| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Division of Public Health Services and Research, Department of Operative Dentistry (C.D.M., J.L.R.), Department of Clinical and Health Psychology (M.E.R.), Division of Cardiovascular Medicine, Department of Medicine (D.S.), University of Florida, Gainesville, Florida.
Address reprint requests to: Cynthia D. Myers, PhD, Division of Public Health Services and Research, P.O. Box 100404 UFHSC, Gainesville, FL 32610-0404. Email: cynthm{at}ufl.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODRESULTSDISCUSSIONCONCLUSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
METHODS: Fifty female and 54 male healthy, young adults completed the Bem Sex Role Inventory and then underwent a cold pressor task. Blood pressure was assessed before and during pain testing.
RESULTS: Univariate analyses indicated significant sex-related differences in pain threshold and pain tolerance. Baseline SBP was positively related to pain tolerance but did not explain sex differences, in accord with previous research. The Bem Sex Role Inventory demonstrated a relationship with pain, but did not explain sex differences.
CONCLUSIONS: We suggest that context-specific measures of gender are needed to assess gender-related pain behaviors in specific situations. Results from the current study support our contention that gender is part of sex as commonly measured. Also, blood pressure does not appear to fully account for sex-related differences in pain.
Key Words: Sex differences • gender role socialization • experimental pain • blood pressure • pain threshold • pain tolerance
Abbreviations: SBP = systolic blood pressure; DBP = diastolic blood pressure; HR = heart rate; BSRI = Bem Sex Role Inventory; FEM = femininity score; MASC = masculinity score.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONCONCLUSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Sex-related differences in blood pressure are emerging as one potential biological explanation of sex-related differences in pain. Many studies report a continuous, inverse relationship between resting blood pressure and pain sensitivity (6–11), and women generally have lower resting blood pressure than men. Furthermore, experimental studies have indicated that sex-related differences in thermal pain report were no longer statistically significant after adjusting for resting blood pressure (7, 12). However, results of experimental studies are mixed with regard to the sex-specific relationship between blood pressure and pain, with one study (7) finding resting blood pressure to be inversely related to ischemic and thermal pain sensitivity in men but not women, while other studies found resting blood pressure to be inversely related to pain sensitivity in women but not men using thermal (8) and electrical (13) stimuli.
Blood pressure reactivity may explain sex-related differences in experimental pain report. Women and men differ in their blood pressure responses to acute stress and pain, with men generally showing greater blood pressure increases (14–16). Although a number of studies have examined the relationship between blood pressure reactivity and pain sensitivity, with most finding a negative association (6, 16–18), no study has examined whether blood pressure reactivity explains sex differences in pain report.
Females and males have different learning histories regarding pain behavior (19), potentially contributing to sex-related differences in experimental pain report through gender-role socialization. Consistent with gender stereotypes, boys reported less fear of pain and less attention paid to pain (20), and used fewer affective descriptors of pain (21), relative to girls. In an analogue study of imagined pain (22), men indicated they would respond to pain with embarrassment and reluctance to disclose it, whereas women would respond to pain with anxiety and a higher likelihood of disclosing it. A significant effect for sex of experimenter was found in one experimental pain study deliberately accentuating gender-relevant cues in the experimenters’ appearance (23), but not in studies without deliberate accentuation of gender cues (24, 25). Collectively, these results suggest that gender role socialization influences pain behavior, but they do not involve direct measurement of the gender role construct.
The Bem Sex Role Inventory (26) is an established measure of the gender role construct operationalized as stereotypically feminine and masculine global personality traits. According to Bem’s gender schema theory (27, 28), the BSRI permits identification of individuals who are particularly motivated across situations to evaluate their personal adequacy by how well their behavior conforms to societal definitions of masculinity and femininity (29). The BSRI has been used as a measure of the gender construct in two published experimental pain studies (25, 30).
Using the BSRI, Otto and Dougher (25) found a significant interaction between gender score and sex on mechanical pressure pain threshold. High masculine males reported higher pain thresholds than low masculine males, whereas gender score was not related to pain threshold in women. By contrast, Fillingim, Edwards, and Powell (30) reported no relationship between gender and experimental pain report, although femininity did correlate with self-report of clinical pain experienced by subjects during the month before the experimental pain session.
The current study was an initial examination of the relative contribution of both psychosocial and physiological measures to sex-related differences in experimental pain report. Specifically, we sought to determine whether the relationship between sex and experimental pain report was explained by either SBP at rest or during pain task, or by gender-role socialization as measured by the BSRI. We also attempted to replicate the results obtained by Otto and Dougher (25) regarding gender-related differences in pain responding, using a different pain stimulus.
|
METHOD |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONCONCLUSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Apparatus
Blood pressure was measured with a Dinamap Critikon automatic blood pressure monitor. Pain was induced in a 2.5 cubic foot cold pressor bath fitted with a screen dividing ice from water and a pump circulating water continuously through the ice, thereby maintaining a constant water temperature of 1 to 3 degrees Celsius.
Measures
Questionnaires.
Participants completed a demographics questionnaire and the BSRI (26). The BSRI consists of 60 adjectives or adjective phrases: 20 feminine, 20 masculine, and 20 neutral, nonscored items, each of which is accompanied by a self-rating scale ranging from 1 = "never or almost never true" to 7 = "always or almost always true." A femininity total score and a masculinity total score was calculated for each participant and converted to a standardized difference score (31). Reliability and validity of the BSRI have been well established (31–33).
Pain.
Pain threshold was measured as the time in seconds to a participant’s report of pain onset during the cold pressor task. Pain tolerance was measured as the total time elapsed while a participant’s hand remained in the cold water.
Blood pressure.
Baseline blood pressure was calculated as the mean of three BP measures during the resting baseline period, which was defined as the period in which three consecutive diastolic blood pressure readings within 10 mm Hg were obtained. The initial BP reading taken after hand immersion served as task BP, and BP reactivity was calculated as task BP minus baseline BP.
Procedure.
Participants were instructed to abstain from caffeine for the two hours before their appointment. In individual sessions, a female experimenter wearing professional garb and a white lab coat administered all measures. After participants completed the questionnaires, the experimenter placed an appropriately sized blood pressure cuff on the participants’ dominant arm with the microphone placed over the brachial artery. Thereafter, the cuff inflated automatically at one-minute intervals. After three consecutive DBP readings within 10 mm Hg had been obtained, the following instructions for the cold pressor task were spoken:
In a moment I will ask you to place your nondominant hand, palm facing down, in the water on this side of the divider (gesture to the ice-free side of the divider) until the water reaches one inch above your wrist. Once your hand is in the water, when you first experience pain, please say the word ‘pain.’ Then keep your hand in the water until you have reached your tolerance for pain. At that time, withdraw your hand from the water and say ‘pain limit.’
Ten seconds before the next blood pressure measurement, participants were signaled to place their nondominant hand in the water. Pain threshold and tolerance were obtained. Participants maintaining hand immersion for five minutes were asked to withdraw the hand from the water. Blood pressure readings were recorded until cardiovascular parameters returned to approximately baseline. The experimenter then removed the blood pressure cuff and debriefed participants.
Analytic strategy.
Zero order Pearson Product Moment Correlations were calculated to examine relationships between continuous variables (SBP, DBP, HR, FEM, MASC, pain threshold, pain tolerance), using a Hochberg (modified Bonferroni) procedure to correct for multiple comparisons to reduce the likelihood of a Type I error (34). Univariate analyses were conducted to determine mean sex differences in the above variables. Effect sizes (Cohen’s) were calculated to determine the magnitude of correlational relationships and sex differences. By Cohen’s convention (35), effect sizes of approximately 0.20 are small, 0.50 are moderate, and 0.80 are large. Three sets of hierarchical regression analyses were conducted. First, sex, gender, and their interaction term product vector were regressed on pain threshold and tolerance, with the standardized FEM-MASC T score (31) serving as the gender variable, to attempt replication of Otto and Dougher (25). A second set of hierarchical regressions assessed relationships between blood pressure, sex, gender, and pain. A third set of hierarchical regressions assessed relationships between heart rate, sex, gender, and pain. A p < .05 was the criterion for statistical significance.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONCONCLUSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
|
|
R2 = .06, F(1,102) = 6.31, p = .01), and there was no significant contribution by the gender by sex interaction term ( R2 = .001, F(1,102) = .15, p = .67). These results are in contrast to Otto and Dougher, who found a main effect of gender, and a gender by sex interaction. In the second regression on threshold, sex was block one, gender was block two, and their interaction term was the final block. R2 for the full model was 0.088, p = .03. Sex-predicted pain threshold (R2 = .09, F(1,102) = 9.55, p < .01); there was no significant contribution by gender ( R2 = .001, F(1,102)= .15, p = .70), or by the interaction term ( R2 = .001, F(1, 102) = .15, p = .70), replicating Otto and Dougher.
In the first regression on tolerance, we entered gender as block one, sex as block two, and gender by sex as the final block. R2 for the full model was 0.20, p < .01. Gender (R2 = 0.07, F (1,102) = 8.18, p < .01) and sex ( R2 = 0.13, F(1,102) = 15.82, p < .01) significantly predicted pain tolerance, and there was no significant contribution by the interaction term ( R2 = 0.00, F(1,102) = 0.03, p = .86), consistent with Otto and Dougher. In the second regression on tolerance, sex was entered as block one, gender as block two, and their interaction term as the final block. R squared for the full model was 0.20, p < .01. Sex-predicted pain tolerance (R2 = 0.20, F (1,102) = 24.69, p < .01), and there was no significant contribution by gender ( R2 = 0.005, F(1,102) = 0.60, p = .44) or by the interaction term ( R2 = .00, F(1,102) = .03, p = .86), replicating Otto and Dougher.
Potential blood pressure and gender effects on pain threshold and tolerance were analyzed in a second set of hierarchical regression analyses. In the first regression on threshold, sex was the first block, baseline SBP was the second block, gender (FEM-MASC T score) was the third block, and a gender by sex interaction term was the fourth block. R2 for the full model was 0.09, p = .04. Sex-predicted threshold (R2 = 0.09, F(1,102) = 9.55, p < .01), beyond which there was no significant contribution by baseline SBP ( R2 = 0.007, F(1,102) = 0.77, p = .38), gender ( R2 = 0.00, F(1,102) = 0.04, p = .83), or the interaction term ( R2 = 0.001, F(1,102) = 0.06, p = .81). In the second regression on threshold, sex was entered as the first block, SBP reactivity as the second block, gender as the third block, and a gender by sex interaction term as the fourth block. R2 for the full model was 0.09, p = .05. Sex-predicted threshold (R2 = 0.09, F(1,102) = 9.55, p < .01), beyond which there was no significant contribution by SBP reactivity ( R2 = 0.09, F(1,102) = 0.12, p = .73), gender ( R2 = 0.08, F(1,102) = 0.17, p = .68), or the gender by sex interaction term ( R2 = 0.09, F(1,102)= 0.15, p = .70). When pain tolerance replaced pain threshold in these two regression models, sex was the only significant predictor of tolerance. R2 for the models was 0.21, p < .01, and 0.20, p < .01, respectively. No gender or blood pressure variable or interaction term made a significant contribution above and beyond that of sex. In the third set of regression analyses, equivalent analyses using heart-rate data yielded equivalent results to those obtained here, with sex being the only significant predictor of pain threshold or tolerance.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONCONCLUSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Similar to previous studies, we found an inverse relationship between SBP at rest and pain sensitivity as indicated by pain tolerance. Mechanisms posited to account for this inverse relationship include central descending pathways, beta-endorphin responses, and baroreceptor sensitivity (9, 11, 36). We observed sex differences in resting SBP, however, the strength of the resting SBP-pain relationship was not different in women and men. Thus, our data differ from two previous studies finding a resting SBP-pain relationship only in women (8, 13), and one study finding a resting SBP-pain relationship only in men (7). We did not control for menstrual phase, which might obscure SBP-pain relationships in women (37, 38), but the same is also true for the previous studies. Because there is no consistent evidence that resting SBP accounts for sex differences in pain report, future studies should evaluate underlying physiologic mechanisms (39).
Neither systolic nor diastolic blood pressure reactivity differed by sex of participant, and these were unrelated to pain measures. One possible explanation for our data concerns our measure of reactivity. Several participants withdrew their hand from the cold water before one minute, necessitating the use of a single data point as the task value. Although the reduced reliability of our measure might account for our results, it is worth noting that the reactivity-pain literature is mixed, with some studies finding a direct relationship between SBP reactivity and pain (6, 16–18), and others finding an inverse relationship (eg, 40, 41). The reactivity-pain relationship requires clarification, with attention to the direction, strength, and mechanisms of any such relationship.
A limitation of the current study concerns the fact that baseline blood pressure was measured during the same session as pain testing. It would have been preferable to measure blood pressure at a separate session to control for potential anticipatory anxiety effects. Additionally, a possible alternative explanation for the finding of sex-related differences in pain response may be sex differences in resting hand temperature or cutaneous, cold-induced vasomotor responses.
Our study did not replicate the main finding of Otto and Dougher (25), who found that high masculine men reported higher pain thresholds than low masculine men, whereas there was no such relationship in women. One possible explanation pertains to the differing pain stimuli in the two studies. Whereas we used a cold pressor bath, Otto and Dougher applied mechanical pressure. A second possibility concerns definitions and measurement of pain. We defined pain threshold as time to self-report of first experiencing pain, and pain tolerance as time to hand withdrawal from the cold pressor. Otto and Dougher used a single, seven-point scale with the following response options: 1 = slight pressure, 2 = moderate pressure, 3 = slight discomfort, 4 = moderate discomfort, 5 = slight pain, 6 = moderate pain, and 7 = pain is no longer tolerable, and chose the time taken to produce responses of "5" and "7" as pain threshold and pain tolerance, respectively. The psychometric properties of Otto and Dougher’s scale are unknown. It is possible that female and male participants responded differently to semantic distinctions among the response options, contributing to the significant interaction effect, which was obtained only on pain threshold. We suggest that future studies examine gender-related differences in the semantics of pain measurement.
Otto and Dougher (25) and the current study used similar measures of pain tolerance, and both studies found that gender independently predicted pain tolerance when entered into a hierarchical regression equation before sex. However, if sex was entered into the equation first, gender no longer predicted pain tolerance in either study. This finding indicates that gender as measured by the BSRI (26) accounts for some of the variance in the sex-pain relationship and, thus, the sex-pain relationship is not only indicative of first-order biology but also reflects acquired differences.
This study did not find the BSRI (26) to make a robust, independent contribution to the observed differences in experimental pain report between women and men, similar to the two previous studies (25, 30). We suggest three possible explanations for this. First, gender-role socialization may be irrelevant to pain report. However, past research suggests that this is unlikely (19–23). Second, the BSRI only assesses one aspect of gender, which may not be robustly related to pain report. Other aspects of gender, for example attitudes, interests, and behaviors, may provide additional explanations for sex-related pain differences. Third, while the BSRI reliably measures global, stereotypically feminine (eg, expressivity) and masculine (eg, instrumentality) traits (42), it may not assess context dependent aspects of gender (43) that are elicited specifically by the experimental pain situation. In support of this contention, Fillingim et al. (30), found that the BSRI was associated with clinical pain report but not experimental pain report. Relatedly, al’Absi and Rokke (44) found that trait anxiety did not predict experimental pain report, but anxiety specific to an experimental pain task did predict pain report. This suggests that trait measures, such as those used for gender role (ie, BSRI) and trait anxiety, might be useful predictors of global pain experiences, such as clinical pain, but poor predictors of specific pain situations, ie, experimental pain. Measures of situation specific gender-related pain behaviors are needed to test for associations in specific pain situations, ie, experimental pain (45).
The relationship of gender to sex is a complex one (46). Results from the current study highlight our contention that gender is part of sex as commonly measured. The dichotomous variable "sex" is often used as a proxy for all presumed biological aspects of being female or male, and statistical analyses employing this variable might obscure the component of sex that comes from social learning. Future work might substitute actual physiological markers of interest (ie, gonadal hormones) for the dichotomous variable of sex as being more reflective of the actual mechanisms of interest, and thereby facilitate distinguishing between the influences of first-order biological mechanisms and social learning mechanisms involved in pain perception.
|
CONCLUSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONCONCLUSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
|
ACKNOWLEDGMENTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONCONCLUSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Dr. Sheffield was funded by Grant 1R01 HL64580 from the National Heart, Lung and Blood Institute, Bethesda, Maryland.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONCONCLUSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
Received for publication February 1, 2000.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODRESULTSDISCUSSIONCONCLUSIONNOTESACKNOWLEDGMENTSREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  M Rueda-Sanchez and L. Diaz-Martinez Prevalence and Associated Factors for Episodic and Chronic Daily Headache in the Colombian Population Cephalalgia, March 1, 2008; 28(3): 216 - 225. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. J. Gedney, T. L. Glover, and R. B. Fillingim Sensory and Affective Pain Discrimination After Inhalation of Essential Oils Psychosom Med, July 1, 2004; 66(4): 599 - 606. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Lowery, R. B. Fillingim, and R. A. Wright Sex Differences and Incentive Effects on Perceptual and Cardiovascular Responses to Cold Pressor Pain Psychosom Med, March 1, 2003; 65(2): 284 - 291. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
