Placenta 117 (2022) 21–27

Contents lists available at ScienceDirect

Placenta
journal homepage: www.elsevier.com/locate/placenta

The risk factors associated with placenta previa: An umbrella review

Ensiyeh Jenabi a, *, Zohreh Salimi a, **, Saeid Bashirian b, Salman Khazaei c, Erfan Ayubi b
a
Autism Spectrum Disorder Research Center, Hamadan University of Medical Sciences, Hamadan, Iran
b
Social Determinants of Health Research Center, Hamadan University of Medical Sciences, Hamadan, Iran
c
Research Center for Health Sciences, Hamadan University of Medical Sciences, Hamadan, Iran

A R T I C L E I N F O A B S T R A C T

Keywords: We evaluated in this umbrella review a systematic collection from meta-analyses conducted on risk factors
Placenta previa associated with placenta previa.
Risk factor We searched PubMed, Scopus, and Web of Science until April 2021 assessing the risk factors associated with
Umbrella review
placenta previa. We calculated summary effect estimates odds ratio, relative risk, 95% CI, heterogeneity I2, 95%
prediction interval, small-study effects, excess significance biases, and sensitive analysis. The quality of the meta-
analyses was evaluated with AMSTAR 2.
We included nine studies in the present umbrella review. Seven risk factors including prior induced abortion
(OR 1⋅36, 95% CI: 1⋅02, 1⋅69), prior spontaneous abortion (OR 1⋅77, 95% CI: 1⋅60, 1⋅94), male fetus (OR 1⋅2,
95% CI: 1⋅2, 1⋅3), smoking (OR 1⋅42, 95% CI 1⋅30, 1⋅54) (RR 1⋅27, 95% CI: 1⋅18, 1⋅35) advanced maternal age
(OR 3⋅16, 95% CI: 2⋅79, 3⋅57), cesarean (OR 1⋅60, 95% CI: 1⋅44, 1⋅76) and ART (singleton pregnancy) (RR 3⋅71,
95% CI: 2⋅67, 5⋅16) were graded as highly suggestive evidence (class III). Endometriosis (OR 3⋅03, 95% CI: 1⋅50,
6⋅13) and maternal cocaine use (OR 2⋅9, 95% CI: 1⋅9, 4⋅3) were graded as risk factors with weak evidence (class
IV).
This study provides suggestive evidence about prior spontaneous abortion, prior induced abortion, male fetus,
smoking, advanced maternal age, cesarean section, and assisted reproductive techniques (singleton pregnancy)
as risk factors associated with placenta previa.

1. Introduction
Placenta previa is an obstetric complication with a prevalence of 5.2
per 1000 pregnancies [1]. In this condition placenta partially or fully
obstructs the internal orifice of the cervix [2]. Placenta previa was
classified into four types include low-lying placenta, marginal, partial,
and complete placenta previa [3].
The pathogenesis of the disease is not clearly understood. Placenta
previa is connected with adverse outcomes of maternal and fetal,
including adherence of the placenta, antepartum hemorrhage, post­
partum hemorrhage, malpresentation, intrauterine growth restriction
(IUGR), thrombophlebitis, preterm labor, and septicemia [2,4].
The systematic review and meta-analysis research indicated that risk
factors for placenta previa were maternal advanced age, assisted
reproductive technology, smoking, prior cesarean section, and endo­
metriosis [2,5–7]. However, these investigations were frequently
restricted to a specific issue and their results could be affected by biases,
such as excess significance bias and publication bias [8]. Furthermore,
hierarchies of evidence have not been defined across different factors.
Therefore, we evaluated in this umbrella review a systematic collection
from meta-analyses and systematic reviews conducted on risk factors
associated with placenta previa.
2. Materials and methods
PRISMA reporting guideline was used in this umbrella review [9].
The stages of screening (title-abstract and full paper), data extraction,
and methodological appraisal of included studies were carried out by
two independent authors. The protocol of the present umbrella review
was prospectively registered in PROSPERO with the number:
CRD42021252802.

* Corresponding author.
** Corresponding author.
E-mail addresses: en.jenabi@yahoo.com, e.jenabi@umsha.ac.ir (E. Jenabi), z.salimi@edu.umsha.ac.ir (Z. Salimi), bashirian@umsha.ac.ir (S. Bashirian),
salmankhazaei61@gmail.com (S. Khazaei), aubi65@gmail.com (E. Ayubi).

https://doi.org/10.1016/j.placenta.2021.10.009
Received 10 August 2021; Received in revised form 8 October 2021; Accepted 12 October 2021
Available online 20 October 2021
0143-4004/© 2021 Elsevier Ltd. All rights reserved.
E. Jenabi et al.
2.1. Search strategy
We systematically searched PubMed, Scopus, and Web of Science
until April 26, 2021. The search strategy including search terms used is
included in the S1. Two authors (EJ and SK) independently identified
eligible articles and reviewed the title-abstract and full texts (Fig. 1).
Forward and backward searching was performed for the identified pa­
pers, i.e. the references of eligible studies were manually searched to
identify more articles, and the authors of the eligible papers were con­
tacted for their potential related papers (published or under publica­
tion). Any disagreement was solved by SB.
2.2. Inclusion and exclusion criteria
In the present umbrella review, all meta-analyses that had focused on
assessing the risk factors associated with placenta previa, regardless of
the publication date or the status of publication were included, if they
had provided meta-analyses based on observational studies (cohort,
case-control, and cross-sectional studies). The meta-analyses included
the studies that the diagnosis of placenta previa was approved by ul­
trasound during pregnancy.
The articles that did not determine the risk factors of diagnosing
placenta previa were excluded. In addition, systematic reviews that did
not constitute a meta-analysis were excluded. Also, if the information
needed for reanalysis was not included or could not be retrieved
otherwise, the meta-analysis was excluded. We excluded animal studies
and genetic studies. Both journal papers and conference full papers were
accepted, but conference abstracts were excluded. To avoid overlap,
when more than one meta-analysis discussed the same risk factor(s),
only one meta-analysis was selected: In the first round, we prioritized
the meta-analyses that have adjusted for confounder variables. In the
next round (if needed), we prioritized the meta-analyses that had the
highest quality according to AMSTAR 2. When two or more meta-
analyses existed for one research topic, we prioritized the most recent
Fig. 1. The process of the included meta-analyses in umbrella review.
22
Placenta 117 (2022) 21–27
meta-analysis which normally has the largest number of studies. We
have listed the meta-analyses excluded in the stage full papers screening
in the S2.
2.3. Data extraction
For this umbrella review, EJ and SK performed the database
searching and identified the included meta-analyses. EJ extracted the
information and EA checked the extracted data. Disagreement between
them was resolved by SB. EJ and SK assessed the quality of the studies
and ZS performed the statistical analyses. Study investigators were
contacted for unreported data or missing data.
Two authors independently assessed each eligible article and
extracted the below information: First author’s name, publication year,
risk factor, information regarding the effect size, heterogeneity, sample
size, number of study estimates, p-values, study design, and methodol­
ogy of the included papers, their participant demographics and baseline
characteristics, metrics used in the original analyses (odds ratio, related
risk, hazard ratio), and every risk factor relevant to placenta previa.
Also, the included papers in each review were retrieved if all the
necessary information for sensitivity analysis and prediction interval of
every meta-analysis was not included, such as the 2 × 2 table of OR or
RR. Storage of information and records were on Microsoft Excel and
statistical analyses were performed using R version 4.0.5 and packages
such as Metafor, ConfoundedMeta, and xlsx.
Each eligible meta-analysis was reanalyzed using the extracted in­
dividual study estimates based on metrics in the original meta-analyses.
The software of R version 4.0.5 was used for the analyses. Metafor
package was used for the analyses along with other packages such as
ConfoundedMeta and xlsx. Odds ratio and relative risk of all reported
risk factors for placenta previa were obtained over all the meta-analyses
using a random-effect model, and the summary effect estimates and p-
values of eligible meta-analyses were reanalyzed. Statistical significance
was calculated at p < 0⋅05. Cochran’s Q test was conducted for
E. Jenabi et al.
calculating the I2 statistic for heterogeneity between studies (I2>50%
indicates high heterogeneity) [10].
We estimated the 95% prediction interval, the range in which we
expect the effect of the risk factor will lie for 95% of future studies [11],
using Metafor codes, by extracting the 2 × 2 table of exposure-outcome
of each of the studies included in each meta-analysis.
Using the regression asymmetry test by Egger et al., We evaluated the
effects of small studies, such as large articles that have more conserva­
tive results than smaller articles [12]. There was a small study effect if
the Egger test had p < 0⋅1.
For the statistically significant meta-analysis, we used the test for
excess significance bias, which compared the expected numbers versus
the observed number of statistically significant individual studies (p-
value < 0⋅05) [13].
We performed the sensitivity analysis for the included meta-analyses
to account for potential methodological limitations of observational
studies that might result in spurious significance, using Mathur’s
method [14]. In this method, magnitudes for bias factor and con­
founding association strength are obtained, employing a given threshold
of acceptable effect size and a satisfying percentage of studies included
in the meta-analysis that pass that threshold, and judgments are made
about the robustness of the meta-analysis to confounding based on those
values. Credibility ceilings are not employed for sensitivity analysis in
this review, as the soundness of that method is argued by Mathur et al.
[15].
For each eligible article, EJ and SK independently evaluated the
quality of the meta-analyses using AMSTAR 2 [16]. In this form, 16
items are included that address:
1 PICO consideration in the research question and inclusion criteria
2 Protocol establishment beforehand. Any deviations?
3 Explaining if/why only certain study designs were included
4 Comprehensive search
5 Two persons performed the search
6 Two persons extracted the data
7 Providing the list of exclusion with reasons
8 Presenting all details of the included papers
9 Proper technic for assessing the risk of bias
10 Reporting sources of funding
11 Appropriate statistical methods
12 Assessment of the potential impact of risk of bias in individual
studies on the results of the meta-analysis
13 Assessment of the potential impact of risk of bias in individual
studies on the discussion of the meta-analysis
14 Discussion of heterogeneity of the results of the meta-analysis
15 Investigation of publication bias. Have they influenced the
results?
16 Reports of potential conflict of interest
Each of the above items was scored as yes (Y), partial yes (PY), or no
(N). Out of the 16 items, items 2, 4, 7, 9, 11, 13, and 15 are the critical
items. The reviews are rated as high, moderate, low, or critically low,
based on the scores of the critical and non-critical items, as Table 1
shows. Disagreement between the assessors was resolved by SB.
Table 1
Guide to rating the systematic reviews, based on AMSTAR2.
Rating # non-critical items scored as N # critical items scored as N
High ≤1 0 according to random effect calculations
Moderate >1 0 Weak evidence (class IV)
Low Any 1 Non-significant
Critically low Any >1
23
Placenta 117 (2022) 21–27
2.4. Determining the credibility of evidence
Based on previous umbrella reviews [17,18], the strength of the
evidence of risk factors for placenta previa was divided into five classes:
convincing (class I), highly suggestive (class II), suggestive (class III),
weak (class IV), and not significant (NS) (Table 2). The p-value of the
random-effect model was <0.05, the number of placenta previa cases,
heterogeneity as I2, small-study effects, excess significance bias, effect
estimate under 10% credibility ceiling, and 95% prediction interval.
Sensitivity analyses were conducted to confirm the robustness of the
associations. We run this analysis to look for potential spurious signifi­
cance in the reported results that have been caused by any potential
biases or confounders that have been missed when running the
experiments.
3. Results
We identified 572 studies from different databases until April 26,
2021. In total, we included nine meta-analyses that were eligible for
inclusion in the present umbrella review (Fig. 1). These nine eligible
studies provided 13 meta-analyses included [2,5,7,19–24] with 154,483
placenta previa cases with 36,095,408 participants. Meta-analyses were
based on cohort or case-control design. In the present umbrella review,
159 original studies were included (112 cohort studies and 47 studies
based on case-control).
Fourteen risk factors were identified in the included reviews,
namely: prior induced abortion, prior spontaneous abortion, advanced
maternal age, assisted reproductive techniques (ART) (singleton preg­
nancy), ART (twin pregnancy), cesarean section, endometriosis, uterine
leiomyoma, smoking, maternal cocaine use, male fetus, chronic hyper­
tension, pregnancy-induced hypertension, and preeclampsia.
In the present umbrella review, out of the 14 associations, 10 asso­
ciations were statistically significant using the random-effects model,
eight included at least 1000 placenta previa cases, six reported hetero­
geneity (I2) less than 50%, three had small study effects, and six had
excess significance bias. Factors about ART (multiple pregnancy) could
not be calculated based on the extracted study estimates of the indi­
vidual studies (Table 4).
According to our sensitivity analyses, the results of eight meta-
analyses were relatively sensitive to unmeasured confounding, so that
a bias factor less than 1.75 in each of their included studies would be
able to reduce the percentage of studies with a true odds ratio of greater
than 1.1 to less than 20%. These factors that are sensitive to unmeasured
confounding are prior induced abortion, uterine leiomyoma, prior
Table 2
Determining the credibility of the evidence for meta-analyses of observational
studies.
classification Criteria
Convincing evidence 1 More than 1000 cases
(Class I) 2 Significant summary associations (p < 10− 6) per
random-effects calculations
3 No evidence of small-study effects (Egger<0.1)
4 No evidence of excess of significance bias
5 Prediction intervals not including the null value 6.
Largest study nominally significant (p < 0.05)
7 Not large heterogeneity (I2 < 50%)
8 Robust results based on sensitivity analysis
Highly suggestive 1 More than 1000 cases
evidence (Class II) 2 Significant summary associations (p < 10− 6) per
random-effects calculation
3 Largest nominally significant study (p < 0.05)
Suggestive evidence 1 More than 1000 cases
(Class III) 2 Significant summary associations (p < 10− 3)
1 All other associations with p < 0.05
1 All associations with p < 0.05
associations (NS)
E. Jenabi et al. Placenta 117 (2022) 21–27

spontaneous abortion, male fetus, preeclampsia, maternal cocaine use,
chronic hypertension, smoking, and pregnancy-induced hypertension,
while the four latter ones were not statistically confirmed as a risk factor.
The other four factors (advanced maternal age, ART (singleton preg­
nancy), cesarean section, and endometriosis) were relatively robust to
unmeasured confounding, considering a bias factor of more than 1.9 for
each of their included studies is needed to reduce the proportion of
studies with a true odds ratio greater than 1.1 to less than 10% (20% in
the case of smaller meta-analyses) (Table 4).
Seven risk factors including prior induced abortion (OR 1⋅36, 95%
CI: 1⋅02, 1⋅69), prior spontaneous abortion (OR 1⋅77, 95% CI: 1⋅60,
1⋅94), male fetus (OR 1⋅2, 95% CI: 1⋅2, 1⋅3), smoking (OR 1⋅42, 95% CI:
1⋅30, 1⋅54) (RR 1⋅27, 95% CI: 1⋅18, 1⋅35) advanced maternal age (OR
3⋅16, 95% CI: 2⋅79, 3⋅57), cesarean section (OR 1⋅60, 95% CI:
1⋅44–1⋅76) and ART (singleton pregnancy) (RR 3⋅71, 95% CI: 2⋅67,
5⋅16) were graded as suggestive evidence (class III). Endometriosis (OR
3⋅03, 95% CI: 1⋅50, 6⋅13) and maternal cocaine use (OR 2⋅9, 95% CI 1⋅9,
4⋅3) were graded as risk factors with weak evidence (class IV) (Table 3,
Fig. 2).
Chronic hypertension (OR 1⋅5, 95% CI 0⋅8, 2⋅7), pregnancy-induced
hypertension (OR 0⋅4, 95% CI: 0⋅2, 0⋅5), uterine leiomyoma (OR 1.49,
95% CI: 0.64, 2.35), ART (multiple pregnancy), and pre-eclampsia (OR
0⋅9, 95% CI: 0⋅5, 1⋅4) were not confirmed as risk factors for placenta
previa (not significant). Only four meta-analyses had effect sizes larger
than 2 (advanced maternal age, ART (singleton pregnancy), endome­
triosis, and maternal cocaine use), which were in classes III and IV, re­
gard to the strength of their evidence (Table 3).
The quality of all meta-analyses, based on AMSTAR 2, was critically
low (Tables 3 and S3).
4. Discussion
4.1. Main findings
To the best of our knowledge, this is the first umbrella review on
previously published systematic reviews and meta-analyses that aimed
to evaluate risk factors for placenta previa. Evidence from the umbrella
review suggests that prior induced abortion, prior spontaneous abortion,
male fetus, smoking, advanced maternal age, cesarean section, and
assisted reproductive techniques (singleton pregnancy) are associated
with an increased risk of placenta previa.
4.2. Interpretation
The definition of the type of placenta previa may be different across
individual studies in each meta-analysis. Clearly define consistent and
reliable studied variables is an important issue when interpreting and
analyzing an umbrella review [25]. Here, the important question is that
whether the diagnosis of placenta praevia from images (ultrasonography
and/or magnetic resonance imaging, MRI) and different classification
(complete, partial, marginal placenta praevia, low-lying placenta) are
similar across included individual studies in each meta-analysis. For
example among individual studies in the conducted meta-analysis by
Karami et al. [21] about the association between prior abortion and
placenta previa, in the one included study the diagnosis of placenta
praevia was based on both ultrasonography and MRI [26] but in
another, it was based on sonographic imaging [27].
The mechanisms involved in the association with risk factors for
placenta previa are unknown. However, there are hypotheses in this
regard. Coping with nicotine-induced hypoxia among women smokers
can increase the effective level for gas exchange. Therefore, the placenta
with larger parts is more likely to cover the internal cervix and lead to
placenta previa [28]. The injury and scarring to the myometrium and
endometrium of the uterus during previous abortions and cesarean may
influence the low implantation of the placenta in the uterus in subse­
quent pregnancies [2]. The underlying mechanisms about the associa­
tion between ART and placenta previa in singleton pregnancies are
unknown. There is a hypothesis that ARTs methods including drugs
applied to stimulate ovulation or maintain pregnancy in the early stages
of pregnancy or maternal factors related with infertility can increase the

Table 3
Risk factors for included meta-analysis in umbrella review.
Risk factors Source (year) Number of Number of Study design Effect metrics Random effect Credibility of AMSTAR2
population included studies summary estimate evidence quality

Prior induced Karami, 2017 62,459 10 Cohort/case- Odds ratio 1.36 (1.02, 1.69) Class III Critically low
abortion control
Prior spontaneous Karami, 2017 58,713 16 Cohort/case- Odds ratio 1.77 (1.60, 1.94) Class III Critically low
abortion control
Advanced maternal Martinelli, 21,961,192 23 Cohort Odds ratio 3.16 (2.79, 3.57) Class III Critically low
age 2018
ART * (singleton Qin, 2016 984623 12 Cohort Relative risk 3.71 (2.67, 5.16) Class III Critically low
pregnancy)
ART * (multiple Qin, 2015 4796 8 Cohort Relative risk 1.52 (0.94, 2.44) Cannot be Critically low
pregnancy) found**
Cesarean section Gurol-Urganci, 399,674 37 Cohort/case- Odds ratio 1.60 (1.44, 1.76) Class III Critically low
2011 control
Endometriosis Zullo, 2017 44,759 10 Cohort Odds ratio 3.03 (1.50, 6.13) Class IV Critically low
Uterine leiomyoma Jenabi, 2019 255,886 9 Cohort/case- Odds ratio 1.49 (0.64, 2.35) Class III Critically low
control
Smoking Shobeiri, 2016 9,094,443 21 Cohort/case- Odds ratio/ 1.42 (1.30, 1.54) Class III Critically low
control Relative risk 1.27 (1.18, 1.35)
Maternal cocaine use Faiz, 2003 55,562 3 Cohort/case- Odds ratio 2.9 (1.9, 4.3) Class IV Critically low
control
Male fetus Faiz, 2003 798,119 7 Cohort/case- Odds ratio 1.2 (1.1, 1.3) Class III Critically low
control
Chronic hypertension Faiz, 2003 152428 3 Cohort/case- Odds ratio 1.5 (0.8, 2.7) NS Critically low
control
Pregnancy-induced Faiz, 2003 171475 3 Cohort/case- Odds ratio 0.4 (0.2, 0.5) NS Critically low
hypertension control
Preeclampsia Faiz, 2003 37,922 3 Cohort/case- Odds ratio 0.9 (0.5, 1.4) NS Critically low
control

*ART: Assisted reproductive techniques.

**- Factors could not be calculated based on the extracted study estimates of the individual studies.
Class I: Convincing; class II: Highly suggestive; class III: Suggestive; class IV: Weak; NS: Not significant.

24
E. Jenabi et al. Placenta 117 (2022) 21–27

Table 4
The credibility of the evidence for meta-analyses included observational studies.
Risk factors Number summary small-study Excess of Prediction Largest study Heterogeneity Sensitivity Classification
of cases associations (p- effects (p- significance intervals nominally (I2%) analysis
value) per random- value for bias (p-value) significant (p
effects calculations Egger) -value)

Prior induced 2946 P < 0.0001 0.904 0.001 0.94–3.14 P < 0.01 59.2 T = 1.644, Class III
abortion G = 2.674
Prior 3036 P < 0.0001 0.770 0.038 1.0–6.55 P < 0.0001 0.0 T = 1.107, Class III
spontaneous G = 1.452
abortion
Advanced 113990 P < 0.0001 0.228 6.350 1.91–3.21 P < 0.0001 97.9 T = 2.113, Class III
maternal age G = 3.646
ART (singleton 2571 P < 0.00001 0.178 0.065 1.15–10.53 P < 0.001 81 T = 2.879, Class III
pregnancy) G = 5.205
ART (multiple 11 P = 0.09 0.91 * * None of the 0.0 * NS
pregnancy) studies had
significant
results
Caesarean 9532 P < 0.0001 0.040 1.930 1.01–4.19 P < 0.001 86.1 T = 2.060, Class III
section G = 3.537
Endometriosis 489 P = 0.0002 0.830 0.181 0.54–28.01 P = 0.005 75.8 T = 3.053, class IV
G = 5.567
Uterine 1271 P < 0.0001 0.097 0.139 1.12–4.14 P < 0.01 65.6 T = 1.738, Class III
leiomyoma G = 2.872
Smoking 16878 P < 0.0001 0.056 0.002 0.89–2.06 P < 0.01 34.6 T = 1.493, Class III
G = 2.352
Maternal 359 P < 0.0001 0.448 0.250 1.69–3.92 P < 0.0001 0.0 T = 1.738, class IV
cocaine use G = 2.872
Male fetus 3620 P < 0.001 0.443 0.027 0.99–1.51 P < 0.001 41.9 T = 1.107, Class III
G = 1.451
Chronic 722 P = 0.200 0.430 0.011 0.67–3.59 P = 0.327 17.32 T = 1.738, NS
hypertension G = 2.872
Pregnancy- 620 P = 0.451 0.788 3.52 0.10–4.56 P = 0.499 82.12 T = 1.676, NS
induced G = 2.740
hypertension
Preeclampsia 445 P = 0.546 0.334 0.032 0.30–2.27 P = 0.230 50.21 T = 1.451, NS
G = 2.261

ART: Assisted reproductive techniques.

Class I: Convincing; class II: Highly suggestive; class III: Suggestive; class IV: Weak; NS: Not significant.
*- Factors could not be calculated based on the extracted study estimates of the individual studies.

Fig. 2. Summary estimates of meta-analyses of potential environmental risk factors for placenta previa.

risk of adverse pregnancy outcomes [29,30]. However, this topic de­ interpreted with caution under any circumstances because these studies
serves further investigation. are affected by several types of biases. In other words, the validity of
The evidence from syntheses of observational studies should be umbrella review is the function of the validity of the included studies.

25
E. Jenabi et al.
For example, OR of 1.42 for the effect of smoking on placenta previa may
be far from causation, although it is along with the suggestive level of
credibility. As can be understood from the sensitivity analysis a degree
of substantial unmeasured confounding exists in the association be­
tween prior abortion and placenta previa T = 1.493, G = 2.352. It
argued that potential covariates in the latter association can be as
follow; the age of mother, multiparity, prior ART, prior cesarean section,
and prior operations on uterine cavity [27], however, included studies in
the meta-analysis by Karami et al. [21] have no attempted to adjust all
possible potential confounders.
Although etiology and biological mechanisms underlying placenta
previa are still not completely defined, however, some risk factors seem
to be superior. Some studies suggest that 37.5% of placenta previa is
related to previous cesarean section [31,32] and even in one study by
Matalliotakis et al. [33] 66% of women with placenta previa had a
history of previous cesarean section. With an emphasis on the effect of
cesarean section, this umbrella review showed that cesarean section
may increase the risk of placenta previa (suggestive; class III). When
considering the effect of cesarean section on the placenta previa, some
issues seem to be considered. e.g. the association between numbers of
previous cesarean sections and the subsequent development of placenta
previa or interaction of previous cesarean section and other risk factors.
The included meta-analysis in the umbrella review did not present in
a unique category of ‘ART and overall estimates were presented based on
ART in singleton and ART in multiple pregnancy. The overall estimated
effect of ART in multiple pregnancy (1.52) is more toward null
compared to that of ART in singleton pregnancies (3.71). It seems that
the effect of ART in singleton pregnancy and multiple pregnancy on
placenta previa are stronger than its effect on other pregnancy-related
complications. According to a meta-analysis by Qin et al. [22] sum­
mary effect estimates for the association between ART and
pregnancy-induced hypertension and ART and gestational diabetes
mellitus in singleton pregnancies were 1.30 and 1.31, respectively.
Moreover, according to a previous meta-analysis [34], the effect of ART
in multiple pregnancy on other pregnancy-related complications is
much closer to the null value. For example, 1.08 for preterm birth or
1.04 for low birth weight and even negative association were suggested
e.g. 0.85 for small for gestational age.
Our umbrella review indicates the associations of ART in singleton
pregnancy and endometriosis with placenta previa are similar in direc­
tion and strength (summary effect of 3.71 for ART and 3.03 for endo­
metriosis). The possible explanation for this similarity in the association
may be because infertility women with endometriosis require ART to
conceive. Moreover, one meta-analysis showed that the risk of placenta
previa among patients with endometriosis undergoing ART is three
times of those without endometriosis [35].
Although the credibility for the effect of smoking on placenta previa
was suggestive (class III), however, the results are still confusing due to
potential biases such as excess significance bias. Moreover, the differ­
ence between summary OR (1.42) and RR (1.27) was negligible, denotes
the probabilities of developing the placenta previa are small enough
among both smoking and non-smoking level. Due to the possible dose-
response manner of smoking with placenta previa [36], dose-response
meta-analysis can be helpful to determine the true effect of smoking
on placenta previa.
This umbrella review synthesizes the evidence on most determinants
of placenta previa. Since the etiology of placenta previa is multifactorial
[37,38], it may also feel to find best evidence on other determinants on
placenta previa such as gravidity, parity, history of drug abuse, and
previous placenta previa, etc … Although in the present review placenta
previa was the outcome of interest, however, the effect of placenta
previa on adverse outcomes has been noted in the literature [39,40] and
more systematic reviews are needed to evaluate the effect of placenta
previa on other adverse outcomes.
Clinicians should consider and communicate with these risk factors
when counseling their patients. Our umbrella review provides data that
26
Placenta 117 (2022) 21–27
can be used to reassure women, refer to pre-conception counseling
clinics or antenatal clinics.
4.3. Strengths and limitations
The main strength of this is the first umbrella review on determinants
of placenta previa. However, some limitations should be considered in
the mind when interpreting the findings. First; although the coverage
rate of three databases PubMed, Scopus, and Web of Sciences is high,
however, selection bias due to unretrieved studies might affect the
observed results. Second, confidence in the results of the review is still a
concern because of the critically low rate of each meta-analysis. Third,
observed associations are not equal to causality.
5. Conclusion
In conclusion, this umbrella review provides suggestive evidence
about prior spontaneous abortion, prior induced abortion, male fetus,
smoking, advanced maternal age, cesarean section, and assisted repro­
ductive techniques (singleton pregnancy) as risk factors associated with
placenta previa.
Funding
This study supported by Hamadan University of Medical Sciences
with code: 140002211132.D:\MYFILES\ELSEVIER\YPLAC\00004508
\S-CEEDITING\gs1
Declaration of competing interest
All authors report no conflict of interest.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.placenta.2021.10.009.
References
[1] J.A. Cresswell, C. Ronsmans, C. Calvert, V. Filippi, Prevalence of placenta praevia
by world region: a systematic review and meta-analysis, Trop. Med. Int. Health 18
(6) (2013) 712–724.
[2] A. Faiz, C. Ananth, Etiology and risk factors for placenta previa: an overview and
meta-analysis of observational studies, J. Matern. Fetal Neonatal Med. 13 (3)
(2003) 175–190.
[3] S. Bi, L. Zhang, Z. Wang, J. Chen, J. Tang, J. Gong, et al., Effect of types of placenta
previa on maternal and neonatal outcomes: a 10-year retrospective cohort study,
Arch. Gynecol. Obstet. 304 (1) (2021) 65–72.
[4] M. Karami, E. Jenabi, B. Fereidooni, The association of placenta previa and assisted
reproductive techniques: a meta-analysis, J. Matern. Fetal Neonatal Med. 31 (14)
(2018) 1940–1947.
[5] F. Shobeiri, E. Jenabi, Smoking and placenta previa: a meta-analysis, J. Matern.
Fetal Neonatal Med. 30 (24) (2017) 2985–2990.
[6] U.V. Ukah, D.A. De Silva, B. Payne, L.A. Magee, J.A. Hutcheon, H. Brown, et al.,
Prediction of adverse maternal outcomes from pre-eclampsia and other
hypertensive disorders of pregnancy: a systematic review, Pregnancy hypertension
11 (2018) 115–123.
[7] K.G. Martinelli, É.M. Garcia, ETd Santos Neto, SGNd Gama, Advanced maternal age
and its association with placenta praevia and placental abruption: a meta-analysis,
Cad. Saúde Pública 34 (2018), e00206116.
[8] J.P. Ioannidis, Why most published research findings are false, PLoS Med. 2 (8)
(2005) e124.
[9] D. Moher, L. Shamseer, M. Clarke, D. Ghersi, A. Liberati, M. Petticrew, et al.,
Preferred reporting items for systematic review and meta-analysis protocols
(PRISMA-P) 2015 statement, Syst. Rev. 4 (1) (2015) 1–9.
[10] W.G. Cochran, The combination of estimates from different experiments,
Biometrics 10 (1) (1954) 101–129.
[11] J.P. Higgins, S.G. Thompson, D.J. Spiegelhalter, A re-evaluation of random-effects
meta-analysis, J. Roy. Stat. Soc. 172 (1) (2009) 137–159.
[12] M. Egger, G.D. Smith, M. Schneider, C. Minder, Bias in meta-analysis detected by a
simple, graphical test, BMJ 315 (7109) (1997) 629–634.
[13] J.P. Ioannidis, T.A. Trikalinos, An exploratory test for an excess of significant
findings, Clin. Trials 4 (3) (2007) 245–253.
E. Jenabi et al.
[14] M.B. Mathur, T.J. VanderWeele, Sensitivity analysis for unmeasured confounding
in meta-analyses, J. Am. Stat. Assoc. 115 (529) (2020) 163–172.
[15] M.B. Mathur, T.J. VanderWeele, Controversy and debate on credibility ceilings.
Paper 1: fundamental problems with the "credibility ceiling" method for meta-
analyses, J. Clin. Epidemiol. 127 (2020) 208–210.
[16] B.J. Shea, B.C. Reeves, G. Wells, M. Thuku, C. Hamel, J. Moran, et al., Amstar 2: a
critical appraisal tool for systematic reviews that include randomised or non-
randomised studies of healthcare interventions, or both, BMJ 358 (2017).
[17] J.H. Kim, J.Y. Kim, J. Lee, G.H. Jeong, E. Lee, S. Lee, et al., Environmental risk
factors, protective factors, and peripheral biomarkers for ADHD: an umbrella
review, The Lancet Psychiatry 7 (11) (2020) 955–970.
[18] J.Y. Kim, M.J. Son, C.Y. Son, J. Radua, M. Eisenhut, F. Gressier, et al.,
Environmental risk factors and biomarkers for autism spectrum disorder: an
umbrella review of the evidence, The Lancet Psychiatry 6 (7) (2019) 590–600.
[19] I. Gurol-Urganci, D.A. Cromwell, L.C. Edozien, G.C. Smith, C. Onwere, T.
A. Mahmood, et al., Risk of placenta previa in second birth after first birth cesarean
section: a population-based study and meta-analysis, BMC Pregnancy Childbirth 11
(1) (2011) 1–10.
[20] E. Jenabi, B. Fereidooni, The uterine leiomyoma and placenta previa: a meta-
analysis, J. Matern. Fetal Neonatal Med. 32 (7) (2019) 1200–1204.
[21] M. Karami, E. Jenabi, Placenta previa after prior abortion: a meta-analysis,
Biomedical Research and Therapy 4 (7) (2017) 1441–1450.
[22] J. Qin, X. Liu, X. Sheng, H. Wang, S. Gao, 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. 105 (1)
(2016) 73–85, e6.
[23] J. Qin, H. Wang, X. Sheng, D. Liang, H. Tan, J. Xia, Pregnancy-related
complications and adverse pregnancy outcomes in multiple pregnancies resulting
from assisted reproductive technology: a meta-analysis of cohort studies, Fertil.
Steril. 103 (6) (2015) 1492–1508, e7.
[24] F. Zullo, E. Spagnolo, G. Saccone, M. Acunzo, S. Xodo, M. Ceccaroni, et al.,
Endometriosis and obstetrics complications: a systematic review and meta-
analysis, Fertil. Steril. 108 (4) (2017) 667–672, e5.
[25] P. Fusar-Poli, J. Radua, Ten simple rules for conducting umbrella reviews, Evid.
Base Ment. Health 21 (3) (2018) 95–100.
[26] S. Sumigama, C. Sugiyama, T. Kotani, H. Hayakawa, A. Inoue, Y. Mano, et al.,
Uterine sutures at prior caesarean section and placenta accreta in subsequent
pregnancy: a case-control study, BJOG An Int. J. Obstet. Gynaecol. 121 (7) (2014)
866–874. ; discussion 75.
27
Placenta 117 (2022) 21–27
[27] F. Shobeiri, E. Jenabi, M. Karami, S. Karimi, Determinants of placenta previa: a
case-control study, Biomedical Research and Therapy 4 (6) (2017) 1411–1419.
[28] G. Mónica, C. Lilja, Placenta previa, maternal smoking and recurrence risk, Acta
Obstet. Gynecol. Scand. 74 (5) (1995) 341–345.
[29] M. Hansen, J.J. Kurinczuk, N. De Klerk, P. Burton, C. Bower, Assisted reproductive
technology and major birth defects in Western Australia, Obstet. Gynecol. 120 (4)
(2012) 852–863.
[30] A. Tavaniotou, C. Albano, J. Smitz, P. Devroey, Impact of ovarian stimulation on
corpus luteum function and embryonic implantation, J. Reprod. Immunol. 55 (1–2)
(2002) 123–130.
[31] G. Garmi, R. Salim, Epidemiology, etiology, diagnosis, and management of
placenta accreta, Obstetrics and gynecology international 2012 (2012) 873929.
[32] D. Getahun, Y. Oyelese, H.M. Salihu, C.V. Ananth, Previous cesarean delivery and
risks of placenta previa and placental abruption, Obstet. Gynecol. 107 (4) (2006)
771–778.
[33] M. M, A. V, G.N. G, E. N, A.E. P, I. M, Association of placenta previa with a history
of previous cesarian deliveries and indications for a possible role of a genetic
component, Balkan J. Med. Genet.: BJMG 20 (2) (2017) 5–10.
[34] J. Qin, H. Wang, X. Sheng, D. Liang, H. Tan, J. Xia, Pregnancy-related
complications and adverse pregnancy outcomes in multiple pregnancies resulting
from assisted reproductive technology: a meta-analysis of cohort studies, Fertil.
Steril. 103 (6) (2015) 1492–1508, e1-1508.
[35] M.L. Gasparri, K. Nirgianakis, K. Taghavi, A. Papadia, M.D. Mueller, Placenta
previa and placental abruption after assisted reproductive technology in patients
with endometriosis: a systematic review and meta-analysis, Arch. Gynecol. Obstet.
298 (1) (2018) 27–34.
[36] A.S. Handler, E.D. Mason, D.L. Rosenberg, F.G. Davis, The relationship between
exposure during pregnancy to cigarette smoking and cocaine use and placenta
previa, Am. J. Obstet. Gynecol. 170 (3) (1994) 884–889.
[37] D.A. Miller, J.A. Chollet, T.M. Goodwin, Clinical risk factors for placenta previa-
placenta accreta, Am. J. Obstet. Gynecol. 177 (1) (1997) 210–214.
[38] L. Tuzović, J. Djelmis, M. Ilijić, Obstetric risk factors associated with placenta
previa development: case-control study, Croat. Med. J. 44 (6) (2003) 728–733.
[39] E.J. Jung, H.J. Cho, J.M. Byun, D.H. Jeong, K.B. Lee, M.S. Sung, et al., Placental
pathologic changes and perinatal outcomes in placenta previa, Placenta 63 (2018)
15–20.
[40] E. Weiner, H. Miremberg, E. Grinstein, Y. Mizrachi, L. Schreiber, J. Bar, et al., The
effect of placenta previa on fetal growth and pregnancy outcome, in correlation
with placental pathology, J. Perinatol. 36 (12) (2016) 1073–1078.