Lactate production without hypoxia in skeletal muscle during electrical cycling: Crossover study of femoral venous-arterial differences in healthy volunteers

Background Aim of the study was to compare metabolic response of leg skeletal muscle during functional electrical stimulation-driven unloaded cycling (FES) to that seen during volitional supine cycling. Methods Fourteen healthy volunteers were exposed in random order to supine cycling, either volitional (10-25-50 W, 10 min) or FES assisted (unloaded, 10 min) in a crossover design. Whole body and leg muscle metabolism were assessed by indirect calorimetry with concomitant repeated measurements of femoral venous-arterial differences of blood gases, glucose, lactate and amino acids. Results Unloaded FES cycling, but not volitional exercise, led to a significant increase in across-leg lactate production (from -1.1±2.1 to 5.5±7.4 mmol/min, p<0.001) and mild elevation of arterial lactate (from 1.8±0.7 to 2.5±0.8 mM). This occurred without widening of across-leg veno-arterial (VA) O2 and CO2 gaps. Femoral SvO2 difference was directly proportional to VA difference of lactate (R2 = 0.60, p = 0.002). Across-leg glucose uptake did not change with either type of exercise. Systemic oxygen consumption increased with FES cycling to similarly to 25W volitional exercise (138±29% resp. 124±23% of baseline). There was a net uptake of branched-chain amino acids and net release of Alanine from skeletal muscle, which were unaltered by either type of exercise. Conclusions Unloaded FES cycling, but not volitional exercise causes significant lactate production without hypoxia in skeletal muscle. This phenomenon can be significant in vulnerable patients’ groups.

Introduction Functional electrical stimulation-assisted cycling (FES cycling) is a method originally developed over 30 years ago for patients with spinal cord injury [1]. It uses computer-driven electrical pulses delivered by transcutaneous electrodes and directly activating muscle contractions, independently on functionality of the physiological pathway between upper motoneuron and the neuromuscular junctions. The method is now commercially available in the form of both stationary and mobile devices [2], used by patients with a wide range of conditions incl. spinal cord injury [3], stroke [4,5], and multiple sclerosis [6]. FES cycling was demonstrated to improve cardiovascular fitness, insulin sensitivity [7] bone density and muscle strength [2,8].
In recent years, FES-cycling has become particularly attractive for sedated critically ill patients. Early mobilization is the only intervention, which can partially prevent the development of intensive care unit-acquired weakness [9][10][11][12][13][14]-the major long-term consequence in the survivors of protracted critical illness [15,16]. Muscle atrophy [17,18] and dysfunction [18] occur very early in the critically ill and FES cycling can help to deliver exercise before the patient can co-operate with a physiotherapist [19].
Although FES cycling seems to be feasible in intensive care unit patients [19], before its effect on meaningful clinical outcomes can be tested in the critically ill and other vulnerable patients groups, important physiological questions need to be addressed. Metabolic efficacy (i.e. power output divided by metabolic cost) of the FES cycling is typically very low, around 5-10%, as compared to 25-40% in volitional cycling [20][21][22]). This is likely due to non-physiological pattern of muscle activation, where large muscle groups are activated simultaneously rather than small well-coordinated units [2,23]. Despite FES cycling increases cardiac output [24] and leg blood flow to the same extent [25] or even more [26] than volitional cycling and consequently oxygen delivery to the muscle should be normal, there are features suggesting early switch to anaerobic metabolism: early fatigue [23,27], rapid intramyocellular glycogen depletion [28], increase of respiratory quotient (RQ) >1 [20] and even a mild increase in arterial lactate levels [29]. Increased lactate production could be caused by microcirculation impairment during electrically stimulated asynchronous contraction [30] or by a mismatch between glycogenolysis activated by electrical stimulation [31] and pyruvate oxidation.
Nonetheless, a direct evidence of the presence of anaerobic metabolism in skeletal muscle during FES cycling is lacking. In addition, whilst the influence of volitional resistance exercise on amino acid metabolism has been extensively studied [32][33][34][35][36] there is no such data for FES cycling, although one study demonstrated activation of anabolic signalling in electrically stimulated gastrocnemius muscle in a rat [31]. These questions may be particularly relevant before FES-assisted exercise is introduced to critically ill patients, who are in profound protein catabolism and may be less able to clear lactate from systemic circulation.
In light of this we conducted a crossover study of volitional and FES supine cycling in healthy postprandial volunteers, where we combined indirect calorimetry with across-leg venous-arterial (VA) difference studies. We hypothesized that FES-cycling as compared to light volitional exercise would lead to increased production of lactate in correlation with widening of VA-CO 2 gap (as the measure of anaerobic metabolism), and with increased amino-acid efflux from skeletal muscle during exercise.

Study subjects
Our experimental group consisted of 14 young (31±8 years), non-obese (23.7±3.7kg/m 2 ) healthy volunteers (gender M/F = 11/3). University Hospital Kralovske Vinohrady's Research Ethics Board reviewed the protocol and approved the study. Prior to the enrolment, all subjects gave their written informed consent in accordance with the Declaration of Helsinki.

Overview of study design
The study was performed during two visits performed 1 week apart. Subjects were asked to attend the visit at 08:00 AM after an overnight fast. In between these visits, the subjects were advised to take their usual diet and avoid strenuous exercise. During the first visit, the volunteers underwent a physical examination and body composition measurement. After 30 min bed rest, their energy expenditure was measured using indirect calorimetry with a ventilated canopy system. Afterwards, in each subject's VO 2MAX was determined on a cycle ergometer with stepwise load by 25 W increments until exhaustion. During the second visit, subjects were given a standardized breakfast containing 70 g of carbohydrates, 10 g protein and 15 g of fat. Afterwards, femoral vein and radial artery were cannulated. After 30 min rest, the subjects were exposed in random order to one of two supine exercise protocols, separated by 3 hours rest. Both protocols begun with baseline measurements (AV difference studies and calorimetry) followed by 5 min of passive cycling. Then, the subjects either performed three 10 min cycles of volitional cycling (at 10, 25 and 50 W, respectively) separated by 5 min of passive cycling (Group A), or FES cycling (Group B). The exercise protocols are outlined in Fig 1.

Methods
Indirect calorimetry and body composition assessment. Resting energy expenditure and RQ were measured after overnight (12 h) fast and 30 min bedrest using canopy as a mixing chamber with 10 sec sampling (Quark RMR device, Cosmed, Italy). To determine peak oxygen uptake (VO 2max ) exhaustive exercise test was performed in each subject on an electromagnetically braked bicycle ergometer Ergoline Ebike (Ergoline Gmbh, Germany). After 5 min warmup period, a workload of 50W was initiated and increased by 25 W every minute continuously until fatigue despite the verbal encouragement. Oxygen uptake was measured using mask, breath-by-breath, 10 sec sampling period (Quark RMR device, Cosmed, Italy. ECG was monitored continuously. Gas analysers (container 5% CO 2 , 16% O 2 and room air) and flow analyser were calibrated prior to each measurement. Body fat was assessed using bioimpedance analysis (NutriGuard 2000, Bodystat, Germany).
Cannulations. Femoral vein was cannulated 2-3 cm below inguinal ligament under ultrasound guidance. In order to avoid the admixture of blood from saphenous and pelvic veins [37], a single-lumen central venous catheter (B-Braun, Germany) was inserted retrogradely to the depth of 10-15 cm so that the tip was deep in the femoral muscular compartment. For arterial sampling, we used a 22 F catheter (BBraun, Germany) inserted into the radial artery.
Cycling protocols. For both volitional and FES cycling we used RT-300 bikes (Restorative Therapies Ltd., USA) and the exercise was performed in supine position. Volitional cycling consisted of three 10 min intervals of active cycling: 10W (13 revolutions/min, resistance 7 N/ m), 25W (31 revolutions/minute, 7.6 N/m), 50W (35 revolutions/min, and resistance 13.4 N/ m). These period were preceded (warm up) and separated by 5 min of passive cycling at 25 revolutions/min. FES cycling: Three pairs of transcutaneous electrodes (3 x 4", Restorative Therapies, Ltd., USA) electrodes were applied on each leg over quadriceps, hamstrings and gluteus maximus muscles, as per manufacturer's instructions. Prior to electrode placement, we measured the thickness of fat layer between the skin and muscle by ultrasound. After 5 min passive warm up (25 revolutions/min), the target speed was changed to 30 revolutions/min and stimulation gradually (1%/s) started to achieve 25 mA. Then, in each subject, the stimulation current was gradually increased to reach subjectively tolerated maximum. Oxygen uptake was measured continuously in both volitional and FES assisted cycling using mask breath-by-breath system (Quark RMR device, Cosmed, Italy). Gas analysers (container 5% CO 2 , 16% O 2 and room air) and flow analyser were calibrated prior to each measurement.
Laboratory methods. Arterial and venous blood samples were analysed for blood gases, lactate and haemoglobin using POCT analyser Cobas b221 (Roche Diagnostics Limited, USA). For other analysis blood samples were centrifuged and frozen at -80˚C until analysed. Serum creatine kinase and myoglobin was measured in a certified institutional laboratory (Cobas system, Roche Diagnostics Ltd., USA). Serum amino acid concentration in arterial/venous blood was analysed using capillary electrophoresis as described [38].

Calculations and statistics
Metabolic efficacy. Metabolic efficacy of volitional cycling was calculated as power output divided by the increase of energy expenditure [2]. Veno-arterial gap in the total content of carbon dioxide (ctCO 2 gap) was calculated according to equations used in ABL 900 Analyser (by Radiometer, Copenhagen, Denmark).
where ctCO 2 (B) = CO 2 content in blood in mmol/L; ctCO 2 (P) = CO 2 content in plasma in mmol/L and equals to 0.23 x pCO2 + cHCO 3 -(P); pCO 2 is partial pressure in kPa, ctHb = haemoglobin content in mmol/L. ctCO 2 (P). pH ERY = estimated intracellular pH in red blood cells, which equals to 7.19+0.77 x (pH-7.4)+0.035 x (1-SO 2 ), where SO 2 is haemoglobin saturation with oxygen; and finally pK ERY is a negative decadic logarithm of bicarbonate dissociation constant: pK ERY ¼ 6:125 À logf1 þ 10 ½pHEryÀ 7:84À ð0:06�SO 2 Þ� g Blood flow. In both FES and volitional cycling, leg oxygen uptake represents a relatively fixed proportion (76±8% and 78±9%, respectively) of whole-body oxygen uptake [39]. Therefore, an index of blood flow through the leg was calculated as whole-body oxygen consumption divided by the difference of oxygen content in arterial and femoral-venous blood. Blood oxygen content was calculated in mmol/L as 0.00983 � pO 2 + SO 2  Statistics. We used linear mixed effect model for 2x2 crossover design processed with software Stata 15 (Stata Corp., LLC, U.S.A.) [40,41]. The model consists of fixed and random part. In the fixed part, the model contained following parameters: (1) Sequence, i.e. order in which subject performed volitional and FES cycling protocols. Had this parameter been significant, a carry-over effect would have been present; (2) Period, basal vs. active, a parameter exploring the effect of the exercise, regardless whether volitional or FES; (3) Treatment, exploits the difference between volitional and FES cycling; and (4) Interaction Period#Treatment exploits whether FES cycling differs from volitional cycling during exercise period. Random part of the model contains subject number in order to take into account repeated measurements. Binary data are showed as frequency + %, continuous data as means ± SD. P value <0.05 was considered as significant. Whenever another test was used we specified this in the text. Sample size determination was performed prior commencement of the protocols with VA lactate difference as a primary outcome.

Characteristics, tolerability and signs of muscle damage
All 14 subjects finished the protocol without adverse events; baseline (visit 1) calorimetry data are available for 13 subjects only due to a technical problem. Baseline characteristics are outlined in Table 1. Sequence parameter of linear mixed effect model was not significant in any of analysed parameters (p = 0.14-0.94), so we assume no carry over effect from previous cycling protocol.

Metabolic efficacy of volitional vs. FES cycling
Metabolic efficacy of volitional cycling was 39.2±5.6%. Unloaded FES cycling led to an increase of metabolic rate to 138±29% from baseline, which was comparable to the increase with 25 W volitional exercise (124±23%). See Fig 2. Energy gain from anaerobic glycolysis was negligible or negative for volitional cycling and 5.0±6.2 W for FES cycling.

Exploring muscle metabolism during FES cycling
VA differences of both O 2 and CO 2 contents (ctO 2 and ctCO 2 ) tended to widen with volitional exercise (Fig 3A and 3B), whilst the opposite trend was seen for FES cycling. In line, there was no change in oxygen saturation of haemoglobin in femoral venous blood neither with volitional exercise (from 63.9±12.7% to 64.3±8.7%), whilst there was an increase after FES cycling (from 62.6±11.3 to 70.3±8.7%; p = 0.02). Across-leg respiratory exchange ratio (i.e. the ratio between VA differences of CO 2 and O 2 contents) although different at baseline (Fig 3C) tended to increase with volitional cycling, but this change was not significant. There was no change from baseline in across-leg glucose uptake of glucose (FES -5.5±3.9 to -5.9±3.6mmol/ min; volitional -7.0±3.6 to -6.9±6.1mmol/min). Whole body RQ increased with FES cycling (0.88±0.02 to 0.95±0.02, p = 0.001, but did not change with volitional exercise (0.87±0.02 to 0.85±0.02, p = 0.55; See Fig 3D) and only FES cycling led to an increase in across-leg lactate VA differences and production (from -1.1±2.1 to 5.5±7.4 mmol/min, p<0.001 vs. from -0.9 ±1.1 to -0.4±1.2 mmol/min, p = 0.70 Fig 3E) with very high inter-individual variability (See

Analysing lactate production
With FES cycling, there was a significant positive correlation between VA lactate difference and femoral venous haemoglobin saturation with oxygen (Spearman's R 2 = 0.6, p = 0.002, Fig  3G). Lactate producers had smaller veno-arterial difference in CO 2 content of the blood (R 2 = 0.3, p = 0.046, Fig 3H), effectively ruling out oxygen delivery problem. Subjects with femoral VA lactate difference >0.5 mmol/L ("lactate producers", n = 5, see Fig 3F) were compared with the rest of the group (n = 9) but no difference was found besides lactate having higher RQ at baseline (0.94±0.06 vs., 0.86±0.07, p = 0.034). Of note, stimulation current used during FES cycling was not different in lactate producers (42±10 vs. 44±16 mA, p = 0.87).

Amino acid metabolism
As expected in postprandial volunteers, at baseline resting skeletal muscle was taking up branched-chain amino acids (BCAAs) whilst producing Alanine (Ala). Skeletal muscle only produced Glutamine (Gln) at baseline in the volitional cycling group, otherwise the change was not significantly different from zero (Fig 4). Neither type of exercise led to a significant change of amino acid metabolism, but it is apparent from Fig 4 that with volitional cycling there was a trend to an increase in Ala production and a decrease of glutamine production, whilst after FES cycling no such a trend was apparent (across-leg amino acid exchange remained unaffected). Uptake of BCAAs continued and did not change with either type of exercise (p = 0.83 and p = 0.86).

Discussion
The major finding of our study is that unloaded supine FES cycling leads to lactate production without signs of muscle hypoperfusion, as low blood flow through exercising limbs would have caused femoral venous haemoglobin desaturation (Esaki et al., 2005;Sun et al., 2016) and widening of VA-CO 2 gap [42], which were not observed in our subjects. Moreover, there was a significant positive correlation between across-leg lactate production and femoral venous oxygenation, suggesting that subjects producing lactate did so whilst extracting less oxygen from (and producing less CO 2 into) the local circulation. There was a marked interindividual variability in metabolic response to FES cycling: some subjects responded to FES similarly to volitional cycling, whilst others produced so much lactate that it elevated systemic (arterial) lactate concentrations well above the normal range. We have not found any convincing characteristics of the subjects producing lactate during FES, although they seemed to be oxidizing more carbohydrates at baseline. Notably there was no correlation between the amplitude of stimulation current used and the production of lactate. Tissue dysoxia and femoral venous desaturations are known to accompany lactate production during high intensity volitional exercise (i.e. > approx. 60% VO 2 MAX ) [43,44,45], at which oxidative phosphorylation becomes oxygen dependent. At lower exercise intensities, there is a concomitant lactate production in fast twitch glycolytic muscle fibres and consumption in slow twitch fibres [46] and-as seen in our subjects-during a steady low intensity volitional exercise, skeletal muscle may become a net lactate consumer [47].
The most obvious explanation of FES-driven lactate production would be tissue dysoxia, occurring despite adequate flow of oxygenated blood through major vessels. Non-physiological asynchronous contractions of large muscle units activated by FES [2,23] could have caused an inhomogeneous perfusion at the level of microcirculation, with hypoxic regions and units with luxurious perfusion acting as functional AV shunts. The increase in whole-body RQ with FES cycling, would support the presence of some degree of anaerobic metabolism, but it could also be explained by impaired fatty acid oxidation with the preference of carbohydrate substrates [39] or by primary increased ventilation. The major argument against microcirculatory impairment and anaerobic lactate generation is the absence of widening of venous-arterial CO 2 gap. Carbon dioxide is produced also anaerobically and released from bicarbonate as the consequence of buffering acid load in hypoxic tissue, and because CO 2 diffuses rapidly even from poorly perfused tissue, VA-CO 2 gap is regarded as a very sensitive marker of tissue hypoxia caused by impaired microvascular flow [48]. Not only VA CO 2 gap was not widened after FES cycling, but in was inversely proportional to lactate production. Moreover, the 138±29% increase in the whole body oxygen consumption after FES-cycling observed by us and others [49] would also argue against major oxygen delivery problem.
Lactate production without tissue dysoxia may occur as a result of the dysbalance between pyruvate production from glycolysis and its conversion to acetyl-CoA and oxidation in tricarboxylic acid cycle [46,47]. Muscle contraction instantly triggers, via the increase in Ca 2+ [IC] , glycogenolysis and glycolysis, producing pyruvate. Sudden increase in cytosolic pyruvate concentration shifts the near-equilibrium reaction: Pyruvate + Glutamate $ Alanine + 2-oxoglutarate, rightwards. Alanine is increasingly released during exercise and 2-oxoglutarate is believed to increase the functional capacity of tricarboxylic acid cycle [50] allowing for increase in oxidative ATP production. BCAAs uptake in skeletal muscle continues or even increases during exercise, providing carbons for oxidative pathways and nitrogen for Alanine and Glutamine formation (Fig 4D). Although non-significant, we have observed some trends to these responses after volitional cycling, but no rearrangement at all of amino acid metabolism was seen with FES exercise. Glycolytic compartment is known to respond much faster compared to oxidative phosphorylation and a rapid increase in cytosolic pyruvate concentration could lead to lactate release from cells even in the absence of tissue hypoxia [46]. Moreover, FES cycling compared to volitional exercise is known to activate glycogenolysis and glycolysis disproportionally faster than oxidative pathways [20,39]. In light of this, our data are consistent with aerobic lactate generation due to a dysbalance between pyruvate generation from glycogenolysis and glycolysis and its oxidation in citric acid cycle. Indeed, skeletal muscle is not a  metabolically homogenous tissue [47] and FES may preferentially trigger muscle contraction in glycolytic fast twitch fibres, whilst lactate oxidizing slow fibres may have been less sensitive to electrical stimulation. The sensitivity of different muscle fibres to external stimulation is unknown and remains to be studied, but a higher sensitivity of fast twitch fibres would be in keeping with the finding, that a long-term external electrical stimulation of a denervated muscle restores its mass and contractile power, but not fatigability [51].
From clinical point of view we found important the absence of venous haemoglobin desaturation during FES-cycling as decreased central venous saturation impairs systemic oxygenation in patients with a degree of intrapulmonary shunt. Mild lactic acidosis could be of concern in patients with impaired lactate clearance (e.g. liver failure). Unloaded FES cycling led to VO 2 response comparable to 25W volitional exercise, which would represent a very significant exercise load for critically ill patients, who tend to have even higher metabolic cost for a given power output [52] and only tolerated cycling at 3-6 W in one study [52]. Lastly, although the absence of laboratory signs of muscle damage and amino acid release is reassuring, the positive association of post-exercise serum myoglobin with stimulation current amplitude suggest a risk of muscle damage from the use of stimulation currents above 70mA, which are often needed to elicit visible contractions in sedated critically ill patient, perhaps due to their impaired muscle excitability [16].
The major weakness of our study is that we have not used direct measurements of leg blood flow and tissue oxygenation. We only use indirect indices, which prevents us from drawing any conclusions about the influence of FESCE on blood flow, which might have been altered, eg. by altered function of muscle pump. However, effects of FES exercise on leg blood flow are known [17,25] and the main finding of the study, i.e. lactate production without evidence of tissue hypoxia, can be supported by across-leg VA differences alone. Muscle tissue oxygen concentrations are known to be closely reflected by femoral venous oxygen content [43,53].
In conclusion, we have demonstrated that 10 min of supine FES cycling in healthy volunteers leads to production of lactate without features suggestive oxygen consumption/delivery mismatch, which are known to accompany lactate production during high intensity voluntary exercise [42,43]. Despite a significant increase in systemic oxygen consumption (proportional to 25W of volitional exercise) and unaltered across-leg glucose uptake with FES cycling, we have not observed the rearrangement of amino acid metabolism towards anaplerosis.
Supporting information S1