First-trimester diagnosis and management of pregnancies implanted into the lower uterine segment Cesarean section scar

Authors

  • D. Jurkovic,

    Corresponding author
    1. Early Pregnancy and Gynaecology Assessment Unit, Department of Obstetrics and Gynaecology, King's College Hospital, London, UK
    • Early Pregnancy and Gynaecology Assessment Unit, 6th Floor, Ruskin Wing, King's College Hospital, Denmark Hill, London SE5 8RX, UK
    Search for more papers by this author
  • K. Hillaby,

    1. Early Pregnancy and Gynaecology Assessment Unit, Department of Obstetrics and Gynaecology, King's College Hospital, London, UK
    Search for more papers by this author
  • B. Woelfer,

    1. Early Pregnancy and Gynaecology Assessment Unit, Department of Obstetrics and Gynaecology, King's College Hospital, London, UK
    Search for more papers by this author
  • A. Lawrence,

    1. Early Pregnancy and Gynaecology Assessment Unit, Department of Obstetrics and Gynaecology, King's College Hospital, London, UK
    Search for more papers by this author
  • R. Salim,

    1. Early Pregnancy and Gynaecology Assessment Unit, Department of Obstetrics and Gynaecology, King's College Hospital, London, UK
    Search for more papers by this author
  • C. J. Elson

    1. Early Pregnancy and Gynaecology Assessment Unit, Department of Obstetrics and Gynaecology, King's College Hospital, London, UK
    Search for more papers by this author

Abstract

Objective

To describe first-trimester ultrasound diagnosis and management of pregnancies implanted into uterine Cesarean section scars.

Methods

All women referred for an ultrasound scan because of suspected early pregnancy complications were screened for pregnancies implanted into a previous Cesarean section scar. The management of Cesarean section scar pregnancies included transvaginal surgical evacuation, medical treatment with local injection of 25 mg methotrexate into the exocelomic cavity and expectant management.

Results

Eighteen Cesarean section scar pregnancies were diagnosed in a 4-year period. The prevalence in the local population was 1 : 1800 pregnancies. Surgical treatment was used in eight women and it was successful in all cases. The respective success rates of medical treatment and expectant management were 5/7 (71%) and 1/3 (33%). Five women (28%) required blood transfusion and one woman (6%) had a hysterectomy.

Conclusions

Cesarean section scar pregnancies are more common than previously thought. When the diagnosis is made in the first trimester the prognosis is good and the risk of hysterectomy is relatively low. Copyright © 2003 ISUOG. Published by John Wiley & Sons, Ltd.

Introduction

Implantation of the placenta into the uterine scar following a previous Cesarean section is associated with severe late pregnancy complications such as placenta previa and placenta accreta1. In women with a history of multiple Cesarean sections the incidence of placenta previa is ten times higher compared with women who had vaginal deliveries2. When the placenta is implanted over the uterine scar, it is abnormally adherent in 30–40% of cases, which often causes uncontrollable hemorrhaging at delivery2, 3. Abnormally inserted placentae are responsible for 50–65% of all obstetric hysterectomies, 66% of which have a history of previous Cesarean sections4, 5. The clinical significance of this complication is emphasized by the Confidential Enquiries into Maternal Deaths, which showed that 72% of all maternal deaths associated with placenta previa in the UK between 1991 and 1999 occurred in women who had at least one previous Cesarean section6.

Placenta accreta can be diagnosed by ultrasound in the third trimester of pregnancy. The diagnostic criteria include an absent decidual interface between the placenta and myometrium as well as unusual dilatation of the vessels under the placental implantation site7. Although the reported accuracy of ultrasound diagnosis in the third trimester is high8, the late detection of placenta accreta is of limited value, as it does not help to prevent serious maternal morbidity associated with this condition.

There are only a handful of reports describing the first-trimester diagnosis and management of pregnancies implanted into Cesarean section scars9–22 (Table 1). In this report we describe 18 cases of pregnancies implanted into the previous lower segment Cesarean section scar, which were diagnosed in our unit over a 4-year period. All were detected in the first trimester and we discuss the controversies surrounding both the diagnosis and clinical management of this condition.

Table 1. Treatment of Cesarean scar pregnancy: summary of the literature
Ref.Gravidity & ParityPrevious LSCS (n)hCG (IU/L)ViabilityDeficient myometriumTreatmentSuccessOutcome
  1. Ref., reference; hCG, human chorionic gonadotropin; G, gravida; P, para; N/K, not known; MTX, methotrexate; US, ultrasound; LSCS, lower segment Cesarean sections; KCl, potassium chloride.

 9N/KN/KN/KN/KNoLaparotomy, hysterotomy and curretageYesN/K
10G2 P11N/KYesNoExpectantNoEmergency LSCS hysterectomy at 35 weeks
11G2 P115789YesNoLocal MTXNoLaparotomy; pregnant 4 months later—normal intrauterine pregnancy
12G3 P2262 000YesYesExpectantNoGiven local MTX and KCl; 3 months to ultrasound resolution
13G4 P2112 000YesYesSystemic MTXYesResolution of ß hCG in 3 months; pregnant 8 months later—normal intrauterine pregnancy
14G4 P1154NoYesLaparoscopic surgeryYesFollow-up visit after 1 year—all well
15G4 P2224 940N/KYesLaparotomyYesPregnant 6 months later—normal twin intrauterine pregnancy
16G8 P517250NoYesLocal glucose, systemic MTXYes8 weeks to resolution
17G3 P2121 866YesYesLaparotomy, hysterotomy, uterine artery ligation, systemic MTXYesFollow-up visit after 4 months—all well
18G6 P1123 328YesYesLocal MTXYesResolution of hCG in 2 months
19G3 P22N/KNoYesLaparotomyYesPregnant 2 months later—normal intrauterine pregnancy
20G2 P1169 000NoYesUS-guided aspirationNoGiven systemic MTX resolution of hCG in 3 weeks
21G3 P2219 755YesYesLaparotomyYesN/K
22G5 P4411 306YesNoSystemic MTXYesN/K

Methods

This study was set in a dedicated tertiary referral early pregnancy assessment ultrasound unit in an inner city hospital. The unit serves a racially mixed population with a high level of socioeconomic deprivation. Women with suspected early pregnancy complications were either self-referred for assessment or they were referred to the unit by a local general practitioner, the hospital accident and emergency department or one of the hospital consultants. Tertiary referrals of women with ectopic pregnancies were also received from other hospitals within the UK.

All women attending the early pregnancy unit had a positive urine pregnancy test, which is routinely performed prior to clinical and ultrasound assessment (Clearview HCG II™, Unipath, Bedford, UK). The test is a monoclonal-based antibody test, which, according to the manufacturer's specifications, has a sensitivity of 99% at a urine β-human chorionic gonadotropin (β-hCG) level greater than 25 IU/L. All women were assessed by gynecologists who were fully trained in transvaginal sonography. A full history was always taken initially and entered into a clinical database (PIA Fetal Database, Version 3.23, Viewpoint Bildverarbeitung GmbH, Munich, Germany). When appropriate, clinical examination was also carried out by the attending physician, including a vaginal speculum examination. A transvaginal ultrasound scan was then performed to examine the viability and location of the pregnancy. Implantation into the previous Cesarean section scar (Figure 1) was diagnosed when the following criteria were satisfied: (1) empty uterine cavity; (2) gestational sac located anteriorly at the level of the internal os covering the visible or presumed site of the previous lower uterine segment Cesarean section scar; (3) evidence of functional trophoblastic/placental circulation on Doppler examination, which was defined by the presence of an area of increased peritrophoblastic or periplacental vascularity on color Doppler examination, and high-velocity (peak velocity > 20 cm/s), low-impedance (pulsatility index < 1) flow velocity waveforms on pulsed Doppler examination23; (4) negative ‘sliding organs sign’, which was defined as the inability to displace the gestational sac from its position at the level of the internal os using gentle pressure applied by the transvaginal probe.

Figure 1.

Four-week gestational sac implanted into the lower segment Cesarean section scar.

In all women with the diagnosis of pregnancy implanted into the lower segment Cesarean section scar the appearance of the myometrium at the implantation site was also examined. The scar was described as deficient whenever there was a visible gap in the myometrium of the anterior uterine wall and the pregnancy was seen bulging towards the urinary bladder (Figures 2 and 3).

Figure 2.

Longitudinal section of the uterus showing a 7-week gestational sac implanted into a previous Cesarean section scar and protruding towards the urinary bladder.

Figure 3.

Transverse section of the uterus at 6 weeks' gestation showing a gestational sac with a live embryo protruding through the left side of the Cesarean section scar.

A blood sample was obtained from each woman to ascertain the full blood count, blood group, cross-match and serum β-hCG (World Health Organization, Third International Reference 75/537). In women who were considered suitable for medical treatment with methotrexate, baseline liver and renal function tests were also performed.

Women were informed about the poor understanding of the natural history and clinical significance of first-trimester pregnancies implanted into previous Cesarean section scars. Clinical symptoms, pregnancy viability, gestational age and evidence of myometrial deficiency determined the management in each individual case. Women with minimal clinical symptoms who had small non-viable pregnancies were considered suitable for all management options including expectant management. Women with a viable pregnancy > 7 weeks' gestation and those with signs of abnormal placentation involving the myometrium were offered medical treatment with local injection of methotrexate. Surgical evacuation of pregnancy was offered to all women with a pregnancy < 7 weeks, those who experienced heavy bleeding and those in whom non-surgical treatment failed. However, the management plan was sometimes modified depending on the woman's own views and preferences and in some cases on the opinion of the referring consultant obstetricians and gynecologists.

Medical treatment involved an injection of 25 mg methotrexate directly into the pregnancy. The injection was administered transvaginally under continuous ultrasound guidance using a 20-G needle. Antibiotic prophylaxis of a single intravenous dose of 1.5 g cefuroxime and 500 mg metronidazole was given to all women. All procedures were performed on an outpatient basis under mild analgesia (50 mg pethidine and 10 mg metoclopramide intravenously). In cases with detectable embryonic cardiac activity, embryocide was performed first by the intracardiac injection of 0.1–0.2 nmol potassium chloride (KCl).

Surgery involved the use of suction curettage under ultrasound guidance. In cases complicated by heavy intraoperative bleeding, a 16–22-G Foley catheter was inserted at the level of the implantation site and inflated with 30–90 mL saline in an attempt to achieve hemostasis by compression. The catheter was left in situ for 12–24 h and then gradually deflated and removed.

Follow-up consisted of weekly outpatient clinical assessment and measurements of serum hCG levels. Once hCG levels declined to < 25 IU/L an ultrasound examination was performed to assess the size of the retained products of conception. Ultrasound examinations were then arranged on a monthly basis until it was confirmed that all pregnancy tissue had been spontaneously expelled or absorbed. All women who were planning future pregnancies were encouraged to attend for an early scan in order to assess the location of pregnancy.

Results

Over the 4-year period 18 245 women with suspected early pregnancy complications attended the early pregnancy unit. Eighteen women were found to have pregnancies implanted into a previous lower segment Cesarean section scar. Eight of these women were referred from other hospitals, and ten lived locally. As only 1–2% of women with early pregnancy problems were referred from other units, the prevalence of Cesarean scar pregnancy in the local population of women attending the early pregnancy assessment unit was approximately 1 : 1800. The mean number of previous Cesarean sections in women with scar pregnancies was 1.9 (range, 1–4) and 13/18 (72%) of these women had undergone previous multiple Cesarean sections (Table 2).

Table 2. Characteristics of 18 women with first-trimester Cesarean scar pregnancies
Case no.Maternal age (years)ConceptionGravidity & ParityPrevious LSCS (n)Gestation (weeks)Initial hCG level (IU/L)Gestational sac diameter (mm)ViabilityCRL (mm)Deficient myome- triumManagementSuccessComplicationsTime to resolution (weeks)Future pregnancy
  1. hCG, human chorionic gonadotropin; CRL, crown–rump length; IVF, in-vitro fertilization; N/K, not known; N/A, not applicable; D&C, dilatation and curettage; KCl, potassium chloride; LSCS, lower segment Cesarean sections.

 139SpontaneousG3, P1111762014Non-viableN/AYesD&C, Foley catheterYesIntraoperative hemorrhage < 1000 mLN/ANot trying
 233SpontaneousG2,P11 4973011Non-viableN/ANoD&CYesNoneN/ANot yet conceived
 342SpontaneousG6,P11 646229Non-viableN/ANoD&CYesNoneN/ANormal pregnancy 5 months later
 427SpontaneousG4,P2210N/K34Non-viableN/AYesD&CYesNoneN/ANormal pregnancy 9 months later
 537SpontaneousG5,P222328325Non-viableN/AYesD&C, Foley catheterYesIntraoperative hemorrhage < 1000 mLN/ANot trying
 634SpontaneousG4,P22 6N/K20Viable2.4NoD&CYesNoneN/ANot yet conceived
 735SpontaneousG6,P54 611 18410Viable3YesD&CYesNone10Not trying
 830SpontaneousG3,P2214N/K25Non-viableN/ANoD&C, Foley catheterYesIntraoperative hemorrhage < 1000 mLN/ANot trying
 9 SpontaneousG4,P31623 70019Viable18.7YesLocal methotrexateYesNone6Not trying
1038SpontaneousG7,P22 818 09016Viable10.8YesLocal methotrexateNoHemorrhage > 1000 mL, D&C, blood transfusionN/ANot trying
1134SpontaneousG4,P21 615 5407Viable2.3YesLocal methotrexateYesNone6Normal pregnancy 6 months later
1239SpontaneousG5,P22 538239ViableN/AYesLocal methotrexateYesNone10Not trying
1343spontaneousG7,P22 664 29925Viable2.3YesLocal methotrexateNoHemorrhage > 1000 mL, embolization D&C, blood transfusionN/ANot trying
1443IVFG5,P33 992 88037Viable23.8YesLocal methotrexate + KClYesNoneN/KAwaiting IVF
1536Spontaneous (heterotopic)G10,P33 768 80027Viable16YesLocal KCLYesHemorrhage at 31 weeks > 1000 mL Emergency LSCSN/ANot trying
1635SpontaneousG8,P33 530137Non-viableN/ANoExpectantNoMissed miscarriage at 10 weeks Systemic methotrexateN/KNormal pregnancy 6 months later
1735SpontaneousG5,P22 913.58Viable1.6NoExpectant for 4 monthsYesProlonged bleeding < 1000 mL16Normal pregnancy 3 months later
1838IVFG3,P22 6N/K12Viable2.3N/KExpectantNoSevere hemorrhage at 17 weeks > 1000mL HysterectomyN/A 

The gestational age of the Cesarean scar pregnancies ranged between 4 and 23 weeks. On ultrasound examination, 11/18 (61%) pregnancies had evidence of cardiac activity. However, all pregnancies diagnosed at > 12 weeks' gestation consisted only of retained placental tissue following first-trimester embryonic demise. Eight (44%) women with Cesarean scar pregnancies were initially treated surgically, seven medically (39%), and three expectantly (17%). Surgical management was successful in all cases, although three of eight (38%) women suffered significant bleeding (500–1000 mL) which required the insertion of a Foley catheter into the cervix in order to achieve hemostasis. There were no cases of retained products of conception following surgical treatment.

In the group of women who were managed medically the success rate was 5/7 (71%). Two women in this group required surgical intervention and blood transfusion due to heavy vaginal bleeding (> 1000 mL). A woman with heterotopic intrauterine and Cesarean scar pregnancies was treated only with local injection of KCl into the scar pregnancy. The intrauterine pregnancy remained viable and progressed normally until the third trimester. At 31 weeks she suffered a major antepartum hemorrhage and an emergency Cesarean section was performed. At the operation the uterine scar appeared thin and severely deficient. Bleeding was originating from the left lateral edge of the scar where a small amount of retained trophoblast tissue was found. Histological examination confirmed the presence of first-trimester trophoblast. However, due to the special circumstances of this case the management was classified as successful.

In women successfully treated with methotrexate the hCG resolution time was between 6 and 10 weeks. None of the women in this subgroup experienced any side effects, which could be attributable to the medication.

Expectant management was successful in one of three cases (33%). This woman had an incomplete miscarriage at 10 weeks. Another women, who was diagnosed with a scar pregnancy at 5 weeks' gestation, had a follow-up scan in another hospital which showed that embryonic demise occurred at 10 weeks. She was given systemic methotrexate and the pregnancy resolved without need for further intervention. The third woman who was managed expectantly had a spontaneous miscarriage at 17 weeks' gestation. She suffered a severe hemorrhage requiring an emergency hysterectomy.

At the time of writing, seven of the women with previous Cesarean scar ectopic pregnancies had tried for another pregnancy. Five of them (71%) have been successful, all achieving normal singleton intrauterine pregnancies.

Discussion

The prevalence of first-trimester Cesarean scar pregnancies was much higher in the population of women attending our early pregnancy unit than would be expected from the review of individual case reports from the literature. This may be explained by the liberal use of transvaginal sonography in our unit, which facilitates the diagnosis of abnormal uterine implantation. In addition, the appearance of a previous Cesarean section scar is routinely examined in our unit in both pregnant and non-pregnant women. An effort to exclude implantation into the uterine scar is always made in all women with a history of previous Cesarean section attending for an early pregnancy scan.

Previous studies have shown that women with a history of Cesarean section are more likely to experience miscarriage, ectopic pregnancy, abruptio placentae, placenta previa and placenta accreta24. Some of these complications, such as tubal ectopic pregnancy, are likely to be a result of postoperative infection and adhesions, which may occur after any abdominal surgery. Our study showed that 44% of Cesarean scar pregnancies ended in spontaneous first-trimester miscarriage. This rate would probably be higher if all the women with a viable pregnancy on the initial scan were managed expectantly. However, the mean maternal age was 36 years in this study, which may also have been a significant contributing factor to this high miscarriage rate.

We have already referred to the high prevalence of placenta previa and accreta in women with multiple Cesarean sections1–5. In 10 of 14 cases (71%) of first-trimester Cesarean scar pregnancies reported in the literature, the authors observed that the pregnancy penetrated deep into the myometrium (Table 1). Typically, the anterior uterine wall appeared thin in these cases and the pregnancy was seen bulging through the anterior uterine wall, reaching to within a few millimeters from the bladder. We saw similar findings in 11/18 (61%) women in our study (Figures 2 and 3). Most previous reports classified these cases as intramural pregnancies. However, the term intramural pregnancy implies that the pregnancy is completely confined to the myometrium, which is not the case with the majority of scar pregnancies. As the first-trimester findings closely resemble the appearance of placenta and myometrium in cases of placenta accreta later in pregnancy, the term ‘abnormally adherent trophoblast/placenta’, may be more appropriate.

The propensity of the Cesarean scar pregnancy to invade deep into the myometrium could be explained by the characteristics of uterine Cesarean scars in non-pregnant women25. The majority of scars are well-healed (Figure 4), but in a small number of women the anterior uterine wall is deficient (Figure 5). These deficient scars are more likely to occur after multiple Cesarean sections due to fibrosis, which leads to poor vascularity of the lower uterine segment and impaired postoperative healing. The scar surface area is increased in women with multiple sections, which in turn increases the chance of a pregnancy implanting into the scar. This may explain both the high number of previous multiple Cesarean sections and the high proportion of abnormally inserted placentae in cases of scar implantation.

Figure 4.

Longitudinal section of a non-pregnant uterus showing a well-healed lower segment Cesarean section scar.

Figure 5.

Longitudinal section of a retroverted non-pregnant uterus showing a highly deficient lower segment Cesarean section scar.

There is no agreement on the best method and criteria to diagnose Cesarean scar pregnancies. However, transvaginal ultrasound has been used in most studies and is likely to emerge as a future gold standard for the diagnosis of scar implantation. Diagnosis is relatively simple early in pregnancy, but as the pregnancy progresses, the distinction between Cesarean scar, cervical pregnancy and low intrauterine pregnancy becomes more difficult. Vial et al.19 proposed a list of sonographic criteria which should be present in order to make the diagnosis of a Cesarean scar ectopic pregnancy. They included: (i) the trophoblast must be mainly located between the bladder and the anterior uterine wall; (ii) no fetal parts must be visible in the uterine cavity; (iii) on a sagittal view of the uterus running through the amniotic sac, a discontinuity in the anterior wall of the uterus should be identified. However, in order to avoid false-positive diagnosis of scar implantation in cases of incomplete miscarriages being expelled from the uterine cavity, we feel that in women with non-viable pregnancies, Doppler ultrasound and ‘sliding organs sign’ should be used to confirm the diagnosis of a scar pregnancy. In cases of miscarriages of intrauterine pregnancies the gestational sac appears avascular, reflecting the fact that the sac has become detached from its implantation site, whereas in Cesarean section scar pregnancies the gestational sac appears well-perfused on Doppler examination.

With limited experience of Cesarean scar pregnancies in the first trimester, it is difficult to decide on optimal management in individual cases. The majority of women in previous reports were treated either surgically or medically, whilst two women were managed expectantly. It is interesting that the authors of the previous case reports who chose surgical treatment all opted for the transabdominal approach, using either laparoscopy or open laparotomy to excise the pregnancy from the Cesarean section scar. Although the treatment was successful in all cases, we achieved a similar success rate using a transcervical approach, which is simpler and has a quick postoperative recovery. Analysis of published case reports showed that the success of methotrexate, when used as the initial treatment option, was 80%. One woman in whom treatment failed had an open laparotomy and excision of pregnancy. These results are similar to the 71% success rate of medical treatment in our study.

There are two reports of expectant management in the literature. One of these women required additional treatment with methotrexate12, whilst the other had an emergency Cesarean hysterectomy at 35 weeks' gestation10. Expectant management failed in two of three cases in our study, one women requiring methotrexate and the other an emergency hysterectomy. These results indicate that expectant management of viable scar pregnancies carries a significant risk of emergency hysterectomy if the pregnancy progresses beyond the first trimester.

Conclusions

The prognosis of a Cesarean section scar pregnancy diagnosed in the first trimester appears to be much better than the prognosis of placenta previa/accreta detected in the third trimester. However, given the uncertainties about the diagnostic criteria and prognosis, the main question is whether any form of treatment is justified in these cases. We feel that an experienced operator using transvaginal sonography should be able to establish with a high degree of accuracy the diagnosis of scar pregnancy and detect an abnormal adherence of trophoblast to the myometrium. Current data also indicate that expectant management is rarely successful and is particularly unsuitable for women with viable scar pregnancies. It is also important to emphasize that the maternal morbidity and length of follow-up both increase with gestation.

Myometrial involvement is a particularly unfavorable feature, which makes surgical termination of pregnancy very difficult. In addition, it is very likely, although unproven, that these women will develop placenta previa/accreta should the pregnancy be allowed to progress into the third trimester. In the absence of reliable scientific data we believe that each woman should be presented with all available information and given the opportunity to decide on the management of her pregnancy. Should she opt for treatment, a local injection of methotrexate and transcervical aspiration of pregnancy should be used in preference to a laparoscopy or laparotomy.

Ancillary