Placenta previa and the risk of intrauterine growth restriction (IUGR): a systematic review and meta-analysis

Sources

We performed a Medline, EMBASE, Google Scholar, Scopus, ISI Web of Science and Cochrane database search, as well as PubMed (www.pubmed.gov) search until the end of December 2018 from the last 30 years using the following Boolean search criteria: “(Placenta previa OR previa) AND (intrauterine growth restriction OR IUGR) OR (fetal growth restriction OR FGR) OR (small for gestational age OR SGA) OR (low birth weight OR LBW)”. Other than restricting the search to human studies and to articles in English, other limiting categorical terms used were the restriction to studies with singleton pregnancies and without any evidence of aneuploidy, congenital anomalies or placental trophoblastic invasion (accreta spectrum). The reference lists and bibliographies of included studies were then searched for other salient and pertinent manuscripts. Finally, manual searches of studies belonging to research teams having prior publications on placenta previa and growth restriction were undertaken, and other pertinent studies retrieved. This review was modeled on the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement, and search flowchart depicting the search strategy is illustrated in Figure 1.

Figure 1:

PRISMA flowchart for the systematic review of studies.

Study selection

All authors independently examined the electronic search results for reports of possibly relevant studies and those reports were retrieved and analyzed in further detail. Published studies were eligible for inclusion if they compared the risk of IUGR/SGA between cases of placenta previa and cases of normal placental location. In undertaking our search, we defined IUGR/SGA as birth weight below the 10^th percentile, regardless of the terminology used in individual studies. Indeed, we included studies, which used terminology such as “small for gestational age”, “fetal growth restriction” and “low birth weight”, provided that they referred to neonates born under the 10^th percentile corrected for gestational age. In addition, we made no distinction between the mode of diagnosis of placenta previa, the placental orientation, the presence of antepartum bleeding, nor the year or country of origin of the study in question.

All studies were assessed following predetermined quality criteria including the presence of a power calculation, the unit of analysis used and risk of bias. Risk of bias was assessed using the Cochrane Handbook for Systematic Reviews of Interventions (results not shown). We attempted to contact one author to obtain supplementary information regarding their study, but did not receive a response. As such, we did not include the study in question into our analysis for lack of complete information.

We used odds ratios (OR) and a fixed effects (FE) model with the Mantel-Haenszel method to calculate weighted estimates and their 95% confidence intervals (CI) where appropriate. A forest plot and a funnel plot are provided for visualization of the results (Figures 2 and 3). Statistical heterogeneity between results of studies was examined by inspecting the scatter in the data points on the graphs and the overlap of CIs, and by checking the χ² and I² statistics. The Review Manager 5.3.5 software (Version 5.3. Copenhagen: The Nordic Cochrane Centre, The Cochrane Collaboration, 2014) was used to combine data for the meta-analysis. This meta-analysis was exempt from Institutional Review Board (IRB) approval because of the nature of the research design (review article), as well as the lack of use of identified patient data.

Figure 2:

Forest plot – risk of IUGR/SGA in placenta previa.

Figure 3:

Funnel plot.

Clinical implications of the findings

The management of pregnancies at risk of IUGR includes the provision of prenatal screening, the adequate assessment of fetal anatomy, the continuous monitoring of fetal biometry throughout gestation and, in select cases, the provision of anticoagulation or anti-platelet therapy [24]. As with most normal pregnancies, before a placenta previa is diagnosed, the vast majority of patients will have undergone regular prenatal screening and evaluation of fetal anatomy. Given the need to re-assess placental location and its distance to the internal os in the third trimester to plan for delivery, fetal growth is already likely to be estimated at that time again. Moreover, given the substantial risk of antepartum hemorrhage, whether the provision of aspirin and anticoagulation in women with placenta previa increases such risk has not been studied. Therefore, at the present time, the findings of this study are unlikely to alter the current clinical management of these pregnancies.

That said, once IUGR is established, fetal Doppler studies are recommended to screen for signs of placental insufficiency and fetal distress [24]. Depending on the etiology and severity of IUGR, delivery is likely to be induced around 38–39 weeks of gestation. However, given the risk of catastrophic hemorrhage, the Eunice Kennedy Shriver National Institute of Child Health and Human Development and the Society for Maternal-Fetal Medicine recommend that pregnancies with uncomplicated placenta previa be delivered between 36+0 and 37+6 weeks of gestation [25]. Therefore, at the present time, there is no evidence that specifically monitoring fetal growth or Doppler studies with serial ultrasound examinations is useful in cases of placenta previa. Given the findings of this study, which only suggest a mild increase in the risk of IUGR/SGA, we do not propose that further screening with serial scans be undertaken at this time. Nevertheless, what the findings of this study may in fact prompt is a more comprehensive counseling to patients, as well as to the NICU team, who may care for the neonates with this condition, often born in the late-preterm or early-term period.

Strengths and limitations

The strengths of this study are multiple and include the vast number of patients analyzed, the homogeneity of the inclusion criteria across studies, as well as the use of proper statistical techniques. Similarly, to the best of our knowledge, ours is the first and only meta-analysis in the literature addressing the association between placenta previa and growth restriction, an increasingly important and relevant question in clinical medicine today. On the other hand, the limitations of our study are several and worth mentioning. First, while all of the studies included in the meta-analysis addressed the outcome of IUGR, the definitions are not strictly identical and as such, they may introduce bias into our estimates. Indeed, the literature is notoriously heterogeneous with regard to terminology about growth restriction, with some reports considering growth under the 10^th percentile as diagnostic while others do so in cases under the fifth or third centile. That said, all weight estimates were sonographically determined and the same criteria were used for all patients in each individual study. Likewise, though individual studies may use a different terminology, be it IUGR or SGA, all refer to fetuses whose growth falls below the 10^th percentile for gestational age. Second, though the pooled number of patients is high, statistical heterogeneity between studies was large as well, with an I² of 94%. Third, given the retrospective nature of the study designs included in the meta-analysis, adjusting for different confounders in individual studies may have introduced information bias. To mitigate these, we conducted sensitivity analyses looking at univariate and multivariate estimates in individual studies, and our findings and conclusion remained unchanged. Finally, despite the robustness of our findings, we cannot prove a causal relationship between placenta previa and IUGR/SGA.