Perinatal complications and time to subsequent pregnancy after open, laparoscopic, and hysteroscopic myomectomy: a retrospective cohort study

Abstract

Background

We aimed to evaluate the risk of perinatal complications in subsequent pregnancies after different types of myomectomy, viz. open, laparoscopic, or hysteroscopic. Moreover, we investigated whether the time interval from myomectomy to subsequent pregnancy (TIMP) is a risk factor for perinatal complications.

Methods

This retrospective cohort study analyzed data from the vast Japanese health insurance JMDC database between January 2008 and July 2024. We identified primiparous women and excluded participants based on the following criteria: age < 20 years at delivery, diagnosis of adenomyosis, multiple pregnancy, or history of repeated myomectomy using different approaches. The occurrence of placenta accreta spectrum (PAS), placenta previa, uterine rupture, gestational hypertension/preeclampsia, and placental abruption was compared among women who underwent open, laparoscopic, or hysteroscopic myomectomy and those in the control group. Subsequently, for each myomectomy procedure, we compared the TIMP between women with and without each perinatal complication. Fisher’s exact test and multivariable logistic regression models were employed.

Results

Among the 27,129 eligible women, 140, 305, and 97 underwent open, laparoscopic, and hysteroscopic myomectomy, respectively. The proportion of PAS was the highest in the hysteroscopic group (5.2%), followed by the control (1.8%), open (1.4%), and laparoscopic (1.3%) groups. After adjustment, there was no association between PAS and hysteroscopic myomectomy (adjusted odds ratio, 1.86; 95% confidence interval, 0.75–4.63). Uterine rupture after myomectomy was observed only in the laparoscopic surgery group (1.0%); this difference among the four groups was statistically significant (Fisher’s exact test, P = 0.001), although a robust adjusted analysis was not feasible due to the low incidence rate. The proportion of gestational hypertension/preeclampsia was the highest in the hysteroscopic group (17.5%); however, a similar trend was observed as for PAS (adjusted odds ratio, 1.30; 95% confidence interval, 0.74–2.27). The incidences of placenta previa and placental abruption did not differ significantly among the groups. The TIMP was the shortest after hysteroscopic myomectomy, followed by laparoscopic and open myomectomy. Although the number of outcomes was small, which constrained clinical interpretation, there were no perinatal complications associated with TIMP.

Conclusion

Our study suggested the potential risk of uterine rupture after laparoscopic myomectomy. The optimal TIMP remains unclear. The risks of perinatal complications after myomectomy should be assessed and addressed at the individual level based on the specific myomectomy method, but further research on the optimal TIMP is warranted.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12884-025-08617-6.

Background

Myomectomy is a widely used fertility-sparing surgical procedure performed to remove fibroids while preserving the uterus [1, 2]. The aim of this procedure is to alleviate fibroid-related symptoms, such as pain and heavy menstrual bleeding, and improve pregnancy rates in affected women [3, 4].

Despite these benefits, pregnancies after myomectomy are associated with potential perinatal complications. Notably, studies have shown increased risks of uterine rupture and placental abnormalities, such as placenta previa and placenta accreta spectrum (PAS) disorders, in such cases [5–8]. Most studies on perinatal complications have focused on the open or laparoscopic approaches. Hysteroscopic myomectomy is a surgical approach that targets submucosal fibroids. A previous study showed that hysteroscopic myomectomy was associated with a higher incidence of PAS compared with open or laparoscopic procedures [6]. These adverse outcomes are thought to be the result of impaired healing of the myometrial incision, uterine scar formation, or endometrial damage [5, 6]. The integrity of the uterine scar may be influenced by factors including the surgical technique, number and location of fibroids removed, and potentially, the time interval from myomectomy to subsequent pregnancy (TIMP) [9]. A sufficient TIMP is often recommended to allow for adequate uterine healing; however, an excessively prolonged TIMP raises concerns about fibroid recurrence or age-related decline in ovarian function, which could have a deleterious effect on future fertility [10, 11]. Regarding the impact of TIMP on perinatal complications, some studies suggest that a TIMP of < 12 months could increase the risk of uterine rupture [8], whereas others indicate a possible link between a TIMP of < 12 months and an increased risk of PAS [12]. Additionally, several studies have suggested that a longer TIMP may increase the risk of hypertensive disorders of pregnancy [9].

Despite the known risks associated with myomectomy and implicative findings regarding the TIMP, a significant gap exists in our understanding of these dynamics, specifically after hysteroscopic myomectomy. Most existing studies on the TIMP have predominantly focused on the open or laparoscopic approaches. Consequently, data on the relationship between TIMP and the risk of perinatal complications following hysteroscopic myomectomy are scarce.

PAS reportedly occurs more frequently when the TIMP after open and laparoscopic surgery is shorter [12], and the risk of PAS is higher after hysteroscopic myomectomy than after open and laparoscopic myomectomy [6]. Thus, it is crucial to determine whether the TIMP affects this risk similarly after hysteroscopic procedures. Identifying the specific risk factors for perinatal complications after all types of myomectomies and understanding the role of TIMP in each surgical approach are essential for optimizing postoperative management, patient counseling, and pregnancy planning.

In the present study, we aimed to evaluate the risk of perinatal complications in pregnancies conceived after myomectomy, stratified by the surgical approach (laparotomy, laparoscopy, or hysteroscopy). We also investigated whether TIMP is a risk factor for perinatal complications.

Methods

Data source

This retrospective cohort study was conducted using data extracted from the JMDC Health Insurance Claims Database (JMDC Inc., Tokyo, Japan), a comprehensive collection of health insurance claims from multiple health insurers across Japan [13]. This database aggregates health insurance information from Japanese large-sized company employees and their families, enabling tracking of individual patients across various medical institutions. It is an anonymized repository of diagnoses recorded according to the International Statistical Classification of Diseases and Related Health Problems, 10th Revision (ICD-10), and original Japanese codes; medical procedures; drugs prescribed for both inpatient and outpatient visits; and findings of medical examinations. An anonymous family identifier enables data linkage among family members.

Participants

We included women whose first child was born between January 2008 and July 2024. The exclusion criteria were as follows: (i) maternal age < 20 years at delivery; (ii) unavailable data on the child’s year and month of birth; (iii) < 3 years of data available before childbirth; (iv) diagnosis of adenomyosis (ICD-10 code: N800); (v) multiple pregnancy (ICD-10 code: O30); (vi) history of repeated myomectomy using different approaches; (vii) maternal diagnosis of preterm birth (ICD-10: O601, O603) but no corresponding diagnoses for the child (ICD-10: P072 or P073); or (viii) children with both diagnoses of preterm birth before 28 weeks of gestation and preterm birth after 28 weeks of gestation. Within the JMDC database, individuals can be tracked even when they are transferred between hospitals or visit multiple clinics. However, changes in insurance enrollment cannot be tracked. Consequently, considering the patient follow-up rate in the database [14] and the reported TIMP [15, 16], we determined the minimum follow-up period before delivery. Adenomyosis is reportedly associated with infertility and adverse obstetric outcomes [17–20]. Although adenomyomectomy is sometimes performed for symptom management in conditions such as treatment-resistant dysmenorrhea while preserving fertility [21], it is currently considered an advanced medical procedure, which is not covered by universal healthcare in Japan. Consequently, based on the data contained in the insurance claims database, it was unclear whether the procedure was performed. Patients with adenomyosis were excluded to eliminate the possibility that they may have undergone adenomyomectomy. Additionally, we excluded mothers (ix) who were presumed to have undergone myomectomy after the onset of the index pregnancy. Eligible mothers were categorized into four groups: (i) those who did not undergo any myomectomy before the index delivery (the control group), (ii) those who underwent open myomectomy, (iii) those who underwent laparoscopic myomectomy, and (iv) those who underwent hysteroscopic myomectomy. Each myomectomy was defined using the Japanese medical procedure code. Robotic or robot-assisted myomectomy was not included in this study, as it is currently not covered by national health insurance in Japan; therefore, relevant data are unavailable in the claims database.

Subsequently, we estimated the shortest and longest TIMP for the three myomectomy groups as follows. First, to determine the shortest TIMP, we calculated the estimated date of pregnancy onset by subtracting the estimated gestational age at delivery from the estimated delivery date. The gestational age of children with the diagnosis of “preterm birth before gestational age of 28 weeks” (ICD-10 code: P072) was designated as 27 weeks and 6 days, which was the estimated maximum gestational age. The gestational age of children with the diagnosis of “preterm birth after 28 weeks of gestational age” (ICD-10 code: P073) was designated as 36 weeks and 6 days. The gestational age of non-premature children was set as 41 weeks and 6 days. The delivery date was estimated using a previously reported protocol [22]. When data on the eligible delivery date were unavailable, the first day of the child’s birth month and year was designated as the estimated delivery date. Using the estimated date of pregnancy onset, the shortest TIMP was calculated by subtracting the date of myomectomy from the date of pregnancy onset. When the exact date of myomectomy was missing but the year and month were available, the procedure date was imputed as the 15th day of that month. Second, the longest TIMP was calculated similarly using each estimated minimum gestational age and the same delivery date determination protocol. When data on the eligible delivery date were unavailable, the last day of the child’s birth month and year was designated as the estimated delivery date (Fig. 1).

Fig. 1.

Method for estimating the postoperative contraceptive interval. The date of pregnancy onset was estimated by subtracting the estimated gestational age at delivery from the estimated delivery date. For the shortest interval, gestational age was estimated as follows: Preterm birth < 28 weeks: Estimated at 27 weeks and 6 days. Preterm birth ≥ 28 weeks: Estimated at 36 weeks and 6 days. Term birth: Estimated at 41 weeks and 6 days. For the longest interval, gestational age was estimated as follows: Preterm birth < 28 weeks: Estimated at 22 weeks and 0 days. Preterm birth ≥ 28 weeks: Estimated at 28 weeks and 0 days. Term birth: Estimated at 37 weeks and 0 days. The delivery date was estimated per Ishikawa et al.’s (22) protocol. If data on the eligible delivery date were unavailable, it was designated as the first day of the child’s birth month/year for the shortest interval and the last day for the longest interval. The postoperative contraceptive interval was calculated by subtracting the date of myomectomy from the estimated date of pregnancy onset

Outcomes

The diagnosis of PAS was designated as the primary outcome. The secondary outcomes included a diagnosis of placenta previa, uterine rupture, gestational hypertension/preeclampsia, and placental abruption.

Variables

The following variables were identified: maternal age at delivery, female infertility, and obstetrical complications (chronic hypertension, superimposed preeclampsia, threatened preterm delivery, preterm birth before/after 28 weeks of gestation, premature rupture of membranes, preterm premature rupture of membranes, postpartum hemorrhage, diabetes mellitus defined as pre-gestational or overt diabetes in pregnancy, and gestational diabetes mellitus). We also extracted information on the delivery methods (total cesarean section and emergency cesarean section). Previous studies have suggested that intrauterine surgery, such as dilation and curettage or surgical abortion, may increase the risk of PAS due to potential damage to the endometrium and myometrium [23, 24]. Therefore, we extracted the following surgical histories: hysteroscopic lysis of uterine adhesion, endometrial ablation, surgical abortion, dilation and curettage, endometrial curettage, and hysteroscopic polypectomy. All intrauterine surgery histories, except for hysteroscopic myomectomy, were collectively evaluated for the presence or absence of previous intrauterine surgery. Previous cesarean section was not considered an independent variable because the study included only primiparous women. Definitions for all outcomes and variables are provided in Table S1.

Statistical analysis

Continuous variables were presented as means with standard deviations. Categorical variables were presented as frequencies and proportions. Absolute standardized differences were used to assess intergroup differences among the four groups. An absolute standardized difference of < 0.1 indicated well-balanced variable distribution among the groups [25]. We performed six pairwise comparisons and calculated the maximum absolute standardized differences.

First, we compared the occurrence of the perinatal complications among the four groups. We used Fisher’s exact test and multivariable logistic regression analyses. For multivariable analyses, we adjusted the following baseline characteristics: maternal age at delivery, chronic hypertension as the history of hypertension, history of diabetes mellitus, previous intrauterine surgery, and female infertility. For the analysis of PAS, we additionally adjusted for placenta previa. For uterine rupture and placental abruption, we were unable to adjust for these baseline characteristics due to the presence of groups with zero events. Second, we compared the TIMP among the three myomectomy procedures. We selected some perinatal complications that were significant in the initial crude analysis and assessed their association with TIMP using multivariable logistic regression. We adjusted for the same baseline characteristics for the first analysis, but some variables were excluded due to their low prevalence.

All statistical analyses were performed using STATA/SE version 18 (STATA, College Station, TX, USA).

Ethics

This study was approved by the Research Ethics Committee of the Faculty of Medicine at the University of Tokyo [approval number: 10862-(1), approval date: June 13, 2018]. Informed consent was not required as the data were de-identified. This study was performed in accordance with the Declaration of Helsinki.

Results

We identified 963,180 women from the JMDC claims database between January 2008 and July 2024. Based on the eligibility criteria, 27,129 women were included in the final analysis. Of these, 140 underwent open myomectomy, 305 underwent laparoscopic myomectomy, and 97 women received hysteroscopic myomectomy (Fig. 2). The baseline characteristics of the study participants across the four groups are presented in Table 1. The mean age was the highest in the open group (38.4 years), followed by the hysteroscopic (38.2 years), laparoscopic (37.3 years), and control (33.3 years) groups, respectively. The prevalence of preexisting female infertility was higher in the myomectomy groups (open surgery, 84%; laparoscopic surgery, 79%; hysteroscopic surgery, 85%) than in the control group (49%). The incidence of chronic hypertension was the highest in the open surgery group (12%), followed by the hysteroscopic (9.3%), laparoscopic (8.9%), and control (4.9%) groups, respectively. The prevalence of diabetes mellitus was the highest in the open group (1.4%), followed by the laparoscopic (1.0%), the control (0.5%), and the hysteroscopic (0.0%) groups, respectively. The frequency of preterm delivery before 28 weeks of gestation was low in all groups. The frequency of preterm delivery after 28 weeks of gestation was the highest in the open group (5.0%), followed by the hysteroscopic (4.1%), laparoscopic (3.3%), and control (2.9%) groups, respectively. Compared with the control group (14%), a history of intrauterine surgery (except for hysteroscopic myomectomy) was more frequent in the myomectomy groups: 21% in the open group, 20% in the laparoscopic group, and 26% in the hysteroscopic group. The frequency of cesarean delivery was the highest in the open group (90%), followed by the laparoscopic (86%), hysteroscopic (41%), and control (26%) groups, respectively.

Fig. 2.

Flow diagram of patient enrollment. A total of 963,180 women were initially identified. After excluding candidates based on age, data availability, specific diagnoses, and history of myomectomy, 27,129 women were enrolled. Participants were categorized into the control group (n = 26,587) and three myomectomy groups: open (n = 140), laparoscopic (n = 305), and hysteroscopic (n = 97)

Table 1.

Baseline characteristics

	Control group		Open myomectomy group		Laparoscopic myomectomy group		Hysteroscopic myomectomy group		Max ASD
	n = 26,587		n = 140		n = 305		n = 97
Maternal age at delivery, mean (SD)	33.3	4.7	38.4	3.9	37.3	3.8	38.2	4.8	1.18
Female infertility, n (%)	13,098	49	117	84	241	79	82	85	0.81
Maternal morbidity
Chronic hypertension, n (%)	1,301	4.9	17	12	27	8.9	9	9.3	0.26
Superimposed preeclampsia, n (%)	42	0.2	0	0.0	1	0.3	0	0.0	0.08
Threatened preterm birth, n (%)	10,050	38	50	36	114	37	39	40	0.09
Preterm birth before 28 weeks of gestation, n (%)	13	0.0	0	0.0	0	0.0	0	0.0	0.03
Preterm birth after 28 weeks of gestation, n (%)	775	2.9	7	5.0	10	3.3	4	4.1	0.11
Premature rupture of membranes, n (%)	5,267	20	13	9.3	13	4.3	15	16	0.49
Preterm premature rupture of membranes, n (%)	294	1.1	1	0.7	4	1.3	0	0.0	0.16
Postpartum hemorrhage, n (%)	4,822	18	28	20	51	17	18	19	0.08
Diabetes mellitus, n (%)	136	0.5	2	1.4	3	1.0	0	0.0	0.17
Gestational diabetes mellitus, n (%)	2,380	9.0	19	14	42	14	13	13	0.15
History of intrauterine surgery, n (%)	3,675	14	30	21	61	20	25	26	0.3
Total Caesarean section, n (%)	6,817	26	126	90	263	86	40	41	1.72
Emergency caesarean section, n (%)	4,218	16	16	11	31	10	19	20	0.27

SD standard deviation, AMD absolute standardized difference

Table 2 shows the frequency of outcomes. The primary outcome, viz. PAS occurred most frequently in the hysteroscopic group (5.2%, 5/97), followed by the control (1.8%, 487/26,587), open (1.4%, 2/140), and laparoscopic groups (1.3%, 4/305). Placenta previa occurred in 3.0% (809/26,587), 2.9% (4/140), 3.3% (10/305), and 1.0% (1/97) of women in the control, open, laparoscopic, and hysteroscopic groups, respectively. Uterine rupture occurred most frequently in the laparoscopic group (1.0%, 3/305), followed by the control group (0.03%, 9/26,587), but not in the open and hysteroscopic groups; this difference among the four groups was statistically significant (P = 0.001). Gestational hypertension/preeclampsia occurred most frequently in the hysteroscopic group (18%, 17/97), followed by the open (11%, 15/140), laparoscopic (10%, 31/305), and control groups (10%, 2654/26,587). The occurrence of placental abruption was 0.6% (170/26,587), 1.3% (4/305), 1.0% (1/97) in the control, laparoscopic, and hysteroscopic groups, respectively; and no cases in the open group.

Table 2.

Frequency of outcomes

Outcome	Control group		Open myomectomy group		Laparoscopic myomectomy group		Hysteroscopic myomectomy group		P-value
	n = 26,587		n = 140		n = 305		n = 97
Placenta accreta spectrum, n (%)	487	1.8	2	1.4	4	1.3	5	5.2	0.13
Placenta previa, n (%)	809	3.0	4	2.9	10	3.3	1	1.0	0.8
Uterine rupture, n (%) *	9	0.03	0	0.0	3	1.0	0	0.0	0.001
Gestational hypertension and preeclampsia, n (%)	2,654	10	15	11	31	10	17	18	0.12
Placental abruption, n (%)	170	0.6	0	0.0	4	1.3	1	1.0	0.28

^* Statistically significant difference among four groups (Fisher's exact test, p < 0.05)

Table 3 shows the results of the logistic regression analysis. Before adjustment for the baseline characteristics, PAS was more likely to occur in the hysteroscopic group [crude odds ratio (OR), 2.91; 95% confidence interval (CI), 1.18–7.20] than in the control group; however, the significant relationship was attenuated and did not reach statistical significance after adjustment [adjusted OR, 1.86; 95% CI, 0.75–4.63]. A similar trend was observed regarding gestational hypertension/preeclampsia in the hysteroscopic group [crude OR, 1.92; 95% CI, 1.13–3.24; adjusted OR, 1.30; 95% CI, 0.74–2.27, respectively]. The adjusted risks of placenta previa did not differ significantly among the four groups.

Table 3.

Odds ratios for the control group

Outcome	Open myomectomy group n = 140				Laparoscopic myomectomy group n = 305				Hysteroscopic myomectomy group n = 97
	crude OR (95% CI)	P-value	adjusted OR (95% CI)	P-value	crude OR (95% CI)	P-value	adjusted OR (95% CI)	P-value	crude OR (95% CI)	P-value	adjusted OR (95% CI)	P-value
Placenta accreta spectrum	0.78 (0.19–3.15)	0.72	0.51 (0.12–2.06)	0.34	0.71 (0.26–1.92)	0.50	0.50 (0.18–1.35)	0.17	2.91 (1.18–7.20)*	0.021	1.86 (0.75–4.63)	0.18
Placenta previa	0.94 (0.35–2.54)	0.90	0.71 (0.26–1.93)	0.50	1.08 (0.57–2.04)	0.81	0.86 (0.46–1.63)	0.65	0.33 (0.05–2.38)	0.27	0.25 (0.04–1.82)	0.17
Gestational hypertension and preeclampsia	1.08 (0.63–1.85)	0.77	0.62 (0.35–1.09)	0.098	1.02 (0.70–1.48)	0.92	0.70 (0.47–1.03)	0.07	1.92 (1.13–3.24)*	0.015	1.30 (0.74–2.27)	0.36

OR odds ratio, CI confidence interval

^*Statistically significant

Figures S1-S6 depict the median TIMP and its distribution in each myomectomy group. The median (interquartile range, IQR) shortest TIMP was 73.5 (44.5–114) weeks for the open group, 62 (30–118) weeks for the laparoscopic group, and 51 (24–104) weeks for the hysteroscopic group. The TIMP was the shortest after hysteroscopy, followed by laparoscopy and open. Figure S7 shows the distribution of the shortest TIMP among patients with PAS after hysteroscopic myomectomy (median shortest TIMP: 88 weeks, IQR: 66–200); no significant association was observed between the shortest TIMP and PAS in the hysteroscopic group (adjusted OR: 1.01, 95% CI: 1.00–1.02). Figure S8 shows the distribution of the shortest TIMP among patients with uterine rupture after laparoscopic myomectomy (median shortest TIMP: 94 weeks, IQR: 50–140); no significant association was observed between the shortest TIMP and uterine rupture in the laparoscopic group (adjusted OR: 1.00; 95% CI: 0.98–1.02). Figure S9 shows the distribution of the shortest TIMP among patients with gestational hypertension/preeclampsia after hysteroscopic myomectomy (median shortest TIMP: 57 weeks, IQR: 28–108); no significant association was observed between the shortest TIMP and gestational hypertension/preeclampsia in the hysteroscopic group (adjusted OR: 1.00; 95% CI: 0.99–1.01). Consistent results were observed for the longest TIMP and PAS after hysteroscopic myomectomy (adjusted OR: 1.01; 95% CI: 1.00–1.02), the longest TIMP and uterine rupture after laparoscopic myomectomy (adjusted OR: 1.00; 95% CI: 0.98–1.02), and the longest TIMP and gestational hypertension/preeclampsia after hysteroscopic myomectomy (adjusted OR: 1.00; 95% CI: 0.99–1.01).

Discussion

Principal findings

In this study, we evaluated the risk of perinatal complications after open, laparoscopic, and hysteroscopic myomectomies compared with a control group that did not undergo any myomectomy before the index delivery within a single analytic framework. Laparoscopic myomectomy poses a potential risk of uterine rupture in subsequent pregnancies; however, a robust adjusted analysis for uterine rupture was not feasible because no uterine rupture events were observed in the open and hysteroscopic myomectomy groups. Regarding the associations between hysteroscopic myomectomy and both PAS and gestational hypertension/preeclampsia, the significant relationships observed in the crude analysis were attenuated and did not reach statistical significance in the multivariable analysis. A trend emerged toward a shorter TIMP after laparotomy, laparoscopy, and hysteroscopy, in descending order. In addition, PAS after hysteroscopic myomectomy, uterine rupture after laparoscopic myomectomy, and gestational hypertension/preeclampsia after hysteroscopic myomectomy were not associated with TIMP.

Results in the context of what is known

We observed the highest proportion of PAS after hysteroscopic myomectomy; however, there was no association after adjustment, which was inconsistent with the results of previous studies [6, 24]. Similarly, a previous study found that the risk of PAS was higher after open and laparoscopic myomectomies than in the control group [6, 7, 24]; however, our study did not find any such relationship. These may be attributed to the low frequency of PAS in the present study. Hysteroscopic myomectomy can potentially cause more direct damage to the endometrium compared to open or laparoscopic myomectomy, which may lead to abnormal placental implantation and invasion, thereby increasing the risk of PAS in subsequent pregnancies after hysteroscopic myomectomy [5]. This hypothesis is supported by the fact that other surgeries that can directly damage the endometrium have been shown to increase the risk of PAS [24, 26, 27]. While fibroid size, depth, number, and cavity involvement may influence the extent of direct endometrial damage and the subsequent risk of PAS, this information is unavailable in the database utilized in this study. Taking these points into account, this result might reflect limited statistical power rather than a true lack of association.

The higher risk of uterine rupture observed in the laparoscopic group warrants careful interpretation since statistical adjustment was not feasible due to the low frequency of events in the other groups. Uterine rupture is a rare perinatal complication; however, several studies have reported uterine rupture in pregnancies after laparoscopic myomectomy [28–31]. Laparoscopic and open myomectomy are associated with a higher risk of uterine rupture compared with a negative history of prior surgery [7]. Furthermore, the risk was even higher after laparoscopic myomectomy than that after open myomectomy [32, 33]. Scar tissue or inadequate scar healing after previous myomectomy is thought to be a risk factor for uterine rupture [5, 16]. Additionally, it may be influenced by various factors such as the size, number, and location of myomas; use of devices; operator’s surgical experience; and suture technique [29, 34]. Especially after laparoscopic myomectomy, variations in scar thickness depending on the suture technique, use of electrosurgery, or operator experience might contribute to a higher incidence of uterine rupture than that after open myomectomy [31–33]. The risk of uterine rupture following laparoscopic myomectomy may be modulated by the surgical technique employed. The paucity of studies on uterine rupture after hysteroscopic myomectomy [35, 36] suggests that it is a rare complication, which may account for the absence of such cases in the present study.

Preserving the pseudocapsule, the specialized myometrium surrounding and nourishing the myoma, during myomectomy is considered important for myometrial healing, thereby reducing the risk of myometrial fibrosis or uterine rupture [37]. Endoscopic approaches, such as hysteroscopy and laparoscopy, facilitate the approach by magnifying the pseudocapsule; however, these techniques require surgical experience and expertise [37]. Meticulous execution of this procedure may reduce these complications, and detailed surgical documentation regarding pseudocapsule management would be valuable for predicting the risk of uterine rupture.

The highest proportion of gestational hypertension/preeclampsia was observed after hysteroscopic myomectomy, although there was no significant association after adjustment. Some reports show the association between hypertensive disorder of pregnancy and myoma itself [38, 39]. Although data regarding the association between myomectomy and HDP are scarce, one study investigated this issue [40] but did not include hysteroscopic myomectomy.

We did not observe a significant difference in the incidence of placenta previa and placental abruption among the groups. Regarding placenta previa, previous studies showed a significant association [6, 8], whereas for placental abruption, most studies have found no such association [6, 8]. In the present study, however, the low number of events raises concern about limited statistical power; further research is warranted on these complications.

Submucosal myomas, the primary indication for hysteroscopic myomectomy, have been reported to negatively impact fertility [41]. Their removal could be considered to improve pregnancy rates [41, 42], or alleviate other symptoms, including heavy menstrual bleeding and anemia [43]. Although other minimally invasive approaches, such as radiofrequency or laser ablation, are available for submucosal myomas, their safety for women desiring future conception has not been established [44, 45]. Therefore, hysteroscopic myomectomy is currently considered the recommended treatment for symptomatic submucosal myomas in women who wish to conceive.

Open or laparoscopic myomectomy may be a necessary intervention for patients with symptomatic subserosal or intramural myomas who desire fertility preservation, particularly when symptoms such as abnormal uterine bleeding or pelvic pain are refractory to medical therapy [46]. Whereas there is insufficient evidence to conclude that removing non-cavity-distorting fibroids improves pregnancy or live birth rates [47, 48]. The choice between open myomectomy and laparoscopic myomectomy depends on several factors, including the size, number, and location of the myomas, as well as the surgeon’s technical expertise [46, 49]. These variables are critical as they determine operative feasibility and intraoperative risks [49].

In the study, we excluded women with adenomyosis because it is unclear from the database whether these patients had received an adenomyomectomy. Previous reports suggest that pregnant women after adenomyomectomy had an increased risk of PAS or uterine rupture [50–53], while the risk of preterm premature rupture of membranes, preeclampsia, and small for gestational age may decrease [50]. Consequently, this exclusion was deemed necessary to appropriately account for this potential confounding factor.

Only a few studies exist on the association between TIMP and perinatal complications, and their findings vary. A shorter TIMP may be associated with the occurrence of PAS [12] or uterine rupture [8], whereas a longer TIMP may be associated with gestational hypertension/preeclampsia [9]. We described the distribution of this interval among different myomectomy approaches and assessed the association of TIMP with some perinatal complications. We observed a trend where TIMP was shortest after hysteroscopic myomectomy, followed by laparoscopic and open myomectomy. Furthermore, we found no significant association between TIMP and PAS or gestational hypertension/preeclampsia following hysteroscopic myomectomy, nor between TIMP and uterine rupture following laparoscopic myomectomy. However, we could not draw concrete clinical conclusions because of the small number of outcomes. The 2023 edition of the Japanese Guidelines for Gynecological Practice avers the difficulty in determining the recommended contraceptive periods. Although Japanese surgeons generally recommend contraceptive periods of approximately 3–6 months, depending on the surgical procedure, the recommendation is not backed by conclusive evidence. Based on findings that uterine healing would be complete after three months [29], there is one report recommending a minimum contraception period of three months [54]. The TIMP may be longer than the actual contraceptive period recommended by surgeons because conception does not immediately follow the cessation of contraception. At present, we cannot suggest an optimal TIMP and must determine it based on other factors, such as the surgical procedure of each patient or patient age, as in the past.

Clinical and research implications

Considering that uterine ruptures are potentially life-threatening complications, their rarity does not diminish the importance of clinical vigilance [55, 56]. Predicting the occurrence of uterine rupture allows for proactive planning, such as optimizing the timing of delivery and ensuring birth occurs in a facility equipped for potential complications such as massive transfusion or emergency surgery.

Our findings suggest that individualized risk assessment based on the myomectomy type may be helpful in guiding preoperative counseling, preconception counseling, and antenatal management. During the preoperative period, it is vital for gynecologists and patients experiencing myoma symptoms to engage in shared decision-making regarding the true necessity of myomectomy, especially when future pregnancies are planned. During the preconception period, obstetricians and gynecologists should counsel patients about the risk of complications during pregnancy associated with prior surgery. During the antenatal period, obstetricians should diligently inquire about any history of myomectomy during routine prenatal visits. When a history of myomectomy is present, obstetricians should meticulously examine myometrial thinning to predict potential uterine rupture. Although further evidence is needed, careful obstetric follow-up of pregnancies after myomectomy may improve maternal and neonatal outcomes.

Strengths and limitations

A major strength of this study was its large-scale data, which enabled a comprehensive comparison of four groups (control, open, laparoscopic, and hysteroscopic myomectomy) and the integration of TIMP analysis across these surgical procedures within a single analytic framework. While the methodology for estimating the date of delivery was adopted from previously established literature [22], the subsequent estimation of TIMP using ICD codes and procedure dates represents a novel approach, particularly given the inherent lack of precise gestational age information in the Japanese claims database. However, some limitations of this study should be acknowledged. First, it is possible that not all postoperative patients were included in the post-myomectomy groups due to the characteristics of the database [14]. Additionally, the sample size might be insufficient to evaluate the association of each surgical procedure and TIMP with perinatal complications. The high proportion of exclusion due to missing data on the children's birth month/year could introduce selection bias and is likely attributable to the Japanese insurance system. In Japan, maternity leave typically begins at 34 weeks of gestation. At this time, some women withdraw from their insurance and join their spouse's dependent coverage, often due to reasons such as resignation or discontinuing social insurance premium payments during maternity leave, though the system regarding these payments has changed recently. This could result in the loss of continuous data up to the child's birth month/year. Second, we could not perform a detailed assessment of the myomectomy procedure, such as the size, number, and location of myomas; suture material; method of uterine closure; and the operator’s surgical experience. Third, we acknowledge that the true TIMP may deviate by several weeks because the gestational age estimated by ICD codes inherently has a few-week interval. To address this limitation, we assessed both the shortest and longest TIMP scenarios and observed consistent trends across this range. Fourth, we summarized the history of previous intrauterine surgeries as a single variable because some procedures were used infrequently. However, the risk of each perinatal complication may differ depending on the type of intrauterine procedure [24]. Fifth, there have been no validation studies on myomectomy or perinatal complications in Japan. Therefore, the possibility of misclassification remains. Sixth, we could not include robotic or robot–assisted myomectomy because these procedures are not covered by national health insurance in Japan. Therefore, although the number was likely small, it was possible that women who received a robotic or robot–assisted myomectomy were included in the control group. Finally, the JMDC database primarily comprises data from employees and their dependents enrolled in health insurance societies of large companies across Japan. This population, which does not include employees of small- and medium-sized enterprises, self-employed individuals, and government employees, can be considered a large corporate-based cohort and may not fully represent the entire female Japanese population due to potential differences in socioeconomic status, regional distribution, and access to healthcare. On the other hand, studies that linked maternal and infant data using the JMDC database reported that maternal age and infant sex were similar to those in the general Japanese population, suggesting that the database is suitable for these studies [57]. Multicenter prospective studies are needed to overcome these limitations.

Conclusions

A key strength of this study is the large-scale evaluation of perinatal complications, comparing three myomectomy procedures against a control group, and assessment of TIMP across these surgical procedures within a single analytic framework. Our study showed that laparoscopic myomectomy may be associated with the potential risk of uterine rupture in subsequent pregnancies. Although the optimal TIMP remains unclear, these findings collectively underscore the critical importance of individualized risk assessment based on the specific myomectomy approach, comprehensive preconception counseling for women with a history of myomectomy, and careful management throughout subsequent pregnancies.

Abbreviations

CI

Confidence interval

ICD-10

International Statistical Classification of Diseases and Related Health Problems, 10th Revision

IQR

Interquartile range

OR

Odds ratio

PAS

Placenta accreta spectrum

TIMP

Time interval from myomectomy to subsequent pregnancy

Authors’ contributions

MF: Conceptualization, Data curation, Formal analysis, Methodology, Visualization, Writing – original draft, Writing – review and editing. OWH: Conceptualization, Data curation, Supervision, Writing – review and editing. MN, RT, HG, YE, MH and YH: Conceptualization, Writing – review and editing. YS: Conceptualization, Data curation, Formal analysis, Methodology, Supervision, Writing – review and editing. HY: Conceptualization, Data curation, Methodology, Supervision, Writing – review and editing.

Funding

This research was funded by the Ministry of Health, Labor and Welfare, Japan (grant number 24FB1001: O.W–H. , and grant number 23AA2003 H. Y.).

Data availability

The data that support the findings of this study are available from JMDC Inc. but access is subject to a license agreement. The data were used for this study under such an agreement. Anyone wishing to request access to these datasets can contact JMDC Inc. at https://www.jmdc.co.jp/en/inquiry/.

Declarations

Ethics approval and consent to participate

All procedures performed in this study conformed with the ethical standards of the institutional and/or national research committee and the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. Informed consent was obtained from all the participants enrolled in the study.

This study was approved by the Research Ethics Committee of the Faculty of Medicine of the University of Tokyo (University of Tokyo IRB number: 3128–6).

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.