This study was conducted as part of a study exploring the metabolic health of postpartum women with a history of GDM. This prospective cohort study was conducted at the Faculty of Medicine Vajira Hospital, Bangkok, Thailand. The study protocol was approved by the Vajira Institutional Review Board (certificate of approval no. 016/2563) and strictly adhered to the Strengthening the Reporting of Observational Studies in Epidemiology guidelines.

The study participants were women with a GDM history who experienced live births between 7 September 2020 and 31 January 2023. To be included in this study, participants were required to be ≥ 18 years of age, have a pregnancy complicated by GDM diagnosed using the Carpenter and Coustan criteria [19], and have undergone blood glucose testing for GDM at our hospital. The exclusion criteria were HIV infection, pregnancy during the 6-month study period, diagnosis of T2DM or MetS at baseline (six weeks postpartum), taking medications or substances that are contraindicated during breastfeeding, such as chemotherapy agents, radioactive substances, illicit drugs, ergotamine, or lithium, and loss during follow-up.

Given that no study has directly investigated the effects of breastfeeding for six months on prediabetes, T2DM, and MetS six months postpartum in women with a history of GDM, we calculated the sample size based on data from a previous study that examined the effect of breastfeeding on glucose intolerance 12–14 months postpartum [20]. To show a 34% decrease in prediabetes/T2DM risk in women practicing breastfeeding for six months (from 78 to 44% [20]) with 80% power at a two-sided significance level of 0.05, at least 64 participants were required (32 who maintained breastfeeding and 32 who did not). Considering a dropout rate of 20%, the total sample size required was 80 (40 in each group).

In the postnatal ward, postpartum women who have been complicated by GDM in their most recent pregnancy (recent GDM) were recruited sequentially. They were informed about the research project and participated in the study. A written informed consent was obtained from all participants before enrollment. Baseline characteristics were extracted from the medical records. The pre-pregnancy body mass index (BMI) was calculated based on the self-reported pregravid weight and the measured height.

On the discharge date, the participants were provided with a mini calendar to record the date of introduction of infant formula or foods/drinks other than breast milk, as well as the date when the mother stopped breastfeeding or expressing milk. They were instructed to bring the calendar with them when they attended follow-up visits.

The participants were scheduled for follow-up visits at 6 weeks and 6 months postpartum. During both visits, the participants were interviewed about the frequency and duration of breastfeeding. In addition, they were asked whether they had provided infant formula or complementary foods to the babies. If so, the time of introduction of the formula or foods/drinks and the frequency of administration and quantity were recorded. Feeding practices were also assessed based on the data noted in the calendar provided to participants and the prospective assessment of breastfeeding status recorded in the hospital’s electronic database during the follow-up visits of the babies at the routine immunization clinics at the completion of the 2nd, 4th, and 6th months after birth. Maternal weight, waist circumference (WC), and blood pressure (BP) were measured by a specially trained nurse using standardized protocols. Venous blood samples were collected at both visits. At six weeks postpartum, blood samples were collected to measure fasting plasma glucose (FPG), fasting lipids such as total cholesterol, triglycerides (TG), low-density lipoprotein cholesterol, high-density lipoprotein cholesterol (HDL-C), and plasma glucose levels 2 h after consuming a 75-g oral glucose load. At six months postpartum, fasting venous blood samples were collected to measure FPG, hemoglobin A1c (HbA1c), and lipid parameters.

Blood samples were collected in the morning after fasting for 12 h overnight. Blood tests for plasma glucose, HbA1c, and lipid levels were performed using standard assays with a well-calibrated analyzer. Our laboratory received approval from the Randox International Quality Assessment Scheme for blood chemistry and HbA1c analyses. The HbA1c test was performed using a standardized and certified assay by the National Glycohemoglobin Standardization Program. This assay can be traced to the reference method used in the Diabetes Control and Complications Trial, thereby ensuring accurate and reliable results.

The outcome measures were prediabetes, T2DM, and MetS. The exposures of interest were lactation duration and intensity. Those with FPG levels of 100–125 mg/dL, 2-h plasma glucose levels of 140–199 mg/dL, or HbA1c levels of 5.7–6.4% were diagnosed with prediabetes [21]. T2DM was defined as FPG ≥ 126 mg/dL, 2-h plasma glucose during a 75-g oral glucose tolerance test ≥ 200 mg/dL, or HbA1c ≥ 6.5% [21]. MetS was diagnosed based on a Joint Interim Statement [22]. A diagnosis was made when three or more of the following were present: (1) WC ≥ 80 cm, (2) systolic BP ≥ 130 or diastolic BP ≥ 85 mmHg or treatment with antihypertensive medication, (3) FPG ≥ 100 mg/dL or treatment with diabetes medication, (4) fasting TG ≥ 150 mg/dL or medication treatment, and (5) HDL-C < 50 mg/dL or medication treatment.

Because this study limited the time for glucometabolic evaluation to six months postpartum, we categorized the lactation duration as six months (maintaining breastfeeding) or less than six months (not maintaining breastfeeding). Given that the WHO and UNICEF recommend that mothers exclusively breastfeed their infants for six months, the intensity of maintaining breastfeeding was divided into EBF and partial breastfeeding. We defined EBF at six months (maintaining EBF) as feeding infants only breast milk for six months, except for drops or syrups containing medicines, vitamins, or mineral supplements. When infants received any other food or liquid during the 6-month breastfeeding period, they were categorized as partial breastfeeding at six months.

Data were analyzed using IBM SPSS Statistics for Windows, Version 28.0 (IBM Corp., Armonk, NY, USA). Categorical variables are presented as numbers and percentages and were compared using the chi-squared or Fisher’s exact tests. Continuous variables were described as means and standard deviations for normally distributed data or as medians and interquartile ranges for non-normally distributed data and were compared using the Student’s t-test or Mann–Whitney U test. To estimate the associations of lactation duration with prediabetes, T2DM, and MetS, generalized linear models were used to control for potential confounders, which were determined a priori based on established associations in a previous study [20] and were limited to age, family history of diabetes, pre-pregnancy BMI, weight change from delivery to six months postpartum, and prediabetes at baseline. The results are reported as risk ratios (RRs) with 95% confidence intervals (CIs). Statistical significance was set at p < 0.05.