Balancing stringency and feasibility: comparative value of disease-activity measures to predict pregnancy outcomes in systemic lupus erythematosus

Abstract

Objective

To compare preconception disease-activity indices—systemic lupus erythematosus Disease Activity Score low disease activity (SLE-DAS LDA), lupus low disease activity state (LLDAS) and SLE-DAS remission—with Definitions of Remission in SLE (DORIS) remission in predicting adverse maternal and fetal outcomes among pregnant women with SLE.

Methods

This retrospective cohort study included 202 pregnancies in 196 women with SLE managed at Shenzhen People’s Hospital between January 2017 and December 2024. Preconception disease activity was categorised using SLE-DAS, LLDAS and DORIS remission criteria. Main outcomes were maternal flares and fetal outcomes, including spontaneous abortion, therapeutic abortion, total fetal loss, preterm delivery and small for gestational age (SGA). Predictive accuracies of indices were compared.

Results

Preconceptionally, 127 pregnancies (62.8%) met LLDAS, 131 (64.9%) met SLE-DAS LDA and 78 (38.6%) achieved DORIS remission. Compared with higher disease activity, SLE-DAS LDA was associated with fewer maternal flares (22.1% vs 45.1%) and therapeutic abortions (6.4% vs 15.7%). LLDAS was associated with lower rates of flare (21.3% vs 45.3%), therapeutic abortion (7.9% vs 17.3%), total fetal loss (19.7% vs 34.2%) and preterm delivery (22.0% vs 25.3%). SLE-DAS and DORIS remission performed similarly for maternal outcomes, while DORIS remission correlated more strongly with favourable fetal outcomes, including lower total fetal loss (15.4% vs 31.5%), preterm delivery (15.4% vs 28.2%) and SGA (9.0% vs 19.4%). Multivariable analyses confirmed that achieving these disease-activity states preconception independently protected against total fetal loss, maternal flare and therapeutic abortion. LLDAS was the best overall predictor, while SLE-DAS LDA was the most attainable and predictive for maternal complications.

Conclusion

SLE-DAS LDA effectively predicts maternal complication, while LLDAS better identifies fetal risk. Remission offers similar protection but is less attainable, suggesting LDA suffices for conception planning. Optimising preconception disease control remains essential and warrants multicentre validation.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Achieving remission or low disease activity (LDA) preconception reduces adverse pregnancy outcomes in SLE, but definitions of these states vary widely and no consensus exists on which tool best predicts pregnancy risk.

WHAT THIS STUDY ADDS

• Lupus low disease activity state (LLDAS) best predicts fetal outcomes, while SLE Disease Activity Score and Definitions of Remission in SLE remission best predict maternal outcomes. LDA offers protection comparable to remission, challenging the requirement for remission preconception.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These results support prioritising LLDAS as a preconception target to reduce fetal complications and suggest guideline updates emphasising achievable LDA over remission. Future work should refine disease-activity thresholds and incorporate multimodal risk assessment.

Introduction

Systemic lupus erythematosus (SLE) is a chronic autoimmune disorder characterised by diverse clinical manifestations, multiorgan involvement and alternating flares and remissions.¹ Rheumatologists use various indices to define disease activity and remission. The Systemic Lupus Erythematosus Disease Activity Index (SLEDAI) assesses disease activity across 24 clinical and laboratory domains.² Modified versions, including the SLEDAI-2000 (SLEDAI-2K), which better captures persistent disease activity³ and the clinical SLEDAI (cSLEDAI), which is more practical for use in clinical settings,⁴ were subsequently developed. Based on these indices, definitions of remission and low disease activity (LDA) were established, such as the lupus low disease activity state (LLDAS)⁵ and the Definitions of Remission in SLE (DORIS).⁶ More recently, the SLE Disease Activity Score (SLE-DAS), developed for improved sensitivity and continuous measurement, was introduced and validated in 2019, with specific criteria for LDA and remission.⁷

SLE primarily affects women of childbearing age and is associated with adverse maternal and fetal outcomes.¹ Studies have demonstrated that patients in remission or LDA are less likely to experience adverse pregnancy outcomes (APOs).^{8 9} Accordingly, preconception counselling and planned pregnancies are essential for women with SLE who intend to conceive. Current recommendations advise achieving remission for at least 6 months before conception; if remission is not achievable, maintaining LDA for the same duration is recommended.10,12 However, no consensus exists regarding the definitions of LDA and remission. The 2023 European Alliance of Associations for Rheumatology recommendations suggest evaluating disease activity using Safety of Estrogens in Systemic Lupus Erythematosus National Assessment (SELENA)-SLEDAI or SLEDAI-2K combined with the British Isles Lupus Assessment Group index.¹² The 2022 Chinese guideline recommends the Systemic Lupus Erythematosus Pregnancy Disease Activity Index (SLEPDAI) and the Physician’s Global Assessment (PGA),¹⁰ while the 2020 American guidelines do not specify particular criteria.¹³ Clinical studies on SLE pregnancy have diverse definitions of LDA and remission, including cSLEDAI,¹⁴ SLEDAI,¹⁵ LLDAS¹⁶ and DORIS remission,¹⁷ as well as other assessments such as PGA and SLEPDAI.¹⁸ Given these inconsistencies, comparing different definitions of SLE disease activity is crucial to identify the most suitable measures for in SLE pregnancies.

Because the classical SLEDAI-2K has undergone repeated validation and underpins several definitions, our retrospective cohort study compared APO rates among patients with SLE having different disease-activity statuses classified by LLDAS, SLE-DAS and DORIS remission to determine the most reliable scoring system for guiding prepregnancy counselling and management.

Methods

Study design and participants

The study population comprised pregnant women with SLE managed at the department of rheumatology and immunology between January 2017 and December 2024. The date of SLE diagnosis and initial symptoms was extracted from medical records. All patients met the 1997 revised American College of Rheumatology classification criteria for SLE.¹⁹ Only women with complete electronic medical records covering at least 1 year before conception through 4 weeks after delivery or pregnancy termination were included. Data were restricted to pregnancies occurring after the diagnosis of SLE, with each pregnancy considered as an independent observation. As this study aimed to analyse the impact of SLE disease activity before conception on pregnancy outcomes, patients with new-onset SLE during pregnancy were excluded. Additionally, elective terminations for personal reasons, multiple gestations and cases with insufficient clinical data were excluded. Ultimately, 202 pregnancies (including multiple pregnancies in three patients) in 196 women (mean age±SD: 29.6±4.57 years) met the eligibility criteria.

Baseline data were collected retrospectively from electronic medical records and included demographic characteristics, medical and reproductive history, SLE clinical history and immunological profile. Clinical parameters assessed included age at disease onset, previous cumulative SLE manifestations and history of antiphospholipid syndrome (APS). Preconception disease-activity status was retrospectively determined using the DORIS,⁶ LLDAS⁵ and SLE-DAS²⁰ criteria. Although objective data (eg, clinical manifestations and serological immunological findings) were reliably documented, the retrospective estimation of the PGA, although standardised, represents a limitation compared with prospective evaluation. Immunological assessments during pregnancy included testing for antinuclear antibodies (ANA), anti-double-stranded DNA, anti-Ro/SS-A, anti-La/SS-B, anticardiolipin antibodies, anti-β2 glycoprotein I antibodies, lupus anticoagulant and complement levels.

Definitions

SLE disease activity and assessment

DORIS⁶: Remission is defined as cSLEDAI=0 and PGA <0.5 (scale 0–3). Serological activity is not considered. Patients may be receiving antimalarials, low-dose glucocorticoids (prednisone ≤5 mg/day) and/or stable doses of immunosuppressants (including biologics).

LLDAS⁵: LLDAS is defined as SLEDAI-2K ≤4, no new lupus disease manifestations since the last evaluation, PGA ≤1 (scale 0–3), glucocorticoid dose (prednisone or equivalent) ≤7.5 mg/day and stable maintenance doses of immunosuppressants or approved biologics (excluding experimental agents) with good tolerability.

SLE-DAS^{7 20}: Remission is defined as SLE-DAS ≤2.08 and prednisolone ≤5 mg/day; LDA is defined as SLE-DAS ≤2.48 and prednisolone ≤7.5 mg/day.

Definition of pregnancy outcomes

Flare during pregnancy^{9 21}: Defined as any of the following: (1) an increase of ≥4 points in the SLEPDAI score from the preconception baseline; (2) an increase in the PGA of ≥1.0 point (scale 0–3) accompanied by an escalation of immunosuppressive therapy; (3) initiation of new immunosuppressive therapy specifically for SLE activity; (4) hospitalisation due to SLE-related complications.

Gestational hypertension, pre-eclampsia and eclampsia²²: Diagnoses of gestational hypertension and pre-eclampsia were based on the criteria of the International Society for the Study of Hypertension in Pregnancy. Gestational hypertension was defined as systolic blood pressure ≥140 mm Hg or diastolic blood pressure ≥90 mm Hg. Pre-eclampsia was defined as gestational hypertension with proteinuria occurring after 20 weeks of gestation. Eclampsia was defined as the occurrence of convulsions in the setting of pre-eclampsia.

Spontaneous abortion and stillbirth²³: Spontaneous abortion was defined as fetal demise before 20 weeks of gestation, and stillbirth was defined as intrauterine fetal demise at or after 20 weeks of gestation.

Early neonatal death²⁴: Death of a liveborn neonate within the first 7 days of life.

Therapeutic abortion²⁵: Elective termination of pregnancy due to maternal life-threatening risk.

Small for gestational age (SGA)^{26 27}: Birth weight below the 10th percentile for gestational age and sex, according to established Chinese neonatal growth standards.

Intrauterine growth restriction (IUGR)²⁸: Fetal weight below the 10th percentile for gestational age, as determined by prenatal growth assessment.

Preterm delivery²⁹: Delivery before 37 completed weeks of gestation.

Statistical evaluation

Feature selection and model establishment

Binary multivariable logistic regression was used to assess and compare the prognostic value of preconception SLE disease-activity indices for APOs. For each index (DORIS, LLDAS, SLE-DAS LDA and SLE-DAS remission), a separate model was fitted and adjusted for key clinical covariates. The strength of the independent associations was expressed as ORs with 95% CIs. A statistically significant association (p<0.05) in the multivariable model was considered evidence of an independent prognostic value for that index.

Statistical description and software

Statistical analyses were performed using SPSS V.22.0 (IBM, Armonk, New York, USA). Continuous variables (eg, age, disease duration) are presented as mean±SD, and categorical variables are expressed as counts and percentages. Normally distributed continuous variables were compared using Student’s t-test, whereas non-normally distributed variables were analysed using the Wilcoxon rank-sum test.

Results

Disease course and remission status

This study included 196 patients with 202 pregnancies between January 2017 and December 2024. Baseline demographic, clinical and laboratory characteristics before conception are summarised in online supplemental table 1. The mean maternal age at delivery was 29.6±4.57 years. The average age at SLE diagnosis was 23.9±5.01 years, and the mean disease duration was 5.7±4.18 years. Mucocutaneous involvement was the most common clinical manifestation (42.6%), followed by lupus nephritis (30.2%). APS was the most frequent comorbidity (8.9%). Regarding autoantibody profiles, 199 patients (98.5%) tested positive for ANA, representing the highest proportion. Lupus anticoagulants were detected in 36 patients (17.8%), and hypocomplementaemia was present in 73% of the cohort.

Distribution of SLE statuses in the cohort

The proportions of patients meeting different definitions of SLE remission and LDA are summarised in online supplemental table 2 and figure 1. 78 patients (38.6%) achieved DORIS remission, of whom only 8 (4.0%) attained complete remission. Clinical remission on corticosteroids was achieved by 11 patients (5.4%), whereas 59 (29.2%) achieved clinical remission off corticosteroids. Using the modified DORIS criteria, 41 patients (20.3%) met the remission threshold. SLE-DAS-based remission was achieved by 46.5% of patients, but only 10 (4.95%) met the more stringent Boolean-based remission criteria. Compared with remission definitions, LDA thresholds were met more frequently: 127 patients (62.8%) achieved modified LLDAS, and 64.9% met the SLE-DAS LDA definition, approximately 1.4 times higher than the proportion meeting SLE-DAS remission.

Figure 1. Components and thresholds of SLE disease activity measures used for preconception assessment. The diagram visualises the key defining components of the three disease activity measures evaluation—definitions of remission and LDA. It highlights the differing thresholds for clinical disease activity indices, permitted corticosteroid dose and PGA that collectively define remission and low disease activity states. The application of these criteria to preconception patient status is detailed in online supplemental table 2. cSLEDAI, clinical Systemic Lupus Erythematosus Disease Activity Index; DORIS, Definition of Remission in SLE; LLDAS, lupus low disease activity state; PGA, Physician’s Global Assessment; SLEDAI-2K, Systemic Lupus Erythematosus Disease Activity Index 2000; SLE-DAS, SLE Disease Activity Score remission.

Maternal and fetal outcomes in patients with differently defined LDA and remission

Adverse maternal and fetal outcomes were compared between patients in LDA before conception, as defined by the SLE-DAS and LLDAS criteria (table 1). Using SLE-DAS, patients in the LDA group had significantly fewer flares (22.1% vs 45.1%, p=0.001) and therapeutic abortions (6.4% vs 15.7%, p=0.037) than those in the high disease activity group. Other APOs were less frequent in the LDA group, though differences were not statistically significant. Using LLDAS, patients meeting the criterion had significantly fewer flares (21.3% vs 45.3%, p<0.001), spontaneous abortions (8.7% vs 10.7%, p=0.019), therapeutic abortions (7.9% vs 17.3%, p=0.037), total fetal losses (19.7% vs 34.2%, p=0.002) and preterm deliveries (22.0% vs 25.3%, p=0.022) than those who did not. Although not statistically significant, all other APOs, except IUGR, were less frequent in the LLDAS group.

Table 1. Adverse pregnancy outcomes by preconception low disease activity status (SLE-DAS vs LLDAS).

	Total	SLE-DAS7 20 Low disease activity			LLDAS5
	(n=202) n (%)	Achieved (n=131) n (%)	Not Achieved (n=71) n (%)	P value	Achieved (n=127) n (%)	Not Achieved (n=75) n (%)	P value
Maternal outcome
Flare during pregnancy	61 (30.2)	29 (22.1)	32 (45.1)	0.001	27 (21.3)	34 (45.3)	0.000
Gestational hypertension	11 (5.4)	6 (4.6)	5 (7.0)	0.522	5 (3.9)	6 (8.0)	0.487
Pre-eclampsia	7 (3.5)	3 (2.3)	4 (5.6)	0.244	3 (2.4)	4 (5.3)	0.282
Eclampsia	2 (1.0)	0 (0.0)	2 (2.8)	0.122	0 (0.0)	2 (2.7)	0.125
Gestational diabetes	8 (4.0)	5 (3.8)	3 (4.2)	1.000	5 (3.9)	3 (4.0)	1.000
Fetal outcome
Spontaneous abortion	19 (9.6)	12 (9.2)	7 (9.9)	0.871	11 (8.7)	8 (10.7)	0.019
Stillbirth	7 (3.5)	4 (3.1)	3 (4.2)	0.698	3 (2.4)	4 (5.3)	1.000
Early neonatal death	2 (1.0)	1 (0.8)	1 (1.4)	1.000	1 (0.8)	1 (1.3)	1.000
Therapeutic abortion	23 (11.4)	12 (9.2)	15 (21.1)	0.017	10 (7.9)	13 (17.3)	0.037
Total fetal loss	51 (25.2)	29 (22.1)	22 (31.0)	0.167	25 (19.7)	26 (34.2)	0.002
SGA26 27	31 (15.3)	17 (13.0)	14 (19.7)	0.204	17 (13.4)	14 (18.7)	0.342
IUGR28	22 (10.9)	16 (12.2)	6 (8.5)	0.412	14 (11.0)	8 (10.7)	0.425
Preterm delivery	47 (23.3)	28 (21.4)	19 (26.8)	0.387	28 (22.0)	19 (25.3)	0.022

IUGR, Intrauterine Growth Restriction; LLDAS, Lupus Low Disease Activity State; SGA, Small for Gestational Age; SLE-DAS LDA, SLE Disease Activity Score Low Disease Activity.

We also evaluated the associations between remission status at conception, as defined by SLE-DAS and DORIS, and adverse maternal and fetal outcomes (table 2). Patients in SLE-DAS remission had significantly lower rates of flares during pregnancy (17.0% vs 41.7%, p=0.001), spontaneous abortions (4.3% vs 13.9%, p=0.019), therapeutic abortions (6.4% vs 15.7%, p=0.037), total fetal losses (14.9% vs 34.1%, p=0.002) and preterm deliveries (16.0% vs 29.6%, p=0.022) than those not in remission. Similarly, patients in DORIS remission had significantly lower rates of flares (16.7% vs 38.7%, p=0.001), spontaneous abortions (3.8% vs 12.9%, p=0.032), therapeutic abortions (3.8% vs 16.1%, p=0.002), total fetal losses (15.4% vs 31.5%, p<0.001) and preterm deliveries (15.4% vs 28.2%, p=0.035) than those not in remission. Additionally, DORIS remission was associated with significantly fewer SGA infants (9.0% vs 19.4%, p=0.046).

Table 2. Adverse pregnancy outcomes by preconception remission status (SLE-DAS vs DORIS).

	Total	SLE-DAS7 20 remission			DORIS6
	(n=202) n (%)	Yes (n=94) n (%)	No (n=108) n (%)	P value	Yes (n=78) n (%)	No (n=124) n (%)	P value
Maternal outcome
Flare during pregnancy	61 (30.2)	16 (17.0)	45 (41.7)	0.000	13 (16.7)	48 (38.7)	0.001
Gestational hypertension	11 (5.4)	4 (4.3)	7 (6.5)	0.487	4 (5.1)	7 (5.6)	1.000
Pre-eclampsia	7 (3.5)	3 (3.2)	4 (3.7)	0.282	2 (2.6)	5 (4.0)	0.709
Eclampsia	2 (1.0)	0 (0.0)	2 (1.9)	0.125	0 (0.0)	2 (1.6)	0.524
Gestational diabetes	8 (4.0)	4 (4.3)	4 (3.7)	1.000	3 (3.8)	5 (4.0)	1.000
Fetal outcome
Spontaneous abortion	19 (9.6)	4 (4.3)	15 (13.9)	0.019	3 (3.8)	16 (12.9)	0.032
Stillbirth	7 (3.5)	3 (3.2)	4 (3.7)	1.000	2 (2.6)	5 (4.0)	0.709
Early neonatal death	2 (1.0)	1 (1.1)	1 (0.9)	1.000	0 (0.0)	2 (1.6)	0.524
Therapeutic abortion	23 (11.4)	6 (6.4)	17 (15.7)	0.037	3 (3.8)	20 (16.1)	0.002
Total fetal loss	51 (25.2)	14 (14.9)	37 (34.1)	0.002	8 (15.4)	43 (31.5)	0.000
SGA26 27	31 (15.3)	12 (16.0)	19 (14.8)	0.342	7 (9.0)	24 (19.4)	0.046
IUGR28	22 (10.9)	12 (12.8)	10 (9.3)	0.425	10 (12.8)	12 (9.7)	0.485
Preterm delivery	47 (23.3)	15 (16.0)	32 (29.6)	0.022	12 (15.4)	35 (28.2)	0.035

DORIS, Definitions of Remission in SLE; IUGR, Intrauterine Growth Restriction; SGA, Small for Gestational Age; SLE-DAS remission, SLE Disease Activity Score remission.

Overall, flares (30.2%) were the most frequent maternal outcome in this cohort, whereas total fetal loss (25.3%) was the most common fetal outcome, followed by preterm delivery (23.3%).

Multivariable logistic regression analyses for predictors of adverse pregnancy outcomes

Given the acknowledged groups overlap, multivariable logistic regression analyses were performed to identify independent predictors of major APOs, adjusting for potential confounders including lupus nephritis, APS, arterial hypertension, primigravida status, SLE duration and maternal age at delivery. The results are presented in figure 2 and online supplemental table 3.

Figure 2. Multivariable-adjusted ORs for predictors of adverse pregnancy outcomes. Forest plots display the ORs and 95% CIs for predictors of (A) total fetal loss, (B) flare during pregnancy and (C) therapeutic abortion across four disease activity states (DORIS, SLE-DAS remission, LLDAS, SLE-DAS LDA). The ORs are denoted by boxes and the 95% CIs by horizontal lines. Predictors with a CI not crossing the vertical line (OR=1) indicate statistical significance. Detailed numerical results for each predictor are provided in online supplemental table 3. All analyses were adjusted for lupus nephritis, APS, arterial hypertension, primigravida, SLE duration and maternal age at delivery. APS, antiphospholipid syndrome; DORIS, Definition of Remission in SLE; LDA, low disease activity; LLDAS, lupus low disease activity state; SLE-DAS, systemic lupus erythematosus Disease Activity Score.

Achieving a state of LDA or remission before conception was consistently identified as a strong independent protective factor against adverse outcomes. Specifically, DORIS remission (OR 0.141 (95% CI 0.056 to 0.357); p<0.001), SLE-DAS remission (0.108 (95% CI 0.044 to 0.264); p<0.001), LLDAS (0.268 (95% CI 0.134 to 0.536); p<0.001) and SLE-DAS LDA (0.217 (95% CI 0.108 to 0.437); p<0.001) were all significantly associated with a reduced risk of total fetal loss. Similarly, these disease-activity states demonstrated a robust protective effect against both flares during pregnancy and therapeutic abortion. Comparisons of the DORIS (0.040 (95% CI 0.012 to 0.139); p<0.001) and SLE-DAS remission (0.034 (95% CI 0.011 to 0.103); p<0.001) criteria consistently revealed the strongest protective effects across both outcomes, underscoring the superior benefit of achieving a stringent remission state. In contrast, LLDAS (0.076 (95% CI 0.035 to 0.163); p<0.001) and SLE-DAS LDA (0.092 (95% CI 0.043 to 0.195); p<0.001) indicated a relatively attenuated but still protective effect. Among clinical covariates, a longer duration of SLE was independently associated with a lower risk of flare during pregnancy across all definitions regarding disease activity (eg, DORIS remission model: 0.893 (95% CI 0.818 to 0.976); p=0.012). A history of lupus nephritis showed a significant protective association against total fetal loss in most models (eg, SLE-DAS LDA model: 0.398 (95% CI 0.170 to 0.933); p=0.034). Conversely, APS, arterial hypertension, primigravida status and maternal age were not consistently or significantly associated with the primary outcomes in the adjusted models. These results emphasise that achieving LDA or remission before conception is an essential, independent protective factor for reducing the risk of major APOs in patients with SLE.

Discussion

We evaluated which preconception SLE disease-activity scoring method most reliably predicts APOs. The most prevalent APOs in our cohort were flares, total fetal loss and preterm delivery. The prevalence of preterm delivery (23.3%) was similar to that reported in multicentre Chinese studies from Beijing³⁰ and Guangzhou³¹ (22.7% and 24.0%, respectively), but notably higher than that in the Predictors of pRegnancy Outcome: biomMarkers In antiphospholipid antibody Syndrome and Systemic lupus Erythematosus (PROMISSE)¹⁸ study (9%). The flare rate in our cohort (30.2%) exceeded those in the Beijing and Guangzhou cohorts (20.4% and 21.4%, respectively). Total fetal loss was also higher in our cohort (25.2%) than in the Beijing, Guangzhou and PROMISSE cohorts (20.3%, 18.8% and 5%, respectively), but comparable to that of a recent US cohort study (26%) involving a diverse patient population.³²

Our findings are consistent with previous studies demonstrating the protective effect of LLDAS against APOs.¹⁶ In our cohort, LDA, as defined by SLE-DAS and LLDAS, performed similarly in predicting maternal complications, although LLDAS had a stronger predictive value for fetal outcomes. This difference may partly reflect glucocorticoid thresholds: SLE-DAS LDA permits prednisolone doses up to 7.5 mg/day, whereas LLDAS uses a lower threshold of ≤5 mg/day and higher glucocorticoid doses have been linked to increased APO risk.^{9 33}

The novelty of the present study lies in providing empirical evidence to refine, rather than replace, established clinical understanding. While the principle of preconception disease control is universally acknowledged,³⁴ our findings quantify the differential benefits of specific, operationalised treatment targets. This evidence shifts the paradigm from a generic recommendation for ‘disease control’ to a data-driven strategy that supports selection of the most appropriate target, thereby enhancing the precision of preconception counselling.

A key implication of our findings is the need for nuanced interpretation of a single disease-activity measure, rather than simultaneous use of multiple scores. We do not advocate for routine dual scoring in clinical practice. The incremental value lies in understanding the prognostic significance of specific components within a chosen tool, with the glucocorticoid threshold being a pivotal differentiator. For instance, when using the SLE-DAS, achieving its LDA criterion provides reassurance against maternal flares, yet the permitted glucocorticoid dose may signal a need for heightened fetal surveillance.³⁵ Conversely, meeting LLDAS, with its stringent steroid limit, offers stronger foetal protection.³⁶ Thus, clinicians can leverage a single index while appreciating that its constituent elements convey distinct implications for maternal risk versus fetal risk.

Previous studies have reported contradictory findings regarding the predictive value of disease remission in patients with APOs. Some research has shown that remission is more protective than LDA,³⁷ while others have reported similar protective effects for DORIS and LLDAS.^{8 14} In our cohort, SLE-DAS and DORIS remission were equally effective for predicting maternal outcomes; however, DORIS remission was superior in discriminating fetal outcomes. Our findings are consistent with those of Nakai et al, who demonstrated the protective value of DORIS remission against flares, iatrogenic abortion and low birth weight.¹⁷ DORIS remission reportedly protects against damage accrual only when maintained for more than two consecutive years.³⁸ In Nakai’s study, the mean age at pregnancy was relatively high (34 years), and the mean disease duration was long (5.7 years). Further studies are therefore needed to validate the protective value of DORIS remission in younger pregnant patients with SLE. Additionally, SLE-DAS remission, SLE-DAS LDA and LLDAS matched the predictive accuracy of DORIS remission for maternal outcomes. For fetal outcomes, these three definitions were nearly equivalent to DORIS remission, except that DORIS remission outperformed SLE-DAS LDA in predicting total fetal loss. This finding aligns with prior research showing that DORIS remission, SLE-DAS remission, LLDAS and SLE-DAS LDA are similarly associated with patient-reported outcomes.³⁹

Despite the overall protective roles of LDA in remission, unexpected trends were observed. First, gestational diabetes mellitus occurred more frequently in patients meeting the SLE-DAS LDA criterion. This may result from increased cumulative glucocorticoid exposure, as glucocorticoids can induce hepatic gluconeogenesis⁴⁰ and impair pancreatic beta-cell function.⁴¹ Second, IUGR was relatively common in patients achieving LLDAS. This may be attributable to persistently elevated lupus anticoagulant levels despite low systemic disease activity or remission, which can impair placentation and cause endothelial dysfunction, leading to IUGR.³³ Moreover, excessive glucocorticoid use can increase placental vascular resistance and impair fetal growth.⁴²

Critically, our observations that >20% individuals achieving SLE-DAS LDA experienced flares and nearly 15% suffered fetal loss underscore a fundamental limitation (table 1): attaining a LDA state does not confer absolute protection against adverse outcomes. These data compel a more nuanced interpretation of the clinical utility of these diagnostic algorithms. The persistence of adverse events in a significant proportion of patients in LDA underscores the multifactorial nature of pregnancy outcomes in SLE.⁴³ Factors beyond the scope of current disease-activity indices—such as pre-existing organ damage, the presence and profile of antiphospholipid antibodies, subclinical placental pathology and other non-inflammatory mechanisms—are likely to play a substantial role in determining residual risk.⁴⁴

This nuanced risk profile supports personalised treatment targets based on individual patient priorities. For instance, striving for the more readily attainable SLE-DAS LDA may be a pragmatically sufficient goal for patients primarily concerned with minimising maternal flare risk.²⁰ Conversely, for those with heightened concerns regarding fetal outcomes, our data support targeting the more stringent LLDAS, which is more strongly associated with fetal protection.⁴⁵ This facilitates a shared decision-making conversation, moving from a generic directive to a tailored plan aligned with patient-specific risks and concerns.

Based on these findings, we propose several clinical recommendations. For predicting maternal complications, the SLE-DAS is recommended, as it offers prognostic accuracy equivalent to other definitions while being the most achievable target. For fetal outcomes, LLDAS is preferred, as it is most clearly distinguished APO risk between those who did and did not achieve the target and it was the second most achievable state. Since all remission definitions performed similarly to LDA but were less frequently attained, remission may not be a necessary prerequisite for conception in patients with SLE. It is also well established that SLE onset during pregnancy in adulthood is associated with more aggressive disease manifestations and higher APO rates compared with childhood-onset or prepregnancy-onset SLE.⁴⁶ Interestingly, recent research has shown that APO incidence does not correlate with disease duration if LDA is achieved before conception, further supporting the clinical importance of preconception disease-activity assessment.⁴⁷

Current guidelines recommend that patients with SLE enter pregnancy in remission or LDA.^{9 34} However, in our cohort, only 64.9% and 62.8% of pregnancies met the preconception SLE-DAS LDA and LLDAS criteria, respectively, highlighting the need for more effective prepregnancy disease control strategies. Furthermore, even among patients who achieved SLE-DAS LDA, 22.1% experienced flares and nearly 15% had total fetal loss, indicating that current targets have limitations and that further research is warranted to develop optimal conception guidance for patients with SLE. Recent efforts to apply novel methods, such as machine learning, to predict APOs have shown promising potential.

Notably, a significant proportion of patients who had achieved LDA or remission status before conception still experienced APOs. This observation underscores the multifactorial nature of pregnancy outcomes in SLE, suggesting that factors beyond the scope of current disease-activity indices play a substantial role in determining risk. Such factors are likely to include pre-existing organ damage, subclinical placental pathology and other non-inflammatory mechanisms.^{9 48} To move beyond simple unadjusted comparisons and to better isolate the independent effect of disease activity, we performed multivariable logistic regression analyses adjusting for key potential confounders, including the antiphospholipid antibody status and a history of lupus nephritis. Results of these analyses provide crucial statistical reinforcement, confirming that achieving preconception LDA or remission—whether defined by LLDAS, DORIS or SLE-DAS criteria—is a powerful and independent protective factor against major adverse outcomes. However, the models also clarify that disease activity, while a major determinant, is not the sole driver of risk. The persistent, although not always statistically significant, trend towards increased risk associated with antiphospholipid antibody positivity in our multivariable models highlights its potential role as an independent risk driver, particularly for fetal loss, and warrants further investigation in larger cohorts. Collectively, these findings advocate for a comprehensive risk stratification strategy in clinical practice that rigorously manages disease activity while simultaneously addressing other significant non-inflammatory risk factors, such as APS, to optimally improve pregnancy outcomes in women with SLE.

This study had some limitations. First, its single-centre, retrospective design may introduce selection bias and limit the generalisability of our findings to the broader SLE population, as results may be influenced by local treatment protocols and specific patient demographics. Second, although the overall cohort was of a respectable size, the number of patients in key subgroups—particularly those achieving more stringent states such as DORIS remission—was modest. This limits statistical power for certain comparisons and increases uncertainty around the corresponding estimates. Third, while objective components (eg, clinical manifestations and serological data) are reliably documented, the retrospective assignment of the PGA, though standardised, remains a limitation compared with prospective collection. Finally, the potential for unmeasured confounders to have influenced both disease activity and pregnancy outcomes cannot be excluded. Consequently, external validation of our findings in larger, prospective, multicentre cohorts from diverse backgrounds is required to confirm their robustness and generalisability.

In conclusion, this study compared multiple preconception disease activity indices for their ability to predict maternal and fetal outcomes in SLE pregnancies. Our findings suggest that SLE-DAS is the most reliable predictor of maternal complications, whereas LLDAS is the strongest predictor of fetal outcomes. Similarly, SLE-DAS and DORIS remission were effective for predicting maternal outcomes, with DORIS remission showing superiority for fetal outcomes. Therefore, we posit that the clinical value of these targets lies not in being perfect predictors, but in providing standardised, evidence-based benchmarks for preconception counselling. They represent a treatment state that, while not eliminating risk entirely, significantly reduces the probability of adverse outcomes. Thus, striving for these states should be viewed as a necessary, although not always sufficient, component of prepregnancy optimisation. However, remission should not be considered a necessary prerequisite for conception, as LDA is more achievable and provides comparable protective effects. These results underscore the need for enhanced prepregnancy disease control strategies, given that many pregnancies did not meet optimal disease-activity thresholds.

Data availability statement

Data are available upon reasonable request.