Quality Assessment of Raw Human Milk Under Prolonged Household Freezing: Titratable Acidity, Energy Content, and Microbiological Safety
Description
This dataset was generated from a prospective observational cohort study designed to evaluate the quality of raw human milk (RHM) stored under household freezing conditions for up to 150 days. The study was conducted in July 2025 at the Nutrition and Dietetics Laboratory of the Faculdade Israelita de Ciências da Saúde Albert Einstein (FICSAE), São Paulo, Brazil. The research was motivated by the discrepancy between Brazilian regulations, which recommend a maximum freezing period of 15 days for expressed human milk, and international guidelines that allow storage for three to twelve months for healthy full-term infants. The dataset was created to provide empirical evidence on the behavior of selected milk quality indicators during prolonged domestic freezing. Milk samples were obtained from 28 lactating adults and analyzed at seven storage time points: 1, 15, 30, 60, 90, 120, and 150 days after collection. Three quality indicators were assessed at each time point: titratable acidity (°D), measured by the Dornic titration method; energy content (kcal/L), estimated by the crematocrit technique; and total coliform occurrence, determined by the presumptive phase of the Brilliant Green Bile Lactose (BGBL) broth method. The dataset includes raw analytical measurements, derived variables, participant-level summaries, sociodemographic and lactation-related characteristics, and a complete data dictionary. Data are available in both wide and long formats to facilitate descriptive, longitudinal, and inferential analyses. Statistical output files generated in Jamovi are also provided to support reproducibility. The data show longitudinal variation in titratable acidity during storage, while energy content remained measurable throughout the 150-day period. Total coliform occurrence was infrequent and recorded as a binary outcome (Absent/Present). Summary statistics for all three indicators are included in the dataset documentation. Missing observations (21.9%) reflect variability in expressed milk volume among participants and are documented at the individual level. This dataset can be used to investigate changes in human milk quality during domestic frozen storage, validate statistical models of acidity and energy trajectories, perform secondary analyses involving donor characteristics and milk quality parameters, and support evidence-based discussions regarding human milk storage recommendations. To the authors’ knowledge, it represents the first Brazilian longitudinal dataset evaluating raw human milk quality at seven storage time points extending to 150 days under household freezing conditions.
Files
Steps to reproduce
A prospective observational cohort study was conducted at the Nutrition and Dietetics Laboratory of the Faculdade Israelita de Ciências da Saúde Albert Einstein (FICSAE), São Paulo, Brazil, in July 2025. Lactating adults were recruited through social networks and community dissemination channels. Eligible participants were aged 18 years or older, had been lactating for at least 15 days, and possessed a personal breast pump. Written informed consent was obtained from all participants. Before milk expression, breast pump components were cleaned with water and soap, sanitized by immersion in boiling water for 15 minutes, and air-dried. Milk was expressed using personal electric or manual breast pumps and collected directly into sanitized glass containers. Immediately after expression, samples were refrigerated at ≤4 °C. Within 24 hours, milk was aliquoted under laminar airflow into sterile 10 mL screw-cap tubes using automatic pipettes with sterile tips. Aliquots were assigned to seven predefined storage time points: day 1, 15, 30, 60, 90, 120, and 150. The day-1 aliquot remained refrigerated and was analyzed within 24 hours. All remaining aliquots were stored at −18 °C in the freezer compartment of a domestic two-door refrigerator until analysis. Before laboratory testing, frozen aliquots were thawed in a water bath not exceeding 37 °C and gently homogenized throughout the thawing process. Three analyses were performed at each time point from the same thawed aliquot: Titratable acidity was measured in triplicate using the Dornic titration method with 3 mL of milk per replicate. Results were expressed as degrees Dornic (°D); Energy content was determined in triplicate using the crematocrit method with 1 mL of milk per replicate. Cream column and total column measurements were recorded, and cream content, fat content, and energy content (kcal/L) were calculated according to the rBLH/FIOCRUZ protocol; Total coliform occurrence was assessed using the presumptive phase of the Brilliant Green Bile Lactose (BGBL) broth method. Four milliliters of milk were inoculated into BGBL broth containing Durham tubes. Results were recorded as Absent or Present according to gas formation after incubation. Sociodemographic, lactation, and milk-storage information were collected through a structured questionnaire administered at the collection session. Raw analytical data, derived variables, participant-level summaries, and a data dictionary were compiled into a spreadsheet database. Statistical analyses were performed using Jamovi v2.3 with the GAMLj module. Continuous outcomes were analyzed using generalized mixed models with Gamma distribution and identity link function and Spearman correlation analysis. Associations involving coliform occurrence were evaluated using Fisher’s exact test, Mantel-Haenszel linear trend test and Mann-Whitney U test. All statistical outputs are provided as supplementary PDF files.