Female hormonal factors and the development of anti-citrullinated protein antibodies in women at
3. Female hormonal factors on rheumatoid arthritis progression
3.3 The role of reproductive factors in the progression of rheumatoid arthritis .1 Introduction
3.3.2 Submitted article: Role of reproductive and menopausal factors in functional and structural progression of rheumatoid arthritis. Results from the SCQM cohort
behalf the physicians of the Swiss Clinical Quality Management Program for Rheumatoid
15
33
40
47
54
Page 1 of 29 Rheumatology
1 2
3 Original article
4 5 6 7
8 Role of reproductive and menopausal factors in functional and structural progression
9
10 of rheumatoid arthritis. Results from the SCQM cohort.
11 12 13
14 D. Alpizar-Rodriguez 1, * F. Förger 2, D. S. Courvoisier1, C. Gabay 1, A. Finckh 1; On
16 17
18 19 Arthritis3.
20 21 22 23
24 1Division of Rheumatology, Department of Internal Medicine Specialties, University
25
26 Hospitals of Geneva; 2Inselspital, Bern; 3SCQM Foundation, Zurich; Switzerland.
27 28
29 30 Correspondence to:
31 32
Deshiré Alpizar-Rodriguez1, MD
34
35 Division of Rheumatology
36 37 University Hospital of Geneva,
38
39 26, Ave Beau-Sejour
41
42 CH 1211 Geneva, Switzerland
43
44 Phone/fax= +41223723769/ +41223723535
45
46 Deshire.AlpizarRodriguez@hcuge.ch
48
49 50 Content: Word count: 2866/3500, abstract: 244/250, ref: 42/50, figures and tables: 6/6
51 52
53 Key word: rheumatoid arthritis, reproductive, hormones, epidemiology, disability
55
56 evaluation
9
16
32
39
Rheumatology Page 2 of
29
1 2
3 Abstract
4
5 6 Objectives: To study the impact of female reproductive and menopausal factors on
7 8
functional and structural joint-damage progression in women with RA.
10
11 Methods: This is an observational cohort study of RA patients enrolled in the Swiss RA
12
13 registry (SCQM-RA). Information about female hormonal factors, such as pregnancies,
14
15 menopause and hormonal therapy, were retrospectively retrieved using a specific
17
18 questionnaire. The primary outcome was functional disability progression (HAQ) and the
19
20 secondary outcome radiographic joint-damage progression. We compared the functional
21 22 progression between pre- and post-menopausal women using a multilevel regression model
23 24
25 for longitudinal data, adjusting for potential confounders, such as baseline age, disease
26
27 duration, DAS28 and treatment.
28 29 Results: A total of 1667 women were analyzed, of which 1025 (61%) were post-
30 31
menopausal. Participants had a median of 6 HAQ assessments [IQR 3 to 10] during 5.1
33
34 years [IQR 2.2-9.8] of follow-up. At baseline post-menopausal women had higher HAQ
35
36 and erosion scores than pre-menopausal women. The evolution of HAQ scores over time
37
38 differed between pre- and post-menopausal women (p<0.001), with a less favorable
40
41 evolution in post-menopausal women, particularly with earlier age at menopause. Among
42
43 pre-menopausal women, being in the active parous period and having pregnancies before
44 45 RA diagnosis was associated with a more favorable evolution of function. Erosion
46 47
48 progression did not differ between pre- and post-menopausal women.
49
50 Conclusions: In women with RA, functional disability progression differed between pre-
51 52 and post-menopausal women. The more favorable evolution of function in pre-menopausal
53
54 women was not explained by disease duration, age or radiographic damage.
15
24
Page 3 of 29 Rheumatology
1 2
3 Rheumatology key messages
4 5 6 7
8 • Functional disability progresses more rapidly in post-menopausal than in pre-
9
10 menopausal women with RA
11 12 13
14 • RA patients with earlier age at menopause have a worse progression of functional
16
17 disability
18 19 20
21 • The more favorable evolution of function in pre-menopausal women is not explained
22
23 by shorter disease duration, younger age or less radiographic damage
25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51
8
15
22
31
38
45
Rheumatology Page 4 of
29
1 2
3 Introduction
4 5 Rheumatoid arthritis (RA) is more prevalent in women, with a female to male ratio up to 4
6 7
to 1 in younger age groups, but near to 1:1 after 60 years of age [1-4]. RA has a major
9
10 impact on women’s health, with an increased risk of osteoporosis, of cardiovascular
11
12 diseases, of malignancies and of overall mortality [1, 3]. In addition, RA disease activity,
13
14 severity and treatment response appear to be worse in women than in men, with women
16
17 having 3 times more rapid progression of the RA related disability than men [5-8]. Disease
18
19 severity in women may be influenced by socio-economic factors, genetic factors,
20
21 environmental exposures, and treatment-related effects [9-11].
23 24 25
26 The role of reproductive factors, such as parity, menopause and hormonal therapy on the
27 28 severity of disease are still controversial. Pregnancy may be associated with a lower disease
29 30
activity in half of women, but the postpartum period is associated with an increase in
32
33 activity in up to 90% of cases [12-14]. First symptoms of the disease have been associated
34
35 with menopause, and an increased risk of RA with earlier age at menopause [4, 15, 16].
36
37 Although in some studies post-menopausal women who developed RA had lower joint
39
40 scores than pre-menopausal patients, post-menopausal period has been related to higher
41
42 disease activity and more rapid progression of disability [17-19]. Some studies suggest that
43
44 hormonal replacement therapy (HRT) reduces the risk of RA in post-menopausal women
46
47 [20]. Randomized studies report discrepant effects of HRT on RA severity, with some
48
49 demonstrating no difference while others report significant improvement in disease activity
50
51 and slower progression of radiographic joint destruction [21, 22].
6
22
29
52
Page 5 of 29 Rheumatology
1 2
3 Functional disability and radiographic joint damage progression represent the cumulative
4
5 effect of disease activity in RA and allow to study the long-term effect of reproductive
7
8 factors. The aim of this study was to analyze the association between female reproductive
9
10 and menopausal factors on functional and radiographic progression in women with RA.
11 12 13 14
15 Patients and methods
16
17 Study design and population
18 19 We performed an analysis within a prospective, longitudinal cohort study of patients with
20
21 RA, the Swiss Clinical Quality Management in Rheumatoid Arthritis (SCQM-RA). The
23
24 registry has been described in detail elsewhere [23]. Patients enrolled are followed-up with
25
26 1 to 4 visits yearly and have radiographic assessments at regular intervals. In addition,
27
28 clinical information is updated every time a patient’s disease status requires a change in
30
31 antirheumatic treatment [23, 24]. Because patients are enrolled into the registry by
32
33 rheumatologists generally before starting their first effective antirheumatic therapy, patient-
34 35 reported outcomes such as HAQ typically improve during the first years of follow-up. The
36 37
38 collection of patient data for the SCQM register was approved by a national review board
39
40 and all individuals willing to participate sign an informed consent form before enrolment,
41 42 in accordance with the Declaration of Helsinki.
43 44 45 46
47 Data collection
48
49 The current analysis included patients and disease characteristics collected from March
50
51 1996 through July 2017. Information about female hormonal factors, such as pregnancies,
53
54 breastfeeding, menopause and hormonal treatments were retrospectively retrieved using a
55
56 specific questionnaire applied once between July 2015 and July 2017. The inclusion criteria
6
22
29
45
52
Rheumatology Page 6 of
29
1 2
3 for this analysis were RA as diagnosed by a board-certified rheumatologist, information
4
5 about reproductive factors, and the availability of at least 2 consecutive Health Assessment
7
8 Questionnaire (HAQ) assessments.
9 10 11 12
Study outcomes definitions
13 14
15 The primary outcome was functional disability progression using the Health Assessment
16
17 Questionnaire (HAQ), which predicts future disability and mortality. The minimal
18
19 important change of this instrument has been determined to be between -0.22 and - 0.24
20
21 [25]. A secondary outcome was the rate of radiographic disease progression. The
23
24 radiographic damage was assessed on digitalized radiographs of hands and feet using the
25
26 Ratingen erosion score that evaluates thirty-eight hand and foot joints [24, 26]. If expressed
27
28 as percentage of the joint surface destroyed by erosions, the Ratingen score has a minimal
30
31 detectable radiographic change of 3.5% [27]. For the analysis of radiographic joint damage,
32
33 we included only individuals with at least two consecutive available sets of radiographs. If
34 35
specific joints were sporadically missing on radiographs, we used a linear extrapolation
36 37
38 approach to impute the missing data [28].
39 40 41
42 Exposures of interest
43
44 All female hormonal factors were self-reported by the participants. Number and dates of
46
47 pregnancies and/or miscarriages were retrieved. Parous women were defined as ≥ 1
48
49 pregnancy or miscarriage ever. Multiparous women were defined as women with more than
50
51 1 pregnancy or miscarriage. Active parous period was operationally defined as the 10 years
53
54 period following the first pregnancy or miscarriage. Post-menopausal status was defined as
55
56 cessation of menses for 12 months, using the following question: ‘Have you had a period
6
22
45
Page 7 of 29 Rheumatology
1 2
3 (menses) in the last 12 months?’, and if the answer was negative, age at menopause was
4
5 requested with the following question: ‘How old were you, when you had your last
7
8 menstrual period? [29]. Early age at menopause was defined as menopause occurring
9
10 before 45 years of age [15, 16, 30]. Women in the early post-menopausal period were
11 12
defined as those within less than 6 years after menopause [31]. Early post-menopausal
13 14
15 period may be relevant on functional decline because is characterized by the presence of
16
17 vasomotor symptoms [32]. We further asked post-menopausal women for history of
18 19 hysterectomy, with or without oophorectomy. The use of hormonal replacement therapy
20
21 (HRT) and the duration of use in years after menopause was also assessed. The
23
24 reproductive span was defined as the number of years between the age at menarche and the
25
26 age at menopause [33].
27 28 29 30
31 Statistical analysis
32
33 Baseline disease and reproductive characteristics were compared between pre- and post-
34 35 menopausal women using descriptive statistics. Baseline was the date at first HAQ
36 37
38 assessment for the primary outcome and the date at first radiograph for the secondary
39
40 outcome. Some reproductive variables were considered permanent, such as age at first
41 42 menstrual period, age at last menstrual period or the dates of pregnancies. Other
43
44 reproductive variables could change over time (i.e. pregnancies or presence of
46
47 contraception use) and were analyzed as time-varying variables. Thus, the number of
48
49 pregnancies could increase over time, depending on the dates of pregnancies during follow-
50
51 up.
6
22
45
Rheumatology Page 8 of
29
1 2
3 Other variables and potential confounding factors considered in the analysis were age,
4
5 current tobacco smoking, ever alcohol consumption, body mass index (BMI), disease
7
8 duration, baseline disease activity score (DAS 28), baseline conventional disease modifying
9
10 antirheumatic drugs (cDMARDS) and biologic treatment. We used multiple imputation for
11 12 sporadically missing covariates.
13 14 15 16
17 The rate of disability progression and erosion progression were compared between pre- and
18
19 post-menopausal women using a multilevel regression model for longitudinal data,
20
21 adjusting for potential confounders. We further compared the disability progression
23
24 between three groups: women during pre-menopause, during the early post-menopausal
25
26 period and late post-menopausal period.
27 28 29 30
31 Among pre-menopausal women, the rate of disability and erosion progression was
32
33 compared between parous and nulliparous women. Among parous women, we further
34 35
explored if the disability and radiographic progression was different during the active
36 37
38 parous period than after this period, between women who became pregnant before or after
39
40 RA diagnosis and between women with only one pregnancy and multiparous women.
41 42 Among post-menopausal women, the rate of disability and erosion progression was
43
44 compared between women with an early age versus an older age at menopause and between
46
47 HRT-users and non-users.
48 49 50 51
6
22
29
45
52
Page 9 of 29 Rheumatology
1 2
3 Results
4
5 We included 1667 women in the analysis (see Fig. 1). Participants had a median of 6 HAQ
7
8 assessments [Interquartile range (IQR) 3 to 10] during a median of 5.1 [IQR 2 - 10] years of
9
10 follow-up.
11 12 13 14
15 Baseline characteristics
16
17 Table 1 displays the baseline characteristics of women by menopausal status. Post-
18 19 menopausal women were older, had higher BMI, worse HAQ and more erosions than pre-
20
21 menopausal women.
23 24 25
26 Reproductive characteristics
27
28 The reproductive characteristics of women are shown in Table 2. At baseline, 1025 (61%)
30
31 women were post-menopausal and 642 (39%) were pre-menopausal. Among pre-
32
33 menopausal patients, 316 (49%) were parous at baseline and during follow-up 58 more
34 35 women had a pregnancy and/or a miscarriage. Among the pre-menopausal women, 276
36 37
38 became post-menopausal during follow-up. Post-menopausal women had an older age at
39
40 menarche and more pregnancies, which probably reflects a time-trend. HRT use was
41 42 reported by 282 (28%) women at baseline and 59 more patients began to use HRT during
43
44 follow-up.
46 47 48
49 Progression of functional disability
50
51 HAQ score change over a 5-year period was different between pre- and post-menopausal
53
54 women (p<0.001), with larger improvement in HAQ scores in pre-menopausal than in post-
55
56 menopausal women with a change in HAQ of -0.12 (95% CI: -0.14 to -0.08) and -0.03
6
24
31
47
54
Rheumatology Page 10 of
29
1 2
3 (95% CI: -0.06 to -0.01) respectively (Fig. 2). In the analysis with three groups, functional
4
5 disability improvement was less pronounced in women during the late post-menopausal
7
8 period (HAQ score change, -0.03 (95% CI: -0.05 to 0.00) compared to pre-menopausal
9
10 women (p<0.001). Functional disability progression was not different between women
11 12 during the early post-menopausal period (HAQ score change, -0.06 (95% CI: -0.11 to -
13 14
15 0.01) and pre-menopausal women (p=0.43) (Fig. 3).
16 17 18
19 Functional disability in pre-menopausal women
20 21 During the pre-menopausal period, functional disability did not differ between parous and
22
23 nulliparous during follow-up, with a change of HAQ of -0.15 (95% CI: -0.21 to -0.10) and -
25
26 0.13 (95% CI: -0.19 to -0.09) over 5 years respectively, p=0.27; Supplementary Fig. S1.
27 28 29
30 In a sub-analysis, functional disability during the active parous period had greater
32
33 improvements in functional disability over time than after this period (-0.19 (95% CI: -0.28
34
35 to -0.11) and -0.15 (95% CI: -0.22 to -0.09) over 5 years respectively, p=0.04,
36 37 Supplementary Fig. S2A). Women with pregnancies before RA diagnosis had greater
38 39
40 improvement in HAQ scores over time compared to those with pregnancies only after
41
42 diagnosis ( -0.21 (95% CI: -0.27 to -0.15) and -0.04 (95% CI: -0.14 to -0.06) over 5 years
43 44 respectively, p<0.001, Supplementary Fig. S2B). Functional disability progression did not
45 46
differ between women with only one pregnancy and multiparous (p = 0.63) (Supplementary
48
49 Fig. S2C).
50 51 52
53 Functional disability in post-menopausal women
55 56
6
22
29
45
52
Page 11 of 29 Rheumatology
1 2
3 Women who developed menopause earlier had worse functional progression over time than
4
5 women with an older age at menopause (increase in HAQ of +0.03 (95% CI: -0.01 to
7
8 +0.08) versus -0.05 (95% CI: -0.08 to -0.02) over 5 years respectively, p<0.001,
9
10 Supplementary Fig. S3A). Women who used HRT during the post-menopause did not
11 12 differ in functional progression over time compared to non-users (p=0.66) (supplementary
13 14
15 Fig. S3B).
16 17
18 19 Erosion progression
20
21 Erosion progression was analyzed in 1062 patients with available sequential radiographs, of
23
24 which 630 (59%) were post-menopausal. The erosion progression did not differ between
25
26 pre- and post-menopausal women (1.61% (95% CI: 1.12-2.10) and 1.95% (95% CI: 1.6 to
27
28 2.3) over 5 years, respectively, p=0.76, see Fig. 4). However, radiographic progression was
30
31 greater in nulliparous than in parous women (1.87% (95% CI: 1.25-2.49) and 0.97% (95%
32
33 CI: 0.34 to 1.6) over 5 years, p=0.003). No differences in radiographic progression were
34 35 found between other pre-menopausal or post-menopausal women subgroups.
36 37 38 39
40 Discussion
41 42 We studied the long-term impact of female reproductive factors on functional disability and
43
44 radiographic progression in women with RA. In this cohort of patients with RA, functional
46
47 disability progression differed between pre- and post-menopausal women, with greater
48
49 improvement in HAQ scores over time in pre-menopausal women, not explained by disease
50
51 duration or age. Among pre-menopausal women, being in the active parous period and
53
54 having had a pregnancy before the diagnosis of RA was associated with a positive long-
55
56 term impact on disability. Among post-menopausal women, we observed that functional
6
29
45
52
Rheumatology Page 12 of
29
1 2
3 disability progression was worse in women with an early age at menopause. Radiographic 4
5 progression between post-menopausal and pre-menopausal was not different. Among pre- 7
8 menopausal patients, parous women tended to have less structural joint damage over time 9
10 compared to nulliparous women. No differences on radiographic progression were found 11 12 between other pre- or post-menopausal subgroups. The differential progression in
13 14
15 functional disability and structural joint damage suggests that the functional decline during 16
17 the post-menopause is not primarily driven by joint damage, but by others phenomenon 18
19 occurring during the post-menopausal period.
20 21 22 23
24 A worse progression of functional disability in post-menopausal women has been reported 25
26 in a recent large longitudinal study as well [19]. The study included more than eight 27
28 thousand women, of which more than sixty percent were post-menopausal who had 30
31 significant functional disability decline compared with pre-menopausal women. Several 32
33 mechanisms have been proposed to explain worse functional progression in women after 34 35 menopause. Menopause has been linked to an increase in pro-inflammatory cytokines [34].
36 37
38 In addition, during menopause women report more musculoskeletal pain, although a causal 39
40 association between estrogen deficiency and musculoskeletal pain has not been revealed 41
42 [35]. Vasomotor symptoms, depression, sleep disturbance that affect quality of life during 43
44 post-menopause could also influence HAQ scores in post-menopausal women with RA 46
47 [32]. Interestingly, we found worse functional disability progression in women during the 48
49 late post-menopausal period, typically characterized by less menopausal symptoms, but 50
51 more comorbidities, such as osteoporosis or other chronic diseases that could explain our 53
54 finding.
55 56
However early age at menopause has also been associated with milder disease, notably in a retrospective study that included women older than 45 years with a mean age at RA diagnosis of 63.4 years [37].
6
22
29
45
52
Page 13 of 29 Rheumatology
1 2
3 The median age at menopause in the SCQM cohort was 49 years, which corresponds to the
4
5 age previously reported among RA patients [18, 36]. Early age at menopause was self-
7
8 reported by 25% of post-menopausal patients. Previous studies stated 17% to 20% of early
9
10 age at menopause in women with RA [30, 36, 37]. An age at menopause younger than 45
11 12 years has been associated with an increased risk of RA onset [15, 16, 36]. In this study we
13 14
15 found worse progression of functional disability in women with early age at menopause.
16 17 18 19 20
21 We did not find a difference in functional disability over time
23
24 associated with HRT use. Our results corroborate the findings of the Women’s Health
25
26 Initiative (WHI) Hormone Therapy Trial, which analyzed the effects of HRT on both the
27
28 risk of new-onset RA and the severity of established RA and did not find a significant
30
31 difference in the severity of RA. However, this analysis included only a small subset of
32
33 prevalent RA patients (n= 63) and incident cases (n=105) [38]. In contrast, a large
34 35 observational study reported less functional declined in postmenopausal patients with ever-
36 37
38 use of HRT[19], but HRT use in an observational setting may be associated with well-
39
40 established selection-biases [39].
41 42 43
44 The improvement in disease activity during pregnancy and the decline in disease activity
46
47 and flares in more than one third of patients despite more intense medication during the
48
49 postpartum period could explain why we did not find differences in function over time
50
51 between parous and nulliparous pre-menopausal women [12, 13, 40]. It is not surprising
53
54 that the effect of parity on long-term disease outcomes is variable and controversial. A
55
56 large Swedish cohort study has reported that parous women who developed RA during
6
22
45
52
Rheumatology Page 14 of
29
1 2
3 reproductive age have more disability over time compared to nulliparous women,
4
5 particularly ACPA-negative women. They did not find association between parity and
7
8 severity among ACPA-positive women [13]. In this study, we could not analyze the impact
9
10 of ACPA status because of missing data. In the Norfolk Study Register, women with
11 12
pregnancies after the onset of arthritis reported more favorable functional progression over
13 14
15 5 years [6]. In this analysis, on the contrary, we found more functional disability
16
17 progression over time in women who had a first pregnancy after the diagnosis of RA. This
18 19 finding could be explained by less intensive antirheumatic treatments before, during and
20
21 after pregnancy in Swiss patients [41]. We adjusted for baseline treatment, however the
23
24 change in treatments during pregnancy and the post-partum period was not available for
25
26 these analyses.
27 28 29 30
31 The strengths of our study were a large sample size and a long follow-up time. We used
32
33 widely accepted definitions of reproductive factors and we restricted the analysis of parity
34 35 to pre-menopausal women to reduced recall bias. We explored effect modification in
36 37
38 clinically relevant subgroups. The groups were analyzed with well-validated functional
39
40 disability and radiographic damage scores. However, the study has several limitations,
41 42 such as the self-reported female hormonal factors. We did not adjust by some possible
43
44 confounders, such as education or socio-economic status, however socioeconomic status in
46
47 SCQM patients was reported as very homogeneous in previous publications [24]. We
48
49 observed in general small changes in functional disability and radiographic progression
50
51 over time. The change in HAQ scores between groups over time was below the minimal
53
54 important change threshold and the difference in Ratingen score was far from the minimal
55
56 detectable radiographic change [25, 27]. Patients are typically enrolled in the SCQM
women with RA during the post-menopausal period may justify a closer medical follow-up, particularly those with an early age at menopause, because of the increased risk of comorbidities and potentially less favourable
6
29
Page 15 of 29 Rheumatology
1 2
3 cohort when they receive an adequate antirheumatic treatment or at the time of change in
4
5 DMARD, which could explain a limited power to observe large changes in function or joint
7
8 damage. In addition, we could face a so-called ‘index bias’ previously described in the
8 damage. In addition, we could face a so-called ‘index bias’ previously described in the