Hypoalgesia after exercises with painful vs. non-painful muscles in healthy subjects – a randomized cross-over study

Participants

Pain-free individuals aged 18–50 years, adept in Danish and naive to hypertonic saline injections, were invited to participate in the study. Individuals were recruited by notifications at the University College Lillebaelt of Southern Denmark and through social media platforms. Exclusion criteria were any pain for more than two weeks within the last three months, any pain on the testing days, known mental illnesses, neurological diseases, inflammatory rheumatoid diseases or circulatory diseases in the form of heart or lung disease, any surgery within the last three month, pregnancy, addictive behavior to any kind of euphoric substances or opioids and consumption of alcohol on the day of participation. In the beginning of the first session, before the experiments, participants were familiarized with the PPT assessments and were shown how to perform the wall squat exercise without actually performing the exercise. In addition, demographic information (age, sex, weight and height) was obtained.

Randomization

A researcher (HVB, not involved in recruitment, the practical experiments and the data analyses) randomized and counterbalanced the session-order (Painful and Control), which were sealed in opaque envelopes. The researchers conducting the experiments (ES, LAM, AKM) were blinded to the sequence allocation until after participants were randomized, and thus not blinded to injection type during the PPT assessments. Participants and the researcher (SH) conducting the data analyses were blinded to the order of sessions.

Interventions

Participants attended two sessions, lasting approximately 30 min each, separated by one week. This between-sessions timeframe was chosen to avoid possible carry-over effects such as delayed-onset muscle soreness normally affecting pain variables for less than a week following acute exercise [14]. The experimental procedures are illustrated in Figure 1. In both sessions, manual PPTs at both thighs and the left neck/shoulder were assessed at baseline. Following these assessments in the first session, participants were randomized into one of two order of sessions: (A) injection with hypertonic saline (Painful) in session 1 and injection with isotonic saline (Control) in session 2, or (B) injection with isotonic saline (Control) in session 1 and injection with hypertonic saline (Painful) in session 2. Then, an intra-muscular bolus injection of 1 mL sterile hypertonic saline (5.8%) or isotonic saline (0.9%) was given into the midline of the right quadriceps femoris (QF) muscle 20 cm proximal to the base of the patella. Injections were made manually using a 1 mL plastic syringe (Becton Dickinson, Madrid, Spain) with a disposable needle (27G, 0.40 × 38 mm, Misawa Medical Industry, Ibaraki-ken, Japan). The injections were administered by the same researcher (ES, LAM, AKM) across participants.

Figure 1:

Experimental procedures during both sessions. Manual pressure pain thresholds (PPTs) were assessed at both quadriceps femoris muscles and the left upper trapezius muscle. Pain intensity was assessed by numerical rating scales (NRS: 0–10) immediately before and after the randomized and counterbalanced saline injections, after 1, 2, and 3 min of wall squat, and 1 min after the wall squat. Sessions were separated by one week.

Next, participants performed an isometric wall squat exercise for a maximum of 3 min or until fatigue: Participants were instructed to stand upright with their back against the wall, feet parallel and shoulder-width apart and hands by their sides, lowering their back down the wall until a knee flexion angle of approximately 100° was reached and to maintain this position for 3 min. This exercise condition has previously been shown to induce a robust hypoalgesic EIH response in groups of pain-free individuals [4].

Outcomes

The primary outcome was the between-session difference in EIH response at the injected thigh muscle assessed as the absolute change in PPT (i.e., PPT immediately after exercise minus PPT at baseline). Secondary outcomes were the between-session differences in EIH responses at the contralateral non-injected thigh muscle and the contralateral neck/shoulder muscle as well as pain intensity in the thighs.

PPTs were assessed before the injection, after the injection and after the wall squat exercise using a handheld pressure algometer (Somedic Type II, Sweden, Horby) with a stimulation probe of 1 cm² placed perpendicularly to the skin and with the participant sitting on a plinth with foot support and arms resting on the thighs. Three sites were located and marked for the PPT assessments: Site 1; the right QF muscle (injected thigh), 15 cm proximal to the base of patella, so the assessment was in closeness, proximity to the injection, but not directly over the injection. Site 2; the left QF muscle (control thigh), 15 cm proximal to the base of patella. Site 3; the left upper trapezius muscle, 10 cm from the acromion in direct line with the 7th cervical vertebra. The pressure was increased at ∼30 kPa/s until the participant defined the pressure as the first sensation of pain and pressed a button. Two PPT assessments were performed at each site, and the average for each site was used for statistical analyses. Twenty-second intervals between assessments were kept.

Pain intensity in both thighs were rated by the participants using two separate 0–10 numerical rating scales (NRS), with 0 defined as “no pain” and 10 “as worst imaginable pain”. Pain intensities were rated at baseline, after the injections as well as at 1, 2, and 3 min into the wall squat exercise and 1 min after the wall squat exercise.

As reported in the pre-registration protocol, cuff-induced pain sensitivity and self-reported physical activity were collected at baseline in both sessions (prior to assessments of PPTs, injections and exercises). These measures will be reported in a later paper on the associations between self-reported physical activity and pain sensitivity.

Statistical analysis

The estimated number of participants were based on an a priori two-tailed sample size estimation. An average EIH effect size of 0.7 following short-term isometric exercise in healthy pain-free individuals has been reported in a previous meta-analysis [2]. As the effect of muscle pain on the EIH response has not been investigated previously, we decided on a medium between-intervention effect size of 0.5 (difference in EIH response at the thigh after exercise with experimental muscle pain compared to exercise without experimental muscle pain), a power of 0.80, and an alpha of <0.05. Using G*power version 3.1.9.2 (Düsseldorf, Germany), 34 participants were required in this within-subject repeated-measures study.

Results, in text and figures are presented as mean ± standard deviation (SD) with 95% confidence intervals (95% CI) unless stated otherwise. To test for normality, all data variables were inspected using QQ-plots and histograms, and analyzed using Shapiro-Wilks test. If QQ-plots, histograms (extreme values) and Shapiro-Wilks tests (p≤0.05) indicated non-normal distributed data non-parametric statistical analyses were applied.

Participant characteristics

Thirty-five participants were recruited and tested in this study as one participant became unwell (dizziness and had to lie down for approx. 5–6 min) after one of the injections and was unable to initiate the wall squat test immediately after the injection. Thirty-four participants (age: 25.5 ± 4.4 years [range: 20–46]; BMI: 24.4 ± 4.3 [range: 18.9–40.9]; 12 women) completed both sessions and were included in the analyses.

Baseline PPTs in the two sessions are presented in Table 1. No significant main effect of sessions (F (1,66)=2.62, p=0.12) or interaction between sessions and sites (F(2,66)=0.53, p=0.59) were found indicating no between-sessions order effect on baseline PPTs.

EIH outcomes

Increases in PPTs (i.e., EIH) at all assessment sites were observed after the wall squat exercise in both sessions except in the injected thigh after the Control injection as indicated by the confidence interval overlapping zero (Table 1).

Exploratory outcomes

Summary of results

This randomized controlled cross-over study is the first study to investigate the influence of experimental muscle pain in the exercising thigh muscle during exercise on the subsequent EIH responses following acute exercise. No between-session difference in EIH responses after exercise with a painful muscle compared to exercise with non-painful muscles were found. Moreover, ratings of muscle pain intensity after the painful injection were not associated with the subsequent EIH responses following exercise. These findings suggest that pain per se in the exercising muscle does not reduce the hypoalgesic effects of exercise in pain-free individuals. The strengths of this study include pre-registration of the protocol, a priori sample size estimation, randomization and allocation concealment, and statistical analysis performed by a blinded investigator.

General discussion of findings

PPTs increased after the wall squat exercise in the non-painful session, although EIH was not robust in the injected thigh, which is in agreement with previous studies in pain-free individuals [4]. Studies of EIH in individuals with different pain conditions suggest that the hypoalgesic EIH response after acute exercise is attenuated when exercising painful muscles [9, 10] and therefore the current findings were unexpected. There are several possible explanations for these conflicting results: The painful hypertonic saline injection in the thigh muscle induced moderate pain and local hypersensitivity compared with baseline, which is in agreement with a previous study using the same needle size, concentration of hypertonic saline, and volume injected in the pelvic area [13]. However, the range in pain intensity in the current study spans from almost no pain (NRS: 1) to very intense pain (NRS: 9) which indicate large inter-individual pain responses to the same noxious stimuli. This may have affected the results and warrant further investigation. The pain intensity induced by the hypertonic saline injection was similar to pain intensity after the control injection 2 min into the exercise (Table 2). Increased pain intensity throughout the whole exercise might have had influenced the EIH response. Also, in the current study, the hypersensitivity was not significantly larger after the painful injection than after the control injection. The equivocal results may be related to the site of injection (muscle vs. ligament), which is supported by a previous study by Drew and colleagues who observed no differences in hypersensitivity after hypotonic saline injections in thigh muscles compared with isotonic control injections [15].

Reduced EIH in individuals with chronic pain has been associated with hypersensitivity, a common finding in painful areas in several clinically painful conditions [16]. The individuals included in this study showed no signs of a reduced pain-inhibits-pain response; in contrast they showed increased PPT in the neck/shoulder muscle and the non-injected thigh after the painful injection. A reduced pain-inhibits-pain response has previously been related to reduced EIH [17, 18]. Fingleton et al. observed reduced EIH after isometric and aerobic exercise in knee osteoarthritis patients who also showed reduced CPM while patients with an inhibitory CPM response also showed a positive EIH response [17]. Further, several studies have suggested shared systemic mechanisms involved in EIH and CPM responses [6, 7]. Collectively, this indicates that the effectiveness of central pain inhibitory mechanisms may be more important for the hypoalgesic response to exercise than the presence of pain itself.

In the current study, non-painful muscles were also exercised (the contralateral thigh) during the wall squat exercise after painful injection possibly activating more segmental or systemic effects counteracting a potential reduced local EIH response from exercising painful muscles, as previous studies have indicated that the EIH response is the net result of an array of locally and systemically acting pain inhibitory mechanisms involving both opioidergic and non-opioidergic mechanisms [2, 19].

The injection itself may have reduced the EIH response as the magnitude of the EIH response in the injected thigh was generally smaller compared to the non-injected contralateral thigh independently of injection type. This may suggest that the bolus injection itself or local intrinsic factors such as different weight bearing strategies or different muscle activation patterns after the injections might have influenced the results. In support of the latter, hypertonic saline injection has previously been shown to redistribute muscle activation patterns during isometric exercise [20].

The results may also be influenced by the duration of pain; injection with hypertonic saline only induced short-lasting pain that was almost normalized 4 min after the injection in contrast to chronic pain states that are consistent. Other pain models e.g., delayed-onset muscle soreness [21] or nerve-growth-factor injections [22] that induce longer lasting pain should be investigated in the future in relation to EIH.

The results may also be influenced by the distribution of pain. As highlighted in a recent EIH review [12], individuals with more widespread pain often show attenuated EIH both locally and systemically, while individuals with more localized pain show attenuated local EIH only when exercising painful muscles. In the current study, pain distribution after the painful injection was not assessed. However, a previous study injecting hypertonic saline into the thigh muscle observed a rather localized pain distribution [15], whereas studies demonstrating reduced EIH often have included individuals with more widespread pain [18, 23, 24].

The current results are in line with Kadetoff & Kosek [25] who found similar EIH responses after a one-legged isometric knee extension exercise to exhaustion in individuals with fibromyalgia compared with pain-free individuals, suggesting that the hypoalgesic response to acute exercise may be unrelated to the presence of clinical pain itself during exercise. Additionally, longitudinal studies reporting no change in EIH in spite of reduced clinical pain following long-term exercise programs [26, 27] or surgery [28] in knee pain populations, suggest that clinical pain and EIH are not directly related. Only a few studies have examined EIH calculated as change in clinical pain after acute exercise as highlighted in a recent review [12]. Future research on e.g., the change in clinical pain (instead of change in pain sensitivity measures) after acute exercise and its relation to improved clinical pain after long-term exercise programs are warranted.

The current findings indicate that the EIH responses are comparable after exercise performed with painful and non-painful muscle in pain-free individuals. Several studies that have investigated EIH, report that the utilized exercise is perceived as moderately painful [4, 29], and larger hypoalgesic responses are found after painful compared to non-painful exercises in pain-free individuals [5], suggesting that pain during exercise may be important for EIH. Further, previous studies have reported a positive correlation between CPM responses and EIH responses in healthy individuals [7, 30] suggesting shared mechanisms of CPM and EIH and that the individual’s ability to pain inhibition is a trait rather than only related to exercise. This may be different in some chronic pain populations; Coombes and colleagues [31] showed that painful isometric exercises above the individual’s pain threshold caused increased clinical pain responses during and after exercise in people with lateral epicondylalgia, while no increase in pain was observed with non-painful exercises performed below the pain threshold. This indicates an interaction between pain and EIH; painful exercises of higher exercise intensity may result in more input to central pain facilitatory mechanisms resulting in a net balance of pain facilitation and subsequently post-exercise increased pain sensitivity and clinical pain. Although acute hyperalgesia to exercise is common in musculoskeletal pain populations [12] studies on change in clinical pain after acute exercise are limited [12], and a recent systematic review with meta-analysis concluded that painful long-term exercise programs may be more beneficial than non-painful exercise for short-term, but not medium- and long-term, pain management in chronic musculoskeletal pain populations [32] highlighting the complexity between the acute exercise pain response and long-term outcome.

Limitations

This study is strengthened by its randomized cross-over design. However, in addition to the limitations discussed in the previous sections, this study did not include a no-exercise control condition as recommended in a recent review [33]. Also, the study was powered to detect a medium between-intervention effect and thus not adequately powered to detect smaller differences.