For this study we selected 1514 of 2367 patients with AS composing the REGISPONSER database, of which 1131 (74.7%) were men and 383 (25.3%) women. The full details on the methodology and comprehensive compilation of the data included in REGISPONSER1,9 have been published previously.

Each center had a trained rheumatologist who was responsible for patient assessment, data collection and compliance with the inclusion criteria. For our study we considered the following variables: age (years), gender, ethnicity (caucasian/other), first-degree family history (yes, no), smoking status (smoker or former smoker/non-smoker), marital status (married/single), type of work (sedentary, moderate, severe), SA disability and presence of human leukocyte antigen HLA-B27. Likewise, we collected the prior and current medications used by each patient (including the use of methotrexate, sulfasalazine, leflunomide, infliximab and etanercept) and improvement in pain with nonsteroidal antiinflammatory drugs (NSAIDs). Other data collected included the presence of symptoms such as peripheral arthritis, enthesitis, uveitis and coxitis, the presence of hip affection and systemic symptoms (cardiac, renal, neurological or pulmonary involvement, all categorized as yes or no). To measure disability, patients were asked to complete the “Bath Ankylosing Spondylitis Functional Index” (BASFI).10 To measure disease activity, we considered the following quantitative variables: erythrocyte sedimentation rate (ESR), C-reactive protein (CRP) and the self-completion of the “Bath Ankylosing Spondylitis Disease Activity Index” (BASDAI).11 Each patient underwent anteroposterior pelvic X-ray, anteroposterior and lateral lumbar spine and cervical spine X-ray, and were scored according to the “Bath Ankylosing Spondylitis Radiographic Index – Spinal” (BASRI-Spinal),12 which measures structural damage.

We performed a descriptive analysis of the clinical, epidemiological, radiographic and laboratory variables by gender and duration of disease. To do that, variables were stratified into 4 groups (0–9 years, 10–19, 20–39 and 40 years or more of disease progression). Subsequently, the first subgroup was analyzed by creating a new variable (0–4 years, 5–7 and 8–10 years). We calculated the mean and standard deviation for quantitative variables and absolute frequencies and percentages for qualitative variables. A bivariate test was performed using the Student's t test for independent data for quantitative variables and the chi-square test for qualitative variables. The values of “P” were adjusted by the Finner test. Subsequently, we performed a univariate factorial ANOVA with a Sidak adjustment for multiple comparisons to establish the differences in BASDAI, BASFI, BASRI, ESR and CRP by gender and duration of disease.

All contrasts were bilateral and those considered significant had a P<0.05.

Table 1 shows the sociodemographic and clinical differences between men and women with AS of the 1514 patients selected. We found that there were significant differences between both groups in the mean age and the onset of symptoms at inclusion, being lower, in both cases, in men. There were also significant differences in the mean duration of disease, again lower in women. Regarding the existence of a history of AS among first-degree relatives, family forms were more frequent in the female group. For other sociodemographic characteristics there were no differences between the two groups, as occurred in the clinical expression, except for the presence of low back pain, which was more common among men.

Regarding treatment, there was no significant differences between genders in the drugs used previously. At the onset of the disease, the proportion of men using sulfasalazine was greater 87.9% than women (95% CI, 86.26–89.54) vs 83.3% (95% CI, 81.73–84.87), P=0.03. The improvement in pain of those patients taking NSAIDs was significantly more favorable in men.

To assess disease activity we considered BASDAI, ESR and CRP. Compared to men, women had a higher average BASDAI score, regardless of the time of disease progression, with significant differences in all cases, except in those patients whose disease had a mean time interval of 20–39 years. This difference was accentuated in the group of 40 or more years of disease progression, where we observed that both the average score of the BASDAI as well as ESR and CRP levels were significantly higher in the women than in the men: BASDAI (cm) 5.25 (2.06) (95% CI, 4.18–6.32) vs 4.01 (2.39) (95% CI, 3.62–4.40), P=.033; ESR (mm/h) 36.56 (26.37) (95%, 29, 14–43.97) vs 17.85 (17.08) (95% CI, 15.01–20.70), P<.001, and CRP (mg/l) of 19.60 (30.44) (95%, 13.39–25.81) vs 8.42 (10.69) (95% CI, 6.03–10.81), P=.001.

Both the average BASDAI score and CRP levels were significantly increased in the men related to the years of disease progression.

Radiological findings were more severe in men, with a mean score of 6.66 in BASRI – spinal (3.47) (95% CI, 6.45–6.87) vs 4.60 (2.67) (95% CI, 4.33–4.87) in women, P<.001. Adjusting for the duration of the disease, we found that the mean score of BASRI – spinal worsened significantly in both groups over the years, resulting in all cases in worse outcomes in men. In the subgroup of 0–10 years we saw that BASRI – spinal significantly increased over the years in men. When comparing both groups we confirmed that, from the early stages of the disease, the mean BASRI – spinal score was higher in men, except in the group of 5–7 years where there were no differences.

Globally, the BASFI increased significantly with time of disease progression although we cannot say that there were gender differences.

The authors declare that procedures conformed to the ethical standards of the committee responsible for human experimentation and were in accordance with the World Medical Association Declaration of Helsinki.

The authors declare that they have followed the protocols of their workplace regarding the publication of data from patients and all patients included in the study have received sufficient information and gave their written informed consent to participate in this study.

The authors have obtained informed consent from patients and/or subjects referred to in the article. This document is in the possession of the corresponding author.