Reproductive trends in females with inflammatory joint disease

Inflammatory joint diseases (IJD), here including rheumatoid arthritis (RA), spondylarthritis (SA), psoriatic arthritis (PsA) and juvenile idiopathic arthritis (JIA) are characterised by chronic, systemic inflammation and may also affect internal organs. Previous studies have shown various effects of IJD on birth outcome [1‐4]. Especially, low birthweight in offspring and preterm birth have been identified as prominent findings. Over time, improvements in diagnosis and treatment, and also change in registration of routines may influence birth outcomes [5, 6]. Before 1990 several females with IJD were dissuaded from having children [7, 8], and there were little experience with monitoring and medical treatment in pregnancy. In a study published in 1991 27 % of the women had been dissuaded from pregnancy by health professionals, family or other patients [7]. Since the end of the 1990ies new treatment options like the tumor necrosis factor (TNF)-α inhibitors have made more women able to achieve remission and thus enabled them to plan their pregnancies. Possible changes in pregnancy outcomes along with improved treatment options have not been published.

In the present study we wanted to assess secular trends in the number of births and the effects of IJD on preterm birth and birth weight in the newborns. The data were provided by the Medical Birth Registry of Norway from 1967 to 2009.

7 %. The mean age at delivery for females with IJD was higher than among the reference women through all periods, and parity was higher from 1990 onwards. In the period 2000–2009 more females with IJD were smoking at time of conception than among the references (Table 1). For offspring of IJD mothers the crude mean birth weight (BW) was significantly lower than in controls through all periods. However, the mean BW difference decreased from −61 g in the 1970ies to −17 g after year 2000, and was no longer significant for the last period (Table 2). The risk of BW < 2500 g was significantly higher in newborns of IJD women from 1980 to 1999 with an aOR varying from 1.4 (95 % confidence interval (CI) 1.2, 1.8) to 1.2 (95%CI 1.03, 1.5), but not from year 2000 (aOR 1.2 (95 % CI 0.9, 1.4)). Similarly, the risk of SGA births was higher for mothers with IJD from 1967 to 1999, and the aOR varied from 1.5 (95 % CI 1.2, 1.9) to 1.2 (95 % CI 1.1, 1.4). After year 2000 the risk of SGA was not significantly different between IJD females and references (aOR 1.1 (95%CI 0.9, 1.2)). However, the risk of prematurity was significantly higher among offspring of IJD mothers from 1980 onwards (Table 3). The risk varied from 1.5 (95 % CI 1.2, 1.8) in the 1980ies to 1.3 (95 % CI 1.1, 1.5) from year 2000.

Since 1967 births among females with IJD have significantly increased. The low number of births during the earliest period may have several explanations. In accordance with previous reports females with IJD may have been advised not to have children [7, 8]. After 1990 new treatments have evolved and knowledge has been gathered about the use of anti-inflammatory drugs before and during pregnancy. Thus more women with inflammatory active disease have been able to plan pregnancy. Biological agents like the TNF-α inhibitors, that target cytokines and immune cells, were introduced in Norway in 1999. TNF-α inhibitors are effective inhibitors of disease activity and are now recommended used until conception [12]. High inflammatory disease activity has a negative effect on sexual function [13], and better treatment options may also improve the sexual function and the possibility to achieve pregnancy. Risk of developing IJD increases with age, and the age of primiparous Norwegian women has increased since 1967 and is currently 28.5 years [14]. Thus, the higher proportion of births in females with IJD may be caused by the general increase in maternal age as well as improved documented safety of medical treatments.

We observed that the mean BW in newborns of IJD women approached that of the reference women during the study period, and the mean difference was no longer statistically significant after year 2000. The proportion of newborns with BW <2500 grams declined from 6 % during 1967 – 1979 to 4 % after year 2000 in females with IJD. Correspondingly, the SGA proportion declined from 17 % to 8 %. Interestingly, these declines coincide with the introduction of better treatment options. Low BW has been associated with high disease activity at conception and during pregnancy [1, 2]. TNF-α inhibitors may be administrated up to time of conception [15, 16] and may help more women with inflammatory active IJD to stay in or close to remission until they conceive. Better knowledge of the use of other anti-inflammatory drugs in pregnancy such as prednisone, hydroxychloroquine and sulfasalazine may also have influenced the results [16, 17]. For females with IJD Norwegian national guidelines suggest regularly surveillance in pregnancy from general practitioners, rheumatologists, obstetricians and other specialists when necessary [18]. The improved care taking and monitoring of pregnant women with IJD may have influenced our results. Ultrasound surveillance of fetal growth has become more widespread over the study period, and the use of ultrasound to detect intrauterine growth restriction may have had an impact on low BW and SGA. The proportion of preterm deliveries from 1979 onwards was 7 % in IJD women compared to around 5 % in reference women. Thus, the observed difference in preterm birth has persisted over the years despite the introduction of new effective drugs. In the earliest period (1967–1979) the number of cases was insufficient for interpretation. In a Danish population based study spanning from 1977 to 2008, offspring of females with RA had a 1.5 higher risk of preterm birth than in the general population [19]. The study did not examine secular trends. The newborns of RA women also had a slightly lower BW [19]. The observed increased risk of preterm birth was present both in females diagnosed with RA and in those with preclinical RA. The latter group had not been exposed to any anti-rheumatic medication. The study hypothesized that potential disease-induced factors influenced the observed differences in fetal growth and preterm birth. These factors are still unknown. A higher rate of preterm birth was also reported in another study of females with RA spanning from 2001 to 2009. Among 46 pregnancies 28 % delivered prior to 37 weeks [20]. The study did not find any associations between preterm birth and active disease at conception or during pregnancy. However, discontinuation of medication because of pregnancy was associated with a significantly earlier gestational age at delivery. We did not have information about drug exposure in pregnancy, and isolating the effect of disease versus treatment is difficult in our material. A few studies have indicated that use of prednisone, even in low dose may induce preterm delivery [2, 21]. Prednisone has been one of the most frequently used medications to treat inflammation in pregnant arthritic females during the last decades, but probably not in the late sixties and early seventies when the use of prednisone was sparsely documented in pregnancy [22‐25]. More women with IJD were smoking at time of conception during 2000–2009, but we did not have information about smoking before year 2000. Thus, it is unknown how smoking might have influenced the results overall.

An advantage of the present study was the availability of population based data from a national registry. This limits selection bias of IJD cases and ensures representative references.

Among potential limitations we cannot exclude the possibility of misclassification bias. However, a validation study of a selection of rheumatic diagnoses in MBRN including inflammatory arthritic diagnoses, reported that 97 % of the diagnoses were correct [26]. We excluded unspecified arthritic diseases, inflammatory connective tissue diseases and vasculitis in the selection of IJD cases (Additional file 1). The birth outcomes of mothers with specific IJD diagnoses were not possible to study separately. During the period with ICD-8 coding, all IJD diseases were grouped together, and the number of cases in each diagnostic group of ICD-10 coding was not large enough. Another possible limitation is underreporting of diagnoses. MBRN has validated 169 females with rheumatic diagnoses in the period 1967–1995. Overall, 10 % of pre-pregnant diagnoses were underreported. However, underreporting was most pronounced during the earliest period (1967–1976) when it was 45 % [26]. Underreporting may certainly have led to an underestimate of births in IJD women especially during the earliest study decade. Finally, we lacked detailed information such as socioeconomic status and clinical data regarding body mass index, disease activity and medication in pregnancy. Information about smoking habits was only available after 1999.

An increasing number of births among females with inflammatory joint disease have been observed between 1967 and 1999. Birth weight was lower in newborns of mothers with inflammatory joint disease than in the general population until 1999, but not later. However, an increased preterm birth rate seems to persist.

aOR, adjusted Odds ratio; BW, birthweight; CI, confidence interval; ICD, international classification of diseases; IJD, inflammatory joint disease; JIA, juvenile idiopathic arthritis; MBRN, medical birth registry of Norway; PSA, psoriatic arthritis; RA, rheumatoid arthritis; SA, spondylarthritis; SD, standard deviation; SGA, small for gestational age; TNF, tumor necrosis factor.