Pre-eclampsia is more common in women with active psoriatic arthritis in pregnancy: a population-based study

Abstract

Objective

To investigate the association between peripheral disease activity and pre-eclampsia, gestational hypertension, preterm birth and fetal growth in women with psoriatic arthritis (PsA).

Methods

Data from a Norwegian nationwide register (RevNatus) were linked with data from the Medical Birth Registry of Norway (MBRN). Cases were singleton births in women with PsA with available disease activity assessment (n=109) included in RevNatus 2010 to 2019. Singleton births registered in MBRN during the same decade served as population controls (n=575 798). Disease activity was assessed by Disease Activity Score based on 28 joints using C reactive protein (DAS28-CRP) in 2nd and 3rd trimester. Active PsA was defined as DAS28-CRP≥2.6 (n=34) and inactive PsA as DAS28-CRP<2.6 (n=75).

Results

Pre-eclampsia was most frequent in women with active PsA (3/34, 8.8%), with a risk difference of 6.1% (95% CI 0.3 to 20.3, p=0.036) compared with population controls (2.6%). Gestational hypertension occurred in 2/34 (5.9%) of women with active PsA, with a risk difference of 4.2% (95% CI 0.0 to 17.4, p=0.065) compared with population controls (1.7%). Pre-eclampsia and gestational hypertension occurred in similar proportions in women with inactive PsA (1.3%, p=0.59 and 2.7%, p=0.24, respectively) and population controls. The occurrence of preterm birth and abnormal fetal growth was comparable in cases and population controls.

Conclusion

Hypertensive disorders of pregnancy occurred more often in women with active, but not inactive PsA. We found no increased risk for preterm birth or abnormal fetal growth in women with PsA.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• There are scarce and conflicting results concerning the risk for hypertensive disorders, preterm birth and abnormal fetal growth in women with psoriatic arthritis (PsA).

• Data on disease activity during pregnancy are limited and call for further research.

WHAT THIS STUDY ADDS

• Information on the risk for hypertensive disorders, preterm birth and fetal weight in women with active PsA and women with inactive PsA compared with population controls.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• Findings of active PsA as a risk factor for hypertensive disorders in pregnancy may motivate improved pregestational counselling as well as targeting inactive disease during pregnancy.

Introduction

Psoriatic arthritis (PsA) is a heterogenous disease, in which psoriasis or nail dystrophy usually precedes other symptoms such as peripheral arthritis, inflammation of the spine, enthesitis and dactylitis.^{1 2} A recent study in Norway reported an incidence of PsA in adults ≥18 years of 26/100 000 person years, with a female-male ratio of 1.2:1 and a peak at 50–55 years.³

The total burden of PsA is affected by cardiovascular, metabolic and psychological comorbidities which occur more often in this patient group than in the general population.⁴ Chronic hypertension, obesity and diabetes are conditions associated with PsA that are risk factors for hypertensive disorders of pregnancy and preterm birth.

Hypertensive disorders of pregnancy include chronic hypertension, gestational hypertension and pre-eclampsia.⁵ Globally, hypertensive disorders of pregnancy cause considerable maternal and perinatal morbidity and mortality.⁵ In Norway, the prevalence of pre-eclampsia is low and there has been a decrement over the last decade from 3.4% in 2007–2010 to 2.7% in 2015–2018.⁶ Increasing use of labour induction and aspirin has probably contributed to this decline.⁶ The revised two-stage placental model of pre-eclampsia argues that placental syncytiotrophoblast stress triggers the hypertensive disorders occurring both early and late (before and after 34 weeks). The model integrates maternal age, parity, diabetes, chronic hypertension, obesity, assisted reproductive technology (ART), maternal autoimmune disease and other known risk factors, describing them as potential accelerators in the development of the clinical manifestations of pre-eclampsia.⁷

Preterm birth is defined as birth before 37 weeks of gestation.⁸ Preterm birth may be spontaneous or medically indicated.⁹ Maternal age, smoking, obesity, maternal stress and intra-amniotic infection or inflammation are risk factors for spontaneous preterm birth.¹⁰ In the Nordic countries, preterm birth occurs in about 6%.⁸ As in other high-income countries, most of these children reach adulthood. Still, there is evidence for increased multimorbidity in adolescence when born preterm, conditions that may affect physical and mental health during the entire lifespan.¹¹

A recent review concluded with an increased risk for pre-eclampsia and preterm birth in women with PsA,¹² but found no association with PsA and small for gestational age (SGA). One of the studies reported an unadjusted increased risk for hypertensive disorders of pregnancy and an increased risk for moderate preterm birth (32–36 weeks) in women with PsA.¹³ This study reported information on disease activity markers and was able to discuss disease activity as a risk factor for pregnancy complications.

The objective of the present study was to investigate a possible association of active inflammatory disease in pregnant women with PsA and the occurrence of pre-eclampsia, gestational hypertension, preterm birth and abnormal fetal growth. We hypothesised that active PsA in the 2nd and 3rd trimesters increases the risk for pregnancy complications.

Material and methods

In this prospective cohort study, we linked data from the Medical Birth Registry of Norway (MBRN) with data from the RevNatus registry.

The MBRN is a national health registry with mandatory registration of all births and pregnancies ended after 12 weeks of gestation. Information about maternal health, obstetric history and complications during pregnancy is registered by the midwives and doctors during the free prenatal care offered to all pregnant women in Norway. Birth outcomes are registered without delay, and after birth, all collected information is reported electronically from all Norwegian labour wards (or health personnel who assist with births outside an institution), within 1 week after discharge.¹⁴ Inflammatory rheumatic disease in the mother is coded according to the International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-10),¹⁵ and is part of the data about maternal health before and during pregnancy. Singleton live births recorded in MBRN 2010–2019 qualified for inclusion in this study. Mothers registered with the diagnosis PsA according to the ICD-10 codes M07.0, M07.1, M07.2, M07.3 (+L40.5) contributed to the cases. All remaining mothers that did not have a rheumatic disease according to the ICD-10 code classification contributed to the population controls, see online supplemental table S1.

RevNatus is a Norwegian nationwide quality register operated by The Norwegian National Network for Pregnancy and Rheumatic Diseases. Female patients above 16 years of age diagnosed by a rheumatologist with a rheumatic disease qualify for inclusion before or during pregnancy. All the local outpatient rheumatology clinics include eligible patients. They have a follow-up in each trimester of the pregnancy, around week 10–12 in the 1st trimester, week 20–22 in the 2nd trimester and week 32–34 in the 3rd trimester, and at 6 weeks, 6 months and 12 months after birth. Not all women attend all visits for various reasons. Demographic variables and obstetric history are recorded at inclusion before or during pregnancy. Disease activity assessment, medication use, laboratory status and self-reported health status are recorded at each follow-up. Pregnancy outcome is registered 6 weeks after birth and breastfeeding is assessed at the visits after birth.

Patient involvement

Two patient representatives were involved in the development of the RevNatus registry and in the associated projects based on data from RevNatus.

Variables

Background variables included maternal age at birth, parity, smoking in pregnancy, body mass index (BMI), diabetes, chronic hypertension and ART. Outcome variables comprised pre-eclampsia, gestational hypertension, preterm birth and z-score for birth weight denoted as SGA and large for gestational age (LGA). These variables were derived from MBRN.

Patient group variables containing educational status and disease-specific information, including disease activity assessment, medication use and patient-reported health-related quality of life (HRQoL), were retrieved from RevNatus.

Disease activity assessment

The Disease Activity Score with C reactive protein (DAS28-CRP-3) was used to assess disease activity during pregnancy in women with PsA, as there is no disease activity assessment validated for PsA in pregnancy.¹⁶ DAS28-CRP-3 is a compound score comprising tender and swollen joints among 28 joints and CRP.¹⁷ It has been validated for use in pregnant women with rheumatoid arthritis (RA).¹⁸ The European Alliance of Associations for Rheumatology has defined four disease categories ranging from 0 to 10: remission (<2.6), low disease activity (≥2.6 but ≤3.2), moderate disease activity (>3.2 but ≤5.1) and high disease activity (>5.1).¹⁹ Inactive PsA was defined as DAS28-CRP-3<2.6 in 2nd and/or 3rd trimester and active PsA as DAS28-CRP-3≥2.6 in 2nd and/or 3rd trimester.

Health-related quality of life

RAND-36²⁰ and RAND-12²¹ are two questionnaires covering eight HRQoL domains, the latter using 12 of the items from RAND-36. The items are scored 0–100, with a higher score indicating a higher HRQoL. In RevNatus, RAND-36 was used until 2016 and RAND-12 from 2019. The domains bodily pain, physical function, general health, mental health and vitality were selected in the present study.

Outcomes

Gestational hypertension was defined as new onset elevated blood pressure ≥140/≥90 mm Hg⁵ after gestational week 20 without proteinuria. Pre-eclampsia was defined as gestational hypertension accompanied by proteinuria (≥1+ on urine dipstick with a minimum of two measurements, or by urine protein ≥0.3 g/24 hours or total protein/creatinine ratio >0.3). Severe manifestations of pre-eclampsia, including eclampsia, HELLP syndrome (haemolysis, elevated liver enzymes and low platelets) and severe gestational hypertension were included in the reported cases of pre-eclampsia.

Preterm birth was defined as birth <37 weeks of gestation and late preterm birth as birth ≥34 weeks and <37 gestational weeks.

SGA was defined as fetal weight <10th percentile or <2.5th percentile. LGA was defined as fetal weight >90th percentile or >97.5th percentile. SGA and LGA were calculated by using the z-score for birth weight based on birth weight, gestational age and gender.

Statistical analysis

We reported descriptive statistics for the inactive PsA group, the active PsA group and population controls, as well as disease-related characteristics of the inactive PsA group and active PsA group. Pairwise group comparisons of the total PsA group with population controls, the inactive PsA group with population controls, the active PsA group with population controls and the active PsA group with the inactive PsA group were performed using an independent samples t-test for continuous variables and the Pearson χ² test, the Fisher’s exact test, the unconditional z-pooled test²² or the Fisher’s mid-p test²³ for dichotomous variables.

Proportions and risk differences for the main outcomes: pre-eclampsia, gestational hypertension, preterm birth, late preterm birth, SGA and LGA were calculated comparing the total PsA group, the inactive PsA group and active PsA group one at a time with population controls. A subgroup analysis of obese women was also performed as described for the total group. Missing data were handled using available case analysis, that is, each analysis includes the women with available data on the included variables. We calculated 95% CIs for risk differences using Newcombe’s method.²⁴ Two-sided p<0.05 was considered to represent statistical significance, and 95% CIs are reported where relevant. However, due to multiple hypotheses, p values between 0.01 and 0.05 should be interpreted with caution.²⁵ The statistical analyses were performed using IBM SPSS Statistics for Windows, V.29.0.1, STATA MP V.18, https://www4.stat.ncsu.edu/~boos/exact/ and R V.4.2.0 (2022-04-22 ucrt).

Results

There were 126 singleton births in women with PsA registered in RevNatus during this period. Five women were not possible to link to MBRN because of incomplete birth date information,²⁶ resulting in 121 singleton births. Disease activity assessment was missing in above 60% of 1st trimester visits, while the majority of women had disease activity assessment in 3rd trimester. 92 of the disease activity assessments were in 3rd trimester, while 17 were in 2nd trimester, in cases where disease activity assessment was missing in 3rd trimester. There was missing information on disease activity in the 3rd and 2nd trimesters in 12/121 (10%) of the pregnancies, leaving 109 singleton births to constitute the cases. There were 75 women with inactive PsA and 34 women with active PsA. In women with active PsA, 13 women had low disease activity and 21 had moderate disease activity. There were none with high disease activity.

There were 575 798 singleton births serving as population controls, after excluding singleton births in women diagnosed with inflammatory rheumatic diseases according to ICD-10, see online supplemental table S1.

Table 1 outlines characteristics of population controls and patient groups with inactive PsA and active PsA. A higher proportion of women with inactive PsA were ≥35 years (24/75, 32%) compared with population controls (115 077/575 798, 20%). Both women with inactive PsA and active PsA had statistically significantly higher proportions of obesity (13/50, 26.0% and 10/24, 41.7%) and ART (12/75, 16.0% and 5/34, 14.7%) compared with population controls (49167/399 708, 12.3% and 20121/575 798, 3.5%, respectively). There were no statistically significant differences between either of the patient groups compared with population controls concerning parity, smoking, diabetes or chronic hypertension.

Table 1. Characteristics of population controls and women with inactive or active psoriatic arthritis, reported as n (%) unless specified as mean (SD).

Characteristic	Population controls	Inactive PsA DAS28<2.6*	P value†	Active PsA DAS28≥2.6*	P value†
Singleton births 2010–2019	575 798	75		34
Maternal age (years), mean (SD)	30.6 (5.1)	32.3 (5.0)	0.004	31.7 (4.5)	0.23
<35	460 720 (80.0)	51 (68.0)	0.014	26 (76.5)	0.76
≥35	115 077 (20.0)	24 (32.0)		8 (23.5)
Missing	0	0		0
Parity
No children	244 354 (42.4)	30 (40.0)	0.76	13 (38.2)	0.75
≥1 child	331 444 (57.6)	45 (60.0)		21 (61.8)
Missing	0	0		0
Smoking in pregnancy	34 237 (6.7)	4 (5.6)	1.0‡	5 (16.1)	0.054‡
Missing	67 663	3		3
BMI first trim, mean (SD)	24.4 (4.8)	25.9 (5.6)	0.026	29.4 (5.1)	<0.001
≥25.0	138 056 (34.5)	24 (48.0)	0.064	20 (83.3)	<0.001
≥30.0	49 167 (12.3)	13 (26.0)	0.006	10 (41.7)	<0.001‡
Missing	176 090	25		10
Diabetes§	25 924 (4.5)	4 (5.3)	0.58‡	3 (8.8)	0.20‡
Missing	0	0		0
Hypertension, chronic	3154 (0.5)	1 (1.3)	0.34‡	0	1.0‡
Missing	0	0		0
ART	20 121 (3.5)	12 (16.0)	<0.001‡	5 (14.7)	0.006‡
Missing	0	0		0

Bold values are statistically significant (p< 0.05)

^*In 2nd or 3rd trimester.

^†P value for patient group compared to population controls.

^‡Fisher’s exact test.

^§Pregestational or gestational.

ART, assisted reproductive technology; BMI, body mass index; DAS28, Disease Activity Score based on 28 joints; PsA, psoriatic arthritis.

The disease-related characteristics of the two patient groups are described in table 2. Women with active disease had statistically significantly lower educational levels and shorter mean disease duration compared with women with inactive disease. A higher proportion of women with inactive PsA had erosive disease compared with women with active PsA. There were no significant differences in the use of medication between the two groups. Concerning prednisolone, the dosage was not available. Around 13% reported prednisolone use in the 1st trimester, dropping to around 10% in the 2nd and 3rd trimester with no significant differences between women with active and inactive PsA (data not shown). More than half of the women with PsA reported use of methotrexate before pregnancy. About 45% reported use of tumour necrosis factor inhibitors (TNFi) before pregnancy, while approximately 1/4 reported use of TNFi in pregnancy.

Table 2. Clinical characteristics of PsA grouped according to disease activity in 2nd or 3rd trimester, reported as n (%) unless specified as mean (SD).

Characteristic	Inactive PsA DAS28<2.6	Active PsA DAS28≥2.6	P value*
Singleton births 2010–2019	75	34
Educational level
Low (10–13 years)	24 (32.4)	23 (69.7)	<0.001
High (≥14 years)	50 (67.6)	10 (29.4)
Missing	1	1
Disease criteria fulfilled†	65 (94.2)	30 (100)	0.22‡
Missing	6	4
Disease duration, mean (SD)	7.3 (6.1)	5.0 (3.3)	0.024
Missing	13	6
Erosive disease	11 (17.2)	3 (10.7)	0.45‡
Missing	11	6
Prednisolone in pregnancy	12 (18.2)	6 (21.4)	0.74‡
Missing	9	6
SSZ in pregnancy	6 (8.2)	4 (12.5)	0.54‡
Missing	2	2
TNFi in pregnancy	18 (27.3)	7 (23.3)	0.70‡
Missing	9	4
MTX before pregnancy	32 (56.1)	14 (53.8)	0.87‡
Missing	18	8
Prednisolone before pregnancy	8 (14.0)	4 (15.4)	0.90‡
Missing	18	8
SSZ before pregnant	8 (14.0)	4 (15.4)	0.90‡
Missing	18	8
TNFi before pregnant	25 (43.9)	12 (46.2)	0.87‡
Missing	18	8
Bodily pain, mean (SD)	62.6 (26.1)	46.6 (26.9)	0.010
Missing	14	7
Physical function, mean (SD)	69.8 (25.8)	46.4 (22.1)	<0.001
Missing	13	7
General health, mean (SD)	58.2 (17.6)	46.9 (18.9)	0.004
Missing	13	7
Mental health, mean (SD)	79.0 (14.0)	73.0 (17.7)	0.096
Missing	15	7
Vitality, mean (SD)	42.3 (23.3)	33.9 (17.3)	0.097
Missing	15	7

^*P value for active compared to inactive disease.

^†The classification criteria for psoriatic arthritis (CASPAR).

^‡The unconditional z-pooled test.

DAS28, Disease Activity Score based on 28 joints; MTX, methotrexate; PsA, psoriatic arthritis; SSZ, sulfasalazine; TNFi, tumour necrosis factor inhibitor.

Women with active PsA reported higher levels of bodily pain (ie, lower scores), lower physical function and poorer general health compared with women with inactive PsA. Both groups reported good mental health, but low levels of vitality.

Main outcomes in the total group of women with PsA were not significantly different from population controls, see table 3.

Table 3. Hypertensive disorders, preterm birth and abnormal fetal growth, expressed as proportions and risk differences.

Outcome	Population controls	PsA	Risk difference (95% CI)	P value*
Singleton births 2010–2019	575 798	121
Pre-eclampsia†	15 162 (2.6)	5 (4.1)	1.4 (−0.9 to 6.6)	0.20
Missing	0	0
Gestational HT	9644 (1.7)	4 (3.3)	1.6 (−0.4 to 6.5)	0.10
Missing	0	0
Preterm birth	27 955 (4.9)	8 (6.6)	1.7 (−1.5 to 7.6)	0.34
Missing	0	0
Late preterm birth	19 919 (3.5)	5 (4.2)	0.7 (−1.7 to 6.0)	0.53
Missing	0	0
SGA, 10 percentile	47 146 (8.3)	7 (5.8)	−2.4 (−5.4 to 3.3)	0.36
Missing	5674	1
LGA, 90 percentile	48 794 (8.6)	15 (12.5)	3.9 (−0.8 to 11.0)	0.12
Missing	5674	1
SGA, 2.5 percentile	9106 (1.6)	2 (1.7)	0.1 (−1.1 to 4.3)	0.58
Missing	5674	1
LGA, 97.5 percentile	13 547 (2.4)	2 (1.7)	−0.7 (−1.9 to 3.5)	0.88
Missing	5674	1

^*Fisher’s mid-p test.

^†Pre-eclampsia, eclampsia and haemolysis, elevated liver enzymes and low platelets syndrome.

HT, hypertension; LGA, large for gestational age; PsA, psoriatic arthritis; SGA, small for gestational age.

Women With active PsA, there was a higher proportion of pre-eclampsia compared with population controls (8.8% vs 2.6%), with a risk difference of 6.1 (95% CI 0.3 to 20.3). The proportion of gestational hypertension was also higher (5.9% vs 1.7%), with a risk difference of 4.2 (95% CI 0.0 to 17.4).

There were similar proportions of hypertensive disorders in women with inactive PsA compared with population controls. Preterm birth did not occur more frequently in either group of women with PsA compared with population controls (see table 4).

Table 4. Hypertensive disorders and preterm birth in population controls and women with active or inactive psoriatic arthritis, expressed as proportions and risk differences.

	Total	Pre-eclampsia*	%	Risk difference (95% CI)	P value†
Population controls	575 798	15 162	2.6
Active PsA‡	34	3	8.8	6.1 (0.3 to 20.3)	0.036§
Inactive PsA‡	75	1	1.3	−1.4 (−2.5 to 4.5)	0.59§
	Total	Gestational HT	%	Risk difference	P value
Population controls	575 798	9644	1.7
Active PsA‡	34	2	5.9	4.2 (−0.0 to 17.4)	0.065§
Inactive PsA‡	75	2	2.7	1.0 (−0.9 to 7.5)	0.24§
	Total	Preterm birth (< 37 weeks)	%	Risk difference (95% CI)	P value
Population controls	575 798	27 955	4.9
Active PsA‡	34	1	2.9	−2.0 (−4.4 to 10.0)	0.84§
Inactive PsA‡	75	6	8.0	3.1 (−1.2 to 11.5)	0.14§†
		Late preterm birth (34 to <37 weeks)
Population controls	575 798	19 919	3.5
Active PsA‡	34	1	2.9	−0.6 (−3.0 to 11.4)	0.82§
Inactive PsA‡	75	4	5.5	1.9 (−1.4 to 9.7)	0.26§

^*Pre-eclampsia, eclampsia and haemolysis, elevated liver enzymes and low platelets syndrome.

^†Statistical significance p<0.01 to adjust for multiplicity.

^‡In 2nd or 3rd trimester.

^§Fisher’s mid-p test.

PsA, psoriatic arthritis.

Two events (a case with pre-eclampsia and a case with preterm birth) were not captured in the cases, as these events were in women with missing disease activity assessment.

There were no differences in small for gestational weight (SGA) or large for gestational weight (LGA) when comparing the two disease groups with population controls (see table 5).

Table 5. Small and large for gestational age in offspring of population controls and women with active or inactive psoriatic arthritis, expressed as proportions and risk differences.

	Total	SGA (<−1.3), 10 perc	%	Risk difference (95% CI)	P value‡
Population controls	570 124	47 146	8.3
Active PsA*	33	1	2.9	−5.2 (−7.7 to 7.0)	0.43†
Inactive PsA*	75	5	6.7	−1.6 (−5.4 to 6.4)	0.75
	Total	LGA (>1.3), 90 perc	%	Risk difference (95% CI)	P value
Population controls	570 124	48 794	8.6
Active PsA*	33	5	15.2	6.6 (−1.9 to 22.4)	0.16†
Inactive PsA*	75	8	10.7	2.1 (−3.1 to 11.1)	0.47
	Total	SGA (<−2.0), 2.5 perc	%	Risk difference	P value
Population controls	570 124	9106	1.6
Active PsA*	33	1	3.0	1.4 (−1.1 to 13.7)	0.25†
Inactive PsA*	75	0	–	−1.6 (−1.6 to 3.3)	0.49†
	Total	LGA (>2.0), 97.5 perc	%	Risk difference	P value
Population controls	570 124	13 547	2.4
Active PsA*	33	1	3.0	0.7 (−1.8 to 12.9)	0.37†
Inactive PsA*	75	1	1.3	−1.0 (−2.1 to 4.8)	0.85†

^*In 2nd or 3rd trimester.

^†Fisher’s mid-p test.

^‡Statistical significance p<0.01 to adjust for multiplicity.

LGA, large for gestational age; PsA, psoriatic arthritis; SGA, small for gestational age.

In a subanalysis of obese women (BMI≥30.0), women with active PsA had a higher proportion of pre-eclampsia compared with population controls (20.0% vs 5.4%), with a risk difference of 14.6 (CI 0.2 to 45.6), though not of statistical significance (p=0.057). There were no differences in inactive PsA compared with population controls. We found no differences for gestational hypertension, preterm birth and SGA among PsA women and population controls, see online supplemental tables S2 and S3.

Discussion

In this prospective population-based study, there was a higher incidence of hypertensive disorders in women with active PsA in pregnancy, but not in women with inactive PsA in pregnancy compared with population controls. These results are relevant for clinicians and patients, especially women with active disease, as disease activity may be modified before and better controlled during pregnancy, thereby possibly improving pregnancy outcomes. Our results are in accordance with one earlier study.¹³ Another study also reported an increased risk for pre-eclampsia in PsA.²⁷ Due to high missing data on disease activity parameters, they were not able to conclude, though they discussed the possible impact of active disease.²⁷ When not taking disease activity into account (table 3), we did not find an increased risk for hypertensive disorders, as in other studies without information about disease activity.28,31

Hypertensive disorders may originate early or late in pregnancy, with possible development of clinical manifestations at varying gestational ages and with varying severity. We investigated the influence of active disease in the second part of pregnancy, but we do not know how active disease early in pregnancy might impact the evolution of hypertensive disorders. In our population, no cases with pre-eclampsia or gestational hypertension occurred preterm. There was one case of preterm pre-eclampsia in a woman without registered disease activity (result not shown). Among the cases of pre-eclampsia in women with PsA, there were three serious cases of pre-eclampsia (one preterm, one eclampsia and one HELLP syndrome). Most cases of pre-eclampsia in the control group were mild and moderate. Even though small numbers, this indicates that PsA, and especially in women with active PsA, may influence the biology and the placenta in the development of pre-eclampsia. Active rheumatic disease during pregnancy is a maternal factor adding to the total burden of inflammation. It may facilitate the progress to clinically manifest hypertensive disorders.^{7 32}

We did not find an increased risk for either SGA or LGA in women with active PsA or inactive PsA. This is in contrast to a Swedish population-based study.^{30 31} This may be due to the late onset of hypertensive disorders in our patient population, not affecting offspring growth, and is reassuring.

We found that the occurrence of preterm birth was similar in women with PsA and population controls, irrespective of disease activity. The highest proportion of preterm birth was noted in women with inactive PsA, thus no signals of impact of disease activity on preterm birth. For the entire PsA group (table 3), the proportions of preterm birth were 6.6% compared with 4.9% in population controls. Though imprecise point estimates, the numbers are indicative of an increased relative risk, though lower than in previous studies.1328,31 Two recent prospective population-based studies with a probable overlap of cases reported an increased risk for preterm and moderately preterm birth (32–36 weeks) in women with PsA.^{30 31} In addition, two large prospective studies with mixed patient populations including women with PsA found an increased risk of preterm birth and late preterm birth (32–36 weeks).^{28 29} Less than 5% of the patients included in these studies were diagnosed with PsA and may therefore not reflect the risk in women with PsA. The prevalence of preterm birth varies worldwide and is lowest in the Nordic countries.^{9 10} This is important to consider when comparing our results to studies performed elsewhere.

Pre-eclampsia can serve as a competing risk for preterm birth, as pre-eclampsia is associated with mortality. Pre-eclampsia mortality may be maternal or neonatal. There has not been any maternal mortality due to pre-eclampsia in Norway since 2012. Neonatal pre-eclampsia mortality is only seen in very preterm neonates. We had only one preterm pre-eclampsia, in a woman without registered disease activity. The total number of pre-eclampsia (five cases) and preterm birth (eight cases) was too small to investigate this issue.

In our patient population, the women with active PsA mainly had low disease activity. This is in accordance with other recent studies describing disease activity in pregnant women with PsA.1333,35 The adherence to international guidelines, as well as the systematic follow-up of patients included in the RevNatus registry according to national recommendations, has the potential to contribute to better disease control.

A large proportion of women with PsA was obese (BMI≥30.0). In the subanalysis of obese women, we found that active PsA increased the risk for hypertensive disorders, whereas inactive PsA did not, when comparing to population controls (see online supplemental tables S2 and S3), though not of statistical significance. The small sample size might lead to an underpowered analysis. However, these results are in accordance with the main results and support the contribution of active PsA in the development of pre-eclampsia.

In women with active PsA, ART was performed in 14.7% and in women with inactive PsA in 16.0%, substantially more often than in population controls (3.5%). Higher maternal age, obesity, active PsA and medication use may all contribute to fertility problems.

More than half of the women reported use of methotrexate before conception. Methotrexate is effective for treating PsA and skin psoriasis, but it is a potential teratogenic drug and should not be used when planning a pregnancy. In women with active disease, the need for effective medication might delay conception, resulting in higher maternal age and reduced fertility. Independent of medication, having a chronic rheumatic disease may hamper the desire to start a family. In 18/74 (24.3%) of the women with inactive PsA and 8/34 (23.5%) of the women with active PsA, it was reported more than 12 months to conception ever, indicating infertility (data not shown).

Treatment exposure, and in particular corticosteroid use, is discussed as a risk factor for hypertensive disorders. In our experience, most women do not use corticosteroids as monotherapy, but in combination with disease-specific treatment to attain better control of the disease in cases of flare. In a recent Nordic study, the risk estimates for pre-eclampsia when stratified by treatment during pregnancy were in line with the overall increased risk of pre-eclampsia in women with PsA, though not reaching statistical significance. In this study, medication was grouped as monotherapy (either corticosteroids, conventional synthetic disease-modifying antirheumatic drug or biologic disease-modifying antirheumatic drug) or combination therapy (three groups). Untreated women had similar risk estimates indicating that medication per se did not increase this risk.²⁷

The number of events in our sample was generally too small to include other variables as adjustment variables, and hence also too small for matched analysis.

Strengths of the study are the prospective population-based design and large patient groups. The population is relatively homogenous, and standardised pregnancy care free of charge is offered to all pregnant women living in Norway.

In Norway, most patients with diagnosed PsA are followed by a rheumatologist either regularly or on demand. All the local outpatient rheumatology clinics include eligible patients in RevNatus, securing access to the register and participation from all geographical regions of Norway. A weakness is unmeasured factors such as long distance to the local clinic, low income or other factors causing a woman not to participate, possibly generating selection bias in the cases.

A limitation of the study is the disease activity assessment. We used DAS28-CRP to assess disease activity, as this is an assessment validated for use in pregnancy in arthritic disease.¹⁸ It is likely that disease activity is under-reported, as PsA is a heterogenic disease that among other manifestations may contain enthesitis, dactylitis as well as inflammation of the spine. This is not registered in DAS28-CRP. Disease activity in PsA was not introduced as a variable in Revnatus until 2016/2017 and could therefore not be used.

For this reason, the potential influence of different PsA phenotypes cannot be assessed. On the other hand, CRP may be elevated in obesity, possibly over-reporting disease activity in these cases. Another limitation is that other studies assessing disease activity have used different indices, making comparison difficult.^{12 13 27}

Another limitation is missing data on BMI and smoking, two variables that possibly influence outcomes. Maternal BMI was missing in 31%. This variable was gradually introduced as the MBRN changed from paper to electronic registration, with registration of BMI in 41% in 2010, increasing to 91% in 2019. As an exploratory analysis, we also carried out analyses adjusting for overweight or obesity. The effect of active PsA is somewhat reduced, partly due to missing BMI data. It still is positive, but with wider CIs and nonsignificant p values due to reduced sample size and adding another covariate (data not shown). Smoking habits were missing in 18% of the population control, with lower missing proportions in the cases. This variable is not mandatory to report, and it is possible that women who smoke are more prone to not reporting smoking habits than women who do not smoke. We reported a higher proportion of smokers in the active disease group, though not of statistical significance.

Overall, the small outcome numbers resulting in reduced power demand a cautious interpretation of the results. Our hypotheses cannot be answered, but our results show tendencies that need to be confirmed in larger cohorts.

Conclusion

The present study indicates that active PsA in pregnancy may contribute to the risk of developing hypertensive disorders during pregnancy. It is encouraging that the risk for preterm birth and growth restriction was not elevated. To improve outcomes further, there is a continuous need for preconception counselling and monitoring during pregnancy to address general risk factors like maternal age and BMI, and to achieve and maintain inactive disease throughout pregnancy.

Data availability statement

Data may be obtained from a third party and are not publicly available.