The study population was drawn from the postpartum clinic of the Fourth People’s Hospital in Zigong City, Sichuan Province, China. This group comprised lactating women who expressed their milk rather than directly nursing their infants due to various reasons. We conducted a retrospective analysis of their subsequent 4 months of breastfeeding. This study received ethical approval from the Institutional Review Board of the Fourth People’s Hospital in Zigong City, Sichuan Province, China (approval number: (003), 2022). The need for obtaining informed consent was waived due to the retrospective nature of the study.

We studied lactating women who had regular follow-up visits at the postpartum clinic of the Fourth People’s Hospital in Zigong City, Sichuan Province, from 07/12/2022 to 01/04/2023. Inclusion criteria for the study population were as follows: ① Women who had regular follow-up visits at the postpartum clinic of the Fourth People’s Hospital in Zigong City, Sichuan Province; ② Inability to breastfeed directly due to various reasons, necessitating milk expression, with a breast milk yield that had reached a balanced supply and did not require artificial feeding; ③ Age $\ge$16 years and the ability to communicate effectively. Exclusion criteria included: ① Inability to express milk or maintain milk emptying due to physiological or pathological factors; ② Presence of severe mental illness or mental health issues preventing cooperation with data collection. The study population was classified into infected and uninfected groups based on current COVID-19 infection status, defined as a positive result in either the COVID-19 nucleic acid test (reverse transcription-polymerase chain reaction/polymerase chain reaction (RT-PCR/PCR)) or the COVID-19 antigen test. Any newly emerging COVID-19 infections in the non-infected group during the follow-up period would resulted in reclassification into the infected group.

Clinical data were collected using self-designed structured network questionnaires, encompassing relevant demographic and clinical characteristics of the participants. Baseline information including age, height, weight, body mass index (BMI), gravidity, parity, labor time (preterm/full-term, delivery mode (vaginal birth/cesarean birth/cesarean birth after vaginal birth), educational level (no undergraduate/undergraduate and above), COVID-19 vaccination history, milking collection method (machine/others), sleep quality, breast pain conditions (using Visual Analogue Scoring (VAS) [22], results were divided into painless (0), mild (1–3), moderate (4–6), and severe (7–10)), breast swelling conditions (no swelling-normal breasts, mild-does not affect maternal rest, moderate-interferes with maternal rest, severe-unable to rest)). For the infected group, we collected the sleep quality [23], breast pain conditions, and breast swelling conditions a month before (1) and after infection (2); for the uninfected group, we collected the information a month before (1) and after enrollment (2). During the follow-up questionnaire, we collected both COVID-19-related data and breastfeeding-related data. The volume of breast milk was documented based on the subjective judgment of the lactating woman. The milk volume 1 week before infection or enrollment was designated as 100%, and subsequent weekly production was recorded and compared to this baseline (due to poor compliance, consecutive recording was not maintained over 2–3 months) (See Supplementary Material for data).

Data were processed and analyzed using SPSS statistical analysis software (version 27.0, IBM Corp., Armonk, NY, USA). Continuous data were assessed using the independent t-test or Mann-Whitney rank sum test, while categorical data were analyzed using the $\chi$${}^2$ test. A significance level of p 0.05 was applied for statistical significance. Logistic regression analysis was employed to determine the odds ratio (OR) of COVID-19 infection associated with decreased breast milk production.

Our study comprised 79 lactating women, with 45 in the infected group and 34 in the uninfected group. The infected group exhibited statistically significant differences in age (younger), educational level (higher), and sleep quality (better) compared to the non-infected group. However, there were no statistically significant differences between the two groups in terms of mothers’ height, weight, BMI, gravidity and parity, delivery mode, delivery time, and COVID-19 vaccine history (Table 1).

In our study, 45 women experienced a decrease in breast milk production during the follow-up period: 31 cases in the COVID-19 infected group and 14 cases in the non-infected group. The rate of breast milk production decrease exhibited a statistically significant difference between the COVID-19 infected and uninfected groups ($\chi$${}^2$ = 6.067, p 0.05). Univariate logistic regression analysis revealed an elevated risk of decreased breast milk production in the COVID-19 infected group compared to the non-infected group (OR 3.163, 95% confidence interval (CI): 1.248–8.017). After adjusting for confounding factors such as age, educational level, and sleep quality using multivariate logistic analysis, the elevated risk of decreased breast milk secretion in the infected group persisted (OR 3.339, 95% CI: 1.134–9.833).

In our research, the average time of COVID-19 infection occurred during the fourteenth week after delivery. The majority of infections were of the mild type. The decline in breast milk production commenced on the fifth day following COVID-19 infection and reached its nadir during the second week post-infection (Fig. 1). Among lactating women experiencing reduced breast milk production 67.7% opted to increase their rest to facilitate breast milk recovery. However, by the end of our retrospective analysis, only 15.2% of women had seen their breast milk production return to normal levels.

Fig. 1.

Fluctuations in milk production after infection of COVID-19 in the infected group. (A) The boxes represent the interquartile range with breast milk production up to 16 weeks after infection. Whiskers present the data range (Q1/Q3 $\pm$ 1.5 IQR). (B) Showing the fluctuations of median breast milk after infection. * represents the median breast milk production among corresponding weeks. IQR, interquartile range.

In our study, 37.8% of infected women opted to discontinue breastfeeding and switch to artificial feeding, while 62.2% of infected women chose to continue breastfeeding. Among those who continued breastfeeding, 25% experienced insufficient breastfeeding during the period of decreased milk production (insufficient means that artificial feeding milk is more than half) (Table 2). During the follow-up period, only 1 case involved a child with sleep disturbances, while the others did not exhibit abnormal symptoms such as irritability, drowsiness, abdominal pain, or diarrhea.

In our study, the proportion of people with sufficient milk decreased from 88.9% to 17.9% after infected, and the proportion of people with insufficient milk increased from 0 to 25%. We speculate that infection with the novel coronavirus may affect the production of breast milk. The average age, height, and weight of children in the COVID-19 infected group who experienced continued breastfeeding did not exhibit a statistically significant difference when compared to the COVID-19 uninfected group at the baseline level. Furthermore, during the 4-month follow-up period, there remained no statistically significant difference in height between the two groups. Concerning infant weight, the infected group was heavier than the uninfected group in the third and fourth months, and this difference was statistically significant (Table 3 and Fig. 2). This study concluded that COVID-19-infected women’s breast milk did not have a significant negative impact on short-term growth and development in their children.

Fig. 2.

The comparison of the growth and development status of the infants. The boxes represent the interquartile range with height/weight up to 4 months after follow-up. Whiskers present the data range (Q1/Q3 $\pm$ 1.5 IQR). (A) Showing the height changes during follow-up. (B) Showing the weight changes during follow-up.

Our research revealed an elevated risk of reduced breast milk production in the COVID-19 infected group compared to the non-infected group (OR 3.163, 95% CI: 1.248–8.017). Even after adjusting for confounding factors, this increased risk persisted (OR 3.339, 95% CI: 1.134–9.833). Among women who continued breastfeeding without interruption, we observed that their breast milk had no adverse impact on short-term growth and development in their children.

Within the first 24 hours postpartum, breast milk production increases rapidly, and it stabilizes at approximately 750–800 mL/day after one month [24]. In our research, 91.1% of women entered into the relatively stable stage. The breast milk secreted by the mammary epithelial cell (MEC), upon the effect of oxytocin (OT), the milk expelled out of the alveoli into the ducts, and simultaneously under the function of milk ejection reflex mediated by OT, ultimately leads to milk ejection [25]. Our research findings indicated that the risk of reduced breast milk production increased following COVID-19 infection. In our study, all women used a breast pump to express milk, and after infection, there were no significant differences in breast symptoms between the infected and non-infected groups. This supports our finding that the reduction in milk production is primarily due to decreased breast milk secretion rather than abnormal breast emptying. Several factors influence breast milk production, including genetics, diet, hormones, and the environment [26]. One possible reason is that the majority of COVID-19 infected women experienced increased body temperature. In our study, 77.8% of women in the infected group had fever (defined as a body temperature exceeding 37.3 °C). This fever can accelerate metabolism and increase the body’s water consumption, which may partly contribute to the reduced breast milk production. Another significant factor is mental health, which can influence the secretion of OT, a hormone crucial for breast milk ejection. Additionally, psychologic distress can lead to elevated cortisol levels and decreased insulin levels in the serum, both of which are associated with reduced milk production [27, 28]. During the COVID-19 pandemic, women’s mental health is significantly affected, with a noticeable increase in symptoms of depression, anxiety, and stress. This increase is particularly pronounced among infected women, who exhibit a higher incidence of depression and anxiety [29, 30, 31, 32]. A meta-analysis showed that the ORs of COVID-19 pandemic for postpartum depression and anxiety symptoms were 1.54 (95 % CI: 1.00–2.36) and 2.56 (95% CI: 1.62–4.04) [33]. Hormones levels, except for those indirectly influenced by the mental factors, were also found to be directly influenced by the COVID-19 infection. One crucial hormone is insulin. Research has reported that mothers with insulin resistance are at an elevated risk of insufficient milk production and delayed lactogenesis [34, 35]. Reports have indicated that COVID-19 can induce insulin resistance in patients. Furthermore, the angiotensin-converting enzyme 2 (ACE2), which serves as the receptor for the severe acute respiratory syndrome (SARS) coronavirus, has been found to be expressed in the endocrine part of the pancreas. This suggests that the virus may enter islets using ACE2 and potentially harm them [36]. The reduction in insulin levels, is associated with decreased milk production. However, a meta-analysis demonstrated that COVID-19 infection has no impact on prolactin (PRL) levels [37]. The influence of COVID-19 infection for other hormones needs further exploration.

Breast milk is rich in energy and nutrients, offering children appropriate amounts of protein, fat, lactose, vitamins, iron, minerals, enzymes, and water. These nutrients are easily digested and absorbed, meeting all the nutritional needs of infants under 6 months of age. In light of this, during the COVID-19 pandemic, international and national health organizations, including the World Health Organization (WHO) and the American Academy of Pediatrics, advocated for mothers to continue breastfeeding, even if they are infected [21]. In contrast, the Chinese consensus differs, as they recommend breastfeeding only when both the PCR test for mothers and breast milk is negative [38]. Regarding the safety of breastfeeding among COVID-19 infected women, there is no explicit evidence demonstrating that the virus can be transmitted through breast milk [18]. Several case reports have identified the presence of viral nucleic acid in breast milk [39, 40]. However, none of these cases have cultured the virus. Therefore, it remains uncertain whether breast milk contains active infectious viruses and whether the virus originates from the milk rather than the infected environment or droplets from infected mothers. According to recommendations from international organizations, when a mother is infected with COVID-19, it is considered safe for an uninfected caregiver to handle feeding. Alternatively, mothers should wear masks and practice breast and hand hygiene. Additionally, breast pumps should be thoroughly cleaned following pumping sessions [19]. Several studies have found that the existence of antibodies against SARS-CoV-2 in breastmilk after being infected or receiving a vaccine, allowing the infant to directly acquire the antibodies against SARS-CoV-2 and confer effective passive immunity [41]. In our research, all the mothers use the pump breastmilk to feed. Following COVID-19 infection, 68.8% of mothers experienced a decrease in breast milk production, reaching its lowest point during the second week after infection. This timeframe is later than the average duration of the worst symptom days among infected individuals (4.9 $\pm$ 2.2). Subsequently, milk production gradually increased. Mothers in this group can attempt to increase milk production by increasing rest intervals and employing localized massage techniques. Our study also revealed that, despite the majority of patients experiencing reduced milk production, only 25% encountered inadequate breastfeeding. Therefore, there is no need for excessive anxiety. Furthermore, our research indicated that breast milk from infected mothers did not negatively impact the height and weight of infants. Based on previous studies, our research supported the continuation of breastfeeding for mothers infected with COVID-19.

The limitations of our study include: first, it is a retrospective study, and retrospective bias is inevitable, potentially impacting result accuracy; second, regarding psychological factors, we intended to collect data on depression and anxiety using standardized scales. However, due to poor compliance among women, we were unable to collect this information.