Bleeding cervical varices in pregnancy: A case report and review of the literature

1 Introduction

The development of cervical varices during pregnancy is rare, with only 20 reported cases in the literature [1–17]. If left untreated, they can cause significant maternal and fetal morbidity. Due to the paucity of data, there is currently no consensus on the diagnosis or management of bleeding cervical varices during pregnancy.

2 Case

A 50-year-old G7 P1051 was transported to our institution via ambulance from an outside hospital at 12 4/7 weeks gestation due to four episodes of vaginal bleeding in the prior four days requiring a transfusion of one unit of blood. She had been followed closely during her transition from reproductive endocrinology and infertility (REI) to her general obstetrician. Her history was significant for one full-term uncomplicated vaginal delivery 21 years ago, then five miscarriages, two of which required a dilation and curettage (D&C), and a hysteroscopy in preparation for in-vitro fertilization (IVF). Pregnancy was achieved by IVF with an egg from a 28-year-old donor.

Ultrasound examination demonstrated a viable intrauterine pregnancy with a crown-rump length consistent with clinical dating. The placenta was located on the posterior uterine wall, and there was a marginal placenta previa with the leading edge covering the internal cervical os. The edge of the placenta appeared to be irregular and lifted up from the myometrium. The endocervical canal appeared to be dilated to 1 cm from the internal to the external cervical os. There was robust maternal venous blood flow by color Doppler waveforms throughout the endocervical canal (Fig. 1). Given the marginal placental separation and threatened miscarriage, the patient remained in the hospital for observation. She was informed that if she were to bleed heavily enough that it threatened her safety, the recommendation was for uterine evacuation via D&C with a possible hysterectomy if the bleeding was unable to be surgically controlled. Given her age, there was additional concern for morbidity due to focal placental adherence.

OPEN IN VIEWER

Fig. 1

Endo-vaginal ultrasonography at the time of diagnosis. (A) Cervix is dilated to external os. (B) Color Doppler signal shows venous vessel enhancement within the endocervical canal.

The following day, she experienced significant hemorrhage with a total estimated blood loss of 3.1 liters requiring a total of five units of packed red blood cells (PRBCs) and one unit of fresh frozen plasma (FFP). She was taken to the operating room for a D&C vs. hysterectomy for maternal hemorrhage. In the operating room, she was found to have active brisk bleeding from the endocervix and was dilated from cervical varicosities within the endocervical canal. The external appearance of the cervix was normal. The walls of the cervix were grasped with ring forceps, and the bleeding from the area of the varicosities was identified and contained with compression from the ring forceps. Ligation of the varicosities was achieved with three interrupted transmural sutures at 12, 8, and 4 o’clock. The bleeding stopped immediately and the fetus still demonstrated strong cardiac activity with a normal amniotic fluid level. Thus, the decision was made to forego the D&C and pursue expectant management. Ultrasound following the procedure showed a cervical canal length of 2.6 cm with no color venous Doppler signal in the endocervix. There was no intra-uterine clot.

On day three of hospitalization, ultrasound examination demonstrated increased color venous Doppler signal in the endocervical canal, suggesting some heightened cervical vascularity compared to the prior day. At this point, the decision was made to place cerclage sutures to tamponade these potential areas of bleeding. This is analogous to the use of such sutures following pregnant cervical conization procedures. An uncomplicated McDonald cerclage placement was performed, and there was no progressive bleeding from the endocervix or elsewhere. Follow up ultrasound demonstrated color venous Doppler signal in a few vessels around the cervix, but this was significantly diminished compared to the exam on admission. Her cervix was otherwise unremarkable (Fig. 2). She was discharged home doing well on hospital day seven. Her REI physician disclosed that her cervix had been normal and without obvious varicosities at the time of the patient’s IVF procedure several months prior.

OPEN IN VIEWER

Fig. 2

Endo-vaginal ultrasonography after cerclage placement.

She was followed by Maternal-Fetal Medicine for the remainder of her pregnancy, which was relatively uncomplicated. Her low-lying placenta resolved by 26 1/7 weeks, although she did have a residual vasa previa. Follow-up ultrasound examinations with color Doppler venous waveforms at 2 to 4-week intervals throughout the rest of the pregnancy noted the persistence of cervical varicosities, although they were stable and regressed over time (Fig. 3). Given her history, the decision was made to have an elective cesarean delivery at 37 weeks. She delivered a viable female infant weighing 3200 grams at 37 2/7 weeks with APGARs of 8 and 8 at 1 and 5 minutes, respectively. Estimated blood loss at delivery was 800 milliliters. The placenta, which was mostly left lateral and posterior, was removed intact and was noted to be bilobed with the cord insertion in the middle (Fig. 4). The patient did well post-operatively and was discharged on hospital day three. Her cerclage sutures were removed without complication at her two-week postpartum visit with no apparent cervical abnormalities or varicosities.

OPEN IN VIEWER

Fig. 3

Endo-vaginal ultrasonography on follow-up visit approximately 6 weeks later. Color Doppler signal shows stable varicosities. Cervical length 2.50 cm.

OPEN IN VIEWER

Fig. 4

Placenta after delivery. (A) Fetal surface bilobed with cord insertion in the middle. (B) Maternal surface bilobed and intact.

Venous varicosities, most commonly located in the legs, vulva, and hemorrhoidal plexus, are a relatively common finding in pregnancy [18, 19] and are thought to be due to hormonal changes, increased blood volume, venous insufficiency, and compression of the inferior vena cava by the enlarging uterus [5, 18]. Cervical varices represent a rare but life-threatening form of venous varicosities. To date, there are only 20 other reported cases in the literature of cervical varices affecting pregnancy (Table 1). Of these cases, most present with vaginal bleeding in the second or third trimester. The case discussed here is unique as it represents one of the earliest gestational ages presenting with profuse vaginal bleeding.

The exact mechanism of the development of cervical varices is incompletely understood, although there appears to be an association with low placental implantation [2, 13–17], maternal age > 35 [1, 13–16], IVF [2, 16], multiple gestations [9, 17], and cervical vascular malformation secondary to in utero diethylstilbestrol (DES) exposure [3, 4]. Of the cases reported in the literature, ten occurred in primigravid women [1, 17]. In our case, the patient was > 35 years old, conceived by IVF, and had a known marginal placenta previa. It is postulated that a low-lying placenta increases drainage of the placental bed into the cervical venous plexus, leading to dilation of the cervical venous system, and eventually to the development of cervical varicosities [7, 16].

According to Tanaka et al. [15], there appear to be two main types of cervical varicosities: external os-type and endocervical, and they present differently in terms of imaging, timing of detection, and optimal treatment for hemorrhage. Speculum exam is more likely to confirm the presence of external os-type varices. Moreover, the exact bleeding site is more likely to be visualized in external os-type varices, thus facilitating the differentiation between bleeding varices and a bleeding placenta previa. Endo-vaginal ultrasound is able to identify enlarged cervical vessels in both subtypes, but is especially useful for endocervical varices. It is important to utilize color Doppler in these scenarios, as varicosities can be missed otherwise. In our case, the bleeding vessels originated from the endocervix, as seen by ultrasonography (Fig. 1), although cervical dilation at the time of presentation facilitated visualization of the varicosities during speculum exam. It was not until the bleeding sites were identified and ligated in the operating room that a bleeding placenta previa was excluded.

Although uncommon, it is important to consider bleeding cervical varicosities as a source of maternal hemorrhage, as timely treatment can be lifesaving to the fetus and mother. Vaginal packing, transfusion of blood and blood products, suture ligation, emergent cerclage placement, tocolysis, and emergency cesarean section have all been reported to control the acute episode of vaginal bleeding [1, 15–17]. Moreover, physical and pelvic rest have been recommended in a large number of cases [2–5, 17]. In one report, pregnancy had to be terminated in order to control life-threatening maternal hemorrhage [9] and in a different case, a cesarean-hysterectomy had to be performed [5]. In our case, ligation of the varicosities with interrupted sutures achieved hemostasis, and a McDonald cerclage was used to tamponade and prevent any further bleeding. Others also have had success with cerclage placement to prevent cervical varices from prolapsing downward and rupturing [2, 4]. Given the risk for severe intrapartum hemorrhage secondary to ruptured cervical varices, it is recommended to proceed with a cesarean delivery. Only one reported case was delivered vaginally without complications [4]. Timing of delivery should be individualized based on maternal and fetal factors. Interestingly, most cases of cervical varices regress postpartum [15].

This case demonstrates an example of massive maternal hemorrhage from bleeding cervical varices diagnosed in the first trimester. Although the bleeding was thought to be due to placenta previa or marginal placental separation, diagnosis following speculum exam was clear. By achieving hemostasis with interrupted suture ligation and preventing further bleeding with a McDonald cerclage tamponade in place, the patient was able to continue pregnant until her planned cesarean delivery at term. This case highlights the importance of including bleeding cervical varices on the differential diagnosis of maternal hemorrhage, especially if the patient has a known low placental implantation. If cervical varicosities are discovered during pregnancy, suture ligation with subsequent cerclage placement is a potential management option to prevent catastrophic bleeding and prolong pregnancy.