Women with ankylosing spondylitis: A review


Recent research has highlighted important differences between men and women in a wide range of areas of medicine, including heart disease, human immunodeficiency virus, depression, responses to pain and pain therapies, and drug pharmacodynamics (1). The influences of sex and gender are proposed to play a role in mediating these observed differences. Furthermore, an emphasis has been placed on the clarification of the terms sex and gender when describing the differences between men and women (2). In a report from the Institute of Medicine, sex is defined as “the classification of living things, generally as male or female according to their reproductive organs and functions assigned by the chromosomal complement,” and gender as “a person's self-representation as male or female, or how that person is responded to by social institutions on the basis of the individual's gender presentation. Gender is shaped by environment and experience” (2). In other words, it is proposed that the term sex should be used to describe differences that are “primarily biological in origin and may be genetic or phenotypic,” and the term gender should be used when “referring to social and cultural influences based on sex” (1). To avoid confusion in the review to follow, when appropriate, the authors have preferentially used either the terms gender or sex from the original references.

This recognition and emphasis of the broad differences between men and women have made it important to consider a recent expert opinion “that not only may physicians need to make diagnostic and treatment decisions based on the sex of the patient, but they will also need to respond to gender differences in how women and men approach their physicians, their own health, and how they communicate their health concerns” (1). Although the general idea that men and women may have different manifestations and sequelae of the same disease may be intuitive, additional knowledge of the specific details in disease differences can assist in diagnostic and treatment issues and also provide hypotheses to test in the research arena.

Ankylosing spondylitis (AS) is one such disease that has demonstrated distinct prevalence and clinical differences between men and women. With the recent treatment advances and the hope of halting the progression of the disease now a possibility, early recognition of the disease with its dependency on understanding the differences between men and women with AS will become increasingly important in the near future.

The pathogenesis of AS in terms of both susceptibility and severity may be affected by both sex and gender factors. However, the exact pathogenesis of AS is still unknown and delineating the roles that either sex or gender factors play in the differences between men and women with AS is still problematic. The differences could be due to sex (i.e., genetic, hormonal, other phenotypic differences) or gender (i.e., society- or culture-related differences in physical activity, delay in diagnosis, environmental influences, infections, smoking, etc.) or a combination of both. Further research on the pathogenesis of AS overall and the cause of the differences between men and women in particular will help to define the roles of sex and gender in patients with AS. In this review, we seek to explore and highlight the research in AS that has been done to date on the differences between men and women with AS.

Prevalence of female AS patients

A bias that AS affects men almost exclusively has existed even through the first half of the twentieth century. Dr. Pierre Marie wrote in his classic description of the disease, “All the cases I have been able to review have been men—is this simply a coincidence?” (3). More recent clinical observations and studies have shown that Dr. Marie's observation may have been just a coincidence, because women clearly do have AS and a large proportion of patients with AS are women.

Several studies since the 1940s have examined the differences in disease prevalence between the genders and have shown that the prevalence among women is higher than initially suspected, although the ratio still favors men (4). The earliest studies from the 1940s estimated the prevalence between the genders to be approximately a 9–10:1 ratio of males to females (5, 6). However, these numbers were clearly affected by selection bias, and likely overestimated the ratio in favor of men. Estimates since the 1970s, following the discovery of HLA–B27 and the development of diagnostic criteria (7–9), have placed the ratio lower, at a 2–3:1 ratio of males to females (4, 10–16). Therefore, men are affected more often by AS but not as overwhelmingly as originally thought. This was best summarized in the study by Hill et al from 1976 that concluded, “The precise ratio is less important than the fact that AS does occur in women and should be included in the differential diagnosis of back pain” (4).

Delay in diagnosis

In the study by Hill et al from 1976, the time delay to diagnosis in their population was estimated at ∼10 years for women versus 3 years for men (4). In a later observation by Calin et al from 1988, the median delay to diagnosis in men was between 5 and 7 years and was between 9 and 14 years in women, until they became nearly equivalent beginning in the mid-1960s (11). This change in the disparity suggests that a timely diagnosis of AS in women from that population improved over time, but overall a significant delay was still present. A more recent investigation by Stone et al comparing juvenile-onset AS with adult-onset AS noted that the delay in diagnosis was greater in juvenile-onset patients (15.3 versus 7.6 years) and that women experienced the longest delays in diagnosis (17). Certainly, patient-selection methods and diagnostic criteria can affect these numbers, but the point appears to be that AS presenting in women is often overlooked for a longer period of time than in men. AS is still regarded as underdiagnosed overall in women (18–21). The delay to diagnosis in women may now be more equivalent to the delay in men, but it is likely that women are still underdiagnosed. The reasons for this underdiagnosis could include a continuing bias of AS being a disease of only men or differences in the disease expression in women that could cause delayed or missed diagnoses.

Clinical differences

Several studies in the literature since the 1950s have compared clinical features between men and women with AS (10, 12, 22–34). Consistent differences that stand out are that women with AS tend to have more cervical spine and peripheral joint pain than men with AS. Some studies evaluated hip and shoulder pain separately from other peripheral joints, and others considered them all together. The most recent study to compare men and women with AS is a cross-sectional study of patients with longstanding AS from the Prospective Study of Outcomes in Ankylosing Spondylitis (PSOAS) cohort (34). In that study, women were more likely than men to have been treated with methotrexate, sulfasalazine, and prednisone. Because these medications are predominately used in the management of the peripheral arthritis of patients with AS, it is suggested that women have more peripheral arthritis. Alternatively, this finding could mean that women with AS are initially being misdiagnosed and treated for seronegative rheumatoid arthritis instead of AS. This same cohort was examined with respect to comparing juvenile- and adult-onset AS differences and similarities. Compared with adult-onset AS, patients with juvenile-onset AS had less severe radiographic axial involvement, similar functional outcomes, more hip disease with a greater need for total hip arthroplasty, and a slightly higher proportion of women (35). However, clinically manifest vertebral fractures appeared to occur more frequently when peripheral joint involvement was present; this finding could not be explained by a gender difference (36).

Radiographic differences

Several studies have also looked at the radiographic differences between men and women with AS (12, 22, 23, 25–28, 31–34, 37–39). These investigations have consistently shown that women tend to have less thoracic and lumbar spinal radiographic severity and in more recent studies, lower Bath Ankylosing Spondylitis Radiology Index (BASRI) scores. The studies that looked specifically at cervical spine and peripheral joint (including hip and shoulder) findings, however, did not show evidence of women having more radiographic damage in these areas. Therefore, studies showing that women have more pain in the cervical spine and peripheral joints did not necessarily correlate with the radiographic changes in those areas. The most recent data, again from the PSOAS cohort, revealed that men with AS have greater spinal radiographic damage as measured by the BASRI (34). In terms of overall radiographic severity, it appears that women with AS tend to have less radiographic spinal damage compared with men with AS.

Differences in functional outcomes

Data from our group (34) and others (40, 41) have failed to show any gender differences in functional impairment in patients with AS. However, the data from the PSOAS cohort revealed some finer nuances. Although radiographic damage did correlate with functional impairment in both men and women, women reported more functional limitations at the same level of radiographic damage.

When looking at patients with juvenile-onset AS, investigators found that women with juvenile-onset AS develop functional impairment earlier than men with juvenile-onset AS, as measured by the Bath Ankylosing Spondylitis Functional Index (17). In evaluating work disability, Ward and Kuzis noted that although there is no difference between men and women with AS in terms of permanent work disability, women with AS were more likely than men to receive disability payments for AS, change their type of work, or report the need for help at work (42).

These studies demonstrate that women with AS are at an equal or greater risk for developing functional limitation and work disability as men, even though their radiographic damage may not be as severe. The findings that women with juvenile-onset AS tend to develop functional impairment earlier and that women with AS tend to require more help at work or changes in the type of work point out the subtle differences in the way AS affects men and women in terms of functional and work-related disability.

Differences in family history

There have been several studies evaluating families with AS revealing gender-specific differences. In 1999, Calin et al noted that the prevalence of AS was higher among the children and siblings of female index cases than male index cases (43). Also, AS prevalence was higher among the sons and brothers of male patients than the daughters and sisters of male patients. There was no difference in sex distribution among affected children or siblings of female AS patients. In a French spondylarthritis cohort, a similar preferential transmission of spondylarthritis from father to son as opposed to the more balanced transmission from mother to children of both sexes was observed (44). When these investigators examined their AS patients separately, they found that the transmission from mothers to sons was nearly twice as high as the transmission from mothers to daughters (45% versus 27%). These findings support the idea of the genetic load effect, in which it is proposed that women require a higher genetic load to develop AS than men do (43). Therefore, the environmental trigger for AS might be more sensitive for a male as opposed to a female who might require more genetic material for that genetic trigger to exert its effect. Based on this hypothesis, a woman with AS would also have more of a genetic load to potentially pass on to her children. However, what exactly this genetic load consists of is still unknown.

In a more recent study, Brophy et al observed that affected offspring of maternal AS cases had more second-degree relatives with AS than did children of men with AS (21). Daughters with AS from parents with AS had greater disease activity and worse functioning than did sons with AS from affected parents. However, children from a mother with AS were comparable with those from a father with AS in terms of disease activity, functioning, and radiographic evidence of disease. Jimenez-Balderas et al observed that maternal family history is associated with earlier-onset AS (45). The most recent data from the PSOAS cohort revealed that women with AS have significantly more first-degree relatives with AS than men with AS (34). These studies also lend support to the genetic load theory and suggest that women need to have more genetic burden than men do in order to express the disease, and also may have more of the genetic load to potentially pass on to their children.

These family history studies suggest real differences in the way that men and women inherit and pass on AS. The relationship between potential environmental triggers and disease expression in men versus women gives rise to speculation as to why this occurs. However, the biologic plausibility and specific pathogenic mechanisms for these effects are still missing from our current understanding.

Effects of pregnancy on AS

The 1989 review of AS and pregnancy by Ostensen and Husby concluded that “pregnancy in general does not improve the symptoms of AS,” and “most AS patients will experience a temporary flare during the first 6 months after delivery, lasting for approximately 2 to 4 months” (46). Several larger case series that have examined pregnancy and AS support these ideas (47–50). In these series, there was no change during pregnancy in disease activity in 31–55% of patients, improvement in 20–38%, and worsening in 24–62%. Postpartum flares of disease activity were seen in 45–92% of patients within 6 weeks to 6 months after delivery.

A more recent study by Ostensen et al investigated cytokine and regulatory molecule levels during and after pregnancy in patients with rheumatic disease, including AS, and found that soluble tumor necrosis factor receptor (sTNFR) levels are higher in pregnant patients compared with nonpregnant patients and that levels of sTNFR drop between the third trimester and 6 weeks postpartum (50). This postpartum decrease could explain the flare of disease activity commonly seen after delivery (50).

These studies of pregnancy and AS show us that there may be changes in disease activity during and after pregnancy, most notably a tendency to flare postpartum. However, the changes may not necessarily be due to the unique hormonal milieu of pregnancy, but may be related to cytokine regulation changes reflecting an as yet unknown link to the underlying disease.

Exogenous hormones in AS

Only 1 study, uncontrolled and unblinded, has addressed the issue of how oral contraceptive (OC) pills or hormone replacement therapy affects women with AS (51). Twelve premenopausal and 5 menopausal female AS patients were followed for 1–8 months; 3 of the 10 premenopausal women who were not already taking OC pills began treatment with the OC pills, and 4 of the 5 postmenopausal women began hormone replacement therapy. After 1 month, peripheral arthritis subsided in all patients and did not recur while receiving hormonal therapy. Chest expansion improved significantly, and the Schober test and lumbar pain improved as well but without statistical significance. The functional class improved in most of the treated patients. The erythrocyte sedimentation rate was unchanged. Moreover, 4 of the treated patients had a disease relapse when hormonal therapy was discontinued with an increase in lumbar pain, recurrence of peripheral arthritis, and worsening of the Schober test and fingers-to-floor distance. Although this study had a small number of patients and the observations were uncontrolled, the results are provocative and suggest that there may be some role for hormones to play in modulating the natural history of the disease or its treatment.

Endogenous hormones in AS

Several studies have examined relative sex hormone levels in men and women with and without AS, and these studies have recently been reviewed and summarized by Gooren et al (52). Slightly higher testosterone levels were reported in male AS patients compared with male controls, but the levels remained within the normal range with no statistically significant difference. The studies that examined serum 17-β estradiol levels in male AS patients versus controls showed no significant differences. Similarly, serum levels of testosterone in female AS patients did not differ from those of female controls, and serum levels of 17-β estradiol were either no different or nonsignificantly lower in female AS patients as compared with controls. These studies failed to show sex hormone differences between those with and without AS within each sex, but they do not address whether the different hormonal milieus between the sexes account for some of the differences observed between men and women with AS.

Environmental effects on AS

Although environmental effects, especially infectious, have been studied for their relationship to the pathogenesis of AS, no environmental factors have been implicated that point to gender differences at the present time. One environmental effect that could possibly explain gender differences is cigarette smoking. It is unknown if AS patients smoke more than the general population (53). However, in the PSOAS cohort it was found that men with AS were more often smokers than women with AS (34). In the study by Averns et al, smokers had statistically significant differences in outcomes of fingers-to-floor distance, Schober test, total spine movement, occiput-to-wall distance, functional index, stiffness, and spine radiograph scores (54). The study by Ward and Kuzis found that smokers were more likely to have permanent work disability and receive disability payments (42). Ward also reported that smoking was associated with the rate of progression of functional disability independent of age, gender, education level, comorbidity, and exercise frequency (40). Doran et al noted that smoking was associated with a worse functional score (39). Recently, Peters et al observed that AS is associated with increased cardiovascular mortality and morbidity (53). They observed that smokers with AS have worse outcomes and men with AS may more often be smokers than women with AS.

Although the evidence is still lacking in terms of the association of smoking with susceptibility for AS, there is evidence that men with AS smoke more often than women with AS. Whether the difference in smoking is linked to the clinical or radiographic dissimilarities between men and women with AS or not remains unproven. However, the association of AS with increased cardiovascular disease combined with the finding of worse functional outcomes in patients with AS who smoke suggests at the very least that smoking cessation should be heavily emphasized to all AS patients.

The genetics of AS

The major known genetic factor for disease susceptibility in AS is HLA–B27, and its prevalence among women with AS is equivalent to men (4). The X chromosome was examined in 2 studies and no linkage of the X chromosome with susceptibility to AS was found (55, 56). In rodents, 2 male-specific (HY) peptides that are controlled by the Y chromosome and bind to HLA–B27 were found to be antigenic, and this finding suggests a possible role for the Y chromosome (57). Human HY antigens have also been discovered (58), but the evidence of any connection to the pathogenesis of AS is lacking.

Although the role of the ANKH gene in AS is controversial, a direct genetic difference in the ANKH gene between men and women with AS has been recently described by Tsui et al (59). In a prior study, Tsui et al found that polymorphisms in the ANKH gene were significantly associated with AS (60), but in another later study Timms et al observed no linkage of ANKH with AS (61). The discrepancy may possibly be due to technique differences, population differences, or power insufficiency (59). As a followup to their prior study, Tsui et al examined 201 Caucasian families with AS and found differences between men and women with AS in the ANKH gene. These investigators noted that 2 intronic markers at the 3′ end of the ANKH gene were significantly associated with AS only in male patients and 2 intronic markers at the 5′ end of the ANKH gene were significantly associated with AS in affected female patients. The investigators concluded that ANKH plays a role in genetic susceptibility to AS, displays gender differences, and “may partly account for the gender differences in the prevalence of AS” (59). Further genetic studies may reveal more genes that are involved in the prevalence and severity of AS.


Women with AS appear to have both similarities and differences in disease expression when compared with men with AS. The causes of the differences are still unknown, although both sex and gender factors may explain some of them. The outcome of the disease in men versus women has only recently been addressed and there is little information available right now to answer this important question. Investigators are just beginning to evaluate, with more precise tools, the potential risk factors (such as environmental or genetic) associated with differences in sex or gender and how these factors may explain variability in outcome as well as in disease expression. For example, gender differences in environmental trigger exposure may have an important influence on disease onset, thus setting the stage for a disease prevention strategy. Furthermore, could differences in how either sex generates intermediate steps in immune pathogenesis or immune recognition lead to the discovery of key biomarkers that are sex specific in preclinical or early clinical disease? Hopefully these hypotheses can be explored and discoveries can be made to allow us to improve our diagnostics, therapeutics, and prognostic information with the ultimate goal of prevention.


Dr. Weisman had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Study design. Lee, Reveille, Weisman.

Acquisition of data. Lee, Reveille, Weisman.

Analysis and interpretation of data. Reveille, Weisman.

Manuscript preparation. Lee, Reveille, Weisman.