INTRODUCTION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

MICRODIALYSIS IS A TECHNIQUE by which interstitial concentrations of various substances may be determined in, e.g., skeletal muscle, adipose tissue, and brain of humans and animals (1, 2, 17, 20). The interstitial fluid is located between cells and capillaries. Accordingly, the interstitial concentration of a given substance may be influenced by cellular uptake or release of the substance as well as by supply and removal via the bloodstream. Therefore, in microdialysis experiments, information about local blood flow is important if one wants to make conclusions about tissue metabolism from measurements of interstitial concentrations (5). However, the methods presently available for estimation of nutritive blood flow are either not suitable for use in humans (the microsphere technique; Ref. 10) or are known to have inherent limitations when used in skeletal muscle (the ¹³³Xe clearance technique; Ref. 22). Hence, a new technique, which can be used to estimate nutritive skeletal muscle blood flow, is highly desirable.

Hickner et al. (11, 13, 15, 16) have suggested that the microdialysis technique can be used to estimate local nutritive skeletal muscle blood flow. A microdialysis probe is placed in skeletal muscle and perfused with a fluid containing a low concentration of ethanol, a portion of which may diffuse out of the probe into the surrounding tissues. The ethanol is subsequently washed away by the circulating blood. The rate at which ethanol diffuses out of the probe is dependent on the probe-to-tissue concentration gradient. Therefore, the higher the blood flow, the more ethanol diffuses out of the probe via the membrane and the less ethanol leaves the probe via the outflow. Thus the outflow-to-inflow ratio (o/i) of ethanol should vary inversely with blood flow (11, 13, 15, 16). In the isolated, perfused cat gastrocnemius muscle, Hickner et al. (13) found ethanol o/i to vary inversely with blood flow in the range from 4 to 99 ml · 100 g¹ · min¹. The higher blood flow rates (>40 ml · 100 g¹ · min¹) were associated with electrically induced muscle contractions. However, in the range that is relevant in humans at rest (1-10 ml · 100 g¹ · min¹), only few determinations were made, and, accordingly, the study does not allow conclusions regarding the ability of the method to detect physiological blood flow changes at rest. Furthermore, the time resolution of the microdialysis technique and its ability to selectively reflect local blood flow are not known, because the method has not been applied during maintenance of constant flow for an extended period of time or during zero blood flow. Also, the response of the method to muscle contractions elicited at constant blood flow remains to be seen. A specific problem is that even a relatively large change in blood flow is accompanied by only a small change in outflow ethanol concentration (13), which means that a very high precision of the ethanol assay is required. Hence, it would be convenient to use a substance that is easier to measure. The present study was carried out to further evaluate the validity of the microdialysis technique in the estimation of blood flow changes in muscle. In addition to ethanol, we used [¹⁴C]ethanol and ³H₂O for o/i determinations.

	MATERIALS AND METHODS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Experimental protocol. The perfused rat hindquarter (18) was used to evaluate the microdialysis ethanol technique at blood flow rates similar to those found in human resting muscle. Four microdialysis probes (CMA/20; membrane length: 10 mm; membrane diameter: 0.5 mm; cat. no. 8309571; CMA Microdialysis, Stockholm, Sweden) were inserted through the skin in four different hindlimb muscles (gastrocnemius and tibialis anterior muscles in right and left hindlimbs) in each rat. The hindquarters of four rats were perfused by using a peristaltic pump (2115, LKB Produkter, Bromma, Sweden) for three periods of 20 min at 21 ± 0 (SE) ml · 100 g hindquarter¹ · min¹ [high-1 (H1)], 6.7 ± 0.3 ml · 100 g hindquarter¹ · min¹ [medium (M)], and 21 ± 0 [high-2 (H2)] ml · 100 g hindquarter¹ · min¹ [high-medium-high (HMH) protocol], respectively. Four other hindquarters were perfused for three periods of 20 min at 7.1 ± 0.1 ml · 100 g¹ · min¹ [medium-1 (M1)], 2.0 ± 0.2 ml · 100 g¹ · min¹ [low (L)], and 7.1 ± 0.1 [medium-2 (M2)] ml · 100 g¹ · min¹ [medium-low-medium (MLM) protocol], respectively. Another four hindquarters were perfused at 1.9 ± 0.1 ml · 100 g¹ · min¹ for 3 h (L) while four hindquarters were studied at a blood flow of 0 ml · 100 g¹ · min¹ (0 flow). The hindquarter perfusate consisted of a Krebs-Henseleit solution with bovine erythrocytes (hematocrit 25%), 4% bovine serum albumin, and a glucose concentration of 5 mM. The mean body weight of the Wistar rats was 386 ± 16 g (n = 16), and the weight of the perfused muscle in the hindquarter was taken to be one-sixth of the body weight (21). At the end of the HMH and MLM experiments, one hindlimb was electrically stimulated for 10 min to elucidate whether contractions per se (blood flow was kept constant at 21 ± 0 and 7.1 ± 0.1 ml · 100 g¹ · min¹ in HMH and MLM experiments, respectively) have any influence on the ethanol o/i ratio. A hook electrode was placed around the sciatic nerve and connected to a stimulator. The stimulation was 200-ms trains at 100 Hz, each impulse in the train being 0.1 ms. The trains were delivered at a rate of 1 train/s at a voltage of 5-20 V. The Danish Animal Experiments Inspectorate approved the experimental protocol.

Microsphere experiments. To verify that the perfusate flows through the capillaries in the perfused rat hindquarter (no shunting), 13 microsphere injections into aorta (¹⁴¹Ce, mean diameter: 15.5 ± 0.1 µm, Dupont de Nemours, Belgium) were performed in separate experiments in five rats (blood flow range: 2.9-16.2 ml · 100 g¹ · min¹) (10). Blood was collected from the inferior vena cava from the time of injection and up to 10 min after the injection, and the percentage of microspheres traveling through arteriovenous shunts and not being trapped in the capillaries was calculated. It was assumed that no arteriovenous shunts <15 µm in diameter existed in the rat hindquarter, since most arteriovenous shunts are >20 µm (3) and since arteriovenous shunts are very rare exceptions in the muscle (6). After the experiments, gastrocnemius and tibialis anterior muscles were cut out, and the number of microspheres in the two muscle groups was calculated to determine whether blood flow differed between muscles.

Microdialysis. Before start of the experiment, the microdialysis probes were flushed with 600 µl of the microdialysis perfusate and then perfused for 10 min before collection of dialysate begun. The microdialysis probes were perfused at 2 µl/min with Ringer acetate containing 4 mM glucose, 1 mM lactate, 10 mM ethanol, 3.9 kBq/ml [¹⁴C]ethanol, and 3.7 kBq/ml ³H₂O by using a high-precision syringe pump (CMA/100, CMA Microdialysis, Stockholm, Sweden). In 0-flow experiments, perfusate did not contain ethanol and [¹⁴C]ethanol. Dialysate was sampled on ice in 300-µl capped glass tubes in 10-min (20 µl; HMH, MLM, and 0-flow experiments) or 20-min periods (40 µl; L experiments), taking the transit time in the outlet tube (2 min) into account. Immediately after sampling, 10 µl of dialysate were pipetted into counting vials, scintillation fluid was added, and vials were corked. Dialysates were counted in a liquid-scintillation counter (2200 CA Packard, Packard Instrument, IL) together with four 10-µl samples of microdialysis perfusate. The rest of the sample was kept at 4°C and analyzed for ethanol (16) within 4 days. In addition to the four probes in each hindquarter, one tube without probe was perfused outside the hindquarter to detect whether evaporation occurred during sampling, storing, and/or analysis.

[¹⁴C]ethanol and ³H₂O in microdialyzed muscles. After all experiments, microdialysis probes were removed, and the microdialyzed muscles were taken out, weighed (1.4 ± 0.1 g), and homogenized with a polytron (PT 3100, Buch & Holm A/S, Denmark) on ice in 4 ml of Ringer acetate for 30 s. The homogenate was centrifuged for 10 min at 4°C (3,800 g), and 2 ml of the supernatant were counted in a liquid-scintillation counter (2200 CA Packard).

	RESULTS

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

Effect of different blood flows on o/i. In the HMH experiment, both ethanol, [¹⁴C]ethanol, and ³H₂O o/i increased significantly from H1 to M and decreased significantly from M to H2 (Fig. 1). In the MLM experiment, all o/i ratios increased significantly from M1 to L, but ratios did not decrease significantly from L to M2 (Fig. 2). In the L experiment, all o/i ratios increased with time for ~60 min and then leveled off (Fig. 3). In the 0-flow experiment, ³H₂O o/i increased with time for a longer period (~140 min) but also in this experiment the increase leveled off (Fig. 4). After leveling off, ³H₂O o/i tended to be higher in the 0-flow than in the L experiment (0-flow: 0.67 ± 0.01, L: 0.61 ± 0.01, time 140-180 min, P = 0.06).

View larger version (16K): [in this window] [in a new window]   Fig. 1.   Ethanol, [14C]ethanol, and 3H2O outflow-to-inflow (Out/in) concentration ratios from microdialysis probes placed in 4 different hindlimb muscles (right and left gastrocnemius and tibialis anterior) in each of 4 rat hindquarters (n = 16 probes) perfused for 3 periods of 20 min at 21 ± 0 (High), 6.7 ± 0.3 (Medium), and 21 ± 0 (High) ml · 100 g1 · min1 (high-medium-high blood flow). Values are means ± SE. * P < 0.05 vs. 15-25 min; # P < 0.05 vs. 15-35 min; $ P < 0.05 vs. 45-55 min;  P < 0.05 vs. 15-25 and 45 min;  P < 0.05 vs. 15-35 and 45 min; + P < 0.05 vs. 15-35 and 55 min.

View larger version (16K): [in this window] [in a new window]   Fig. 2.   Ethanol, [14C]ethanol, and 3H2O Out/in concentration ratios from microdialysis probes placed in 4 different hindlimb muscles (right and left gastrocnemius and tibialis anterior) in each of 4 rat hindquarters (n = 16 probes) perfused for 3 periods of 20 min at 7.1 ± 0.1 (Medium), 2.0 ± 0.2 (Low), and 7.1 ± 0 (Medium) ml · 100 g1 · min1 (medium-low-medium blood flow). Values are means ± SE. * P < 0.05 vs. 15 min; # P < 0.05 vs. 15-25 min; $ P < 0.05 vs. 15-35 min.

View larger version (16K): [in this window] [in a new window]   Fig. 3.   Ethanol, [14C]ethanol, and 3H2O Out/in concentration ratios from microdialysis probes placed in 4 different hindlimb muscles (right and left gastrocnemius and tibialis anterior) in each of 4 rat hindquarters (n = 16 probes) perfused for 3 h at 1.9 ± 0.1 ml · 100 g1 · min1 (constant low blood flow). Values are means ± SE. * P < 0.05 vs. 20 min; # P < 0.05 vs. 20-40 min.

View larger version (10K): [in this window] [in a new window]   Fig. 4.   3H2O Out/in concentration ratios from microdialysis probes placed in 4 different hindlimb muscles (right and left gastrocnemius and tibialis anterior) in each of 4 rat hindquarters (n = 16 probes) studied at a blood flow of 0 ml · 100 g1 · min1 (zero blood flow). Values are means ± SE. * P < 0.05 vs. 15 min; # P < 0.05 vs. 15-25 min; $ P < 0.05 vs. 15-35 min;  P < 0.05 vs. 15-45 min;  P < 0.05 vs. 15-55 min; + P < 0.05 vs. 15-65 min; ~ P < 0.05 vs. 15-75 min; § P < 0.05 vs. 15-95 min.

When compared among experiments at identical times (20 and 60 min after start of experiments), all o/i ratios varied inversely with blood flow (Fig. 5).

View larger version (27K): [in this window] [in a new window]   Fig. 5.   Comparison among experiments. Ethanol, [14C]ethanol, and 3H2O Out/in concentration ratios from microdialysis probes placed in gastrocnemius and tibialis anterior muscles of perfused rat hindquarters. Data are taken from 4 different experiments (high-medium-high, medium-low-medium, low, and zero flows) at identical times (20 and 60 min) after start of experiments. Out/in ratios at t = 20 min taken from high-medium-high, medium-low-medium, and zero-flow experiments are average values of out/in ratios at t = 15 min and 25 min. Similarly, out/in ratios at t = 60 min are average values of out/in ratios at t = 55 min and 65 min. High, medium, and low blood flows corresponded to 21 ± 0, 7.1 ± 0.1, and 1.9 ± 0.1 ml · 100 g1 · min1, respectively. For each bar, n = 16 probes. Values are means ± SE. * P < 0.05 vs. out/in at other blood flows at the same time point.

Effect of contractions per se on o/i. In probes placed in the contracting hindlimb, o/i ratios were lower during contractions than before contractions, although differences were significant only for [¹⁴C]ethanol and ³H₂O (Table 1). In probes placed in the noncontracting hindlimb, o/i ratios were not changed significantly by the contractions in the contralateral hindlimb (Table 1).

Ethanol, [¹⁴C]ethanol, and ³H₂O o/i ratios. In none of the experiments (HMH, MLM, and L) did overall ethanol o/i and [¹⁴C]ethanol o/i differ significantly (Figs. 1-3). ³H₂O o/i paralleled ethanol and [¹⁴C]ethanol o/i ratios, but ³H₂O o/i was significantly lower than both ethanol and [¹⁴C]ethanol o/i values (Figs. 1-3).

[¹⁴C]ethanol and ³H₂O in microdialyzed muscles. After all experiments, dialyzed muscles contained [¹⁴C]ethanol and ³H₂O. Generally, muscle contained more isotope after longer lasting experiments with low flow (L and 0-flow) than after shorter lasting experiments with high flow (HMH and MLM). After the HMH experiment, the content of indicator substance in muscle water relative to the content in microdialysis perfusate was 1.0 ± 0.1 and 0.8 ± 0.1% for [¹⁴C]ethanol and ³H₂O, respectively (P < 0.0001 between isotopes, n = 16). Corresponding values for the other experiments were as follows: MLM (n = 16): 2.3 ± 0.5 and 2.1 ± 0.4% (P = 0.05 between isotopes); L (n = 15): 5.6 ± 1.1% (P < 0.05 vs. HMH and MLM) and 5.5 ± 1.1% (P < 0.05 vs. HMH); 0-flow (³H₂O, n = 16): 5.4 ± 0.7% (P < 0.05 vs. HMH and MLM).

1

1

1

1

Evaporation of ethanol, [¹⁴C]ethanol, and ³H₂O. During sampling, storing, and analysis, evaporation of ethanol, [¹⁴C]ethanol, and ³H₂O occurred to some extent as o/i ratios from the tube outside the hindquarter were 0.94 ± 0.01 (n = 82 samples from 12 tubes), 0.95 ± 0.01 (n = 81), and 0.96 ± 0.01 (n = 81) (P > 0.05), respectively. Four hindquarters were excluded from the study because of the evaporation of ethanol during storing and/or analysis. In samples from tubes outside these hindquarters, o/i ratios of ethanol, [¹⁴C]ethanol, and ³H₂O were 0.83 ± 0.03 (n = 23 samples from 4 tubes), 0.96 ± 0.01 (n = 22, P < 0.05 vs. ethanol o/i), and 0.96 ± 0.01 (n = 22, P < 0.05 vs. ethanol o/i), respectively.

	DISCUSSION

TOP ABSTRACT INTRODUCTION MATERIALS AND METHODS RESULTS DISCUSSION REFERENCES

The present study shows that the microdialysis ethanol technique can qualitatively detect differences in skeletal muscle blood flow within and between groups at blood flow rates similar to those found in human resting muscle (Figs. 1, 2, and 5). It is important, however, to notice the rather long time constant of the method, which impairs detection of rapid oscillations in blood flow, especially at low blood flow rates. Furthermore, interpretive caution must be taken during exercise, because contractions per se may decrease o/i of the indicator substance. [¹⁴C]ethanol and ³H₂O may replace ethanol in the estimation of blood flow by microdialysis because the labeled substances are easier to measure and yield similar results compared with ethanol.

In the present study, a decrease in blood flow was always associated with an increase in o/i both within and between experiments (Figs. 1, 2, and 5). Correspondingly, in the HMH experiment, an increase in blood flow was associated with a decrease in o/i (Fig. 1), but this was not the case in the MLM experiment (Fig. 2). Looking carefully at Fig. 2, however, a small nonsignificant decrease in [¹⁴C]ethanol and ³H₂O o/i ratios was seen when blood flow was increased. The lack of a significant decrease in o/i in the last period of the MLM experiment probably reflects that the time constant for changes in tissue concentration of constantly supplied indicator substances upon changes in blood flow is long, relative to the duration of the experiment, and is even longer with lower blood flow. In accordance with this view, o/i increased for ~60 and 140 min in experiments in which blood flow was maintained at 1.9 ml · 100 g¹ · min¹ (approximate flow in L periods) and 0 ml · 100 g¹ · min¹, respectively (Figs. 3 and 4). In MLM experiments, the ML period lasted 40 min, and the o/i ratio would be expected to rise also in the final M period (M2) if the increase in flow from L to M had had no effect (Fig. 2). The fact that o/i did not increase indicates that the change in flow did have an influence. In line with the present findings at constant flow, others have also found an increase in ethanol o/i in the beginning of an experiment, both in muscle (12, 15, 16) and in adipose tissue (8), and they found leveling off to occur at 20-60 min after probe implantation. Also, according to a mathematical model derived by Wallgren et al. (23), the time required to reach steady state becomes longer as blood flow is lowered. According to their Eq. 5 (23), the time to reach steady state at a blood flow of 1.9 ml · 100 g¹ · min¹ should be 53 min, which fits very well with the ~60 min found in the present study.

Also in adipose tissue, the microdialysis ethanol technique has been shown to be able to detect small changes in blood flow (7). In the basal state as well as on stimulation by external heating, adipose tissue blood flow was estimated by both the microdialysis ethanol and the ¹³³Xe clearance technique. Blood flow estimated by ¹³³Xe clearance varied inversely with o/i of ethanol, both within and between subjects. Hence, in muscle and adipose tissue, the microdialysis ethanol technique can detect physiological differences in blood flow both within and between groups. A prerequisite for the technique to detect differences in blood flow between conditions is, however, that only blood flow, not the diffusibility of the indicator substance, differs.

One condition for which a change of diffusibility may be a problem is exercise. Thus, in the present study, we found that o/i decreased during muscle contractions, even though overall blood flow was kept constant (Table 1). This could be due to the movement of the probe in the tissue during contractions, due to "stirring" of the tissue, or due to an increase in the distribution volume of the indicator substance and, in turn, creation of a steeper concentration gradient of the indicator substance around the probe. However, it cannot be excluded that the decrease in o/i during contractions was due to a redistribution of blood flow to contracting muscles, even though overall blood flow to the hindquarter was constant. That no decrease in flow in the muscles of the noncontracting leg occurred speaks against redistribution, as judged from the fact that o/i was unchanged in these muscles (Table 1). Still, nutritive blood flow to the contracting muscles may have increased, even though total blood flow to the muscles was constant. In favor of o/i being, in fact, decreased by muscle contractions per se is a very recent study performed in humans by Rådegran et al. (19), from which it was concluded that ethanol o/i during exercise reflects leg movements and muscle contractions per se rather than local skeletal muscle blood flow. In any case, our findings indicate that detection of blood flow changes during exercise by the microdialysis ethanol technique, as previously done (12, 14, 20), may be problematic.

In two previously published studies (12, 13), the microdialysis ethanol technique was evaluated during muscle contractions (13). The extent to which an effect of contractions per se on o/i influenced the results is uncertain. In one study, Hickner et al. (13) evaluated the microdialysis ethanol technique in the isolated, perfused cat gastrocnemius muscle and found ethanol o/i to vary inversely with blood flow in the range from 4 to 99 ml · 100 g¹ · min¹. To obtain blood flows >40 ml · 100 g¹ · min¹ without reaching excessively high perfusion pressures, the muscle was stimulated electrically. In another study in which young men performed intermittent isometric contractions (12), ethanol o/i varied inversely with skeletal muscle blood flow measured by ¹³³Xe clearance.

A mathematical model for measuring absolute blood flow in skeletal muscle with the microdialysis ethanol technique has been developed by Wallgren et al. (23). The model has been used to convert ethanol o/i to absolute blood flow values both at rest and during exercise (12). The model assumes a steady-state situation. However, the present study has demonstrated that the time to reach steady state for ethanol o/i after a change in blood flow is long at physiological blood flow levels. Furthermore, the model does not take into consideration the change in diffusibility of ethanol that presumably occurs with exercise. Also, if one wants to estimate absolute blood flow from ethanol o/i of a small number of microdialysis probes, the variation in ethanol diffusibility between microdialysis probes due to differences in the structure of the tissue surrounding the microdialysis membrane is a problem. It appears that even with the use of the proposed mathematical model the microdialysis technique can seldom be used to quantify muscle blood flow.

Theoretically, the microdialysis probe might affect the blood flow in the region around the probe. However, in adipose tissue, it has been shown that the probe does not affect local microcirculation as determined by the ¹³³Xe clearance technique (7). Because the ¹³³Xe clearance technique is known to have inherent limitations when used in skeletal muscle (22), we could not perform the same control experiment.

If a substance is supplied by microdialysis to a tissue in which its diffusibility is low, the removal of the substance by the blood will be restricted by diffusion rather than by the blood flow. It follows that high diffusibility and, accordingly, low molecular weight should be advantageous for substances used in the estimation of blood flow by microdialysis. As suggested first by Hickner et al. (11, 13, 15, 16), ethanol is one candidate. We evaluated whether ³H₂O, which has an even smaller molecular weight and would be expected to also have a higher diffusion coefficient in the tissue, might be a better candidate. In line with higher diffusibility, ³H₂O o/i was always lower than ethanol's o/i (Figs. 1-3). However, as ³H₂O and ethanol o/i curves were always parallel, no apparent advantage was revealed for the use of ³H₂O compared with ethanol as indicator substance.

For both ³H₂O and [¹⁴C]ethanol, tissue concentrations among experiments varied inversely with blood flow. However, tissue concentrations were quite low (<6%) relative to perfusate concentrations, suggesting that diffusion significantly limited removal of both indicators by the bloodstream. Still, the indicator concentrations in the tissue will wane with distance from the microdialysis probe, and, accordingly, the measured low average tissue concentration does not accurately reflect the extent to which removal by blood vessels close to the probe is restricted by diffusion. The mean radius around the microdialysis probe of the muscle within which the indicator substance concentration was determined was 7 mm (estimated from mean muscle tissue weight and microdialysis probe length). Although data are not directly comparable, the relative tissue concentration of indicator substance in the present study was apparently higher than expected from the concentration estimated in a study by Hickner et al. (13), who, with a control probe 3-5 mm from the ethanol-perfused microdialysis probe, found an ethanol concentration <0.1% of the perfusate ethanol concentration.

The view that o/i is not solely determined by tissue blood flow is underlined by the findings at zero flow. As would be expected, in steady-state o/i was higher than in experiments in which blood flow was present. However, the steady-state o/i was well below 1 in the zero-flow experiment (Fig. 4), illustrating that indicator supplied by microdialysis can be removed by other routes than by microdialysis outflow and blood flow. Apparently, mere diffusion within the tissue represents a significant means of indicator elimination. This is probably partly responsible for the fact that o/i for ³H₂O was always lower than these of [¹⁴C]ethanol and ethanol (Figs. 1-3).

With respect to the ability of the microdialysis ethanol technique to detect changes in blood flow, it must be pointed out that during the present conditions a tripling in blood flow was accompanied by a change in indicator substance o/i of only ~0.05 (Figs. 1, 2, and 5). Hence, the assay for the indicator substance must be very precise. In the various experiments, ethanol and [¹⁴C]ethanol o/i ratios never differed (Figs. 1-3), showing, accordingly, that [¹⁴C]ethanol can be used just as well as ethanol as indicator substance. ³H₂O and [¹⁴C]ethanol are radioactive, which is an advantage in the sense that they can be measured with high precision. Dialysates can be pipetted and scintillation liquid added right after sampling and, therefore, no storage is needed. Ethanol evaporates easily, and much care must be taken during sampling, storage, and analysis of samples. In the present study, 4 out of 20 hindquarters had to be excluded because of the evaporation of ethanol during storage and/or analysis. Data on ³H₂O and [¹⁴C]ethanol from the same hindquarters did not warrant exclusion. Considering the high precision required for the ethanol concentration, due to the potentially small changes in ethanol o/i and the risk of evaporation during storage and analysis, ethanol is less feasible as indicator substance compared with ³H₂O and [¹⁴C]ethanol. The fact that ³H₂O and [¹⁴C]ethanol are radioactive is a disadvantage because of the radiation risk, but this risk is very low as <4 kBq/ml of perfusate are added. Another indicator substance, which has been used in microdialysis experiments but which, so far, has not been evaluated, is [¹³C]urea (9). Theoretically, [¹³C]urea is not as good an indicator substance as ethanol and ³H₂O, since [¹³C]urea is a larger molecule. This is in agreement with a study in which the permeability rate constant over the microdialysis membrane in rat muscle was shown to be twice as high for ³H₂O as for [¹⁴C]urea (4).

The purpose of the present study was to evaluate the validity of the microdialysis technique in the estimation of blood flow changes in muscle tissue. It was not our intention to compare o/i ratios in different muscle groups. In the microsphere experiments, it was found that blood flow was 35% higher in gastrocnemius than in tibialis anterior muscle (P = 0.02, data not shown). Correspondingly, in microdialysis experiments, overall o/i ratios were lower in gastrocnemius than in tibialis anterior muscle (ethanol: 0.56 ± 0.02 vs. 0.61 ± 0.02, n = 24 pairs of muscle, P < 0.01; [¹⁴C]ethanol: 0.55 ± 0.01 vs. 0.60 ± 0.01, n = 24 pairs of muscle, P < 0.0001; ³H₂O: 0.51 ± 0.02 vs. 0.57 ± 0.0, n = 31 pairs of muscle, P < 0.0001). However, differences in flow and o/i ratios between muscle types were small. Furthermore, the percent change in o/i with changes in blood flow did not differ between muscles (data not shown). These findings and the fact that conclusions would not have been different if muscles had been analyzed separately justify that o/i ratios from gastrocnemius and tibialis anterior muscles were pooled.

In conclusion, the microdialysis technique can be used qualitatively to detect changes in skeletal muscle blood flow, even at levels similar to those in human resting muscle blood flow. However, at low blood flow rates, it takes at least 60 min for the indicator substance supplied by the microdialysis perfusate to reach a steady state within the tissue. The long time constant for changes in tissue concentration of constantly supplied indicator substances on changes in blood flow impairs detection of rapid oscillations in blood flow, in particular at low blood flow rates. Caution must also be exerted when the microdialysis ethanol technique is used to detect blood flow changes during exercise, as contractions per se may decrease the indicator substance o/i. [¹⁴C]ethanol and ³H₂O may replace ethanol in the estimation of blood flow by microdialysis, because the labeled substances are easier to measure and yield similar results compared with ethanol.