Breastfeeding determinants in Egyptian mothers with systemic lupus erythematosus or rheumatoid arthritis: a retrospective cohort study

Abstract

Objective

Breastfeeding prevalence and challenges among women of systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA) is under-researched especially in the Middle East-North Africa region. This study aimed to assess breastfeeding initiation, duration and predictors of early discontinuation (<6 months post partum) among Egyptian mothers with SLE or RA.

Methods

This multicentre retrospective cohort study included 320 pregnancies: 62 SLE (105 pregnancies), 71 RA (110 pregnancies) and 59 healthy mothers (105 pregnancies). Data on pregnancy history, breastfeeding intent, initiation, duration and weaning reasons were collected.

Results

Exclusive breastfeeding was lowest in SLE (29.9%) vs RA (50.6%) and controls (60%, p<0.001). Continuation beyond 6 months was significantly lower in SLE (36.2%) and RA (33.6%) vs controls (81%, p<0.001). Postpartum depression independently predicted discontinuation in SLE (adjusted OR (aOR)=0.06, 95% CI 0.01 to 0.6) and RA (aOR=0.34, 95% CI 0.13 to 0.9). Multivariable generalised estimating equation confirmed SLE reduced breastfeeding odds versus controls (aOR=0.41, p=0.040).

Conclusion

Breastfeeding is significantly less prevalent among Egyptian mothers with SLE and RA when compared with control group. Targeted educational programme and support may help improve breastfeeding rates in SLE/RA mothers.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• International studies indicate women with SLE and rheumatoid arthritis (RA) face reduced breastfeeding rates due to medication safety concerns, disease flares, physical limitations and mental health challenges such as postpartum depression. Data on these factors in Egyptian women are lacking.

WHAT THIS STUDY ADDS

• This is the first study to examine breastfeeding patterns among Egyptian women with SLE and RA, highlighting the impact of postpartum depression, disease activity and lack of support on breastfeeding decisions.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• These findings underline the importance of integrating multidisciplinary care and breastfeeding support into postpartum care for women with autoimmune diseases. The results can guide healthcare providers and inform national strategies to improve breastfeeding outcomes in this population.

Introduction

Breastfeeding is widely recognised by organisations such as the WHO as an optimal method for infant nutrition and development, recommending exclusive breastfeeding for the first 6 months of life, with continued breastfeeding alongside complementary foods up to 2 years or beyond.¹ However, for women suffering from chronic inflammatory diseases such as systemic lupus erythematosus (SLE) and rheumatoid arthritis (RA), breastfeeding presents unique and significant challenges.² Both diseases disproportionately affect women of childbearing age, and postpartum periods are often associated with an increased risk of disease flares, necessitating adjustments to medication regimens.^{3 4} Pain, fatigue and physical limitations associated with disease activity can make holding the baby and maintaining breastfeeding challenging.⁵

While some studies suggest that breastfeeding might influence disease activity in RA, leading to flares, especially after a first pregnancy,⁶ other studies have shown lower disease activity in breastfeeding patients⁷ or found no significant difference in disease activity between breastfeeding and non-breastfeeding groups in SLE.⁸

Despite the growing body of research on breastfeeding in women with rheumatic diseases from various international contexts,8,11 there is a notable scarcity of specific studies addressing these challenges in the Egyptian population. Given that breastfeeding patterns and healthcare contexts can differ significantly across regions, particularly between developed and developing countries, understanding the unique experiences and needs of women with SLE and RA in Egypt is crucial. This manuscript aims to investigate the initiation and duration of breastfeeding among patients with SLE and RA in Egypt, to identify predictors of discontinuation of breastfeeding before 6 months post partum and analyse relevant influencing factors.

Patients and methods

Study design and setting

This multicentre, retrospective, cohort study was conducted at the Rheumatology Departments of Minia, Mansoura and Al-Azhar University Teaching Hospitals in Egypt between August 2024 and June 2025.

Study population

This study consecutively enrolled women with SLE or RA who had experienced at least one pregnancy resulting in a live birth after the onset of their disease. If multiple pregnancies were reported, all pregnancies that occurred after diagnosis of the disease were included to account for repeated lactation outcomes from the same individual. Controls were healthy mothers without autoimmune diseases, recruited from paediatric outpatient waiting areas. All patients with SLE met the 2019 American College of Rheumatology/European Alliance of Associations for Rheumatology (ACR/EULAR) classification criteria,¹² and patients with RA met the 2010 ACR/EULAR classification criteria.¹³

Exclusion criteria included coexisting autoimmune diseases known to influence pregnancy or lactation, such as Sjögren’s syndrome, systemic sclerosis or autoimmune thyroiditis; missing medical records; or inability to reliably recall lactation details (eg, duration, exclusivity, initiation timing).

Data collection

A standardised data form was used to collect demographic variables (age at enrolment, educational level, body mass index), obstetric history (number of pregnancies, maternal age at delivery, gestational age, delivery mode) and clinical data (disease duration, disease activity and medications). Spontaneous pregnancy loss (‘miscarriage’) was defined as non-viable pregnancy ending before 20 weeks of gestation, based on obstetric documentation.

Disease activity during pregnancy and post partum was assessed from medical records, including physician documentation, clinical signs, relevant labs and imaging and validated disease activity indices. For SLE, the Systemic Lupus Erythematosus Disease Activity Index and physician global assessment were used. For RA, the Disease Activity Score in 28 joints was applied. Disease status per pregnancy was categorised as active or inactive.

Medication data included both SLE-specific and non-SLE-specific drugs. Use of immunosuppressants, corticosteroids, hydroxychloroquine and lactation-incompatible medications was documented.

Neonatal and maternal intensive care unit (ICU) admissions and any perinatal complications were recorded. Lactation variables included intention to breastfeed, skin-to-skin contact, initiation within 24 hours, exclusive breastfeeding, duration and reasons for weaning. Exclusive breastfeeding was defined as breastfeeding without formula supplementation. Drug compatibility with lactation was assessed using the 2022 EULAR recommendation.¹⁴

Ethical considerations

The study complies with the Declaration of Helsinki. Informed consent was obtained from all participants.

Statistical analysis

All statistical analyses were performed using Jamovi V.2.6.2. Continuous variables are reported as medians with IQRs, and categorical variables as frequencies and percentages. Group comparisons (SLE vs RA vs controls) were done by Kruskal-Wallis test for continuous variables; χ² or Fisher’s exact test for categorical variables. Bonferroni-adjusted pairwise comparisons (α=0.0167) were conducted where applicable. Two-group comparisons (breastfed vs non-breastfed within SLE/RA cohort) were done by Mann-Whitney U test for continuous variables; χ² or Fisher’s exact test for categorical variables.

Binary logistic regression was conducted to assess predictors of breastfeeding. Variables with p<0.1 in univariable analysis were entered into multivariable models. No additional variable selection was applied, and adding clinical variables such as maternal age and parity did not materially alter the results. Model performance was assessed using the area under the receiver operating characteristic curve (AUC), and multicollinearity was assessed using the variance inflation factor (VIF); VIF >5 was excluded. ORs and 95% CIs were reported, and a p value <0.05 was considered statistically significant. The final multivariable logistic regression model demonstrated excellent discriminatory power in the SLE group (AUC=0.896; 85% accuracy) and moderate discriminatory power in the RA group (AUC=0.72; 70% accuracy).

To account for multiple pregnancies per mother, multivariable generalised estimating equations (GEE) were applied with a logit link and exchangeable correlation structure. Candidate predictors were first screened in univariable analyses, and all variables with p<0.10 were then entered simultaneously into the multivariable GEE models. Separate models were constructed for the overall cohort, the SLE subgroup and the RA subgroup. This approach ensured that adjusted ORs (aORs) presented for each predictor reflect estimates obtained while jointly accounting for all other retained predictors, as well as for within-mother clustering. Multicollinearity was evaluated using VIFs, with no evidence of problematic overlap. Model convergence was achieved in all cases.

Results

Baseline characteristics of study participants

This study included a total of 320 pregnancies in 62 SLE mothers (105 pregnancies), 71 RA mothers (110 pregnancies) and 59 healthy mothers (105 pregnancies). The median age at enrolment was 31 years (11.3) in SLE, 40 years (12.5) in RA and 32 years (9.5) in controls. In terms of educational level, healthy mothers received higher education (controls 72.9%, SLE 17.7%, RA 12.7%, p<0.001). SLE mothers were more likely to suffer from spontaneous pregnancy losses (SLE 54.8%, RA 36.6%, controls 11.9%, p<0.001) (table 1).

Table 1. Demographics of the included patients with SLE and RA and healthy controls.

	SLE n=62	RA n=71	Control n=59	P value
Age	33 (11.3)	40 (12.5)	32 (9.5)	<0.001*
BMI	27.3 (5.2)	27.18 (5.1)	27.9 (7.6)	0.13
Educational level
Low	13 (21%)	24 (33.8%)	4 (6.8%)	<0.001*
Middle	38 (61.3%)	38 (53.5%)	12 (20.3%)
High	11 (17.7%)	9 (12.7%)	43 (72.9%)
Number of pregnancies	3 (2)	3 (3)	2 (1.5)	<0.001*
Number of children	2 (1)	3 (2)	2 (1.5)	0.004*
Parity
Primipara	15 (24.2%)	16 (22.5%)	15 (25.4%)	0.93
Multipara	47 (75.8%)	55 (77.5%)	44 (74.6%)
History of spontaneous pregnancy losses	34 (54.8%)	26 (36.6%)	7 (11.9%)	<0.001*

Educational levels: low=elementary school, medium=high school/vocational education, high=college/university. Data represented in median (IQR) for quantitative variables and number (%) for qualitative variables.

^*P value is statistically significant at <0.05. Group comparisons were performed by Kruskal-Wallis test for ordinal, continuous variables and χ2 test for categorical variables.

BMI, body mass index; RA, rheumatoid arthritis.

Pregnancy and lactation characteristics among SLE, RA and control mothers

Table 2 compares disease activity and management during pregnancy and post partum in SLE and RA mothers. Disease duration during pregnancy was longer in patients with RA (RA 3.75 (6) years vs SLE 3 (4) years, p=0.006). Active disease during pregnancy occurred more frequently in SLE (56.2%) vs RA (23.6%, p<0.001). No significant difference was noted in the rate of postpartum active disease between SLE (75.2%) and RA (84.5%, p=0.09).

Table 2. Disease activity and management during pregnancy and post partum in SLE and RA mothers.

	SLE n=105	RA n=110	P value
Disease duration during pregnancy (years)	3 (4)	3.75 (6)	0.006*
Disease activity during pregnancy	59 (56.2%)	26 (23.6%)	<0.001
Drugs used during pregnancy
Hydroxychloroquine	76 (72.4%)	72 (65.5%)	0.27
Glucocorticoids	78 (74.3%)	59 (53.6%)	0.002*
Anti-tumour necrosis factor	0 (0%)	1 (0.9%)	0.35
Azathioprine	19 (20.2%)	1 (0.9%)	<0.001*
Sulfasalazine	0 (0%)	3 (2.7%)	0.11
Disease activity post partum	79 (75.2%)	93 (84.5%)	0.09
Drugs used post partum
Hydroxychloroquine	89 (84.8%)	93 (84.5%)	0.97
Glucocorticoids	92 (87.6%)	90 (81.8%)	0.24
Anti-tumour necrosis factor	0 (0%)	2 (2.4%)	0.23
Azathioprine	21 (23.3%)	5 (6%)	0.001*
Leflunomide	1 (1.1%)	10 (11.9%)	0.004
Methotrexate	1 (1.1%)	24 (28.6%)	<0.001
Cyclosporine	3 (3.3%)	0 (0%)	0.25
Cyclophosphamide	8 (8.9%)	0 (0%)	0.007*
Mycophenolate mofetil	4 (4.4%)	0 (0%)	0.12
Baricitinib	0 (0%)	2 (2.4%)	0.23
Disease manifestations during pregnancy/lactation
Arthritis	78 (74.3%)	98 (89.1%)	0.005*
Mucocutaneous	69 (65.7%)	1 (0.9%)	<0.001*
Cardiovascular	9 (8.6%)	0 (0%)	0.001*
Haematological	40 (38.1%)	0 (0%)	<0.001*
Neurological	7 (6.7%)	0 (0%)	0.006*

Data represented in median (IQR) for quantitative variables and number (%) for qualitative variables.

^*P value is statistically significant at <0.05. Group comparisons were performed using Fisher’s exact test for categorical variables. For continuous, non-normally distributed data, a Mann-Whitney U test was performed.

RA, rheumatoid arthritis.

As shown in online supplemental table 1, the median maternal age at the time of pregnancy was significantly lower in the SLE group and control group (25 years both) compared with the RA group (30 years, p<0.001 overall; SLE vs RA p<0.001; SLE vs control p=0.98; RA vs control p<0.001).

Gestational age at delivery differed significantly between groups. Preterm birth (<37 weeks) occurred in 25.7% of SLE pregnancies, compared with 10% in RA and 7.6% in controls (p<0.001). Mode of delivery was comparable, with caesarean section (C-section) rates of 60% in SLE, 53.6% in RA and 62.9% in controls (p=0.37).

Maternal ICU admission was significantly more frequent in the SLE group (14.3%) compared with RA (0.9%) and controls (1.9%) (p<0.001). Perinatal complications were also significantly more common in SLE pregnancies (46.7%) than in RA (13.6%) or control pregnancies (14.3%) (p<0.001). Notably, pre-eclampsia occurred in 21.9% of SLE pregnancies compared with 1.8% in RA and 5.7% in controls (p<0.001), while eclampsia was observed only in the SLE group (15.2%) and once in the control group (1%), with no cases in RA (p<0.001).

Neonatal outcomes were also less favourable in the SLE group. Neonates born to SLE mothers had the lowest median birth weight (2800 g (IQR 500)), compared with 3000 g (IQR 300) in RA and 3000 g (IQR 350) in controls (p<0.001). Neonatal intensive care unit (NICU) admission occurred in 7.6% of neonates born to SLE mothers, 4.5% in RA and 1.9% in controls (p=0.15 overall). Postpartum depression affected 34.3% of SLE mothers, 26.7% of RA and 30.5% of controls, with no statistically significant difference (p=0.23).

A high proportion of women across all groups reported an intention to breastfeed, though significantly lower in the SLE group (81%) compared with RA (93.6%) and controls (97.1%) (p<0.001). However, breastfeeding initiation within the first 24 hours was significantly lower in SLE mothers (72.4%) than in RA (83.6%) and controls (88.6%) (p=0.008). Skin-to-skin contact postdelivery occurred in 71.4% of SLE deliveries, 85.5% of RA and 75.2% of controls (p=0.038).

The mode of infant feeding differed substantially across groups. Exclusive breastfeeding in the first 6 months was reported by 29.9% of SLE mothers, compared with 50.6% of RA and 60% of controls (p<0.001). Formula feeding was most common in the SLE group (43.8%), compared with 24.5% in RA and only 6.7% in controls. Mixed feeding was similarly distributed (26.7% SLE, 24.5% RA, 33.3% controls).

Breastfeeding duration was shortest in the SLE group. Only 36.2% of SLE mothers continued breastfeeding beyond 6 months, compared with 33.6% of RA mothers and 81% of controls (p<0.001). Early discontinuation was also more frequent in SLE: 47.6% of mothers ceased breastfeeding within the first 3 months, compared with just 10.5% in controls.

Maternal and neonatal characteristics according to breastfeeding status within SLE and RA cohorts

In the SLE cohort, breastfeeding mothers (n=31) had significantly more favourable clinical and perinatal profiles compared with those who did not breastfeed (n=74). Active disease during pregnancy was observed in 25.8% of breastfeeding mothers versus 68.9% of non-breastfeeding mothers (p<0.001). Similarly, postpartum disease activity was lower in the breastfeeding group (54.8% vs 83.8%, p=0.02). Glucocorticoid use was also significantly less frequent among breastfeeding mothers (38.7% vs 89.2%, p<0.001) (table 3).

Table 3. Comparison between breastfeeding and formula feeding children in SLE and RA.

	SLE			RA
	Breastfed infants (n=31)	Non-breastfed infants (n=74)	P value	Breastfed infants (n=56)	Non-breastfed infants (n=54)	P value
Maternal age, median (IQR)	25 (9.5)	25.5 (5)	0.99	30 (9)	28 (6.75)	0.37
High educational level	5 (16.1%)	12 (16.2%)	0.99	6 (10.7%)	5 (9.3%)	0.8
Parity (primipara)	2 (6.5%)	13 (17.6%)	0.14	10 (17.9%)	6 (11.1%)	0.32
Disease duration during pregnancy (years)	2 (3.25)	3 (4.38)	0.25	5 (7.6)	3 (3.8)	0.22
Disease activity during pregnancy	8 (25.8%)	51 (68.9%)	<0.001*	9 (16.1%)	17 (31.5%)	0.06
Drugs used during pregnancy
Hydroxychloroquine	17 (54.8%)	59 (79.7%)	0.009*	33 (58.9%)	39 (72.2%)	0.14
Glucocorticoids	12 (38.7%)	66 (89.2%)	<0.001*	29 (51.8%)	30 (55.6%)	0.69
Anti-TNF agents	0 (0%)	0 (0%)	–	0 (0%)	1 (2.1%)	1
Azathioprine	2 (6.7%)	17 (26.7%)	0.025*	0 (0%)	1 (1.9%)	0.5
Sulfasalazine	0 (0%)	0 (0%)	–	3 (5.4%)	0 (0%)	0.24
Disease activity post partum	17 (54.8%)	62 (83.8%)	0.002*	44 (78.6%)	49 (90.7%)	0.08
Gestational weeks
Preterm (<37 weeks)	3 (9.7%)	24 (32.4%)	0.015*	3 (5.4%)	8 (14.8%)	0.1
Mode of delivery
C-section	15 (48.4%)	48 (64.9%)	0.12	28 (50%)	31 (57.4%)	0.44
Maternal admission to ICU post partum	0 (0%)	15 (20.3%)	0.007*	0 (0%)	1 (1.9%)	0.49
Neonatal weight (g)	3000 (700)	2800 (500)	0.15	3000 (0)	3000 (500)	0.08
Neonatal sex
Male	18 (58.1%)	41 (55.4%)	0.8	26 (46.4%)	20 (37%)	0.32
Neonatal admission to NICU	3 (9.7%)	5 (6.8%)	0.61	0 (0%)	5 (9.3%)	0.03*
Postpartum depression	1 (3.2%)	35 (47.3%)	<0.001*	10 (17.9%)	22 (40.7%)	0.008*
Intention to breastfeed	31 (100%)	54 (73%)	0.001*	55 (98.2%)	48 (88.9%)	0.045*
Skin-to-skin contact after delivery	27 (87.1%)	48 (64.9%)	0.021*	52 (92.6%)	42 (77.8%)	0.03*
Breastfeeding initiation	28 (90.3%)	48 (64.9%)	0.008*	54 (96.4%)	38 (70.4%)	<0.001*
Compatible drug with lactation	29 (93.5%)	56 (75.7%)	0.03*	42 (75%)	33 (61.1%)	0.12

Continuous variables (maternal age, disease duration, neonatal weight) are presented as median (IQR) and compared using Mann-Whitney U test. Categorical variables are presented as n (%) and compared using χ2 test or Fisher’s exact test, as appropriate.

^*Statistical significance defined as p<0.05.

C-section, caesarean section; ICU, intensive care unit; NICU, neonatal intensive care unit; RA, rheumatoid arthritis; TNF, tumour necrosis factor.

Neonatal outcomes followed a similar trend. Preterm birth was less reported in the breastfeeding group (9.7% vs 32.4%, p=0.015), and maternal ICU admission occurred only among non-breastfeeding mothers (0% vs 20.3%, p=0.07). Breastfeeding initiation within 24 hours was more frequent in the breastfeeding group (90.3% vs 64.9%, p=0.021), as was skin-to-skin contact (87.1% vs 64.9%, p=0.021).

In the RA cohort, breastfed infants (n=56) had better neonatal outcomes compared with non-breastfed infants (n=54). NICU admissions occurred in none of the breastfed infants, compared with 9.3% in the non-breastfed group (p=0.03). Postpartum maternal depression was less observed among breastfeeding mothers (17.9% vs 40.7%, p=0.08). Similar to the SLE group, breastfeeding initiation (96.4% vs 70.4%, p<0.001) and skin-to-skin contact (92.6% vs 77.8%, p=0.03) were more frequent among mothers who breastfed.

Predictors of breastfeeding

As presented in table 4, univariable logistic regression among 105 SLE pregnancies identified several factors significantly associated with breastfeeding. Significant negative predictors included active disease during pregnancy (unadjusted OR=0.16, 95% CI 0.06 to 0.40, p<0.001), use of hydroxychloroquine (OR=0.31, 95% CI 0.13 to 0.76, p=0.01), glucocorticoids (OR=0.08, 95% CI 0.03 to 0.21, p<0.001), active postpartum disease (OR=0.24, 95% CI 0.09 to 0.60, p=0.003), preterm delivery (OR=0.22, 95% CI 0.06 to 0.81, p=0.02), perinatal complications (OR=0.22, 95% CI 0.09 to 0.58, p=0.002) and postpartum depression (OR=0.04, 95% CI 0.005 to 0.29, p=0.002). Additionally, early breastfeeding initiation (OR=5.1, 95% CI 1.40 to 18.23, p=0.01) and skin-to-skin contact after delivery (OR=3.66, 95% CI 1.15 to 11.59, p=0.03) were positively associated with breastfeeding.

Table 4. Regression of breastfeeding in SLE (univariable and multivariable binomial logistic regression).

	SLE
	Unadjusted			Adjusted
	OR	95% CI	P value	OR	95% CI	P value
Maternal age	0.98	0.98 to 1.06	0.57	–	–	–
Parity (multipara)	3.09	0.65 to 14.60	0.15	–	–	–
Disease activity during pregnancy (yes)	0.16	0.06 to 0.403	<0.001*	0.28	0.07 to 1.09	0.07
Drugs used during pregnancy (yes)
Hydroxychloroquine	0.31	0.13 to 0.76	0.01*	1.84	0.33 to 10.14	0.49
Glucocorticoids	0.08	0.03 to 0.21	<0.001*	0.06	0.01 to 0.32	0.001*
Disease activity post partum (yes)	0.24	0.09 to 0.601	0.003*	1.44	0.29 to 7.15	0.66
Gestational weeks (preterm)	0.22	0.06 to 0.81	0.02*	0.75	0.12 to 4.79	0.76
Maternal admission to ICU post partum (yes)	1.65×10−8	0 to infinity	0.99	–	–	–
Neonatal admission to NICU (yes)	1.48	0.33 to 6.61	0.61	–	–	–
Perinatal complications (any)	0.22	0.09 to 0.58	0.002*	0.88	0.16 to 4.81	0.88
Postpartum depression (yes)	0.04	0.005 to 0.29	0.002*	0.06	0.01 to 0.61	0.02*
Intention to breastfeed (yes)	6.6×107	0 to infinity	0.99	–	–	–
Skin-to-skin contact after delivery (yes)	3.66	1.15 to 11.59	0.03*	1.27	0.20 to 8.04	0.80
Breastfeeding initiation	5.1	1.40 to 18.23	0.01*	1.23	0.16 to 9.4	0.84

Multivariable model included variables with p<0.10 in univariable analysis. Model AUC=0.896; accuracy=85%.

^*P value is statistically significant at <0.05.

AUC, area under the receiver operating characteristic curve; ICU, intensive care unit; NICU, neonatal intensive care unit.

In the multivariable model, only glucocorticoid use (aOR=0.06, 95% CI 0.01 to 0.32, p=0.001) and postpartum depression (aOR=0.06, 95% CI 0.01 to 0.61, p=0.02) remained independent predictors of lower breastfeeding likelihood. Active disease during pregnancy showed a negative trend (aOR=0.28, p=0.07).

The univariate analysis of patients with RA is shown in table 5. Postpartum depression (OR=0.32, 95% CI 0.13 to 0.76, p=0.01), skin-to-skin contact (OR=3.7, 95% CI 1.12 to 12.36, p=0.03) and breastfeeding initiation (OR=11.37, 95% CI 2.50 to 52.36, p=0.002) were significantly associated with breastfeeding status. The final multivariate model retained only postpartum depression (aOR=0.34, 95% CI 0.13 to 0.9, p=0.02) and breastfeeding initiation (aOR=7.5, 95% CI 1.4 to 40.6, p=0.02) as independent predictors. Intention to breastfeed showed a non-significant positive trend (aOR=3.63, p=0.32).

Table 5. Regression of breastfeeding in RA (univariable and multivariable binomial logistic regression).

	RA
	Unadjusted			Adjusted
	OR	95% CI	P value	OR	95% CI	P value
Maternal age	0.97	0.91 to 1.04	0.44	–	–	–
Parity (multipara)	0.58	0.19 to 1.71	0.32	–	–	–
Disease activity post partum (yes)	0.37	0.12 to 1.15	0.09	0.51	0.15 to 1.73	0.28
Gestational weeks (preterm)	0.33	0.08 to 1.30	0.11	–	–	–
Perinatal complications (any)	0.60	0.20 to 1.82	0.37	–	–	–
Postpartum depression (yes)	0.32	0.13 to 0.76	0.01*	0.34	0.13 to 0.87	0.02*
Intention to breastfeed (yes)	6.88	0.80 to 59.10	0.08	3.63	0.29 to 46.19	0.32
Skin-to-skin contact after delivery (yes)	3.71	1.12 to 12.36	0.03*	1.54	0.37 to 6.39	0.55
Breastfeeding initiation (yes)	11.37	2.50 to 52.36	0.002*	7.51	1.39 to 40.6	0.02*

Multivariable model included variables with p<0.10 in univariable analysis. Model AUC=0.722; accuracy=70%.

^*P value is statistically significant at <0.05.

AUC, area under the receiver operating characteristic curve; RA, rheumatoid arthritis.

Multivariable GEE analysis

Table 6 presents the results of the multivariable GEE models evaluating predictors of breastfeeding while adjusting for clustering of pregnancies within the same mother. In the overall cohort, postpartum depression was significantly associated with lower likelihood of breastfeeding (aOR=0.36, 95% CI 0.20 to 0.65, p=0.001), contrary to both intention to breastfeed (aOR=3.68, 95% CI 1.52 to 8.93, p=0.004) and early breastfeeding initiation (aOR=3.47, 95% CI 1.42 to 8.48, p=0.006), which were associated with increased likelihood of breastfeeding. SLE mothers had significantly lower odds of breastfeeding compared with healthy controls (aOR=0.41, 95% CI 0.17 to 0.96, p=0.04), while the difference between RA and controls did not reach statistical significance (aOR=0.49, 95% CI 0.23 to 1.04, p=0.06).

Table 6. Multivariable GEE models for predictors of breastfeeding with adjusting for multiple pregnancies in the same mother.

	Predictor variable	Adjusted OR	95% CI	P value
A. Overall cohort (n=320)	Participant group (SLE vs control)	0.41	0.17 to 0.96	0.04*
	Participant group (RA vs control)	0.49	0.23 to 1.04	0.063
	Postpartum depression	0.36	0.20 to 0.65	0.001*
	Intention to breastfeed	3.68	1.52 to 8.93	0.004*
	Breastfeeding initiation	3.47	1.42 to 8.48	0.006*
	Gestational weeks (preterm)	0.90	0.91 to 6.66	0.07
	Neonatal weight at delivery	0.99	0.99 to 1.00	0.09
	Perinatal complications	1.54	0.70 to 3.37	0.28
	Educational level (high)	0.72	0.35 to 1.50	0.38
	Skin-to-skin contact after delivery (yes)	1.27	0.56 to 2.86	0.57
	Mode of delivery (C-section)	1.15	0.64 to 2.10	0.64
	Maternal age	1	0.95 to 1.05	0.99
Subgroup analysis
B. SLE pregnancies (n=105)	Glucocorticoids	0.035	0.003 to 0.38	0.006*
	Disease activity during pregnancy (yes)	0.16	0.025 to 1.06	0.06
	Postpartum depression	0.06	0.002 to 1.30	0.07
	Skin-to-skin contact after delivery (yes)	2.06	0.12 to 12.17	0.43
	Drug incompatible with lactation	1.84	0.40 to 8.54	0.43
	Breastfeeding initiation	0.63	0.12 to 3.48	0.60
	Azathioprine	1.60	0.26 to 10.03	0.61
	Hydroxychloroquine	1.68	0.14 to 19.90	0.68
	Gestational weeks (preterm)	0.85	0.23 to 3.13	0.80
	Disease activity post partum (yes)	0.92	0.17 to 4.89	0.93
C. RA pregnancies (n=110)	Breastfeeding initiation	5.59	1.81 to 17.33	0.003*
	Postpartum depression	0.42	0.15 to 1.14	0.09
	Intention to breastfeed	2.37	0.66 to 8.51	1.19
	Parity	3.22	0.67 to 15.38	0.14
	Maternal age	0.94	0.86 to 1.03	0.19
	Skin-to-skin contact after delivery	2.13	0.56 to 8.01	0.27
	Perinatal complications	0.85	0.17 to 4.13	0.84
	Disease activity post partum	0.42	0.10 to 1.74	0.23

Multivariable GEE models were constructed using a logit link function and exchangeable working correlation structure. aORs, 95% CIs and p values are presented. For each analysis (overall cohort, SLE, RA), all predictors with p<0.10 in univariable analyses were entered simultaneously. Estimates reflect associations adjusted for the other retained predictors and for clustering of repeated pregnancies within mothers.

^*P value is statistically significant at <0.05.

aORs, adjusted ORs; C-section, caesarean section; GEE, generalised estimating equation; RA, rheumatoid arthritis.

In the SLE subgroup, glucocorticoid use during pregnancy remained an independent predictor of reduced breastfeeding likelihood (aOR=0.035, 95% CI 0.003 to 0.38, p=0.006). Postpartum depression and disease activity during pregnancy demonstrated non-significant negative trends. In the RA subgroup, early breastfeeding initiation was significantly associated with higher breastfeeding rates (aOR=5.59, 95% CI 1.81 to 17.33, p=0.003). Other variables, including postpartum depression, maternal intention to breastfeed and parity, did not reach statistical significance in this model.

Discussion

To our knowledge, this is the first multicentre study to comprehensively evaluate breastfeeding practices among Egyptian mothers with SLE or RA. While more than 80% of patients with SLE and 90% of patients with RA intended to breastfeed, only one-third of SLE mothers and approximately half of RA mothers were able to exclusively breastfeed. This is markedly lower than the rate among healthy controls. The WHO and the American Academy of Pediatrics recommend exclusive breastfeeding in the first 6 months after birth, and continuing breastfeeding along with appropriate complementary foods up to 2 years of age.¹

Most SLE and RA mothers initiated breastfeeding in the first 24 hours; however, only a third were able to continue breastfeeding at 6 months, which is a significantly lower rate than breastfeeding among healthy controls in our study. These findings align with previous international studies demonstrating variable but consistently reduced breastfeeding continuation among patients with autoimmune rheumatic diseases.

In Norway, a cohort of 101 SLE women had a 91% initiation rate that declined to 54% at 6 months, comparable to the general population.⁸ Similarly, in Italy, 71.9% of 43 patients with SLE initiated breastfeeding, but half discontinued within 3 months.¹⁰ Studies from the USA (Ikram et al² and Noviani et al¹⁵) found that about 49% of SLE mothers breastfed at their postpartum visit, with earlier discontinuation compared with mothers with other autoimmune conditions. In Argentina, the breastfeeding initiation rate was 80.6% among SLE mothers, lower than population norms.¹⁶ A Chinese study reported the lowest observed rate of breastfeeding in patients with SLE at 36.7%, compared with 86.7% in non-SLE mothers.¹¹

For patients with RA, early discontinuation is also common. A Dutch cohort found 57% stopped breastfeeding by 6 weeks post partum.⁵ However, improved outcomes were reported in the Preconception Counseling in Active RA (PreCARA) study, where 70.2% initiated breastfeeding and 60.2% continued at 6 weeks, comparable to national averages.¹⁷ Similar findings came from Romanian,¹⁸ US¹⁹ and Italian²⁰ studies, which reported breastfeeding continuation rates between 51% and 84%, with discontinuations over time.

In our study, the most frequent reason for early discontinuation of breastfeeding was concern over medication safety, even when the prescribed drugs were considered compatible with lactation. This observation is in line with previous studies.^{2 5 15 16 19} This echoes findings from the pregnancy and lactation autoimmune network (PLAN) registry, where 33% of mothers avoided breastfeeding due to medication fears, including hydroxychloroquine and biologics.¹⁹ This highlights a general fear of taking medication while breastfeeding and the lack of consistent counselling for patients with rheumatic diseases despite the regular updates and guidelines on the safety of medication use during pregnancy and lactation.^{14 21}

This highlights the need for structured, multidisciplinary counselling during the antenatal and postpartum periods, involving rheumatologists, obstetricians and lactation consultants. Providing written educational materials, reinforcing verbal discussions with evidence-based drug safety resources and integrating shared decision-making frameworks into routine care may empower mothers to continue breastfeeding without unnecessary cessation of breastfeeding.

In our adjusted multivariable analysis using GEE, three variables emerged as independent predictors of breastfeeding: postpartum depression, intention to breastfeed and early initiation. Mothers who lacked breastfeeding intention or delayed initiation had significantly lower odds of breastfeeding. Postpartum depression was a strong and consistent predictor of discontinuation, even after adjusting for sociodemographic and obstetric factors.

Importantly, SLE remained independently associated with lower odds of breastfeeding compared with controls, even after adjusting for confounders. In contrast, RA was not significantly associated in the fully adjusted model, suggesting more favourable breastfeeding dynamics or fewer disease-specific barriers.

In the SLE subgroup, GEE models showed that glucocorticoid use during pregnancy significantly predicted lower breastfeeding rates, possibly reflecting either direct treatment effects or more severe disease activity. While disease activity and postpartum depression trended towards reduced breastfeeding, they did not reach statistical significance.

In our cohort, more than half of the patients with SLE had active disease during pregnancy and experienced more frequent perinatal complications, preterm labour and a higher rate of maternal and NICU admissions compared with RA and control groups. Additionally, SLE mothers experienced postpartum depression more frequently, which likely interfered with breastfeeding. This aligns with prior studies showing associations between breastfeeding discontinuation and complications such as pre-eclampsia, preterm delivery and emergency C-sections in patients with SLE.^{2 8}

On the other hand, patients with RA in our study had more favourable pregnancy outcomes and were able to breastfeed their babies at a rate initially comparable to the control group. However, the breastfeeding rate declined significantly at 6 months to 33.6%, compared with 81% of the control group. In the RA subgroup analysis, delayed breastfeeding initiation was the only independent predictor of breastfeeding discontinuation. Other factors, including postpartum depression, parity and maternal age, did not reach statistical significance but may warrant further exploration in larger samples. Physical difficulties due to arthritis constitute another challenge to breastfeeding in patients with RA. Joint pain, fatigue and difficulty holding the baby in a breastfeeding position have been reported as causes for early weaning.²² Postpartum flares may also necessitate reintroduction of immunosuppressive therapy, further complicating lactation. Prior studies have shown associations between functional disability and early breastfeeding cessation.⁵

Whether breastfeeding worsens RA activity remains debated. Some studies report increased disease activity post partum with lactation,^{6 7 9} while others suggest lower disease activity among breastfeeding mothers.⁵ Prolactin has been implicated in exacerbating lupus activity and having a proinflammatory role in RA.²³ In our cohort, SLE mothers who breastfed had statistically significant lower disease activity flares than the non-breastfeeding group, whereas this pattern was not observed in patients with RA.

Our study identifies postpartum depression as a predictor of breastfeeding discontinuation in Egyptian SLE/RA mothers. This underscores an urgent need for mental health integration into postpartum care. Also, it is essential to recognise that breastfeeding decisions are highly personal and influenced by diverse medical, psychological and social factors. In some cases, patients may also feel pressured to breastfeed, experiencing guilt if they are unable or choose not to.²² To address these challenges, tailored, person-centred interventions are essential. These should provide accurate, accessible information on breastfeeding-compatible medications and foster a supportive, non-judgemental environment where mothers feel empowered to make informed feeding decisions aligned with their health needs and personal values.

Strengths of our study include its multicentre design, the inclusion of a healthy control group and the detailed categorisation of exclusive and partial breastfeeding. However, limitations must be acknowledged. The retrospective design introduces recall bias and the lack of infant health outcome data in relation to maternal breastfeeding patterns. Standardised assessments of disease-related organ damage were not consistently available across participating centres and could not be analysed in this study. Future prospective research should incorporate validated damage indices to better understand their potential influence on breastfeeding outcomes. Moreover, although we adjusted for multiple clinical and demographic variables, the possible impact of unmeasured confounders such as maternal employment, access to breastfeeding support, cultural practices, personal beliefs and coexisting medical or psychosocial conditions cannot be excluded. These factors may indirectly affect both breastfeeding behaviour and disease course and should be carefully addressed in future prospective investigations.

Conclusion

SLE and RA mothers had lower rates of breastfeeding and shorter breastfeeding duration compared with control groups. Postpartum depression, lack of breastfeeding intention and delayed initiation were consistent predictors of early cessation. In SLE, glucocorticoid use and higher disease activity may further impair breastfeeding outcomes. Multidisciplinary support, education on lactation-compatible medications and integration of mental health services are essential to support informed maternal choices and improve maternal-infant well-being.

Data availability statement

Data are available upon reasonable request.