Background

Throughout the years, it has always been clear to scientists that the primary endpoint in reproductive medicine was the healthy baby and that all other endpoints would be considered only a surrogate [1, 2]. The published infertility clinical trials have rarely reported clear data about the possible harm of the medical, surgical and biological procedures for enhancing fertility [3, 4], as well as giving very little relevance to long-term effects of those procedures on maternal and offspring health. Only 4.8 % and 5.7 % of randomized controlled trials (RCTs) on infertility treatments reported on perinatal and maternal outcome [5]. This is probably because obstetrical and infant care are delivered by other providers and patients are lost to follow-up.

Notwithstanding these limitations, more and more data available in the literature seem to demonstrate that pregnancy following infertility treatments are at higher risk of adverse pregnancy and perinatal outcomes when compared with those after natural conception (NC) independently from scientific approach [6], and this is particularly true for pregnancies achieved thanks to high technology infertility treatments [5]. The majority of this risk is a “pure” iatrogenic risk due to the high rates of multiple births, i.e. 41.1 % of the United States (US) infants conceived with assisted reproductive technologies (ART) were born as multiple-birth infants compared with only 3.5 % of infants among the general birth population [7]. The rate of multiple deliveries following ART represents about 18.7 % of total multiple-birth infants [7]. However, because also singleton infants conceived with ART are at higher risk of preterm birth (PTB) and low birthweight (LBW) [7], other determinants cannot be excluded. Patients’ characteristics including the infertility state [8, 9] and many preconception risk factors for subfertility [1012] can largely increase the absolute and relative risk of obstetric morbidity.

Although systematic reviews with and/or without meta-analysis have been published on specific topics, at the moment no comprehensive review is available in the literature, discussing the impact on maternal and perinatal outcome of each element and/or clinical/biological choice which comprise the high technology infertility treatments. Based on these considerations, the aim of the current document was to comprehensively review in a systematic fashion the hitherto published evidences regarding the effects of high technology infertility treatments on the obstetric risk of patients with female and couple infertility. The effects on the risk of chromosomal abnormalities and malformations of the offspring was not a study aim.

Materials and methods

The methodology used for the current systematic review consisted of searching all available articles for each specific issue to explore the relationship between high technology infertility treatments and pregnancy and perinatal complications. High technology infertility treatments were considered as all interventions for fertility enhancement including manipulation of female gametes. The Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) Statement [13] was followed but after comprehensive search all the authors agreed to prepare the document in a narrative fashion in consideration of the multifaceted aspects to discuss.

Multiple strategies were used to search and identify relevant demographic, epidemiological, clinical and experimental studies. Sociological online libraries (IBSS, SocINDEX), Institute for Scientific Information, PubMed, Web of Science and Google Scholar were consulted. Only articles written in English were considered. The search was conducted independently by two authors (S.P. and S.S.). The literature available up to August 2016 was captured, including all available studies which reported data about the relationship between each fertility technique and the related obstetric and perinatal complications, matching every intervention with every potential obstetric disorder and perinatal health impact, as shown in Table 1. Additional journal articles were identified from the bibliography of the studies included. At study design, all the authors agreed to exclude from final analysis data of pregnancy and perinatal complications related to: 1. prepregnancy maternal factors; 2. low technology interventions (intervention aimed to enhance fertility without any manipulation of female gametes, ie. lifestyle intervention programs, insulin sensitizing drugs, ovulation inductors, macro- and micro-supplements, intrauterine insemination, etc.); 3. specific high technology infertility treatments for male factor [including intracytoplasmic sperm injection (ICSI), specific techniques for sperm selection and/or retrieval, etc.]; 4. ovarian tissue/ovary and uterus transplantation; 5. chromosomal abnormalities and malformations of the offspring. The choice to consider a study relevant in order to be included in the current review was arbitrarily taken by each author, even if an evidence-based hierarchy was used. Exploratory studies on mechanisms of action and/or pathogenesis of any complication were included only in absence of available clinical data. Data on the efficacy of each procedure were reported only as necessary for the study aim. Any disagreement or uncertainty was resolved by discussion to reach a consensus.

Table 1 Key words used to study the relationship between specific procedure of ARTs for female or couple infertility and obstetric and perinatal outcomes

The available evidence about the relationship between high technology infertility treatments and adverse obstetric and perinatal outcomes was analyzed assessing the level of evidence according to the Oxford Centre for Evidence-Based Medicine (OCEM)-Levels of Evidence 2011 guidelines [14] and the quality of evidence according to the Grading of Recommendations Assessment, Development, and Evaluation (GRADE) system [15].

Results

Figure 1 is the flow diagram of the systematic review including the numbers of studies screened, assessed for eligibility, and included in the review, with reasons for exclusions at each stage [13]. Table 2 summarizes the main risks for obstetric and perinatal adverse outcomes in women who receive ART and specific procedure for high technology infertility treatments according to the level [14] and the quality [15] of available evidence.

Fig. 1
figure 1

PRISMA 2009 [13] flow diagram

Table 2 Levels and quality of the evidences available about the relationship between specific procedure for ARTs and risk of the main obstetric and perinatal adverse outcomes

Overall ART-related complications

Singleton pregnancies

Many data from systematic reviews with and without meta-analyses have demonstrated that ART singletons are at increased risk of pregnancy and perinatal complications. In 2004, an initial systematic review of controlled studies found a relative risk of 3.27 (95 % CI 2.03 to 5.28) for early-preterm birth (EPTB) (< 32 weeks) in singleton pregnancies from assisted conceptions [16]. Singleton pregnancies resulting from in vitro fertilization (IVF) presented higher rates of worse obstetric outcome, compared with NC singletons of couples matched for maternal age, showing an increase in perinatal mortality [odd ratio (OR) 2.40, 95 % confidence interval (CI) 1.59 to 3.63], PTB at < 33 weeks’ gestation (OR 2.99, 95 % CI 1.54 to 5.80), PTB at < 37 weeks’ gestation (OR 1.93, 95 % CI 1.36 to 2.74), very-low preterm birth (VLBW) (<1500 g) (OR 3.78, 95 % CI 4.29 to 5.75), and SGA (OR 1.59, 95 % CI 1.20 to 2.11) [17]. It is noteworthy that this review included in the final analysis studies in which the control arm was composed of NCs and after ovulation induction strategies. In 2012, an extensive systematic review with meta-analysis of 20 matched and 10 unmatched cohort studies, most having high quality and adjusted for important confounders, concluded that singleton IVF/intracytoplasmic sperm injection (ICSI) pregnancies are associated with higher risks of antepartum hemorrhage [relative risk (RR) 2.49, 95 % CI 2.30 to 2.69], pregnancy-induced hypertension (PIH)/preeclampsia (PE) (RR 1.49, 95 % CI 1.39 to 1.59), gestational diabetes mellitus (GDM) (RR 1.48, 95 % CI 1.33 to 1.66), caesarean section (RR 1.56, 95 % CI 1.51 to 1.60), PTB (RR 1.54, 95 % CI 1.47 to 1.62), LBW (RR 1.65, 95 % CI 1.56 to 1.75), SGA (RR 1.39, 95 % CI 1.27 to 1.53), admission to neonatal intensive care unit (NICU) (RR 1.58, 95 % CI 1.42 to 1.77), and perinatal mortality (RR 1.87, 95 % CI 1.48 to 2.37) than NC [18]. Also in this last meta-analysis, only in some of the included studies were the pregnancies resulting from ovulation induction excluded from non-IVF/ICSI conceptions [18].

The most recently published meta-analysis about the risk of pregnancy-related complications and adverse perinatal outcomes in singleton pregnancies obtained with ART involves 50 cohort studies for a total of 161,370 ART singleton compared with 2,280,241 NC singleton pregnancies [19]. This meta-analysis revealed that the singleton ART pregnancies experienced a significantly increased risk of placenta previa (RR 3.71, 95 % CI 2.67 to 5.16), placental abruption (RR 1.83, 95 % CI 1.49 to 2.24), antepartum hemorrhage (RR 2.11, 95 % CI 1.86 to 2.38), and postpartum hemorrhage (RR 1.29, 95 % CI 1.06 to 1.57). The increased risk for PIH, GDM, caesarean sections, PTB, LBW, SGA and perinatal mortality was confirmed and resulted not different from previous meta-analytic data [19]. Of note, the risk of EPTB (RR 2.12, 95 % CI 1.73 to 2.59) and of VLBW (RR 2.12, 95 % CI 1.84 to 2.43) was two-fold higher in ART conceptions than in NC [19]. The results of this study are consistent with those of the previous reviews but it presents important strengths, such as the large sample size, 64 % of the included studies were considered of high methodological quality and the association between ART and obstetric risk persisted and remained statistically significant in sensitivity analysis based on various exclusion criteria [19]. Nevertheless, as relevant biases, patients who achieved a pregnancy with ovulation induction and/or intrauterine insemination (IUI) were included in the NC category, resulting in an underestimation of the association between ART and adverse outcomes. When data were restricted to studies that did not include these patients in the NC group, the risk of GDM, placental abruption, PTB, EPTB, LBW, VLBW, and perinatal mortality resulted in a further increase. Unfortunately, more than half of the included studies did not specify whether they were included, thereby restricting this subgroup analysis [19]. Finally, recent retrospective studies confirmed in ART pregnancies a risk of PIH/PE about 20 % higher (aOR 1.17, 95 % CI 1.10 to 1.24) [20] and demonstrated a risk of peripartum hysterectomy about six-fold increased (OR 5.98 95 % CI 2.18 to 16.40) in comparison with non-ART pregnancies [21].

Multiple pregnancies

Initial systematic reviews with meta-analyses [16, 17] demonstrated that ART twins had a higher risk of PTB compared with NC twins but the perinatal mortality, contrary to singletons, was unchanged [17] or reduced [16].

A recent meta-analysis including 39 cohort studies explored the risk of adverse pregnancy outcomes in ART conceptions compared with NC in a sample of 146,008 multiple births [153]. Also the odds for SGA (OR 1.81, 95 % CI 1.26 to 2.60), PTB (OR 1.34, 95 % CI 1.08 to 1.66), and caesarean section (OR 2.71, 95 % CI 2.23 to 3.30) were increased [153]. Meta-regression for the covariate of age suggested that risk was independent of age [153]. Similarly, also the fourth meta-analysis [154] including 11 retrospective and prospective cohort studies confirmed that OD increase the risk of PE (in comparison with homologous IVF cycles) of about 70 % and that neither multiple pregnancies nor patient age can explain that effect by meta-regression analysis.

Finally, the last systematic review with meta-analysis [155] included, after search for original studies reporting at least five OD pregnancies with a control group of pregnancies conceived by conventional IVF/ICSI or NC and case series with > 500 cases, 35 studies reporting one or more pregnancy and perinatal complications were analyzed. The risk of PIH (aOR 2.30, 95 % CI 1.60 to 3.32), PE (aOR 2.11, 95 % CI 1.42 to 3.15), LBW (aOR 1.53, 95 % CI 1.16 to 2.01), PTB (aOR 1.75, 95 % CI 1.39 to 2.20) and CS (aORs 2.20, 95 % CI 1.85 to 2.60) was higher in OD than in IVF singleton pregnancies, whereas in multiple pregnancies only the incidence of PIH (aOR 2.45, 95 % CI 1.53 to 3.93) and PE (aOR 3.31, 95 % CI 1.61 to 6.80) was increased [155]. The risk of PE (aOR 2.94, 95 % CI 2.29 to 3.76), PTB (aOR 2.30, 95 % CI 1.09 to 4.87), LBW (aOR 1.94, 95 % CI 1.10 to 3.41) and CS (aOR 2.38, 95 % CI 2.01 to 2.81) was also increased in OD vs. NC singleton pregnancies [155]. Postpartum hemorrhage resulted increased in OD vs. IVF both in singleton (aOR 2.40, 95 % CI 1.49 to 3.88) and multiple (aOR 4.91, 95 % CI 1.22 to 19.83) pregnancies. No difference was detected in terms of GDM [155].

One of the largest and better controlled cohort study published on perinatal outcomes of children born after OD included 375 OD babies, and clinical data compared with three control cohorts of children, i.e. NC (33,852 babies matched by date and year of birth) and born after either IVF (11,060 singletons, and 6,532 twins) or ICSI (5,866 singletons, and 3,101 twins) [156]. An increased risk of PTB (aORs 1.8, 95 % CI 1.2 to 2.3; aOR 2.5, 95 % CI 1.7 to 3.6; and aOR 3.4, 95 % CI 2.3 to 4.9, respectively) and LBW (aOR 1.4, 95 % CI 0.9 to 2.2; aOR 1.8, 95 % CI 1.2 to 2.8; and aOR 2.6, 95 % CI 1.7 to 4.0, respectively) was detected in OD pregnancies vs. control pregnancies [156]. Of note, the risk of PE was also increased three-fold in OD pregnancies (aOR 2.9, 95 % CI 1.8 to 4.6; aOR 2.8, 95 % CI 1.7 to 4.5; and aOR 3.1, 95 % CI 1.9 to 4.9) [156]. The risk remained higher also after adjusting for maternal characteristics and after sub-analysis for twin pregnancies. Moreover, when the perinatal risk was adjusted for maternal PE the results improved, demonstrating no direct effect of OD on perinatal outcomes [156]. These data have been recently confirmed by a systematic review [157] showing that OD is an independent risk factor for PIH/PE, especially in twin pregnancies, and that its effect on fetal birthweight or growth is minimal after adjusting for obstetric complications. Finally, a very recent national registry study confirmed that OD recipients are more likely to have PTB (aOR 1.28, 95 % CI 1.12 to 1.46) and VPTB (aOR 1.30, 95 % CI 1.03 to 1.64) when compared with autologous patients, whereas the risk of having a SGA baby (aOR 0.72, 95 % CI 0.58 to 0.89) and of perinatal death (aOR 0.29, 95 % CI 0.09 to 0.94) was lower after adjusting data for gestational age [158].

Sperm donation

Sparse available data in the literature seems to suggest an increased risk of hypertensive disorders during pregnancy, with a specific increase of the PE risk, in nulliparous and in multiparous pregnancies with changed paternity [159, 160]. In this view, it could be hypothesized that the exposure to the paternal semen before conception has a protective effect, whereas the use of donor insemination after a previous pregnancy with paternal semen could increase the risk with an immune mechanism similar to hypertensive disorders seen in OD pregnancies. Unfortunately, at the moment data on sperm donation regard low technology interventions [11], and showed that the use of donor sperm in IUI cycles is associated with a risk of perinatal complications lower to those of the children born after partner sperm IUI and comparable to those of the NC children [44].

Embryo donation

Very little is known about the relationship between embryo donation and pregnancy and perinatal complications. In fact, available data can be extrapolated from infertile populations who conceived after mixed procedures of gamete and embryo donations [159]. At the moment, it very difficult to draw conclusions on the obstetric risk in women who had an embryo donation because the number of confounders and biases is so frequent and the detrimental effect of the double gamete donation can be only supposed. Commonly, the recipients are highly selected women with few medical comorbidities but who had had probably many previous ART failed attempts and a longer time-to-pregnancy. In addition, embryos donated are almost always frozen embryos. Finally, because embryo donation is more cost-effective than oocyte donation in case of male factor [161], the comparison in terms of pregnancy complication can be favorable for pregnancies obtained after embryo donation.

Surrogate pregnancy

A systematic review on gestational surrogacy has been very recently published, and the pregnancy and perinatal outcomes of gestational carriers compared, when possible, with those of standard IVF and OD cycles [162]. The incidence of PIH ranged from 4.3 to 10 % and from 2.9 and 7.4 %, and the incidence of placenta previa and/or placental abruption from 1.1 to 7.9 % and from 1.1 to 3.7 %, respectively, in singleton and twin surrogate pregnancies [162], and resulted not different from those observed in IVF pregnancies and lower than that usually reported in OD pregnancies (ranged from 16 to 40 %) [163]. Three cases of hysterectomies related to delivery were also reported in gestational carriers and were due to uterine atony, placenta accreta and uterine rupture [162]. In surrogate singletons, the incidence of PTB and of LBW resulted, respectively, ranging from 0 to 11.5 % and from 0 to 11.1 % [162]. When compared to control groups, the risk of PTB and LBW was not different from IVF singletons (incidence of PTB of 14 % and of 13.6–14.0 %, respectively) and the risk of LBW was also not different from OD singletons (incidence of 14.0 %) [162]. However, a very recent US cohort study [164] underlined that the increased risk of PTB (aRR 1.14, 95 % CI 1.05 to 1.23) observed in gestational carriers is significantly influenced not only by multiple pregnancies but also by OD.

Discussion

Despite the level and the quality of the current evidence it is generally suboptimal due to the presence of biases, confounders and limitations in study design (Table 3), current comprehensive review confirms that subfertile women who conceived after the use of high technology infertility treatments have an overall increased risk of pregnancy and perinatal complications, and highlights that every single step and/or procedure can play an independent and crucial role. In addition, several concomitant risk factors are frequently present in the same woman and influence the clinical and biological strategy of treatment. Thus, it is virtually impossible to define the weight of each reproductive treatment’s phase determining the whole patient’s risk. In fact, the infertility condition represents a bias per se in every study dealing with infertility treatments [11] and the presence of many confounding factors cannot be always adequately controlled through multivariate analysis because in many studies they are not clinically available, missing, or not collected.

Table 3 Specific biases, confounders and limitations present and/or declared in the available studies

Ideally, the knowledge of the pathogenesis of the increased risk of pregnancy and perinatal complications in women who receive high technology infertility treatments and of the specific mechanisms of action could be crucial for preventing them. Unfortunately, few data are available regarding the biological explanations of that increased risk. Many mechanisms have been hypothesized and regard the alterations of the early placentation including not only alterations in endometrial receptivity, genetic and epigenetic mechanisms of implantation, invasion and growth of the trophoblast but also genetic and/or epigenetic alterations of oocyte/embryos due to biological manipulations (extended culture, culture media, techniques of cryopreservation, etc.), and immunological intolerance in case of OD because the fetal genome is allogenic to the carrier [11, 90, 127, 142, 165167]. Thus, in the next future, still remains the need and an effort should be made to understand the reasons of these risks in order to minimize or prevent them.

However, from the clinical point of view, the priority is not to precisely estimate the amount of the obstetric risk but to recognize the presence of one or more risk factors (infertility and subfertility causes, patient’s characteristics and specific ARTs-associated risks), to correct those modifiable and to strictly follow the resulting pregnancies with appropriate prenatal cares. In fact, the delay in receiving prenatal care increased the PTB risk, while more-frequent use of prenatal care significantly improved the birth weight among pregnancies at high risk including subfertile women [168]. Recently, a large nationwide population-based study demonstrated that an adequate and intensive prenatal care reduces the risk of adverse pregnancy outcomes in women with history of infertility [169]. Specifically, less than six prenatal visits (compared with equal or more than six prenatal visits) and prenatal visits performed after the 12th week of gestation (compared with prenatal visits performed at or before the 12th week of gestation) are related with a risk lower of VLBW neonates (aOR15.1, 95 % CI 8.8 to 25.8; aOR 2.1, 95 % CI 1.2 to 3.8; respectively), LBW neonates (aOR 2.1, 95 % CI 1.5 to 2.5; aOR 1.6, 95 % CI 1.3 to 1.9; respectively), and preterm birth (aOR 2.2, 95 % CI 1.9 to 2.6; aOR 1.1, 95 % CI 0.9 to 1.3; respectively) [169]. That data, however, were limited to singleton pregnancies.

At the moment, well established strategies to identify, follow and manage infertile patients and/or patients who have receive an infertility treatment are lacking and only few papers suggest potential strategies of management consisting in a generic close pregnancy monitoring and diagnostic testing [137, 157]. Physicians should assess the pregestational risk of infertile women before start any fertility enhancement treatment and discuss with the couples the increased risk for maternal and perinatal complications in a view of risk to benefit ratio and the potential alternatives, suggesting also to avoid any medical intervention in case of high-risk patients [137, 170], such as in case of women of very advanced reproductive age (> 55 years) or advanced reproductive age (> 45 years) with medical conditions [171].

Conclusion

Subfertile women who conceived after the use of high technology infertility treatments are at increased risk of pregnancy complications, and every single/specific step and/or procedure can play an independent and crucial role. Thus, all infertile patients scheduled for high technology infertility treatments should be clearly informed of that increased obstetric and perinatal risk in case of pregnancy, regardless of multiple pregnancy. A careful preconceptional counselling aimed to optimize the general health status of the pre-pregnant women is needed (to stop smoking, reduce BMI in overweight/obese patients, and so on), identifying and treating modifiable reproductive disorders [11] and, finally, an effort should be made to optimize the infertility treatments (milder stimulation, OHSS prevention, elective SET) in order to prevent or reduce the risk of pregnancy complications in these infertile women. Finally, further large cohort prospective studies are required to clarify the contribution of each single factor on pregnancy and perinatal outcomes.