Study design and setting

This retrospective cohort study used the Diagnosis Procedure Combination (DPC) database, a national administrative claim database for acutecare inpatients in Japan. This database does not include data from private clinics, as they operate under a different reimbursement system. As of 2023, the DPC database covered 1,761 facilities and 483,459 beds, representing approximately 55% of all general hospital beds in Japan [24,25]. Although this database does not cover all medical institutions, it encompasses nearly all academic hospitals and regional perinatal centers. Because intensive tocolytic management for preterm labor—specifically intravenous administration—is primarily centralized in these advanced-care facilities, this dataset strongly reflects the nationwide clinical trends in Japan. For each patient, the database provides comprehensive data, including demographics, diagnoses coded by the International Classification of Diseases 10th revision (ICD-10), daily records of medical procedures and medications, admission/discharge dates, and clinical outcomes [25]. Specifically for the perinatal field, it captures essential information such as gestational age at admission and details of delivery during hospitalization. The reliability and validity of the DPC database for clinical epidemiological research have been well-established in previous studies [26–28].

Study population

The study participants were pregnant women aged 22 ⁺ 0–36 ⁺ 6 weeks of gestational age who were hospitalized between April 1, 2012, and March 31, 2023, and received ritodrine hydrochloride injection. In Japan, tocolytic therapy is indicated for not only “preterm labor” (defined as regular painful contractions with cervical changes) but also “threatened preterm labor,” which may involve uterine contractions alone regardless of the cervical status. As this study used a nationwide administrative database, the study population encompassed the full clinical spectrum of conditions for which intravenous ritodrine was clinically indicated according to Japanese practice standards.

Patients with unknown gestational age and those transferred from another hospital were excluded, as their prior treatment status could not be ascertained. For research purposes, the dataset was first accessed on 05/02/2025.

Because the DPC database records medication data on a per-day basis, precise hourly durations cannot be determined. To ensure a clear distinction between short-term and prolonged therapy, we categorized patients into two groups: the AT group, including those receiving ritodrine for ≤2 days, and the MT group, including those receiving ritodrine for ≥4 days. Patients with a 3-day duration were excluded from the analysis to avoid potential misclassification, as this group likely contained a mix of both acute and maintenance treatment profiles.

Variables

The variables analyzed in this study included maternal age; the Charlson Comorbidity Index [29,30]; obstetric diseases—such as preterm premature rupture of membranes (pPROM), hypertensive disorder of pregnancy (HDP), placenta previa, and cervical insufficiency; pregnancy status—specifically gestational age and singleton pregnancy—along with academic hospital status, perinatal medical center designation, hospital region, and fiscal year at admission. Academic hospitals were identified as DPC Group 1 (university hospitals). Perinatal medical centers were defined as facilities that claimed the comprehensive perinatal intensive care unit management fee. The hospitals were categorized into six regions based on geographical location: Hokkaido-Tohoku, Kanto, Chubu, Kansai, Chugoku-Shikoku, and Kyushu-Okinawa. Because these institutional categories may overlap, we assessed the variance inflation factor (VIF) to detect potential multicollinearity among variables in the multivariable models. All VIF values were within acceptable limits, confirming the independent contribution of each variable.

Statistical analysis

First, we measured the AT utilization rate on both a patient-by-patient and a hospital-by-hospital basis. We analyzed annual trends in AT utilization rates using the Jonckheere-Terpstra test, to evaluate the presence of monotonic trends over time. Next, patients were divided into groups according to the method of tocolytic agent administration, after which patient and hospital characteristics were summarized using appropriate descriptive statistics. Continuous variables are expressed as means with standard deviations (SD) and were compared between groups using the Mann–Whitney U test. Categorical variables are presented as frequencies with percentages and were compared using the chi-square test. Logistic regression analysis (forced entry method) was then performed with AT utilization as the dependent variable. Results are presented as adjusted odds ratios (aOR) with 95% confidence intervals (CI). SPSS version 30 (IBM, Chicago, Illinois, USA) was used for all analyses, and the significance level was set at 5% (two-tailed test).

Ethical considerations

The Institute of Science Tokyo Faculty of Medicine Ethics Review Committee approved this study (approval number: M2000-788; date of approval: December 19, 2023). The committee waived the requirement for informed consent because all personal information was excluded, and de-identified data were provided to the researchers for secondary use. This study was conducted in accordance with the principles of the Declaration of Helsinki.

Participants

During the study period, 191,201 patients were identified who received ritodrine infusion. We first established a population of 164,407 patients from 720 hospitals to analyze annual trends. This cohort excluded those with unknown or ineligible gestational age (<22 or ≥ 37 weeks) and patients transferred from other facilities, ensuring the analysis reflected the full course of inpatient management (Fig 1). For the subsequent analysis of factors associated with tocolytic protocols, we further excluded patients who received ritodrine for exactly 3 days. This was done to minimize misclassification between the two distinct treatment strategies. Consequently, 156,356 patients were included in the multivariable logistic regression analysis to compare the AT and MT.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Flowchart of study population.

The annual trends by duration of ritodrine administration were based on 164,407 patients from 720 hospitals who were hospitalized and received intravenous ritodrine hydrochloride from April 2012 to March 2023, excluding patients with a pregnancy of 37 weeks or more, less than 22 weeks, unknown gestational age, and those admitted for maternal transfer after discharge. The analysis of factors associated with AT included 156,356 patients, excluding patients with a mixed AT and MT who received ritodrine for exactly 3 days.

https://doi.org/10.1371/journal.pone.0351293.g001

The tocolysis duration significantly decreased over time (Fig 2a). The Jonckheere–Terpstra test confirmed significant directional shifts at both ends of the duration spectrum. The proportions of patients in the groups receiving ≥ 14 days of tocolysis showed a significant decreasing trend, while those receiving ≤ 2 days significantly increased. In contrast, no significant trends were observed in the intermediate groups with tocolysis durations of 3–13 days (Fig 2a, S2 Table). The AT utilization rate by hospitals also showed a monotonically increasing trend, although the variation between hospitals increased over time (Fig 2b).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Infusion period of ritodrine hydrochloride.

a: Trends in ritodrine infusion duration among all study participants (n = 164,407). Stacked area chart showing the annual proportions of tocolytic treatment durations from 2012 to 2022. The vertical axis represents the percentage of patients, and the horizontal axis denotes the fiscal year. The area for acute tocolysis (AT, ≤ 2 days) demonstrates a significant increasing trend (Jonckheere-Terpstra test: p = 0.036). The utilization rate of AT (≤2 days) showed an increasing trend (Jonckheere-Terpstra test: p = 0.036). b: Annual trends and inter-hospital variation in AT utilization rates (n = 720 hospitals). The box-and-whisker plots illustrate the distribution of AT utilization rates (≤ 2 days) across facilities for each fiscal year. The horizontal line within each box represents the median, while the × symbol indicates the mean utilization rate. The box boundaries indicate the interquartile range (IQR), and the whiskers extend to the minimum and maximum values (excluding outliers). While a significant increasing trend was observed (Jonckheere-Terpstra test: p = 0.016), the widening of the boxes and whiskers reflects the increasing diversity in clinical protocols among institutions.

https://doi.org/10.1371/journal.pone.0351293.g002

Table 1 summarizes the patient characteristics categorized by tocolytic duration. The mean maternal age was similar between the two groups (AT: 31.92 ± 5.48 years; MT: 31.76 ± 5.41 years), although the difference reached statistical significance due to the large sample size. Notably, the mean gestational age at admission was significantly higher in the AT group compared with the MT group (31.89 ± 4.41 weeks vs. 29.02 ± 4.40 weeks, p < 0.001). Regarding obstetric complications, patients with pPROM or HDP were more frequently managed with AT. Conversely, multiple pregnancies, placenta previa, and cervical insufficiency were more prevalent in the MT group. Institutional characteristics also influenced treatment, with academic hospitals and perinatal centers showing a higher tendency toward AT. Additionally, the AT group had a higher rate of antenatal corticosteroid administration (23.90% vs. 21.99%, p < 0.001), while the concomitant use of magnesium sulfate was more frequent in the MT group (23.30% vs. 19.05%, p < 0.001).

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Characteristics of the participants according to ritodrine hydrochloride infusion (N = 156,356).

https://doi.org/10.1371/journal.pone.0351293.t001

The results of the multivariable logistic regression analysis are presented in Fig 3 and S3 Table. Institutional factors significantly influenced the choice of protocol; admission to academic hospitals (aOR: 1.16; 95% CI: 1.12–1.20) and perinatal centers (aOR: 1.12; 95% CI: 1.08–1.15) were positively associated with AT use. We also observed a distinct regional gradient, with the Kyushu-Okinawa region showing the strongest association with AT (aOR: 1.79; 95% CI: 1.70–1.90) compared to the reference region. Furthermore, the likelihood of AT utilization increased progressively over time, particularly from 2018 onward, reaching an aOR of 1.55 (95% CI: 1.45–1.66) in 2022 compared to 2012 (all p < 0.001). Regarding maternal obstetric complications, pPROM (aOR: 3.18; 95% CI: 3.06–3.30) and HDP (aOR: 1.52; 95% CI: 1.43–1.62) were associated with AT. In contrast, multiple pregnancies (aOR: 0.64; 95% CI: 0.61–0.67) and placenta previa (aOR: 0.67; 95% CI: 0.63–0.71) were associated with MT. For every one-week increase in gestational age at admission, the odds of receiving AT increased by 1.13 times (95% CI: 1.12–1.14).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Forest plot of factors associated with the utilization of acute tocolysis.

A multivariable logistic regression analysis was performed to identify factors associated with AT (≤ 2 days). The dependent variable was the presence of AT (≤ 2 days). The model was adjusted for the following covariates: preterm labor management (magnesium sulfate use and antenatal corticosteroid administration), healthcare provider factors (academic hospital and perinatal medical center), region (reference: Hokkaido-Tohoku), fiscal year, and maternal obstetric complications (pPROM, HDP, gestational age, placenta previa, multiple pregnancy, and cervical insufficiency). *1: magnesium sulfate, *2: antenatal corticosteroid, *3: preterm premature rupture of membranes, *4: hypertensive disorder of pregnancy.

https://doi.org/10.1371/journal.pone.0351293.g003

Trends in AT adoption

International guidelines generally recommend AT because there is no robust evidence that tocolysis beyond 48 hours improves neonatal outcomes [7,11].Conversely, adherence to guidelines is reported to be low, and in actual clinical practice, MT after AT and the administration of tocolytic agents for threatened preterm labor are still performed [12–14,31,32]. Rousseau et al. identified barriers to guideline adherence, including physicians over 50 years of age, those working in non-university hospitals, and those fearing change [31]. Additionally, Lee et al. stated that patient opinion plays a major role in obstetricians’ decision-making, and that maintenance medications may be prescribed based on the patient’s request [32].

In Japan, ritodrine is the first-line drug for preterm labor, and its use for MT aligns with both domestic guidelines and the package insert [19]. The JSOG guidelines, revised every three years since 2008, have played a key role in maintaining Japan’s low preterm birth rate, even as maternal aging and the use of assisted reproductive technologies have increased [17]. The JSOG guidelines first introduced the international trend of AT utilization in 2014 by citing the European Medicines Agency’s 48-hour restriction and the Food and Drug Administration’s warnings. While the JSOG did not formally restrict the duration of tocolysis to 48 hours—maintaining that the decision should be based on clinical context and maternal safety—the repeated inclusion of these international perspectives in subsequent revisions (2017, 2020, and 2023) likely increased awareness among Japanese clinicians. This likely contributed to the shift toward AT, particularly in academic centers where physicians are more exposed to international evidence. The rate of ritodrine injection use in Japan decreased by 15.3% from 2017 to 2018 and increased to 47.4% by 2022 [21]. Multivariate logistic regression analysis suggested that the association with AT has strengthened over time since 2018, likely due to the influence of guideline descriptions such as this. Furthermore, patients treated at academic hospitals and perinatal centers were strongly associated with AT. Many of the committee members who create the guidelines are doctors employed at such facilities, and it is speculated that their frequent exposure to diverse information and publications from Japan and abroad influences their choice of AT. This finding is consistent with the results of Rousseau, et al. [31].

Interestingly, ACS administration was not a significant predictor of AT in our study results; the administration rates were 23.90% in the AT group and 21.99% in the MT group. This lack of association reflects the unique clinical landscape in Japan, where tocolytic duration and ACS administration are often handled as independent clinical decisions. Unlike in many Western countries, where tocolysis is primarily indicated as a 48-hour bridge for ACS, Japanese clinical practice has historically focused on long-term pregnancy prolongation. Consequently, the choice of AT in Japan may be driven more by factors such as the physician’s adherence to global evidence-based trends or the severity of symptoms, rather than being specifically coordinated with the ACS administration window.

Regional and institutional variations in tocolytic practice

This study showed that regional factors are associated with AT use. Previous research has reported that differences in the provision of perinatal medical care between urban and rural areas, as well as availability of obstetricians and gynecologists, can influence medical decision-making and care for mothers and children [33,34]. It is hypothesized that in areas with low population density and vast land area, factors such as long travel distances for obstetric checkups and births, difficulty in transporting the mother due to snow in winter, and availability of obstetricians and pediatricians can influence the choice of MT [35,36]. These findings indicate the importance of selecting the most appropriate treatment for mothers and children based on local healthcare infrastructure and regional conditions.

Maternal and obstetric conditions influencing tocolysis

The utilization rate of AT has increased over the past 11 years; however, the preterm birth rate in Japan has remained stable at 5.48%–5.75% for those under 37 weeks of gestation and 0.25%–0.26% for those under 28 weeks of gestation [16]. It is noteworthy that while the utilization of long-term tocolysis decreased and that of AT increased over the 11-year period, the preterm birth rate in Japan remained stable, being significantly lower than that in the United States, which ranges from 9.7% to 10.4% [36]. Although the contribution of tocolysis beyond 48 hours to a further reduction in preterm birth rates remains controversial, the stability of Japan’s preterm birth rate, despite a decrease in long-term tocolysis, suggests that the shift toward shorter-term management has not compromised overall neonatal outcomes.

Enengl et al. identified placenta previa as a specific indication for MT [13]. Similarly, our findings indicate that placenta previa is more strongly associated with MT. In contrast, conditions such as pPROM and severe HDP, which require careful management due to the potential adverse effects of prolonged gestation on maternal and fetal outcomes, are more strongly associated with AT. In addition, the early gestational age at admission tends to be more closely linked to MT. These findings suggest that the adoption of AT is not uniform across facilities and that treatment options are often selected based on the maternal condition and clinical context.

Importance of shared decision-making in tocolysis

Major international guidelines, including those from the WHO, ACOG, and NICE, emphasize that tocolytic therapy must involve shared decision-making (SDM). This process requires providing pregnant women and their families with comprehensive information regarding the risks and benefits of both AT and MT, potential neonatal outcomes, and the limitations of survival at various gestational ages [7,11]. While the present study—relying on administrative data—could not directly evaluate the quality of these counseling processes, our findings highlight a significant gap: there is still a lack of evidence-based frameworks to guide SDM that account for regional healthcare infrastructure, financial resources, and geographical constraints. Future research should prioritize addressing these factors to facilitate more personalized and informed decision-making in tocolytic management [9].

Strengths and limitations of the study

A major strength of this study is its longitudinal design, covering an 11-year period across 720 hospitals, which ensures high representativeness of the clinical landscape in Japan. Furthermore, our methodology—using administrative claims data to analyze long-term trends and factors influencing clinical practice—offers a scalable model for similar investigations in other countries.

However, several limitations must be acknowledged. First, the retrospective nature of this cohort study introduces the potential for “confounding by indication”; patients receiving different treatment durations likely varied in clinical severity or other unmeasured factors that influenced the initial choice of therapy. Second, the use of the DPC database limits access to granular clinical details, such as cervical length, fetal fibronectin levels, or the specific qualitative rationale behind a physician’s decision. Third, the DPC exclusively covers acute care hospitals and excludes private clinics. While some low-risk patients may receive initial treatment at clinics, our study focuses on facilities where critical decisions such as transition to AT and corticosteroid administration are primarily made. Thus, our findings provide a crucial benchmark for high-risk perinatal care in Japan. Finally, the generalizability of our results may be limited; ritodrine is not available in the United States, and the Japanese definition of “threatened preterm labor” is broader than that per international standards. Therefore, while our study findings reliably reflect Japanese clinical practice, the comparative effectiveness of these trends with regard to long-term neonatal outcomes requires further investigation.

Implications for clinical practice and future research

The findings of this study have important implications for clinical practice and education. By identifying factors that influence the use of AT, this study provides evidence for obstetricians to tailor treatment choices to guidelines and patient conditions. These results may also support shared decision-making. Furthermore, these findings may raise awareness among trainees about unresolved challenges in routinely performed tocolysis. Future prospective studies should compare maternal and neonatal outcomes of long-term versus short-term tocolysis in Japan and include economic evaluations of prolonged therapy. Therefore, to maintain and improve the quality of maternal and child health, further research is needed to establish protocols and evaluate available tocolytic drugs, taking into consideration work style reforms for doctors, shortages of obstetricians and pediatricians, and geographical issues.