The Association of Lupus Nephritis with Adverse Pregnancy Outcomes among Women with Lupus in North America

Abstract

Objectives:

We evaluated the association of lupus nephritis (LN) and adverse pregnancy outcomes in prospective cohorts of pregnant women with SLE (Systemic Lupus Erythematosus).

Methods:

We conducted a patient-level pooled analysis of data from 3 cohorts of pregnant women with SLE. Pooled logistic regression models were used to evaluate the association of LN and adverse pregnancy outcomes. Odds ratios and 95% confidence intervals were calculated using a fixed effect model by enrolling cohort.

Results:

The pooled cohort included 393 women who received care at clinics in the United States and Canada from 1995 to 2015. There were 144(37%) women with a history of LN. Compared to women without LN, those with LN had higher odds of fetal loss (OR: 1.90; 95% CI: 1.01, 3.56) and preeclampsia (OR: 2.04; 95% CI: 1.01, 4.13). Among the 31 women with active nephritis (defined as urine protein ≥ 0.5gram/24hr) there was a higher odds of poor pregnancy outcome (OR: 3.08; 95% CI: 1.31, 7.23) and fetal loss (OR: 6.29; 95% CI: 2.52, 15.70) compared to women without LN.

Conclusions:

In this pooled cohort of women with SLE, a history of LN was associated with fetal loss and preeclampsia. Active nephritis was associated with poor pregnancy outcome and fetal loss.

Introduction

Compared to the general population, pregnant women with systemic lupus erythematosus (SLE) are at increased risk for adverse pregnancy outcomes, including fetal and maternal mortality, stroke, venous thromboembolism, preterm birth, Cesarean section, low birth weight, congenital defects, and preeclampsia.^1–8 Lupus nephritis (LN), a common manifestation of SLE, is also an important risk factor for adverse pregnancy outcomes.⁹ Women with LN have higher rates of preeclampsia, gestational hypertension, preterm delivery, neonatal death, small for gestational age, and intrauterine growth restriction than women with SLE without LN. ^{6,10,11,12,13} Active LN has also been associated with hypertensive disorders of pregnancy, higher risk for SLE flare, with preterm birth, severe infections.^{3,11,14 14,15}

While previous studies have established the association of active LN and adverse pregnancy outcomes, many of these studies were conducted at single centers with homogenous patient populations.^11,16–19 It is well documented that the there is a higher incidence of lupus nephritis among Black, Hispanic, and Asian patients.²⁰ Black and Hispanic patients with LN often have greater disease severity and are at higher risk for End Stage Renal Disease (ESRD) requiring renal replacement therapy.^21,22 A study of claims data suggested that pregnant Black and Hispanic women may have double the odds of preeclampsia, preterm birth, and acute renal failure compared to White women. ²³

In this study we evaluated the association between LN and adverse pregnancy outcomes in a pooled analysis of three prospective cohorts of pregnant women with SLE in North America.

Methods

We searched for prospective cohorts of pregnant women with a diagnosis of SLE using MEDLINE (via PubMed), Embase (via Elsevier) and the Cochrane Database of Systematic Reviews between the years 1995 and 2015. We identified 46 publications describing prospective cohorts of pregnant women with SLE (Figure 1). After excluding duplicate cohorts and cohorts that did not collect relevant data on pregnancy outcomes, we attempted to contact 27 principal investigators (PI) to request patient-level data. We obtained patient-level data from 7 prospective cohorts of pregnant women with SLE. This analysis was restricted to cohorts in North America: Duke University, Johns Hopkins University, and the University of Toronto. We included only participants who established care prior to 20 weeks’ gestation to minimize selection bias based on referral for complicated pregnancy, and randomly selected only one pregnancy per participant to include in the pooled analysis. To increase the generalizability of our results, we included all eligible women regardless of SLE disease activity and severity, kidney disease severity and immunomodulatory therapy use.

Figure 1:

Flow Diagram

Measures

Exposures of interest included race and lupus nephritis. Maternal race was based on self-report and stratified as White, Black, Asian, Hispanic, and Other. Because there were few patients in the Asian, Hispanic, and Other racial and ethnic groups, we excluded them from our analysis to focus on the comparison of Black and White women (Figure 1). LN was defined by a documented history of nephritis and was further classified as active or inactive based on the presence or absence of proteinuria ≥ 0.5gram/24hr prior to 20 weeks’ gestation. Adverse pregnancy outcomes of interest included fetal loss (defined as loss at any time during the gestational period), preterm birth (birth before 37 weeks’ gestation), preeclampsia (based on physician diagnosis), and a composite poor pregnancy outcome (outcome of fetal loss, preeclampsia, or preterm birth).

Statistical Analyses

Summary statistics are presented as mean ± SD for continuous data and percentages for categorical data. Categorical variables were compared using chi-squared test and continuous variables by ANOVA. In separate analyses, LN was considered as a 2-level categorical variable (any history, yes or no) and as a 3-level categorical variable (no history, inactive LN, active LN). Logistic regression models with fixed effects by medical center estimated the association of LN with adverse pregnancy outcomes, adjusted for maternal age and race. All analyses were conducted in SAS 9.4 (Cary, North Carolina). The Duke Institutional Review Board determined this study was exempt (Pro00066961).

Results

The pooled cohort included 393 women, including 148 women (38%) of Black race and 245 women (62%) of White race. The mean age at delivery was 30.6 years (SD 5.4). Black women in the pooled cohort were younger with a mean age of 28.9 (SD 5.7). One hundred and forty-four women (37%) had a history of lupus nephritis, with only 31 (8%) women meeting the study criteria for active nephritis. Black women were more likely to have a history of LN than white women. The majority of women were prescribed immunomodulatory medications during pregnancy, with 266 women (70%) prescribed hydroxychloroquine, 190 women (50%) prescribed corticosteroids, and 61 (16%) prescribed azathioprine and 214 (58%) received aspirin. Almost half of women with active LN received corticosteroids, azathioprine and aspirin (S1).

Adverse pregnancy outcomes were common, with 152 (43%) of women experiencing a poor pregnancy outcome (Table 2). These included 50 (13%) women with fetal loss, 88 women (26%) with preterm birth, and 44 (14%) women with preeclampsia. Preterm birth, preeclampsia, and poor pregnancy outcome, though not fetal death, were more common among Black than White women. Women with a history of LN had higher rates of adverse pregnancy outcomes than women without this history. In logistic regression models adjusted for maternal age and race, women with a history of LN, had almost twice the odds of fetal loss, preterm birth, preeclampsia, and poor pregnancy outcomes (Table 3). Much of this increase in adverse outcomes, however, was due to very high rates of complications in women with active LN, not inactive LN (Figure 2, Table 3). When adjusted for maternal age and race, women with inactive LN did not have more frequent pregnancy complications. Women with active LN, however, had six times the odds of fetal loss (OR: 6.29; 95% CI: 2.52, 15.70) and three times the odds of poor pregnancy outcome (OR: 3.08; 95% CI: 1.31, 7.23) compared to women without a history of LN (Table 3).

Table 2:

Overall Pregnancy Outcomes of Included Cohorts by Race

	Overall	Black	White
	n=393	n=148	n=245	p-value
Maternal Age at Delivery	30.6 (5.4)	28.8 (5.7)	31.7 (4.9)	<0.0001
Duration of SLE, years (n=388)	7.4 (5.3)	6.8 (4.9)	7.7 (5.5)	0.09
Antiphospholipid Syndrome	37/344 (11%)	12/141 (9%)	25/203 (12%)	0.3
High Disease Activity *	122/392 (31%)	58/147 (39%)	64/245 (26%)	<0.005
History of Lupus Nephritis
No history	249 (63%)	76 (51%)	173 (71%)	<0.0001
Inactive nephritis	112 (28%)	47 (32%)	65 (27%)
Active nephritis	31 (8%)	25 (17%)	6 (2%)
History of nephritis but unknown activity during pregnancy	1 (0.3%)	0 (0%)	1 (0.4%)
Pregnancy Outcomes:
Poor Pregnancy Outcome	152/351 (43%)	73/136 (54%)	79/215 (37%)	0.002
Fetal loss	50 (13%)	21 (14%)	29 (12%)	0.5
Preterm birth	88/337 (26%)	43/126 (34%)	45/211 (21%)	0.01
Preeclampsia	44/310 (14%)	24/118 (20%)	20/192 (10%)	0.02

^*Defined as Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) >4 ever during pregnancy

Table 3.

Association of history of lupus, active and inactive lupus nephritis with pregnancy outcomes, adjusted for age and maternal race.

OR (95% CI)	No History of LN	History of LN	No History of LN	Inactive LN	Active LN
Poor Pregnancy Outcome	1.0 (ref)	1.59 (1.00, 2.51)	1.0 (ref)	1.36 (0.83, 2.22)	3.08 (1.31, 7.23)
Fetal loss	1.0 (ref)	1.90 (1.01, 3.56)	1.0 (ref)	1.18 (0.56, 2.50)	6.29 (2.52, 15.70)
Preterm birth	1.0 (ref)	1.52 (0.90, 2.57)	1.0 (ref)	1.42 (0.81, 2.48)	2.28 (0.84, 6.20)
Preeclampsia	1.0 (ref)	2.04 (1.01, 4.13)	1.0 (ref)	1.86 (0.88, 3.94)	3.10 (0.98, 9.79)

Figure 2:

Incremental increase in adverse outcomes based on lupus nephritis diagnosis and disease activity

Discussion

We found that women with active LN had higher rates of poor pregnancy outcome, fetal loss, preterm birth and preeclampsia compared to women without a history of LN and inactive LN. In multivariable analysis, women with a history of LN had higher odds of preeclampsia and fetal loss, compared to women without nephritis. We suspect this finding was driven largely by the higher odds of these outcomes in women with active LN. Compared to women without nephritis, women with active LN had higher odds of poor pregnancy outcome and fetal loss. Although not statistically significant in this smaller subgroup, preeclampsia and preterm birth were also more common in women with active LN compared to women without nephritis. Our results support previous studies demonstrating active nephritis is a risk factor for adverse pregnancy outcomes among women with SLE.^19,3

We found high rates of adverse pregnancy outcomes. We suspect the high rates of adverse pregnancy outcomes were due to the broad inclusion of women into lupus pregnancy cohorts regardless of SLE and LN disease activity, immunosuppressant use, medical comorbidities and kidney function. For instance, over one-third of women in the combined cohort had high disease activity and not all women with active LN were prescribed azathioprine and corticosteroids during pregnancy. Almost a third of women in our cohort did not receive hydroxychloroquine, which has been associated with reduced risk of adverse pregnancy outcomes such as preterm birth and hypertensive disorders of pregnancy.^24,25 The low rates of azathioprine and corticosteroids among women with active LN may explain the higher frequencies of adverse pregnancy outcomes compared to other studies with more selective criteria, such as the PROMISSE study.⁵

We also observed that Black women had higher rates of high disease activity and active LN compared to White women (Table 2). Black women also had higher frequencies of poor pregnancy outcome, preterm birth, and preeclampsia. While the increase in prior and active LN in Black women contributes to these poor outcomes, even without a history of LN, it appears that Black women with SLE have higher rates of preterm birth and preeclampsia than white women with SLE. Our results are consistent with previous studies that demonstrate a disproportionate amount of Black and Hispanic women with LN and higher disease severity compared to White women. ^26,27 A strength of this study is the high proportion of Black pregnant women with SLE, a group that is often under-represented in clinical research.^21,28

Few studies have evaluated racial disparities in pregnancy outcomes among women with SLE and LN. The LUMINA study, a multiethnic study of patients with SLE, revealed a 5-fold odds of adverse pregnancy outcomes in women with kidney involvement prior to pregnancy. In the PROMISSE study, Black women with antiphospholipid antibodies had 10-fold the odds of adverse pregnancy outcomes, however no racial differences were observed in adverse pregnancy outcomes in women without antiphospholipid antibodies, with adjustments for education, community income, and community education. Our study provides additional insight regarding race and adverse pregnancy outcomes, though we do not have complete data on the other social determinants of health that could be modifying the risk posed by race. Further studies are needed in non-white populations to understand the underlying causes for increased poor pregnancy outcomes in women of color and how they can be addressed.

Our study had several limitations. We did not have complete kidney biopsy data for all pregnancies to determine WHO class of LN. It has previously been established that pregnant women with World Health Organization (WHO) proliferative classes (III and IV) of LN have higher rates of preeclampsia, infection, hemorrhage and SLE flares. Advanced kidney disease of any etiology has been associated with adverse pregnancy outcomes.²⁹ Carmona et al. also found babies born to women with WHO Class III and IV disease had lower birthweights compared to women with Class II and V disease and women without nephritis.³⁰ We did not have preconception data available for all participants including markers of disease activity such as anti-double-stranded DNA, complements, and proteinuria. Information on preexisting medical comorbidities and family medical history was not available. Although we provide information on immunomodulatory medications, we did not have complete data on other medications that could influence pregnancy outcomes, including calcium and low molecular weight heparin.

In this study we offer a real-world glimpse of pregnancies in women with SLE and LN. Our study also represents one of few multicenter studies evaluating adverse pregnancy outcomes in women with LN,^31,32 with a high representation of Black women. Much of the literature on pregnancy outcomes in women with SLE and LN are from single centers with limited generalizability for clinical practice. Further scholarship on prescribing practices, medication adherence, and racial disparities among pregnant women with LN is needed to develop interventions targeted at improving pregnancy outcomes.

Table 1:

Baseline Characteristics Included Cohorts by Medical Center

	Pooled Cohort	Duke	Hopkins	Toronto
Included pregnancies (1/woman)	n=393	n=96	n=248	n=49
Maternal Race (self-reported)
Black	148 (38%)	56 (58%)	85 (34%)	7 (14%)
White	245 (62%)	40 (42%)	163 (66%)	42 (86%)
Maternal Age at Delivery	30.6 (5.4)	29.7 (5.5)	30.8 (5.2)	31.4 (5.6)
SLE History:
Duration of SLE pre-pregnancy, years (n=388)	7.4 (5.3)	7.9 (5.1)	6.9 (5.2)	8.7 (5.7)
Antiphospholipid Syndrome (n=344)	37 (11%)	4 (4%)	33 (13%)	NR
History of Lupus Nephritis	144 (37%)	29 (30%)	106 (43%)	9 (18%)
Medications in Pregnancy
Hydroxychloroquine	266/379 (70%)	86 (90%)	166 (67%)	14/35 (40%)
Prednisone	190/379 (50%)	49 (51%)	120 (48%)	21/35 (60%)
Azathioprine	61/380 (16%)	23 (24%)	28 (11%)	10/36 (28%)

NR: Not reported