Pregnancies after vaginal radical trachelectomy (RT) in patients with early invasive uterine cervical cancer: results from a single institute

Radical trachelectomy (RT) with pelvic lymphadenectomy, originally reported by Dargent et al. in 1988 [1], has become a new treatment option for young patients in Japan with early invasive uterine cervical cancer who desire preservation of their fertility [2‐4]. This operative modality is recently performed even during pregnancy [5]. Recent data indicate that the oncologic results following RT are comparable to the results following standard radical hysterectomy (RH) [6‐8].

Almost 15 years has passed since we started vaginal RT in our institute, and we have performed 71 vaginal RTs so far. Among those who underwent this operation, there have been 28 pregnancies in 21 patients. However, as was reported in various studies, with pregnancy after RT there still remains a high risk of preterm labor and subsequent preterm birth. This is because of the lack of mechanical support of the residual cervix and ascending infection and chorioamnionitis (CAM) caused by disruption of the endocervical glands and reduced secretion of mucus after the radical operation [9, 10]. Although the optimal obstetrical management of these patients remains controversial, the follow-up modality for pregnant patients after vaginal RT in our department has been improved year by year. However, they still usually need long-term obstetrical treatments over the course of the pregnancy.

We believe it is important to examine this technique not only from the oncological viewpoint but also from the obstetrical one to improve both the oncological and the obstetrical prognoses of patients with this disease. In this report, we summarize the clinical profile of pregnancy after vaginal RT in our institute and review the obstetrical management of pregnancy after vaginal RT. This is the first cumulative study of pregnancy after vaginal RT in Japan.

Twenty-eight pregnancies in 21 patients after vaginal RT with pelvic lymphadenectomy in our institute between 2007 and 2018 are reviewed in this study. The study protocol was approved by the review board and the ethics committee in our institute (approval number 05–26). A written informed consent was also obtained from each patient.

The clinical characteristics of the patients are summarized in Table 1 and the clinical indications for vaginal RT in our hospital are summarized in Table 2. All of the patients except one had International Federation of Gynecology and Obstetrics (FIGO) 2009 stage 1A2-1B1 uterine cervical cancer. Clinical stage 1B2 and higher stages are beyond the indication for RT in our hospital, although RT was performed for a patient with stage IIA who strongly desired the preservation of fertility after repeated explanations to her and her family. Adjuvant chemotherapy using paclitaxel and carboplatin was also used for her during pregnancy and after cesarean section. Assessment of pelvic lymph node metastasis was performed by laparotomy or laparoscopy at the time of vaginal RT. For pregnant patients, it was performed by magnetic resonance imaging (MRI) before the operation and the lymph node sampling at the time of the operation.

Vaginal RT itself was performed based on a modification of the procedure of Dargent et al. Amputation height was preoperatively determined by transvaginal ultrasonography and careful search of lesion area by colposcopy for each patient.

The procedure started with lymphadenectomy by laparotomy or laparoscopy, and then vaginal RT was performed. Briefly, a rim of vaginal mucosa was delineated circumferentially and excised so that the anterior and posterior mucosae could cover the cervix. The vesicovaginal space was defined laterally on each side. Then the descending branches of the uterine arteries and the cardinal ligaments were cut at the level of Type II hysterectomy. After this procedure, we usually amputated the uterine cervix 10 mm below the internal os of the uterus. A nylon suture was placed around the residual cervix to support cervical strength, and Strumdorf sutures were placed to cover the surface of the cervix. Follow-up of the cancer was performed by periodical checking of cytology, tumor markers, and CT scans. Tumor marker examination and CT scans were not usually performed for pregnant patients.

Vaginal RT for pregnant patients was also conducted according to the same procedure. However, as the patients were pregnant, we confirmed the fetal status by ultrasonography soon after the operation, and careful check of the amount of the bleeding from the vaginal wound and the existence of leakage of amniotic fluid were performed. Additionally, prophylactic continuous intravenous administration of tocolytic agents was performed at least for 1 week after RT.

We performed transabdominal cerclage (TAC) for four patients who experienced second trimester fetal losses due to troubles with the transvaginal cerclage placed at the time of vaginal RT. Details of the operative procedure are described in our previous report [11].

As for the management of pregnant patients after vaginal RT, there are no international guidelines so far. Prevention of cervical and uterine infection, and prevention of preterm uterine contraction are a key to avoid pPROM and the following preterm birth. We recommend that patients enter hospitalization early in the second trimester of the pregnancy even if there are no signs of threatened abortion. Daily vaginal disinfection with 0.3% chlorhexidine, the administration of an ulinastatin suppository, and weekly examination of vaginal flora, the granulocyte elastase level and oncofetal fibronectin in vaginal secretion, the measurement of uterine cervical length by transvaginal ultrasonography, and weekly intramuscular administration of a hydroxyprogesterone caproate depot were performed for the patients under moderate bed rest. Serum CRP and general blood examinations were also done weekly. In cases with a Nugent score > 5, and those that were cervical elastase or fibronectin positive, transvaginal administration of metronidazole was performed. Tocolysis started when the patients felt abdominal pain/tension, or the uterine cervical length shortened to < 2 cm.

Ritodrine hydrochloride and/or magnesium sulfate were intravenously administered as tocolytic agents. Nifedipine was also orally administered. Considering the maternal physical and psychological conditions from the long-term bed rest and tocolysis, and to avoid the sudden occurrence of premature preterm rupture of membrane (pPROM), cesarean section was usually scheduled at 34–35 weeks of pregnancy. For one patient who had recurrence during pregnancy, standard total abdominal RH was performed at the time of cesarean section.

Measurement of uterine cervical length was performed weekly for each patient from 14 weeks of pregnancy until the end of pregnancy. Changes of the cervical length during pregnancy were expressed as “average ± SE” using Microsoft EXCEL 2016 in Fig. 1.

We performed 71 vaginal RTs for patients with early invasive uterine cervical cancer from 2006 to 2018 at Sapporo Medical University. Six of the operations were performed during the period from 16 to 26 weeks of pregnancy (Nos. 8, 12, 15, 16, 17, and 18). Specimens of all the pregnant patients had cancer free margin > 20 mm.

The clinical characteristics of the patients are summarized in Table 1. In total, we experienced 28 deliveries by 21 patients, with 24 live births, including a case of twin pregnancy (patient 14). The overall pregnancy rate among the patients was 29.6%. The median time to each pregnancy after RT was 29.5 months (range 6–123 months). Thirteen pregnancies (46.4%) did not require artificial insemination or assisted reproductive technology. The median time of pregnancy, excluding first trimester abortion, was 34 weeks (range 18–35 weeks).

This study demonstrated both the safety and the difficulties of the obstetrical management of pregnant patients after vaginal RT in our institute. For the approach to the uterine cervix, abdominal RT is performed more frequently than vaginal RT in Japan. Many of the reports thus far in Japan have been those using the abdominal approach [3, 12]. This is the first cumulative report regarding pregnancies after vaginal RT in Japan.

Of the 21 patients, who underwent vaginal RT during pregnancy, one (4.8%) had a recurrence at 24 weeks of gestation and underwent cesarean section with radical hysterectomy. She is currently free of disease. No other patient has had recurrence to date after pregnancy. Our recurrence rate in pregnant women corresponded to the reports of 4.8% after vaginal RT by Plante et al. [9], and 3.7% after abdominal RT by Li et al. [13]. As we reported recently, expansion of vaginal RT even for pregnant patients is a practical option now [5]. We believe that operative procedures of vaginal RT for pregnant patients to remove the uterine cervix are more conservative and simpler than abdominal RT for them because vaginal RT makes it possible to cut the uterine cervix and the parametrium without laparotomy. In addition, the pregnancy rate without artificial insemination or assisted reproductive technology was very high in our study. The conservative approach to the uterus of vaginal RT seems to have made possible this better pregnancy rate.

Most patients who became pregnant after vaginal RT in this study could continue their pregnancy until the time of scheduled cesarean section. As is well known, pregnancy after RT has higher risks of preterm birth. Although there are no international guidelines for the following-up of pregnant patients after RT, the basics of the maternal management for patients after RT are bed rest and the prevention of uterine cervical infection. The necessity of bed rest under hospitalization and its effectiveness are, however, still controversial [14]. Bed rest for a long time is not only a burden for patients but can also be a cause of deep vein thrombosis (DVT). However, we still believe that bed rest is essential to prevent emergencies in patients after RT.

Prevention of infection is another key to improve the obstetrical prognosis.

It is said that 25–60% of preterm births are attributable to maternal infections, and bacterial vaginosis (BV) has been suggested to be a risk factor for preterm birth [15, 16]. BV is a condition in which the normal, lactobacillus-predominant vaginal flora is replaced with anaerobic bacteria such as Gardonerella vaginalis and Mycoplasma hominis [17]. BV has also been reported to be associated with pPROM and CAM [18]. Therefore, the prevention and the treatment of BV during pregnancy might have an impact to improve obstetrical prognosis in patients after RT. We as well as Speiser et al. administered Metronidazole vaginal tablets for the treatment of BV during pregnancy [19]. Ito et al. reported the usefulness of oral probiotics to reduce the occurrence of BV and the following pPROM for a patient after RT [20].

We usually cut the uterine cervix about 10 mm below the internal os of the uterus. Such a large cut of the uterine cervix causes disruption of the endocervical glands, reduces secretion of mucus, and promotes cascades from BV to CAM. Furthermore, it can be a cause of reduced mechanical support of the residual cervix. Therefore, the transvaginal cerclage of the residual cervix, which is usually placed at the time of radical RT, plays an important role in maintaining pregnancy [10]. Troubles with this cerclage can become a cause of second trimester fetal losses and early preterm birth. Patients with a slack cerclage usually suffered from sudden pPROM without elevation of infectious markers such as BV, cervical elastase, and fibronectin. Shepherd et al. also reported the same types of pPROM in patients after RT [21]. How does it occur? The uterine cavity has been believed to be in a sterile condition. However, recent analysis using next-generation sequencing of the 16 s ribosomal RNA gene revealed the existence of various bacteria there [22]. Kyono et al. reported that a considerable percentage of the non-Lactobacillus-dominated microbiota, which is impossible to detect by routine bacterial culture, affects the occurrence of infertility and preterm abortion [23]. We believe such subclinical intrauterine infections, which are undetectable by routine bacterial culture examination, might exist in patients with a slack cerclage after RT.

As for the treatment of patients with preterm uterine contraction, there are differences between Japan and other western countries. We still use long term administration (> 72 h) of ritodrine hydrochloride and/or magnesium sulfate for them, but most western countries do not undergo long term tocolysis for them. Tocolytic agents are usually used for at most 72 h. Instead, progesterone gels or pessaries will be used for preterm uterine contraction as described in a review by Tirlapur et al. [24]. Ritodrine hydrochloride has not been used in western countries except Japan due to its complications anymore. However we believe, if it is used carefully and with lower doses (50-200 μg/min), it might be effective to improve obstetrical prognosis of patients after RT without serious complications.

What can gynecologic oncologists do to reduce such obstetrical troubles? Transabdominal uterine cervical cerclage (TAC), originally proposed by Benson and Durfee in 1965, actually seems to improve the obstetrical prognosis for patients who have troubles with the cerclage [11].

Large amputation of the uterine cervix can be a cause of preterm birth. Therefore, more conservative and less invasive operations for young patients are in great demand all over the world. However, there are only a few reports on less invasive operations for patients with early invasive uterine cervical cancer so far. Schmeler et al. reported that parametrium invasion occurred in less than 1% of patients with 1A2-1B1 with a < 2 cm tumor without lymphovascular invasion [25]. And our study revealed that a simple trachelectomy 20 mm not 10 mm below the internal os with pelvic lymphadenectomy might be possible for stage 1A2 patients [26]. These results might indicate the possibility of more conservative trachelectomy such as simple trachelectomy with a shorter amputation of the uterine cervix which lead to the improvement of the obstetrical prognosis for stage 1A2 patients [26, 27].

The cooperation of gynecologic oncologists and obstetricians thus has a great impact on improving the obstetrical prognosis for patients after RT. In addition, it is important to consolidate these high risk pregnant patients in large medical facilities that can provide comprehensive care.

We have performed 71 vaginal RTs for patients with early invasive uterine cervical cancer, and have experienced 28 pregnancies in 21 patients so far. Both the obstetrical prognosis and oncological prognosis after vaginal RT have become favorable for pregnant patients after vaginal RT under the appropriate management.