Maternal probiotic milk intake during pregnancy and breastfeeding complications in the Norwegian Mother and Child Cohort Study

We hypothesized that free-market probiotic milk intake during pregnancy prevents breastfeeding complications and thus promotes longer breastfeeding by stabilizing healthy breast flora. The aim of this study was to evaluate whether intake of probiotic milk products during pregnancy is associated with less breastfeeding complications, i.e., mastitis, medication for mastitis, sore nipples or other problems, or associated with longer breastfeeding duration (no cessation of any or predominant breastfeeding before 4 months).

The MoBa is a prospective, population-based pregnancy cohort study conducted by the Norwegian Institute of Public Health [17]. Participants were recruited from all over Norway from 1999 to 2008, and 41% of invited women consented to participate. Follow-up is conducted by questionnaires at regular intervals and by linkage to pregnancy and birth records in the Norwegian Medical Birth Register (NMBR) [18]. All questionnaires (Q) are available on the website of the Norwegian Institute of Public Health [19].

This study is based on version 10 of the quality-assured data files released for research in 2017.

Out of 114,240 births registered in MoBa, all singleton pregnancies with live births after gestational week 37 + 0 were included in the study. Women had to have filled in questionnaires Q1 on general health and lifestyle, Q2 on dietary habits during pregnancy, and Q4 on follow-up 6 month postpartum. As a quality measure of Q2, a food frequency questionnaire (FFQ), only women reporting an energy intake between 4.5 and 20 megajoules (MJ) daily were included. Mothers with reported autoimmune disease or cancer were excluded, as were babies born with serious malformations. Only the first pregnancy enrolled in MoBa was included in the analyses, to avoid repeated assessments of the same mother. After exclusion of women who did not initiate breastfeeding, 57,134 mother–baby pairs remained (see Fig. 1).

Maternal probiotic milk intake during the first half of pregnancy was self-reported in the MoBa FFQ, a semi-quantitative questionnaire designed to record dietary habits. FoodCalc [20] and the Norwegian Food Composition Table [21] were used to calculate food and nutrient intakes. A validation study of the FFQ showed that, relative to a dietary reference method and several biological markers, the MoBa FFQ produces a realistic estimate of habitual intake and is a valid tool for ranking pregnant women according to high and low intakes of energy, nutrients, and food [22].

The FFQ asked specifically about intake frequency of two probiotics-containing dairy products produced by Tine SA, Oslo, Norway (Product A: Biola^®, containing La-5, Bb12 and LGG; and product B: Cultura^®, containing La-5 and Bb12). Responses ranged from “never” to “8 or more times per day”, with a total of 11 response alternatives. One glass was specified as 250 ml. The bacteria count in these beverages is indicated as a minimum of 2 × 10⁸ CFU of LGG and Bb12 and 2 × 10⁷ CFU of La-5 per 1 mL product A and a minimum of 2 × 10⁸ CFU of Bb12 and 2 × 10⁷ CFU of La-5 per 1 mL product B. These products were the only widely consumed probiotic products available on the Norwegian market at the time of the study. The MoBa FFQ included questions about the use of dietary supplements and an open text field for reporting supplements other than those listed. Very few women (fewer than 0.5%) reported consumption of probiotic supplements, and intake of probiotics from supplements was thus not considered in this study.

Probiotic milk intake from both sources combined was studied as a dichotomous variable (yes/no) as well as in tertiles of probiotics intake (low, medium, and high).

In Q4, administered 6 month postpartum, women were asked whether they had consulted a doctor, midwife, or health visitor during the first month after delivery for breast-related problems, specified as “mastitis”, “sore nipples”, or “breastfeeding problems”, and whether they had received “medication for mastitis”. In addition to these four variables, a combined variable, “any breastfeeding problems”, was created.

Furthermore, mothers reported the specifics of their babies’ nutrition, both during the first week of life, choosing between breast milk, sugar water, water, and different types of formula, and then monthly, choosing between breast milk and different types of formula. Two breastfeeding variables were created based on the WHO definitions [23]: “cessation of predominant breastfeeding before 4 months” and “cessation of any breastfeeding before 4 months”. “Exclusive breastfeeding” could only be reported for the first week of life, as the questionnaires did not specifically ask about ingestion of water, water-based drinks, and fruit juices later on. “Predominant breastfeeding” refers to infants either given only breast milk (exclusive breastfeeding) or breast milk and water-based drinks, but infants should not be fed with solid food, non-human milk or formula. This group thus includes those who were breastfed exclusively during their first week of life. “Any breastfeeding” refers to infants either given only breast milk (exclusive breastfeeding), predominantly breastfed (see above) or partially breastfed, i.e., given solid food, formula, or non-human milk in addition to breast milk [23].

Confounders were selected a priori. Maternal age was registered in the NMBR and used as a continuous variable. The following variables were self-reported in Q1: maternal pre-pregnancy BMI was calculated based on reported height and pre-pregnancy weight and used as a categorical variable (< 18.5, 18.5–24.9, 25–29.9, > 30 kg/m²). Maternal education was categorized as < 13, 13–16, ≥ 16 years. Family income was recorded as neither, one or both partners earning > 300,000 Norwegian Crowns (NOK)/year. Maternal smoking categories during pregnancy were never, occasionally, and daily.

Based on the FFQ, daily fibre and energy intake were considered as continuous variables, while non-probiotic yoghurt and milk consumption were calculated as described above for probiotic milk intake.

All analyses were performed using IBM^® SPSS^® Statistics version 25. Maternal characteristics, related to probiotic milk consumption, were analysed using Pearson’s chi-square test or the two-sided Fisher’s exact test. The Kruskal–Wallis test was used to study whether there was a statistical difference between the amounts of probiotic milk intake in the different maternal characteristic categories.

The associations between being a probiotic milk consumer and breastfeeding complications and duration were studied by logistic regression analysis, unadjusted and adjusted for the confounders described above. Missing data were given a category of their own.

Sensitivity analyses were performed for the association between education and reported breastfeeding complications and duration, as well as between reported breastfeeding complications and breastfeeding duration.

In our study population, 20,901 (36.6%) women were probiotic milk consumers. The median daily intake (interquartile range) among consumers was 54 ml/day (IQR 22–179).

Probiotic milk consumers had lower parity, although they were older. They had higher educational levels and income as well as more health-conscious behaviour, with higher fibre intake and less smoking than non-consumers (see Table 1).

During the first month after delivery, a total of 4675 (8%) women in the study population contacted healthcare services for mastitis and 3127 of these (6% of the total study population) received medication for the mastitis. Sore nipples were reported by 3595 women (6%) and 3665 (6%) reported other breastfeeding problems. In total, 8788 women (15%) contacted healthcare services for any breastfeeding problems during the first month after delivery. Of the study population, 22,235 women (39%) ceased to predominantly breastfeed and 6567 women (11%) stopped any breastfeeding before the baby reached the age of 4 months.

Probiotic milk intake was significantly associated with a higher incidence of breastfeeding complications (except for medication-treated mastitis), as well as a lower frequency of breastfeeding cessation before 4 months. Odds ratios (OR) became less pronounced after adjustment, but results remained significant (except for medication-treated mastitis, see Table 2).

However, there was no dose–response association between amount of probiotic milk intake either with breastfeeding complications or duration (see Supplemental Table 1).

Sensitivity analyses were performed to better understand the contradicting results of probiotic milk consumers having a higher incidence of breastfeeding complications, while also breastfeeding longer. First, the known association between breastfeeding complications and earlier cessation of breastfeeding was confirmed in this study population (see Table 3). Associations remained the same when studying the subgroups of probiotic milk consumers and non-consumers separately (data not shown).

Second, higher maternal education was associated with higher incidence of reported breastfeeding complications during the first month after delivery, as well as with longer duration of breastfeeding (see Supplemental Table 2).

To the best of our knowledge, this is the first study to examine the possible effect of probiotic milk intake during pregnancy on breastfeeding complications and duration in an epidemiologic population-based setting. Strengths of this study are its size, with 57,134 women included, the comprehensive information on lifestyle and socioeconomics and the prospective design with registration of probiotic milk intake before possible breastfeeding complications might occur. The MoBa FFQ has been extensively validated [22, 37, 38]. However, several limitations need to be considered when interpreting the results. The outcome was based on the women’s self-reported data referring to if and why they contacted health-care services. As discussed above, this might have introduced bias, as better educated women are known to both breastfeed longer and having a lower threshold for contacting health-care services [28]. Even if a clinical follow-up of all women would have been desirable, it is unfeasible in a population-based study like MoBa. Since exposure is self-reported in a semi-quantitative FFQ, there is no information on the exact intake of probiotic bacterial count or measurement of actual bacterial count and type in the breast milk. However, the reported median daily intake in this study is comparable to the reported intake in most of the published RCTs [6-8]. The FFQ is answered at gestational week 22 and there is no comparable information on probiotics intake after delivery. However, it can be assumed that the reported habitual intake of probiotic milk products in pregnancy is a proxy of the continued habitual intake during breastfeeding. Although pregnancy is a time when most women think a lot about healthy eating, the major changes occur for intake of alcohol and coffee, while their core diet largely remains unchanged [34]. Maternal probiotic milk intake was also asked for in a less comprehensive way in MoBa Q1 and Q3 regarding the time period from before pregnancy to answering Q3 in pregnancy week 32, showing that most women continued to consume probiotic milk products as before pregnancy [16].

Administration of antibiotics, common as prophylaxis in obstetrics, or as treatment for manifest infection, might have interfered with the probiotic effect; this type of datum is not available in the MoBa data set. However, stratifying women by vaginal delivery or delivery by caesarean section, when prophylactic antibiotics treatment is usually given, did not change the results (data not shown). Unfortunately, MoBa provides no data on the exact time when the mother attracts breastfeeding complications or stops breastfeeding. Therefore, the possibility of reverse causality explaining the results—women breastfeeding longer having more time at risk for breastfeeding complications—cannot be completely excluded. However, as the analysis was restricted to women who initiated breastfeeding as well as to breastfeeding complications reported during the first month after delivery, we judge the risk for reverse causality as very low. Women were asked what, but not how, they fed their children; whether they were breastfeeding and/or bottle-feeding pumped breast milk might have affected the risk of developing breastfeeding complications. Despite adjustment for relevant confounders, residual confounding is probably still part of this association, as discussed above.