SEARCH

SEARCH BY CITATION

Keywords:

  • Meta-analysis;
  • multiple sclerosis;
  • pregnancy;
  • systematic review

Abstract

  1. Top of page
  2. Abstract
  3. Background
  4. Results
  5. Discussion
  6. Conclusion
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethical approval
  10. Funding
  11. Acknowledgements
  12. References
  13. Supporting Information

Please cite this paper as: Finkelsztejn A, Brooks J, Paschoal F, Fragoso Y. What can we really tell women with multiple sclerosis regarding pregnancy? A systematic review and meta-analysis of the literature. BJOG 2011;118:790–797.

Background  Although several papers report on pregnancy and multiple sclerosis (MS), no systematic review of the literature has been carried out. Neurologists and obstetricians need to have proper information to discuss with women presenting with MS who consider pregnancy.

Objectives  Literature review and meta-analysis of data on pregnancy in women with MS.

Search strategy  The present work followed the recommendations of the PRISMA Statement. Using the PICO framework, the authors independently searched for the terms ‘pregnancy’ OR ‘gestation’ OR ‘pregnant’ AND ‘multiple sclerosis’ OR ‘MS’ in the following databases: EMBASE/Excerpta Medica, Medline, Pubmed, Scopus, Index Medicus, Biomed Central, Ebsco Fulltext, LILACS, Scielo and the Cochrane Database of Systematic Reviews. Selection criteria: only papers presenting original work with analysis of at least one of the outcomes among pregnant women with MS were included.

Data collection and analysis  Two independent workers performed the literature review. All the authors selected and read the relevant papers. Two other authors summarised data for analysis.

Main results  Twenty-two papers reporting on 13 144 women with MS and their pregnancies were analysed. A significant decrease in relapse rate was observed during pregnancy, followed by a significant increase after delivery. Miscarriages, low birthweight, prematurity, neonatal death and malformations were assessed among these women and their offspring. There seems to be a regional influence on the rates of caesarean sections and abortions among women with MS. Neonatal death and malformation rates did not seem to be particularly high.

Authors’ conclusions  The present work provides evidence-based data that can be discussed with women with MS and their relatives when pregnancy is considered by these families.


Background

  1. Top of page
  2. Abstract
  3. Background
  4. Results
  5. Discussion
  6. Conclusion
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethical approval
  10. Funding
  11. Acknowledgements
  12. References
  13. Supporting Information

Multiple sclerosis (MS) is most prevalent among women of fertile age so poses the extra challenge of guiding these women with regard to pregnancy. It is a chronic neurological disease, manifesting by clinical and subclinical relapses of central nervous system demyelination. The disease comprises inflammation, demyelination and degenerative features.

Over recent decades, many papers have been published on the subject of pregnancy and MS, and the days when women were discouraged from having children seem to have passed.1 However, it seems that termination of pregnancy because of MS or maternal exposure to MS drugs still occurs in many countries. Faced concomitantly with pregnancy and MS, neurologists and obstetricians still base their medical approach on knowledge provided by one or more separate studies.

It is unlikely that a single study can answer a clinical question, no matter how well designed it is. Properly conducted systematic reviews and meta-analyses may provide the least biased cumulative state of the art regarding scientific knowledge on a specific subject. By pooling the results from several studies, a summary of results with statistically greater precision is made available in an attempt to answer a given question.2

The purpose of the present systematic review and meta-analysis was to provide a summary of the many studies carried out in different countries. Some authors had reported on pregnancy and MS before the Poser diagnostic criteria were established in 1983.3 However, it was only after these criteria were published that women with definite MS could be considered for inclusion in any study. Therefore, the present paper has concentrated on papers published since 1983, although the authors were aware that the discussion of pregnancy and MS is not a new one.

In 1983, Poser and Poser4 reported on the postpartum negative effect on the relapse rate. In their report, there was a lack of evidence for any negative influence from MS on the offspring. However, there was a high prevalence of miscarriages and termination of pregnancy. In 1984, Korn-Lubetzki et al.5 reported on a pattern of relapse remission during pregnancy followed by postpartum relapse exacerbation, in a retrospective assessment of 338 women with MS. The authors commented that this pattern was similar to that observed in other putative diseases with a possible immunological component.

In 1998, a large prospective study was performed by Confavreux et al.,6 assessing the natural history of MS in pregnant women. This study included data from 12 European countries and over a hundred researchers. It confirmed the initial observations regarding the lower relapse rate during pregnancy, with a puerperal increase in the relapse rate. Moreover, in 2004, Vukusic et al.7 published further data on this same group of women, showing that after this initial increase in relapses, the relapse rate returned to pre-pregnancy values.

Several other studies (to be discussed below) confirmed this finding, which became classic and virtually undisputable. However, other data on the mother and child have been published with more contradictory results. Disturbing data regarding low weight among newborns, shortened gestational period, high frequency of abortions and caesarean sections and presence of child malformations in some reports were at odds with other authors’ data, which did not show such findings. In fact, it is very likely that the rates of abortions and caesarean sections have been greatly influenced by other factors beyond the disease itself. For example, legal, cultural and religious attitudes towards abortion are different in many countries, which may have an effect on the abortion rates reported by different authors. Caesarean sections are also more frequently performed in some countries than in others, independently of underlying diseases in the mother. Additionally, it is possible that attitudes regarding the initial fears of malformations relating to drugs used by mothers may have changed over the years. Therefore, a systematic review and meta-analysis of all these outcomes may show the influence of medical attitudes that are related to the country and the time when the study was performed. Nonetheless, the need for a systematic review on the subject encouraged the present authors to proceed with this study.

Objective

The aim of the present paper was to assess the current evidence regarding the short-term effects of pregnancy on MS and vice versa. Through this, guidance for neurologists and obstetricians for situations in which a woman with MS wants to become (or already is) pregnant would be provided.

Search strategy

Published data were used in this study and in the preparation of this paper; hence no ethical approval was required.

The present work followed the recommendations of the PRISMA Statement.8 Using the PICO framework,9 the authors independently searched for the terms ‘pregnancy’ OR ‘gestation’ OR ‘pregnant’ AND ‘multiple sclerosis’ OR ‘MS’ in the following databases: EMBASE/Excerpta Medica, Medline, Pubmed, Scopus, Index Medicus, Biomed Central, Ebsco Fulltext, LILACS, Scielo and the Cochrane Database of Systematic Reviews. Abstracts of articles in any language containing these words in English (in the title, key words or abstract) were independently reviewed by two authors (JBBB and FMPJ).

The outcomes assessed were: (1) the effect of pregnancy on MS relapse rate; (2) pregnancy complications and delivery mode; (3) the relative risk of prematurity and low birthweight; (4) the prevalence of malformations; and (5) the breastfeeding index.

  •  Qualitative data: after screening the title and abstract, all texts not presenting data on outcomes relating to the short-term effects of pregnancy on MS and vice versa were discarded. Papers specifically discussing the effects of MS drug treatments on pregnancy were also discarded.
  •  Quantitative data: only papers presenting original work with analysis of at least one of the aforementioned outcomes among women with MS were included. Reference lists from these papers were read to search for other possible relevant papers. Abstracts from scientific meetings, review papers, anecdotal case reports, duplicate papers and editorials were excluded.

Every effort was made to obtain the full text of all relevant papers. These articles were individually read by all of the authors, who summarised the results in an Excel file for subsequent analysis using the Comprehensive Meta Analysis (CMA) software, version 2 (Biostat, Englewood, NJ, USA).

Although publication bias must be checked on meta-analysis, the present study did not allow for such evaluation. The assessed outcomes were rates and prevalence of given events, therefore there were no ‘negative’ results as might have been expected with drug trials. Data were collected and pooled for the following outcomes: relapse rates (before pregnancy, during pregnancy and after delivery), abortion rate, caesarean section rate, low birthweight (<2500 g) rate, prematurity (<38 weeks) rate and breastfeeding rate. The summarised effect for each of these outcomes was calculated for fixed and random effects, using the inverse variance method. The Q-statistic was calculated as the ratio of variation among studies, considering the internal error in each study. The I2 statistic defined the proportion of the observed variance, reflecting the real differences in the effect of the size. The Q and I2 statistical values measured the degree of heterogeneity among the studies. These values are presented together with the summarised effect in each one of the graphs.

Figures in the form of graphs were presented for the summarised effects of both the fixed and the random effects. Statistical calculations for each of the outcomes were carried out using proportions. In addition to the results presented as summarised effects, the percentages of results were also presented. All confidence intervals were 95%.

When the authors presented data on the relapse rates during or after pregnancy considering different trimesters,6,10 the final result considered in the meta-analysis was that of an average for the whole studied period.

Percentages of reported malformations and neonatal deaths were collected and are presented separately (see Table S1).

All discussions and disagreements were settled in a meeting between two of the authors (AF and YDF).

Data collection and analysis

The active search found 83 studies. Six papers (one in Dutch, one in Polish, three in Russian and one in Finnish) with no abstract were excluded after our attempts to contact the journals concerned did not produce any response.

Of the 77 potentially eligible studies identified by the search strategy, 54 papers were excluded after evaluation of the full text. These excluded papers did not conform to the criteria used for this systematic review because they consisted of single case reports, reviews or editorials.

In summary, 23 original papers have been published since 1983 assessing the short-term effects of pregnancy on MS, and vice versa.4–6,10–29 Three papers from Dahl et al.21–23 probably included many women from the same database, but analysed different control groups and outcomes. Two of these papers were included,21,22 but the 2008 paper23 was discarded because it exclusively provided data on different phases of MS in relation to pregnancy complications. Figure S1 presents the flow chart of the literature review and paper selection.

The findings of the 22 papers analysed are summarised (see Table S1) and in Figures 1–8.

image

Figure 1.  Relapse rate in the year previous to pregnancy in women with MS.

Download figure to PowerPoint

image

Figure 2.  Relapse rate during pregnancy in women with MS.

Download figure to PowerPoint

image

Figure 3.  Relapse rate after delivery in pregnant women with MS.

Download figure to PowerPoint

image

Figure 4.  Rate of miscarriages and pregnancy termination among pregnant women with MS.

Download figure to PowerPoint

image

Figure 5.  Rate of deliveries by caesarean section in women with MS.

Download figure to PowerPoint

image

Figure 6.  Rate of preterm (<38 weeks) deliveries in pregnant women with MS.

Download figure to PowerPoint

image

Figure 7.  Rate of low birthweight (<2500 g) among pregnant women with MS.

Download figure to PowerPoint

image

Figure 8.  Summary of breastfeeding rate among pregnant women with MS.

Download figure to PowerPoint

Results

  1. Top of page
  2. Abstract
  3. Background
  4. Results
  5. Discussion
  6. Conclusion
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethical approval
  10. Funding
  11. Acknowledgements
  12. References
  13. Supporting Information

As described in the Methods section, both fixed and random effects were assessed for the outcomes and are presented below and in the figures.

Effect of pregnancy on MS relapse rate

During the year preceding pregnancy, the women presented 0.435 ± 0.021 relapses/year (variance 0.000; range 0.394–0.476; 95% CI 0.40–0.72; Figure S1). During pregnancy, the relapse rate decreased to 0.182 ± 0.012 (variance 0.000; range 0.158–0.206) or 0.26 relapses/year (95% CI 0.19–0.32; Figure 1). After delivery, the relapse rate increased to 0.703 ± 0.024 (variance 0.001; range 0.656–0.750) or 0.758 relapses/year (95% CI 0.64–0.87; Figure 2). The data for these calculations were obtained from 1221 pregnancies reported in 13 papers (Table S1). The relapse rate significantly decreased during pregnancy in relation to the preceding year (P < 0.0001). After delivery, the relapse rate was significantly higher than it had been during the year preceding pregnancy (P < 0.0001). The period for evaluation of the postnatal relapse rate varied among authors from 3 to 12 months.

Pregnancy complications—abortion rates by miscarriages and termination of pregnancy

The abortion rate was found to be 23.80% (range 24.90–31.80%) or 27.90%; 95% CI 23.30–32.70%. These results did not only include miscarriages because in some studies, termination of pregnancy may have been considered by the women and their doctors because of the use of MS-related drugs. Except for a Brazilian study, which showed a lower abortion rate (4.08%),28 all other studies showed abortion rates between 20 and 30% among women with MS.

Other pregnancy complications were not fully reported and consisted of anecdotal data that could not be considered for meta-analysis.

Delivery mode

The caesarean section rate was 41.50% (range 40.50–42.40%) or 21.4% (95% CI 11.20–36.90%; Figure 5). The large numbers of women assessed in a national birth registration database by Kelly et al.29 clearly influenced these results. However, in all the other studies, the caesarean rate was between 9.6 and 41.10%, hence showing the possibility of a higher risk of caesarean section among women with MS. However, cultural and geographical influences on the caesarean rates cannot be excluded.

Risk of prematurity

The prematurity rate was found to be 10% (range 8.80–11.40%) or 10.20% (95% CI 8.10–12.90%). The results were consistent in all studies (Figure 6).

Risk of low birthweight

The risk of low birthweight was found to be 5.80% (range 4.80–6.90%) or 5.7% (95% CI 4.2–7.8%). The results were consistent throughout the studies (Figure 7).

Prevalence of malformations and neonatal death

Two clearly described malformations were reported: one case each of cleft lip and Down syndrome. The reported rates of unspecified malformations and neonatal deaths ranged from 1.13 to 6.25% of all births, with an average of 3.03% (Table S1). Four neonatal deaths were reported among 1081 deliveries (0.37%). Malformations and neonatal deaths were not clearly defined as being disease-related or drug-related in any of the papers.

Breastfeeding index

The event rate was 56.50%, ranging from 52.50 to 60.50, or 54.5% (95% CI 38.1–70.0%). Results varied in the countries reporting these rates. In no paper was there a comparison with the national expected breastfeeding rate, although cultural aspects might have influenced these results (Figure 8).

Discussion

  1. Top of page
  2. Abstract
  3. Background
  4. Results
  5. Discussion
  6. Conclusion
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethical approval
  10. Funding
  11. Acknowledgements
  12. References
  13. Supporting Information

It is not difficult to understand the reasons for anxiety among women with MS. The uncertainties regarding disease evolution and the limited success of therapies are only some of the motives for worry among these women and their relatives. This inflammatory and degenerative neurological disease poses an extra challenge for women, their relatives and health professionals when these young women consider pregnancy. Multiple sclerosis is not a hereditary disease, and the risk of having children who may develop the disease is small.30 However, the effects of MS on pregnancy outcomes and vice versa need to be fully discussed with women and their relatives. These women are mostly young, with a life-long disabling neurological disease. Presentation of clear, evidence-based data is imperative for those involved in pregnancy among women with MS.

The present literature review and meta-analysis showed that women with MS do not seem to present a significantly higher risk of obstetric and neonatal complications. It is possible that there was a relatively higher prevalence of caesarean sections, abortions, prematurity and low birthweight. However, these rates did not reach levels that would signify great concern and may, in many cases, be a reflection of cultural, religious, geographical and regional backgrounds. The increased postnatal relapse rate seems to be the only real risk for these women. With awareness of this risk, neurologists and obstetricians can work together at the time of delivery. There is no international consensus on the approach to this situation and different countries adopt different measures. Obviously, adequate disease control before conception and pregnancy planning are of great importance for these women.

The main strength of the present work was the number of papers on this subject. Likewise, the large number of women studied, from many countries and databases, rendered the results very reliable. The vast majority of the results were similar across all studies.

The main weakness of the present work was the excessively large number of women studied in one of the papers,29 in comparison with the other articles, which could produce a final result more representative of one particular country. However, regarding the outcomes studied in our review, this paper by Kelly et al.29 only provided data on low birthweight rate and delivery mode. Their results on these outcomes were similar to those of other studies. Another negative point is the inherent difficulty in obtaining all the data from all the studies because the methodology varied among the published papers.

Conclusion

  1. Top of page
  2. Abstract
  3. Background
  4. Results
  5. Discussion
  6. Conclusion
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethical approval
  10. Funding
  11. Acknowledgements
  12. References
  13. Supporting Information

The large number of papers reporting cases of pregnancy among women with MS shows the worldwide importance of the subject. There has been no previous systematic review and meta-analysis on these papers. The present work provides evidence-based data that can be presented and discussed with women with MS and their relatives, when pregnancy is considered by these families.

Disclosure of interests

  1. Top of page
  2. Abstract
  3. Background
  4. Results
  5. Discussion
  6. Conclusion
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethical approval
  10. Funding
  11. Acknowledgements
  12. References
  13. Supporting Information

There are no conflict of interests to declare. All authors have received support from pharmaceutical companies for participation in congresses.

Contribution to authorship

  1. Top of page
  2. Abstract
  3. Background
  4. Results
  5. Discussion
  6. Conclusion
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethical approval
  10. Funding
  11. Acknowledgements
  12. References
  13. Supporting Information

AF conceived the idea and designed the study. JBBB and FMPJ performed the literature review and selected relevant papers. All the authors read the potentially relevant papers and selected those for review. AF and YDF analysed data and AF carried out the meta-analysis. YDF wrote the paper.

Details of ethical approval

  1. Top of page
  2. Abstract
  3. Background
  4. Results
  5. Discussion
  6. Conclusion
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethical approval
  10. Funding
  11. Acknowledgements
  12. References
  13. Supporting Information

All data were collected from published papers reporting on unidentified women; therefore no ethical approval was required.

References

  1. Top of page
  2. Abstract
  3. Background
  4. Results
  5. Discussion
  6. Conclusion
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethical approval
  10. Funding
  11. Acknowledgements
  12. References
  13. Supporting Information

Supporting Information

  1. Top of page
  2. Abstract
  3. Background
  4. Results
  5. Discussion
  6. Conclusion
  7. Disclosure of interests
  8. Contribution to authorship
  9. Details of ethical approval
  10. Funding
  11. Acknowledgements
  12. References
  13. Supporting Information

Figure S1. Flow chart for literature review and selection of papers for meta-analysis.

Table S1. Data for meta-analysis. Total number of pregnancies = 13 144.

FilenameFormatSizeDescription
BJO_2931_sm_FigS1.doc25KSupporting info item
BJO_2931_sm_TableS1.doc70KSupporting info item

Please note: Wiley Blackwell is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.