Dr A Ash, Consultants’ Office, Women’s Health, 10th Floor, North Wing, St Thomas’ Hospital, Lambeth Palace Road, London SE1 7EH, UK. Email Alok.Ash@gstt.nhs.uk
Caesarean scar pregnancy is one of the rarest forms of ectopic pregnancy. Little is known about its incidence and natural history. With increasing incidence of caesarean section worldwide, more and more cases are diagnosed and reported. Transvaginal ultrasound and colour flow Doppler provides a high diagnostic accuracy with very few false positives. A delay in diagnosis and/or treatment can lead to uterine rupture, major haemorrhage, hysterectomy and serious maternal morbidity. Early diagnosis can offer treatment options of avoiding uterine rupture and haemorrhage, thus preserving the uterus and future fertility. Management plan should be individually tailored. Available data suggest that termination of pregnancy is the treatment of choice in the first trimester soon after the diagnosis. Expectant treatment has a poor prognosis because of risk of rupture. There are no reliable scientific data on the risk of recurrence of the condition in future pregnancy, role of the interval between the previous caesarean delivery and occurrence of caesarean scar pregnancy, and effect of caesarean wound closure technique on caesarean scar pregnancy. In this article, we aim to find the demography, pathophysiology, clinical presentation, most appropriate methods of diagnosis and management, with their implications in clinical practice for this condition.
The secondary rise of repeat caesarean delivery in the western world has been associated with an increase in complications of placentation in subsequent pregnancies, e.g., placenta accreta and its subtypes.1 Embryo implantation in a previous caesarean scar (CS) resulting in a caesarean scar pregnancy (CSP) is another rare but potentially catastrophic complication of a previous caesarean birth.
The first case of a CS ectopic pregnancy was reported in English medical literature in 1978.2 Since then, there are only 19 cases published until 2001.3 But over the past 5 years, there has been a substantial increase in the number of CSP published in the English language literature. This may reflect a ‘true’ increase in its incidence because of the rising caesarean section rate worldwide or an ‘apparent’ one as result of better detection of the CSP by liberal use of transvaginal ultrasound scan. This review article aims to give an overview on the subject.
Materials and methods
We searched MEDLINE/PubMed electronic database using medical subheading search words such as ‘caesarean scar pregnancy’; ‘caesarean scar’; ‘caesarean scar complications’; ‘caesarean scar implantations’; ‘ectopic pregnancy’; ‘caesarean scar ectopic pregnancy’; ‘pregnancy, previous caesarean scar’; and ‘unusual ectopic pregnancies’ to obtain a comprehensive list of articles concerning this condition from January 1966 to October 2006. We also searched using ‘Scholar Google’ search engine in the internet (www.scholar.google.com) for articles on this topic since 1987. The citations were cross-checked via related article and Link in PubMed. Additional articles were obtained from the cross-references of the relevant publications for bibliographical purpose. Of all the articles found, we chose only those on CSP published in English language journals and the English abstracts of original articles in other languages from MEDLINE. Altogether we found 58 citations on CSP (total 161 cases); eight of these were case series comprising between 8 and 18 cases and three were reviews, one of which was a systematic review. The rest were case reports (of one to three cases) with or without a review of the literature (Figure 1).
The incidence of CSP is unknown, as very few cases have been reported in the literature. Jurkovic et al.4 have estimated a prevalence of 1:1800 in their local population of women attending the early pregnancy assessment unit. A recent case series5 estimates an incidence of 1:2226 of all pregnancies, with a rate of 0.15% in women with a previous CS and a rate of 6.1% of all ectopic pregnancies in women who had at least one caesarean delivery. The gestational age at diagnosis ranged from 5+0 to 12+4 weeks (mean 7.5 ± 2.5 weeks), and the time interval between the last caesarean section and the CSP was 6 months to 12 years in this series. CSP has been described in spontaneously conceived pregnancy as well as after in vitro fertilisation (IVF) and embryo transfer.6 IVF-associated heterotopic CSP, a rarer event, has also been described, both with twins7 and triplets.8 No particular predilection has been reported for maternal age or parity.
Little information is available on the natural history of this condition. Very few of these pregnancies reported in the literature progressed beyond first trimester4,9 as almost all are terminated during this period. It is likely that if a developing pregnancy in a caesarean section scar were to continue to the second or third trimesters, there would be a substantial risk of uterine rupture with catastrophic haemorrhage, with a high risk of hysterectomy causing serious maternal morbidity and loss of future fertility. There is also a danger of invasion of the bladder by the growing placenta. A pregnancy that protrudes through the scar, if viable, can implant on other abdominal organs and continue to grow as a secondary abdominal pregnancy.10,11 However, if the pregnancy continues within the uterus, the risk of placenta accreta is significantly increased, up to three- to five-fold.12,13 CSP progressing to 35 weeks of gestation has been described, but this case was complicated by massive haemorrhage and disseminated intravascular coagulopathy at CS, requiring a life-saving hysterectomy.9
Predisposing factors, mechanism and pathophysiology
The exact cause and mechanism is not well understood. Implantation of a pregnancy within the scar of a previous caesarean section is different from an intrauterine pregnancy with placenta accreta.3 Placenta accreta is characterised by the absence of decidua basalis and varying degrees of invasion of the myometrium by trophoblastic tissue. The pregnancy is essentially within the uterine cavity. In CSP, the gestation sac is completely surrounded by myometrium and the fibrous tissue of the scar, quite separate from the endometrial cavity.14 The most probable mechanism that can explain scar implantation is that there is invasion of the myometrium through a microtubular tract between the caesarean section scar and the endometrial canal.3,15 Such a tract can also develop from the trauma of other uterine surgery, e.g. curettage, myomectomy, metroplasty, hysteroscopy and even manual removal of placenta.16 Damage to the decidua basalis during uterine surgery can persist in the endometrium in the form of tiny dehiscent tracts or minute wedge defects. A CSP is more aggressive in its behaviour than placenta praevia accreta because of its early invasion of the myometrium.6 Pathological findings after a total hysterectomy suggest that the villi are not merely penetrating the myometrium but are bound with or implanted in it.17 Vial et al.18 proposed two different types of CSPs. The first is an implantation on the prior CS with progression towards the cervicoisthmic space or the uterine cavity. Such a CSP may progress to a viable birth but with the risk of a life-threatening bleeding (see below, Expectant Treatment). The second is a deep implantation into a CS defect growing towards the bladder and abdominal cavity, a type that is more prone to rupture.
It is uncertain whether the risk of CSP is related to the number of previous caesarean sections. Some case series have reported that between 50 and 72% of CSPs occur after two or more prior caesarean sections.4,19 Some4 believe multiple CSs are a strong risk factor for CSP because of increased scar surface area, but others20,21 argue against such a correlation. A recent systematic review found that 52% cases followed one previous caesarean section, 36% after two and 12% after three or more previous caesarean sections.22 No particular order or sequence between the previous caesarean sections and subsequent pregnancies for any risk of CSP has been reported in the literature. Data on any correlation between the CSP and the indication of the caesarean delivery that preceded this abnormal implantation are inadequate, but, interestingly, breech presentation is the most frequent (31%) among the publications that documented the indications.22 We could not find any reference in the literature whether surgical technique (e.g. a single noninverting running suture technique for uterine closure instead of a double layer closure) might have any role on the causation of CSP. The impact of the time interval between the previous caesarean sections and the subsequent CSP implantation is also not clear. Some CSPs occurring within a few months of a caesarean section suggest that an incomplete healing of the caesarean sections scar may contribute to scar implantation,10 but this theory cannot explain the CSP that has been reported as late as 12 years after a prior caesarean sections.5
CSP may present from as early as 5–6 weeks5 to as late as 16 weeks.23 A light, painless vaginal bleeding is usually the early presenting symptom in 39%. Approximately 16% of women complain of accompanying mild to moderate pain and 9% complain of only abdominal pain.22 It can be an incidental finding in an asymptomatic woman (37%).22 Severe acute pain with profuse bleeding implies an impending rupture. Collapse or haemodynamic instability strongly indicates a ruptured CSP. Clinical examination in stable women is usually unremarkable. The uterus may be tender if the CSP is in the process of rupture.
Ultrasound is the first-line diagnostic tool for CSP. By far, the majority of the CSPs have been diagnosed by transvaginal scan (TVS) in the early weeks of pregnancy. A sagittal view along the long axis of the uterus through the gestation sac can localise a CSP with confidence. The following ultrasound criteria have been put forward for the diagnosis of a CSP.3,15
• an empty uterine cavity, without contact with the sac
• a clearly visible empty cervical canal, without contact with the sac
• presence of the gestation sac with or without a fetal pole with or without fetal cardiac activity (depending on the gestation age) in the anterior part of the uterine isthmus, and
• absence of or a defect in the myometrial tissue between the bladder and the sac (Figure 2).
In addition, there should be no adnexal mass or free fluid in the pouch of Douglas, unless the CSP has ruptured.
The thickness of the intervening myometrium between the gestation sac and the bladder has been shown to be less than 5 mm in two-thirds of the cases.24 There may be a discontinuity in the lower anterior wall of the uterus, with a bulging pregnancy sac protruding through the gap (Figure 3).25 TVS on its own has a diagnostic sensitivity of 86.4% (95% CI 0.763–0.9050).22 To reduce the risk of a false diagnosis, Maymon et al.19,21 recommend a combined approach: a TVS to obtain the fine details of the gestation sac and its relation to the scar followed by a meticulous abdominal scan with a full bladder. The latter provides a ‘panoramic view’ of the uterus and an accurate measurement of the distance between the gestation sac and the bladder. Jurkovic et al.4 have described a negative ‘sliding organ sign’, defined as inability to displace the gestational sac from its position at the level of the internal os by gentle pressure applied by the transabdominal probe, as diagnostic.
Additional diagnostic information can be obtained by colour flow Doppler to show distinct circular peritrophoblastic perfusion surrounding the gestation sac6,15,18 that can help delineate the CSP sac with location of the placenta in relation to the scar and proximity to the bladder.4 With pulsed Doppler functions, more information on the flow pattern of the peritrophoblastic vasculature can be obtained.5 Typically, a prominent high-velocity (peak velocity > 20 cm/second), low-impedance (pulsatility index < 1) flow velocity waveforms can be demonstrated, consistent with normal early pregnancy.26
Transvaginal three-dimensional (3-D) power Doppler ultrasound has been used to enhance the diagnostic accuracy of a CSP.27,28 Combination of the multiplanar views and surface-rendered images helps identify subtle anatomical details of a well-developed trophoblastic shell around the gestational sac.27 The thin myometrium between the gestational sac and the bladder wall can be recognised with confidence. Furthermore, peritrophoblastic flow surrounding the CSP may be illustrated by 3-D power Doppler. The gestational sac is neither distorted nor collapsed, which are common observations on 3-D ultrasound imaging of a miscarriage. Ballooning of the endocervical canal, usually associated with a cervical pregnancy, is also absent. Chou et al.29 have described a more sophisticated new 3-D colour Doppler imaging technique (termed 3-D-virtual organ computer-aided analysis [VOCAL]) to monitor the quantification of changes of uterine neovascularisation characteristics before and after successful treatment of CSP. They claim that this is a superior technique to display the CSP sac volume as well as to reveal a more detailed spatial angioarchitecture pattern than conventional colour Doppler image.
Magnetic resonance imaging
Magnetic resonance imaging (MRI) has been used as an adjunct to ultrasound scan15,21,24,25,30,31 (Figure 4). Both sagittal and transverse T1- and T2-weighted MRI sequences can clearly show the gestational sac embedded in the anterior lower uterus.
MRI is superior in the assessment of the pelvic structures because of improved differentiation of soft tissue, spatial resolution and the possibility of a multiplanar imaging.25 MRI can measure the volume of the lesion and thus help assess the indication and success of local methotrexate (MTX), with an added advantage that it can also improve intraoperative orientation.25 However, a major limitation of MRI is its long acquisition time. Many authors21,24 do not routinely recommend MRI, as they feel TVS combined with colour flow Doppler is very reliable in diagnosing a CSP; MRI may be reserved for cases where TVS and colour flow Doppler are inconclusive.
It allows the cervix and the uterine cavity to be distended, with relatively little trauma, with the finding of a normal and empty uterine cavity together with the pregnancy tissues at the lower corpus.32
Laparoscopy has been used for diagnosis of CSP.33–35 The uterus is usually seen normal sized or bulky (depending on the gestation age) with the CSP arising as a hillock with a ‘salmon red’ ecchymotic appearance, bulging the uterine serosa from the previous caesarean section scar behind the bladder36 (Figure 5). The fallopian tubes and the ovaries are seen normal.
Histology of excised CSP alone and in a hysterectomy specimen reveal interstitial trophoblasts within the fibromuscular tissue of the old caesarean section scar.25,27,37 The myometrium usually thins out to merge with the thin and fibrous scar of previous caesarean section. The placental attachment in the lower segment may lack both decidua basalis and myometrium, merely consisting of some connective tissue.9 These microscopic features coupled with absence of surrounding endocervical glands confirm a CSP and rules out a cervical pregnancy.27 Immunostaining with β-human chorionic gonadotrophin (β-hCG) and desmin confirm the presence of trophoblast cells within smooth myometrial muscle fibres19 (Figure 6).
Spontaneous miscarriage in progress and a cervicoisthmic pregnancy can be difficult to distinguish from a CSP. In a cervicoisthmic pregnancy, unlike a CSP, there would be a layer of healthy myometrium visible between the bladder and the gestation sac15 and bleeding as the presenting symptom is much heavier. A transabdominal ultrasound performed with a well-distended bladder can establish the diagnosis of a cervical pregnancy if the uterus is empty with an endometrial stripe, the gestation sac present dominantly within the cervix, giving an hour-glass shape to the uterus with a ballooned cervical canal.27,38 In case of a spontaneous miscarriage in progress, the gestation sac should be seen in the cervical canal on TVS, and on colour flow Doppler, the sac should appear avascular, indicating that the sac has been detached from its implantation site, in contrast to the well-perfused CSP located in its unique site.39 An exponential fall in the serum β-hCG level would also be noticed in a miscarriage. Rarely, a CSP can be misdiagnosed as a trophoblastic tumour if ultrasound scan indicates no evidence of pregnancy within the uterus but detects a highly vascularised mixed mass infiltrating the myometrium in presence of a high serum β-hCG level.24 Diagnosis of a CSP is relatively easy early in pregnancy, but as the pregnancy progresses, the distinction between CSP, cervical pregnancy and a lowly implanted intrauterine pregnancy becomes more difficult.4
Because of the rarity of the condition, majority of CSPs are case reports or small case series reported in the literature, with no consensus on the preferred mode of treatment. Generally, termination of pregnancy (TOP) in the first trimester is strongly recommended, as there is a high risk of subsequent uterine rupture, massive bleeding and life-threatening complications. At this gestation, the embryo is soft and fragile; vascularity of the placental bed, depth of placental implantation and risk of invasion of the bladder are all considerably less than those later in pregnancy. Treatment objectives should be to perform feticide prior to rupture, to remove the gestation sac and to retain patient’s future fertility. Gestational age and viability, evidence of myometrial deficiency and clinical symptoms at presentation have been considered by various authors to determine the management.
Conservative medical treatment
Systemic administration of MTX is a standard treatment for tubal ectopic pregnancy. There should be no reason to doubt its efficacy on CSP. Furthermore, systemic administration of the drug has not been reported to cause any adverse effects such as nausea, stomatitis, alopecia, pneumonitis, etc.22 in the treatment of CSP. Indeed, CSPs have been shown to respond well to it (dose of 50 mg/m2), especially in those with β-hCG levels < 5000 miu/ml.30,40 Conservative medical treatment is appropriate for a woman who is pain free and haemodynamically stable with an unruptured CSP of <8 weeks of gestation and a myometrial thickness < 2 mm between the CSP and the bladder.21 All women considered suitable for MTX treatment should have prior baseline full blood count and liver and renal function tests performed. They must be agreeable to surgery if medical treatment fails or if the CSP ruptures.37 Owing to the short half-life of MTX (10 hours), a systemic regimen needs repeated doses. For these reasons, some authors41 prefer direct intrasac injection of MTX (see below) as it achieves a high concentration locally and therefore interrupts the pregnancy more rapidly, as opposed to an extremely slow absorption of pregnancy following the systemic administration of the drug. But pharmacokinetics of the drug has not shown any advantage of local injection over its systemic administration in regard to serum drug level and systemic toxicity.42
Local injection of embryocides
This has been successfully reported with local injection of MTX,15,40,43 potassium chloride,15,44 hyperosmolar glucose36 and crystalline trichosanthin.24 Under ultrasound guidance, MTX can be injected locally to the gestation sac via transabdominal or via transvaginal route. Transabdominal route requires a longer needle, used with caution not to penetrate the bladder wall, and does not require any anaesthesia. The transvaginal approach allows for a shorter distance to the gestation sac with minimal risk of bladder injury. The usual technique for injection of MTX uses 20- to 22-gauge needle. But the lumen of these needles may be too narrow to allow concurrent aspiration of the embryo. To overcome this, 16-gauge double-lumen oocyte-retrieval IVF needles have been used with success.5,32 These thicker double-lumen needles ensure better aspiration of the trophoblastic tissue via one lumen and injection of MTX through the other. But the procedure may require general anaesthesia.5,32 There is no agreement on the use of prophylactic antibiotics during the procedure.
Combined medical treatment
Combined medical treatment in varying regimens has been described by many authors, e.g. local injection of 8 mEq potassium chloride (2 mEq/ml) followed by 60 mg of MTX injected into the gestation sac,15,41 direct injection of 3 ml of 50% glucose plus oral MTX (2.5 mg three times a day for 5 days),36 multi-dose systemic MTX (1 mg/kg) with alternate day folinic acid rescue,30,32,37 failed systemic MTX followed by successful local MTX,45 simultaneous intravenous and intra-amniotic injections of MTX on day 1 followed by two repeat doses on day 4, plus calcium folinate (30 mg orally/day) from day 5 to 11,46 cervical injection of crystalline trichosanthin (1.2 mg) followed by oral mifepristone (50 mg orally every 12 hours for 3 days) or intramuscular MTX,24,25,47 and systemic MXT followed by oral mifepristone.24
Medical treatment combined with surgical sac aspiration
Medical treatment—systemic or local, single agent or combined regimen—can interrupt the pregnancy, but symptoms can continue with bleeding, sometimes heavy. Also, it is difficult to rule out some scar dehiscence already developing at the time of treatment, as the very thin myometrium could be in a state of prerupture.43 Medical treatment has therefore been combined with surgical aspiration of the sac in some cases. Various sequences of combination have been described:
• local potassium chloride → TVS-guided sac aspiration → local MTX injection → intramuscular MTX injection39
• sac aspiration (transvaginal or transabdominal) → local MTX injection5,48
• sac aspiration under ultrasound guidance → systemic MTX47
• systemic MTX → sac aspiration by vaginal route → local MTX.32
A supportive, secondary, salvage treatment, e.g. hysteroscopic or laparoscopic resection of the CSP under direct vision, has been indicated in some cases.18,40 Selective embryo reduction by potassium chloride and primary sac aspiration have been performed in heterotopic IVF twin and triplet pregnancies, respectively.7,8
Medical treatment alone or in conjunction with needle aspiration can avoid unnecessary laparotomy and preserve the woman’s fertility, but it requires time and patience. It may take 4–16 weeks for β-hCG to drop to normal.15,49,50 Close follow up with serial β-hCG monitoring is therefore essential. Daily monitoring during the hospital stay and weekly thereafter until a level of <5 miu/ml has been recommended.5 Ideally, such follow up should include ultrasound scan evidence of resorption of any residual pregnancy tissue. It is hard to predict when the CS mass completely resolves after a conservative treatment. In some cases, it has been found to take several months to a year.5,15,49 One possible explanation is that the scanty venous flow within the fibrous scar tissue makes the resorption of the residual trophoblastic tissue difficult. A second mechanism might be related to the proliferation of collagen fibres or fibrous tissue in the isthmic portion of the uterus in response to myometrial injury induced by placental villi invasion.5
Serial transvaginal colour flow Doppler is useful for monitoring the response to medical treatment and appears to correlate well with β-hCG levels.5 The high-velocity, low-amplitude, turbulent flow remains prominent without much change during the course of follow up till the β-hCG level returns to normal. Until then, these women with such flow characteristics should be warned of the risk of uterine rupture or internal bleeding.43 Furthermore, a high peak systolic velocity on Doppler could be a clear warning to the surgeon should a surgical procedure be undertaken. Chou et al.29 have used 3-D VOCAL imaging system for follow up.
It is now well recognised that for a tubal ectopic pregnancy, a higher failure rate of medical treatment is associated with a gestation age of ≥9 weeks, a fetal pole >10 mm, presence of embryonic cardiac activity and a serum β-hCG concentration >10 000 miu/ml.51 However, no similar or consistent pattern could be observed for a CSP.37 But, systemic MTX appears to be more effective in women with β-hCG level <5000 miu/ml. For levels exceeding this, intralesional injection and/or other additional interventions are indicated.22 Continuation of fetal cardiac activity or growth of the sac with rising β-hCG concentration indicate failure of medical treatment. Analysis of published case reports show that medical treatment with MTX is successful in 71–80% of cases, with 6% of women requiring hysterectomy.4
To date, a total of 12 cases of CSP managed by uterine curettage as primary therapy have been reported in the English medical literature,52,53 and another four cases have been cited in a case series of eight CSPs.34 A review of the literature by Arslan et al.52 shows that uterine curettage was either unsuccessful or caused complications in eight out of nine women, requiring surgical treatment, and in a case series of eight CSPs, Wang et al.34 had four secondary referrals after failed curettage, thus indicating a failure rate of 70% (12/17).
The gestation sac of a CSP is not actually within the uterine cavity, and the chorionic villi implant into the caesarean section scar of the lower segment. Therefore, not only the trophoblastic tissue is unreachable by the curette but also such attempts can potentially rupture the uterine scar leading to severe haemorrhage and cause more harm.3,24 Profuse bleeding during the procedure and absence of chorionic villi in the specimen obtained by curettage must prompt immediate laparoscopy/laparotomy.
Blind uterine curettage as a primary treatment for CSP is therefore insufficient and should be discouraged. But some authors argue for suction curettage under ultrasound guidance in selected cases where the gestation is ≤7 weeks and the myometrial thickness anterior to the CSP is ≥3.5 mm.52,53 Notwithstanding this, the lack of direct visualisation, risk of a local haematoma formation and the need for a prolonged β-hCG follow up remain the major drawbacks,33,52 and more reports are needed to rationalise this treatment modality.52 Conversely, serious bleeding after an otherwise uncomplicated dilatation and evacuation procedure for a missed or therapeutic abortion in a woman with a previous CS should raise alarm for a CSP.54
Various haemostatic measures have been used successfully as an adjunct to conservative treatment of viable CSPs for the prevention and control of profuse bleeding, e.g. local injection of vasopressin,20 intrauterine balloon tamponade by Foley catheter (30–90 ml of balloon capacity for 12–24 hours),4,20 Shirodkar suture,55 prior selective bilateral uterine artery embolisation48,56,57 and bilateral uterine artery liagtion.28,33,58,59
In 2005, Wang et al.49 have described a successful treatment of CSP by operative hysteroscopy and suction curettage, the first of its kind reported in English language literature. At 4-week follow up, serum β-hCG level became normal, with restoration of normal echotexture of the uterus on ultrasound scan. Clinical follow up at 3 months did not reveal any complication. The authors have since reported hysteroscopic management of six more cases with success in all of them, with no complication and no blood transfusion.33 They conclude that this procedure offers an important alterative treatment for CSP, with a short operative time (mean 36.7 ± 20.8 minutes), less blood loss (mean 50.0 ± 0.0 ml), short postoperative stay (mean 1.1 ± 0.9 days) and a rapid return of the pregnancy test to negative (<4 weeks, mean 22 days). Most importantly, the fertility is conserved after the surgery. The procedure requires general anaesthesia, operative skill and facilities. Direct visualisation of the CSP with meticulous coagulation of the blood vessels at the implantation site is crucial to prevent severe intraoperative haemorrhage. Chao et al.60 have described a successful hysteroscopic management of a CSP after failed curettage and MTX treatment. The success rate of this modality (8/8 = 100%) is encouraging; however, the total number of cases managed is small, and it is too early to know whether operative hysteroscopy can be performed in presence of heavy bleeding or with unstable vital signs.
Lee et al.,61 the first to perform a successful laparoscopic resection of a CSP, have reported four such cases followed by others.33,35 Operative laparoscopy should be performed only after a prior TVS confirms the diagnosis. Laparoscopic findings as described above must corroborate with the features of a CSP. The CSP mass is incised and the pregnancy tissue removed in an endobag. Bleeding can be minimised by local injection of vasopressin (1 unit/ml, 5–10 ml),34 haemostasis achieved by bipolar diathermy34,61 and the uterine defect closed with endoscopic suturing. In trained hands, laparoscopic treatment is safe and less time consuming (mean 113.8 ± 32.0 minutes, range 75–120 minutes),33,34 blood loss is limited (mean 200.0 ± 108.0 ml, range 50–200 ml)33,34 and recovery is fast with minimal hospital stay (mean 3.0 ± 0.0 days).33 This method is suitable only if the woman is stable and appropriate facilities available with a trained surgeon. In the event of difficulty in achieving haemostasis or uncontrollable bleeding, the operation must be converted to a laparotomy.
An experienced endoscopist should make the choice between a laparoscopy and hysteroscopy; an operative hysteroscopic approach should be chosen for the CSP that grows inwards toward the uterine cavity, while a laparoscopy is more justified for a deeply implanted CSP growing towards the abdominal cavity and bladder.18,33
Primary open surgical treatment
Laparotomy followed by wedge resection of the lesion (hysterotomy) should be considered in women who do not respond to conservative medical and/or surgical treatments, present too late or if facilities and expertise for operative endoscopy are not available. Laparotomy is mandatory when uterine rupture is confirmed or strongly suspected. This conventional low-tech surgery, which is available in all hospitals, has the advantage of complete removal of the CSP and simultaneous repair of the scar, followed by a quick return of the β-hCG to normal level within 1–2 weeks. Some consider this as the best treatment option,3,18 as the excision of the old scar not only avoids the possibility of residual trophoblasts being left in situ,27 but also removes the microtubular tracts and thus reduces the risk of recurrence. This approach, however, inflicts a larger surgical wound, longer hospital stay and longer recovery time, with a possible higher risk of a future placenta praevia accreta.11
If the woman does not wish to have a TOP and wants to continue the pregnancy, and there is sonographic evidence of the sac growing towards the uterine cavity (the first variety of the CSP as described by Vial et al.18), an expectant treatment may be considered. It must be stressed, however, that the minimum thickness of the myometrium anterior to the CSP sac to warrant safety of a continuing pregnancy is unknown. Some authors21,33,36 question the prudence and safety of continuing a CSP. In their experience, the prognosis of an uneventful term pregnancy is very poor. If expectant treatment is intended, a detailed care plan must be documented with the risks and benefits of pregnancy terminated versus continued. The woman must be appropriately counselled by a senior clinician with a clear explanation that continuation of pregnancy will almost invariably lead to uterine rupture with serious consequences. Progress of the pregnancy must be closely monitored as an inpatient with regular imaging. An elective delivery by caesarean section around 28–30 weeks with antenatal corticosteroid administered 24–48 hours before delivery would appear to be a reasonable compromise. Even as an elective procedure, massive haemorrhage may occur at caesarean delivery, with the risk of hysterectomy. In women having expectant treatment, emergency operative delivery must be instituted without delay in the event of any features indicating uterine rupture. Herman et al.9 expectantly managed a CSP which was suspected at 7 weeks. An elective caesarean section was planned at 36 weeks, but severe abdominal pain necessitated an emergency caesarean section at 35 weeks. Smith et al.23 described a case, first diagnosed at 16 weeks and managed expectantly, as the woman refused a TOP. She required a laparotomy and hysterotomy because of a ruptured CSP at 20 weeks. Both these cases were complicated by massive haemorrhage requiring multiple units of blood transfusion (the first case also needed a hysterectomy). Expectant treatment failed in two of the three women in a series of 18 cases reported by Jurkovic et al.,4 one woman requiring MTX and the other an emergency hysterectomy.
These results indicate that expectant treatment of viable CSPs either fails or carries a significant risk of rupture requiring laparotomy and hysterectomy.
In a review until August 2002, hysterectomy was performed in 7 out of 19 cases.3 From then, six more hysterectomies have been reported in the literature either as a primary procedure or because other treatment modalities failed.5,25,37,47,54 This shows that CSP is a potentially serious condition despite advances in many of the diagnostic techniques and therapeutic measures.
Future pregnancy prospects and management
Uneventful viable intrauterine pregnancies have been reported after all modalities of conservative management of a CSP.3,12,21,43,62 The largest series62 shows a 50% incidence, with a mean interval of 13.3 months (range 3–34 months) between the previous CSP and subsequent pregnancy. One woman died of uterine rupture in a future pregnancy at 38+3 weeks.
With the advent of sonohysterography, post-CS uterine wall integrity can be detected in a nonpregnant state.4,63,64 CS defect, defined by the presence of fluid within the incision site,63 or any filling defect (niche), defined as a triangular anechoeic structure at the site of the scar,65 or a gap in the anterior lower uterine segment myometrium at the site of prior caesarean deliveries64 should raise alarm for uterine scar complication in subsequent pregnancies. This is important for women at risk, such as those with a history of ectopic pregnancy, placental pathology or multiple CS.4,19 This might help the clinician in counselling the woman and alert the sonographer to look for the possibility of a CSP during early pregnancy scanning.
Owing to the insufficient data, it is difficult to advise on the future risk of recurrence. To date, only three cases of recurrent CSP have been reported in the world literature,55,62,66 one of whom had three recurrences!55 Likewise, no reliable statistics exist on the safe interval between a CS and any risk to a subsequent pregnancy. Although there is little data to suggest any danger in conceiving soon after treatment with MTX, some women may prefer to wait at least 6 months to minimise any potential teratogenic effect in a new pregnancy.67 Despite a 10-hour half-life, MTX can still be found in the liver and kidneys months later.68 Some authors recommend future pregnancies be avoided for more than 3 months and probably 1 or 2 years.5 In absence of any evidence base for this, such advice must be empirical. These women can use any contraceptive of their choice including oral contraceptives to prevent a new pregnancy while myometrial integrity was unproven.50 Once pregnant, they should have an early TVS in order to assess the location of the new pregnancy. Indeed, it should be a ‘routine practice’ to examine the appearance of a previous caesarean section scar in every early pregnancy assessment unit.4 The rest of the pregnancy should be closely monitored as they may be at risk of placenta accreta. There is no evidence to suggest other complications, e.g. fetal growth restriction. In absence of any obstetric indication for a caesarean section, these women should deliver vaginally. For women with a history of ruptured CSP or a very thin or absent myometrium between the CSP and bladder in the current pregnancy, it seems rational to recommend a caesarean section once a reasonable fetal maturity is achieved, aided by antenatal corticosteroid, to avoid the possible risk of a spontaneous uterine rupture.5
Implications for clinical practice
Although caesarean section is a common operation, CSP is comparatively a rare occurrence. However, its incidence is likely to increase with the increasing caesarean rate worldwide. Every healthcare professional managing early pregnancy complications should maintain a heightened awareness of the possibility of a CSP. Without a high index of suspicion and correct early diagnosis, like other ectopic pregnancies, this abnormal implantation can lead to uterine rupture and/or hysterectomy, with consequent maternal morbidity and loss of future fertility. Furthermore, if the location of such an implantation is misdiagnosed, a common gynaecological operation such as dilatation and curettage can be met with catastrophic haemorrhage.6,54 Ultrasonographers should follow strict imaging criteria to assess any pregnancy implanted on a previous caesarean section scar. Trainee doctors should be aware of the potential dangers of a CSP in the early pregnancy assessment unit. Senior clinicians must be involved in patient counselling and making appropriate management plan and follow up. Service providers must be equipped with the modern diagnostic imaging and operative facilities, with appropriately trained staff. Owing to the lack of reliable data on the best evidence, women should be given information on the various treatment options available to date to enable them to make an informed choice.
Research is needed in the following areas to explore:
• any correlation or association between caesarean section surgical technique and CSP
• impact of time interval between previous caesarean section and CSP
• comparison between treatment options with outcome
• prediction of susceptibility to rupture and its timing.
Embryo implantation in the region of a previous caesarean section scar is a rare but potentially catastrophic complication of a previous caesarean birth. The exponential rise in its incidence during the past 5–6 years may be a true increase in incidence because of rising caesarean section rate worldwide or an apparent increase as a result of more liberal use of TVS in early pregnancy. Little is known about its exact mechanism and natural history. The number and indication of previous caesarean sections do not seem to affect the rate of CSP. Bleeding and/or pain in early pregnancy are the usual presenting symptoms, but approximately one in three women may be asymptomatic. Every woman with a previous caesarean section presenting to the early pregnancy unit should have a routine check of the caesarean section scar appearance. This review shows that most authors adopt a widely accepted set of criteria of ultrasound diagnosis by TVS and colour flow Doppler, which yields a high diagnostic accuracy—this is expected to emerge as a future gold standard.4
Current data indicate that expectant treatment is rarely successful. Even though most authors recommend early intervention with TOP, there is no universal agreement on the best or most preferred treatment modality. It is therefore difficult to decide on the optimal management. Patient counselling and briefing, although vital, may be limited by this lack of reliable data.
This review, we believe, is up to date and extensive to provide data to the clinicians and women to help raise awareness in this contemporary, potentially risky clinical condition. However, these data cannot reliably guide clinicians regarding optimal management and advice in areas such as future risk of recurrence, role of the interval between previous caesarean delivery and the occurrence of CSP, and the effect of caesarean section wound closure technique on CSP.
Research is required in this subject. Setting up regional centres or multicentre collaboration would encourage robust evidence-based studies essential for making recommendations for practice. Until then, one has to rely on ‘good practice points’ based on anecdotal case reports and small case series.