Ethical approval (approval number: [2024] 142) was obtained for this analysis and each participant signed informed consent. This study examined the institutional database encompassing all women who gave birth at the Third Affiliated Hospital of Guangzhou Medical University (Guangzhou Medical Centre for Critical Pregnant Women) from January 2016 to December 2020. In total, 38,439 women were included in the analysis. Women lacking SPPH information (n = 52) were excluded from the analysis. Additionally, pregnancies with missing values comprising ≥ 10% of covariates, such as pregestational body mass index (BMI) (n = 4,084), were removed. Subsequently, missing values for covariates comprising < 10% were imputed using the random forest method. Thus, our final analysis included a total of 34,303 pregnant women ( Figure 1 ).

The outcome of the research focused on SPPH, defined as an estimated hemorrhage of ≥ 1,000 mL within 24 hours after childbirth. Blood loss assessment was based on a combination of visual estimation, gravimetric method, estimated blood loss volume, shock index, and hemoglobin levels, as documented by the attending healthcare providers, including physicians, midwives, and nurses. The final estimated value was recorded by the resident physician who attended the delivery.

The diagnosis of hypothyroidism during pregnancy was based on the Guidelines on Diagnosis and Management of Thyroid Diseases during Pregnancy and Postpartum (14, 15). Overt hypothyroidism was defined as serum thyroid-stimulating hormone (TSH) levels above the trimester-specific reference ranges with free thyroxine (FT4) levels below the corresponding lower limit, while subclinical hypothyroidism was defined as elevated TSH levels with normal FT4. Measurements of TSH and FT4 were performed employing a chemiluminescent immunoassay with reagents from Abbott Laboratories. Trimester-specific reference ranges are shown in Supplementary Tables 1 , 2 . Laboratory test reports from hospitals of equivalent level were also acceptable. Cases of hypothyroidism were ascertained using the corresponding International Classification of Diseases (ICD) codes from the electronic health records. Regardless of thyroid peroxidase antibody (TPOAb) status (positive or negative), all pregnant women diagnosed with hypothyroidism routinely received levothyroxine (LT4) therapy, with dose adjustment to maintain TSH within the guideline-recommended levels.

To reduce potential biases in our analysis, we chose covariates based on previous research findings and clinical plausibility. The analysis incorporated the following covariates: age, pregestational body mass index (BMI), education level, history of radiation exposure, previous cesarean delivery, number of fetuses, gestational age, birth weight, head circumference of newborns, stillbirth, hypertensive disorders of pregnancy, HELLP syndrome, diabetes, intrahepatic cholestasis of pregnancy (ICP), anemia, thrombocytopenia, placenta previa, placental abruption, placenta accreta spectrum (PAS), polyhydramnios, fetal anomalies, malposition, cesarean delivery, prolonged second stage of labor, and fetal distress. Full list of covariates investigated in analysis are provided in Supplementary Table 3 .

Continuous data with a normal distribution were presented as mean (± standard deviation [SD]) and were compared using the independent samples t-test. Variables with skewed distributions were expressed as medians [first quartile (P25) and third quartile (P75)] and were compared using the nonparametric Wilcoxon rank-sum test. Categorical variables were expressed as percentages (%) and were compared using either the chi-square test or Fisher’s exact test, as appropriate. We utilized three logistic regression models to explore the impact of overt or subclinical hypothyroidism on SPPH. Model 1 did not adjust for any covariates. Model 2 made adjustments for age, pregestational BMI, education level, history of radiation exposure, previous cesarean delivery, number of fetuses, gestational age, and birth weight. Model 3 further adjusted for head circumference of newborns, stillbirth, hypertensive disorders of pregnancy, HELLP syndrome, diabetes, ICP, anemia, thrombocytopenia, placenta previa, placental abruption, PAS, polyhydramnios, fetal anomalies, malposition, cesarean delivery, prolonged second stage of labor, and fetal distress based on Model 2. Subsequently, we conducted an interaction analysis, finding that age and history of radiation exposure acted as potential modifiers in the relationship between hypothyroidism and SPPH. Considering the relatively small number of patients with SPPH in the history of radiation exposure, stratified analysis was done only for age. To examine the influence of age on the association between hypothyroidism and SPPH, we stratified the study cohort based on age, employing a cut-off threshold of 35 years. Furthermore, sensitivity analyses were conducted to evaluate the robustness of our findings; In one analysis, participants who delivered twins or triplets were excluded, while in the other sensitivity analysis, participants with a history of radiation exposure and an unknown radiation exposure history were removed. All p values reported were two-sided, with a significance level established at 0.05. Statistical analyses were performed using IBM SPSS Statistics 26.

The baseline characteristics of the study population are shown in Table 1 . In our analysis, we included 34,303 pregnant women, among whom 913 were diagnosed with SPPH, and the average age of the participants with SPPH was 32.77 ± 4.84 years. Overweight women (pregestational BMI exceeding 23.9) were more frequently observed in the SPPH group. Low educational level, history of radiation exposure, and previous cesarean delivery were more common among pregnant women with SPPH. Furthermore, the SPPH group exhibited higher rates of hypertensive disorders of pregnancy, diabetes, anemia, thrombocytopenia, placenta previa, placental abruption, PAS, cesarean delivery, prolonged second stage of labor, and hypothyroidism. Baseline characteristics stratified by thyroid status are provided in Supplementary Table 4 .

Table 2 illustrates the association between hypothyroidism and SPPH using three logistic regression models. In model 1, a positive association between hypothyroidism and SPPH was observed (odds ratio [OR] = 1.539, 95% confidence interval [CI]: 1.109–2.134, p = 0.009). Upon dividing hypothyroidism into overt hypothyroidism and subclinical hypothyroidism, it was found that compared to controls, the OR for SPPH with overt hypothyroidism and subclinical hypothyroidism were 1.594 (95% CI: 1.123–2.261, p = 0.009) and 1.246 (95% CI: 0.510–3.045, p = 0.630), respectively. After adjusting for age, pregestational BMI, education level, history of radiation exposure, previous cesarean delivery, number of fetuses, gestational age and birth weight in model 2, it was found that the association between hypothyroidism and increased SPPH risk persisted (OR = 1.477, 95% CI: 1.056–2.066, p = 0.023). Consistent with the findings of model 1, overt hypothyroidism was associated with an increased risk of SPPH (OR = 1.587, 95% CI: 1.109–2.271, p = 0.012) while subclinical hypothyroidism didn’t have a prominent impact on SPPH (OR = 0.993, 95% CI: 0.397–2.483, p = 0.988), compared to controls. Moreover, model 3 further adjusted for head circumference of newborns, stillbirth, hypertensive disorders of pregnancy, HELLP syndrome, diabetes, ICP, anemia, thrombocytopenia, placenta previa, placental abruption, PAS, polyhydramnios, fetal anomalies, malposition, cesarean delivery, prolonged second stage of labor, and fetal distress based on model 2. Subsequently, we observed a consistent finding that hypothyroidism (OR = 1.609, 95% CI: 1.111–2.329, p = 0.012), particularly overt hypothyroidism (OR = 1.688, 95% CI: 1.137–2.507, p = 0.009), was correlated with an elevated risk of SPPH. Nevertheless, subclinical hypothyroidism showed no significant association with SPPH.

We conducted an interaction analysis and found that age and history of radiation exposure acted as potential modifiers in the relationship between hypothyroidism and SPPH (p for interaction < 0.05). Due to the relatively small number of patients with SPPH in the history of radiation exposure, stratified analysis was performed only for age. Table 3 indicates the correlation between hypothyroidism and SPPH stratified by age, using 35 years as the cut-off (p for interaction < 0.05). The ORs were stronger in women aged < 35 years (OR = 2.412, 95% CI: 1.583–3.673, p < 0.001) compared to those aged ≥ 35 years (OR = 0.755, 95% CI: 0.364–1.567, p = 0.450). For participants aged < 35 years, the ORs for SPPH with overt hypothyroidism and subclinical hypothyroidism were 2.492 (95% CI: 1.585–3.916, p < 0.001) and 2.006 (95% CI: 0.671–5.996, p = 0.213), respectively. For women aged ≥ 35 years, the ORs for SPPH with overt and subclinical hypothyroidism were 0.807 (95% CI: 0.372–1.749, p = 0.586) and 0.487 (95% CI: 0.058–4.119, p = 0.509), respectively.

We conducted several sensitivity analyses to verify the stability of our results ( Table 4 ). In one sensitivity analysis, we excluded participants who delivered twins or triplets. Additionally, another sensitivity analysis eliminated participants with a history of radiation exposure and those with an unknown radiation exposure history. The sensitivity analyses indicated that overt hypothyroidism remained a substantial influence on elevating the risk of SPPH, whereas subclinical hypothyroidism was not found to exert a notable impact on SPPH.