The effect of breastfeeding on treatment outcomes in in-vitro fertilization frozen embryo transfer cycles

Abstract

Purpose

While breastfeeding is known to impact unassisted conception, its effect on fertility treatment outcomes remains unclear. This study aimed to evaluate the effect of breastfeeding on in-vitro fertilization (IVF) frozen embryo transfer (FET) outcomes.

Methods

A retrospective cohort study at a single, urban, Canadian Fertility center, including women who underwent an FET(2015–2023). Breastfeeding women (study group) were compared to non-breastfeeding controls, matched 1:1 for age, treatment protocol, and year. We compared outcomes between the two groups. Primary outcomes were pregnancy rate (PR), clinical pregnancy rate (CPR),and live birth rate (LBR). Secondary outcomes included peak endometrial thickness prior to progesterone supplementation, biochemical PR and miscarriage rate. Multivariate logistic regression controlled for confounders. Two sub-analyses within the breastfeeding group examined treatment outcomes by breastfeeding duration (≤ 12 months; > 12 months) and prolactin levels (normal;elevated).

Results

Overall, 118 women met inclusion criteria (59 breastfeeding, 59 controls). Breastfeeding women, compared to controls,were characterized by a younger median age at initial oocyte collection (34.8 years(31.9–37.11) vs. 35.85 years(33.2–39.2),p = 0.044); higher median parity (1(1–1) vs. 0(0–1), p < 0.0001);and higher rates of irregular cycles (35.59% vs. 18.64%,p = 0.016);and elevated prolactin levels (16.95% vs. 1.69%,p = 0.003). Utilizing multivariate logistic regression adjusting for breastfeeding status,parity, age at initial oocyte collection, and cycle regularity, only maternal age at oocyte collection was associated with PR (adjusted odds ratio 0.84, 95% confidence interval 0.75–0.94,p = 0.003). Both sub-analyses revealed similar PRs, CPRs and LBRs between groups.

Conclusions

Breastfeeding during an FET cycle was not associated with treatment outcomes, including PR,CPR, or LBR, providing cautious reassurance to physicians and patients.

Supplementary Information

The online version contains supplementary material available at 10.1007/s10815-025-03556-9.

Introduction

Breastfeeding, described as the feeding of human milk to the infant, is known for its myriad health benefits for both the breastfeeding mother and the offspring, as well as for the mother-infant relationship [1, 2]. These benefits for the breastfed infant include both short- and long-term advantages. Short-term advantages include decreased risks of respiratory infections [3], otitis media [4], neonatal mortality, and sudden infant death syndrome [5], as well as provision of vital nutrients and immune protection, reducing diarrhea [6] and other illnesses [7]. Long-term infant benefits include a lower mean blood pressure, lower prevalence of diabetes, improved cognitive development, and a decreased risk of becoming overweight in adulthood [1, 2, 8]. Benefits for breastfeeding mothers include facilitation of postpartum recovery, reduction in the likelihood of breast and ovarian cancers, and assistance in natural birth spacing, which is especially advantageous in areas with limited healthcare access [7, 8].

Currently, the World Health Organization and the American Academy of Pediatrics (AAP) recommend exclusive breastfeeding for a minimum of six months after delivery, followed by continuation of breastfeeding for up to two years or more after introducing complementary food [3, 9, 10]. Since the duration of this feeding practice may be several years, often during a woman’s fertile years, the question arises whether breastfeeding may affect subsequent in-vitro fertilization (IVF) pregnancy outcomes. Whereas the association between breastfeeding and unassisted conception is known, since breastfeeding may induce hypothalamic amenorrhea, thereby inhibiting ovulation [11], data is scarce regarding the effect of breastfeeding on fertility treatment outcomes in the population with infertility, requiring IVF to conceive. These patients’ treatment outcomes may be affected through several mechanisms: 1) high prolactin levels that may impair progesterone secretion and alter hormonal conditions for adequate embryo implantation [11], with adverse effects on follicular or oocyte development during IVF [12–14]; 2) some hormonal activity of prolactin, even with normal hormone levels. An additional point to consider when contemplating the use of fertility drugs to conceive, is the effect of some of these drugs on the offspring, if passed through breast milk. While some are not harmful to the nursing infant but may reduce the mother’s milk supply, others may potentially be dangerous, with unknown effects on the nursing infant (such as Letrozole (an Aromatase inhibitor) and Dehydroepiandrosterone (DHEA)) [15].

These concerns, combined with the paucity of data in the literature regarding the effect of breastfeeding on IVF treatment outcomes, has led various fertility clinics worldwide to recommend cessation of breastfeeding before commencing IVF treatment, whereas others enable treatment on the condition that prolactin levels have normalized, and menses have resumed. This practice may additionally vary among physicians within the same clinic, most likely due to lack of true evidence regarding any effect of breastfeeding on IVF outcomes.

Hence, the aim of our study was to examine the effect of breastfeeding on frozen embryo transfer (FET) cycle outcomes in a population of patients with infertility at a single center.

Materials and methods

Study design and participants

This was a single-center retrospective cohort study conducted at an urban fertility center in Canada, including women who underwent an FET during 2015–2023. The study group consisted of women undergoing FET with a documented positive breastfeeding status (to any extent, i.e., any breastfeeding (exclusive or partial)) according to the patient’s statement at the last patient-physician medical encounter prior to FET cycle commencement). The control group consisted of non-breastfeeding women, matched in a 1:1 ratio to patients in the study group according to age, treatment protocol (hormone replacement therapy FET versus natural cycle FET, detailed below), and year of treatment. Each patient in both the study and control groups was included only once during the study period, contributing one FET cycle each. Notably, women who stated they were breastfeeding but intended to stop prior to treatment commencement; who were counselled to stop breastfeeding prior to the FET cycle; or who reported breastfeeding cessation prior to or during the FET cycle, were excluded from the analysis. Additionally, all breastfeeding women in the study were actively breastfeeding or pumping breast milk for their infant, with no use of donor breast milk.

Frozen embryo transfer protocols

The FET regimens were determined based on physician discretion, cycle regularity and patient preference. Notably, a regular cycle was defined as any cycle between 21–35 days that the patient stated as regular; any cycles not meeting these criteria were considered irregular. Two standard FET regimens were employed: (i) the programmed protocol, also called the hormone replacement therapy (HRT) protocol, which included oral administration of micronized 17β-estradiol at a starting dose of 2 mg three times daily. 17β-estradiol dose was increased by 2–4 mg increments added vaginally, according to physician discretion, when endometrial thickness was below 7 mm on day 11 after starting treatment; (ii) the modified natural cycle (NC) protocol, which included a transvaginal ultrasound on day 10 of a patient’s menstrual cycle, and every few days thereafter as needed, to assess for endometrial thickness and the presence of a dominant follicle. Luteinizing hormone (LH) and progesterone levels were taken during the NC protocol visits, and ovulation triggering was performed through administration of 250 mcg subcutaneous recombinant-hCG when the leading follicle reached 17 mm or more and endometrial thickness was 7 mm or more, with a tri-laminar appearance. If LH and progesterone levels indicated spontaneous LH surge prior to ovulation triggering, FET was planned accordingly, with or without an additional trigger shot.

As described, in both protocols, endometrial thickness and the presence or absence of a dominant follicle was monitored using serial transvaginal ultrasonography, with the addition of serum estradiol, LH and progesterone measurements as needed, with 17β-estradiol doses adjusted as required in the HRT group. When endometrial thickness reached 7 mm or more with a trilaminar appearance, FET was planned, with the transfer of a frozen-thawed cleavage stage embryo (day 2–3) or a blastocyst (day 5–6). In the HRT-FET group, vaginal progesterone (400 mg daily) or intra-muscular progesterone (100 mg every other day) were added according to the planned day of transfer; and in the NC-FET group, no progesterone was given prior to or after embryo transfer, unless otherwise indicated by the treating physician (supplemented NC-FET), with the embryo transfer timed according to the time of ovulation.

Data collection

Data encompassed patient demographics, infertility workup, prolactin levels (measured in ug/l, taken in the morning at least two hours after awakening), and cycle characteristics. For the breastfeeding group, median breastfeeding duration was also assessed. Notably, documented prolactin levels were the last prolactin levels taken prior to FET cycle commencement and had to be, per clinic guidelines, within one year of cycle commencement. In cases of breastfeeding patients, this time span obligatorily included the breastfeeding period.

Outcomes

The primary outcomes included pregnancy rate (defined as a positive β-human chorionic gonadotropin (βhCG)), clinical pregnancy rate (defined as the visualization of a fetal pole with a positive heartbeat), and live birth rate. Secondary outcomes included peak endometrial thickness prior to progesterone supplementation; biochemical pregnancy rate (defined as a positive βhCG that declined without visualization of a gestational sac on ultrasound); and miscarriage rate (defined as an arrested pregnancy following visualization of a gestational sac or additional fetal structures on ultrasound examination). Treatment outcomes were compared between the study and control groups. Additionally, two sub-analyses were performed within the breastfeeding group, according to: i) breastfeeding duration (defined as the duration between the birth of the breastfeeding infant and the start of the FET treatment cycle) – baseline characteristics and treatment outcomes were compared between women who breastfed for ≤ 12 months and those who breastfeed for > 12 months at the time of the FET; ii) prolactin level—baseline characteristics and treatment outcomes were compared between women with a normal prolactin level and those with an elevated prolactin level (defined as values above 26.7 ug/l, according to our center’s lab reference range) at the time of the FET.

Statistical analyses

Categorical data were presented by the number of cases and percentages, while continuous data were described using the median and interquartile range (IQR). Categorical data were subjected to Pearson Chi-square analysis and Fisher’s exact test, while continuous data underwent the Student’s t-test and Mann–Whitney-Wilcoxon test as appropriate. Multivariate logistic regression analysis and Analysis of Variance (ANOVA) were conducted to explore associations between breastfeeding and categorical and continuous FET treatment outcomes, respectively. Logistic regression was conducted through the estimation of adjusted odds ratio (aOR) and 95% confidence intervals (CI). The ANOVA and regression models adjusted for the potential confounding effects of maternal demographics, pre-existing clinical characteristics, and concurrently occurring conditions where p < 0.05 on the chi-squared tests. Computations were performed using SAS Software, Version 9.4 (Cary, North Carolina, United States).

Ethics

The study received approval from the ethics committee of the McGill University Health Center (MUHC) Reproductive Center (MUHC REB 2024–10239). Informed consent was waived due to the retrospective design of the study.

Results

Cohort characteristics

Overall, 118 women met the inclusion criteria. Of them, 59 women were in the study group, and 59 women were in the control group. Median breastfeeding duration was 14 months (IQR 11–20 months). Table 1 presents the baseline, treatment and cycle characteristics of the study and control groups. Women in the study group, compared to controls, were characterized by a younger median age at initial oocyte collection (34.8 years (inter-quartile range (IQR)) 31.9–37.11 vs. 35.85 (33.2–39.2), p = 0.044); higher median parity (1 (1–1) vs. 0 (0–1), p < 0.0001); a higher rate of previous cesarean delivery (CD) (37.29% vs. 11.86%, p = 0.002); a higher rate of an infertility diagnosis of unexplained infertility (30.51% vs. 13.56%, p = 0.003); a higher rate of elevated prolactin levels (16.95% vs. 1.69%, p = 0.003); and a higher rate of irregular cycles (35.59% vs. 18.64%, p = 0.016).

Table 1.

Baseline and cycle characteristics of the cohort undergoing frozen embryo transfer cycles

Variable	Breastfeeding (N = 59)	Not breastfeeding (controls) (N = 59)	p-value*
Embryo transfer cycle numbera	3 (2–4)	3 (2–4)	0.777
Maternal age at FET (years)	38.10 (35.7–40.3)	38.10 (36.11–40.7)	0.99
Maternal age at initial oocyte collection (years)	34.8 (31.9–37.11)	35.85 (33.2–39.2)	0.044
Breastfeeding duration (months)	14 (11–20)	NA	NA
Body mass index (kg/m2)	22.82 (21.35–25.74)	25 (21.22–31.89)	0.331
Parity	1 (1–1)	0 (0–1)	< 0.0001
Previous CD	22 (37.29)	7 (11.86)	0.002
Fibroids on baseline ultrasound	11 (18.64)	14 (23.73)	0.487
Endometriosis on baseline ultrasound	2 (3.39)	1 (1.69)	1
Infertility etiology:			0.003
Ovulatory disorder	7 (11.86)	3 (5.08)
Male factor	15 (25.42)	16 (27.12)
Mechanical/tubal	5 (8.47)	4 (6.78)
Advanced maternal age Unexplained	4 (6.78) 18 (30.51)	6 (10.17) 8 (13.56)
Combined	1 (1.69)	14 (23.73)
Diminished ovarian reserve	6 (10.17)	1 (1.69)
PGT	0 (0)	3 (5.08)
Endometriosis	2 (3.39)	1 (1.69)
Other	1 (1.69)	3 (5.08)
RIF	1 (1.69)	4 (6.78)	0.364
Irregular cyclesb	21 (35.59)	11 (18.64)	0.016
Prolactin level (ug/l)	13.75 (9.85–26.9)	8.8 (6.1–11)	0.0003
Elevated prolactin	10 (16.95)	1 (1.69)	0.003
Treatment:			1.0
HRT-FET	49 (83.05)	50 (84.75)
NC-FET	10 (16.95)	9 (15.25)
17β-estradiol daily dose (mg) (where applicable)	6 (6–6)	6 (6–6)	0.752
Progestin type:			0.73
Unknown	5 (8.47)	2 (3.39)
Vaginal (400 mg daily)	38 (64.41)	39 (66.1)
Intra-muscular (100 mg every other day)	5 (8.47)	9 (15.25)
Vaginal + intra-muscular	4 (6.78)	3 (5.08)
No progesterone	7 (11.86)	6 (10.17)
Number of embryos transferred	1 (1–1)	1 (1–1)	0.137
Embryo day at transfer:			0.057
Unknown	1 (1.69)	0 (0)
Cleavage	0 (0)	5 (8.47)
Blastocyst	58 (98.31)	54 (91.53)
Embryo quality (cleavage or blastocyst):			0.648
Goodc	57 (96.61)	56 (94.92)
Moderated	2 (3.39)	3 (5.08)

Continuous variables are presented as median (inter-quartile range (IQR)) and categorical variables are presented as n (%); *Categorical data were subjected to Pearson Chi-square analysis and Fisher’s exact test, and continuous data underwent the Student’s t-test and Wilcoxon test as appropriate

Abbreviations: FET – frozen embryo transfer; NA – non applicable; CD – cesarean delivery; PGT – preimplantation genetic testing; RIF – repeated implantation failure (defined as a negative βhCG following the transfer of three or more good-quality blastocysts); HRT-FET – hormone replacement therapy frozen embryo transfer; NC-FET – natural cycle frozen embryo transfer

^aEmbryo transfer cycle number – including both fresh and frozen embryo transfer cycles

^bIrregular cycles – a regular cycle was defined as any cycle between 21–35 days that the patient stated as regular; any cycles not meeting these criteria were considered irregular

^cGood-quality embryo – defined as an embryo with at least three cells on day 2 and six cells on day 3 with < 20% anucleate fragments and no apparent morphological abnormalities, or as a blastocyst with an expansion grade of 3 and above, and an inner cell mass and trophectoderm grade of A or B (for both), as defined by Gardner and Schoolcraft’s criteria[23]

^dModerate-quality embryo – any embryo not meeting criteria for good-quality embryos (as above) and considered adequate for embryo transfer

FET treatment outcomes

Table 2 depicts the FET treatment outcomes in the study and control groups. In the univariate analysis, peak endometrial thickness prior to progesterone supplementation was found to be higher in breastfeeding women compared to controls (9.85mm (8.6–11) vs. 8.6mm (8.1–10), p = 0.006), with comparable pregnancy, biochemical pregnancy, clinical pregnancy, miscarriage, and live birth rates.

Table 2.

Treatment outcomes of the frozen embryo transfer cohort

Variable	Breastfeeding (N = 59)	Not breastfeeding (controls) (N = 59)	cOR (95% CI)	p-value*
Peak endometrial thickness (mm)	9.85 (8.6–11)	8.6 (8.1–10)	N/A	0.006
Positive β-hCG	32 (54.24)	30 (50.85)	0.94 (0.51–1.73)	0.854
Biochemical pregnancya	8 (13.56)	10 (16.95)	1.25 (0.46–3.39)	0.579
Miscarriageb	4/32 (12.5)	2/30 (6.67)	0.53 (0.09–3.13)	0.438
Clinical pregnancyc	24 (40.68)	20 (33.9)	0.83 (0.42–1.67)	0.382
Live birthd	12 (20.34)	10 (16.95)	1.25 (0.49–3.17)	0.814

Continuous variables are presented as median (inter-quartile range (IQR)) and categorical variables are presented as n (%); *Categorical data were subjected to Pearson Chi-square analysis and Fisher’s exact test, and continuous data underwent the Student’s t-test and Wilcoxon test as appropriate

Abbreviations: cOR – crude odds ratio; CI – confidence interval; N/A – non-applicable; β-hCG – β-human chorionic gonadotropin

^aBiochemical pregnancy rate – calculated out of all of those who had an embryo transfer

^bMiscarriage rate – calculated out of all positive β-hCGs

^cClinical pregnancy rate – calculated out of all of those who had an embryo transfer

^dLive birth rate—calculated out of all of those who had an embryo transfer

ANOVA and logistic regression analyses

We utilized ANOVA and multivariate logistic regression analysis to explore the associations between breastfeeding and the following FET treatment outcomes: peak endometrial thickness prior to progesterone supplementation, pregnancy rate, clinical pregnancy rate, and live birth rate. In the multivariate logistic regression and ANOVA adjusting for breastfeeding status, parity, maternal age at initial oocyte collection, and cycle regularity, breastfeeding status was not found to be associated with peak endometrial thickness (p = 0.279), pregnancy rate (aOR 0.5, 95% CI 0.16–1.58, p = 0.239), clinical pregnancy rate (aOR 2.8, 95% CI 0.47–16.52, p = 0.257), or live birth rate (aOR 2.8, 95% CI 0.47–16.52, p = 0.257) (Table 3). The only factor found to be associated with pregnancy rate was initial maternal age at oocyte collection, with advancing age associated with decreased pregnancy rates (aOR 0.84, 95% CI 0.75–0.94, p = 0.003).

Table 3.

Multivariate logistic regression of factors associated with pregnancy rate, clinical pregnancy rate, and live birth rate, in frozen embryo transfer cycles

Pregnancy rate
Variable	aOR (95% CI)	p-value
Breastfeeding versus no-breastfeeding	0.5 (0.16–1.58)	0.239
Maternal age at initial oocyte collection	0.84 (0.75–0.94)	0.003
Parity	1.8 (0.73–4.44)	0.198
Cycle regularitya	1.13 (0.43–3.0)	0.804
Clinical pregnancy rate
Variable	aOR (95% CI)	p-value
Breastfeeding versus no-breastfeeding	2.8 (0.47–16.52)	0.257
Maternal age at initial oocyte collection	0.93 (0.79–1.1)	0.409
Parity	0.69 (0.22–2.12)	0.511
Cycle regularitya	1.81 (0.36–8.98)	0.471
Live birth rate
Variable	aOR (95% CI)	p-value
Breastfeeding versus no-breastfeeding	1.55 (0.43–5.66)	0.506
Maternal age at initial oocyte collection	0.94 (0.84–1.06)	0.312
Parity	1.29 (0.47–3.51)	0.619
Cycle regularitya	1.36 (0.43–4.37)	0.602

Abbreviations: aOR – adjusted odds ratio; CI – confidence interval

^aCycle regularity – a regular cycle was defined as any cycle between 21–35 days that the patient stated as regular; any cycles not meeting these criteria were considered irregular

Sub-group analyses

Supplemental Table 1 presents the first sub-analysis, depicting baseline and cycle characteristics and treatment outcomes of breastfeeding women who breastfed for ≤ 12 months compared to those who breastfed for > 12 months at the time of treatment. Notably, data regarding breastfeeding duration was not available for one patient, such that there were 21 patients evaluated in the group who breastfed for ≤ 12 months, and 37 patients in the group who breastfed for > 12 months at the time of FET. As presented in the table, there was an increased rate of no progesterone use during the luteal phase in the extended breastfeeding duration group (> 12 months), compared to patients breastfeeding for ≤ 12 months (16.22% with no progesterone use vs. 4.76%, respectively, p = 0.046). This coincides with the higher rate of NC-FET, which often is prescribed with no luteal progesterone support, in the extended breastfeeding group compared to the ≤ 12 months breastfeeding group, though this difference was not statistically significant (21.62% vs. 9.52%, respectively, p = 0.301). Other baseline and cycle characteristics were similar between groups. Regarding treatment outcomes, in the univariate analysis, median peak endometrial thickness was found to be higher in the extended breastfeeding group compared to those breastfeeding ≤ 12 months (10mm (9–11) vs. 8.8mm (8.3–11), p = 0.039), with comparable pregnancy, biochemical pregnancy, miscarriage, clinical pregnancy, and live birth rates.

Supplemental Table 2 presents the second sub-analysis, depicting baseline and cycle characteristics and treatment outcomes of breastfeeding women with normal prolactin levels (n = 30) compared to those with elevated prolactin levels (n = 10). Notably, data regarding prolactin levels at the time of treatment was not available for 19 out of 59 breastfeeding patients. As presented in the table, the breastfeeding group with elevated prolactin, compared to the breastfeeding group with normal prolactin, was characterized by a higher rate of irregular cycles (60% vs. 20%, p = 0.011) and a lower rate of male factor infertility as the infertility etiology (0% vs. 36.67%, p = 0.018). Other baseline and cycle characteristics were similar between groups. Regarding treatment outcomes, there were no detectable between-group differences in the univariate analysis with respect to peak endometrial thickness, pregnancy, biochemical pregnancy, miscarriage, clinical pregnancy, and live birth rates.

Discussion

Main findings

We aimed to evaluate the effect of breastfeeding on FET cycle treatment outcomes at a single center. Following multivariate logistic regression, breastfeeding status was not found to be associated with peak endometrial thickness prior to progesterone supplementation, pregnancy rate, clinical pregnancy rate, or live birth rate. In the univariate analysis of the first sub-analysis, extended breastfeeding > 12 months was associated with higher median peak endometrial thickness compared to those breastfeeding ≤ 12 months, with comparable pregnancy, biochemical pregnancy, miscarriage, clinical pregnancy, and live birth rates. In the univariate analysis of the second sub-analysis comparing breastfeeding women with normal versus elevated prolactin levels, we found no detectable between-group differences in peak endometrial thickness, pregnancy, biochemical pregnancy, miscarriage, clinical pregnancy, and live birth rates.

Interpretation

As previously noted, the effect of breastfeeding on FET treatment outcomes has not been examined to date. However, hyperprolactinemia (due to a prolactinoma, for example) has been examined concerning its effect on spontaneous conception and pregnancy complications [11, 16]. Prolactin, an extremely versatile hormone, has been demonstrated to have regulatory functions in the reproductive system, immune system, and during breastfeeding [17]. Notably, a previous study [18] examining the effect of high prolactin levels on endometrial morphology in young post-pubertal and adult mice rendered hyperprolactinemic by long-term treatment with metoclopramide, demonstrated that both young and adult metoclopramide-treated mice presented an increased number of endometrial glands compared to their respective controls. They concluded that metoclopramide-induced hyperprolactinemia causes mouse endometrium proliferation, mainly in young animals. However, this association has only been demonstrated in a mouse model, and, importantly, not all breastfeeding women necessarily display increased prolactin levels.

Similarly, the lack of association in our study between breastfeeding and pregnancy and clinical pregnancy rates may simply represent the capability of IVF-FET cycles to circumvent the possible negative effect on spontaneous conception brought on by pituitary inhibition of ovulation in cases of high prolactin levels. Moreover, notably, in our multivariate logistic regression, cycle regularity (which may reflect normal/elevated prolactin levels) was not found to be associated with pregnancy rate or clinical pregnancy rate, further supporting this point. It seems like the hormonal milieu brought on by breastfeeding, whether affecting prolactin levels and cycle regularity or not, does not affect the chances of implantation and clinical pregnancy in FET cycles. Since both embryonic and endometrial factors play a pivotal role in embryo implantation, and the embryo in cases of FET cycles is a fixed factor separate from breastfeeding status, our results may imply that breastfeeding does not significantly affect the endometrium, thus exerting no significant impact on embryo implantation.

The only factor we found to be associated with pregnancy rate was initial maternal age at oocyte collection, with advancing maternal age associated with decreased pregnancy rates. This finding accords with known literature regarding the link between advancing maternal age and declining fertility [19].

We found an increased median peak endometrial thickness in the extended breastfeeding group (> 12 months) compared to those breastfeeding ≤ 12 months in the sub-analysis stratified by breastfeeding duration. This may also be explained by the findings of the aforementioned study [18] carried out on mice, that demonstrated increased endometrial proliferation in the face of hyperprolactinemia, which may display itself differentially in extended versus shorter-term breastfeeding duration. Notably, it has been shown that breast stimulation in long-term lactating women continues to produce a slight increase in serum prolactin levels [20]. However, one must note that prolactin levels in both breastfeeding duration groups were comparable, questioning the plausibility of this explanation. Whether breastfeeding duration exerts a true effect on endometrial thickness during FET cycles, with any resultant impact on pregnancy, clinical pregnancy, or live birth rates, will need to be explored in future large prospective studies.

Strengths and limitations

Our study has several strengths. First, to the best of our knowledge, it is the first and only study to date to evaluate the effect of breastfeeding on IVF-FET treatment outcomes, thus providing scientific grounds on which to base future patient counseling in such cases. Second, our study’s single-center design and matching of controls according to year of treatment ensured similar treatment strategies, laboratory techniques, and medical follow-up of these patients during the study period, enabling a more accurate comparison between groups.

Our study is not without limitations. First, due to its retrospective nature, some data are missing, such as the extent of breastfeeding (exclusive versus partial). Additionally, breastfeeding duration varied between included women in the study group, and possible differences in endocrine regulation may be present in short-term (≤ 1 year) versus extended breastfeeding (> 1 year), potentially affecting our results. Moreover, we could not verify with certainty that patients who stated they were breastfeeding actually were breastfeeding, exactly how extensive any breastfeeding was, and if they stopped breastfeeding during the process. Having said that, this bias is likely to have been reduced by our rigorous inclusion of patients who were not told to stop breastfeeding prior to treatment commencement and did not express their wish to do so, and due to the relatively short duration between statement of breastfeeding and treatment start. Moreover, one must note that even if some patients stopped breastfeeding at some point during or prior to treatment, breastfeeding’s hormonal effect on ovulation and menstruation could take 1–2 months to dissipate [21], rendering these patients more similar to actual breastfeeding patients in that respect, though we do recognize this as a limitation of the study. Additionally, inherent diurnal variation in prolactin levels may pose another limitation (with plasma concentration being highest during sleep and lowest during the day [22]), though this could be assumed to be distributed similarly between the study and control groups. Second, due to our single-center design, our findings may not be generalizable to other centers. Lastly, due to our small sample size, we may have been underpowered to detect actual differences between groups, if such exist. Nevertheless, due to the varying practice worldwide with respect to initiation of a FET cycle in breastfeeding patients, this single-center study provides a basis with which to counsel and advise physicians and patients. Repetition of this study in a large prospective database would be ideal, though may be inherently difficult to conduct. Future studies may circumvent some of these limitations by implementing routine documentation detailing the extent of breastfeeding (exclusive vs. partial). Furthermore, conducting these studies with a prospective design to evaluate treatment cycle outcomes could help overcome the biases associated with retrospective studies.

Conclusions

In conclusion, we did not find any negative association between breastfeeding during a FET cycle and treatment outcomes, including peak endometrial thickness prior to progesterone supplementation, pregnancy rate, clinical pregnancy rate, or live birth rate. Our findings may cautiously reassure physicians and breastfeeding patients alike, that FET treatment outcomes do not seem to be affected by concurrent breastfeeding. Given that breastfeeding tends to occur during a woman’s reproductive years, this finding may help alleviate physician and patient anxiety when contemplating conception through FET, while allowing the mother to continue to provide her breastfeeding child with breastfeeding’s myriad health benefits. Future large prospective studies should be conducted to further elucidate the effect of breastfeeding on FET treatment outcomes.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

AH, UA, and WB designed the study and helped with data acquisition, analysis, and interpretation. AH, NK, VB, TS and ES participated in data acquisition, analysis, interpretation, and article revision for critical intellectual content. All authors contributed intellectually to the writing or revision of the manuscript, approved the final version, and are accountable for all aspects of the work.

Funding

This research received no specific grant from public, commercial, or not-for-profit funding agencies.

Data availability

All data supporting the findings of this study are available within the paper.

Declarations

Ethics approval

This study was performed in line with the principles of the Declaration of Helsinki. The study received approval from the ethics committee of the McGill University Health Center (MUHC) Reproductive Center (MUHC REB 2024–10239).

Informed consent

Informed consent was waived due to the retrospective design of the study.

Competing interest

The authors have no relevant financial or non-financial interests to disclose.