Thermosensitivity of muscle: high-intensity thermal stimulation of muscle tissue induces muscle pain in humans

Abstract

Small-calibre afferent units responding to thermal stimuli have previously been reported to exist in muscle. The question as to whether these receptors in humans mediate subjective thermal sensations from muscle remains unresolved. The aims of the present study were to determine in humans whether intramuscular injection of warm and cold isotonic saline elicits temperature sensations, muscle pain or any other sensations. In 15 subjects, no thermal sensations assessed on a temperature visual analogue scale (VAS) could be detected with intramuscular injections of isotonic saline (1.5 ml) into the anterior tibial muscle at temperatures ranging from 8 to 48 °C. The same subjects recorded strongly increasing scores on a temperature VAS when thermal stimuli in the same intensity range were applied to the skin overlying the muscle by a contact thermode. However, i.m. isotonic saline of 48 °C induced muscle pain with peak scores of 3.2 ± 0.8 cm on a VAS scale ranging from 0 to 10 cm. Using the the McGill pain questionnaire a subgroup, of subjects qualitatively described the pain using the ‘thermal hot’ and ‘dullness’ word groups. Temperature measurements within the muscle during the stimulating injections showed that the time course of the pain sensation elicited by saline at 48 °C paralleled that of the intramuscular temperature and far outlasted the injection time. The present data show that high-intensity thermal stimulation of muscle is associated with muscle pain. High-threshold warm-sensitive receptors may mediate the pain following activation by temperatures of 48 °C or more. Taken together, the data indicate that thermosensation from a given volume of muscle is less potent than nociception.

The functional importance of activity in small-calibre afferent nerve fibres from muscle (thin myelinated group III and unmyelinated group IV) is not completely understood. The proportion of group III and IV afferents is approximately 45 % of all afferent nerve fibres in the nerve from the lateral gastrocnemius-soleus (GS) muscle in cat (Stacey, 1969; Mitchell & Schmidt, 1983), but this figure may vary between muscles (Richmond et al. 1976; Abrahams, 1986).

Approximately 33% of group III and 43 % of group IV muscle afferents appear to be nociceptive as they are excited exclusively or predominantly by noxious chemical or mechanical stimuli (Mense & Meyer, 1985) or by contractions under ischaemic conditions (Mense & Stahnke, 1983). Muscle nociceptors might also be excited by noxious heat or cold stimuli because some muscle nociceptors were reported to respond to temperatures above 41 °C or below 25 °C (Iggo, 1961; Hertel et al. 1976; Kumazawa & Mizumura, 1977; Mense & Meyer, 1985). Group III and IV receptors are located around the small arterioles in muscle tissue and connective tissue.

Besides nociceptive group III and IV afferent units, there are also muscle receptors with small-calibre afferent fibres that are not characterised as nociceptive because they can be excited by moderate innocuous stimuli such as physiological stretch, contractions or light touch. These receptors have been suggested to function as: (i) contraction-sensitive receptors which may be involved in cardiovascular adjustments during exercise, so-called ergoreceptors (McCloskey & Mitchell, 1972; Kniffki et al. 1978; Mense & Stahnke, 1983; Kaufman et al. 1984; Mense & Meyer, 1985; Tallarida et al. 1990; Adreani et al. 1997), the contraction-sensitive endings might also mediate subjective force sensations; (ii) low-threshold mechanosensitive receptors which are likely to mediate pressure sensations from deep structures (Mense & Meyer, 1985) and (iii) thermosensitive receptors (Hertel et al. 1976; Kumazawa & Mizumura, 1977; Mense & Meyer, 1985). The latter units might be involved in thermoregulation (Jessen et al. 1983). However, the question still remains as to whether these receptors might also function in humans as peripheral sensors for subjective temperature sensations from muscle.

In humans, electrical, mechanical and chemical stimuli have been used to induce activity in small-calibre muscle nerve afferents. The majority of studies dealt with the induction and assessment of experimental muscle pain. Sensory manifestation of muscle pain was cramp-like, diffuse aching pain, which was felt locally in the stimulated muscle and as pain referred to distant somatic structures (Kellgren, 1938; Travell & Simons, 1982; Mense, 1994; Arendt-Nielsen et al. 1997; Graven-Nielsen et al. 2001). Electrical stimuli have been applied to muscle nerves, muscle nerve fascicles, or intramuscularly to study basic aspects of deep pain in humans (Torebjörk et al. 1984b; Vecchiet et al. 1988; Simone et al. 1994; Marchettini et al. 1996; Arendt-Nielsen et al. 1997; Laursen et al. 1997; Svensson et al. 1997). Electrical stimulation is, however, a non-physiological stimulus bypassing receptor transduction and depolarising the receptive ending or afferent fibre directly. Ischaemia and exercise have also been used to induce muscle pain (Lewis, 1932; Asmussen, 1956; Newham, 1988; Howell et al. 1993; Bajaj et al. 2000). Intramuscular injections of algesic substances such as capsaicin, bradykinin, serotonin, potassium chloride, l-ascorbic acid, acid phosphate buffer, and hypertonic saline have likewise been used to induce muscle pain and thereby excitation of thin muscle nerve afferents (for reviews see Arendt-Nielsen, 1997; Graven-Nielsen et al. 2001). Thus it appears that electrical, mechanical and chemical excitation of group III and IV muscle afferents has been widely explored in humans whereas the effects of thermal stimuli have never been studied.

In studies on skin thermoreceptors supplied by Aδ or C-fibres, warm receptors (Konietzny & Hensel, 1975; Hallin et al. 1981), cold receptors (Torebjörk, 1974; Adriaensen et al. 1983; Campero et al. 1996), and heat nociceptive endings (Gybels et al. 1979; Hallin et al. 1981; Van Hees & Gybels, 1981; Adriaensen et al. 1983; Treede et al. 1995; Campero et al. 1996) have been described in humans, and the neural activity was found to correlate with the thermal sensations. The warm receptors start to fire action potentials at ∼30 °C, they increase their frequency with rising temperature, and saturate at ∼43 °C, where heat nociceptors start to become active. Temperatures below 30 °C excite cold receptors (McCleskey, 1997). To date there is no comprehensive information on the existence of subjective perception of temperatures from muscle although such information might be relevant in conditions of fever, exercise or exposure to cold or warm environments.

This study presents the psychophysical responses to thermal stimulation of muscle tissue in humans by injections of temperature-controlled isotonic saline.

METHODS

Psychophysical response to thermal stimulation of muscle tissue in humans

Fifteen non-medicated subjects participated (9 males, 6 females; mean age, 23.9 years, range, 21–29 years). Manual palpation of the investigated areas was performed to exclude subjects with signs of tender or sore muscle tissues. The study was conducted in accordance with the Declaration of Helsinki, approved by the local ethics committee, and written informed consent was obtained from all participants prior to inclusion.

Thermal stimulation of muscle tissue in humans

Muscle tissue was thermally stimulated by injection of warmed or cooled saline. Intramuscular (i.m.) injections of sterile isotonic and hypertonic saline (1.5 ml) were given over 20 s (270 ml h⁻¹) into the tibialis anterior (TA) muscle. The injections were performed with a 27G needle (15 mm injection depth) attached to a 10 ml syringe. To avoid cutaneous sensations, the injection site was anaesthetised with intradermal injections of 0.2 ml lidocaine (lignocaine, 10 mg ml⁻¹) 1 min before the i.m. injection. Three sites were injected in each TA muscle: 12, 15 and 18 cm distal to the apex patellae. Previously, it was shown that muscle pain induced by hypertonic saline (i.e. excitation of group III and IV afferents) is easily reproducible when injections are separated by 2 cm and given at 6 min intervals (Graven-Nielsen et al. 1997b). Five injections of isotonic saline of different temperatures and one injection of hypertonic (5.8 %) saline were given in a sequence allowing at least 10 min without any sensation between each injection. The injection sequence was randomised.

The isotonic saline was warmed or cooled by immersing bottles of sterile saline in water baths or keeping them at room temperature or in the refrigerator. The target temperatures of the isotonic saline were 8, 18, 28, 38 and 48 °C. The hypertonic saline was injected at 18 °C. Three thermostatically regulated water baths set to 30, 42 and 55 °C were used to obtain the 28, 38 and 48 °C target temperatures. The water baths were set at higher temperatures than the target temperatures to compensate for the temperature loss due to subsequent handling (e.g. drawing up into the syringe, etc.). A target temperature of 8 °C was attained by cooling the saline in a refrigerator set to 5 °C. A target temperature of 18 °C was attained by keeping the saline at room temperature. The temperature of the saline in the syringe (7–8 ml) was measured (DM852, Ellab A/S, Rødover, Denmark) immediately after the injection was completed. Before all injections, the skin temperature was recorded a few centimetres away from the injection site. The subject was not in the room where the saline was prepared and was given no information regarding saline temperature and concentration of the injection.

Psychophysical assessments

The intensity of saline-induced pain was continuously scored on a 10 cm electronic visual analogue scale (VAS) where 0 cm indicated ‘no pain’ and 10 cm ‘most intense pain’. The pain intensities were sampled every second by a computer. The area under the VAS time curve, the maximum pain intensity rating and the duration of pain were evaluated. Moreover, the delay from the start of injection to the onset of pain was determined (VAS onset). The subject completed a Danish or English version of the McGill pain questionnaire (MPQ) (Melzack, 1975; Drewes et al. 1993) to provide a qualitative description of the pain sensations induced. Word groups from the MPQ chosen by at least one-third of the subjects were analysed.

Unpleasantness was scored on a VAS ranging from 0 cm or ‘not unpleasant at all’ to 10 cm or ‘extremely unpleasant’. The temperature sensation after each injection was assessed on a VAS ranging from 0 cm or ‘painful cold’, through 5 cm or ‘neutral’, to 10 cm or ‘painful heat’.

Assessment of the skin sensitivity to thermal stimulation in humans

To ensure that the subjects were able to differentiate between different thermal stimuli, the skin sensitivity to contact heat stimulation was tested with a 4 cm² thermode (Somedic AB, Sweden). The sensitivity tests of the skin covering the left and right TA muscle were performed before the injections. Three stimulation intensities (8, 28 and 48 °C) were applied with 1 °C s⁻¹ increase and decrease rates starting from a baseline temperature of 30 °C. All stimulations were repeated three times (4–8 s interval), and for each stimulus intensity, VAS scores for pain, unpleasantness, and temperature sensation (see previous section) were obtained. VAS scores from left and right leg were averaged.

Assessment of sensitivity to pressure in humans

Pressure pain thresholds (PPTs) were determined with an electronic pressure algometer (Somedic AB, Farsta, Sweden) equipped with a 1 cm² circular probe. Force was applied at 30 kPa s⁻¹ in a simple, continuous, ascending series. The subject was instructed to press a button at the precise moment that the pressure stimulation elicited ‘just noticeable’ pain. PPTs were recorded directly on the site of the i.m. injections. The mean of three trials (minimum 30 s interval) defined the PPT. PPTs were recorded before (after skin anaesthesia) and after the saline injection (after eventual pain had vanished).

Intramuscular temperature profile during thermal stimulation in humans

In an additional experiment including a subgroup of subjects (n = 5), the i.m. temperature was recorded in parallel with the VAS scores of pain after six injections of saline. A flexible temperature probe (0.8 mm, Ellab A/S) connected to a thermometer (DM852, Ellab A/S) was used to record the i.m. temperature. The thermometer was modified so that a computer could sample (200 Hz) the temperature. The temperature signal was accordingly low-pass filtered at 100 Hz. To insert the probe, a plastic catheter (18G, 45 mm long) with a side-arm was introduced into the TA muscle and remained in place for the duration of the experiment. The main catheter was capped with a self-sealing bung to allow needle insertions. The temperature probe was inserted via a hypodermic needle into the main catheter and protruded 5–8 mm into the TA muscle. Six injections of saline were given via the side-arm of the catheter, this ensured delivery to the same tissue site. This technique enabled i.m. temperature to be recorded in exactly the area where the saline was applied.

Statistical analysis

The majority of PPTs and VAS scores did not pass the Kolmogorov-Smirnov normality test (P < 0.05). For PPTs two-way analysis of variance (ANOVA) for repeated measures with time (before and after) and injection types as factors was appropriate, but not possible with non-parametric tests. After logarithmic transformation, the PPTs passed the Kolmogorov-Smirnov normality test (P > 0.05). The transformed PPTs were tested using two-way ANOVA followed by the parametric Student-Newman-Keul's (SNK) test to correct for multiple comparisons. Logarithmic transformation of VAS scores was not suitable because transformation of zero values is not defined. Therefore, the non-parametric Friedman analysis of variance for repeated measures was used for VAS scores. When this test gave significant results, it was followed by the non-parametric SNK test. Spearman's correlation coefficient was used to describe correlation between parameters. The data are presented as means and standard error of the mean (s.e.m.). The level of significance was set at P < 0.05.

RESULTS

Pain and unpleasantness scores in response to thermal muscle and skin stimulation in humans

The temperature of the saline measured immediately after the injections was significantly different among all target temperatures from 8 to 48 °C (Fig. 1A; Friedman: P < 0.001; SNK: P < 0.05). The skin temperature measured before each injection was not significantly different between the different injections and was 31.6 ± 0.1 °C on average.

Figure 1. Pain and unpleasantness scores.

Mean (±s.e.m., n = 15) VAS peak pain scores (A) and VAS unpleasantness scores (B) after injections of isotonic saline (1.5 ml) with target temperatures of 8–48 °C and injections of hypertonic saline (1.5 ml, 18 °C, ‘hyp’). * Significant difference compared with all other injections (SNK: P < 0.05).

Injection of isotonic saline of 48 °C caused significantly higher VAS peak pain than isotonic saline of lower temperatures (Fig. 1A; Friedman: P < 0.001; SNK: P < 0.05). The VAS peak pain (Fig. 1A), duration of pain, VAS area (Table 1) and unpleasantness scores (Fig. 1B) were higher for 18 °C hypertonic saline than for all isotonic saline injections at the temperatures used (Friedman: P < 0.034; SNK: P < 0.05). Thus, the main finding concerning pain is that there was a clear and significant peak in the VAS pain following injection of isotonic saline at 48 °C (Fig. 1A).

Table 1.

VAS pain scores

Target temperature	VAS onset (s)	VAS duration (s)	VAS area (cm s)
8 °C	19.7 ± 2.2	77.7 ± 16.3	69.0 ± 12.7
18 °C	34.6 ± 4.4	59.5 ± 18.4	71.9 ± 26.8
28 °C	39.6 ± 6.0	61.8 ± 21.8	86.4 ± 45.1
38 °C	26.4 ± 3.0	51.5 ± 13.1	75.5 ± 30.0
48 °C	16.8 ± 1.7	76.3 ± 15.4	153.1 ± 38.2
18 °C (hypertonic)	23.5 ± 2.5	384.2 ± 32.0*	1563.5 ± 214.6*

Values are expressed as means ±s.e.m. (n = 15).

^*Significantly different compared with all other target temperatures (Friedman: P < 0.03; SNK: P < 0.05).

Only after injections of 48 °C isotonic and 18 °C hypertonic saline, were word groups selected by more than 1/3 of subjects. The ‘thermal, hot’ word group was selected by 40 % of subjects after 48 °C isotonic saline and by none of the subjects after 18 °C hypertonic saline. The ‘dullness’ word group was chosen by 47 and 53 % of subjects after 48 °C isotonic saline and hypertonic saline, respectively. ‘Punctate pressure’ (47 % of subjects), ‘incisive pressure’ (47 %), ‘evaluative’ (73 %) and ‘spatial distribution’ (67 %) word groups were selected exclusively after hypertonic saline.

Skin stimulation at 48 °C caused significantly more pain (VAS, 5.9 ± 0.6 cm) and unpleasantness (VAS, 4.5 ± 0.6 cm) than stimulation at 28 °C (VAS pain and unpleasantness, 0.0 ± 0.0 cm) and 8 °C (VAS pain, 0.1 ± 0.0 cm; VAS unpleasantness, 0.4 ± 0.2 cm; Friedman: P < 0.001; SNK: P < 0.05).

Thermosensation following thermal stimulation of muscle and skin in humans

With regard to subjective temperature sensations no significant difference was seen between any of the injections; the mean score was 5.2 ± 0.1 cm corresponding to the neutral sensation (Fig. 2A). In contrast, the same thermal intensity applied to the skin induced a progressive increase in thermosensation from 8 to 48 °C (Fig. 2B; Friedman: P < 0.001; SNK: P < 0.05).

Figure 2. Thermosensation.

Mean (±s.e.m., n = 15) VAS temperature scores after thermal i.m. (A) and skin (B) stimulation. Intramuscular stimulation consisted of isotonic saline injections (1.5 ml) with target temperatures of 8–48 °C and injections of hypertonic saline (1.5 ml, 18 °C, ‘hyp’). The skin was stimulated with a contact probe at 8, 28 and 48 °C. * Significant difference compared with all other stimulus intensities (SNK: P < 0.05).

Hyperalgesia in response to pressure after thermal muscle stimulation

ANOVA revealed that the PPTs were decreased after all injections compared with before injection (Fig. 3; ANOVA: degrees of freedom (dF) = 1, F > 7.2, P < 0.019). Although not significant, the PPTs after injection of 48 °C isotonic saline tended to be more decreased than after the other injections.

Figure 3. Sensitivity to pressure.

Mean (±s.e.m.) pressure pain threshold (PPT) before and after injections of isotonic saline (1.5 ml, ‘iso’) with target temperatures of 8–48 °C and injections of hypertonic saline (1.5 ml, 18 °C, ‘hyp’). * Significant difference compared with before the injection (SNK: P < 0.05).

Intramuscular temperature profile during thermal muscle stimulation

The i.m. peak temperature was significantly different in all injections except the isotonic and hypertonic saline injections with target temperatures of 18 °C (Fig. 4A and Table 2; Friedman: P < 0.05; SNK: P < 0.05). The same applied for the temperature of the saline in the syringe measured immediately after the injection. Accordingly, a high correlation between the peak temperature measured in the syringe and intramuscularly was found (Spearman: R = 0.98, n = 30, P < 0.0001). The temperature of the saline contained in the syringe immediately after the injection was on average 1.6 ± 1.5 % higher than the i.m. peak temperature.

Figure 4. i.m. temperature and pain intensity.

Mean profiles (n = 5) of i.m. temperature (A) and pain intensity (B) over time following injection of isotonic saline (1.5 ml) with target temperatures of 8–48 °C and an injection of hypertonic saline (1.5 ml, 18 °C, ‘hyp’). The filled bar below the abscissa indicates the time of injection. The peak i.m. temperature differs clearly between the different target temperatures.

Table 2.

Temperatures and VAS peak pain scores after i.m. injections

Target temperature	i.m. baseline temperature (°C)	i.m. peak temperature (°C)	Syringe temperature (°C)	VAS peak pain (cm)
8 °C	33.3 ± 0.4	4.8 ± 0.4*	4.8 ± 0.3*	2.0 ± 0.7
18 °C	32.5 ± 0.7	19.1 ± 0.9**	18.1 ± 1.0**	0.8 ± 0.4
28 °C	33.8 ± 0.4	28.3 ± 0.4*	28.1 ± 0.2*	0.7 ± 0.5
38 °C	32.6 ± 0.3	37.7 ± 0.4*	38.1 ± 0.4*	0.7 ± 0.2
48 °C	32.7 ± 0.5	49.1 ± 0.8*	49.6 ± 0.4*	2.4 ± 1.4
18 °C (hypertonic)	31.9 ± 0.8	18.3 ± 1.2**	17.7 ± 1.1**	5.2 ± 0.9*

Values are expressed as means ±s.e.m. (n = 5).

^*Significantly different compared with all other target temperatures (Friedman: P < 0.03; SNK: P < 0.05).

^**Significantly different compared with all other target temperatures, except 18 °C (SNK: P < 0.05).

The VAS peak pain was significantly higher for hypertonic saline compared with cold and warm isotonic saline injections (Fig. 4B and Table 2; Friedman: P < 0.029; SNK: P < 0.05). Similar to the results of the main psychophysical experiment, a tendency towards increased pain following 48 °C isotonic saline injections was seen. However, the increase was not significant probably due to the limited number of subjects tested. In contrast to the injection of hypertonic saline at 18 °C, there was a clear relation between peak pain intensity and peak i.m. temperature for isotonic saline at 8 and 48 °C (Fig. 5). Interestingly a more extreme i.m. temperature was needed to initiate the pain sensation than to maintain pain when the i.m. temperature returned to baseline levels. This relation was not as pronounced after injections of hypertonic saline.

Figure 5. i.m. temperature versus pain intensity.

Hysteresis curves of mean (n = 5) i.m. temperature versus pain intensity following injection of isotonic saline (1.5 ml) with target temperatures of 8 and 48 °C and an injection of hypertonic saline (1.5 ml, 18 °C, ‘hyp’). Arrows indicate time progress after injection. The relation between peak pain intensity and maximum/minimum i.m. temperature is clearly seen for isotonic saline at 8 and 48 °C. When the i.m. temperatures returned to baseline levels after isotonic saline injections, pain was also induced at temperatures that did not cause pain when the temperature approached the peak temperature.

DISCUSSION

The present human study shows for the first time that i.m. injection of 48 °C isotonic saline induced muscle pain in contrast to isotonic saline at lower temperatures (8–38 °C). In contrast to thermal skin stimuli, the i.m. injections of warm and cold isotonic saline did not cause any thermosensation assessed on a temperature VAS.

Muscle pain induced by thermal stimulation

In line with previous studies, hypertonic saline induced intense muscle pain which was characteristically deescribed as deep and diffuse (Graven-Nielsen et al. 1997a; Svensson et al. 1998; Stohler et al. 2001). Moreover, muscle pain was elicited by injections of isotonic saline at a temperature of 48 °C. This pain is not thought to be due to the injection procedure as such because in previous studies injections of isotonic saline at room temperature induced very slight pain with intensities below 1 cm VAS (Svensson et al. 1998; Graven-Nielsen et al. 1998a, b; Babenko et al. 1999; Wang et al. 1999). Compared with muscle pain induced by hypertonic saline, the pain duration for isotonic saline-induced muscle pain was significantly shorter. The intramuscular temperature profiles recorded in the present study illustrate the short duration of the stimuli even though the temperatures were significantly above or below the baseline intramuscular temperature. This explains the short duration of the pain. Nevertheless, the pain outlasted the time it took to inject the heated saline indicating that the thermal energy was needed to induce pain and that the pain was not induced by mechanical stimulation during the injection. Previous studies employing isotonic saline injections at room temperature often used an infusion rate of 90 ml h⁻¹ or below (Svensson et al. 1998; Graven-Nielsen et al. 1998b; Babenko et al. 1999; Wang et al. 1999), whereas in the present study the injection was given at a rate of approximately 270 ml h⁻¹. Although the mechanical stimulation of muscle tissue caused by the faster injection was probably stronger in the present study than in previous studies, similar pain intensities were reported (Svensson et al. 1998; Graven-Nielsen et al. 1998a, b; Babenko et al. 1999; Wang et al. 1999). This indicates that the mechanical stimulation associated with the injection procedure is not relevant to the elicited sensation.

To our knowledge, there are no previous studies assessing the sensitivity of muscle tissue to noxious thermal stimulation in humans. Indirect evidence suggests, however, that intense heating of deep tissue might cause pain since microwave-induced hyperthermia (> 42 °C) used in therapy of deep-seated tumours occasionally results in pain from the treated structures (Matsuda et al. 1991).

The muscle pain induced by hypertonic saline and by isotonic saline at 48 °C is presumably mediated by two different mechanisms. Hypertonic saline-induced muscle pain is most probably due to an unspecific excitation of nociceptors whereas the thermal effects are probably mediated by heat-sensitive nociceptors (Hertel et al. 1976; Mense & Meyer, 1985) similar to those present in the skin. The fact that a high proportion of heat-sensitive nociceptors respond to bradykinin compared with the low-threshold thermosensitive receptors likewise suggests a nociceptive function, even though responsiveness to bradykinin is not specific for nociceptors (Mense, 1996). Interestingly, 40 % of the subjects in this study used the terms in the ‘thermal hot’ category to describe their pain sensations following intramuscular injection of 48 °C saline. This may indicate that in fact the muscle receptors responding to noxious heat may represent nociceptors that mediate heat sensations. However, the thermal sensation after 48 °C isotonic saline was not dominating, as it could not be reliably detected on VAS scores of the thermal sensations. Other receptors respond with a long latency and thresholds at high temperatures (Mense, 1996). In such cases, an indirect mechanism of excitation such as release of algesic agents (including bradykinin) from muscle tissue during long-lasting noxious heat stimulation has to be taken into consideration. Similarly the temperature-pain curve (Fig. 5) after 48 °C injections showed that a higher i.m. temperature was needed to initiate the pain sensation compared with the temperature needed to maintain pain when the i.m. temperature returned to baseline levels. Sensitisation by algesic substances may explain such a phenomenon. Actually, the frequent combination of sensitivity to noxious heat and to bradykinin (Mense, 1996) may reflect such a mechanism.

Although the spatial temperature distribution after an injection of heated isotonic saline is not known, it probably stimulates the same volume as the injection of hypertonic saline. Nevertheless, hypertonic saline activates a broader population of nociceptors (unspecific excitation) whereas thermal stimuli are a more specific modality. Thus, the number of activated nociceptors after injection of heated saline is most probably smaller than that following injection of hypertonic saline. The small number of excited nociceptors and short duration of increased intramuscular temperature after injection of heated isotonic saline compared with hypertonic saline might explain the great difference in muscle pain intensity since spatial and temporal summation of neural activity has been shown to be important for muscle pain. Intraneural microstimulation of muscle nociceptive afferents causes muscle pain that is dependent on the stimulation time (temporal summation) (Marchettini et al. 1996) and the number of stimulated afferents (spatial summation) (Simone et al. 1994). Likewise, the effects of repeated injections of hypertonic saline and electrical stimulation were shown to be affected by temporal summation (Arendt-Nielsen et al. 1997; Graven-Nielsen et al. 1997b).

Increased sensitivity to pressure was found after all injections irrespective of the injected saline concentrations or temperature. This change probably reflects tissue sensitisation or a minor trauma caused by the injection procedure per se as reported previously after both injections of isotonic and hypertonic saline (Graven-Nielsen et al. 1998b). The observed tendency towards further sensitisation after thermal stimulation with saline at 48 °C may reflect receptor sensitisation similar to that seen in skin receptors after stimulations above 47 °C (Torebjörk et al. 1984a; Pedersen & Kehlet, 1998).

The clinical importance of muscle pain induced by 48 °C thermal stimuli might be limited at first glance. The present study did not, however, assess the pain threshold temperature. A tendency towards increased pain at 38 °C compared with injections at room temperature was present (Fig. 1A). This finding suggests that the pain threshold is considerably lower than 48 °C and probably not far above the physiological range of intramuscular temperatures. From the relation between i.m. temperature and pain intensity (Fig. 5) initial pain (pain threshold) was recorded far below 48 °C. Moreover, a prolonged duration of elevated intramuscular temperature might increase the pain intensity due to temporal summation in contrast to the short-lasting stimuli used in the present study. These considerations might be of relevance in clinical conditions of muscle pain associated with fever.

Thermosensation

In our study, it was not possible to elicit cold or warm sensations from muscle assessed on a VAS although the necessary neural apparatus of thermosensitive group III and IV afferents exist (Hertel et al. 1976; Kumazawa & Mizumura, 1977; Mense & Meyer, 1985). In contrast thermal skin stimulation at 48 °C evoked both thermal and pain sensations assessed on VAS.

There are two possible explanations for the fact that in human subjects innocuous thermal stimuli did not elicit subjective temperature sensations. (1) The discharges of the low-threshold thermosensitive muscle receptors are used for thermoregulatory purposes and do not reach consciousness. This explanation means that there is no subjective temperature sense from muscle. (2) The number of receptors reached by the injected volume of saline was too small to produce the necessary spatial summation in the central nervous system to elicit subjective sensations. To test this hypothesis, larger volumes of muscle tissue must be thermally stimulated. Actually, there is indirect evidence for innocuous thermosensation from deep structures as a rise in deep-body temperature has been reported to generate a vague sensation of warmth strongly associated with the affective component of thermal comfort or discomfort (Benzinger, 1970). A small increase in skin and muscle temperature caused by a counterirritant (Eucalyptamint) applied to the skin did, however, not cause any subjective sensations of thermal character (Hong & Shellock, 1991). Likewise, a survey showed that despite multiple symptoms of fever (e.g. hyperthermia, discomfort) among 529 adults who thought that they had fever, only 13 % were correct (Einterz & Bates, 1997). Thus, the sensation of deep-body temperature seems not to be a prominent one, which is in contrast to thermosensation from skin where an increase of 0.2 °C from 36 to 36.2 °C was detected in more than 60 % of the trials (Handwerker et al. 1982).

Conclusion

For the first time high-intensity thermal stimulation of human muscle was found to induce muscle pain probably mediated by polymodal muscle nociceptors. The data from the present investigation strongly suggest that thermosensation from muscle is less potent than nociception if a small volume of muscle tissue is stimulated.