Outcome of Cesarean scar pregnancy managed expectantly: systematic review and meta-analysis
ABSTRACT
enObjective
To explore the outcome in women managed expectantly following the diagnosis of Cesarean scar pregnancy (CSP).
Methods
An electronic search of MEDLINE, EMBASE and ClinicalTrials.gov databases was performed utilizing combinations of relevant medical subject headings for ‘Cesarean scar pregnancy’ and ‘outcome’. Reference lists of relevant articles and reviews were hand-searched for additional reports. Observed outcomes included: severe first-trimester vaginal bleeding; clinical symptoms (abdominal pain, vaginal bleeding) requiring treatment; uncomplicated miscarriage; complicated miscarriage requiring intervention; first- or second-trimester uterine rupture or hysterectomy; third-trimester bleeding, uterine rupture or hysterectomy; maternal death; incidence of abnormally invasive placenta (AIP); prevalence of placenta percreta; ultrasound signs suggestive of AIP; and live birth. Meta-analyses of proportions using a random-effects model were used to combine data. Cases were stratified based on the presence or absence of embryonic/fetal heart activity at the time of diagnosis.
Results
A total of 17 studies (69 cases of CSP managed expectantly, 52 with and 17 without embryonic/fetal heart beat) were included. In women with CSP and embryonic/fetal heart activity, 13.0% (95% CI, 3.8–26.7%) experienced an uncomplicated miscarriage, while 20.0% (95% CI, 7.1–37.4%) required medical intervention. Uterine rupture during the first or second trimester of pregnancy occurred in 9.9% (95% CI, 2.9–20.4%) of cases, while hysterectomy was required in 15.2% (95% CI, 3.6–32.8%) of all cases. Forty (76.9% (95% CI, 65.4–86.5%)) women progressed to the third trimester of pregnancy, of whom 39.2% (95% CI, 15.4–66.2%) experienced severe bleeding. Finally, 74.8% (95% CI, 52.0–92.1%) had a surgical or pathological diagnosis of AIP at delivery and around two-thirds (69.7% (95% CI, 42.8–90.1%)) of them had placenta percreta. In women with CSP but no embryonic/fetal cardiac activity, an uncomplicated miscarriage occurred in 69.1% (95% CI, 47.4–87.1%) of cases, while surgical or medical intervention during or immediately after miscarriage was required in 30.9% (95% CI, 12.9–52.6%). Uterine rupture during the first trimester of pregnancy occurred in 13.4% (95% CI, 2.7–30.3%) of cases, but hysterectomy was not required in any case.
Conclusions
CSP with positive embryonic/fetal heart activity managed expectantly is associated with a high burden of maternal morbidity including severe hemorrhage, early uterine rupture, hysterectomy and severe AIP. Despite this, a significant proportion of pregnancies complicated by CSP may progress to, or close to, term, thus questioning whether termination of pregnancy should be the only therapeutic option offered to these women. Expectant management of CSP with no cardiac activity may be a reasonable option in view of the low likelihood of maternal complications requiring intervention, although close surveillance is advisable to avoid adverse maternal outcome. Copyright © 2017 ISUOG. Published by John Wiley & Sons Ltd.
Resumen
esResultado del embarazo sobre cicatriz de cesárea tratado de forma expectante: revisión sistemática y metaanálisis
Objetivo
Investigar el resultado en mujeres tratadas de forma expectante después del diagnóstico de embarazo sobre cicatriz de cesárea (CSP, por sus siglas en inglés).
Métodos
Se realizó una búsqueda electrónica en las bases de datos MEDLINE, EMBASE y ClinicalTrials.gov utilizando combinaciones de encabezados de temas médicos relevantes para ‘embarazo sobre cicatriz de cesárea’ y ‘resultado’. Para encontrar más informes se realizó una búsqueda manual en la bibliografía de cada artículo. Los resultados observados incluyeron: hemorragia vaginal grave en el primer trimestre; síntomas clínicos (dolor abdominal, hemorragia vaginal) que requirieron tratamiento; aborto sin complicaciones; aborto con complicaciones que requirieron intervención; ruptura uterina o histerectomía en el primer o segundo trimestre; hemorragia, ruptura uterina o histerectomía en el tercer trimestre; muerte materna; prevalencia de placenta invasiva (AIP, por sus siglas en inglés); prevalencia de placenta percreta; indicios de ultrasonido que sugieren AIP; y nacimiento vivo. Para combinar los datos se utilizó un metaanálisis de proporciones con un modelo de efectos aleatorios. Los casos se estratificaron en función de la presencia o ausencia de actividad cardíaca del embrión o del feto en el momento del diagnóstico.
Resultados
Se incluyeron un total de 17 estudios (69 casos de CSP tratados de forma expectante, 52 con presencia de latido del embrión/feto y 17 con ausencia de este). En mujeres con CSP y actividad cardíaca del embrión/feto, el 13,0% (IC 95%, 3,8–26,7%) experimentaron un aborto espontáneo sin complicaciones, mientras que el 20,0% (IC 95%, 7,1-37,4%) requirieron intervención médica. La rotura uterina durante el primer o segundo trimestre del embarazo ocurrió en un 9,9% (IC 95%, 2,9–20,4%) de los casos, mientras que en un 15,2% (IC 95%, 3,6–32,8%) de todos los casos se requirió histerectomía. Cuarenta mujeres (76,9% (95% CI, 65,4–86,5%)) llegaron al tercer trimestre del embarazo, de las cuales el 39,2% (IC 95%, 15,4–66,2%) experimentaron una hemorragia severa. Finalmente, el 74,8% (IC 95%, 52,0–92,1%) recibió un diagnóstico quirúrgico o patológico de AIP en el momento del parto y alrededor de dos tercios (69,7% (IC 95%, 42,8–90,1%)) tenían placenta percreta. En mujeres con CSP pero sin actividad cardíaca del embrión o del feto, en el 69,1% (95% CI, 47,4–87,1%) de los casos se produjo un aborto espontáneo sin complicaciones, mientras que en un 30,9% se requirió intervención quirúrgica o médica durante o inmediatamente después del aborto (95% CI, 12,9–52,6%). La rotura uterina durante el primer trimestre del embarazo ocurrió en un 13,4% (IC 95%, 2.7–30.3%) de los casos, pero en ningún caso se requirió histerectomía.
Conclusiones
La CSP con actividad cardíaca positiva del embrión o del feto tratada de forma expectante se asocia con una alta carga de morbilidad materna que incluye hemorragia grave, ruptura uterina temprana, histerectomía y AIP grave. A pesar de esto, una proporción significativa de embarazos complicados por CSP pueden llegar a término, o cerca de este, lo que cuestiona si la terminación del embarazo debería ser la única opción terapéutica ofrecida a estas mujeres. El tratamiento de forma expectante de la CSP sin actividad cardíaca puede ser una opción razonable, en vista de la baja probabilidad de complicaciones maternas que requieren intervención, aunque se recomienda una intensa vigilancia para evitar resultados maternos adversos.
摘要
zh剖宫产瘢痕妊娠期待治疗的结局:系统综述和meta分析
目的
探讨诊断为剖宫产瘢痕妊娠(Cesarean scar pregnancy,CSP)妇女经期待治疗后的结局。
方法
在MEDLINE、EMBASE和ClinicalTrials.gov数据库中进行“Cesarean scar pregnancy”和“outcome”相关医学主题词组合检索。手式检索相关文章及综述中的参考文献以发现其他研究。纳入的观察结局包括:严重的孕早期阴道出血;需治疗的临床症状(腹痛、阴道出血);单纯流产;需要干预的合并并发症的流产;孕早中期子宫破裂或子宫切除术;孕晚期出血、子宫破裂或子宫切除术;孕产妇死亡;异常侵入性胎盘 (abnormally invasive placenta,AIP)发生率;胎盘植入发生率;提示AIP的超声征象;活产。采用随机效应模型进行比例的meta分析以整合数据。基于诊断时是否存在胎芽或胎心活动对于病例进行分层。
结果
共计纳入17 项研究(69例接受期待治疗的CSP,52例有胎芽或胎心搏动,17例无)。在有胎芽或胎心活动的CSP妇女中,13.0%(95% CI,3.8%~26.7%)患者出现单纯流产,而20.0%(95% CI,7.1%~37.4%)患者需要药物干预。在孕早中期子宫破裂发生率为9.9%(95% CI,2.9%~20.4%),而15.2%(95% CI,3.6%~32.8%)患者需要行子宫切除术。40例进入孕晚期的妇女[76.9%(95% CI,65.4%~86.5%)],其中39.2%(95% CI,15.4%~66.2%)发生严重出血。最后,74.8%(95% CI,52.0%~92.1%)在分娩时手术或病理诊断AIP,约三分之二[69.7%(95% CI,42.8%~90.1%)]患者发生胎盘植入。在无胎芽或胎心活动的CSP妇女中,单纯流产发生率为69.1%(95% CI,47.4%~87.1%),而其中30.9%(95% CI,12.9%~52.6%)流产时或流产后立即需要手术或药物干预。孕早期子宫破裂发生率为13.4%(95% CI,2.7%~30.3%),而无病例需行子宫切除术。
结论
有胎芽或胎心活动的CSP妇女接受期待治疗与更高的孕产妇死亡率相关,包括严重出血、早期子宫破裂、子宫切除术以及严重AIP。尽管如此,仍有相当大比例的合并CSP的妊娠妇女几近分娩,因此质疑中断妊娠是否为此类妇女的唯一治疗选择。尽管建议对于无胎心活动的CSP妇女进行密切监测以避免孕产妇的不良结局,但鉴于需要干预的孕产妇并发症可能性较低,故接受期待治疗仍为此类妇女的理想选择。
INTRODUCTION
The considerable rise in the rate of Cesarean delivery (CD) over the last two decades has led to an increase in the occurrence of short- and long-term complications associated with CD, such as hysterectomy, postpartum hemorrhage, uterine rupture, abnormally invasive placenta (AIP), ectopic pregnancy and infertility1. A peculiar and rare complication that can occur in a subsequent pregnancy after CD is the implantation of the gestational sac in the hysterotomy scar, known as Cesarean scar pregnancy (CSP)2.
If left untreated, CSP may lead to severe hemorrhage, uterine rupture and the need for hysterectomy2, 3. The prenatal diagnosis of CSP is based on the presence of a gestational sac at the site of the previous uterine incision and the presence of an empty uterine cavity and cervix, as well as thin myometrium adjacent to the bladder4. However, prenatal diagnosis of CSP is not always achieved, many cases being misdiagnosed as threatened or missed/incomplete miscarriage or simply intrauterine pregnancy. Such a misdiagnosis may lead to sharp curettage for presumed failed pregnancy, which can result in profuse bleeding and emergency surgical intervention, on occasion ending with hysterectomy4.
A multitude of treatment modalities have been proposed for the management of CSP, but the optimal approach in terms of patient safety and clinical effectiveness has yet to be determined3. Despite this uncertainty, it is common clinical practice to offer termination of pregnancy in view of the reported high rate of morbidity associated with this condition3. The recently proposed association between CSP and AIP complicates the counseling of women with prenatally diagnosed CSP and raises the dilemma of whether termination of pregnancy should be the only therapeutic option offered5, 6. Although evolution towards early severe hemorrhage or uterine rupture may seem the most probable scenario, the reported association with AIP raises questions about the proportion of CSPs progressing to viable birth, and how to identify cases that will have a successful pregnancy outcome and be amenable to postnatal treatment. The majority of published series on CSP include mainly symptomatic cases or those undergoing emergency treatment, thus making it difficult to estimate the actual burden of maternal morbidity associated with this condition.
The aim of this systematic review was to explore the outcome of women with a diagnosis of CSP undergoing expectant management.
METHODS
This systematic review was performed according to a protocol designed a priori and recommended for systematic reviews7, 8. An electronic search of MEDLINE, EMBASE and ClinicalTrials.gov databases was performed on 23 February 2017, utilizing combinations of the relevant medical subject heading (MeSH) terms, keywords and word variants for ‘Cesarean scar pregnancy’ and ‘outcome’. Reference lists of relevant articles and reviews were hand-searched for additional reports. Before data extraction, the review was registered with the PROSPERO International Prospective Register of Systematic Reviews (Registration No: CRD42017064596) following the PRISMA guidelines for protocols (PRISMA-P)9, 10.
Study selection, data collection and data items
Observed outcomes included: severe first-trimester vaginal bleeding; clinical symptoms, such as abdominal pain, vaginal bleeding or both; symptoms requiring either surgical or medical intervention; uncomplicated miscarriage that resolved spontaneously; complicated miscarriage requiring medical or surgical intervention; first- or second-trimester uterine rupture or hysterectomy; third-trimester bleeding, uterine rupture or hysterectomy; maternal death; incidence of AIP; prevalence of placenta percreta; ultrasound signs suggestive of AIP; and live birth.
Only studies including women diagnosed with CSP by ultrasound who were managed expectantly were considered suitable for inclusion in the present systematic review. Furthermore, the analysis was stratified according to the presence or absence of embryonic/fetal heart activity at the time of diagnosis.
CSP was defined as the gestational sac or trophoblast presenting within the dehiscence/niche of the previous Cesarean section scar or implanted on top of it. CSP was diagnosed according to the following criteria11: (1) visualization of an empty uterine cavity; (2) detection of the placenta and/or gestational sac embedded in a hysterotomy scar; (3) triangular gestational sac filling the niche of the scar; (4) thin (1–3 mm) or absent myometrial layer between the gestational sac and the bladder; (5) closed cervix and empty endocervical canal; (6) presence of an embryonic/fetal pole and/or yolk sac with or without heart activity; (7) presence of a prominent, and at times rich, vascular pattern at or around the chorionic sac and placenta.
All abstracts identified through the search were reviewed independently by two authors (F.D., G.C.). Agreement regarding potential relevance was reached by consensus with a third reviewer (D.B.). Full-text copies of articles deemed relevant were obtained and the same two reviewers independently extracted data regarding study characteristics and pregnancy outcomes. Inconsistencies were discussed and consensus was reached or the dispute was resolved by discussion with another author. If more than one study was published for the same cohort with identical endpoints, the report containing the most comprehensive information on the population was included to avoid overlapping populations.
Only full-text articles were considered eligible for inclusion. Conference abstracts and single-case reports were excluded to avoid publication bias. Studies published before 2000 were not included, as we considered that advances in prenatal imaging techniques and improvements in the diagnosis and management of CSP make these less relevant. Studies reporting only cases of CSP undergoing treatment were excluded. Finally, studies not providing a clear classification of the anomaly were not considered suitable for inclusion in the current review.
Quality assessment of the included studies was performed using the Newcastle–Ottawa Scale (NOS)12. According to the NOS, each study is judged on three broad perspectives: (a) selection of the study groups, which includes evaluation of the representativeness of the exposed cohort, selection of the non-exposed cohort, ascertainment of exposure and demonstration that the outcome of interest was not present at the start of the study; (b) comparability of the groups, which includes evaluation of the comparability of cohorts on the basis of the design or analysis; and (c) ascertainment of the outcome of interest, which involves evaluation of the type of assessment of the outcome of interest and length and adequacy of follow-up. According to the NOS, a study can be awarded a maximum of one star for each numbered item within the selection and outcome categories. A maximum of two stars can be given for comparability12.
Statistical analysis
For quantification of the incidence of explored outcomes, meta-analyses of proportions using a random-effects model were used to combine data13. Funnel plots displaying the outcome rate from individual studies vs their precision (1/standard error) were carried out with an exploratory aim. Tests for funnel-plot asymmetry were not used when the total number of publications included for each outcome was less than 10; in such cases, the power of the tests is too low to distinguish chance from real asymmetry14, 15. Between-study heterogeneity was explored using the I2 statistic, which represents the percentage of between-study variation that is due to heterogeneity rather than chance16. Statistical analysis was performed using StatsDirect statistical software (StatsDirect Ltd 2013, Cheshire, UK).
RESULTS
Study selection and characteristics
A total of 872 articles were identified from the search, of which 65 were assessed in full text for eligibility for inclusion (Table S1) and 17 were finally included in the systematic review (Table 1, Figure 1)5, 17-32. These studies included a total of 273 women with CSP, of whom 69 were treated expectantly because the parents declined termination of pregnancy, prenatal diagnosis of CSP was missed or the physicians opted for expectant management. The prevalence of cases with and without fetal heart beat detected at diagnosis was 52/69 (75.4% (95% CI, 63.5–84.9%)) and 17/69 (24.6% (95% CI, 15.1–36.5%)), respectively.
| Study | Country | Study design | Period analyzed (years) | GA at diagnosis (weeks) | Total cases of CSP (n) | CSP cases managed expectantly (n) |
|---|---|---|---|---|---|---|
| Gao (2016)17 | China | Retrospective | 2009–2012 | 7.3 (5.4–12) | 9 | 1 |
| Michaels (2015)18 | USA | Retrospective | 2000–2012 | 6.8 ± 1.6 | 34 | 11 |
| Zosmer (2015)19 | UK | Prospective | 2011–2013 | 8 + 2 (7 + 5 to 11 + 6) | 27 | 10 |
| Timor-Tritsch (2015)20 | USA | Retrospective | 2009–2014 | 5 + 4 to 14 | 60 | 13 |
| Ko (2015)21 | China | Retrospective | 2004–2013 | 6.7 (4.7–11.8) | 22 | 4 |
| Riaz (2015)22 | USA | Retrospective | 2008–2015 | 5 + 2 to 10 + 0 | 20 | 5 |
| Cheng (2014)23 | Taiwan | Retrospective | 2000–2012 | 6.8 ± 1.3 | 48 | 2 |
| Timor-Tritsch (2014)5 | USA | Retrospective | NS | 5 + 4 to 9 + 2 | 2 | 2 |
| Calì (2013)24 | Italy | Retrospective | 2004–2012 | 8.3 (7–9) | 5 | 5 |
| Sinha (2012)25 | UK | Retrospective | NS | 6 | 2 | 1 |
| Stirnemann (2011)26 | France | Prospective | 2008–2010 | 11 to 13 + 6 | 6 | 6 |
| Liu (2010)27 | China | Retrospective | NS | 7 + 2 to 5 + 5 | 2 | 2 |
| de Vaate (2010)28 | The Netherlands | Retrospective | NS | 5 + 1 to 6 + 6 | 4 | 1 |
| Yan (2007)29 | China | Retrospective | NS | 7.5 (5–10) | 4 | 1 |
| Maymon (2004)30 | Israel | Retrospective | 1995–2002 | 7 (6–9) | 8 | 2 |
| Jurkovic (2003)31 | UK | Retrospective | NS | 4–23 | 18 | 2 |
| Singh (2002)32 | Australia | Retrospective | NS | 10 | 2 | 1 |
- Only first author of each study is given.
- Data are given as median (range), mean ± SD or median.
- GA, gestational age; NS, not specified.
Quality assessment of the included studies according to NOS for cohort studies is shown in Table 2. Most of the included studies showed an overall good rate in the selection and comparability of the study groups. The main weaknesses of the studies were their retrospective design, small sample size, lack of stratification of the analysis according to levels of beta-human chorionic gonadotropin (β-hCG) and size and position of the gestational sac, and large heterogeneity in gestational age at diagnosis, outcomes observed and management protocols.
| Study | Selection | Comparability | Outcome |
|---|---|---|---|
| Gao (2016)17 | ★★ | ★ | ★★ |
| Michaels (2015)18 | ★★ | ★ | ★★ |
| Zosmer (2015)19 | ★★ | ★ | ★★ |
| Timor-Tritsch (2015)20 | ★★ | ★ | ★★ |
| Ko (2015)21 | ★★ | ★ | ★★ |
| Riaz (2015)22 | ★★ | ★ | ★★ |
| Cheng (2014)23 | ★★ | ★ | ★ |
| Timor-Tritsch (2014)5 | ★★ | ★★ | ★★ |
| Calì (2013)24 | ★★ | ★ | ★★ |
| Sinha (2012)25 | ★★ | ★ | ★ |
| Stirnemann (2011)26 | ★★ | ★ | ★ |
| Liu (2010)27 | ★★★ | ★★ | ★★ |
| de Vaate (2010)28 | ★★ | ★ | ★★ |
| Yan (2007)29 | ★★ | ★ | ★★ |
| Maymon (2004)30 | ★★ | ★ | ★★ |
| Jurkovic (2003)31 | ★★ | ★ | ★★ |
| Singh (2002)32 | ★★ | ★ | ★★ |
- Only first author of each study is given.
- A study can be awarded a maximum of one star for each numbered item within the selection and outcome categories.
- A maximum of two stars can be given for comparability.
CSP with embryonic/fetal heart beat
Fifty-two women with CSP and embryonic/fetal heart activity undergoing expectant management were included in the present systematic review. Severe vaginal bleeding during the first trimester of pregnancy occurred in 12.9% (95% CI, 4.7–24.3%) of women for whom this outcome was reported in the included studies (Table 3, Figure S1). Other clinical symptoms occurred in 37.3% (95% CI, 21.6–54.4%) of women, while 20.0% (95% CI, 7.1–37.4%) of all cases required surgical or medical intervention. 13.0% (95% CI, 3.8–26.7%) of the affected women experienced an uncomplicated miscarriage and did not require any additional intervention, while in 20.0% (95% CI, 7.1–37.4%) of the cases, miscarriage was complicated and treatment was required. Uterine rupture during the first or second trimester of pregnancy occurred in 9.9% (95% CI, 2.9–20.4%) of cases, while hysterectomy was required in 15.2% (95% CI, 3.6–32.8%) of women (Table 3).
| Outcome | Studies (n) | Pregnancies (n/N) | Raw proportion (95% CI) (%) | I2 | Pooled proportion (95% CI) (%) |
|---|---|---|---|---|---|
| CSP with fetal heart beat | |||||
| First/second trimester | |||||
| Severe bleeding | 11 | 4/39 | 10.26 (2.9–24.2) | 0 | 12.90 (4.7–24.3) |
| Other symptoms | 8 | 10/28 | 35.71 (18.6–55.9) | 0.7 | 37.27 (21.6–54.4) |
| Symptoms requiring intervention | 10 | 8/39 | 20.51 (9.3–36.5) | 31.1 | 20.00 (7.1–37.4) |
| Uncomplicated miscarriage | 10 | 4/39 | 10.26 (2.9–24.2) | 18.1 | 13.03 (3.8–26.7) |
| Complicated miscarriage | 10 | 8/39 | 20.51 (9.3–36.5) | 31.1 | 20.00 (7.1–37.4) |
| Uterine rupture | 10 | 3/39 | 7.69 (1.6–20.9) | 0 | 9.91 (2.9–20.4) |
| Hysterectomy | 10 | 6/39 | 15.38 (5.9–30.5) | 39.2 | 15.21 (3.6–32.8) |
| Third trimester | |||||
| Severe bleeding | 8 | 8/27 | 29.63 (13.7–50.2) | 51.6 | 39.23 (15.4–66.2) |
| Uterine rupture | 10 | 2/34 | 5.88 (0.7–19.7) | 0 | 10.20 (2.8–21.6) |
| Hysterectomy | 11 | 23/40 | 57.50 (40.9–73.0) | 58.6 | 60.59 (36.9–81.9) |
| Maternal death | 11 | 0/40 | 0 (0–8.8) | 0 | 0 (0–13.4) |
| Occurrence of AIP | 11 | 30/40 | 75.00 (58.8–87.3) | 58.8 | 74.82 (52.0–92.1) |
| Occurrence of placenta percreta | 8 | 17/29 | 58.62 (38.9–76.5) | 56.3 | 69.67 (42.8–90.1) |
| Ultrasound signs of AIP | 8 | 27/29 | 93.10 (77.2–99.2) | 0 | 90.31 (78.2–97.8) |
| Live birth | 11 | 40/40 | 100 (92.2–100) | 0 | 100 (86.6–100) |
| CSP without fetal heart beat | |||||
| First trimester | |||||
| Severe bleeding | 10 | 3/17 | 17.65 (3.8–43.4) | 0 | 22.16 (7.7–41.4) |
| Other symptoms | 9 | 5/14 | 35.71 (12.8–64.9) | 2.7 | 36.41 (16.3–59.3) |
| Symptoms requiring intervention | 10 | 4/17 | 23.53 (6.8–49.9) | 10.8 | 26.04 (9.4–47.3) |
| Uncomplicated miscarriage | 10 | 12/17 | 70.59 (44.0–89.7) | 9.7 | 69.09 (47.4–87.1) |
| Complicated miscarriage | 10 | 5/17 | 29.41 (10.3–56.0) | 9.7 | 30.91 (12.9–52.6) |
| Uterine rupture | 10 | 1/17 | 5.88 (0.1–28.7) | 0 | 13.36 (2.7–30.3) |
| First-/second-trimester hysterectomy | 10 | 0/17 | 0 (0–19.5) | 0 | 0 (0–25.5) |
| Maternal death | 10 | 0/17 | 0 (0–19.5) | 0 | 0 (0–25.5) |
- AIP, abnormally invasive placenta.
Of the 52 women with CSP and embryonic/fetal heart activity, 40 (76.9% (95% CI, 65.4–86.5%)) women progressed to the third trimester of pregnancy. 39.2% (95% CI, 5.4–66.2%) of women had severe bleeding during the third trimester while uterine rupture occurred in 10.2% (95% CI, 2.8–21.6%) of cases. Of the women with CSP progressing to the third trimester, 74.8% (95% CI, 52.0–92.1%) had a surgical or pathological diagnosis of AIP at delivery and around two–thirds (69.7% (17/29); 95% CI, 42.8–90.1%) of AIP cases were placenta percreta. AIP was diagnosed prenatally in the majority of cases included in the present systematic review (90.3% (27/29); 95% CI, 78.2–97.8%). Overall, hysterectomy during CD was required in 60.6% (23/40; 95% CI, 36.9–81.9%) of women who had CSP. There were no cases of maternal death (Table 3).
CSP without embryonic/fetal heart beat
Seventeen women affected by CSP in which no embryonic/fetal heart activity could be documented were included in this systematic review.
Severe vaginal bleeding occurred in 22.2% (95% CI, 7.7–41.4%) of cases; overall, clinical symptoms affected 36.4% (95% CI, 16.3–59.3%) of women and 26.0% (95% CI, 9.4–47.3%) required surgical or medical intervention (Table 3, Figure S2). The majority of women affected by CSP with no embryonic/fetal cardiac activity experienced an uncomplicated miscarriage (69.1% (95% CI, 47.4–87.1%)), while surgical or medical intervention during or immediately after miscarriage was required in 30.9% (95% CI, 12.9–52.6%) of cases. Uterine rupture during the first trimester of pregnancy occurred in 13.4% (95% CI, 2.7–30.3%) of cases, while there were no cases of first-trimester hysterectomy or maternal death among the cases included in the review.
DISCUSSION
Summary of findings
This systematic review showed that women affected by CSP with positive embryonic/fetal cardiac activity had a high risk of severe bleeding and clinical symptoms requiring surgical or medical intervention during the first trimester of pregnancy. Uncomplicated miscarriage occurred in 13.0% of women, while in 20.0% medical or surgical intervention was required. Uterine rupture occurred in 9.9% of cases and hysterectomy was required in 15.2% during the first or second trimester of pregnancy. About a third of CSPs progressing to the third trimester experienced bleeding, and three–quarters had a surgical or pathological diagnosis of AIP at delivery, mostly placenta percreta.
The large majority (69%) of CSP cases without embryonic/fetal cardiac activity experienced an uncomplicated miscarriage, while surgical or medical intervention during or immediately after miscarriage was required in 31% of cases. The risk of uterine rupture and hysterectomy during the first trimester of pregnancy was negligible in CSP cases without embryonic/fetal cardiac activity.
Strengths and limitations of the study
This is the first meta-analysis assessing the outcome of women with CSP undergoing expectant management. The strengths of this study are its robust methodology for identifying all possible studies for inclusion, assessing data quality and synthesizing all suitable data.
The small number of cases in the majority of the included studies, their retrospective non-randomized design, different follow-up periods and lack of stratification of the analysis according to serum β-hCG levels and size and position of the gestational sac represent the main limitations of this systematic review. In addition, assessment of publication bias was difficult because of the nature of the outcome (the numerator for many explored outcomes being 0), which limits the reliability of funnel plots, and because of the scarce number of individual studies, which strongly limits the reliability of formal tests. The level of evidence for this type of study is very low. Another limitation relates to selection bias in the studies as, in most of the included studies, selection of the patients for expectant management was not performed in a controlled or randomized manner. Therefore, most severe and symptomatic patients are likely to have received non-expectant management.
Nevertheless, this review represents the best published estimate of the occurrence of adverse outcomes in women with CSP undergoing expectant management. This is important, as counseling of parents based on small studies that are subject to publication bias may be inadequate.
Implications for clinical practice
Increasing CD rates over the past two decades and advances in prenatal imaging have led to an increase in the number of patients with CSP requiring obstetric management. Despite the high burden of maternal morbidity associated with this condition, CSP is commonly misdiagnosed. The ultrasound criteria for CSP have been redefined recently, proposing transvaginal ultrasound as the imaging technique of choice at this early gestation and magnetic resonance imaging to be of no significant help11.
CSP is diagnosed in the presence of an early gestational sac and/or placenta in close proximity to the previous hysterotomy scar/niche in a patient with previous CD and a positive pregnancy test. Before 7 weeks' gestation, the gestational sac may assume the shape of the niche in which it implants. Furthermore, an absent or thin myometrial layer between the gestational sac and the bladder wall is usually present in this scenario. At color Doppler examination, blood flow around the gestational sac can usually be seen. After 7 weeks, if the patient continues the pregnancy, the sac slowly ‘moves’ towards the uterine cavity and gradually changes shape and assumes an intracavitary position that may lead to its misdiagnosis as an intrauterine pregnancy. Therefore, after 7 weeks, examiners should use high-resolution grayscale and color Doppler ultrasound to visualize the exact site of placental insertion, since this will stay ‘anchored’ at its original implantation site, providing the correct diagnosis (Figure S3)4.
This systematic review suggests that expectant management may be a reasonable option for CSP with no detectable embryonic/fetal heart activity, as 70% of the included cases did not experience any major maternal complication and had an uncomplicated miscarriage. While there were no cases of uterine rupture or first-trimester hysterectomy, prompt treatment was required in 30% of the cases. Serial assessment of serum β-hCG levels is warranted until normalization. Prenatal imaging can alert care providers to adverse outcomes. For example, an enhanced residual myometrial vascularity may require more aggressive treatment, such as uterine artery embolization, but it is not possible to completely rule out all short-term complications (Figure 2)33.
Conversely, CSP with embryonic/fetal heart activity was associated with a high burden of first- and second-trimester complications if left untreated. About 20% of women required surgical or medical intervention during the first trimester, 10% experienced early uterine rupture and 15% required a hysterectomy. Of the pregnancies progressing to the third trimester, about three-quarters developed AIP, mainly placenta percreta. For these cases, the optimal management is still to be defined and women should be counseled about short- and long-term risks. Assessment of the location of the gestational sac and residual myometrial thickness should be performed to identify the risk of AIP because it may help in stratifying women at higher risk of complications (Figure 2). Finally, delivery in a center with expertise in the surgical management of severe AIP should be arranged.
Unfortunately, in the present review it was not possible to stratify the analysis according to serum β-hCG levels, number and shape of previous CD scar(s) and relationship between the ectopic gestational sac and myometrium, thus it is not entirely certain whether all CSPs with positive embryonic/fetal heart activity are invariably associated with an adverse outcome or whether there is a subset of CSPs with better short- and long-term outcome despite the presence of a heart beat at the time of diagnosis.
The findings of the present review also confirmed the proposed association between CSP and AIP. Identification of women affected by CSP who are at higher risk of adverse outcome is an important issue. Several studies that have attempted to stratify the risk of AIP in women with a CSP have been published recently. Kaelin Agten et al.34 showed that CSP implanted ‘on the scar’, defined as a placenta implanted partially or fully on top of a well-healed scar, had a substantially better outcome compared with patients in which the CSP implanted into a deficient or dehiscent scar.
Calì et al.35, 36 showed that the relationship between the gestational sac of the CSP, previous Cesarean scar and the anterior uterine wall thickness can be used to predict not only the evolution of the CSP towards the most severe types of AIP, but also the clinical outcome of these women. However, there is still a need for an adequately powered study assessing the role of ultrasound in the prediction of short- and long-term complications associated with CSP35, 36.
The optimal treatment of CSP is still to be determined. The majority of published studies include only symptomatic or unstable patients and currently there has been no randomized trial comparing the different treatment options, thus making it difficult to extrapolate any objective evidence regarding the best approach3.
Large prospective studies are needed in order to elucidate the actual burden of maternal morbidity in women with CSP undergoing expectant management and to ascertain whether prenatal imaging can identify women with CSP at higher risk of developing complications.
