PMDD Spotlight: Women With Histories of Abuse: A Clinically Meaningful Subgroup?

Susan S. Girdler, PhD


November 26, 2007


Sexual and physical abuse rates for women in the United States are sobering. Population-based national surveys indicate that 13%-27% of women experience childhood sexual abuse.[1] When adult sexual abuse, battering, and other forms of physical abuse are included, more than one third of women from the general population have had these traumatic experiences.[2]

Although the links between histories of abuse and major psychiatric illness as well as medical illness have been well established, an association of abuse histories with premenstrual distress, particularly premenstrual dysphoria, has now also been documented in several community-based samples.[3,4,5] More compelling evidence for a link between abuse and premenstrual dysphoria comes from the results of a recent study showing that a history of trauma, or posttraumatic stress disorder (PTSD) may be a risk factor for the development of premenstrual dysphoric disorder (PMDD).

In a large community sample of women aged 18-24 years followed for 48 months, Perkonigg and colleagues[6]reported that among women who were free of PMDD at baseline, those with a trauma history (including sexual abuse, physical abuse, and severe accidents) were 4 times more likely to develop PMDD at follow-up than were nontraumatized women. The strongest association of trauma with the development of PMDD was seen for childhood sexual abuse, with an odds ratio = 6.7. While this study is limited by the use of retrospective symptom reports to establish the PMDD diagnosis, the results are provocative and reinforce conceptualization of PMDD as a stress-related disorder.

Animal models provide ample evidence that severe or repeated stress exposure early in life leads to long-term changes in neurobiology, as reflected in alterations in stress-responsive systems that persist into adulthood. Our laboratory has had the opportunity to examine the association of abuse histories with alterations in stress-responsive neuroendocrine axes in women with PMDD. In our studies all women: (1) meet strict prospective DSM-IV criteria for PMDD; (2) are assessed for histories of sexual and physical abuse using a validated structured interview;[7] (3) are medically healthy and not taking any medications including psychotropic, thyroid, or hormonal agents; and (3) are free of current Axis I disorders, including PTSD, as determined by structured interview, although a history of lifetime Axis I disorders is permissible with the stipulation that Axis I disorders are in full remission.

Thyroid axis function. An early study of ours[8] examining thyroid axis function in PMDD documented what had been fairly consistently reported by others -- that, as a group, women with PMDD do not differ from non-PMDD controls in mean thyroid hormone concentrations, but do exhibit greater variability in thyroid axis hormones. One possible explanation for these consistent findings was that for a subgroup of women with PMDD, dysregulation in the hypothalamic-pituitary-thyroid (HPT) axis is clinically relevant.

Given the evidence that the HPT axis is stress-responsive, especially to traumatic stress, we set out to examine the association of sexual abuse (SA) with HPT-axis measures. In 28 women with PMDD compared with 25 non-PMDD control women, we found that significantly more PMDD women reported a history of SA compared with control women (10/28 or 36% vs 3/25 or 12%). Thus, we compared 3 groups for HPT-axis measures: PMDD women with SA (n = 10), PMDD with no SA (n = 18), and non-PMDD controls with no SA (n = 22).

We found that the greater variability in HPT-axis measures, particularly thyroid-stimulating hormone (TSH), which had been previously documented in PMDD, exclusively resided in the PMDD subgroup with SA. In addition to the greater variance in TSH, the PMDD/SA group showed a highly unusual HPT-axis profile with significantly elevated total tri-iodothyronine (TT3) and thyroxine-binding globulin (TBG) concentrations; TSH levels were 75% greater than the other groups (although not statistically different due to their greater variance in TSH). No group differences existed in free T3 (FT3), free or total thyroxine (FT4, TT4) or in reverse T3.

When we examined the ratios of TT3 and FT3 to FT4, we found that PMDD women with SA had significantly greater ratios than the other 2 groups, a pattern consistent with increased conversion of T4 to T3, at least in periphery. Whatever the mechanism, there appeared to be something systematically different in the PMDD women with SA shifting from a T4 economy toward a T3 economy.

The behavioral significance of these findings was suggested by the results of our multiple regression analyses indicating that for PMDD women as a whole, greater TT3 concentrations independently predicted greater premenstrual depression, anxiety, and anger ratings.

The case for the HPT axis findings in our work having behavioral significance is further bolstered by the remarkable similarity of our results in PMDD women with SA histories those reported by Mason and colleagues in men with combat-related PTSD[9,10] and those in women with childhood SA-related PTSD,[11] data that also showed markedly elevated TT3 and elevated TT3/FT4 and FT3/FT4 ratios consistent with increased conversion, elevated TBG, and correlations of TT3 to PTSD symptoms.

It is remarkable that PMDD women with SA histories and men with combat-related PTSD show highly similar and unusual changes in the HPT-axis because our PMDD women did not have current PTSD, nor differed from the nonabused PMDD women with PTSD-like symptoms as assessed with the Minnesota Multiphasic Personality Inventory. Although this study provided some of our first evidence for long-term persistence in biologic disturbance following trauma in PMDD, it was limited by too few non-PMDD controls with SA to determine whether the HPT-axis disturbance associated with SA is unique to women in whom PMDD develops.

Sympathetic nervous system and adrenergic receptor function. While histories of abuse are associated with long-term sequelae for all women, subsequent research of ours suggests that the biologic correlates of prior abuse may be uniquely different for PMDD women relative to non-PMDD women. In 2 separate, small PMDD cohorts (N = 25 and 28, respectively), and consistent with community-based reports, we found that the PMDD women were more likely to have a history of sexual abuse or physical abuse (about 60% averaged across the 2 cohorts) compared with non-PMDD controls (about 30% averaged across the 2 cohorts).

In our studies on biologic correlates of abuse, plasma neuroendocrine (using indwelling venous catheter) and cardiovascular measures were assessed following an extended baseline rest period and also in response to a standardized mental stressor battery. We have shown that both PMDD and non-PMDD women with histories of sexual or physical abuse have lower circulating plasma norepinephrine (NE) coupled with greater heart rates (HR) at rest and during mental stress.[12,13] Additionally, there was a tendency for histories of abuse to be associated with lower plasma cortisol levels at rest and following stress.

Despite these similarities between abused PMDD and abused non-PMDD women, we also obtained evidence for diagnosis-related differences in the biologic profiles associated with abuse histories. For example, only PMDD women with abuse histories had significantly greater systolic and diastolic blood pressure and vascular resistance levels, both at rest and in response to stress relative to never abused PMDD women, while abuse histories did not influence BP or vascular responses in non-PMDD women.

We have also examined underlying adrenergic receptor (AR) mechanisms that may be associated with the unique cardiovascular profiles seen in PMDD women with abuse histories. Using the standardized isoproterenol (nonselective beta-adrenergic receptor antagonist) sensitivity test[14] we assessed myocardial beta-1 and vascular beta-2 AR sensitivity in 28 PMDD and 28 non-PMDD controls. We found that only for PMDD women were histories of abuse associated with alterations in AR function since only PMDD women with abuse had greater beta AR responsivity than never abused PMDD women, while non-PMDD women did not differ in beta AR responsivity as a function of abuse history. We interpreted this pattern of results to reflect differential up-regulation in beta AR function in PMDD women with abuse histories resulting from low circulating agonist concentrations (ie, low plasma norepinephrine).

A pathophysiologic role for enhanced beta AR function in PMDD comes from several lines of evidence: (1) a study by Gurguis and colleagues[15] using neutrophil membrane beta-2 AR models, showed greater beta-2 AR density in PMDD women relative to non-PMDD controls, and that luteal phase beta-2 AR density predicted luteal phase anxiety in PMDD; (2) a 3-month, placebo-controlled study[16] comparing fluoxetine with propranolol (nonselective beta AR antagonist) to treat symptoms in PMDD found that while fluoxetine was superior to propranolol in reducing overall mean symptom severity, propranolol was more efficacious for certain symptoms, including depression, headache, crying, heart pounding, fatigue, dizziness, forgetfulness, nausea, and cravings; and (3) in small studies,[17,18,19] propranolol has been shown to be effective in the treatment and prevention of PTSD, a condition that may be pathogenically related to the development of PMDD.[20]

Enhanced beta AR function in PMDD women with histories of abuse may also have implications for somatic symptoms in PMDD. Although significant emotional distress is a required feature of PMDD based on DSM criteria, and emotional symptoms have the greater impact on function for most PMDD women, premenstrual pain represents an important feature of this disorder and contributes to the overall dysfunction associated with PMDD.[21]

There is overwhelming evidence for a link between abuse histories and clinical pain, and animal models indicate an important role of beta ARs in mediating increased pain perception.[22,23] Our own work as well as work by others has documented hyperalgesia to experimental noxious stimuli in PMDD women relative to non-PMDD controls,[24] although pain sensitivity was not examined as a function of abuse history in that prior research. Nevertheless, we have recently documented that PMDD women with histories of abuse have greater headache severity ratings than never-abused PMDD women[13] and unpublished observations from our laboratory also suggest that they have greater premenstrual backache and cramping than never-abused PMDD women.

Finally, we have preliminary evidence from a randomized, placebo-controlled, cross-over study that used a long-term clonidine challenge test (alpha-2 adrenergic receptor agonist) to investigate alpha-2- AR mechanisms as a function of sexual abuse histories in 12 PMDD women (5 with SA histories).[25] We reported that clonidine was associated with significantly greater reductions in resting and stress-induced BP and HR in the PMDD women with SA compared with PMDD women with no SA, suggesting that abuse experiences may modify presynpatic alpha-2- AR function in PMDD.

This interpretation is limited by the small sample size and also by the absence of a non-PMDD control group. Nevertheless, potential pathophysiologic relevance of our findings comes from studies relating alpha-2 AR function directly to symptoms in PMDD.[15,26] Moreover, our findings described above for greater BP and vascular resistance levels in abused PMDD women that were not seen in the abused non-PMDD group, may well reflect vascular alpha AR mechanisms since pharmacologic blockade studies have shown that individual differences in vascular resistance at rest and during mental stress are mediated, at least in part, by vascular alpha ARs.[27]

The persistent dysregulation in HPT-axis and adrenergic physiology seen later in life in PMDD women with abuse histories is a remarkable feature of our research findings. Although our studies have included women with abuse at any age, approximately 80% of our samples have consisted of women who experienced childhood or adolescent abuse. Thus, many years, if not decades, had elapsed since the majority of the women in our studies had experienced abuse. This is consistent with the work of Wang and Mason[28] showing persistent disturbance in HPT-axis function in male combat veterans 40-50 years following combat and consistent with the work of Heim and colleagues[29] showing altered hypothalamic-pituitary-adrenal axis responses to stress in adult women with childhood sexual abuse.

It has been suggested that a genetic vulnerability coupled with early life or severe stress may contribute to persistent alterations in neurobiologic systems that are known to be stress responsive, resulting in dysregulation in stress responsiveness to even mild stressors in adulthood and forming the basis for the development of mood disorders.[29] Consistent with the study of Perkonigg and colleagues[6] that showed childhood sexual abuse to be a powerful predictor of the development of PMDD, the possibility exists that given a genetic predisposition for PMDD,[30] then persistent disturbance in stress responsive systems (eg, HPT-axis and adrenergic physiology) following traumatic stress may form the underlying neurobiologic basis for the development of PMDD.

Regardless of mechanisms contributing to the development of PMDD, the results of our research on abuse histories support more recent conceptualizations that PMDD represents a heterogeneous disorder with respect to underlying pathophysiology.[31] While early life exposure to sexual or physical abuse appears to lead to persistent and long-term dysregulation in stress-responsive systems for all women, even in the absence of current Axis I illness, PMDD women may be more vulnerable to the effects of abuse on adrenergic physiology and possibly HPT-axis function.

The possibility exists that PMDD women with histories of abuse, who also exhibit more severe emotional and physical symptom ratings than never-abused PMDD women,[13] represent a clinically distinct subgroup with respect not only to causal factors, but also with respect to therapeutic approach. The fact that data have shown a 40% nonresponse rate to selective serotonin reuptake inhibitors in PMDD[21,32,33] supports this contention, as does the placebo-controlled study described early that demonstrated superior efficacy of propranolol to treat specified PMDD symptoms.[16] Indeed, expert guidelines on the diagnosis and treatment of PMDD urge not only a careful medical history, including thyroid hormone profiles, but also the assessment of prior and current abuse[34,35,36] before initiating treatment for PMDD. Our research suggests that is prudent advice.

Supported by an independent educational grant from Bayer


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.