Statistical analysis
Descriptive data was expressed as median and interquartile range (IQR) and in proportions (absolute and relative frequencies). Spearman correlation coefficient (Rho) was calculated as a measure of the degree of the relationship between each immunoglobulin type measured in colostrum and maternal serum, and between each immunoglobulin type measured in mature milk and maternal serum. Linear regression analysis was performed to assess the relationship between maternal symptoms and immunoglobulin levels. Coefficients, 95% confidence interval (CI) and p-value were reported. To graphically represent the relationship between antibody levels and time lapse to disease, we adjusted a locally weighted scatterplot smoothing (LOWESS) regression which, essentially, joins together the curves generated by several regressions derived from different segments of the timeframe23.
Lastly, to see how each immunoglobulin type evolved from the colostrum to the mature milk, paired t-test was used. This analysis was performed separately for women with active disease and those who had recovered from SARS-CoV-2 infection. In all cases mean of the differences, 95% CI and p-value were reported. Level of significance was set up at 0·05.
The statistical software R version 4·0·3 (Vienna, Austria) was used for all data analyses24.