Clinical features and outcomes of peripartum obstetric patients admitted to the intensive care unit: A nationwide inpatient database in Japan.

Between July 2010 and March 2022, we identified 8,184 obstetric patients from 195 institutions who were admitted to ICUs during the antepartum period or within one week postpartum. Patient characteristics, stratified by antepartum and postpartum status, are summarized in Table  1 . The median age was 34 years (IQR, 30–38 years), and 46.8% of the patients were aged 35 years or older. Cesarean delivery was performed in 53.2% of the patients. As shown in Fig.  1 , hemorrhage was the leading cause of ICU admissions (52.6%), followed by hypertensive disorders of pregnancy (16.7%) and cardiovascular disease (2.1%). Table  2 presents the procedures performed on obstetric patients admitted to the ICU among antepartum and postpartum status. Blood transfusions were administered to 71.5% of patients, mechanical ventilation to 28.0%, and transcatheter arterial embolization to 18.0%. Table 1 Characteristics and clinical features of obstetric patients admitted to the ICU between antepartum and postpartum patients Variables Overall (n = 8184) Antepartum (n = 5645) Postpartum (n = 2539) P-value Age, years, median (IQR) 34 (30–38) 34 (30–38) 34 (30–37)  < 0.001 Age category, years, n (%)  < 0.001   < 20 67 (0.8) 47 (0.8) 20 (0.8)  20–24 357 (4.4) 231 (4.1) 126 (5.0)  25–29 1360 (16.6) 929 (16.5) 431 (17.0)  30–34 2570 (31.4) 1721 (30.5) 849 (33.4)  35–39 2711 (33.1) 1879 (33.3) 832 (32.8)   ≥ 40 1119 (13.7) 838 (14.8) 281 (11.1) BMI, kg/m 2 , n (%)  < 0.001   < 18.5 277 (3.4) 180 (3.2) 97 (3.8)  18.5–24.9 4534 (55.4) 3074 (54.5) 1460 (57.5)  25.0–29.9 2145 (26.2) 1,546 (27.4) 599 (23.6)   ≥ 30 725 (8.9) 581 (10.3) 144 (5.7)  Missing 503 (6.1) 264 (4.7) 239 (9.4) Gestational age at hospital admission, weeks, n (%)   < 28 369 (4.5) 369 (6.5) –  28 to < 32 520 (6.4) 520 (9.2) –  32 to < 37 1903 (23.3) 1903 (33.7) –  37 to < 42 2391 (29.2) 2391 (42.4) –   ≥ 42 18 (0.2) 18 (0.3) –  Missing 2983 (36.4) 444 (7.9) – Instrumental delivery, n (%) 311 (3.8) 311 (5.5) – Cesarean delivery, n (%) 4356 (53.2) 4356 (77.2) – Patient conditions and comorbidities at admission, n (%)  Hypertension 471 (5.8) 364 (6.4) 107 (4.2)  < 0.001  Gestational diabetes mellitus 325 (4.0) 289 (5.1) 36 (1.4)  < 0.001  Uterine myoma 317 (3.9) 279 (4.9) 38 (1.5)  < 0.001  Premature rupture of the membranes 223 (2.7) 201 (3.6) 22 (0.9)  < 0.001  Overt diabetes mellitus 96 (1.2) 62 (1.1) 34 (1.3) 0.35  Ovarian tumor 69 (0.8) 53 (0.9) 16 (0.6) 0.16  Multiple pregnancy 24 (0.3) 21 (0.4) 3 (0.1) 0.049  Ovarian endometrioma 20 (0.2) 19 (0.3) 1 (0) 0.012  Endometriosis 13 (0.2) 11 (0.2) 2 (0.1) 0.22  Adenomyosis 18 (0.2) 14 (0.2) 4 (0.2) 0.42  Cardiac arrest on arrival 9 (0.1) 0 (0) 9 (0.4)  < 0.001 Hospital admission characteristics, n (%)  Admission to teaching hospital 7582 (92.6) 5162 (91.4) 2420 (95.3)  < 0.001  Emergency admission 6335 (77.4) 3978 (70.5) 2357 (92.8)  < 0.001  Ambulance use 3511 (42.9) 1684 (29.8) 1827 (72.0)  < 0.001 ICU, intensive care unit; IQR, interquartile range; BMI, body mass index Fig. 1 Causes of ICU admission in obstetric patients Table 2 Procedures performed on obstetric patients admitted to the ICU between antepartum and postpartum patients Procedures Overall (n = 8184) Antepartum (n = 5645) Postpartum (n = 2539) P-value Blood transfusion 5850 (71.5) 3682 (65.2) 2168 (85.4)  < 0.001 Mechanical ventilation 2293 (28.0) 1530 (27.1) 763 (30.1) 0.006 Transcatheter arterial embolization 1470 (18.0) 644 (11.4) 826 (32.5)  < 0.001 Hysterectomy without uterine rupture 612 (7.5) 331 (5.9) 281 (11.1)  < 0.001 Hysterectomy with uterine rupture 551 (6.7) 317 (5.6) 234 (9.2)  < 0.001 Renal replacement therapy 166 (2.0) 109 (1.9) 57 (2.2) 0.35 Chest compressions 101 (1.1) 48 (0.9) 53 (2.1)  < 0.001 Extracorporeal membrane oxygenation 93 (1.1) 62 (1.1) 31 (1.2) 0.63 Resuscitative endovascular balloon occlusion of the aorta 68 (0.8) 27 (0.5) 41 (1.6)  < 0.001 Craniotomy hematoma removal 37 (0.5) 30 (0.5) 7 (0.3) 0.11 Defibrillation 31 (0.4) 21 (0.4) 10 (0.4) 0.88 Burr hole drainage 22 (0.3) 16 (0.3) 6 (0.2) 0.70 Data are shown as n (%) ICU, intensive care unit Characteristics and clinical features of obstetric patients admitted to the ICU between antepartum and postpartum patients ICU, intensive care unit; IQR, interquartile range; BMI, body mass index Causes of ICU admission in obstetric patients Procedures performed on obstetric patients admitted to the ICU between antepartum and postpartum patients Data are shown as n (%) ICU, intensive care unit Outcomes for obstetric patients admitted to the ICU, stratified by antepartum and postpartum status, are shown in Table  3 . The overall in-hospital mortality was 1.1% (94/8,184) and ICU mortality was 0.7% (61/8,184), while the fetal death was 3.6%. The median number of days from hospital admission to delivery was 2 days (IQR, 1–7 days). Among antepartum patients, 79.2% were admitted to the ICU on the day of delivery. Table 4 presents a comparison of detailed diagnoses as causes of ICU admission among survivors and non-survivors. Hemorrhage was more common in survivors, whereas cardiovascular disease, cerebrovascular disease, infection, pulmonary disease, injury, and suicide were more frequent in non-survivors. Among more specific diagnoses, notable examples more prevalent in non-survivors included amniotic fluid embolism, peripartum cardiomyopathy, and intracranial hemorrhage. Patient characteristics and procedures, compared between survivors and non-survivors, are summarized in Additional files 5 and 6 , respectively. Table 5 presents in-hospital mortality stratified by the cause of ICU admission. Among these specific conditions, in-hospital mortality was as follows: intracranial hemorrhage, 20.5% (16/78); pulmonary embolism, 13.5% (5/37); sepsis due to streptococcus, 12.5% (1/8); amniotic fluid embolism, 11.3% (9/80); and sepsis due to all causes, 10.8% (8/74). Table 3 Outcomes for obstetric patients admitted to the ICU between antepartum and postpartum patients Outcomes Overall (n = 8184) Antepartum (n = 5645) Postpartum (n = 2539) P-value In-hospital mortality, n (%) 94 (1.1) 53 (0.9) 41 (1.6) 0.008 ICU mortality, n (%) 61 (0.7) 29 (0.5) 32 (1.3)  < 0.001 Fetal death, n (%) 292 (3.6) 207 (3.7) 85 (3.3) 0.47 Length of hospital stay, days, median (IQR) 11 (8–20) 13 (9–25) 8 (5–12)  < 0.001 Length of ICU stay, days, median (IQR) 1 (1–2) 1 (1–2) 1 (1–2) 0.42 Days from hospital admission to ICU admission, days, median (IQR) 2 (1–4) 2 (1–8) 1 (1–1)  < 0.001 Days from hospital admission to delivery, days, median (IQR) 2 (1–7) – Days from delivery to ICU admission, n (%)   ≥ 3 days before 289 (5.1) –  2 days before 131 (2.3) –  1 day before 543 (9.6) –  Same day 4472 (79.2) –  1 day after 87 (1.5) –  2 days after 14 (0.2) –   ≥ 3 days after 109 (1.9) – For consistency, if two events occurred on the same day, the interval was recorded as one day ICU, intensive care unit; IQR, interquartile range Table 4 Detailed causes of ICU admission in obstetric patients between survivors and non-survivors Cause of ICU admission Overall (n = 8184) Survivor (n = 8090) Non-survivor (n = 94) P-value Hemorrhage 4306 (52.6) 4274 (52.8) 32 (34.0)  < 0.001  Uterine atony 1376 (16.8) 1370 (16.9) 6 (6.4) 0.007  Disseminated intravascular coagulation 1165 (14.2) 1150 (14.2) 15 (16.0) 0.63  Placenta previa 716 (8.7) 715 (8.8) 1 (1.1) 0.008  Placental abruption 581 (7.1) 579 (7.2) 2 (2.1) 0.059  Placenta accreta 526 (6.4) 525 (6.5) 1 (1.1) 0.033  Hemorrhagic shock 288 (3.5) 286 (3.5) 2 (2.1) 0.46  Uterine rupture 209 (2.6) 207 (2.6) 2 (2.1) 0.79  Pelvic hematoma 152 (1.9) 152 (1.9) 0 (0) 0.18  Uterine inversion 130 (1.6) 130 (1.6) 0 (0) 0.22  Amniotic fluid embolism 80 (1.0) 71 (0.9) 9 (9.6)  < 0.001  Shock during or following labor and delivery 38 (0.5) 36 (0.4) 2 (2.1) 0.017  Hemorrhage during or following labor and delivery 16 (0.2) 16 (0.2) 0 (0) 0.67 Hypertensive disorders of pregnancy 1366 (16.7) 1351 (16.7) 15 (16.0) 0.85  Preeclampsia 1133 (13.8) 1118 (13.8) 15 (16.0) 0.55  Eclampsia 286 (3.5) 286 (3.5) 0 (0) 0.064 Cardiovascular disease 170 (2.1) 165 (2.0) 5 (5.3) 0.027  PPCM 68 (0.8) 65 (0.8) 3 (3.2) 0.011  Pulmonary edema 35 (0.4) 35 (0.4) 0 (0) 0.52  Cardiomyopathy without PPCM 21 (0.3) 19 (0.2) 2 (2.1)  < 0.001  Heart failure 21 (0.3) 20 (0.2) 1 (1.1) 0.12  Aortic dissection 16 (0.2) 16 (0.2) 0 (0) 0.67  Acute myocardial infarction 6 (0.1) 6 (0.1) 0 (0) 0.79  Endocarditis 9 (0.1) 9 (0.1) 0 (0) 0.75 Cerebrovascular disease 150 (1.8) 131 (1.6) 19 (20.2)  < 0.001  Intracranial hemorrhage 78 (1.0) 62 (0.8) 16 (17.0)  < 0.001  Subarachnoid hemorrhage 44 (0.5) 42 (0.5) 2 (2.1) 0.034  Cerebrovascular dissection 23 (0.3) 23 (0.3) 0 (0) 0.60  Ischemic stroke 10 (0.1) 9 (0.1) 1 (1.1) 0.009 Infection 102 (1.2) 92 (1.1) 10 (10.6)  < 0.001  Sepsis due to all causes 74 (0.9) 66 (0.8) 8 (8.5)  < 0.001  Urinary and genital infections 27 (0.3) 26 (0.3) 1 (1.1) 0.21  Sepsis due to streptococcus 8 (0.1) 7 (0.1) 1 (1.1) 0.003 Pulmonary disease 59 (0.7) 54 (0.7) 5 (5.3)  < 0.001 Pulmonary embolism 37 (0.5) 32 (0.4) 5 (5.3)  < 0.001  Asthma 17 (0.2) 17 (0.2) 0 (0) 0.66  Aspiration 5 (0.1) 5 (0.1) 0 (0) 0.81 Injury, poisoning, and other external causes 43 (0.5) 41 (0.5) 2 (2.1) 0.031 Suicide 16 (0.4) 15 (0.4) 1 (2.7) 0.021 Data are shown as number (%) unless otherwise specified ICU, intensive care unit; PPCM, peripartum cardiomyopathy Table 5 Cause-specific in-hospital mortality in obstetric patients Cause of ICU admission Mortality Deaths/Total cases (%) Hemorrhage 32/4306 (0.7)  Uterine atony 6/1376 (0.4)  Disseminated intravascular coagulation 15/1165 (1.3)  Placenta previa 1/716 (0.1)  Placental abruption 2/581 (0.3)  Placenta accreta 1/526 (0.2)  Hemorrhagic shock 2/288 (0.7)  Uterine rupture 2/209 (1.0)  Pelvic hematoma 0/152 (0)  Uterine inversion 0/130 (0)  Amniotic fluid embolism 9/80 (11.3)  Shock during or following labor and delivery 2/38 (5.3)  Hemorrhage during or following labor and delivery 0/16 (0) Hypertensive disorders of pregnancy 15/1366 (1.1)  Preeclampsia 15/1133 (1.3)  Eclampsia 0/286 (0) Cardiovascular disease 5/170 (2.9)  PPCM 3/68 (4.4)  Pulmonary edema 0/35 (0)  Cardiomyopathy without PPCM 2/21 (9.5)  Heart failure 1/21 (4.8)  Aortic dissection 0/16 (0)  Acute myocardial infarction 0/6 (0)  Endocarditis 0/9 (0) Cerebrovascular disease 19/150 (12.7)  Intracranial hemorrhage 16/78 (20.5)  Subarachnoid hemorrhage 2/44 (4.5)  Cerebrovascular dissection 0/23 (0)  Ischemic stroke 1/10 (10.0) Infection 10/102 (9.8)  Sepsis due to all causes 8/74 (10.8)  Urinary and genital infections 1/27 (3.7)  Sepsis due to streptococcus 1/8 (12.5) Pulmonary disease 5/59 (8.5)  Pulmonary embolism 5/37 (13.5)  Asthma 0/17 (0)  Aspiration 0/5 (0) Injury, poisoning, and other external causes 2/43 (4.7) Suicide 1/16 (6.2) ICU, intensive care unit; PPCM, peripartum cardiomyopathy Outcomes for obstetric patients admitted to the ICU between antepartum and postpartum patients For consistency, if two events occurred on the same day, the interval was recorded as one day ICU, intensive care unit; IQR, interquartile range Detailed causes of ICU admission in obstetric patients between survivors and non-survivors Data are shown as number (%) unless otherwise specified ICU, intensive care unit; PPCM, peripartum cardiomyopathy Cause-specific in-hospital mortality in obstetric patients ICU, intensive care unit; PPCM, peripartum cardiomyopathy

This retrospective observational study utilized a nationwide inpatient database comprising discharge summaries and healthcare utilization information from acute care hospitals in Japan. This study was approved by the Institutional Review Board of The University of Tokyo [approval number: 3501-(5)]. Because all data were de-identified, the requirement for patient informed consent was waived. We used the Japanese Diagnosis Procedure Combination (DPC) inpatient database for the period from July 2010 to March 2022. The database contains discharge abstracts and detailed healthcare service information derived from both medical records and reimbursement claims from more than 1,000 acute care hospitals, covering approximately 90% of tertiary care emergency hospitals in Japan. All academic hospitals are required to contribute to the database, while community hospitals provide data voluntarily [ 17 ]. The database contains the following variables for each patient: sex; age; diagnoses, comorbidities at admission, and complications after admission, according to International Classification of Diseases Tenth Revision (ICD-10) codes and Japanese text; procedures recorded according to the Japanese medical procedure codes; device codes; drug codes; referral from other facilities; intensive care unit admission; teaching hospital; pregnancy status (pregnant or not); gestational age at the time of hospital admission and delivery during hospitalization; date of admission, procedures, and discharge; readmission within one week after delivery; and discharge status (in-hospital mortality, transportation to other hospital, or discharge to home). Validation of the various diagnostic records in this database was conducted to a high standard [ 18 , 19 ]. The sensitivity and specificity of the primary diagnosis in the database were 78.9% and 93.2%, respectively [ 19 ]. Patients were eligible for inclusion if they met all of the following criteria: (i) either antepartum at the time of hospital admission or assigned any of the ICD-10 codes listed in Additional file 1 ; (ii) delivered during hospitalization or within one week before hospital admission; and (iii) admitted to the ICU during hospitalization. We defined the ICU as a separate unit providing critical care services with at least one physician on site 24 h per day, around-the-clock nursing, necessary equipment to care for critically ill patients, and a nurse-to-patient ratio of at least 1:4 ( Additional file 2 ). This definition corresponds to the criteria required to claim higher levels of ICU management fees in Japan. It is also comparable to the Level 2 or Level 3 ICU classification defined by the World Federation of Societies of Intensive and Critical Care Medicine [ 20 ]. Patients characteristics included age, body mass index (BMI), gestational age at the time of hospital admission, instrumental delivery, cesarean delivery, comorbidities at hospital admission, admission to teaching hospital, emergency admission to hospital, transport by ambulance at the time of admission to hospital (including both inter-facility transfers and transport from home to hospital), and cause of ICU admission. Age was categorized into < 20, 20–24, 25–29, 30–34, 35–39, and ≥ 40 years. BMI was categorized into < 18.5, 18.5–24.9, 25.0–29.9, or ≥ 30 kg/m 2 . Gestational age at the time of hospital admission was categorized as < 28, 28 to < 32, 32 to < 37, 37 to < 42, or ≥ 42 weeks [ 21 , 22 ]. Definitions of comorbidities at admission and causes of ICU admission are provided in Additional files 3 and 4 . Procedures included blood transfusion, mechanical ventilation, transcatheter arterial embolization, hysterectomy (with or without uterine rupture), renal replacement therapy, chest compressions, extracorporeal membrane oxygenation, resuscitative endovascular balloon occlusion of the aorta, craniotomy for hematoma removal, defibrillation, and burr hole drainage. Burr hole drainage refers to a neurosurgical procedure performed to evacuate intracranial hematomas, typically subdural hematomas. Outcomes of the present study included maternal death (in-hospital mortality, ICU mortality, and cause-specific in-hospital mortality), fetal death (ICD-10 code, O364), length of hospital stay, length of ICU stay, days from hospital admission to ICU admission, days from hospital admission to delivery, and days from delivery to ICU admission. If two events occurred on the same day, the interval was recorded as one day for consistency. The patient characteristics, procedures performed, and outcomes of obstetric patients admitted to the ICU were described separately for antepartum and postpartum patients, based on their status at the time of hospital admission, and compared between the groups. The causes of ICU admission were compared between survivors and non-survivors. Categorical data were presented as numbers and percentages. Continuous data were presented as the median and interquartile range (IQR). Binary and categorical data were compared using the chi-square test. Continuous data were compared using the Wilcoxon rank sum test. Two-sided p-values of < 0.05 were considered significant. All statistical analyses were conducted using Stata/SE 18 (College Station, TX).

Pregnant and postpartum women exhibit distinct characteristics from non-pregnant women, facing the risk of serious illness due to multi-organ pathophysiological changes and pregnancy-related conditions, such as hypertensive disorders, post-partum hemorrhage and placental abruption. In recent years, the risk of severe illness in pregnant women has increased, driven by factors such as advanced maternal age, a rise in morbid obesity, abnormal placental positioning, and an increase in multiple pregnancies due to assisted reproductive technologies [ 1 – 7 ]. The World Health Organization reported that approximately 287,000 maternal deaths occurred worldwide in 2020, with 3 to 75 critically ill obstetric patients for every maternal death [ 8 , 9 ]. Healthcare providers relatively often manage critically ill obstetric patients in intensive care units (ICU) [ 10 ], understanding the epidemiology and clinical characteristics of obstetric ICU admissions is essential for improving maternal mortality and severe morbidity. Large database studies on critically ill obstetric patients in ICUs have been conducted in countries such as France, the Netherlands, Canada, China and Australia and New Zealand using large databases [ 11 – 15 ]. In Japan, our previous study of 750 obstetric patients from 61 centers using the Japanese Intensive care PAtient Database (JIPAD) provided insight into the epidemiology of obstetric patients admitted to ICUs [ 16 ]. Obstetric patients accounted for 0.41% of all ICU admissions, with an ICU mortality of 0.7% [ 16 ]. However, the study had the following limitations: only approximately 25% of all ICUs in Japan are registered with JIPAD, thus the study population may not represent the whole patients in Japan. Additionally, the leading cause of ICU admission was categorized as"pregnancy-related disorders postoperatively,"encompassing a variety of conditions, the specific details of which remain unclear. The aim of the present study was to provide descriptive epidemiological data for obstetric patients admitted to the ICU, including patient characteristics, causes of ICU admission, procedures performed, and patient outcomes. We used a nationwide inpatient database to comprehensively illustrate the critical care system for pregnant and postpartum women in Japan.

This nationwide study provides valuable insights into the clinical characteristics and outcomes of obstetric ICU admissions in Japan. Hemorrhage was the leading cause of ICU admission, and the overall in-hospital mortality was 1.1%. These findings emphasize the need for targeted interventions to reduce maternal mortality and improve outcomes for critically ill obstetric patients.

The present study provides a comprehensive overview of the epidemiological characteristics of obstetric patients admitted to the ICU in Japan. Hemorrhage was the most common cause of ICU admission, with 71% requiring blood transfusions and 28% requiring mechanical ventilation. The overall in-hospital mortality for obstetric patients admitted to the ICU was 1.1%. Compared to survivors, non-survivors were less likely to have hemorrhage-related diagnoses but had a higher prevalence of conditions such as amniotic fluid embolism, cardiovascular and cerebrovascular disease, infection, pulmonary disease, trauma, and suicide. Cause-specific in-hospital mortality was 20.5% for intracranial hemorrhage, 13.5% for pulmonary embolism, 12.5% for sepsis due to streptococcus, 11.3% for amniotic fluid embolism, and 10.8% for sepsis of all causes. In our study, hemorrhage emerged as the primary cause of ICU admissions among obstetric patients, accounting for a half of the patients—representing the highest proportion among high-income countries. Previous studies have reported hemorrhage-related ICU admissions at rates of 11% in Australia and New Zealand, 18.8% in the United States (Maryland), 34.2% in France, and 45.5% in the Netherlands [ 11 , 12 , 15 , 23 ]. A recent systematic review indicated that 31.5% of obstetric ICU admissions in high-income countries were due to hemorrhage [ 10 ]. In studies conducted after 2010 examining the causes of ICU admission for obstetric patients, hemorrhage was the leading cause in 28% (10/36) of studies conducted in low-income country settings, while it was the leading cause in 74% (17/23) of studies conducted in high-income country settings [ 10 ]. The high proportion of hemorrhage-related ICU admissions in Japan may be influenced by the country’s healthcare infrastructure. Approximately 930,000 births occur annually across 2,500 delivery facilities [ 24 , 25 ], with about half taking place in private obstetric clinics. These clinics often lack transfusion capacity, ICUs, or high-dependency units (HDUs), and detailed national data on blood product availability remain scarce. In cases of obstetric hemorrhage, delayed access to transfusion or definitive care may necessitate interfacility transfers to tertiary centers. However, such transfers are not always timely, potentially worsening patient condition at the time of ICU admission. Similar structural issues have been reported in Europe, where many obstetric units lack transfusion readiness [ 26 ] and rely on uncoordinated transfers to higher-level hospitals [ 27 ]. Importantly, these systemic limitations may not be unique to hemorrhage but may also compromise timely and effective care for other life-threatening obstetric complications. Hemorrhage was the predominant cause of ICU admissions in the present study. Notably, the proportion of patients with hemorrhage was lower among non-survivors than among survivors, suggesting that ICU access and appropriate treatment may improve survival in patients with severe hemorrhage. In contrast, amniotic fluid embolism was more prevalent among non-survivors, underscoring its rapid progression and high fatality rate [ 28 ]. Additionally, other critical conditions—such as cardiovascular disease (e.g., cardiomyopathy), cerebrovascular disease, infection (e.g., sepsis), pulmonary disease (e.g., pulmonary embolism), trauma and suicide—were more frequently observed in non-survivors compared to survivors. These findings suggest that while ICU care plays a crucial role in improving outcomes for hemorrhage-related cases, certain conditions may require enhanced preventive strategies, rapid transport systems, and specialized, time-sensitive interventions designed to mitigate their high mortality risk and further reduce maternal mortality. The in-hospital mortality rate of obstetric ICU patients in our study was 1.1%, which is slightly higher than that reported in the JIPAD study (0.7%) [ 16 ]. This difference may reflect variations in population coverage; JIPAD includes data from approximately 25% of selected ICUs, mainly university hospitals, whereas the DPC database used in the present study covers about 90% of tertiary emergency hospitals, providing a more comprehensive national perspective. Internationally, our mortality is comparable to or slightly lower than that reported in other high-income countries, such as Australia and New Zealand (0.7%), Canada (1.3%), France (1.3%), Maryland (1.8%), and the Netherlands (3.5%) [ 11 – 13 , 15 , 23 ]. Although differences in admission criteria and case mix complicate direct comparisons, the proportion of patients requiring mechanical ventilation—28% in our study—serves as a surrogate marker of illness severity. This rate is higher than that in Australia and New Zealand (19%, mortality 0.7%) [ 15 ] and similar to the Netherlands (35%, mortality 3.5%) [ 12 ], suggesting that obstetric ICU mortality in Japan is relatively low given the severity of cases. Cause-specific in-hospital mortality in the present study was generally within range of previously reported values. For intracranial hemorrhage, the in-hospital mortality in the present study (20.5%) was higher than the 10.3% reported in a nationwide hospital-based study from Japan [ 29 ], but lower than the rates exceeding 30% reported in a single-center study from China [ 30 ] and 68% in a systematic review including severe cases [ 31 ]. For pulmonary embolism, the observed mortality (13.5%) was similar to the 11.1% reported in a recent multicenter study from Singapore [ 32 ]. In contrast, mortality from sepsis due to streptococcus in the present study (12.5%) was higher than the 7% reported in a recent systematic review [ 33 ]. Likewise, mortality associated with sepsis of all causes exceeded the 1–8% range described in previous nationwide hospital-based observational studies from the United Kingdom [ 34 ] and the Netherlands [ 35 ]. Conversely, mortality due to amniotic fluid embolism (11.3%) was lower than the 15–20% reported in population-based studies from Australia [ 36 ] and the United Kingdom [ 37 ]. These discrepancies may reflect differences in disease severity, inclusion criteria, and whether only ICU-admitted patients were analyzed or all hospitalized patients were included. During the study period, 94 in-hospital deaths occurred among obstetric patients admitted to the ICU, whereas the Japan Maternal Death Exploratory Committee (JMDEC) reported 573 maternal deaths nationwide [ 38 ]. Although the JMDEC includes deaths up to one year postpartum and our study does not capture all ICU admissions of obstetric patients across Japan, this discrepancy suggests that many severely ill patients may have died before reaching ICU. In the United Kingdom, such gaps in early recognition, referral, and transport—termed"fractionated care"—have contributed to preventable maternal deaths [ 39 ], and similar issues may exist in Japan. Notably, Japan has established a robust national system for out-of-hospital cardiac arrest, featuring standardized protocols, designated centers, and coordinated prehospital triage [ 40 ]. A comparable framework tailored to maternal emergencies could help reduce delays in accessing life-saving care in other countries. The MaCriCare study in Europe reported limited ICU or HDU availability and uncoordinated transfers in obstetric care [ 26 , 27 ]. Japan also lacks national ICU guidelines specific to pregnant and postpartum women, and related practices vary across institutions. Most ICUs would not have standard protocols for obstetric consultation, fetal monitoring, uterotonic use, delivery planning, or breastfeeding support. Access to such care depends on each facility’s resources. Given the rarity and severity of these cases, developing obstetric-specific ICU protocols and promoting simulation-based multidisciplinary training are urgently needed. The present study has several limitations. First, reasons for ICU admission were inferred based on primary diagnoses due to the lack of specific information regarding ICU admissions in the DPC database, which may introduce interpretative subjectivity. Although the database has undergone rigorous validation, its accuracy has not been specially assessed in the obstetric population [ 18 , 19 ]. Second, the DPC database does not include structural information such as ICU bed capacity, staffing levels, or classification by unit size or type. Therefore, stratified analyses based on ICU characteristics could not be performed. Third, without data on overall ICU admissions, we could not calculate the proportion of obstetric patients among total ICU admissions. Fourth, we were unable to determine whether hemorrhagic events occurred intrapartum or postpartum period due to the absence of time-specific clinical data within the database. Lastly, our database does not contain detailed information regarding the exact causes of death. Approximately 35% of patients died after discharge from the ICU. This limitation constrains our ability to determine the precise timing, setting, and clinical context of deterioration and death in critically ill obstetric patients. Future research should aim to integrate more granular data on the location of maternal death, classification of hemorrhage, and incorporate extended longitudinal follow-up to enable a more comprehensive understanding of these events.

Additional file 1. Additional file 1.