Is there an Increased Risk for Unfavorable Obstetric Outcomes in Women with Endometriosis? An Evaluation of Evidences

Data for adverse outcomes were collected as dichotomous data, and the results are presented as odds ratios (ORs) with 95% confidence interval (CI).

The present study was carried out at the Hospital Israelita Albert Einstein, São Paulo, state of São Paulo, Brazil, according to the PRISMA methodology for systematic reviews. To identify relevant articles to be included in the study, a review of the literature was done using the PubMed, Web of Science and Scopus databases. The keywords used were: pregnancy outcome , pregnancy complications , obstetrical complications , obstetrics , obstetric outcomes and endometriosis. The search period was from May 1946 to October 2017. A manual review of cross-referenced articles completed the survey, which was last performed on August 30, 2018.

The search identified a total of 2,468 articles, ranging from May 1946 to October 2017, of which 1,630 were from PubMed, 738 from Scopus and 68 from Web of science, 459 duplicates and 585 revisions or published in other languages. By limiting the research to only English-written articles, and excluding duplicate articles and systematic reviews, we found 1,358 articles. The initial selection was done by reading the titles and abstracts of the articles. All case-control, prospective or retrospective cohort studies evaluating obstetric outcomes in women diagnosed with endometriosis were included. Fifty articles were read in full. No date limit was imposed, and two reviewers (Annicchino G. and Podgaec S.) independently nonblindly assessed the eligibility of the articles following the standardized protocol. Disagreement regarding the inclusion of studies were discussed and, by consensus, the articles were included or excluded. The references of these articles were also searched resulting in finding one additional study. A systematic review was performed analyzing the year of publication, number of patients involved in the study, type of study, and the results and conclusions of each study resulting in the selection of 18 studies ( Fig. 1 ). Flowchart of the method used to select the articles (Prisma). All of the articles within the described theme were included, regardless of the age of the patients, type of pregnancy (single or multiple), gestational age or form of conception (natural or artificial). The diagnosis method of endometriosis was not taken into consideration for exclusion or inclusion purposes; it could be clinical, surgical or histopathological. These 18 studies were classified according to their relevance using the NOS scale (scores of 0–9 stars). Studies with NOS ≥ 4 were regarded as moderate quality and ≥ 8 were regarded as high-quality. According to this evaluation, 5 studies of greater impact and quality of superior evidence and 13 studies of moderate quality of evidence were selected ( Supplementary Material Appendix 1 ). Table 1 exhibits the data of the control groups, how the disease was diagnosed, and other particularities of the studies. Tables 2a , 2b and 2c display the studies in which the obstetric outcomes were studied in relation to endometriosis and their statistical relevance. Abbreviations: DIE, Deep infiltrative endometriosis; FIV, in vitro fertilization. Abbreviation: NR, not relevant. Abbreviation: NR, not relevant. Abbreviation: NR, not relevant. Obstetric outcomes – higher risk in patients with endometriosis.

Studies were selected using the following predetermined inclusion criteria: [i]women who had a diagnosis of endometriosis during or before pregnancy compared to a control group of women without the diagnosis [ii] any outcomes of interest in the present pregnancy, and [iii] observational, cohort or case-controlled human study design that were reported in English. The primary outcomes of the present study were determined previously and included the following adverse obstetric and perinatal outcomes: abortion, ectopic gestation, fetal loss, pre-eclampsia, bleeding during pregnancy, placental retention, placenta previa, premature placental abruption, premature membranes rupture, preterm labor, cesarean section, postpartum hemorrhage, preterm delivery, small for gestational age (SGA) fetus, stillborn neonate and neonatal death. The secondary outcomes were the presence of any other clinically important adverse pregnancy outcomes reported in the literature. Information extracted from each study included: the country where the research was done; the name of the cohort study; duration and sample size; inclusion criteria; definition of reference or control group; endometriosis diagnostic criteria; obstetric or neonatal outcomes; demographics to which the studies were adjusted.

Studies that assessed smaller control groups than those mentioned above also showed a correlation between unfavorable obstetric outcomes and women diagnosed with endometriosis. The oldest of them evaluated 174 women with endometriosis and compared it to the same number of women without diagnosis. 32 The authors examined the possibility of higher rates of abortion in the case group, but did not observe this correlation. Mekaru et al 27 also reached this result after evaluating a group of 108 pregnant women who had previously undergone laparoscopy to investigate infertility. In contrast, Saraswat et al, 19 Turocy et al, 21 Santulli et al, 22 Aris, 26 and Hjordt Hansen et al 31 published results showing increased risk of abortion in women with endometriosis. When evaluating the correlation between transient hypertensive disease during pregnancy and endometriosis, most studies did not report this association, as shown in the studies by Mannini et al, 18 Saraswat et al, 19 Lin et al, 24 Conti et al, 25 Aris, 26 Mekaru et al, 27 Hadfield et al, 29 and Brosens et al. 30 While specifying the obstetric outcome for pre-eclampsia, some results did show statistical significance when related to the diagnosis of endometriosis, as described by Berlac et al, 16 Glavind et al 17 and Brosens et al. 30 But the results reported by Saraswat et al, 19 Conti et al 25 and Aris 26 disagreed as they show negative association between endometriosis and the outcome in question. 9 16 17 18 19 24 25 26 27 29 30 Many authors also evaluated the relationship between pregnant women with endometriosis and placental disorders such as premature rupture of membranes, placenta previa and premature placental abruption. Berlac et al, 16 Conti et al, 25 and Harada et al, 33 who included premature rupture of membranes in the studied outcomes, concluded that women with endometriosis are a risk group to present these pathologies. The investigated placental outcomes were linked to the diagnosis of endometriosis in the studies by Berlac et al, 16 Mannini et al, 18 Saraswat et al, 19 Fujii et al, 23 Lin et al 24 and Vercellini et al 28 and Harada et al, 33 showing a strong correlation among them. Premature placental abruption was identified to be associated with endometriosis in the studies by Berlac et al 16 and Harada et al. 33 However, Lin et al 24 did not find the same association. 16 18 19 23 24 25 28 29 30 31 32 33 34 Many authors place newborn-related outcomes among the unfavorable obstetric results to be evaluated in women with endometriosis. All of the studies that evaluated preterm birth found a positive correlation, as reported by Berlac et al, 16 Glavind et al, 17 Saraswat et al, 19 Stephansson et al, 20 Fujii et al, 23 Jacques et al, 34 Lin et al, 24 and Conti et al. 25 Berlac et al 16 and Conti et al 25 included SGA fetuses, and also obtained a positive correlation with increased risk in women with endometriosis. The delivery route was also evaluated in many studies included in the present systematic review. Glavind et al, 17 Saraswat et al, 19 Jacques et al 34 and Lin et al 24 reported higher risk of cesarean delivery in women diagnosed with endometriosis. However, Mannini et al 18 and Mekaru et al 27 did not report this association. 17 19 24 27 34

We excluded from the analyses the studies that were not prospective or retrospective cohort or case-control, as well as those not written in English or lacking data. The study selection process, full text screening, and data extraction was conducted independently by two researchers (Annicchino G. and Piccinato C. A.), following the PRISMA guidelines. Disagreements were solved after consulting a third opinion (Podgaec S.).

The quality of the included studies was assessed by the Newcastle-Ottawa Quality Assessment Scale (NOS, scores of 0–9 stars) for the selection of study groups (up to 4 stars/points); comparability of groups (up to 2 stars/points); and, the ascertainment of either the exposure or outcome of interest for case-control or cohort studies, respectively (up to 3 stars/points). 15

Endometriosis is a disease that extends beyond the presence of ectopic endometrial implants. The condition of the endometrium can determine the quality of implantation and placental development, influencing obstetric outcomes, especially preterm birth and placenta previa. More studies paying more attention to the quality of the methodology, with adequate experimental designs and without bias such as in vitro fertilization, are necessary. The information gathered is useful to alert obstetricians and women diagnosed with endometriosis about possible unfavorable obstetric outcomes.

The present systematic review highlighted observational studies, some more robust with larger control groups and others with more restricted groups ( Table 1 ). The highest agreement between studies of greater quality of evidence is the high incidence of preterm birth and placenta previa in patients with endometriosis. 16 18 19 Moderate quality studies also showed endometriosis-diagnosed patients to have more abortion occurrences and cesarean deliveries. 21 22 24 25 31 34 Generally, most studies highlight the impact of endometriosis on unfavorable obstetric outcomes, although only three less relevant case-control group studies found no evidence of higher risk ( Table 2 ). 27 29 32 The causes of higher risk for obstetric complications have not yet been defined, and the underlying pathophysiological factors are still unclear. 35 Despite this, endomyometrial changes present in patients with endometriosis seem to be responsible for several obstetric adverse factors such as abortion, fetal growth restriction, placenta previa, and preterm delivery or SGA infants. 35 More specifically, in relation to the increased incidence of complications such as preterm birth and placenta previa, we can emphasize alterations in endometrial hormonal receptivity, decidualization and remodeling of the spiral uterine arteries and inflammatory state that alter the regulation of the endocrine immune system in patients with endometriosis. 18 36 The reason why some placentas are implanted in the lower segment of the uterus remains under discussion. As gestation progresses, 90% of low-insertion placentas move towards the uterine fundus. The placenta grows preferably towards the best-vascularized area, which is the uterine fundus (trophotropism), and the placenta that remains in the least vascularized area undergoes atrophy. Uterine contractions lead to detachment of this area of the placenta and subsequent bleeding, which further stimulates uterine contraction. 37 Resistance to progesterone and inadequate uterine contractions occurring in women with endometriosis may explain the greater frequency of placenta previa in this subgroup. 18 Endometriosis also leads to a hyperinflammatory state of the endometrium that causes endometrial endocrine immune balance disorder (increase in sex hormones, neurohormones, cytokines and growth factors). This disorder is thought to influence the interaction of decidua/trophoblast and activate the mechanisms of preterm delivery when the imbalance between pro-inflammatory and anti-inflammatory mechanisms of the placenta occurs. 7 These mechanisms found in patients with endometriosis may justify the greater obstetric risks. In the last 20 years, several studies focused on the evaluation of this diversity of obstetric complications, with premature rupture of membranes and cases of placenta previa being the most commonly associated complications in the representative studies. Reviews on the topic report the same aforementioned complications in women diagnosed with endometriosis as those found in the present systematic review. However, in our review, we perceived that among all analyzed studies, there is great statistical heterogeneity and their quality was not taken into account. Therefore, to be able to derive significant conclusions from our study we chose to assign greater importance to the quality of each of the studies. This approach minimized the possible biases of each study and allowed a better analysis of the results. 13 At the same time, there is a concern in the literature regarding the results found in a number of different studies (bias, false positives or false negatives). 38 39 In this respect, systematic reviews are able to extract from the studies information on data quality, sample size, possible biases and methodological description. Despite the volume of studies published with endometriosis in pregnancy, there is the need to prepare large studies, with carefully selected control groups (to avoid bias), based on a working hypothesis compatible with existing results from previous reviews, and focusing on the association and risk between endometriosis and unfavorable obstetric outcomes. Nonetheless, our study also presents limitations in view of those of each of the articles analyzed. Inherent characteristics of many studies, such as methodological flaws, lack of histological confirmation of endometriosis, and small number of patients were a frequent finding. 16 32 Another divergence among the studies was whether infertility and in vitro fertilization were considered as a bias ( Table 1 ). It is important to note that endometriosis and infertility may be independent risk factors, since polymorphisms in genes associated with infertility, regardless of endometriosis, are also related to unfavorable obstetric outcomes. 40 To minimize the mentioned limitations of many studies, we used the NOS scale method to classify and assign higher or lower quality to each study as explained. Finally, through the present data compilation, it is possible to direct the search for endometriosis-associated obstetric complications looking for the underlying causes and mechanisms. It also informs on possible guidelines for the clinical care of patients with surgical diagnosis of endometriosis, in order to reduce the rates of comorbidities associated with endometriosis in pregnancy.

Endometriosis is defined by the presence of endometrial (glandular and/or stromal) tissue outside the uterus. The most frequent sites of lesion are the pelvic viscera and the peritoneum, and the disease can be classified as superficial, deep or ovarian and/or peritoneal. The most severe forms can lead to deformities of the Fallopian ducts and may affect the urinary tract and intestinal walls. 1 It is estimated that endometriosis affects 10% of women of reproductive age, is associated with pelvic pain in 30% and causes infertility in 30 to 40%. 2 3 4 In recent years, there has been considerable progress in understanding the pathogenesis, the evolution, the diagnosis and the treatment of the disease. 5 It is important to emphasize that infertility alone is already associated to a greater risk of obstetric complications such as pre-eclampsia, gestational hypertension, prematurity, hemorrhage before delivery and the need of cesarean section. 6 Some studies postulated the association of endometriosis with unfavorable obstetric outcomes, such as pre-eclampsia or spontaneous hemoperitoneum, and the occurrence of sigmoid perforation or appendicitis. 7 8 While not yet clarified, these associations may occur due to endometrial resistance to progesterone, inadequate uterine contractions, excessive stimulation of the endometrium caused by free radicals, changes in the uterine junctional zone, and inflammatory processes causing endometrial, peritoneal and systemic manifestations. 1 9 10 11 12 These mechanisms will be addressed in the discussion of the present study. Two systematic reviews that related endometriosis to gestational risks have been recently published; however, there is a methodological gap regarding the heterogeneity among the groups studied, the confirmatory diagnosis of endometriosis, the sample size of each published study, and the inclusion of patients who were already classified as having high-risk gestations (symptomatic patients who sought out clinics and hospitals). 13 14 Therefore, the goal of our study was to perform a systematic review of the literature to assess whether the presence of endometriosis in fact results in adverse obstetric outcomes. We took the data quality of the analyzed articles into consideration to reach the conclusions.