SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSORS
  9. REFERENCES
  10. APPENDIX A

Objective

To assess the number of live births in women whose systemic sclerosis (SSc) onset occurred during their reproductive years, and to compare this with general population rates.

Methods

Within the Canadian Scleroderma Research Group cohort, we identified 320 women whose SSc symptoms began prior to age 50 years. We determined the number of children born in the years following first onset of symptoms. We summed the years of followup from the time of first symptoms in subjects up to age 50 years (or oldest age attained, if the subject was age <49 years). We applied age-specific birth rates for Canadian women to these years of followup in order to determine the expected number of live births for the period. We then calculated the standardized incidence ratio (SIR) of observed to expected live births.

Results

In the 320 women studied, the number of live births over the interval since symptom onset was below the expected number (111 live births observed versus 140 expected; SIR 0.79, 95% confidence interval [95% CI] 0.65–0.95). This finding was more prominent in women with diffuse cutaneous disease versus limited cutaneous disease. The mean and median numbers of live births were similar across SSc subgroups based on organ involvement or cyclophosphamide exposure. In repeat analyses, including the reproductive period before SSc symptom onset, the ratio of observed to expected births was 1.23 (95% CI 1.13–1.33).

Conclusion

Compared with the general population, fewer live births were noted in women with SSc, but this phenomenon was only apparent in the period after symptom onset.


Introduction

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSORS
  9. REFERENCES
  10. APPENDIX A

There have been relatively few assessments of reproductive issues in systemic sclerosis (SSc), despite the fact that multiple factors may affect the number of children born to women with SSc. These factors potentially include decreased fertility (1), impaired sexual function and/or personal relationships (2), a relative contraindication to pregnancy (e.g., pulmonary hypertension), or a conscious decision to limit family size (3). Many of the studies performed to date have focused on pregnancies occurring before SSc onset.

Our primary objective was to assess the number of live births in a cohort of women whose SSc symptom onset occurred during their reproductive years, and to compare this with general population rates. We explored potential social and clinical factors that might be associated with lower birth rates in these women. In secondary analyses, we also examined pregnancy histories of the patients throughout their lifespan, including the time before scleroderma onset, assessing both live birth rates and fetal losses for this period.

Subjects and Methods

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSORS
  9. REFERENCES
  10. APPENDIX A

The Canadian Scleroderma Research Group is a national cohort of adult SSc patients; the diagnosis in all subjects is clinically confirmed by a rheumatologist. Cohort members complete detailed questionnaires at baseline on demographic and clinical data. Basing our study on this cohort, we assessed the number of children born to women with SSc, evaluating only those women whose onset of SSc symptoms began prior to age 50 years. The subjects reported the number of children they had and the years of their birth, and we determined the number of children born to the women following the first onset of symptoms. This provided the observed number of live births, in order to calculate the standardized incidence ratio (SIR). The SIR is the ratio of the observed number of events in a sample divided by the expected number of events.

We determined the expected number of events as follows. We summed the years of followup from the time of first symptoms in the subjects up to age 50 years (or the oldest age attained, if the subject was age <49 years). We applied age-specific regional birth rates among Canadian women to these years of followup in order to obtain the expected number of births for the period of followup. Given that there have been some changes in age-specific birth rates over time, we applied the relevant calendar period rates for the interval.

In secondary analyses, we examined the potential effect of social and clinical factors on pregnancies. We compared the ratio of observed to expected births in the following subgroups of patients: married women only, subjects with diffuse cutaneous disease versus limited cutaneous disease, and subjects whose age of symptom onset was ≤30 years versus those whose symptom onset occurred between the ages of 31 and 49 years. We also repeated our analyses, including the reproductive period before SSc symptom onset. Finally, we described miscarriages and stillbirths in the sample.

Results

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSORS
  9. REFERENCES
  10. APPENDIX A

There were 320 women with onset of SSc symptoms before age 50 years. The mean ± SD age at time of symptom onset was 34.5 ± 10.4 years (median 36.4). The majority (87.8%) of the subjects were white. Most (n = 280 [87.5%]) of the subjects fulfilled American College of Rheumatology (formerly the American Rheumatism Association) criteria for SSc (4). At the time of cohort entry, the mean ± SD SSc duration was 9.7 ± 8.4 years (median 7.8). In the sample, 70% of subjects had lung involvement according to the disease severity score by Medsger et al (5), and 3% had renal involvement. Thirty-nine subjects (12.2%) had elevated pulmonary pressure (estimated >40 mm Hg) on echocardiogram. A history of cyclophosphamide exposure was present in 28 patients (13 with limited cutaneous disease and 15 with diffuse cutaneous disease). The majority (70%) of subjects were currently married or living common-law, 14.8% were divorced or separated, 8% were widowed, and the remainder (6.4%) were never married or were not otherwise classified. In our sample, 182 subjects had already gone through menopause at a mean ± SD age of 46.2 ± 6.1 years (median 48 years).

In the sample of all 320 subjects, the number of live births over the interval (n = 111) was below that which would be expected (n = 140), although this phenomenon was not clearly demonstrated when we restricted the analysis only to married women (Table 1). There was evidence of fewer births (relative to what would be expected) in subjects with diffuse cutaneous SSc compared with those without diffuse cutaneous disease. Our data did not suggest a difference in the ratio of observed to expected births among women whose disease had begun at an earlier (≤30 years) versus later (31–49 years) age. The results were similar when we truncated followup time at age 45 years as opposed to age 50 years. The reproductive history in our sample, stratified according to organ involvement and cyclophosphamide exposure, is shown in Table 2. There were no differences in the mean or median number of pregnancies or live births in any of these groups.

Table 1. Live births in women with onset of systemic sclerosis before age 50 years: observed and expected births occurring after the first onset of symptoms*
 ObservedExpectedSIR95% CI
  • *

    SIR = standardized incidence ratio; 95% CI = 95% confidence interval.

  • Expected live births calculated from age-specific Statistics Canada general population rates (6).

  • As defined by LeRoy et al (12).

  • §

    Age at onset of first symptoms.

All (n = 320)1111400.790.65–0.95
Married women only (n = 226)92950.970.77–1.17
Diffuse cutaneous disease (n = 128)36510.710.48–0.94
Limited cutaneous disease (n = 192)82890.920.72–1.12
Age ≤30 years at disease onset (n = 106)§1041210.860.69–1.03
Age 31–49 years at disease onset (n = 214)14190.740.35–1.13
Table 2. Reproductive history in women with systemic sclerosis whose symptom onset occurred before age 50 years (n = 320): pregnancies and live births according to organ involvement and cyclophosphamide exposure
 Pregnancies, mean ± SDLive births, mean ± SDNever pregnant, no. (%)
  • *

    As defined by Medsger et al (5).

Lung involvement*   
 Yes (n = 224)2.4 ± 1.61.9 ± 1.468 (30.2)
 No (n = 96)2.2 ± 1.51.7 ± 1.225 (26.0)
Renal involvement*   
 Yes (n = 10)2.8 ± 1.92.2 ± 1.00 (0)
 No (n = 310)2.2 ± 1.61.8 ± 1.386 (27.8)
Elevated estimated pulmonary artery pressure (>40 mm Hg on echocardiogram)   
 Yes (n = 39)2.7 ± 1.92.1 ± 1.510 (25.6)
 No (n = 281)2.2 ± 1.51.8 ± 1.276 (27)
Cyclophosphamide exposure (ever)   
 Yes (n = 28)2.1 ± 1.71.8 ± 1.37 (25.0)
 No (n = 292)2.3 ± 1.61.9 ± 1.379 (27.1)

When we repeated our analyses, including the reproductive period before SSc symptom onset (i.e., all time between age 15 years and the current age), there was a total of 589 births (mean 1.8 births per woman, median 2) compared with an expected number of 479 births, based on population birth rates. The ratio of observed to expected births based on the entire reproductive history was 1.23 (95% confidence interval [95% CI] 1.13–1.33).

Among the 320 women, there were 97 reports of miscarriages and 8 reports of stillbirths, occurring at any time from age 15 years to the subject's current age. These events were similar in the group of individuals with limited cutaneous scleroderma versus those with diffuse cutaneous disease. We did not have information regarding the year of these events or whether the miscarriage was a first or second trimester loss. Statistics Canada rates for fetal loss include only miscarriages reported by hospitals, plus stillbirths (6). Not all of the miscarriages reported by our subjects necessarily would be included in the fetal loss rate as defined by Statistics Canada, since only a proportion of first trimester pregnancy losses would result in a hospital encounter. Presuming that three-quarters of the miscarriages were first trimester events, and that ∼57% of these early losses would require no medical or surgical intervention (7), the number of fetal losses experienced by our subjects, according to the Statistics Canada definition (i.e., involving a hospital encounter), would be ∼40. Applying age- and calendar period–specific fetal loss rates, we would expect ∼17 fetal loss events in our subjects over the observation interval. This would yield an SIR for fetal loss in SSc of 2.3 (95% CI 1.6–3.1) compared with the general population. However, if 90% of the pregnancy losses in SSc were first trimester losses and the majority of these required no hospital contact, it would mean that the SIR would be much closer to unity (1.1).

Discussion

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSORS
  9. REFERENCES
  10. APPENDIX A

Compared with the general population, we found some evidence of decreased live birth rates in women with SSc whose onset of first symptoms began during child-bearing years, particularly in those with diffuse cutaneous disease. We note that this effect was attenuated when we considered only married women. Our estimate of the expected number of births would have been more accurate if we had data on the marital status of the women over time, but unfortunately, we did not. Since divorced and never-married women have lower birth rates than married women, if women with SSc were less likely to be married than the general population, this would have inflated the number of expected births. However, the number of women in our sample who had never married was relatively small (6.4%) compared with the general Canadian population. Theoretically, early menopause could also contribute to low birth rates, but in our sample we could not establish that the mean age at menopause (46.2 years; 95% CI 34.2–58.2) was lower than expected, since the confidence interval for our estimate overlaps with the national average of 51 years (8).

A limitation of our study is that it was not prospective, and the average duration of symptoms was relatively long. This could mean that more severe SSc cases, having diminished survival, were not represented in our sample. However, we do note that we had a relatively large number of individuals with diffuse cutaneous disease and high skin scores. The registry cohort members cover the spectrum of disease from mild to severe, and include a significant number of individuals with renal and pulmonary involvement. Still, there is likely some aspect of survivorship bias; presuming that women with more severe disease might tend to have fewer children, this may actually have biased our primary results toward the null value. Furthermore, assessing events that occurred prior to cohort entry, as we did, is problematic. This may in part explain why, when we considered the entire reproductive history, including the time before disease onset, the number of live births was not less than that which would be expected.

In our sample, although live birth rates appeared to be decreased after symptom onset, there was evidence that before disease onset, women with SSc had higher live birth rates than expected. One might postulate that this reflects an increased risk of SSc among parous women, related to fetal cell transfer and microchimerism (9). However, 2 previous studies actually found that nulliparity (10) was associated with an increased risk of developing SSc (11). In our sample, in those with diffuse cutaneous disease versus limited cutaneous disease, we noted a similar percentage with a history of miscarriage (21.8% versus 22.0%) and therapeutic abortions (11.6% versus 10.4%). Unfortunately, we do not have information regarding whether the fetal losses occurred before or after disease onset.

There was a suggestion that the number of pregnancy losses was increased in our sample compared with the general population. However, there was likely some misclassification bias because we determined fetal losses by self-report, but general population rates are determined from hospitalization data. This could have inflated the number of fetal losses in the SSc cohort compared with the general population, and thus falsely elevated the relative rate for fetal loss in SSc. At least one other study found an increased rate of miscarriages in SSc (1), although a more recent study did not demonstrate this (3).

Several disease-related factors may contribute to reduced live birth rates in women with SSc. Fertility has been shown to be decreased by some investigators, although this is controversial (1). Faced with a chronic illness that may be associated with high morbidity, some women may choose not to have children or may have a relative contraindication to pregnancy, such as pulmonary hypertension. Childbearing decisions in some women might also be influenced by fear of potential adverse fetal outcomes because SSc has been associated with an increased risk of prematurity and infants small for their gestational age (3), or concern of transmitting the disease to children. The changes in appearance, physical limitations, and emotional impact of the disease may impair the personal relationships and sexual lives of these women.

In summary, compared with the general population, we found some evidence of decreased live birth rates in women with SSc whose onset of first symptoms began during child-bearing years, particularly in those with diffuse cutaneous disease.

AUTHOR CONTRIBUTIONS

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSORS
  9. REFERENCES
  10. APPENDIX A

Dr. Bernatsky 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. Bernatsky, Pope, Robinson, Baron.

Acquisition of data. Hudson, Pope, Markland, Robinson, Jones, Docherty, Abu-Hakima, LeClercq, Dunne, Smith, Mathieu, Khalidi, Sutton, Baron.

Analysis and interpretation of data. Bernatsky, Hudson, Pope, Baron.

Manuscript preparation. Bernatsky, Hudson, Pope, Vinet, Smith, Sutton, Baron.

Statistical analysis. Bernatsky.

Patient recruitment/registry entry. LeClercq.

ROLE OF THE STUDY SPONSORS

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSORS
  9. REFERENCES
  10. APPENDIX A

Actelion, Pfizer, and Encysive had no involvement in the study design, data collection, data analysis, and writing of the manuscript, and allowed the authors to publish their findings without restriction.

REFERENCES

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSORS
  9. REFERENCES
  10. APPENDIX A

APPENDIX A

  1. Top of page
  2. Abstract
  3. Introduction
  4. Subjects and Methods
  5. Results
  6. Discussion
  7. AUTHOR CONTRIBUTIONS
  8. ROLE OF THE STUDY SPONSORS
  9. REFERENCES
  10. APPENDIX A

ADDITIONAL MEMBERS OF THE CANADIAN SCLERODERMA RESEARCH GROUP

Additional members of the Canadian Scleroderma Research Group not listed as coauthors include Elzbieta Kaminska, Ariel Masetto, and Sophie Ligier.