Results
Overall, 177 women were recruited, and breast milk samples were collected for analysis from the six recruiting units. Maternal, pregnancy and disease characteristics are shown in Table 1. 29 (16·38%) women had active SARS-CoV-2 infection at the time of delivery. 28 (15·82%), 55 (31·07%) and 65 (36·74%) women acquired the infection in the first trimester (<14 weeks), second trimester (14-28+6) and third trimester (>28+6) of pregnancy, respectively.
148 samples of colostrum were collected between the day of delivery until day 4 postpartum. In 13 cases there was insufficient sample for analysis and therefore colostrum from 135 women was included in the final analysis. Of these, serological status of the women at the time of colostrum collection was available in 124 cases (Table 2). In 83 women milk samples were collected after day 7 postpartum (mature milk). In two samples there was insufficient mature milk for analysis and therefore 81 samples were finally included. Of these, serological status of the women at the time of mature milk collection was available in 66 (Table 2; Figure 1).
IgA and IgG were present in 111/135 (82·23%) and 2/135 (1·48%) of the colostrum samples, respectively. Due insufficient sample, IgM could only be analyzed in 67 colostrum samples, and it was present in 14 (26·42%) cases. While IgG was negative in all mature milk samples, IgA was present in 27/81 (33·33%). IgM could only be analyzed in 59 mature milk samples and all tested negative (Table 2).
70 colostrum samples were tested against rRT-PCR SARS-CoV2. In 8 samples there was insufficient sample for analysis and therefore 61 were finally tested. 2 were inhibited and 59 tested negative.
Breast milk antibody correlation with maternal serological status
Among the 111 cases with positive IgA in the colostrum, maternal blood at the time of colostrum collection tested negative for IgM, IgG and IgA in five cases (Table 4). There was a significant correlation between IgM (rho=0·49; p=0·001), IgA (rho=0·3; p=0·008) and IgG (rho=0·27; p=0·018) measured in the colostrum and maternal serum (Figure 2). There was also a significant correlation between IgA (rho=0·49; p=0·002) measured in the mature milk and maternal serum but there was no association for IgM rho=0·31; p=0·056) and IgG (rho=-0·17; p=0·216) (Figure 3).
Breast milk antibody correlation with severity of SARS-CoV-2 infection
Women with pneumonia had significantly higher IgA (2·370, IQR 0·29-4·44; p=0·03) and IgG (0·19, IQR 0·07-0·31; p=0·001) in the colostrum than those who were asymptomatic. IgG was also higher in women with pneumonia than in women with mild symptoms (-0·19, IQR 0·08-0·31; p=0·001) (Figures 4a and 4b). No other significant correlations were identified.
Breast milk antibody correlation with time interval from active infection
No correlation was found between IgM, IgA or IgG in the colostrum or mature milk and the time interval from acute infection of SARS-CoV-2. However, higher rates of IgA and IgM positivity were noted in the colostrum when SARS-CoV-2 infection occurred during the second and third trimesters of pregnancy although these differences were not statistically different (Table 3).
Antibody evolution from colostrum to mature milk
Paired colostrum and mature milk samples were analyzed for IgA and IgG in 47 women. IgA concentration in the colostrum was higher than in the mature milk (2·42, 95% CI 1·46 to 3·37; p<0·001) while IgG concentration was not significantly different between the two types of milk (0·016, 95% CI -0·0178 to 0·050; p=0·341). Paired colostrum and mature milk samples were analyzed for IgM in 20 women. No significant differences were found between them (1·043, 95% CI -0·016 to 2·102, p=0·053).
12 paired samples corresponded to women with active COVID-19 at the time of delivery. Only IgA concentration was significantly higher in the colostrum than in the mature milk (3·07, 95% CI 1·23 to 4·92; p<0·005). 35 paired samples corresponded to women who had recovered from the disease at the time of delivery. Only IgA concentration was significantly higher in the colostrum than in mature milk (2·19, 95% CI 1·04 to 3·35; p<0·001).
Active SARS-CoV-2 infection at the time of delivery
29 pregnant women had SARS-CoV-2 infection at the time of delivery. Two colostrum samples had insufficient volume for the r-RT-PCR SARS-CoV-2 analysis, and one was inhibited. 26 samples were tested, and all were negative.
Five colostrum samples had insufficient volume for serological analysis and, therefore, 24 colostrum samples were analyzed. IgA and IgG were positive in 18 (75%) and 1 (4·16%) cases, respectively. Due insufficient sample, IgM could only be analyzed in 10 samples, and it was positive in 3 (30%) cases (Table 3). No correlation was found in IgM (rho=0·71; p=0·088), IgA (rho=0·46; p=0·153) and IgG (rho=0·33; p=0·322) between the colostrum and the maternal blood.
Discussion: