Skip to main content
NCBI home page
The Journal of Clinical Hypertension logoLink to The Journal of Clinical Hypertension
. 2016 Nov 7;19(2):173–183. doi: 10.1111/jch.12945

Risk of hypertensive disorders in pregnancy following assisted reproductive technology: overview and meta‐analysis

Costas Thomopoulos 1,✉, George Salamalekis 2, Konstantinos Kintis 3, Iliana Andrianopoulou 1, Helena Michalopoulou 1, George Skalis 1, Stefanos Archontakis 1, Ourania Argyri 1, Costas Tsioufis 3, Thomas K Makris 1, Emmanuel Salamalekis 2
PMCID: PMC8031300  PMID: 28071857

Abstract

The extent of the increased risk of pregnancy hypertensive disorders following assisted reproductive technology (ART) was investigated. PubMed and the Cochrane Collaboration Library were used as data sources to identify and select longitudinal cohorts comparing pregnancies following ART with spontaneously conceived pregnancies, between 1978 and June 2016. Risk ratios and 95% confidence intervals (CIs) of three outcomes, ie, gestational hypertension (GH), preeclampsia (PE), and their sum (PHD), were calculated. Stratification of results by gestation order (singletons and nonsingletons) was pursued, but a separate “all orders” mixed stratification was considered. Sixty‐six longitudinal studies (7 038 029 pregnancies; 203 375 following any ART) were eligible. All outcomes independent of gestation order (“all orders”) were increased following any invasive ART: GH (+79% [95% CI, 24%–157%]) and PE (+75% [95% CI, 50%–103%]) to a greater extent, with smaller increases in PHD (+54% [95% CI, 39%–70%]). The risk of PHD following ART steadily increased independent of gestation order.

Keywords: assisted pregnancy, gestational hypertension, infertility, preeclampsia, pregnancy‐induced hypertension

1. Introduction

Pregnancy‐related hypertensive complications represent the main cause of maternal and fetal morbidity1 and also anticipate early development of cardiovascular disease relatively early after the termination of pregnancy.2 Infertility treatments have increased steadily in the past 20 years and have proven effective in achieving considerable rates of successful conception and live birth rates even among women younger than 35 years.3 Assisted reproductive technology (ART) procedures including the noninvasive method of ovulation induction (OI) achieved by specific pharmacologic agents, are used to induce assisted conceptions. However, these treatments continue to raise criticism because they have been related to adverse obstetric and perinatal outcomes, including pregnancy‐related hypertensive complications (preeclampsia [PE] and gestational hypertension [GH]), at least in the early years of their widespread implementation.4 Although a lowering trend in the risk of pregnancy‐related hypertension in ART pregnancies5 was suggested in the later years as a result of more rigorous clinical and laboratory rules,6 we have previously reported by qualitative synthesis that this risk still remains significant.7 Also, a recent meta‐analysis of selective ART procedures (in vitro fertilization [IVF] and intracytoplasmic sperm injection [ICSI]) restricted to singletons demonstrated that ART was associated with a significant risk of pregnancy‐related hypertensive complications compared with spontaneous pregnancies.8

It is suggested that the activation of the immune system plays an important role in the pathophysiological interplay between ART and hypertensive complications in pregnancy.9 Additionally, different techniques used during ART (OI, intrauterine insemination [IUI; minimally invasive ART], IVF, ICSI, oocyte donation [OD], embryo transfers after fresh or frozen transfers, single or double embryo transfers) may predispose women to abnormal placentation, inadequate uteroplacental circulation, and impaired vasodilation.10 However, coexisting risk factors such as maternal age, body size, parity, the order of gestation, and preconception hypertension could also play a role as either an accelerator or an effect modifier in such an association.5, 11, 12 Finally, the historical absence of a consensus regarding the diagnostic criteria for pregnancy‐related hypertensive complications, especially for PE, further complicates any investigation in this particular field.13

We have conducted a comprehensive overview of cohort studies from the inception of ART procedures in 1978 to 2016 to quantify the relative and absolute risks of different pregnancy‐related hypertensive phenotypes (ie, GH, PE, and their sum [PHD]) following ART‐mediated pregnancies (ie, noninvasive ART [OI] and any invasive ART [IUI, IVF, ICSI, OD, single or double embryo transfer, and transfer after fresh and frozen cycles]) compared with spontaneous pregnancies.

2. Methods

The recommendations of the Preferred Reporting Items for Systematic Reviews and Meta‐Analysis (PRISMA) were adhered to (Figure S1).14

2.1. Data sources and searches

Studies were primarily identified in the PubMed and Cochrane Collaboration Library databases. We reviewed all published studies conducted from 1978 to June 25, 2016. The reference lists of the retrieved articles and previous meta‐analyses were also examined. The search procedure used in PubMed and Cochrane Library is presented in Table S1.

2.2. Study Selection

We included prospective and retrospective cohort studies—both matched and unmatched—providing categorical data on the outcomes of interest. Accordingly, we excluded studies that lacked a spontaneous conception group and studies of head‐to‐head comparisons between different ARTs. We also excluded nonhuman studies, letters to the editor (not containing primary data), case series, editorial articles, commentaries, and articles with no authors listed. Articles that were not in English were included provided that they were abstracted in English. No restriction criteria were imposed with regard to the type or size of the studied population, nor to the type of implemented ART treatment.

2.3. Data extraction, ART definition, and outcomes of interest

The literature search, selection of studies, and extraction of data were performed independently by two investigators (C. Th. and G. Sa.). Disagreements were resolved by discussion. The numeric data tabulated in the articles were used. For each study, we extracted the number of incident pregnancy‐related hypertensive disorders in both groups.

We defined minimally invasive ART as the implementation of IUI; invasive ARTs as the implementation of IVF, ICSI, OD, single or double embryo transfer, and transfer after fresh and frozen cycles; and any invasive ART as the sum of minimally invasive and invasive ART. OI as an infertility treatment (noninvasive ART) was considered separately. However, it was acknowledged that “ovarian stimulation” might be included in different invasive ART protocols.

The outcome of interest was hypertensive disorders in pregnancy, defined as follows: GH, PE, and PHD consisting of the sum of GH and PE when these two phenotypes were reported separately or when aggregated data for all pregnancy‐related hypertensive complications were reported in each individual study. Traditionally, GH is defined as new‐onset elevated blood pressure (≥140/90 mmHg) after 20 weeks of gestation without proteinuria, while the presence of proteinuria (≥0.3 g/24 h) in the context of GH defined PE. After 2013, the American College of Obstetricians and Gynecologists replaced proteinuria as a necessary criterion for PE diagnosis with signs and symptoms of end organ damage.15, 16 However, the definitions of GH and PE of individual studies were guided by guidelines in place at the time of each study.

2.4. Quality assessment

Because of both the complexity of the clinical setting and the high number of confounders with potentiality to introduce selection and detection bias, we used a nonstandard quality assessment procedure. We evaluated the quality of the included studies by assessing the selection, detection, and attrition bias. To control for selection bias, we assessed whether each study provided equal conditions for all participants. Prospective designs were considered of higher quality than retrospective, and matching for maternal age, parity, smoking status, same gestation period for studied groups, ethnicity, body mass index at booking, chorionicity, and order of gestation between exposed and unexposed cohorts was also considered. Studies that matched for more than four of the above characteristics and unmatched studies were considered of lower quality. Because the order of gestation is strongly related to pregnancy‐related hypertensive disorders in spontaneously conceived pregnancies17 we evaluated whether the studies reported stratified results by order of gestation. We assessed whether conditions such as “vanishing embryo” and ovarian stimulation were excluded in the spontaneously conceived pregnancies and finally whether gestation age at recruitment was before or after the 23rd week to control for the early development of pregnancy‐related hypertensive disorders.

To control for detection bias, we assessed whether the outcomes were properly defined and whether they were confirmed by hospital records or national registries or were only based on self‐reported data embedded in questionnaires. Moreover, we evaluated the number of outcomes reported in each individual study, with those reporting two outcomes (GH and PE) being of higher quality than those reporting only one or a sum (ie, GH, or PE or PHD). Excluding hypertension before the 20th week of gestation was also considered an additional factor of higher quality.

Finally, to control for attrition bias, we evaluated the extent of loss to follow‐up by calculating the ratio of the number of individuals lost to follow‐up to the number of patients who completed the study protocol or the extent of the response ratio in studies based on patient interviews. We arbitrarily considered a loss‐patient ratio <10% and a response rate >80% as adequate criteria to reject significant attrition bias. The evaluation and scoring of the 10 bias criteria mentioned above (five criteria for selection, four for detection, and one for attrition) were based on a binomial integer scale ranging from zero to one, with one indicating higher quality. These scores were summed and reflected the overall study quality, with 10 representing the highest quality. High‐ and intermediate‐quality studies were arbitrarily considered those with scores seven or greater and equal to five or six, respectively. The procedure used for the quality assessment is presented in Table S2.

2.5. Data synthesis and analysis

The protocol of the analyses was prespecified. We conducted separate analyses of the selected cohort studies regarding the impact of OI (noninvasive ART) and any invasive ART (IUI, IVF, ICSI, combined IVF/ICSI, OD, single embryo transfer, double embryo transfer, transfer following frozen cycles) on PHD, GH, and PE. All analyses were stratified by order of gestation as follows: singleton, nonsingleton (twins and multiples), and “all orders,” consisting of both singleton and nonsingleton. The studies reporting nonstratified data by order of gestation (mixed order studies) were only considered in the all orders analysis. Sensitivity analyses were performed to include higher‐quality studies and to control for potential confounders.

The relative risk estimates (with 95% confidence intervals [CIs]) were combined using a random effects model, in which the log relative risk of every study was weighted by the reciprocal of the variance of the log relative risk. The proportion of inconsistency across studies that was not explained by chance was quantified with the I 2 statistic. When no significant heterogeneity was detected by the χ2 Q statistic (P>.10), a fixed‐effects model was implemented. Furthermore, the influence of individual studies on the pooled effect sizes was tested by excluding one study at a time; if the point estimate of the combined effect size when a given study was excluded fell outside the CI of the overall estimate risk of all available studies, the study in question was considered to have an excessive influence.

Risk ratios and their 95% CIs were reported using the Mantel‐Haenszel method, and the effects of ART on each individual outcome were illustrated with forest plots in the random effects model. The absolute risk increase (weighted for inverse variance) was also calculated, as was the number of pregnancies needed to treat for the development of one adverse outcome (NNH). The presence of publication bias was investigated graphically by funnel plots of precision (random effects plotting) and the Duval and Tweedie trim and fill method. All statistical analyses were performed using Comprehensive Meta‐Analysis version 2 (Biostat, Englewood, NJ, USA). In each individual analysis, a P<.05 was considered to indicate statistical significance.

3. Results

3.1. Studies and pregnancies

Figure 1 illustrates the steps performed to identify the studies to be included. The reasons for the exclusion of noneligible articles (n=67)E1–E67 are presented in Table S3. Table 1 shows the characteristics of the included studies (n=66, 79 cohorts),S1–S66 with a total of 7 038 029 pregnancies (203 375 by any ART and 6 834 654 naturally conceived). Along with the selected studies, we found nine cohortsS6,S7,S26,S28,S33,S38,S52,S53,S55 of noninvasive ART (5544 pregnancies through OI compared with 218 182 naturally conceived); five cohorts S6,S9,S17,S47,S55 of minimally invasive ART (2439 pregnancies through IUI compared with 11 283 naturally conceived); and 70 cohorts (presented in Table 1) of any invasive ART (197 831 pregnancies through invasive ART compared with 6 616 472 naturally conceived). Thirty‐seven cohorts focused on singleton pregnancies and 32 on nonsingletons. Finally, we identified 10 cohorts that were not stratified by gestation order (mixed order) and 14 cohorts in which the qualification of mixed order was imputed by using the tabulated stratified results for singletons and nonsingletons (Table 1). Two studiesS4,S17 that reported mixed order results were considered singletons because the rate of nonsingletons was too small. In 10 studiesS3,S8,S13,S15,S19,S23,S29,S33,S36,S57 on invasive ART, IUI was also partially used to an unknown extent as an alternative to more invasive ART (eg, IVF), while in six studies, OI and any invasive ART results were not separately reported.S8,S13,S15,S23,S29,S66 In four studies,S6,S9,S47,S55 separate results were provided for IUI and any other invasive ART. Data on incident GH were reported in the cohorts of eight studies, on incident PE in the cohorts of 35 studies, and on incident PHD in the cohorts of 40 studies (Table 1).

Figure 1.

Figure 1

Open in a new tab

Identification process for included cohort studies

Table 1.

Characteristics of Selected Studies

Study Design Total Population (ART vs Spontaneous) Maternal Age (ART Group), y Order of Gestation ART Treatment Outcome Quality
Apantaku et alS1 RCM 88 vs 88 33.5 S IVF/ICSI PE Low
Baxi and KaushalS2 RCU 36 vs 138 28.8 T IVF/ICSI PHD Low
Beltrán Montoya et alS3 RCM 57 vs 57 32.5 T IVF or IUI GH, PE, PHD High
Caserta et alS4 PCM 364 vs 304 36.3 S ICSI PHD High
Caserta et alS5 RCU 138 vs 207 38.5 T IVF/ICSI PHD Intermediate
Chen et alS6 RCM 422 vs 1658313 vs 1212935 vs 3532 33.0 MX OIIUIIVF/ICSI PE High
Daniel et alS7 RCU 72 vs 12110 vs 121 32.0 T OIIVF/ICSI PHD Low
Dayan et alS8 RCU 11247 vs 776623 33.9 S OI, IUI, IVF, IVF/ICSI PE Intermediate
De Geyter et alS9 RCU 37 vs 44356 vs 443147 vs 44358 vs 443 34.4 S IUIIVFICSIFR PHD Low
Delgadillo Barros et alS10 RCM 41 vs 77 34.0 MX, ML, S IVF PHD High
Fan et alS11 RCU 140 vs 21322 vs 213 31.4 T IVFICSI PE Low
Farhi et alS12 PCU 202 vs 587338 vs 587 32.9 S IVFICSI PHD Low
Fritzsimmons et alS13 RCM 128 vs 232 32.2 T, ML, MX OI, IUI, IVF, GIFT PHD High
Gielchinsky et alS14 RCU 27 vs 231 >45.0 MX OD PE Low
Hernández‐Díaz et alS15 RCU 349 vs 4762 NR MX, T, S OI, IUI, IVF, ICSI, GIFT GH, PE, PHD Low
Howe et alS16 RCM 54 vs 54 NR S IVF PHD Intermediate
Hoy et alS17 RCM 1552 vs 7717 NR S IUI PE Intermediate
Isaksson et alS18 RCM 2377 vs 445 33.9 MX, T, S IVF/ICSI PHD Intermediate
Jie et al, 2015S19 PCU 428 vs 2788 32.5 MX Invasive ART, SET PE Low
Källén et alS20 RCU 12186 vs 2000372 33.3 MX IVF/ICSI PE Low
Katalinic et alS21 PCU 2687 vs 7938 28.5 MX, ML, S ICSI PHD Low
Koudstaal et alS22 RCM 307 vs 307 32.8 S IVF PHD Low
Kozinszky et alS23 RCM 359 vs 359 33.5 MX, T, S OI, IUI, IVF PHD Intermediate
Le Ray et alS24 RCU 40 vs 236104 vs 236 45.4 MX IVFIVF+OD PE Low
Luke et alS25 RCU 228 vs 725 33.1 T Invasive ART PE Low
Lynch et alS26 RCU 129 vs 33069 vs 330 30.0 ML OIIVF PE Low
Malchau et alS27 RCU 316 vs 560228652 vs 5602215337 vs 56022 33.5 MX, T, S ODIVFICSI PE, PHD Low
Maman et alS28 RCM 646 vs 1902169 vs 469 34.1 S OIIVF PE Intermediate
Marchand et alS29 RCU 484 vs 13565 NR MX IVF, ICSI, OI, IUI GH, PE, PHD Low
Miyake et alS30 RCU 200 vs 5739 NR MX IVF, ICSI PHD Intermediate
Mohammed and Abdel‐MaaboudS31 RCU 145 vs 175 34.5 T IVF PE Low
Moini, et alS32 PCU 230 vs 170 30.6 T IVF/ICSI PHD Intermediate
Morcel et alS33 RCU 39 vs 17365 vs 173 33.3 T OIIUI, IVF, ICSI PE Intermediate
Najdet, 2016S34 RCU 388 vs 99980426696 vs 999804 NR S ODIVF/ICSI PE Low
Nassar et alS35 RCM 56 vs 112 31.0 T IVF PHD High
Nayeri et alS36 RCM 84 vs 473 29.8 MX IUI, IVF, ICSI PE Intermediate
Ochsenkühn et alS37 RCM 400 vs 400 32.6 MX, T, S IVF, GIFT PHD Intermediate
Olivennes et alS38 RCU 263 vs 5096162 vs 5096 33.6 S OIIVF PHD Low
Opdahl et alS39 RCM 58006 vs 315273 33.3 MX, T, S IVF/ICSI PHD High
Pandian et alS40 RCU 162 vs 32969 30.8 S OI, IVF, ICSI PE Low
Pelinck, et alS41 RCU 307 vs 132 32.6 S IVF PHD Intermediate
Pinborg et alS42 RCU 236 vs 566 32.3 T IVF/ICSI GH, PE, PHD Low
Poikkeus et alS43 RCU 499 vs 15037 33.5 S IVF/ICSI (DET, SET) GH, PE, PHD Low
Porreco et alS44 RCM 50 vs 50 47.0 MX OD PE Intermediate
Reubinoff et alS45 RCM 260 vs 260 32.7 S IVF PHD Intermediate
Rubio‐Cid et alS46 RCU 44 vs 52 36.5 T IVF PE Low
Salha et alS47 RCM 27 vs 2766 vs 2712 vs 12 33.5 MX, S, ML IUIODED GH, PE, PHD High
Saygan‐Karamürsel et alS48 RCU 274 vs 348 31.5 T ICSI PE Intermediate
Sazanova et alS49 RCU 11292 vs 571914 39.1 S Invasive ART, SET, FR PE Low
Schieve et alS50 RCU 3316 vs 157066 NR MX, ML, S Invasive ART PHD Intermediate
Shah et alS51 PCU 65 vs 128 33.4 ML Invasive ART PHD Intermediate
Shevell et alS52 PCU 1222 vs 34286554 vs 34286 34.5 S OIIVF GH, PE, PHD Intermediate
Silberstein et alS53 RCU 1294 vs 1715131974 vs 171513 32.5 S OIIVF PE Intermediate
Simchen et alS54 RCU 42 vs 417 49.2 T IVF+OD PHD Intermediate
Sun et alS55 RCM 777 vs 3103471 vs 18841341 vs 5317 32.9 S OIIUIIVF/ICSI PE High
Tallo et alS56 RCM 101 vs 101 33.0 MX IVF PHD Intermediate
Tanbo et alS57 RCM 355 vs 643 33.0 S Invasive ART GH, PE, PHD High
Tan et alS58 RCM 619 vs 999 34.2 MX, T, S IVF PHD Low
Tepper et alS59 RCU 6547 vs 595168 36.0 S Invasive ART PHD Intermediate
Verlaenen et alS60 RCM 140 vs 140 31.7 S IVF PHD Intermediate
von Düring et alS61 RCU 842 vs 240088 32.9 MX, T, S IVF PE Low
Wang et alS62 RCU 21615/574905 MX, ML, S Invasive ART PHD Low
Watanabe et alS63 RCU 471 vs 2610 38.3 S IVF PE Intermediate
Wolff et alS64 RCU 46 vs 49 42.6 S OD PHD Low
Yang et alS65 RCU 67 vs 143 32.5 T IVF PE Low
Zaib‐un‐Nisa et alS66 RCU 36 vs 96 30.2 T IVF, ICSI, OI PHD Low
Open in a new tab

Abbreviations: ART, assisted reproductive techniques; DET, double embryo transfer; ED, embryo donation; FR, frozen embryo transfer; GH, gestational hypertension; GIFT, gamete intrafallopian transfer; ICSI, intracytoplasmic sperm injection; IUI, intrauterine insemination; IVF, in vitro fertilization; ML, multiple not otherwise specified, MX, mixed (all orders together); NR, not reported; OD, oocyte donation; OI, ovulation induction; PE, preeclampsia; PHD, all pregnancy hypertensive disorders; PCM, prospective cohort matched; PCU, prospective cohort unmatched; RCM, retrospective cohort matched; RCU, retrospective cohort unmatched; S, singleton; SET, single embryo transfer; T, twin. For references, refer to supplementary data (online only) file. Numbers in bold are counted once. For quality assessment, refer to Supplemental Table S2 and Table S4.

Forty‐seven percent (31 of 66) of the included studies were of low quality (scoring from 0 to 4), approximately 38% (25 of 66) were of intermediate quality (scoring from 5 to 6), and approximately 15% (10 of 66) were of high quality (Table 1 and Table S4). Among 25 matched studies, 19 reported results after controlling for at least four from eight relevant confounders considered (Table S4).

3.2. Risk of ART on hypertensive disorders

OI (ie, noninvasive ART) was associated with an increased risk of PE and PHD (Figure 2). The relative risk estimates did not differ in PE incidence between singleton and nonsingleton pregnancies following OI. Although GH was increased in singleton pregnancies following any invasive ART, this finding was not significant in nonsingletons. Finally, all invasive ART procedures were associated with an increased risk of PE and PHD independent of gestation order. A significant difference was observed in the extent of increase in relative risk between singletons and nonsingletons only for PE (P=.005).

Figure 2.

Figure 2

Open in a new tab

Relative risk of pregnancy‐related hypertensive disorders following ovulation induction (noninvasive assisted reproductive technology [ART]) and any invasive ART. Minimally invasive ART cohorts were included in the analysis of invasive ART. Heterogeneity is presented by the I 2 statistic and P value, which is P for the heterogeneity test χ2 Q statistic; P value, P for the heterogeneity between random effects of singleton and nonsingleton pregnancies. GH indicates gestational hypertension; NA, not applicable; OI, ovulation induction; PE, preeclampsia; PHD, all pregnancy‐related hypertensive disorders; RR, Mantel‐Haenszel risk ratios in random effects models

3.2.1. Absolute risk increase

In Table 2, the relative risk changes following ART translated into increased absolute number of hypertensive complications are shown. Although the risk of pregnancy‐related hypertensive complications in naturally conceived pregnancies was almost 2‐ to 2.5‐fold higher in singletons compared with nonsingletons, the absolute risk increase did not differ by gestation order. Thus, in terms of absolute risk, for every 1000 invasive ART procedures, there were 31 and 46 more incident GH cases in all order and singleton pregnancies, respectively. A similar risk increase of incident PE and PHD was observed for any invasive ART independent of gestation order: one PE and one PHD event developed for every 49 and 43 treated pregnancies, respectively.

Table 2.

Absolute Risk Increase and NNH for Pregnancy‐Related Hypertensive Disorders Following ART

Intervention/Outcome All Interventionsa Risk in Naturally Conceived Pregnancies, % Absolute Risk Increase/1000 Pregnancies (95% CI) NNH P Value (Heterogeneity)b
OI (noninvasive ART)/PE
All orders 4.5 25 (12–41) 40 (82, 24) .090
Singletons 4.5 22 (5–43) 46 (185, 23)
Nonsingletons 10.5 82 (20–173) 12 (50, 6)
OI (noninvasive ART)/PHD
All orders 7.4 18 (4–33) 56 (270, 30) .33
Singletons 7.4 19 (3–38) 52 (340, 26)
Nonsingletons 9.9 −25 (−76 to 113) −40 (−14, 9)
Invasive ART/GH
All orders 4.0 31 (10–62) 32 (100, 16) .58
Singletons 4.3 46 (13–97) 22 (77, 10)
Nonsingletons 12.1 27 (−40 to 149) 37 (−25, 8)
Invasive ART/PE
All orders 2.7 20 (14–28) 49 (74, 36) .48
Singletons 2.7 18 (14–21) 57 (70, 48)
Nonsingletons 7.9 27 (8–51) 37 (128, 20)
Invasive ART/PHD
All orders 4.4 23 (17–30) 43 (60, 33) .067
Singletons 4.1 19 (14–25) 52 (70, 40)
Nonsingletons 9.8 32 (17–50) 31 (60, 20)
Open in a new tab

Abbreviations: ART, assisted reproductive technology; GH, gestational hypertension; NNH, number needed to harm every 1000 treated pregnancies; OI, ovulation induction; PE, preeclampsia; PHD, all pregnancy‐related hypertensive disorders. aFor relative risks in this section, refer to Figure 2. b P value for the comparison between singletons and nonsingletons.

3.2.2. Specific invasive ART treatments

The selected studies allowed us to perform analyses on the risk of specific invasive ARTs on pregnancy hypertension outcomes (Table 3). No analyses were performed for the effect of IVF, ICSI, and OD on GH because of the limited evidence. In two studiesS42,S43 reporting mixed data on IVF and ICSI, no significant effect on GH risk (+79% [95% CI, −43 to +563]) was observed.

Table 3.

Relative Risk of Pregnancy‐Related Hypertensive Disorders Following a Specific Invasive ART

Outcome Order of Gestation Intervention No. of Studies Heterogeneity (I 2, P value) Events Assisted Events Spontaneous RR (95% CI) P Value
PE All orders IVF alone 14 97%, <.001 2059/32806 76104/251210 1.68 (1.20–2.35) .25
Singletons 7 54%, .040 532/13780 18180/477056 1.63 (1.37–1.95)
Nonsingletons 6 82%, <.001 518/6130 2130/28557 1.29 (0.90–1.85)
All orders ICSI alone 3 73%, .025 453/8948 2575/56583 1.30 (0.72–2.37) NA
Singletons 1 – – – –
Nonsingletons 3 72%, .030 265/3255 1842/25573 1.32 (0.73–2.38)
All orders IVF/ICSIa 5 0%, .41 1189/27681 28254/1015708 1.47 (1.39–1.56) .32
Singletons 3 0%, .51 1125/27283 28165/1014929 1.48 (1.40–1.57)
Nonsingletons 2 43%, .19 64/398 89/779 1.28 (0.87–1.90)
All orders OD 6 55%, .048 142/912 30456/1056370 3.60 (2.56–5.05) <.001
Singletons 3 0%, .85 75/625 28646/1030840 4.53 (3.65–5.62)
Nonsingleton 2 0%, .43 23/106 1786/25013 2.99 (2.07–4.33)
All orders IUI (minimally invasive ART) 4 0%, .89 155/2402 437/10840 1.66 (1.39–1.99) NA
Singletons 3 0%, .79 147/2079 423/9627 1.66 (1.39–1.99)
Nonsingletons 1 – – – –
PHD All orders IVF alone 13 72%, <.001 1344/17889 6648/98484 1.43 (1.14–1.78) <.001
Singletons 12 32%, .14 773/12251 4320/73535 1.45 (1.26–1.68)
Nonsingletons 5 0%, .70 556/5765 2329/25248 1.03 (0.94–1.13)
All orders ICSI alone 5 45%, .13 986/12188 4175/65294 1.28 (1.12–1.47) .038
Singletons 5 6%, .37 586/8597 1859/40205 1.38 (1.25‐1.52)
Nonsingletons 2 0%, .67 400/3591 2316/25089 1.18 (0.87–1.90)
All orders IVF/ICSIa 6 72%, .002 197/1243 803/16239 1.55 (1.10–2.17) NA
Singletons 1 – – – –
Nonsingletons 5 0%, .65 146/744 208/1202 1.29 (1.06–1.57)
All orders OD 4 60%, .060 97/431 3536/56515 4.13 (2.52–6.77) .65
Singletons 3 0%, .69 50/283 1191/31085 4.21 (3.11–5.70)
Nonsingletons 3 74%, .023 47/148 2345/25430 3.54 (1.80–6.96)
Open in a new tab

Abbreviations: ART, assisted reproductive technology; GH, gestational hypertension; IUI, intrauterine insemination; NA, not applicable; OD, oocyte donation; OI, ovulation induction; PE, preeclampsia; PHD, all pregnancy‐related hypertensive disorders; RR, Mantel‐Haenszel risk ratios for random effects model. aStudies reporting pooled (mixed) results for in vitro fertilization (IVF) and intracytoplasmic insemination (ICSI). Heterogeneity is presented by the I 2 statistic and P value, which is P for the heterogeneity test χ2 Q statistic; P value, P for the heterogeneity between the effects of singleton and nonsingleton pregnancies.

However, IVF treatment alone was significantly associated with an increased relative risk of PE and PHD (Table 3) when all orders were considered together and when the analyses were restricted to singletons. In contrast, in nonsingleton pregnancies, IVF treatment demonstrated no significant effect for both outcomes. We also observed a significant difference in PHD risk estimates following IVF between singleton and nonsingleton pregnancies (P<.001).

ICSI treatment alone was not significantly associated with PE (Table 3), but this finding was limited by the small number of available studies. The effect of ICSI alone on PHD resulted in a significant increase in relative risk. Additionally, a significant difference in the pooled effect sizes between singleton and nonsingleton pregnancies for PHD (P=.038) was observed.

Studies reporting mixed results on IVF and ICSI (Table 3) demonstrated a significant risk increase in both PE and PHD for all gestation orders. However, in singletons, no analysis of PHD was possible (one study),S1 and the PE risk was not significant (two studies).S1,S43

OD in “all order” cohorts resulted in a 3.6 and 4.1 times increased risk of PE and PHD, respectively, compared with naturally conceived pregnancies, a finding that was also confirmed in three cohorts reporting separate data on singletons (Table 3).

PE risk was evaluated in subgroups in two studies,S43,S49 comparing either fresh cycles of single embryo transfer or double embryo transfer with spontaneous conceptions: the risks were +55% (95% CI, +37% to +74%; P heterogeneity, .76) and +67% (95% CI, +42% to +96%; P heterogeneity, .48), respectively. No analyses were performed for transfers after frozen cycles because the two available studiesS9,S49 reported results for different outcomes (PHD and PE, respectively).

Following IUI (Table 3), the relative risk of PE was found to increase by 66%, a finding that was almost identically reproduced in singleton pregnancies, while an analysis of nonsingleton pregnancies was not feasible. Additionally, no analyses were performed to test the effects of IUI on PHD and GH because only two studies on the formerS9,S47 and one on the latterS47 outcome investigated this issue, and they only included a few participants.

3.3. Sensitivity analyses in invasive ART

We observed that higher‐quality studies (n=35, intermediate and high quality together; Table S5) of invasive ART were associated with not different results of increased GH, PE, and PHD risk, compared with all studies presented in Figure 2. Also, among the 19 cohorts with extensive matching (ie, matched for >3 confounders), the results did not change compared with those presented in Figure 2. In the same Table S5, three selective sensitivity analyses were performed for studies including: those matched for baseline BMI,S4,S22,S23 those matched for parity,S1,S4,S6,S10,S13,S18,S22,S23,S28,S32,S37,S39,S44,S47,S55,S56–S58,S60 and those excluding chronic hypertension.S4‐S6,S8,S10,S13,S16,S30,S32,S35,S39,S48,S55,S62,S62–S64 In all cases, invasive ART was associated with increased incidence of PE and PHD beyond matching for baseline BMI or parity, as well as independently of chronic hypertension exclusion.

3.4. Influence of studies on pooled effect sizes, fixed effect models, and publication bias

In some of the analyses, excluding one study at a time, moved the point estimate of the combined effect outside the overall estimate of all available studies. More specifically, in the IUI singleton analysis (Table 3), the study by Hoy and colleaguesS17 was an outlier. In the sensitivity analysis excluding the outlier study, the relative risk increased by 25% (2.06; 1.07–3.98). We observed the same phenomenon with the study by Malchau and colleaguesS27 in the ICSI alone analysis presented in Table 3 for PHD in nonsingletons. No sensitivity analysis was performed because only two studies were available. In the analyses with low heterogeneity (P value >.1), a fixed‐effect model was also applied (Table S6). In most cases, the results were not substantially different from those obtained by random effects analysis.

To assess publication bias, reference is made to Table S7 and Figure S2. Although the graphic representations could not definitively exclude publication bias, significant bias was rejected by the trim and fill method.

4. Discussion

This overview was aimed at investigating whether and to what extent ART treatments as a whole and each individual ART alone increase the risk of pregnancy‐related hypertensive complications compared with spontaneous conceptions. The main finding was that hypertensive disorders in pregnancy, as summarized by the sum of GH and PE, are steadily increased by 54% following any invasive ART. This relative risk increase was extended in both singleton and nonsingleton pregnancies (47% and 33%, respectively). Also, noninvasive ART—namely OI—is associated with increased incidence of hypertensive complications. Although the baseline risk of hypertensive disorders in nonsingleton pregnancies was two to three times higher than that in singletons, the absolute risk increase was not different in higher gestation orders. In singleton pregnancies, all invasive ART modalities considered separately were accompanied by an increased risk of all outcomes. In contrast, in nonsingleton pregnancies, the increase in hypertensive complications was not uniformly increased because additional obstetric complications might mask the underlying PHD.

This is a larger, more focused and comprehensive overview on the association between ART and hypertensive disorders in pregnancy compared with previous meta‐analyses on the same issue. It is larger because of the number of included studies (66 studies on 7 038 029 pregnancies) independent of gestation order, it is more focused because the outcomes are limited to pregnancy‐related hypertensive complications, and it is more comprehensive because it provides an extensive network of analyses to assess both relative risk and change in absolute risk following ART compared with spontaneously conceived pregnancies.

With respect to the two major meta‐analyses restricted only to singletons by Jackson and colleagues18 (8 studies on 219 382 pregnancies; outcome PE) and Pandey and colleagues8 (15 studies on 606 314 pregnancies; outcome PHD), our analysis included 37 studies of singleton cohorts (17 cohorts of 2 961 248 pregnancies; outcome PE and 25 cohorts of 1 785 566 pregnancies; outcome PHD). We considered all 15 studies included in Pandey and colleagues (singletons by IVF or ICSI),8 but for singletons, we included 22 additional cohorts mediated by any invasive ART.S4,S8,S12,S15,S17,S23,S27,S28,S34,S37,S39,S40,S43,S52,S53,S55,S57,S59,S62,S64–S66 Thus, focusing on singletons, despite the larger number of studies, our risk estimates are similar to those reported by Jackson and colleagues18 for PE, ie, +62% in Jackson and colleagues vs 65% in our study, and in Pandey and colleagues8 for PHD, +51% vs +47%, respectively. Our specific analyses on PHD risk following IVF or ICSI in singletons did not yield substantially different results (+45% and +38%, respectively) compared with the analysis by Pandey and coworkers.8 In contrast to the above‐mentioned reviews, we also estimated that the risk of GH following ART in singletons was almost two times higher than that of spontaneous pregnancies. Two recent meta‐analyses from Qin and colleagues19, 20 in a limited number of studies demonstrated that PHD in singleton and multiple pregnancies is increased by 30% and 11%, respectively. These rates are lower compared with the estimates of the present and previous reviews possibly because of the different meta‐analytical methods used including studies selection.

The resulting PE risk following OD was almost four‐fold higher compared with spontaneous pregnancies, which is consistent with the findings of two previous reviews21, 22 focused on the same issue. Compared with these two previous reviews,21, 22 we excluded the study by Henne and colleagues23 because the naturally conceived arm was contaminated by an alternative ART, while compared with the more recent review by Masoudian and colleagues,22 we included three more studies.S14,S34,S54 To determine whether the risk of pregnancy‐related hypertensive complications following OD compared with all other ART procedures is higher or not, a direct head‐to‐head comparison should be used.18 However, we acknowledge that women undergoing OD demonstrate a higher propensity to develop pregnancy‐related hypertension because of their older reproductive age.

We also performed separate analyses of pregnancy‐related hypertensive complications to assess the risk following minimally invasive ART (IUI) and noninvasive ART (OI). In both cases, we observed an increased risk of PE (singletons: +48% for OI and +66% for IUI).

Although the relative risk increase at first glance appears to be lower in nonsingleton than in singleton pregnancies, the baseline absolute risk (the event rate of each individual outcome calculated in the spontaneously induced pregnancy arm) was two‐ to three‐fold higher in the former case. However, the absolute risk change in all hypertensive‐related outcomes following ART did not differ between singleton and nonsingleton pregnancies.

4.1. Strengths and limitations

Our overview and meta‐analyses have both strengths and limitations. The major strengths include the number of studies included; the comprehensive inclusion criteria, which considered all studies of any ART‐mediated pregnancies compared with spontaneous pregnancies, thus avoiding strict but often arbitrary inclusion criteria; and the exclusion of studies of head‐to‐head comparisons between different ARTs because these studies aimed to answer a different clinical question, namely, which ART procedure is associated with a lower rate of pregnancy‐related hypertensive disorders than another. A further strength is that all analyses were stratified by gestation order, as the higher the order of gestation, the higher the event rate. Moreover, pregnancy‐related hypertensive disorders were also considered separately because GH and PE are characterized by different prognosis. Finally, we systematically assessed the absolute risk increase and the “number needed to harm” to provide a better understanding of the risk distribution among singleton and nonsingleton pregnancies characterized by a great difference in baseline outcome risk.

Our analyses also have limitations. In our analyses, we not only considered studies on any invasive ART, but we also included a few studies that enrolled women who opted for minimally invasive ART (IUI) and/or noninvasive ART. However, in different ART studies, the extent of OI implementation in women in the spontaneously conceived arm was unknown. Another limitation is that we included cohort (partial) data of individual studies stratified by order of gestation. To include each individual study in its integrity, we also had to impute a mixed gestation order and conduct separate analyses jointly with other studies that did not provide data stratified by order of gestation (mixed order). Since the definition of discrete hypertensive disorders changes over time, bias on outcome detection remains. However, the present results are not different than those reported in reviews performed 10 years ago.8, 18 Although we ran separate sensitivity analyses controlling for different confounders including the extent of matching procedure used in individual studies, a causal association between ART and pregnancy‐related hypertensive complications is difficult to establish because of the different design of included studies and the different ART protocols applied.

4.2. Clinical and research implications

Although ART is accompanied by an increased relative and absolute risk of pregnancy‐related hypertensive complications, we should also acknowledge the treatment benefits resulting from ART. Since pregnancies especially at advanced age are accompanied by additional cardiovascular risk factors (preconception hypertension, obesity, and diabetes mellitus), medical advice and treatment for these factors before ART treatment might at least partially counterbalance the increased hypertensive propensity of assisted pregnancies. ART pregnancies, especially nonsingleton ones, should be managed under close surveillance not only by obstetricians but also by specialists in pregnancy‐related hypertensive disease to promptly identify warning signs of incident hypertensive complications.

No specific recommendation could be made based on our results regarding which individual ART treatment is associated with a higher risk of pregnancy‐related hypertensive complications because this question should be answered by head‐to‐head comparison studies between different ART procedures and resulting meta‐analysis. However, the body of evidence comparing directly different ART strategies is modest and mostly hidden in the studies identified for the present analysis, which was focused on the comparison between any ART and naturally conceived pregnancies. Also, future studies should control for important confounders (eg, age, parity, booking body mass index, chorionicity, preexisting hypertension, and vanishing embryo phenomenon) and examine GH and PE separately rather than provide pooled results of PHD.

Although in recent years more complex and sophisticated ART procedures have been accompanied by higher rates of successful pregnancy induction, the rate of adverse obstetric events including pregnancy‐related hypertensive disorders remains significant. A more refined research in the field particularly based on integration of data from national ART registries could indicate which clinical practice is associated with better outcomes.

Conflicts of Interest

The authors declare no conflicts of interest regarding this overview and meta‐analyses, but C.Th. declares consultancy fees from AstraZeneca and lecture honoraria from Sanofi, MSD, and Servier. C.Ts. served as a consultant and received modest honoraria from St Jude Medical, but he is not a stakeholder or shareholder in the company.

Authors’ Contributions

The corresponding author (C.Th.) is responsible for the design of the study and preparation of the first draft of the manuscript; C.Th., G.Sa., K.K., and I.A. performed the systematic review of the literature and C.Th. and G.Sa. extracted the data; C.Th. conducted the meta‐analyses; and all authors (C.Th., G.Sa., K.K., I.A., H.M., G.Sk., S.A., O.A., C.Ts., T.Ma., and E.S.) substantially contributed to the interpretation of data, critical revision of the manuscript for important intellectual content, and final approval of the manuscript to be published. The corresponding author (C. Th.) takes responsibility for the integrity of the analyses.

Supporting information

 

Click here for additional data file. (277.6KB, docx)

Acknowledgments

The valuable assistance of Ms Sofia Zisi in the provision of the requested literature is gratefully acknowledged.

Thomopoulos, C. , Salamalekis, G. , Kintis, K. , Andrianopoulou, I. , Michalopoulou, H. , Skalis, G. , Archontakis, S. , Argyri, O. , Tsioufis, C. , Makris, T. K. and Salamalekis, E. (2017), Risk of hypertensive disorders in pregnancy following assisted reproductive technology: overview and meta‐analysis. Journal of Clinical Hypertension, 19:173–183. doi: 10.1111/jch.12945

References

  • 1. Saleem S, McClure EM, Goudar SS, et al., and the Global Network Maternal Newborn Health Registry Study Investigators . A prospective study of maternal, fetal and neonatal deaths in low‐ and middle‐income countries. Bull World Health Organ. 2014; 92: 605–612. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2. Romundstad PR, Magnussen EB, Smith GD, Vatten LJ. Hypertension in pregnancy and later cardiovascular risk: common antecedents? Circulation. 2010;122:579–584. [DOI] [PubMed] [Google Scholar]
  • 3. Malizia BA, Hacker MR, Penzias AS. Cumulative live‐birth rates after in vitro fertilization. N Engl J Med. 2009;360:236–243. [DOI] [PubMed] [Google Scholar]
  • 4. Roberts JM. Pre‐eclampsia: new approaches but the same old problems. Am J Obstet Gynecol. 2008;199:443–444. [DOI] [PubMed] [Google Scholar]
  • 5. Källén B, Finnström O, Lindam A, Nilsson E, Nygren KG, Otterblad Olausson P. Trends in delivery and neonatal outcomes after in vitro fertilization. Hum Reprod. 2010;25:1026–1034. [DOI] [PubMed] [Google Scholar]
  • 6. Magli CM, Van den Abbeel E, Lundin K, Royere D, Der Van Elst J, Gianaroli L, For Committee of the Special Interest Group on Embryology . Revised guidelines for good practice in IVF laboratories. Hum Reprod. 2008;23:1253–1262. [DOI] [PubMed] [Google Scholar]
  • 7. Thomopoulos C, Tsioufis C, Michalopoulou H, Makris T, Papademetriou V, Stefanadis C. Assisted reproductive technology and pregnancy‐related hypertensive complications: a systematic review. J Hum Hypertens. 2013;27:148–157. [DOI] [PubMed] [Google Scholar]
  • 8. Pandey S, Shetty A, Hamilton M, Bhattacharya S, Maheshwari A. Obstetric and perinatal outcomes in singleton pregnancies resulting from IVF/ICSI: a systematic review and meta‐analysis. Hum Reprod Update. 2012;18:485–503. [DOI] [PubMed] [Google Scholar]
  • 9. Redman CWG, Sargent IL. Immunology of pre‐eclampsia. J Reprod Immunol. 2010;63:534–543. [DOI] [PubMed] [Google Scholar]
  • 10. Staff AC, Benton SJ, von Dadelszen P, et al. Redefining pre‐eclampsia using placenta‐derived biomarkers. Hypertension. 2013;61:932–942. [DOI] [PubMed] [Google Scholar]
  • 11. Sibai B, Dekker G, Kupferminc M. Pre‐eclampsia. Lancet. 2005;365:785–799. [DOI] [PubMed] [Google Scholar]
  • 12. Allen VM, Wilson RD, Cheung A. Pregnancy outcomes after assisted reproductive technology. J Obstetr Gynaecol Can. 2006;28:220–223. [DOI] [PubMed] [Google Scholar]
  • 13. Tranquilli AL, Dekker G, Magee L, et al. The classification, diagnosis and management of the hypertensive disorders of pregnancy: a revised statement from the ISSHP. Pregnancy Hypertens. 2014;4:97–104. [DOI] [PubMed] [Google Scholar]
  • 14. Moher D, Liberati A, Tetzlaff J, Altman DG, The PRISMA Group . Preferred Reporting Items for systematic reviews and meta‐analyses: the PRISMA statement. PLoS Med. 2009;6:e1000097. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. American College of Obstetricians and Gynecologists , Task Force on Hypertension in Pregnancy . Hypertension in pregnancy: report of the American College of Obstetricians and Gynecologists’ task force on hypertension in pregnancy. Obstet Gynecol. 2013;122:1122–1131. [DOI] [PubMed] [Google Scholar]
  • 16. American College of Obstetricians and Gynecologists Committee on Obstetric Practice . ACOG practice bulletin, no. 33. Diagnosis and management of preeclampsia and eclampsia. Int J Gynaecol Obstet. 2002;77:67–75. [PubMed] [Google Scholar]
  • 17. Sibai BM, Hauth J, Caritis S, et al. for the National Institute of Child Health and Human Development Network of Maternal‐Fetal Medicine Units. Am J Obstet Gynecol. 2000;182:938–942. [DOI] [PubMed] [Google Scholar]
  • 18. Jackson RA, Gibson KA, Wu YW, Croughan MS. Perinatal outcomes in singletons following in vitro fertilization: a meta‐analysis. Obstetr Gynecol. 2004;103:551–563. [DOI] [PubMed] [Google Scholar]
  • 19. Qin J, Wang H, Sheng X, Liang D, Tan H, Xia J. Pregnancy‐related complications and adverse pregnancy outcomes in multiple pregnancies resulting from assisted reproductive technology: a meta‐analysis of cohort studies. Fertil Steril. 2015;103:1492–1508. [DOI] [PubMed] [Google Scholar]
  • 20. Qin J, Liu X, Sheng X, Wang H, Gao S. Assisted reproductive technology and the risk of pregnancy‐related complications and adverse pregnancy outcomes in singleton pregnancies: a meta‐analysis of cohort studies. Fertil Steril. 2016;105:73–85. [DOI] [PubMed] [Google Scholar]
  • 21. Pecks U, Maass N, Neulen J. Oocyte donation: a risk factor for pregnancy‐induced hypertension. Dtsch Aztebl Int. 2011;108:23–31. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 22. Masoudian P, Nasr A, de Nanassy J, Fung‐Kee‐Fung K, Bainbridge SA, El Demellawy D. Oocyte donation pregnancies and the risk of preeclampsia or gestational hypertension: a systematic review and metaanalysis. Am J Obstet Gynecol. 2016;214:328–339. [DOI] [PubMed] [Google Scholar]
  • 23. Henne MB, Zhang M, Paroski S, Kelshikar B, Westphall LM. Comparison of obstetric outcomes in recipients of donor oocytes versus women of advanced maternal age with autologous oocytes. J Reprod Med. 2007;52:590–595. [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

 

Click here for additional data file. (277.6KB, docx)

Articles from The Journal of Clinical Hypertension are provided here courtesy of Wiley

RESOURCES