different from values at 28°C, P ≤ 0.05.

RFox, RGox, and RPox and their respective contributions to total prodthroughout cold exposure are plotted in Fig.4. Total lipid and CHO utilization increased 3.8-fold (Fig. 4 A; 39 ± 2 mg fatty acids/min at 28°C to 177 ± 17 mg fatty acids/min at 10°C) and 2.2-fold (Fig. 4 B; 165 ± 9 mg glucose/min at 28°C to 358 ± 41 mg glucose/min at 10°C), respectively. Protein utilization was not affected significantly by the change in temperature and averaged 62.1 ± 3.1 mg/min at 28°C and 77.7 ± 5.0 mg/min at 10°C. A trend toward an increase in the relative contribution of lipids and a decrease in the relative contribution of CHO during cold exposure was noticed in the cold. However, these observed changes failed to reach statistical significance (Fig. 4 B;P = 0.10). In contrast, the relative contribution of protein oxidation to total prod decreased significantly throughout cold exposure, from 21.2 ± 0.8% at 28°C to 10.7 ± 0.8% at 10°C.

Fig. 4.

Glucose (●, Eq. 2 ) and lipid (○, Eq. 3 ) utilization rates (protein oxidation was constant at 62.1 ± 3.1 mg/min at 28°C and 77.7 ± 5.0 mg/min at 10°C; A), as well as their relative contribution to total prod(% prod; proteins, □;B) at 28°C and during whole body 10°C. FFA, free fatty acid. Values are means ± SE. * Significantly different from values at 28°C, P ≤ 0.05.

Plasma concentrations.

Changes in plasma concentrations of insulin, glucose, lactate, and NEFA during cold exposure are presented in Fig.5. Insulin and glucose concentrations were not affected by the change in temperature (Fig. 5, Aand B). After 90 min of cold exposure, plasma lactate and NEFA concentrations were increased 1.8-fold (0.90 ± 0.11 to 1.61 ± 0.15 mM, Fig. 5 C) and 2.5-fold (0.21 ± 0.03 to 0.52 ± 0.01 mM, P < 0.001; Fig.5 D), respectively, over control values at 28°C.

Fig. 5.

Plasma insulin (A), glucose (B), lactate (C), and nonesterified fatty acid (D) concentrations at 28°C and during whole body 10°C cold exposure. Values are means ± SE. * Significantly different from values at 28°C, P ≤ 0.05.

CHO oxidation: plasma glucose vs. muscle glycogen.

The amount of [13C]glucose administered provided a strong signal in CO2 and plasma glucose to quantify circulatory glucose oxidation (Fig. 6, Aand B). The change in isotopic enrichment of expired CO2 and plasma glucose ([δ-13C]PDB-1), as well as the calculated values of RGox-plasma and RGox-exo obtained throughout the experiment are plotted in Fig. 6. At 28°C, 60 min after glucose ingestion, RGox-plasma averaged 39.4 ± 2.4 mg/min (RGox-liver was 37.7 ± 2.3 mg/min and RGox-exo only 2.1 ± 0.4 mg/min) and increased progressively throughout cold exposure to reach a maximal value of 107.3 ± 6.1 mg/min (RGox-liv was 84.0 ± 6.0 mg/min and RGox-exo only 9.6 ± 5.5 mg/min; Fig.6 C).

Fig. 6.

Expired CO2 (A) and plasma glucose (B) 13C enrichment [δ13C-to-12C ratio, Pee Dee Belemnitella-1 (PDB-1)] and calculated plasma glucose (RGox-plasma) and exogenous glucose oxidation rates (RGox-exo) (C) measured at 28°C and during whole body 10°C cold exposure. Values are means ± SE. * Significantly different from values at 28°C, P ≤ 0.05.

Table 2 summarizes average values measured at 28°C (90–120 min) and 10°C (210–240 min) for all of the parameters of fuel utilization estimated in this study (RGox-plasma, RGox-mus, RFox, and RPox). Expired CO2 and plasma glucose isotopic composition did not change significantly over the last 30 min at 28°C (t = 90 and 120 min) and over the last 30 min at 10°C (t = 210 and 240 min) (ANOVA, P > 0.05). However, Fig. 6, A and B, suggest that priming of the bicarbonate pool with NaH13CO3would have produced a better isotopic steady state. It is, therefore, possible that RGox-plasma was slightly underestimated. As a result of cold exposure, RGox-plasma and RGox-mus increased by 2.1- and 2.4-fold, respectively. Although RFox increased by as much as 3.8-fold at 10°C, RPox remained constant throughout the experiment. The relative contributions of RGox-plasma and RGox-mus to total prod did not change when subjects were exposed to the cold, whereas that of RFox increased 1.5-fold and that of RPoxdecreased twofold.

View this table:
Table 2.

Absolute oxidation and relative contributions of plasma glucose, muscle glycogen, lipid, and protein oxidation to total heat production at 28°C (90–120 min) and quasi-steady state at 10°C (210–240 min)


Even though plasma glucose oxidation is strongly stimulated during low-intensity shivering (+138%), we show that this fuel only plays a minor role in total prod (10% prod). Also, muscle glycogen oxidation doubled during mild cold exposure, providing 75% of total CHO oxidized. Interestingly, lipids are the most important fuel, showing close to a fourfold increase in oxidation rate and accounting for the production of as much heat as all other metabolic substrates combined.

This study quantifies the RGox-plasma during cold exposure. It shows that RGox-plasma is stimulated in direct proportion to metabolic rate (Table 2, Fig. 3 B) and that its relative contribution to total prodremains constant and low. During low-intensity shivering, the contribution of plasma glucose is as minor as that of proteins (Table2). In contrast, muscle glycogen stores play a more prominent role, providing three times more glucose units for oxidation than the circulation (Table 2).

The [13C]glucose ingestion technique selected for this study allowed us to quantify the role of circulatory glucose as an oxidative fuel to support shivering. Results show that, during mild cold exposure, circulatory glucose plays a more minor role than previously suggested when oxidation was estimated from measurements of Ra Glu (42, 43). As anticipated, neglecting to subtract nonoxidative glucose disposal from Ra Glu caused a significant overestimation of glucose oxidation rates. Under conditions of steady state (i.e., when plasma glucose concentration remains constant over time), Ra Glu and glucose disposal were matched. However, the disposal of glucose can take place through two distinct metabolic pathways: oxidation (RGox-plasma) and storage (or nonoxidative disposal). At low metabolic rates (rest, mild exercise, or low-intensity shivering), nonoxidative disposal represents a significant fraction of Ra Glu, and, therefore, a direct measurement of oxidation is necessary to quantify the role of plasma glucose as an oxidative fuel.

The stimulation of circulatory glucose utilization in the cold was not accompanied by changes in plasma glucose or insulin concentrations (Fig. 5, A and B), as previously observed in several other studies (23, 34, 38, 42, 43). Constant glycemia shows that Ra Glu and glucose disposal were both increased in parallel. The increase in glucose uptake taking place during cold exposure is, therefore, not dependent on changes in plasma insulin. However, it may be the consequence of a cold-induced increase in insulin sensitivity and/or GLUT translocation, as previously proposed for humans (35, 40) as well as animals (32,44). An insulin-independent control of glucose uptake may allow an increase in glucose delivery specifically to shivering muscles rather than indiscriminately to all insulin-sensitive tissues.

The major source of CHO was muscle glycogen, providing three-fourths of all of the glucose oxidized in the cold (Table 2). Even though whole body glycogen stores only represent 1% of total energy stores, shivering studies in humans have shown that glycogen availability, modified through diet or exercise, affects fuel selection (22,33, 47) and possibly body cooling rate (22). In these studies, whereas prod was the same among depleted glycogen, loaded glycogen, and normal glycogen controls immersed in 18°C water, RER values were significantly lower for depleted glycogen than for loaded glycogen and glycogen controls, indicating a compensatory shift to a greater relative use of lipids when glycogen reserves are depleted (22, 33, 47). However, these studies do not provide information on the effect of glycogen availability on the relative importance of RGox-plasma and RGox-mus to total CHO oxidation.

Oxidizing lipids to generate heat.

The dual role of lipids as a heat insulation layer and as a large, energy-dense metabolic fuel (>95% of total energy stored) has been recognized for a long time (31). However, the quantitative importance of lipids as a substrate to support prolonged, low-intensity shivering has been somewhat neglected because, over the last decade, most studies have focused on CHO-dependent prod (16). Our results show that RFox provides 50% of all of the heat produced (Table2). The relative importance of lipid oxidation measured here is consistent with several studies (15, 21-23, 33, 45), whereas many others found that CHO oxidation is dominant (12, 13,20, 36, 37, 39, 41-43). In all of these studies, it is very interesting to note that the reported dominance of either CHO or lipids has no clear link with differences in shivering intensity but seem to be correlated with the cooling protocol. Whereas subjects exposed to cool air used CHO preferentially (∼60% of prod; Refs. 12,13, 20, 36, 37,39-43), those cooled by water immersion or by LCS favored lipid utilization (∼60% of prod; Refs. 15,21-23, 33, 45). Physiological reasons for such a difference are unclear, and further research will be needed to explain it.

If one considers the relatively low metabolic rates reached here during mild shivering (15–20% V˙o 2 max), the observed utilization of lipids is not that surprising. Exercise studies reveal that lipid oxidation predominates for prolonged work at all intensities <50% V˙o 2 max (1,4, 30). Therefore, even at the highest possible metabolic rates reached during maximum shivering [∼5 times resting metabolic rate or ∼40% V˙o 2 max (10)], lipids may still play a significant role in heat generation, if fuel selection patterns are identical between exercise and shivering.

The large increase in lipid utilization observed here during low-intensity shivering (Table 2) is a strategy to spare limited CHO reserves. Any increase in the relative use of lipids allows the maintenance of prod for longer and, therefore, improves chances of survival in the cold. We can estimate theoretical values for maximum cold endurance under the conditions of our experiments, assuming that the relative use of the different fuels remains the same as measured after 2 h of mild shivering. An average adult man would be able to shiver at 245 W (Fig. 2) for ∼20 h under the specific conditions of the present study before depleting muscle glycogen reserves (assuming that 80% of muscle glycogen is available for oxidation; mean muscle glycogen concentration = 100 mmol glucosyl units/kg wet mass; actively shivering muscle mass = 70% of 36 kg; RGox-mus = 18.4 μmol · kg body mass−1 · min−1; Table 2).

The stimulation of RFox (Table 2) found here was accompanied by a twofold increase in circulatory NEFA levels (Fig.5 D) as observed in other studies (34, 37, 38, 42, 43,45). Vallerand et al. (43) found that men exposed to 5°C for 3 h showed a parallel increase in NEFA concentration and NEFA disappearance rate. Together, these observations suggest that the oxidation of circulatory NEFA increased throughout cold exposure. However, the relative contributions of NEFA from the circulation (adipose and liver) and from muscle triacylglycerols to the fourfold increase in total fat oxidation remain to be established.

In conclusion, this study shows that total prod during prolonged, low-intensity shivering is unequally shared among lipids (50%), muscle glycogen (30%), circulatory glucose (10%), and proteins (10%). Therefore, the importance of plasma glucose oxidation is only minor, and future research should focus on lipid and muscle glycogen stores that provide most of the energy for prod.


The authors acknowledge the excellent technical assistance provided by Caroline Proulx, Marie-Catherine Lamoureux, Martin Milot, Charlotte Weber, and Kevin Litchfield. We also thank the subjects of this study for their collaboration.


  • This work was funded by grants from the Natural Sciences and Engineering Research Council of Canada (NSERC Canada) (to J.-M. Weber, G. P. Kenny, D. Massicotte, F. Péronnet, and C. Lavoie). F. Haman is the recipient of a NSERC scholarship.

  • Address for reprint requests and other correspondence: J.-M. Weber, Biology Dept., Univ. of Ottawa, 30 Marie Curie, Ottawa, Ontario, Canada K1N 6N5 (E-mail:jmweber{at}

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

  • First published March 1, 2002;10.1152/japplphysiol.00773.2001


View Abstract