Muscle Pain as a Regulator of Cycling Intensity

Effect of Caffeine Ingestion

Alexander R. Gonglach; Carl J. Ade; Michael G. Bemben; Rebecca D. Larson; Christopher D. Black


Med Sci Sports Exerc. 2016;48(2):287-296. 

In This Article


Experimental Approach

Nine testing sessions were performed over the course of approximately 4 wk with 48 to 72 h separating each testing session. All testing sessions were performed at approximately the same time of day, with participants being instructed to eat a similar light meal (~50% carbohydrates, 30% fat, and 20% protein) 2 h before each testing session. Participants were also encouraged to water liberally the day before and day of each testing session and to refrain from any form of exercise on the day before and day of each testing session. Compliance with these instructions was confirmed by a 24-h diet and exercise recall questionnaire, which was completed at the beginning of each testing session. All experimental procedures were approved by the institutional review board at the University of Oklahoma Health Science Center, and participants provided written informed consent before participation.

During testing session 1, participants were screened and consent provided. They then performed a maximal exercise test on a cycle ergometer to determine their V̇O2peak. Sessions 2 to 4 were identical and were used to familiarize participants to regulating exercise intensity to produce moderate muscle pain and to establish the reliability of distance covered and work performed when exercising in this manner. Sessions 5 and 6 were identical to the familiarization session, but participants ingested either caffeine or placebo 60 min before exercise. Further testing was then performed on participants who improved their performance after caffeine ingestion in session 5 or 6. This subset of participants completed three additional testing sessions during which they regulated their exercise intensity to produce strong muscle pain. Session 7 served as a familiarization, and sessions 8 and 9 occurred after ingestion of either caffeine or a placebo. A schematic outline of the nine testing sessions and the movement of participants through each session are shown in Figure 1.

Figure 1.

Diagram of testing sessions and flow of participants through each session.


Sixteen healthy men age 18 to 35 volunteered for the study. Three participants dropped out of the study because of the time commitment (one participant) or an injury unrelated to the research procedures (two participants) before completion of all testing sessions. Thirteen participants (age 22 ± 5 yr, height 180 ± 4 cm, and weight 76.7 ± 6.8 kg) completed familiarization and the initial protocol at a pain level of "3." Participants were nonsmokers, caffeine naive (<40 mg caffeine ingested per day), had no contraindications to vigorous exercise, and were physically active by ACSM guidelines (accumulating at least 150 min of moderate intensity exercise per week; see p. 8 of ref.[1] Participants were recruited through flyers and word-of-mouth in health and exercise science classes at the University of Oklahoma. A sample size of 13 would allow the detection of a small (SD, 0.31) effect (difference) in mean scores on distance covered and work performed with a power of 0.80 and an alpha level of P < 0.05, assuming a correlation of 0.95 or greater between repeated trials (as determined during familiarization testing; see below) using a two-tailed, dependent measures t test.[29] Participants who experienced an increase in total distance covered during the caffeine condition (i.e., caffeine "responders") of testing sessions 5 and 6 were asked to continue in the study and complete sessions 7 to 9. All 10 "responders" volunteered; however, 1 participant had changed his caffeine consumption (exceeding the ≤40 mg·kg−1·d−1 requirement) and was excluded, yielding a total sample of 9. A sample size of 9 would allow the detection of a moderate (SD, 0.49) effect (difference) in mean scores on distance covered and work performed with a power of 0.80 and an alpha level of P < 0.05.[29]

Assessment of V̇O2peak

Participants performed a peak aerobic capacity test on a cycle ergometer (Excalibur Sport; Lode BV, Groningen, The Netherlands). The cycle ergometer seat height was adjusted to allow for approximately 160° of knee extension at the bottom of the pedal stroke. Body weight, height, and environmental conditions were recorded before exercise. Open circuit spirometry (TrueOne 2400 Metabolic Measurement System, Parvo Medics; Sandy, UT) was used to assess oxygen uptake (V̇O2) and related gas exchange variables. The spirometry system was gas and flow calibrated before each measurement according to manufacturer specifications using a standard calibration gas (16% O2, 4% CO2) and a 3-L flow syringe. V̇O2 and V̇CO2 were standardized to STPD. Participants warmed up for 5 min at 50 W (watt) while being instructed to keep the revolutions per minute (rpm) above 60. Next, the workload increased at a rate of 1 W every 2 s (30 W·min−1). The test was terminated when the participant could not maintain a pedal cadence of 50 rpm. After termination of the test, the workload was reduced to 30 W, and the participant was allowed to cool down for at least 3 min. After the cool down, the participants' color and demeanor was monitored for an additional 5 min. V̇O2, respiratory exchange ratio (RER; V̇CO2/V̇O2), and heart rate (HR) were continuously measured during the test and averaged in 30-s intervals. Ratings of perceived exertion (RPE) and quadriceps muscle pain were provided by the participant using a 6 to 20 Borg scale[8] and a 0 to 10 pain scale[10] at the end of every minute. Pain ratings, a primary outcome measurement for subsequent testing sessions, were made using a previously validated scale[10] using both numerical ratings and verbal anchors as follows: 0—no pain at all, 0.5—very faint pain (just noticeable), 1—weak pain, 2—mild pain, 3—moderate pain, 4—somewhat strong pain, 5—strong pain, 7—very strong pain, and 10—extremely intense pain (almost unbearable). No verbal anchors are associated with numbers 6, 8, and 9, whereas a (·) symbol is provided to describe any pain intensities greater than 10.[10] Gas exchange threshold (GET) was determined from the peak test by visual inspection of plots of the following measures: 1) the initial disproportionate increase in V̇CO2 when V̇CO2 was graphed against V̇O2 (i.e., the V-slope method), 2) an increase in V̇E/V̇O2 when graphed against V̇O2 without a corresponding increase in V̇E/V̇CO2, 3) an increase in FEO2 when graphed against V̇O2 without a corresponding increase in FECO2, and 4) RER, which was used to ensure GET did not occur at an "inflection point," which would be the result of changes in CO2 storage rather than the GET.[5,30] This measure has been validated when compared with traditional measured of GET.[5] V̇O2peak was defined as no further increase in V̇O2 with a corresponding increase in work rate. In lieu of a plateau in V̇O2 with an increase in work rate, the criteria for achievement of a V̇O2peak required attainment of two of three following measures: an RER of 1.10 or greater, an RPE of 18 or greater, and a peak HR within 10 beats (~95%) of the age predicted maximum (220 - age). All participants demonstrated either a plateau or met the secondary criteria.

Pain-based Cycling

Familiarization. The reliability of distance covered and work performed during cycling when participants are asked to regulate their exercise intensity based on ratings of muscle pain has not been established in previous studies, we felt it was important to familiarize our participants to the exercising in this manner and to establish the day-to-day reliability before the caffeine intervention. To do this, participants were asked to perform a single bout of cycling exercise for a total of 23 min on three separate occasions. An overview of the experimental protocol during this 23-min exercise bout is shown is Figure 2. Participants warmed up for 5 min at 50 W. Next, participants cycled for 3 min at the power output that previously elicited a pain rating of "1" during the V̇O2peak test. This was done to serve as a transition period between the warm-up and exercise periods. After the warm-up and transition periods, 15 min of cycling was then performed. During the 15-min bout, participants were asked to regulate their work rate based on their sensations of quadriceps muscle pain. To achieve this, participants adjusted the resistance of the cycle ergometer (which was placed in hyperbolic mode where power output is independent of pedal rate) such that their work rate elicited a pain rating of "3" (moderate pain). Participants were instructed to reach a "3" on the pain scale within the initial 3 min of the exercise period (all subjects complied), as all dependent variables were recorded every 3 min. By setting the power output to a wattage that elicited a pain rating of "1" during the transition period, participants had to consciously alter the work rate to produce and sustain a pain rating of "3." Heart rate, V̇O2, and RER were collected continuously during the bout, whereas work performed (kJ), distance covered (m), rpm, RPE, and quadriceps pain ratings were collected every third minute during the 15-min bout. Reported values for HR, V̇O2, and RER are the average of the previous 1 min during the exercise bout. Participants were blinded to all dependent variables (including power output) except for rpm, as participants were instructed to keep the rpm above 60 at all times. If participants were able to reliably cover a similar distance and perform a similar amount of work (judged by a coefficient of variation of <8.0%) over the final two testing sessions, they were asked to participate in the subsequent caffeine/placebo testing sessions. Distance covered and work performed were reliable over the three familiarization sessions with intraclass correlation coefficient (ICC:[1,3] two-way mixed, single measure model) yielding values of 0.95 and 0.95 for distance and work performed, respectively. The mean coefficient of variation (CV) over the three testing sessions was 4.4% and 4.5%, respectively. Reliability of distance covered (ICC of 0.97, CV of 3.5%) and work performed (ICC of 0.97, CV of 3.4%) improved if only the final two familiarization session were used.

Figure 2.

Schematic diagram of experimental procedures.

Caffeine/Placebo Sessions at a Pain Rating of "3.". Testing was conducted in a double-blind, placebo-controlled manner where participants served as their own control. The exercise protocol was identical that used during the three familiarization sessions. Participants consumed a commercially available caffeine supplement consisting of two pills containing a total of 400 mg of caffeine or a placebo 60 min before beginning exercise. This yielded a caffeine of dose ranging from 4.8 to 5.3 mg·kg−1 of body weight per participant (the mean among the 13 participants was 5.2 mg·kg−1 of body weight) and was based on previous studies examining the effects of a similar (~5 mg·kg−1 of body weight) dose of caffeine on muscle pain and exercise performance.[7,27] Based on previous findings,[20] this dose likely resulted in plasma caffeine concentrations of approximately 30 to 40 μM·L−1 at the onset of exercise. Orange-flavored breath mints (not containing menthol), which were similar in size and appearance to the caffeine pills were used as placebos. Both caffeine and placebo pills were consumed with approximately 25 mL of a zero-calorie orange-flavored sports drink to aid in masking pill content. One researcher (C. D. B.) performed the randomization using an online randomization program (research randomizer) and placed each set of pills into envelopes marked as testing session 1 and session 2. Pills were removed from the envelopes and given to participants by a researcher who did not participate in data collection. Participants were asked to close their eyes, and each pill was placed into their hand and consumed with their eyes closed to aid in masking pill content. All testing was performed by a single researcher (A.R.G.) who was blinded to pill content. After 60 min of quiet rest, participants repeated the cycling protocol as performed during the familiarization sessions.

Familiarization to Cycling at a Pain Rating of "5.". A single cycling session was performed to familiarize the nine participants with cycling at an intensity sufficient to produce a pain rating of "5." The familiarization protocol consisted of 18 min of cycling. The 8-min warm-up consisted of 5 min of cycling at 50 W, followed by 2 min of cycling at the wattage that elicited a pain rating of "1" during their V̇O2peak test, and a final minute of cycling at the wattage that elicited a pain rating of "3." At the conclusion of the warm-up, the participants were then instructed to increase their power output to produce and sustain a pain rating of "5" (strong pain)[10] for 10 min.

Caffeine/Placebo Sessions at a Pain Rating of "5.". Testing again occurred in a double-blind manner, and caffeine and placebo pills were again administered 60 min before beginning exercise as previously described. The exercise protocol was altered from the familiarization session to aid participants in quickly reaching a pain level of "5." An overview can be seen in Figure 2. The protocol lasted 20 min: a 5-min warm-up at 50 W, a 5-min transition stage consisting of 2 min at the wattage that elicited a "1," 2 min at the wattage that elicited a "3," 1 min at the wattage that was halfway between the power output that elicited a "3" and the power output that elicited a "5" during the familiarization trial, and 10 min of cycling at a work rate eliciting a pain rating of "5." The extended warm-up and transition served to protect the participants from a potential "overshoot" of the power output and possible premature fatigue as they attempted to achieve a pain level of "5." Heart rate, V̇O2, RER, work performed, distance covered, RPM, RPE, and quadriceps pain ratings were collected every 2 min during the 10-min bout. Heart rate, V̇O2, and RER values were recorded as the average of the previous 1 min to account for changes in power output as controlled by the participants. Participants were again blinded to all dependent variables (including power output) except for rpm, as participants were instructed to keep the rpm above 60 at all times.

Statistical Analyses

Dependent t tests were used to test for differences in total distance covered and total work accumulated as well as the mean values of power output, RER, HR, RPE, V̇O2, percentage of GET, muscle pain ratings, and RPMs between caffeine and placebo trials. A two-condition (caffeine and placebo) × 5 (time) repeated measures analysis of variance was used to test for differences for heart rate, V̇O2, average power output, RER, RPE, and muscle pain rating across the exercise trials. All data were analyzed for significant condition–time interactions. When simple effects were found, comparisons were made using a Bonferroni correction for multiple comparisons. In the absence of a significant condition–time interaction, main effects were examined. In the case of a significant main effect, main comparisons were made using a Bonferroni correction. Statistical significance was set a priori at α < 0.05.