Early first-trimester transvaginal ultrasound is indicated in pregnancy after previous Cesarean delivery: should it be mandatory?

Abstract

Recent advances in imaging techniques, medical genetics and clinical biology have enabled clinicians to screen for various adverse outcomes, such as aneuploidy1-4, pre-eclampsia5-7 and preterm birth8-10, with potentially huge impact on short- and long-term maternal and fetal wellbeing. Indeed, the introduction of such screening techniques in obstetrics has led to significant reductions in the incidence of certain conditions. For example, first-trimester measurement of nuchal translucency is a well-established screening modality for Down syndrome and other aneuploidies, detecting 64–70% of cases11, and enabling evidence-based counseling and clinical management. The sensitivity of detection improved with the introduction of sequential screening using first-trimester ultrasound and first- as well as second-trimester serum markers (90% detection), and subsequently improved even further with the advent of cell-free DNA testing, which allowed screening as early as 10 weeks' gestation, with 99% sensitivity12. These advances in screening have improved our ability to offer prenatal, or immediate neonatal, targeted therapy1-4. To take another example, routine screening for pre-eclampsia, with thorough maternal history-taking, biochemical-marker measurement and uterine artery Doppler assessment, has improved the prediction of this hypertensive disorder, thus facilitating the use of aspirin prophylaxis in high-risk patients, a course of action which, when initiated prior to 16 weeks of gestation, has proved efficacious in the prevention of pre-eclampsia5-7, 13. The increased rate of Cesarean deliveries throughout the last decade is well documented14. Concomitantly, obstetricians and gynecologists across the world have encountered an increasing number of complications in women conceiving and carrying pregnancies after one or more previous Cesarean sections. These complications involve a range of pathologies, of which pathologically adherent placenta, lately termed placenta accreta spectrum (PAS) by a consensus panel of the International Federation of Gynecology and Obstetrics (FIGO), is the most relevant clinically15. It is also well documented that Cesarean delivery is associated with a significant impact on future pregnancy, including an increased risk of both short- and long-term complications, such as profuse overt and intra-abdominal hemorrhage, PAS, uterine rupture and gravid hysterectomy16-20. PAS is also known by several alternative terms (such as morbidly adherent placenta, abnormally invasive placenta and placental adherence disorder), used by different organizations in various countries, as well as by different journals; for the purposes of this Opinion article, we will use the term suggested by FIGO, i.e. PAS15. This pathological entity may have devastating consequences, and presents affected women and their healthcare providers with unique diagnostic and therapeutic challenges. Yet, there is no established guideline for how and when to screen these women. A proposal for early screening for Cesarean section-related morbidity has been documented previously by El-Refaey et al.21. In the present article, we show why such early screening would be beneficial. This Opinion was conceived by clinicians who, in addition to having academic and research interests in the subject, have had the opportunity to care for a significant number of patients with such unusual pregnancies. Drawing on our experience of affected pregnancies at a variety of gestational ages, it has become obvious that early, accurate clinical and sonographic evaluation of a pregnancy after a previous Cesarean delivery can facilitate not only timely counseling, but also, if needed, gestational age-appropriate individualized pregnancy management. Supported by statements and conclusions found in a wide range of publications, we advocate early (specifically, between 5 and 7 weeks' gestation) evaluation of pregnant women with prior Cesarean section, via early confirmatory urine or serum human chorionic gonadotropin (hCG) testing and an early transvaginal ultrasound (TVS) examination. Our focus in this Opinion is the faulty implantation of a fertilized oocyte into a niche or on a uterine scar created by hysterotomy at the time of a previous Cesarean delivery. This type of unusual gestation is, aptly, termed a Cesarean scar pregnancy (CSP); in the literature, terms such as ‘Cesarean ectopic pregnancy’ and ‘isthmic pregnancy’ are also used. Our agreed view is that the correct term should be CSP, since the implantation is in the lowest area of the cavity and, in contrast to true ectopic pregnancies, a CSP may yield a live neonate. The term ‘ectopic’ should be avoided in these cases, since such descriptors may lead to treatments that are inappropriate for this type of pathology. Many clinicians consider CSP to be a rare occurrence, with its reported rate being from 1 in 800 to 1 in 2500 pregnancies20. However, this reported incidence of CSP is at best an estimation and is influenced by many confounding factors, such as population heterogeneity, Cesarean delivery rates and time to diagnosis14, 16. Regardless of the accuracy of the aforementioned incidence, it is important to anticipate that the rising rate of Cesarean deliveries will likely be accompanied by an increase in the incidence of CSP. Whilst evaluation of a scar from a previous Cesarean section has been reported previously, its clinical value has largely been unrecognized16-18, 22. Yet, failure to identify a CSP (with or without PAS) can have catastrophic consequences. In a systematic review, we included 46 cases from 45 individual articles reporting severe complications from CSP; most of these cases were undiagnosed or misdiagnosed, ultimately leading not only to loss of the pregnancy, but also, frequently, to hysterectomy with loss of fertility, due to PAS and uncontrollable hemorrhage16. One study reported a rate of up to 36% for misdiagnosis of CSP as an early miscarriage in the setting of painful vaginal bleeding, elevated hCG and presence of a cystic mass in the lower uterine segment23. A search of the literature for cases with significant complications resulting from management of both undiagnosed and diagnosed CSPs revealed an increasing number of communications in recent years. The most relevant articles, published between 1978 and the present, are summarized in Table 124-40. In light of such a high rate of misdiagnosis or treatment failure and the lack of a standardized approach to early screening for these abnormal pregnancies, it is all the more concerning that, if an undiagnosed or misdiagnosed CSP is allowed to continue, it can result in loss of the uterus and sometimes even in maternal death25, 33. The early detection of a CSP facilitates early assessment for invasive placentation, which is potentially lethal34. Fortunately, two easily applicable methods are available to facilitate the diagnosis of CSP and differentiate it from a correctly implanted intrauterine pregnancy41-43. These use evaluation of sagittal ultrasound images to determine the location of the center of the gestational sac relative to transverse43 and/or longitudinal41,42 lines that are placed specifically on the image to mark the lower and/or anterior halves of the uterus21, 22. These methods work best if applied in the early first trimester, since after that point, the growing gestational sac begins to fill the uterine cavity, distorting the location of initial implantation, although the placenta and associated vessels remain in place; such assessment has up to 98.9% specificity and 93% sensitivity when used between 5 and 10 weeks' gestation41-43. Using these simple sonographic methods, early diagnosis and evaluation of CSP may provide clues to the natural progression of, and/or best management approaches to, these unusual gestations. For example, it has been demonstrated that for CSP cases without detectable fetal cardiac activity, the majority (approximately 69%) result in spontaneous uncomplicated miscarriage; only about 31% of these cases required intervention44. In contrast, a CSP with a fetal heart beat detected in the first trimester has a very different prognosis. These pregnancies are at high risk for severe bleeding and clinical symptoms requiring first-trimester surgical and/or medical intervention: in a review of the literature, uncomplicated miscarriage occurred in only about 13% of such women, while 20% required surgical management, and in the first and second trimesters of pregnancy, rupture of the uterus occurred in 9.9% of cases and hysterectomy was necessary in 25.2%. Furthermore, about a third of such CSPs progressing to the third trimester were complicated by hemorrhage and, at the time of delivery, 75% had pathologic diagnosis of PAS, mostly placenta percreta44. The use of sonography early in pregnancy to diagnose CSP is not a new concept. Comstock et al.45 stated as early as 2003 that their ‘results suggest that sonographic findings can also be used to identify patients at risk for placenta accreta as early as the first trimester of pregnancy’. In the same year, Jurkovic et al.46 wrote: ‘…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…’. Furthermore, Zosmer et al.47 reported on 10 pregnancies implanted in or on a Cesarean scar, diagnosed in the first trimester and followed to near term. The median age at diagnosis of CSP was 7 weeks and 5 days, though the diagnosis was already suspected at 5 or 6 weeks in at least four cases. They commented on sonographic indicators for prognosis and the natural history of these gestations as follows: ‘Complete implantation within the myometrial defect, bulging of the trophoblast out of the uterine contour and large placental lakes in the first trimester are ultrasound findings that may predict a severe placenta accreta or percreta and consequently a poor outcome.’. Rarely is first-trimester diagnosis of PAS reported, since there is no consensus on which early marker (or combination of markers) for PAS should be considered truly predictive early in pregnancy48, 49. However, several studies have illustrated that low anterior sac position within the niche (or on the hysterotomy scar), thin (or no) myometrial thickness, early appearance of placental lacunae and hypervascularity at the uterus–bladder interface are the most reliable sonographic features at 11–14 weeks associated with subsequent diagnosis of PAS50. A recently published assessment of a two-stage screening and diagnostic protocol for CSP also demonstrated high accuracy for sonographic prediction of PAS in the first trimester (false-positive rate of 0.1%); in that study51, all cases of PAS were associated, at the time of diagnostic assessment as early as 12 weeks in a PAS-specialized clinic, with the sonographic features: intraplacental lacunae, retroplacental arterial-trophoblastic blood flow and irregular placental vascularization on three-dimensional power Doppler. These two studies50, 51 describe ‘classic’ ultrasound findings identified in the late first to early second trimesters for accurately characterizing PAS. We argue that, not only can predictions about PAS be made even sooner (in the early first trimester), but also, such evaluations should be performed routinely to improve patient outcome and expand management options. The same publications which have described several sonographic characteristics for accurate late first-trimester diagnosis of PAS have also demonstrated that cases which go on to develop PAS exhibit first-trimester sonographic characteristics similar to those of CSP50. Additionally, there is a well-documented histopathological connection between CSP and PAS, suggesting not only that these two entities are related fundamentally, but also that CSP is very frequently a precursor to PAS52, 53. Zosmer et al.47 studied several cases of women opting to manage expectantly their CSP, and described the natural course of these gestations: ‘Placenta previa and varying degrees of [morbidly adherent placenta] develop in all patients who continue the pregnancy.’. The presence of previa in the first trimester is rarely reported because it usually resolves in later trimesters; Hill et al.54 found that only 4.8% of cases of placenta previa detected in the first trimester persisted to term. However, previa in the setting of CSP is an entirely different scenario and frequently these abnormally implanted placentae not only fail to resolve, but also hint at underlying invasion51. Furthermore, one of the hallmark diagnostic findings for CSP, namely low anterior gestational sac position between 5 and 7 weeks, correlated 100% of the time with development of PAS41-43. Given this well-documented connection, we argue that CSP, which we know can be diagnosed accurately in the early first trimester, can itself serve as the sought-after early-pregnancy sonographic marker for predicting PAS. While not all CSP cases are ultimately diagnosed with PAS and labeling all CSP cases as ‘PAS until proven otherwise’ might result in false positives, it is important to remember that CSP alone is a fundamentally high-risk pregnancy; this risk is only compounded by the additional diagnosis of invasive placenta. Furthermore, the risk of not diagnosing CSP and/or PAS, thereby missing an opportunity for low-risk management due to a delay in diagnosis, arguably holds far graver consequences for patients than does a false-positive diagnosis in the early first trimester17. To quote El-Refaey et al.21, ‘All women with prior caesarean section should have a 6- to 8-week scan to predict placenta accreta.’. This concept was later developed in a systematic review by Gonzalez and Tulandi55, who stated: ‘Because early diagnosis and treatment are important for the best outcome, every pregnant woman with history of a cesarean delivery should be screened early in the first trimester of pregnancy to rule out this life-threatening complication. Diagnosis can be achieved with ultrasound and Doppler imaging.’. We believe that early diagnosis is paramount to improved patient outcome, and a component of this improvement is the opportunity for optimal management from earlier in gestation. CSP complications and untoward developments are directly correlated with the gestational age at which the CSP is diagnosed and treated, with later diagnosis carrying a higher risk of complications. For example, a study of excessive blood loss during suction evacuation treatment of CSP demonstrated a significant increase in risk of maternal hemorrhage with increasing gestational age, specifically greater than 8 weeks56. These temporal trends prompt us to advocate a TVS scan as early as possible during the first trimester (i.e. at 5–7 weeks), as ultrasound examination during this time span can reveal the location of the early gestational sac, the most reliable sign of a CSP, and potential placental implantation abnormalities. Figure 1 depicts the TVS markers of early first-trimester CSP. Figure 2 shows the ultrasound markers from 8 to 10 weeks. If, following evidence-based counseling, a patient makes the personal decision to terminate a CSP, the necessary procedure for terminating the pregnancy is time-sensitive. Thus, early and accurate diagnosis is paramount to facilitating timely management. At 5–6 weeks' gestation, a local intragestational methotrexate (MTX) or potassium chloride injection with adjuvant systemic administration of MTX or a single-/double-balloon treatment is a simple and effective management protocol for terminating early CSP57-59. Applied between 5 and 8 weeks, the minimally invasive double-balloon procedure has a low complication rate. Even a suction evacuation at 5–7 weeks can be performed relatively safely58. However, with every passing day that the accurate diagnosis and/or decision to terminate is delayed, the gestational sac, the embryo and, most importantly, its vascularization, are growing rapidly. Within days, what might have been a simple, uncomplicated termination can evolve into a much more complex procedure requiring more resources and greater expertise, and associated with higher risk of morbidity for the mother57. A difference of just a couple of weeks (for example, 8–10 weeks compared with 5–7 weeks) increases the complexity of a termination, usually necessitating an operating suite, an anesthesiologist and an expert gynecologic surgeon. It is clearly documented in the literature that termination of CSP at or after 9 weeks decreases the success rate of initial treatment and calls for more involved procedures, such as hysteroscopy, laparoscopy, uterine artery embolization, and/or gravid hysterectomy, all of which by their very nature carry higher risk for complications and even loss of fertility (in cases of hysterectomy)57-60. Early and accurate diagnosis is equally important if, after extensive evidence-based counseling, a patient makes the decision to continue a CSP rather than terminating the pregnancy. In this instance, obstetricians may then often seek the involvement of a maternal-fetal medicine specialist and refer the patient to a hospital center that can provide advanced ultrasound/magnetic resonance imaging, as well as experienced anesthesiologists, skilled operating teams, blood-bank services and intensive neonatal care (in case of premature delivery)50, 59. Such centers are better equipped to manage PAS if this becomes evident. A recent systematic review61 analyzed 13 studies investigating the influence of prenatal diagnosis of PAS on maternal outcome, finding that: ‘Women with a prenatal diagnosis of [abnormally invasive placenta] had less blood loss during surgery ([mean difference], –0.87 L; 95% CI, –1.5 to –0.23 L; P = 0.008). Likewise, the units of [red blood cells] ([mean difference], –1.45; 95% CI, –2.9 to –0.04; P = 0.04) and [fresh frozen plasma] ([mean difference], –1.73; 95% CI, –3.3 to –0.2; P = 0.03) transfused were fewer in women with a prenatal diagnosis of [abnormally invasive placenta] compared with in those with an intrapartum diagnosis…’. It is clear from several published cases and clinical experience that the unpleasant surprise of an undiagnosed invasive placenta at the time of a planned or, even worse, emergency Cesarean section (or other surgical intervention) can have disastrous consequences; for instance, Sivasankar62 described a case of an elective repeat Cesarean section complicated by undiagnosed PAS invading the bladder, which necessitated hysterectomy, bladder repair and massive transfusion. While this may be an uncommon occurrence, most obstetricians can attest to such deliveries haunting them for the duration of their careers. It is fortunate, therefore, that such a scenario can be mitigated, or perhaps avoided entirely, by implementation of routine early first-trimester ultrasound examination after a previous Cesarean delivery. According to the World Health Organization (WHO)63: ‘The CCI Conference on Preventive Aspects of Chronic Disease, held in 1951, defined screening as “the presumptive identification of unrecognized disease or defect by the application of tests, examinations or other procedures which can be applied rapidly.” ’. The WHO suggests various principles to guide the planning of a screening test. The condition for which screening is being considered should be an important health problem with a suitable test or examination that is acceptable to the population, as well as an accepted treatment, already available. Logistically, diagnosis of the condition and administration of the treatment should be possible, including the necessary facilities to enable this. There should be a recognizable latent or early symptomatic stage of the disease, and its natural history, including development from latent to declared disease, should be understood sufficiently. There also needs to be an agreed policy on whom to treat, which takes into consideration the cost: the total cost of screening, including diagnosis and treatment, should be balanced economically in relation to medical expenditure as a whole. Finally, screening should be a continuous process, not just a ‘once and for all’ project. Although the incidence of CSP is significantly lower than that of other obstetric conditions, such as trisomy 21 or pre-eclampsia, for which the importance of first-trimester screening is widely recognized, it represents a life-threatening condition with potentially devastating consequences. Identification of cases at risk of developing CSP or CSP-related conditions would be fundamental to reducing the high burden of maternal mortality and morbidity associated with them. Furthermore, although not yet elucidated completely, the natural history of CSP is becoming clearer. Many CSPs present with acute and unexpected life-threatening symptoms, such as massive hemorrhage or uterine rupture, which commonly lead to emergency intervention, while those progressing beyond the first trimester evolve towards severe PAS disorders44. Prompt identification of the CSP would allow preplanned treatment of the condition, with improved maternal outcome. Finally, although a large variety of elective treatment options for CSP have been reported in the published literature, termination of pregnancy in cases presenting with positive heart beat currently represents the most reasonable option, especially because there is no reliable ultrasound sign with which to identify women who will experience severe symptoms in the first trimester of pregnancy. Conversely, close monitoring is the optimal approach in cases presenting with no heart beat. Ultrasound represents a reliable and well-accepted technique which has been shown to identify accurately CSP and CSP-related adverse effects in the early first trimester. Ultrasound assessment of the uterine scar in women with prior Cesarean section therefore adheres to many of WHO's guidelines required of a screening test44, 57. Several considerations regarding the actual feasibility and long-term impact on maternal health of a screening program for CSP should be raised. Early first-trimester (5–7 weeks of gestation) ultrasound assessment is not undertaken routinely in the majority of developed countries unless there are specific coexisting indications, such as conception by assisted reproductive techniques or previous ectopic pregnancy or CSP, and many professional medical societies advise that the first ultrasound assessment in pregnancy should be performed between 11 and 14 weeks of gestation. Screening for CSP would therefore require the introduction of an additional ultrasound assessment, and its effect on costs for national health services should be balanced against its efficacy in improving the outcome of these women. Furthermore, although reliable in expert hands, ultrasound assessment of the scar after Cesarean section requires standardization with respect to the modality and timing of assessment. Although prenatal diagnosis of CSP is well established, there is still debate about how to identify and counsel those cases not presenting with the classic signs of CSP but showing implantation of the gestational sac between a prior Cesarean section scar and the uterine fundus. The introduction of ultrasound screening for CSP would pose an ethical dilemma for parents and healthcare providers because prenatal imaging cannot completely predict clinical outcome when CSP is detected in asymptomatic women. Although evolution towards acute first-trimester symptoms, such as hemorrhage or uterine rupture, represent the most likely clinical scenario for cases presenting with positive heart beat, the possibility of progression towards severe but potentially treatable conditions, such as PAS, creates the dilemma of how to counsel these women and whether termination of pregnancy should represent the only treatment option offered44. Future researches should focus on developing more accurate prediction models that are able to differentiate women at higher risk of first-trimester symptoms from those in which progression to viability, although potentially complicated by PAS disorders, is more likely. Finally, it should be mentioned that, as a result of the efforts of one of the authors of this Opinion (G.C.), very early first-trimester sonography of pregnant patients with a history of previous Cesarean delivery has been incorporated into obstetrics and gynecology practice guidelines in Italy64; unfortunately, however, this has not yet been incorporated into routine practice worldwide. Currently, there is no widely established guideline for the timing of initiating care and detailed sonographic screening for Cesarean-section related morbidities, such as CSP and PAS. We face the challenge of convincing not only the governing medical community, but also patients, about the importance of modifying pre-existing conventions in early post-Cesarean prenatal care. Here, we have demonstrated that ultrasound is a useful, accurate and reliable tool for raising the suspicion of first-trimester CSP and presumed PAS in the form of a low, anterior gestational sac/embryo/fetus and placenta previa in patients with one or more prior Cesarean deliveries. Furthermore, over the last two decades, we have seen that new screening recommendations for adverse outcomes such as aneuploidy and pre-eclampsia have facilitated earlier intervention and reductions in their incidence. Why then should we not consider an earlier screening method for CSP and PAS, conditions that are potentially life-threatening? Considering the grave potential for morbidity and even mortality in these patients, early first-trimester (i.e. at 5–7 weeks) sonography of pregnant patients with a history of Cesarean section would be an addition to routine prenatal care that would be easy to implement – and indeed should be implemented as soon as possible. Such a shift in the timing of initiation of care would be both prudent and clinically advantageous, not only for maximizing the management options available to patients, but also for optimizing their outcome.