Comparing Sensitivity and the Nociceptive Flexion Reflex Threshold Across the Mid-follicular and Late-luteal Menstrual Phases in Healthy Women

Emily J. Bartley, MS and Jamie L. Rhudy, PhD

on pain20 and sex differences in pain21 and by observations Objectives: Understanding the relationship between the menstrual that clinical pain worsens during the late-luteal (premen- cycle and pain can contribute significantly to our knowledge of strual) phase in women with and without clinical pain dis- pain processing in women. Many early studies suggested that pain orders.10,22–26 Indeed, animal literature suggests that sex sensitivity was enhanced during the luteal phase of the menstrual hormones alter nociceptive reactivity and response to an- cycle relative to the follicular phase; however, these studies were 27 often limited by small sample sizes, lack of ovulation verification, algesics. Given this, the clinical importance of inves- focus on a single pain modality, inadequate assessment of men- tigating menstrual cycle effects on nociceptive processing strual cycle regularity, and low-powered statistical methods. The are likely to be manifold (eg, characterizing pain in women, current study was designed to address these limitations and ex- informing therapeutic interventions for hormonally influ- amine the difference in pain processing between the mid-follicular enced pain conditions, identification of hormonal modu- (days 5 to 8) and late-luteal (days 1 to 6 preceding menses) phases. lators of pain, and understanding sex differences in pain). Methods: Forty-one healthy, regularly cycling women attended A 1999 meta-analysis of experimental pain studies testing sessions that measured pain sensitivity from mechanical pain concluded that pain evoked by most stimuli was enhanced 28 threshold, electrocutaneous pain threshold/tolerance, and ischemia during the luteal phase, relative to the follicular phase. pain threshold/tolerance, as well as McGill Pain Questionnaire More recently, however, Sherman and LeResche1 have qsensory and affective ratings of electric and ischemic stimuli. questioned whether clear conclusions can be drawn from Electrocutaneous stimulation was also used to assess nociceptive these studies, noting several methodological limitations in flexion reflex threshold, a physiological measure of spinal . the procedures used. For example, several studies failed to Results: When analyses were limited to data collected only in the have women monitor multiple menstrual cycles to establish targeted menstrual phases (N = 30), results indicated no menstrual cycle regularity. Many studies also did not verify ovulation. phase effect on any pain outcome (all P’s > 0.05), with the ex- Given that anovulation is common and changes the hor- ception of lower electrocutaneous pain thresholds during the late- monal milieu,5,29,30 this could ultimately alter the associa- luteal phase. No outcomes differed by menstrual phase in the full tion between menstrual phase and pain. Moreover, without sample (N = 41). This indicates nociceptive responding varies little clear identification of the time of ovulation, it is difficult to between the mid-follicular and late-luteal phases. ensure that all participants are being tested at similar times Discussion: The present study suggests that experimental pain pro- during the menstrual cycle (eg, early-follicular, mid-fol- cessing does not significantly differ between the mid-follicular and licular, ovulation, mid-luteal, and late-luteal). This can be late-luteal phases of the menstrual cycle in healthy women. This im- critical given that sex hormones vary considerably both plies hormonal variation across these 2 phases (ie, progesterone) has a between and within follicular and luteal phases. minimal effect on subjective and physiological responses to pain. Furthermore, most studies have used small samples. Key Words: menstrual cycle, pain sensitivity, nociception, noci- Although recruiting large samples in menstrual cycle re- ceptive flexion reflex search is notably difficult, small samples may limit the generalizability of findings and cause low power and spu- (Clin J Pain 2013;29:154–161) rious results. Another issue that may contribute to hetero- geneity in study findings is that many previous studies have he relationship between the menstrual cycle and ex- assessed pain sensitivity from a single stimulus modality. perimental pain has been a topic of much attention,1–19 Different stimulus modalities activate different receptor T 31 in part fueled by interest in the influence of sex hormones types and result in varying pain experiences ; thus, re- sponses to some stimuli may be more susceptible to men- strual effects than others. Of the studies published since Received for publication April 21, 2011; revised September 6, 2011; Sherman and LeResche’s1 critical review, only 1 addressed accepted January 22, 2012. 32 From the Department of Psychology, The University of Tulsa, Tulsa, OK. the methodological limitations of previous studies. That This work was funded by a Future Leaders in Pain Research Grant study by Klatzkin et al32 found that experimental heat pain, awarded to J.L.R. from the American Pain Society. The authors cold pain, and ischemic pain did not differ across menstrual declare no conflict of interest. phases (early-follicular, late-follicular, and luteal) and that Reprints: Jamie L. Rhudy, PhD, Department of Psychology, The University of Tulsa, 800 South Tucker Drive, Tulsa, OK 74104 sex hormones (estradiol and progesterone) did not correlate (e-mail: [email protected]). with measures of pain sensitivity in either menstrual phase. Supplemental Digital Content is available for this article. Direct URL Moreover, results from other human studies assessing the citations appear in the printed text and are provided in the HTML relationship between sex hormones and pain have been and PDF versions of this article on the journal’s Website, 9,13,15,17,19,33 www.clinicalpain.com. contradictory, most likely due to limitations in Copyright r 2013 by Lippincott Williams & Wilkins study methodologies.20

154 | www.clinicalpain.com Clin J Pain  Volume 29, Number 2, February 2013 Clin J Pain  Volume 29, Number 2, February 2013 Menstrual Cycle and Nociceptive Reactivity

The present study was designed to overcome many of women who failed to complete both testing sessions, 3 the limitations noted by Sherman and LeResche1 through: completed the mid-follicular phase, whereas 4 completed (1) recruitment of a relatively large sample size; (2) ver- the late-luteal phase. All participants provided full in- ification of ovulation and timing of phases by luteinizing formed consent to participate and were provided an hon- hormone (LH) surge assessment; (3) confirmation of men- orarium up to $185 for participation (or course credit if strual phase regularity; and (4) assessment of pain sensi- recruited from the University of Tulsa subject pool). tivity from multiple stimulus modalities. Given that clinical pain is often exacerbated during the late-luteal phase, we Menstrual Cycle Monitoring and Phase compared pain responsivity during this phase with a phase Determination associated with low levels of female sex hormones (mid- Participants completed a modified version of the follicular phase). Nociceptive responding was assessed from Prospective Record of the Impact and Severity of Menstrual mechanical pain threshold, ischemia pain threshold/toler- Symptoms (PRISM)36 for daily report of emotional and ance, electrocutaneous pain threshold/tolerance, and noci- physical symptoms, basal body temperature (BBT), and LH ceptive flexion reflex threshold (NFR; a physiological test results. The PRISM calendar contains affective (eg, de- correlate of spinal nociception). In addition, sensory and pressed), behavioral (eg, insomnia), and physical (eg, breast affective pain ratings of suprathreshold electric and ische- tenderness) symptoms that are rated daily for severity mic stimuli were assessed. (0 = absent, 1 = mild, 2 = moderate, 3 = severe). Participants To our knowledge, there have only been 5 studies that completed calendars daily for 3 consecutive menstrual cycles. have assessed experimental pain sensitivity during the mid- To discourage retrospective reporting, participants were asked follicular and late-luteal phases of the menstrual cycle,2,9,15,19,34 to mail in calendars on a weekly basis. Experimental pain with results being mixed. Two studies found enhanced ische- assessments were scheduled within the mid-follicular and late- mic pain sensitivity during the late-luteal phase2,19;whereas luteal phases of the participant’s menstrual cycle during cycles some found no phase differences in ischemia pain,9 thermal 2 and 3. The mid-follicular phase was defined as days 5 to 8 pain,19 cold pressor pain,15 or mechanical pressure pain.34 after menses onset. Of the 37 women who completed the mid- Unfortunately, none of these studies met all of the methodo- follicular testing session, 33 females (89%) were tested during logical recommendations of Sherman and LeResche.1 There- days 5 to 8, with an average of 7.54 days after menses onset fore, it is unclear what differences in pain sensitivity would be (SD = 1.17). The average testing day for the other 4 women noted between the mid-follicular and late-luteal phases. was day 9.5 (range, 9 to 10), because their menses lasted longer than 8 days. The late-luteal phase was defined as days 1 to 6 MATERIALS AND METHODS preceding menses (or approximately 9 to 11 d after ovulation if luteal phase length was unable to be estimated from cycle 1 Participants due to anovulation). The 9- to 11-day timeframe was based on Participants were recruited from the surrounding evidence suggesting the mean luteal phase length for normal, community by means of radio/newspaper advertisement, menstruating women to be approximately 12.4 days (range, 9 flyers, and email distribution. In addition, a few partic- to 15 d).37 Therefore, 9 to 11 days after ovulation should ipants were recruited from the University of Tulsa psy- generally fall within days 1 to 6 preceding menses onset. Of the chology subject pool. Participants were excluded for: <18 38 participants who completed their late-luteal phase ex- years of age; menopausal or postmenopausal; use of hor- perimental session, 30 females (79%) were seen 1 to 6 days mone preparations in the last 6 months; failure to regularly preceding the commencement of their subsequent cycle, with cycle; hysterectomy; pregnant or trying to get pregnant; average timing occurring 4.84 days before menses (SD = 3.24). pregnant or breastfeeding in the previous 6 months; body The average testing day for the other 8 women was 9.25 days mass index >35 (due to potential difficulties obtaining an preceding menses (range, 7 to 15), which was due to difficulties NFR in individuals with high adiposity); history of car- predicting the onset of menses from prior cycles and LH diovascular, neuroendocrine, or neurological disorders; surges. Raynaud’s disease; hypertension; history of ; Verification of ovulation was conducted by BBT and current opioid, antidepressant, or anxiolytic medication LH surge (assessed from self-administered home urine use; or recent psychological trauma as defined by DSM-IV- QTests), and this information was also used for scheduling TR.35 Fifty-three participants met initial inclusion criteria of the late-luteal phase assessment. Participants were asked and agreed to participate; however, 10 women were later to date and retain each positive QTest for confirmation of determined to not meet inclusion criteria (7 anovulatory, 1 results by an experimenter. irregular cycle, 1 using medication, and 1 using hormone preparations). Two women withdrew reporting Apparatus the procedures were too painful. Hence, 41 women were A computer running LabVIEW software (National included in the final analyses. Most were white non- Instruments, Austin, TX) equipped with dual monitors and Hispanic (71%), married (73%), and employed full-time A/D board (National Instruments, PCI-6036E) controlled (56%), with an average age of 31 years (SD = 8.86) and an all stimuli, questionnaire presentation, and data acquisition average education of 15 years (SD = 1.79). All participants and was used for offline data reduction. Physiological sig- had regular menstrual cycles ranging from 21.67 to 37.77 nals and experimental timing were monitored by an ex- days (M = 28.98, SD = 3.28). Average luteal phase length perimenter in an adjacent room by use of a 17” flat panel was 14.74 days (SD = 3.48) as calculated from the date of monitor. Questionnaires were presented by an LCD pro- positive LH surge until 1 day preceding menses onset. jector onto a large screen positioned approximately 2 m in Thirty-four women completed both experimental testing front of the participant. attenuating headphones and sessions, whereas 7 participants only completed 1 testing a video camera allowed the experimenter to communicate session (reported reasons: scheduling, health problems un- with and monitor the participant from an adjoining room. related to menstrual cycle, procedures too painful). Of the 7 Noxious electrocutaneous stimulations were delivered to r 2013 Lippincott Williams & Wilkins www.clinicalpain.com | 155 Bartley and Rhudy Clin J Pain  Volume 29, Number 2, February 2013 the left ankle over the retromalleolar pathway of the sural Electrocutaneous Threshold and Tolerance nerve by use of a Grass Instruments stimulator (Model S88; Electrocutaneous pain threshold and tolerance were West Warwick, RI), stimulus isolation unit (Model SIU8T), assessed by a single-ascending staircase of electric stimuli constant current unit (Model CCU1), and bipolar stim- over the sural nerve. After each electric stimulus, partic- ulating electrode (019-401400; Nicolet, Madison, WI). The ipants rated their sensation using a computer-presented onset/offset of the stimulator was controlled by computer, numerical rating scale oriented vertically.45,46 Arranged and a computer-controlled voltage regulator varied the from bottom to top, the scale was labeled: 0 (no sensation), current to the participant (maximum current = 40 mA). 1 (just noticeable), 25 (uncomfortable), 50 (painful), 75 Mechanical pressure pain was measured with a Wagner (very painful), and 100 (maximum tolerable). Participants Instruments Force Ten FDX Digital Force Gage with a responded by moving an indicator along the scale and 1.1-cm diameter tip (Wagner Instruments, Greenwich, CT). submitted their answers by computer mouse. Electric A Lafayette Instrument Hand Dynamometer (Models stimulations started at 0 mA and increased until the par- 78010 and 78011; Lafayette, IN) and Omron Tru-Gage ticipant rated a stimulus as 100 or until the maximum of blood pressure cuff were used during the ischemic pain 40 mA was achieved. Electrocutaneous pain threshold was procedures. defined as the first stimulus (in mA) rated Z50 on the rating scale, whereas electrocutaneous was Electrode Application and NFR Signal Acquisition defined as the stimulus (in mA) rated 100 (or 40 mA if the Biceps femoris electromyography (EMG) for NFR maximum current was reached). recording was sampled at 1000 Hz and collected/filtered using a Grass Instruments Model 15LT Bipolar Amplifier Ischemic Threshold and Tolerance with Quad AC (15A54) module. Recording electrodes were To assess ischemia pain sensitivity, hand exercises were Ag-AgCl. To apply recording and stimulating electrodes, first conducted at 50% maximal effort with the left hand for the skin was cleaned with alcohol, slightly abraded using 60 seconds with a dynamometer, followed by 15 seconds of NuPrep gel to attain impedances below 5 kO, and then arm elevation for exsanguination. Then, a blood pressure conductive gel (EC60, Grass Instruments) was applied. cuff was affixed to the left arm and inflated to 220 mm/Hg. NFR was recorded from 2 electrodes over the biceps fem- After inflation, participants conducted another 20 hand oris muscle of the left leg 10 cm superior to the popliteal exercises at 1-second intervals. Participants were then asked fossa, and a common ground electrode was placed over the to rate the sensation evoked by the blood pressure cuff on lateral epicondyle of the left femur. The raw biceps femoris the numerical rating scale, and the time until they reached EMG was amplified, bandpass filtered (10 to 300 Hz), and pain threshold (rating of 50 on the numerical rating scale) rectified. and tolerance (rating of 100 on the numerical rating scale) Assessment of Pain and Nociception was assessed. To ensure participant safety, a maximum of 25 minutes was used.19 NFR Threshold Assessment The NFR is a spinally mediated withdrawal reflex McGill Pain Questionnaire-Short Form (MPQ-SF) elicited by activation of Ad fibers after noxious stim- The MPQ-SF47 is a reliable and valid pain rating scale 38,39 ulation. NFR threshold was determined by delivering commonly used in pain research that allows quantitative, trains of five 1-ms rectangular wave pulses at 250 Hz to the multidimensional pain ratings to be obtained in a brief sural nerve with a varying intertrain interval of 8 to 12 period. Participants were asked to rate 11 sensory (eg, seconds to reduce stimulus predictability. The first train throbbing, shooting) and 4 affective (eg, sickening, fearful) began at 0 mA (current) and was increased in 1.5 mA steps pain descriptors on a scale ranging from 0 (none) to 3 until an NFR was detected. The mean biceps femoris EMG (severe). The MPQ-SF was administered twice, once im- activity 90- to 150-ms poststimulation that exceeded mean mediately after electrocutaneous pain tolerance and once EMG activity during the 60-ms prestimulus baseline inter- immediately after ischemia pain tolerance. Sums of sensory 40 val by 1.4 SD was used to define the presence of a reflex. words and affective words were used to compute sensory Prior research has shown that using the 90- to 150-ms and affective pain rating scores, respectively, for electro- timeframe avoids potential contamination by the non- cutaneous and ischemia stimuli. nociceptive RII reflex, startle responses, and voluntary movements.41–42 The stimulus intensity was decreased in Menstrual Cycle-related Symptoms 0.75 mA steps until an NFR was no longer observed. Then, The modified version of the PRISM36 Calendar was this up-down staircase procedure was repeated twice but by used to assess daily behavioral, affective, and physical using 0.5 mA steps. NFR threshold was defined as the symptomatology. Participants were required to rate items average stimulation intensity (in mA) of the last 2 peaks for severity on a 4-point Likert scale ranging from 0 (ab- and troughs of this up-down staircase procedure.40,43 sent) to 3 (severe). Menstrual cycle-related symptoms from the PRISM symptom severity ratings were averaged across Mechanical Pressure Pain 4 days (3 d before testing + day of testing) for each of the Pressure pain threshold was assessed bilaterally by menstrual phases in which participants were tested. applying mechanical pressure on 3 body sites (occiput, lateral epicondyle, and fatty knee pad). These sites were Procedure chosen from tender points used to assess fibromyalgia cri- All procedures were fully approved by the University teria44 and include an assessment of all 4 body quadrants. of Tulsa Institutional Review Board. A brief phone health Pain threshold (in kg) was assessed from each site by in- screening was initially conducted to evaluate inclusion/ creasing the pressure at 1 kg/s until the participant reported exclusion criteria; however, a more comprehensive assess- pain. Pain threshold was defined as the average of each ment was conducted during an initial laboratory session bilateral body site. with those individuals not excluded by the phone screen.

156 | www.clinicalpain.com r 2013 Lippincott Williams & Wilkins Clin J Pain  Volume 29, Number 2, February 2013 Menstrual Cycle and Nociceptive Reactivity

During the initial laboratory session, participants were curred slightly outside of the targeted window. Only results given a complete overview of the study, and informed from the first set of analyses are reported, although both consent was obtained. Then, a brief demographics and produced similar outcomes. The data from the second set of health status form was used to assess inclusion/exclusion analyses are presented as supplementary material for the criteria and attain relevant background information. If interested reader (Supplemental Digital Content 1, http:// deemed eligible, the participant was instructed on how to links.lww.com/CJP/A40). To correct for positive skew, all monitor their menstrual cycle and then randomly assigned ischemic pain variables were log-transformed. Cohen d to a testing order (ie, mid-follicular/late-luteal vs. late- was reported as the effect size for mean comparisons. Sig- luteal/mid-follicular). nificance was set at P < 0.05 (2 tailed). Participants were scheduled for 2 experimental testing sessions, once in the mid-follicular phase (days 5 to 8) and once during the late-luteal phase (1 to 6 d before menses RESULTS onset) of their menstrual cycles. Each experimental testing Menstrual Cycle-related Symptoms session was scheduled at approximately the same time of the Table 1 reports means, SDs, and effect sizes for men- day to control for potential diurnal fluctuations in pain strual cycle-related symptoms for the PRISM calendar for processing. At home, participants monitored 3 complete cy- data collected during the targeted phase windows. Results cles. Cycle 1 was used to establish cycle length and ovulation revealed fatigue (F1,35 = 4.29, P = 0.046) and depression timing. Experimental testing sessions occurred during cycles 2 (F1,33 = 5.66, P = 0.023) were rated of higher severity and 3. Upon arrival to each experimental testing session, a during the mid-follicular phase, whereas breast tenderness complete overview of the procedures was provided, followed was more severe during the late-luteal phase (F1,38 = 4.30, by review of informed consent and health status. Afterwards, P = 0.045). None of the other variables from the PRISM participants were provided instruction for using the numer- calendar varied by menstrual phase (all P’s > 0.05). ical pain rating scale. Mechanical pain threshold was then assessed, followed by application of electrodes. Next, par- Measures of Pain Sensitivity ticipants completed questionnaires and then sat quietly for Table 2 reports means, SDs, and effect sizes for meas- 5 minutes to help habituate to the testing environment. The ures of pain sensitivity for data collected inside the targeted rest of the session included pain testing in the following order: phase windows. There were no significant differences found NFR threshold, emotional controls of nociception, diffuse for any of the pain outcomes (all Ps > 0.05), with the ex- noxious inhibitory controls (DNIC), electrocutaneous pain ception of lower electrocutaneous pain thresholds during the threshold and tolerance, MPQ-SF rating of electrocutaneous late-luteal phase relative to the mid-follicular phase (F1,30 = tolerance, ischemic pain threshold and tolerance assessment, 4.39, P = 0.049). When data from the full sample were con- and MPQ-SF rating of ischemia tolerance. Emotional con- sidered (even if a woman was tested slightly outside of the trols of nociception and DNIC results have been reported 48,49 targeted range), results were identical, except that electro- elsewhere. At the culmination of the session, each par- cutaneous pain threshold was no longer significant (all ticipant was informed to continue menstrual phase mon- Ps > 0.05). itoring until 3 cycles were complete. Participants were provided their honorarium after all study procedures were completed. DISCUSSION The present study assessed the relationship between Data Analysis nociceptive responding and menstrual phase in healthy, All analyses were conducted using the MIXED pro- regularly cycling women and addressed the methodological cedure in SPSS 14.02 (IBM Corporation, Somers, NY).50 limitations of prior research. We recruited a relatively large Menstrual phase (mid-follicular vs. late-luteal) was entered sample (N = 41), verified menstrual timing through LH as a nominal within-subjects variable. No random effects assessment, monitored for 3 consecutive cycles to confirm were specified; therefore, this approach is similar to con- cycle regularity, assessed pain from multiple stimulus mo- ducting 1-way within-subjects analysis of variances except dalities, and used powerful statistical analyses. Although that maximum likelihood (ML) estimation is used instead some studies have suggested enhanced pain during the late- of general linear model (GLM). Although ML generally luteal phase relative to the mid-follicular phase,2,19 we provides similar results as GLM, there are 3 major ad- found no significant differences in any pain outcome be- vantages to the MIXED procedure and ML: (1) cases with tween the mid-follicular and late-luteal phases, with the missing data on the within-subject variable are not excluded exception of lower electrocutaneous pain thresholds during thus taking full advantage of all data points; (2) statistical the late-luteal phase. These findings are also similar to the power is increased, because cases with missing data are not rigorous study by Klatzkin et al32 who tested 49 women in excluded; and (3) the variance-covariance structure can be the early-follicular, late-follicular, and luteal phases. So, fit to the data rather than being fixed to compound sym- when the 2 studies are taken together, this suggests that metry as is true for GLM. The MIXED procedure uses pain may vary little across the menstrual cycle in healthy Satterthwaite estimation for the denominator degrees of women when rigorous methodology is used. freedom that produces noninteger values that vary from As a cautionary note, failure to reject the null hy- analysis to analysis. These degrees of freedom were rounded pothesis can never firmly establish that a relationship does to the nearest integer for ease of reporting. Because some not exist between menstrual phase and nociceptive re- women were tested outside of the targeted range of days for sponding. However, the statistical effect sizes for pain the mid-follicular (days 5 to 8) and late-luteal (days 1 to 6 outcomes were small in both sets of analyses, ranging from preceding menses) phases, the pain data were analyzed in 2 d = 0.03 (affective pain ratings of electric stimuli) to ways: (1) with only data from women tested in the targeted d = 0.29 (electrocutaneous pain threshold). But what is phase windows and (2) with all data, even if testing oc- more, the clinical significance of menstrual effects is likely r 2013 Lippincott Williams & Wilkins www.clinicalpain.com | 157 Bartley and Rhudy Clin J Pain  Volume 29, Number 2, February 2013

TABLE 1. Means, SDs, Observed Ranges, and Effect Sizes of PRISM Calendar Symptoms by Menstrual Phase Mid-Follicular Late-Luteal Mean SD Observed Range Mean SD Observed Range Cohen d Irritable 0.23 0.38 0-1.50 0.18 0.33 0-1.25 0.14 Fatigue* 0.53 0.61 0-2.50 0.26 0.47 0-1.50 0.50 Inward anger 0.10 0.33 0-1.50 0.03 0.11 0-0.50 0.28 Labile mood 0.16 0.33 0-1.50 0.04 0.12 0-0.50 0.48 Depressed* 0.25 0.42 0-1.25 0.07 0.15 0-0.50 0.57 Restless 0.13 0.28 0-1.00 0.14 0.43 0-2.25 0.03 Anxious 0.27 0.48 0-2.00 0.14 0.29 0-1.25 0.33 Insomnia 0.12 0.36 0-1.75 0.02 0.06 0-0.25 0.39 Lack of control 0.08 0.26 0-1.25 0.00 0.00 0-1.25 0.44 Edema 0.13 0.44 0-2.50 0.06 0.18 0-0.75 0.21 Breast tenderness* 0.05 0.15 0-0.75 0.23 0.48 0-1.75 0.51 Abdominal bloating 0.24 0.62 0-3.00 0.18 0.44 0-2.00 0.11 0.28 0.55 0-2.50 0.04 0.17 0-0.75 0.59 Cravings 0.21 0.46 0-2.25 0.22 0.45 0-2.00 0.02 Feel unattractive 0.13 0.35 0-1.75 0.17 0.44 0-1.75 0.10 Guilty 0.08 0.26 0-1.25 0.01 0.05 0-0.25 0.37 Unreasonable behavior 0.04 0.19 0-1.00 0.00 0.00 0-1.00 0.30 Low Self-esteem 0.16 0.32 0-1.00 0.15 0.43 0-1.75 0.03 Nausea 0.06 0.17 0-0.75 0.06 0.28 0-1.50 0.00 Menstrual cramps 0.04 0.18 0-1.00 0.06 0.12 0-0.50 0.13

*P < 0.05. PRISM indicates Prospective Record of the Impact and Severity of Menstrual Symptoms.

to be minimal, given that menstrual phase only explained a Relation to the Prior Literature maximum of 2% of the variance in pain sensitivity in the A 1999 meta-analysis concluded that, relative to the present study (electrocutaneous pain threshold). It is worth follicular phase, the luteal phase is associated with hyper- noting that electrocutaneous pain threshold achieved sig- algesia for mechanical pressure, ischemia, cold pressor, and nificance at P = 0.049 when the analysis was restricted to thermal heat pain; but, electrocutaneous pain had the op- the 30 women tested only in the strict windows we defined posite pattern ( during follicular).28 However, as the mid-follicular and late-luteal phases. However, future in 2006, Sherman and LeResche1 conducted a critical re- research is needed to replicate this finding to determine view of the literature and noted that many of the studies whether this is a true difference that was observed after examining the menstrual cycle and experimental pain suf- eliminating women outside of the testing windows (ie, re- fered from numerous study design limitations and/or moving error variance) or whether it is a spurious result of methodological inconsistencies that made it difficult to the smaller sample size given that only 1 out of 13 pain draw firm conclusions from the data. Of the 34 studies, we outcomes were significant. were able to identify on the topic,2,3,5,7–9,13–15,17–19,32–34,51–69

TABLE 2. Means, SDs, and Observed Ranges of Pain Outcomes by Menstrual Phase Mid-Follicular Late-Luteal Mean SD Observed Range Mean SD Observed Range Cohen d NFR threshold (0-40 mA) 9.31 4.23 2.92-21.31 10.32 5.57 3.04-23.24 0.20 Electrocutaneous pain threshold (0-40 mA) 11.26 5.15 5.05-25.99 10.02 3.07 5.20-17.41 0.29 Electrocutaneous pain tolerance (0-40 mA) 18.55 9.09 6.83-38.00 16.41 6.28 8.10-32.86 0.27 Log ischemic pain threshold (log seconds) 1.63 0.33 1.08-2.33 1.60 0.42 1.04-3.03 0.08 Log ischemic pain tolerance (log seconds) 2.24 0.36 1.43-2.93 2.21 0.38 1.26-3.10 0.08 Mechanical pressure pain threshold (kg/s) 4.56 1.16 2.82-6.90 4.81 1.23 2.66-7.57 0.21 Occiput 4.34 1.13 2.61-7.04 4.62 1.52 1.95-9.20 0.21 Epicondyle 4.40 1.07 2.64-6.47 4.68 1.36 2.30-7.17 0.23 Knee 4.95 1.77 2.71-8.90 5.12 1.72 2.91-11.39 0.10 MPQ (electric) Sensory (0-33) 15.85 6.58 4-29 14.70 5.74 4-29 0.19 Affective (0-12) 4.79 3.39 0-12 4.57 3.21 0-11 0.07 Log MPQ (ischemic) Sensory (log units) 1.19 0.22 0.48-1.49 1.17 0.20 0.48-1.41 0.10 Affective (log units) 0.64 0.33 0-1.11 0.65 0.30 0-1.00 0.03

All Ps > 0.05. MPQ indicates McGill Pain Questionnaire; NFR, nociceptive flexion reflex.

158 | www.clinicalpain.com r 2013 Lippincott Williams & Wilkins Clin J Pain  Volume 29, Number 2, February 2013 Menstrual Cycle and Nociceptive Reactivity only 16 studies verified ovulation through hormone assessment menstrual cycle influences may be more pronounced in to exclude anovulatory cycles and accurately define menstrual clinical populations. For example, Aberger et al66 found that phases,2,7,9,13–15,17–19,32–34,52–55 only 10 studies assessed pain women diagnosed with dysmenorrhea exhibit higher ischemic from multiple stimulus modalities,14,18,19,32,33,52–54,56,57 and pain thresholds and tolerances during the premenstrual only 4 studies used sample sizes >33 (as recommended by phase, relative to the menstrual and postmenstrual phases. Sherman and LeResche).13,32,58,59 Of the 9 studies we found Using rectal balloon distension, Houghton et al70 found that that were published since Sherman and LeResche’s re- pain was greater during the menstrual phase (days 1 to 4) view,13–15,17,18,32–34,69 only Klatzkin et al32 adequately ad- compared with the follicular phase (days 8 to 10) in women dressed the methodological concerns they noted and found no with irritable bowel syndrome. Further, Sherman et al53 significant differences in heat pain, cold pain, and ischemia found lower palpation pain thresholds during the ovulatory pain across the early-follicular, late-follicular, and luteal and late-luteal phases relative to the menstrual and mid-luteal phases. Their study is in accordance with ours; however, our phases in women with temporomandibular disorder. Women study supplements theirs by assessing nociceptive responding with premenstrual dysphoric disorder have also been found through mechanical pressure pain and NFR threshold, using to exhibit greater ischemic pain sensitivity relative to healthy powerful statistical models to test phase differences, and ex- women, an effect that was independent of menstrual phase.71 amining differences between the mid-follicular and late-luteal Taken together, this suggests individuals with clinical pain phases. may be more prone to menstrual phase-related changes in Of the 5 studies that specifically examined nociceptive pain sensitivity or exhibit a generalized hyperalgesia relative responding during the mid-follicular and late-luteal phases to healthy populations. of the menstrual cycle,2,9,15,19,34 4 failed to find phase dif- The current study also provides further insight into the ferences in heat,19 ischemia,9 cold,15 and pressure pain34; potential role of hormones on pain processing. Progesterone whereas 2 found that ischemia pain sensitivity was en- levels are relatively low during the mid-follicular phase and hanced during the late-luteal phase compared with the higher during the late-luteal phase.20 Given that progesterone mid-follicular phase.2,19 However, these studies had the levels should vary between these 2 phases, it seems that this methodological limitations noted by Sherman and LeRe- variation in progesterone may be unrelated to pain sensitivity. sche including small sample size2,9,15,19,34 and failure to Although there is some variation in estradiol between the mid- verify cycle regularity through multiple cycle monitor- follicular and late-luteal phases, this variation is much smaller ing.2,15,34 Moreover, most failed to assess multiple stimulus than progesterone; thus, we are not able to make a clear modalities to determine whether this contributed to the statement about the role of estradiol. However, it is important differences in results.2,9,15,34 Therefore, it is unclear whether to note that there is considerable within-phase variability in our results can be directly compared with them. Nonethe- sex hormone levels in addition to between-phase variability. less, they are generally consistent with our findings. Thus, future studies should measure intraindividual and in- terindividual variability in sex hormones to determine the Menstrual Cycle and Spinal Nociceptive relationships between hormone levels and pain processes. Processing We found that NFR threshold did not differ across Limitations mid-follicular and late-luteal phases of the menstrual cycle As previously noted, the present study had a number of (phase explained 0.4% variance in NFR threshold). Be- strengths such as use of LH tests to verify ovulation, multiple cause NFR is a measure of spinal nociceptive processes phase monitoring for verification of cycle regularity, use of that is under tonic modulation from supraspinal centers,39 powerful statistical models, recruitment of a relatively large this implies that descending modulatory processes are not sample size, and assessment of pain sensitivity from multiple affected by physiological changes associated with these stimulus modalities. Further, the current study is the first to menstrual phases. In contrast, Tassorelli et al8 found that assess pain sensitivity from mechanical pressure and the NFR thresholds were lower (enhanced nociceptive re- physiologically measured NFR having addressed method- sponding) during the early-to-mid-luteal phase (days 6 to 8 ological limitations as suggested by Sherman and LeResche.1 after ovulation) relative to the late-follicular phase (days 8 Nonetheless, some limitations must be acknowledged. First, to 10 from menses onset). Because their testing sessions we did not directly measure sex hormones; thus, statements were closer to ovulation, it is possible that descending about the role of hormones are speculative. Second, pain modulation is influenced by the physiological changes that sensitivity was assessed only during the mid-follicular and occur near ovulation. This notion is supported by a recent late-luteal phases; so, it is possible that variability in sex study18 that found descending inhibition of heat pain by hormones associated with other phases (eg, ovulation) or DNIC was enhanced during ovulation (days 12 to 14) rel- other within-phase periods (eg, mid-luteal) may also influence ative to the early-follicular (days 1 to 3) and mid-luteal pain sensitivity. Third, because of intraindividual and inter- (days 19 to 23) phases. However, it is important to note the individual variability in menstrual cycles, some women were Tassorelli and colleagues study was based on a small sam- not tested within the targeted windows that we defined as the ple of 14 women and they used BBT to verify ovulation, mid-follicular phase (N = 4) and the late-luteal phase (N =8). which is prone to errors. Therefore, it is important for their Thus, our final sample was slightly restricted in size by results to be replicated before firm conclusions are drawn. this problem, although results were nearly identical when compared with the total sample of 41 women. Indeed, results Implications and Future Directions suggested that pain sensitivity varies little between the The current study corroborates findings from a pre- mid-follicular and late-luteal phases of the menstrual cycle vious methodologically rigorous study32 and other studies in healthy women. Fourth, our sample may not represent comparing the mid-follicular and late-luteal phases9,15,34 women more generally. We found few phase-related that suggest that the menstrual cycle may exert little in- changes in menstrual symptoms, with the exception of fluence on pain sensitivity in healthy women. However, increased severity of fatigue and depressed mood during the r 2013 Lippincott Williams & Wilkins www.clinicalpain.com | 159 Bartley and Rhudy Clin J Pain  Volume 29, Number 2, February 2013 mid-follicular phase and increased breast tenderness during 14. Soderberg K, Poromaa IS, Nyberg S, et al. Psychophysically the late-luteal phase. Therefore, the possibility exists that we determined thresholds for thermal and pain recruited participants with less menstrual symptoms than the perception in healthy women across the menstrual cycle. Clin general population. And finally, our results may not gen- J Pain. 2006;22:610–616. 15. Kowalczyk WJ, Evans SM, Bisaga AM, et al. Sex differences eralize to clinical populations. For instance, menstrual phase and hormonal influences on response to cold pressor pain in has been found to influence disease-related symptomato- humans. J Pain. 2006;7:151–160. 10,72 logy and experimental pain responsivity in individuals 16. Johannesson U, de Boussard CN, Brodda Jansen G, et al. with temporomandibular disorder,53 dysmenorrhea,66 and Evidence of diffuse noxious inhibitory controls (DNIC) elicited irritable bowel syndrome.70 by cold noxious stimulation in patients with provoked vestibulodynia. Pain. 2007;130:31–39. CONCLUSIONS 17. Stening K, Eriksson O, Wahren L, et al. Pain sensations to the in normally menstruating women: comparison In sum, the present study indicates that pain sensitivity with men and relation to menstrual phase and serum sex (ie, mechanical pain threshold, electrocutaneous pain thresh- steroid levels. Am J Physiol. 2007;62:R1711–R1716. old/tolerance, ischemia pain threshold/tolerance, sensory/af- 18. Tousignant-Laflamme Y, Marchand S. Excitatory and inhibitory fective ratings of electrocutaneous and ischemic stimuli) and pain mechanisms during the menstrual cycle in healthy women. spinal nociceptive processing (NFR threshold) varies mini- Pain. 2009;146:47–55. mally between the mid-follicular and late-luteal (premenstru- 19. Fillingim RB, Maixner W, Girdler SS, et al. Ischemic but not al) phases of the menstrual cycle. Future studies are needed to thermal pain sensitivity varies across the menstrual cycle. assess the specific association between sex hormones and Psychosom Med. 1997;59:512–520. 20. Fillingim RB, Ness TJ. Sex-related hormonal influences on pain nociceptive processing using rigorous methodologies and to and analgesic responses. Neurosci Biobehav Rev. 2000;24:485–501. determine whether our results extend to clinical populations 21. Fillingim RB. Sex, gender, and pain: women and men really are (eg, premenstrual dysphoric disorder, chronic pain). different. Curr Rev Pain. 2000;4:24–30. 22. Ostensen M, Rugelsjoen A, Wigers SH. The effect of reproduc- ACKNOWLEDGMENTS tive events and alterations of sex hormone levels on the symptoms of fibromyalgia. Scand J Rheumatol. 1997;26:355–360. The authors thank Mary C. Chandler, Kara L. Kerr, 23. Silberstein SD, Merriam GR. Sex hormones and . and Ashley Vincent for their valuable assistance in data J Pain Symptom Manage. 1993;8:98–114. collection and data processing for this study. 24. Kappius RE, Goolkasian P. Group and menstrual phase effect in reported headaches among college students. Headache. 1987;27:491–494. REFERENCES 25. Ader DN, South-Paul J, Adera T, et al. Cyclical mastalgia: 1. Sherman JJ, LeResche L. Does experimental pain response prevalence and associated health and behavioral factors. vary across the menstrual cycle? A methodological review. Am J Psychosom Obstet Gynaecol. 2001;22:71–76. J Physiol. 2006;291:R245–R256. 26. Huerta-Franco MR, Malacara JM. Association of physical 2. Pfleeger M, Straneva PA, Fillingim RB, et al. Menstrual cycle, and emotional symptoms with the menstrual cycle and life- blood pressure and ischemic pain sensitivity in women: a style. J Reprod Med. 1993;38:448–454. preliminary investigation. Int J Psychophysiol. 1997;27:161–166. 27. Craft RM. Modulation of pain by estrogens. Pain. 2007; 3. Tedford WH, Warren DE, Flynn WE. Alteration of shock 132:S3–S12. aversion thresholds during the menstrual cycle. Percept 28. Riley JL III, Robinson ME, Wise EA, et al. A meta-analytic Psychophys. 1977;21:193–196. review of pain perception across the menstrual cycle. Pain. 4. Goolkasian P. Phase and sex effects in pain perception: a 1999;81:225–235. critical review. Psychol Women Q. 1985;9:15–28. 29. Metcalf MG, Skidmore DS, Lowry GF, et al. Incidence of 5. Hapidou EG, De Catanzaro D. Sensitivity to cold pressor pain ovulation in the years after the menarche. J Endocrino. in dysmenorrheic women as a function of menstrual cycle phase. 1983;97:213–219. Pain. 1988;34:277–283. 30. Girdler SS, Pedersen CA, Stern RA, et al. Menstrual cycle and 6. Fillingim RB, Ness TJ. The Influence of Menstrual Cycle and premenstrual syndrome: modifiers of cardiovascular reactivity Sex Hormones on Pain Responses in Humans. Seattle, WA: in women. Health Psychol. 1993;12:180–192. IASP Press; 2000:191–207. 31. Rainville P, Feine JS, Bushnell MC, et al. A psychophysical 7. Cimino R, Farella M, Michelotti A, et al. Does the ovarian cycle comparison of sensory and affective responses to four modalities influence the pressure-pain threshold of the masticatory muscles of experimental pain. Somatosens Mot Res. 1992;9:265–277. in symptom-free women? JOrofacPain. 2000;14:105–111. 32. Klatzkin RR, Mechlin B, Girdler SS.. Menstrual cycle phase 8. Tassorelli C, Sandrini G, Cecchini AP, et al. Changes in does not influence gender differences in experimental pain nociceptive flexion reflex threshold across the menstrual cycle sensitivity. Eur J Pain. 2010;14:77–82. in healthy women. Psychosom Med. 2002;64:621–626. 33. Teepker M, Peters M, Vedder H, et al. Menstrual variation in 9. Straneva PA, Maixner W, Light KC, et al. Menstrual cycle, experimental pain: correlation with gonadal hormones. Neuro- beta-endorphins, and pain sensitivity in premenstrual dys- psychobiology. 2010;61:131–140. phoric disorder. Health Psychol. 2002;21:358–367. 34. Kowalczyk WJ, Sullivan MA, Evans SM, et al. Sex differences 10. LeResche L, Mancl L, Sherman JJ, et al. Changes in temporo- and hormonal influences on response to mechanical pressure mandibular pain and other symptoms across the menstrual cycle. pain in humans. J Pain. 2010;11:330–342. Pain. 2003;106:253–261. 35. APA. Diagnostic and Statistical Manual of Mental Disorders. 11. Mrykalo M. Menstrual cycle phase, gender, and catastroph- 4th Edition Text Revision. Arlington, VA: American Psychi- izing influences on pain perception. Diss Abst Int: B: Sci Eng. atric Association; 2000:463–472. 2005;65:4297. 36. Reid RL. Premenstrual syndrome. In: Leventhal JM, Hoffman 12. Kibler JL, Rhudy JL, Penzien DB, et al. Hormones, menstrual JJ, Keith LG, Taylor PJ, eds. Current Problems in Obstetrics, distress, and migraine across the phases of the menstrual cycle. Gynecology, & Fertility, Vol. Chicago: Year Book Medical Headache. 2005;45:1181–1189. Publishers; 1985:345–353. 13. Okifuji A, Turk DC. Sex hormones and pain in regularly 37. Fehring RJ, Schneider M, Raviele K. Variability in the phases menstruating women with fibromyalgia syndrome. J Pain. of the menstrual cycle. J Obst Gynecol Neonatal Nurs. 2006;7:851–859. 2006;35:376–384.

160 | www.clinicalpain.com r 2013 Lippincott Williams & Wilkins Clin J Pain  Volume 29, Number 2, February 2013 Menstrual Cycle and Nociceptive Reactivity

38. Skljarevski V, Ramadan NM. The nociceptive flexion reflex 57. Amodei N, Nelson-Gray RO. Reactions of dysmenorrheic and in humans: review article. Pain. 2002;96:3–8. nondysmenorrheic women to experimentally induced pain 39. Sandrini G, Serrao M, Rossi P, et al. The lower limb flexion throughout the menstrual cycle. J Behav Med. 1989;12:373–385. reflex in humans. Prog Neurobiol. 2005;77:353–395. 58. Hapidou EG, Rollman GB. Menstrual cycle modulation of 40. Rhudy JL, France CR. Defining the nociceptive flexion reflex tender points. Pain. 1998;77:151–161. (NFR) threshold in human participants: a comparison of 59. Rao SS, Ranganekar AG, Saifi AQ. Pain threshold in relation different scoring criteria. Pain. 2007;128:244–253. to sex hormones. Indian J Physiol Pharmacol. 1987;31:250–254. 41. Dowman R. Possible startle response contamination of the 60. Goolkasian P. An ROC analysis of pain reactions in spinal nociceptive withdrawal reflex. Pain. 1992;49:187–197. dysmenorrheic and nondysmenorrheic women. Percept Psy- 42. France CR, Suchowiecki S. Assessing supraspinal modulation chophys. 1983;34:381–386. of pain perception in individuals at risk for hypertension. 61. Hellstrom B, Lundberg U. Pain perception to the cold pressor Psychophysiology. 2001;38:107–113. test during the menstrual cycle in relation to estrogen levels 43. France CR, Rhudy JL, McGlone S. Using normalized EMG to and comparison with men. Integr Physiol Behav Sci. 2000;35: define the nociceptive flexion reflex (NFR) threshold: further 132–141. evaluation of standardized scoring criteria. Pain. 2009;145:211–218. 62. Robinson JE, Short RV. Changes in breast sensitivity at 44. Wolfe F, Smythe HA, Yunus MB, et al. The American College puberty, during the menstrual cycle, and at parturition. BMJ. of Rheumatology 1990 criteria for the classification of 1977;1:1188–1191. fibromyalgia. Report of the multicenter committee. Arthritis 63. Kuczmierczyk AR, Adams HE, Calhoun KS, et al. Pain Rheum. 1990;33:160–172. responsivity in women with premenstrual syndrome across the 45. France CR, France JL, al’Absi M, et al. Catastrophizing is menstrual cycle. Percept Mot Skills. 1986;63:387–393. related to pain ratings, but not nociceptive flexion reflex 64. Isselee H, De Laat A, Bogaerts K, et al. Long-term fluctuations threshold. Pain. 2002;99:459–463. of pressure pain thresholds in healthy men, normally menstru- 46. Rhudy JL, Williams AE, McCabe K, et al. Affective ating women and oral contraceptive users. Eur J Pain. modulation of nociception at spinal and supraspinal levels. 2001;5:27–37. Psychophysiology. 2005;42:579–587. 65. Drobek W, Schoenaers J, De Laat A. Hormone-dependent 47. Melzack R. The short-form McGill Pain Questionnaire. Pain. fluctuations of pressure pain threshold and tactile threshold of 1987;30:191–197. the temporalis and masseter muscle. J Oral Rehabil. 2002; 48. Bartley EJ, Chandler MC, Rhudy JL. The influence of the 29:1042–1051. menstrual cycle on supraspinal modulation of pain and spinal 66. Aberger EW, Denney DR, Hutchings DF. Pain sensitivity and nociception. Program No. 854.5.W31. 2009. 2009; Society for coping strategies among dysmenorrheic women: much ado Neuroscience. Chicago, IL. about nothing. Behav Res Ther. 1983;21:119–127. 49. Rhudy JL, Bartley EJ. The effect of the menstrual cycle on 67. Giamberardino MA, Berkley KJ, Iezzi S, et al. Pain threshold affective modulation of pain and nociception in healthy women. variations in somatic wall tissues as a function of menstrual Pain. 2010;149:365–372. cycle, segmental site and tissue depth in non-dysmenorrheic 50. Hox JJ. Multilevel Analysis. Mahwah, NJ: Lawrence Erlbaum women, dysmenorrheic women and men. Pain. 1997;71: Associates; 2002. 187–197. 51. Goolkasian P. Cyclic changes in pain perception: an ROC 68. Herren RY. The effect of high and low female sex hormone analysis. Percept Psychophys. 1980;27:499–504. concentration on the two-point threshold of pain and touch 52. Bajaj P, Arendt-Nielsen L, Madsen H. Sensory changes during and upon tactile sensitivity. J Exp Psychol. 1933;16:324–327. the ovulatory phase of the menstrual cycle in healthy women. 69. Vignolo V, Vedolin GM, De Araujo CDRP, et al. Influence of Eur J Pain. 2001;5:135–144. the menstrual cycle on the pressure pain threshold of masticatory 53. Sherman JJ, LeResche L, Mancl LA, et al. Cyclic effects on muscles in patients with masticatory myofascial pain. Oral Surg experimental pain response in women with temporomandibu- Oral Med Oral Pathol Oral Radiol Endod. 2008;105:308–315. lar disorders. J Orofac Pain. 2005;19:133–143. 70. Houghton LA, Lea R, Jackson N, et al. The menstrual cycle 54. Gazerani P, Andersen OK, Arendt-Nielsen L. A human affects rectal sensitivity in patients with irritable bowel experimental capsaicin model for trigeminal sensitization. syndrome but not healthy volunteers. Gut. 2002;50:471–474. Gender-specific differences. Pain. 2005;118:155–163. 71. Girdler SS, Pedersen CA, Straneva PA, et al. Dysregulation of 55. Nguyen P, Lee SD, Castell DO. Evidence of gender differences cardiovascular and neuroendocrine responses to stress in in esophageal pain threshold. Am J Gastroenterol. 1995;90: premenstrual dysphoric disorder. Psychiatry Res. 1998;81: 901–905. 163–178. 56. Veith JL, Anderson J, Slade SA, et al. Plasma beta-endorphin, 72. Anderberg UM, Marteinsdottir I, Hallman J, et al. Variability in pain thresholds and anxiety levels across the human cyclicity affects pain and other symptoms in female fibromyalgia menstrual cycle. Physiol Behav. 1984;32:31–34. syndrome patients. JMusculoskeletPain. 1998;6:5–22.

r 2013 Lippincott Williams & Wilkins www.clinicalpain.com | 161