Sex differences have been reported for many pain-related responses. For instance, women are at greater risk for pain disorders such as fibromyalgia (Unruh, 1996). Women also report more widespread pain, more pain-related affective symptoms (Mullersdorf and Soderback, 2000 and Keefe et al., 2000), and more frequent daily pain (Berkley and Holdcroft, 1999, Barsky et al., 2001 and Bassols et al., 2002). Sex differences have also been investigated in laboratory settings, with women demonstrating lower pain thresholds and higher pain ratings across a variety of noxious stimuli (Riley et al., 1998 and Fillingim, 2000). It has been suggested (Fillingim and Maixner, 1995, Fillingim et al., 1999 and Fillingim et al., 1999b) that these findings are linked, with greater pain sensitivity acting as a risk factor for enhanced clinical pain.
Although sex differences are well-documented, explaining these differences is more challenging. While some researchers emphasize socialization (Fearon et al., 1996) or emotional responsiveness (Riley et al., 2001) as potential mechanisms, others have highlighted biological factors (Berkley, 1997) or pain-coping (Unruh, 1996, Fillingim, 2000 and Myers et al., 2003). Indeed, how one copes with pain consistently predicts important clinical outcomes, including pain severity and disability (Turk and Okifuji, 2002). Generally, the most robust predictor of pain outcomes is catastrophizing (Sullivan et al., 2001), defined as a negative emotional and cognitive response to pain involving elements of magnification, helplessness, and pessimism. Catastrophizing is positively correlated with pain and depression, and some laboratory studies show an association with responses to standardized noxious stimuli (Sullivan et al., 2001).
Additionally, women report more frequent catastrophic cognitions (Sullivan et al., 2001), making catastrophizing a potential contributor to sex differences in pain. In a recent osteoarthritis study, women had higher levels of pain, pain behavior, and disability. Moreover, women reported more catastrophizing, which mediated the relationship between sex and pain-related outcomes after controlling for depression (Keefe et al., 2000). A second study reported that during a cold pressor task, females reported more pain and displayed more pain behavior than males, effects which became non-significant when catastrophizing was controlled (Sullivan et al., 2000b).
The literature, however, is inconsistent, with some studies showing no sex differences in catastrophizing (Unruh et al., 1999 and Edwards et al., 2000). Moreover, while pain-related sex differences in the laboratory are often large (Riley et al., 1998 and Fillingim, 2000), sex differences in clinical pain are inconsistent (Turk and Okifuji, 1999). Additionally, catastrophizing as a mediator of sex differences in day-to-day pain among healthy individuals has not been studied. It is important to investigate such questions in non-clinical samples, before sex differences in pain are confounded by additional factors such as sex differences in pain treatment or sex-specific selection biases. Many researchers have also not controlled for depression when evaluating catastrophizing, which is now a standard in the field (Sullivan et al., 2001). Finally, catastrophizing likely depends on contextual factors such as the threat value of pain, which may differ in laboratory settings versus clinical environments. The present investigation, therefore, studied catastrophizing as a mediator of sex differences in both day-to-day pain and experimental pain.
Our results indicate that sex differences in complaints of painful day-to-day symptoms are accounted for by the significant differences between men and women in reports of catastrophizing. However, the sex differences in catastrophizing do not account for the substantially higher threshold and tolerance for thermal and cold pain observed among men.
The present study builds upon prior work by assessing catastrophizing as a mediator of male–female differences in pain responses both inside and outside of the laboratory. Several prior studies have suggested that catastrophizing at least partially mediates observed sex differences in both clinical pain (i.e. arthritis pain) (Keefe et al., 2000) and in experimental pain responses (Sullivan et al., 2000a and Sullivan et al., 2000b). Our results are consistent with those of the former study, and suggest that catastrophizing plays an influential role in shaping sex differences in pain report in both clinical and non-clinical samples. However, it is somewhat more difficult to reconcile our findings with those of Sullivan and colleagues (Sullivan et al., 2000b), who also used a cold pressor task. They assessed pain intensity ratings and a measure of the duration of pain behaviors during the task. Correlations of between 0.33 and 0.53 were observed for the association between catastrophizing scores and pain responses. In contrast, we observed minimal correlations (r=-0.12 for CPTO and catastrophizing). One other recent cold pressor study evaluated threat appraisals (i.e. conceptually similar to catastrophizing) as a potential mediator of sex differences in cold pain tolerance (Sanford et al., 2002). The association between threat appraisals and cold pain tolerance was only marginally significant (r=-0.15); however, controlling for threat appraisals did slightly reduce the significance of sex difference in cold pressor tolerance times. The small magnitude of this association parallels our findings and suggests that catastrophizing's relationship with experimental pain responses may be contingent on the type of response measured, with ratings of pain intensity and observation of pain behaviors (see Keefe et al., 2000), being most subject to influence by cognitive and affective processes.
We should also note that assessment of catastrophizing should include the additional dimensions of rumination and magnification, as measured by the pain catastrophizing scale (PCS) (Sullivan et al., 1995).
Pain recall is subject to many biases (Haythornthwaite and Fauerbach, 2001 and Stone et al., 2004), and it may be that catastrophizing is more powerful in shaping pain recall than in determining in vivo pain responses. Indeed, in the present study, the CSQ was completed shortly after participants answered questions about recent daily pain, and their responses to questions about catastrophizing may have been influenced by those recently-recalled pain experiences.
The process of catastrophizing is closely tied to the meaning of pain (Sullivan et al., 2001); it seems probable that catastrophic interpretations are more likely to accompany daily pain relative to a brief, controllable stimulus administered in a laboratory, where subjects have foreknowledge about the stimuli (Gracely, 1999). In contrast, clinical pain is more unpredictable and threatening. For example, a study of cancer patients found that those who believed that physical therapy-induced pain was cancer-related showed higher ratings of pain intensity and unpleasantness compared to those who attributed their pain to other factors, highlighting the role of meaning and interpretation (Smith et al., 1998). Another investigation noted that pain intensity ratings are strongly affected by interpretations of stimuli as more or less tissue-damaging, a consideration that would rarely apply to most laboratory settings (Arntz and Claassens, 2004). Finally, a study of fibromyalgia patients reported strong correlations between catastrophizing and clinical pain but no associations with responses to noxious mechanical stimuli administered in a laboratory (Gracely et al., 2004).
Sex-related variation in catastrophizing appears to emerge relatively early in development, probably well before most individuals have had any substantive experience with chronic pain. Catastrophizing is more common among adolescent school girls than boys, and was associated with more pain and pain medication use in a survey of high school students (Bedard et al., 1997). Indeed, this sample consists of healthy college-age students with no history of chronic pain. Unfortunately, while it suggests some consequences of sex differences in catastrophizing, the present study can offer little insight into its causes; such information will require longitudinal studies with long follow-up periods in children and adolescents.
While sex differences in laboratory pain responses are generally robust, catastrophizing does not appear to play a role in producing these male–female discrepancies. Laboratory pain responses are related to the brain's response to noxious stimulation (Coghill et al., 2003), to quality of life (Edwards et al., 2003c), and can prospectively predict clinical pain (Granot et al., 2003). Illuminating the mechanisms producing sex differences in experimental pain responses may therefore have important clinical implications. The fact that catastrophizing did mediate sex differences in day-to-day pain supports prior work in osteoarthritis patients. Importantly, catastrophizing mediated this association even after controlling for negative mood, highlighting catastrophizing's unique effects. Catastrophizing is observed in community residents who report no current pain (Buer and Linton, 2002), is fairly stable across time (Sullivan et al., 2001), and is a risk factor for worsening pain (Keefe et al., 1989 and Haythornthwaite et al., 2003). As such, it certainly warrants further attention as a marker for individuals at high risk for the onset or worsening of pain, as a target of treatment, and as an important variable in shaping group differences in the clinical experience of pain.