Trends in labor induction and perinatal outcomes: A decade comparison before and after the ARRIVE trial Authors: Wing Lam Tsui, Dah-Ching Ding Published: 2026 ABSTRACT ABSTRACTObjectives:This study aimed to compare maternal characteristics, delivery outcomes, and neonatal conditions in 2013 and 2023 at our hospital and to evaluate the association between labor induction and neonatal outcomes.Materials and Methods:A retrospective cohort study was conducted, including 600 women who delivered in 2013 (n = 316) and 2023 (n = 284). Cases involving scheduled cesarean delivery, vaginal birth after cesarean, missing data, or immediate neonatal death were excluded. Maternal demographics, obstetric parameters, and neonatal outcomes were compared. Logistic regression was performed to assess the association between induction of labor and adverse outcomes.Results:The mean maternal age increased significantly from 29.73 to 31.29 years (P < 0.001), and the induction rate more than doubled from 19.94% to 50.70% (P < 0.001). Although cesarean rates were unchanged, vacuum extraction deliveries increased (15.82% vs. 27.46%, P < 0.001), and spontaneous vaginal deliveries decreased (80.70% vs. 68.31%, P < 0.001). Neonatal intensive care unit (NICU) admissions rose significantly (2.53% vs. 7.39%, P < 0.001), while low 1-min Apgar scores (≤6) remained stable. Regression and sensitivity analyses revealed that induction of labor was independently associated with a higher risk of low Apgar scores (adjusted odds ratio = 3.24, 95% confidence interval = 1.27–8.28, P = 0.014), but not with increased rates of NICU admission or vacuum-assisted delivery.Conclusion:Over the past decade, both maternal age and labor induction rates have increased markedly. While cesarean rates remained steady, induction of labor was associated with greater neonatal morbidity, reflected by increased low Apgar scores. These findings emphasize the need for careful clinical judgment and close monitoring during labor induction to optimize perinatal outcomes. INTRODUCTION Labor induction rates have risen significantly in recent decades, with approximately 25%–46% of pregnancies being induced [12]. Common indications include postterm pregnancy, hypertension/preeclampsia, fetal growth restriction, and spontaneous membrane rupture [34]. While evidence supports induction for postterm pregnancies and hypertension/preeclampsia, many other indications lack strong supporting evidence [3]. Induction is associated with increased rates of cesarean sections and postpartum hemorrhage compared to spontaneous labor [4]. The decision to induce should balance the benefits of prompt delivery against the risks of waiting for spontaneous labor onset [2]. Providing accurate information to women undergoing induction is crucial for informed decision-making and empowerment [1]. Further research is needed to establish evidence-based guidelines for many common induction indications [3]. The publication of the ARRIVE trial in 2018 led to significant changes in obstetric practices, particularly regarding elective labor inductions. Multiple studies observed a substantial increase in elective inductions among nulliparous women at term following the trial’s publication [567]. However, the impact on other outcomes was less clear. While some studies found no significant changes in cesarean delivery rates or hypertensive disorders of pregnancy [67], others reported an increase in cesarean deliveries among electively induced women [5]. Fineberg et al. found that labor induction was associated with increased odds of cesarean delivery both before and after the ARRIVE trial [8]. Additionally, some studies noted longer hospital stays, increased blood loss, and lower birth weights in the post-ARRIVE cohort [5]. These findings suggest that the ARRIVE trial’s results may not be universally applicable to general obstetric populations. While induction at 39 weeks may reduce certain complications such as perineal injury, macrosomia, and low Apgar scores [9], it is also associated with higher risks of neonatal issues such as chorioamnionitis, bacterial sepsis, and intracranial hemorrhage [10]. For women with a previous cesarean, induction carries increased risks of adverse perinatal outcomes and cesarean delivery [11]. However, induction does not appear to negatively impact long-term academic achievement, as evidenced by similar third-grade test scores between induced and expectantly managed births [12]. While induction may be beneficial for some women, potential risks should be carefully considered, especially for nulliparous women who may face an increased likelihood of shoulder dystocia [9]. However, there is a lack of information regarding the influence of elective induction in the Chinese community; hence, this study aimed to explore the number of elective inductions in 2013 and 2023 (before and after the ARRIVE trial) and the associated perinatal outcomes. MATERIALS AND METHODS Patient selection This was a single-institution retrospective cohort study conducted at Hualien Tzu Chi Hospital and approved by the Research Ethics Committee of Hualien Tzu Chi Hospital (IRB113-131-B). This study was conducted in accordance with the Declaration of Helsinki. Due to the retrospective nature of the study, the Research Ethics Committee of Hualien Tzu Chi Hospital waived the need for informed consent. This study adhered to relevant ethical standards and guidelines. The patients’ data were retrieved from the electronic medical records stored in the hospital. The inclusion criteria were patients who delivered a child in 2013 or 2023 at our hospital. The exclusion criteria were as follows: (1) scheduled cesarean section (C/S), (2) vaginal birth after a previous C/S, (3) missing data, or (4) immediate death of a newborn after birth or termination cases. The basic characteristics of the recruited participants, including age, body mass index (BMI), parity, gestational age, labor induction, birth weight, presentation, and maternal complications (gestational diabetes mellitus [GDM], hypertension, and preeclampsia), were recorded. Labor induction was indicated with postterm pregnancy, hypertension, intrauterine growth restriction, and prelabor rupture of membranes [13]. Elective induction was not the reason mentioned above. Outcome measures The primary outcome was the number of labor inductions performed in 2013 and 2023. Secondary outcomes included the mode of delivery (normal vaginal delivery [NSD], vacuum extraction delivery [VED], and C/S), maternal blood loss, perineal laceration degrees, and fetal complications (1st minute APGAR ≤ 6 and neonatal intensive care unit [NICU]). Statistical analysis In this study, continuous variables are presented as mean ± standard deviation. Fisher’s exact and Chi-squared tests were used for dichotomous variables, whereas an independent Student’s t-test was used for continuous variables to determine statistical significance. Multivariate logistic regression was performed on the entire cohort, adjusting for study covariates to assess the odds ratio (OR) of the induction of labor associated with the risk of neonatal outcome or VED. Sensitivity analyses were conducted using inverse probability of treatment weighting to account for residual confounding. Stabilized propensity score weights were applied in weighted logistic regression models. E-values were calculated for significant associations to assess the robustness of findings to potential unmeasured confounding. Statistical significance was set at P < 0.05. All statistical analyses were performed using SPSS version 25 (IBM Corp., Armonk, NY, USA). RESULTS Basic characteristics Figure 1 illustrates the selection flow of study participants from deliveries at Hualien Tzu Chi Hospital in 2013 and 2023. A total of 933 cases were initially identified, encompassing all newborn deliveries during the two study periods (January 1–December 31, 2013, and January 1–December 31, 2023). After applying exclusion criteria – including scheduled cesarean sections (n = 203), vaginal births after cesarean section (n = 7), cases with missing data (n = 100), and those with immediate neonatal death or termination (n = 23) – a total of 600 eligible cases were included for analysis. These were divided into two cohorts: 316 deliveries from 2013 and 284 from 2023, representing the final study population used for comparative analyses of maternal and neonatal outcomes. Study flow chart Table 1 summarizes the comparison of maternal characteristics, obstetric parameters, and complications between 2013 and 2023. The mean maternal age significantly increased from 29.7 ± 5.4 years in 2013 to 31.3 ± 5.4 years in 2023 (P < 0.001), while parity and gestational age showed no significant differences. Birth weight was slightly lower in 2023 (2963.8 ± 400.2 g) compared with 2013 (3046.2 ± 460.1 g) (P = 0.019). Maternal BMI remained comparable between the two groups. The overall induction rate was significantly higher in 2023 than in 2013 (50.70% vs. 19.94%, P < 0.001). Inductions at 37–38 + 6 weeks were more common in 2023 (P < 0.001), whereas 39–40-week inductions were more frequent in 2013 (P = 0.003). Elective induction rates were similar between groups (P = 0.410). However, medically indicated induction differed significantly (P = 0.030), with hypertension more common in 2023 (P = 0.001) and postterm pregnancy more common in 2013 (P = 0.004). Regarding maternal complications, GDM increased significantly from 1.6% to 4.9% (P = 0.019), whereas rates of hypertension and preeclampsia did not differ significantly between the two cohorts. Overall, these results indicate that by 2023, mothers were older and more likely to undergo labor induction and experience GDM, despite stable BMI and obstetric outcomes. Table 1Baseline characteristics20132023 P Patient, n (%)316 (52.67)284 (47.33)Maternal age, mean±SD29.73±5.3731.29±5.44<0.001Parity, mean±SD1.67±1.021.79±1.020.159Gestational weeks, mean±SD38.58±1.8138.35±1.410.090Birth weight (g), mean±SD3046.20±460.102963.80±400.200.019BMI, mean±SD26.64±4.0927.12±5.440.228Induction, n (%)63 (19.94)144 (50.70)<0.00137–38+6 weeks of GA21 (33.33)92 (63.88)<0.00139–40 weeks of GA27 (42.8)47 (32.63)0.003Elective induction of labor38 (77.78)130 (82.64)0.410GDM4 (10.52)5 (3.84)0.18Large for date3 (7.89)11 (8.46)1Others1 (2.63)1 (0.76)1Induction with indications25 (39.68)14 (9.72)0.030Hypertension7 (21.42)12 (85.71)0.001IUGR2 (7.14)00.52Nonreassuring fetal status3 (21.42)1 (7.14)0.62Postterm13 (52.00)1 (7.14)0.004ComplicationGDM5 (1.58)14 (4.93)0.019Hypertension7 (2.22)11 (3.87)0.235Preeclampsia7 (2.22)4 (1.41)0.552n: Number, P: P value, SD: Standard deviation, BMI: Body mass index, GDM: Gestational diabetes mellitus, IUGR: Intrauterine growth restriction, GA: Gestational age Baseline characteristics n: Number, P: P value, SD: Standard deviation, BMI: Body mass index, GDM: Gestational diabetes mellitus, IUGR: Intrauterine growth restriction, GA: Gestational age Delivery outcomes and neonatal conditions in 2013 and 2023 Table 2 compares delivery methods, neonatal outcomes, and delivery-related trauma between 2013 and 2023. The proportion of NSD decreased significantly from 80.7% to 68.3% (P < 0.001), while VED increased from 15.8% to 27.5% (P < 0.001), indicating a growing reliance on assisted vaginal delivery. The induction-C/S rate remained stable between the 2 years (3.5% vs. 4.2%, P = 0.635). Neonatal outcomes revealed a higher rate of NICU admissions in 2023 compared with 2013 (7.4% vs. 2.5%, P < 0.001), though low 1-minute Apgar scores (≤6) showed no significant change. Regarding delivery-related trauma, perineal laceration patterns shifted significantly (P = 0.020): first-degree tears became more frequent (3.6%–8.1%), while fourth-degree tears disappeared (2.3%–0%). Although mean blood loss increased from 201.5 ± 155.4 mL to 231.7 ± 265.4 mL, the difference was not statistically significant (P = 0.094). Overall, the findings suggest an increase in the rate of assisted deliveries and NICU admissions in 2023, accompanied by changes in perineal injury profiles, but stable cesarean and Apgar outcomes. Table 2Comparison of delivery methods, neonatal conditions, and delivery-related trauma in 2013 and 20232013 (n=316)2023 (n=284) P Delivery method, n (%) Induction-C/S11 (3.48)12 (4.23)0.635 VED50 (15.82)78 (27.46)<0.001 NSD255 (80.7)194 (68.31)<0.001NICU admission, n (%)8 (2.53)21 (7.39)<0.0011 min APGAR≤6, n (%)16 (5.06)14 (4.93)0.940Delivery-related trauma Perineal laceration, n (%)0.020 016 (5.3)16 (5.9) 111 (3.64)22 (8.12) 2257 (85.1)222 (81.92) 311 (3.64)11 (4.06) 47 (2.32)0Blood loss, mean±SD201.5±155.4231.7±265.40.094C/S: Cesarean section, NICU: Neonatal intensive care unit, VED: Vacuum extraction delivery, NSD: Normal spontaneous delivery, SD: Standard deviation Comparison of delivery methods, neonatal conditions, and delivery-related trauma in 2013 and 2023 C/S: Cesarean section, NICU: Neonatal intensive care unit, VED: Vacuum extraction delivery, NSD: Normal spontaneous delivery, SD: Standard deviation The outcomes associated with the induction of labor Table 3 presents the regression analysis examining outcomes associated with the induction of labor. After adjustment for maternal age, BMI, parity, gestational age, year, and maternal comorbidities (hypertension, diabetes mellitus, and preeclampsia), induction of labor was significantly associated with higher risks of adverse neonatal outcomes. Specifically, induced labor increased the adjusted odds of NICU admission nearly threefold (adjusted OR [aOR] = 2.94, 95% confidence interval [CI] = 1.01–8.52, P = 0.047) and the likelihood of a low 1-min Apgar score (≤6) more than threefold (aOR = 3.24, 95% CI = 1.27–8.28, P = 0.014). However, there was no significant association between induction and the rate of VED (aOR = 0.68, 95% CI = 0.42–1.10, P = 0.117). These findings suggest that while labor induction may not increase the need for assisted vaginal delivery, it is independently associated with a higher risk of transient neonatal distress requiring NICU care. Table 3Regression analysis of outcome associated with induction of laborOutcomeOR95% CI P aOR95% CI P E NICU1.320.64–2.730.4552.941.01–8.520.0475.33APGAR score ≤ 61.710.82–3.580.1553.241.27–8.280.0145.93VED1.040.69–1.560.8600.680.42–1.100.1172.30Adjust: Maternal age, BMI, parity, gestational age week, year, hypertension, diabetes mellitus, preeclampsia. NICU: Neonatal intensive care unit, VED: Vacuum extraction delivery, BMI: Body mass index, CI: Confidence interval, OR: Odds ratio, aOR: Adjusted odds ratio Regression analysis of outcome associated with induction of labor Adjust: Maternal age, BMI, parity, gestational age week, year, hypertension, diabetes mellitus, preeclampsia. NICU: Neonatal intensive care unit, VED: Vacuum extraction delivery, BMI: Body mass index, CI: Confidence interval, OR: Odds ratio, aOR: Adjusted odds ratio Sensitivity analysis In the weighted analysis [Table 4], the exposure (induction of labor) was not significantly associated with NICU admission (OR 1.95, 95% CI 0.68–5.56; P = 0.211), with an E-value of 3.31. However, a significantly increased risk of low Apgar score (≤6) was observed (OR 3.15, 95% CI 1.42–7.03; P = 0.005). The corresponding E-value was 5.75, suggesting that a relatively strong unmeasured confounder would be required to fully explain the observed association. In contrast, the association with VED did not reach statistical significance (OR 0.65, 95% CI 0.41–1.02; P = 0.061), with an E-value of 2.45. Table 4Association between exposure and neonatal outcomes using inverse probability of treatment weighting analysis with corresponding E-valuesOutcomeOR95% CI P E NICU1.950.68–5.560.2113.31APGAR score ≤ 63.151.42–7.030.0055.75VED0.650.41–1.020.0612.45NICU: Neonatal intensive care unit, VED: Vacuum extraction delivery, CI: Confidence interval, OR: Odds ratio Association between exposure and neonatal outcomes using inverse probability of treatment weighting analysis with corresponding E-values NICU: Neonatal intensive care unit, VED: Vacuum extraction delivery, CI: Confidence interval, OR: Odds ratio DISCUSSION The findings highlight significant temporal and clinical changes associated with labor induction and its outcomes. From 2013 to 2023, maternal age increased, and the rate of labor induction more than doubled (from 19.9% to 50.7%, P < 0.001), with a parallel rise in GDM. Although cesarean rates remained stable, vacuum extraction deliveries became more frequent, and spontaneous vaginal deliveries declined. Neonatal outcomes worsened slightly, with NICU admissions rising from 2.5% to 7.4% (P < 0.001), despite no significant difference in low Apgar scores. Regression analyses further demonstrated that induction of labor independently increased the odds of adverse neonatal outcomes – specifically, NICU admission (aOR = 2.94, 95% CI = 1.01–8.52, P = 0.047) and low 1-min Apgar scores (aOR = 3.24, 95% CI = 1.27–8.28, P = 0.014) – after adjustment for confounding factors. Sensitivity analysis revealed the same association between induction of labor and low 1-min Apgar scores. Collectively, these results suggest that while induction of labor has become increasingly common over the past decade, it is associated with higher rates of neonatal morbidity, underscoring the need for careful case selection and perinatal monitoring. Multiple large cohort studies and meta-analyses demonstrate that elective induction of labor at term is not associated with an increased risk of NICU admission compared to spontaneous labor or expectant management. A multicenter analysis of over 140,000 term births found that induction of labor was not associated with increased odds of NICU admission (aOR 0.95, 95% CI 0.89–1.01) compared to spontaneous labor [14]. Systematic reviews and meta-analyses show either no significant difference or a reduced risk of NICU admission with induction, particularly among nulliparous women (OR 0.75, 95% CI 0.63–0.89) [915]. Some population-based studies report that induction of labor may be associated with higher rates of certain neonatal complications (e.g., low Apgar scores, sepsis, assisted ventilation), but causality is unclear and overall NICU admission rates are not consistently increased [101617]. Hospital-level analyses indicate no significant association between higher induction rates and increased NICU admissions; in fact, higher induction rates are linked to improved perinatal outcomes such as reduced stillbirth and mechanical ventilation, without a corresponding increase in NICU admissions [18]. However, our study showed induction of labor was associated with a higher chance of low Apgar score and NICU admission. A large population-based cohort study demonstrated that women undergoing induction of labor had a 20%–50% higher risk of vacuum extraction after adjustment for confounding factors, and this association persisted even in women without recognized medical complications [19]. Vacuum extraction is typically indicated for fetal compromise or failure to progress in the second stage of labor, both of which are more common following induction due to factors such as uterine contractility and cervical readiness [20]. The increased rate of vacuum extraction after induction should be considered when counseling patients about the risks and benefits of labor induction [19]. However, our study did not show induction of labor associated with a higher risk of VED. The rising induction rate observed in our study appears to be multifactorial. Changes in gestational age composition contributed to this trend, with significant differences in inductions occurring at 37–38 + 6 weeks and 39–40 weeks of gestation, suggesting a temporal shift in delivery timing practices. In addition, analysis of induction subtypes demonstrated that medically indicated inductions, particularly those related to hypertensive disorders and postterm pregnancy, accounted for a substantial proportion of the observed differences. In elective indications, such as GDM and large for gestational age, there were no significant differences. These findings suggest that evolving obstetric risk profiles and gestational age management strategies may partly explain the increasing induction rate. Nevertheless, because provider-level decision-making processes, institutional policies, and patient preferences were not captured in this retrospective dataset [212223]. Further prospective and qualitative studies are warranted to better elucidate the underlying drivers of this trend. Despite a substantial increase in labor induction, the cesarean section rate remained stable over time in our study. The World Health Organization suggests that cesarean section rates above approximately 10%–15% at the population level are not associated with further reductions in maternal or neonatal mortality [242526]. This benchmark emphasizes the importance of avoiding unnecessary cesarean deliveries. However, the concurrent increase in VED and decrease in NSD among induced labors suggest a shift in the mode of vaginal delivery [23]. These may have implications for maternal recovery and childbirth experience (higher rates of perineal trauma, postpartum pain, short-term pelvic floor symptoms, and stress) [27]. These findings underscore the importance of evaluating the overall obstetric quality of care, beyond cesarean section rates alone. The increasing use of obstetric interventions, including labor induction, should be interpreted within the context of contemporary perinatal care [28]. The primary objective is to optimize maternal and neonatal outcomes rather than to minimize intervention per se [29]. Current induction practices are increasingly guided by risk stratification, enabling timely delivery while avoiding unnecessary cesarean delivery [3031]. In our cohort, labor induction was associated with lower Apgar scores. Overall, stable perinatal outcomes despite rising induction rates suggest targeted use of obstetric interventions that balance neonatal safety. Importantly, our findings should not be interpreted as evidence that induction of labor increases neonatal risk, nor do they contradict randomized controlled trial data, including the ARRIVE trial. Randomized evidence has demonstrated that appropriately selected induction at term does not worsen perinatal outcomes and may, in some contexts, improve them. In contrast, our study is observational and reflects real-world clinical practice across varying indications and risk profiles. Therefore, the associations observed herein likely reflect underlying maternal–fetal characteristics and clinical contexts rather than a causal effect of induction itself. The strengths of this study include its large sample size and retrospective cohort design, allowing for a comprehensive comparison of labor induction trends and perinatal outcomes in the 2 years. By analyzing data from a single institution, the study ensures consistency in clinical practices and data collection methods, reducing variability. Additionally, the inclusion of multivariate regression analysis strengthens the findings by identifying independent risk factors associated with delivery mode and neonatal outcomes. The study also provides valuable insights into the impact of elective induction on obstetric trends in an Asian population, an area with limited existing research. This study had some limitations. The study’s retrospective nature may introduce selection and recall biases, which could affect the accuracy of the data and limit the ability to establish causal relationships. This study was conducted at a single tertiary referral center in Eastern Taiwan, and the observed outcomes likely reflect local practice patterns, patient characteristics, and institutional expertise. Therefore, caution should be exercised when generalizing these findings to other healthcare settings or regions with different clinical practices or patient populations. Because labor induction is a non-random clinical exposure influenced by underlying maternal and fetal risk profiles, residual confounding from unmeasured or incompletely captured factors may persist despite multivariable adjustment, and the observed associations should therefore be interpreted with caution. This study compares outcomes between two selected calendar years (2013 and 2023) and does not represent a continuous longitudinal assessment of temporal trends. As such, the observed differences may partially reflect year-specific practices, including changes in obstetric protocols, neonatal care standards, or NICU admission criteria, rather than true secular trends. NICU admission is influenced by institutional policies and surveillance intensity and may not uniformly reflect neonatal morbidity; therefore, changes in admission thresholds or monitoring practices over time could have contributed to the observed differences independent of true neonatal risk. Some adjusted estimates were associated with wide CIs, indicating limited precision that may reflect constrained statistical power and the inclusion of multiple covariates relative to the number of outcome events. Accordingly, the potential for model overfitting cannot be fully excluded, and these estimates should be interpreted with caution. CONCLUSION Over the past decade, labor induction has become substantially more common, coinciding with an increase in maternal age and gestational diabetes prevalence. Overall, delivery methods and maternal complications remained stable. Regression analysis revealed that induction of labor is independently associated with a higher risk of transient neonatal compromise, as reflected by increased odds of low Apgar scores. The findings suggest that while labor induction plays a crucial role in modern obstetric practice, its increasing use warrants careful clinical judgment and enhanced perinatal surveillance to balance the benefits of timely delivery with potential neonatal risks. Data availability statement The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. Conflicts of interest Dr. Dah-Ching Ding, an editorial board member at Tzu Chi Medical Journal, had no role in the peer review process of or decision to publish this article. The other author declared no conflicts of interest in writing this paper.