DOI: 10.1111/1471-0528.

12717 General gynaecology

www.bjog.org

Uterine sutures at prior caesarean section and

placenta accreta in subsequent pregnancy: a
case–control study
S Sumigama,a C Sugiyama,b T Kotani,a H Hayakawa,c A Inoue,d Y Mano,a H Tsuda,a M Furuhashi,e
O Yamamuro,f Y Kinoshita,g T Okamoto,h H Nakamura,b K Matsusawa,i K Sakakibara,j H Oguchi,k
M Kawai,l Y Shimoyama,m K Tamakoshi,n F Kikkawaa
a
Department of Obstetrics and Gynaecology, Nagoya University Graduate School of Medicine, Aichi, Japan b Department of Obstetrics and
Gynaecology, Gifu Prefectural Tajimi Hospital, Gifu, Japan c Department of Obstetrics and Gynaecology, Kasugai Municipal Hospital, Aichi,
Japan d Department of Obstetrics and Gynaecology, Hekinan Municipal Hospital, Aichi, Japan e Department of Obstetrics and Gynaecology,
Japanese Red Cross Nagoya Daiichi Hospital, Aichi, Japan f Department of Obstetrics and Gynaecology, Nagoya Daini Red Cross Hospital,
Aichi, Japan g Department of Obstetrics and Gynaecology, Ogaki Municipal Hospital, Gifu, Japan h Department of Obstetrics and Gynaecology,
Chukyo Hospital, Aichi, Japan i Department of Obstetrics and Gynaecology, Anjo Kosei Hospital, Aichi, Japan j Department of Obstetrics and
Gynaecology, Okazaki City Hospital, Aichi, Japan k Department of Obstetrics and Gynaecology, Toyota Memorial Hospital, Aichi, Japan
l
Department of Obstetrics and Gynaecology, Toyohashi Municipal Hospital, Aichi, Japan m Department of Pathology and Laboratory Medicine,
Nagoya University Hospital, Aichi, Japan n Department of Nursing, Nagoya University School of Health Sciences, Aichi, Japan
Correspondence: Dr S Sumigama, Department of Obstetrics and Gynaecology, Nagoya University Graduate School of Medicine,
65 Tsurumai-cho, Showa-ku, Nagoya 466-8550, Japan. Email [email protected]

Accepted 13 January 2014. Published Online 26 March 2014.

Objective To clarify the effects of uterine myometrial suture Results No difference was found comparing single-layer with
techniques at prior caesarean section on the incidence of double-layer closure in the incidence of placenta accreta (37.1
pathologically diagnosed placenta accreta in placenta praevia with versus 39.7%, P = 0.805); however, a significant difference was
prior caesarean section (PPPC). found comparing continuous with interrupted sutures (58.1 versus
29.9%, P = 0.008). Multivariable logistic regression analysis with
Design Case–control study.
stepwise selection for the eight factors meeting the criterion of
Setting Eleven tertiary referral hospitals in central Japan. P < 0.10 in univariate analysis was used, and four independent
factors were selected, as follows: gravidity ≥ 3 (adjusted odds
Population A total of 98 cases of placenta praevia, a history of
ratio, aOR, 3.4, 95% confidence interval, 95% CI, 0.99–11.6,
one or more prior caesarean sections, and a history of uterine
P = 0.050); total praevia (versus non-total, aOR 18.4, 95%
transverse incision and usage of only absorbable thread for
CI 3.2–107.0, P = 0.001); anterior/centre placenta (versus
myometrial sutures at the prior caesarean section. Exclusions were
posterior, aOR 16.4, 95% CI 3.7–72.2, P < 0.001); and continuous
a history of myomectomy or Strassmann’s operation.
sutures (versus interrupted, aOR 6.0, 95% CI 1.4–25.2,
Methods Cases were grouped into a pathologically diagnosed P = 0.015).
placenta accreta group (38 cases) and a no accreta group (60
Conclusions In this limited study, a history of continuous sutures
cases). Clinical characteristics including uterine suture methods at
on the inner side of the uterine wall showed potential to influence
prior caesarean section were compared (single-layer versus
the development of placenta accreta in PPPC patients.
double-layer closure; continuous versus interrupted sutures in the
inner myometrial layer). Keywords Caesarean section, continuous suture, GRADE,
interrupted suture, placenta accreta, uterine closure.
Main outcome measure The incidence of placenta accreta.

Please cite this paper as: Sumigama S, Sugiyama C, Kotani T, Hayakawa H, Inoue A, Mano Y, Tsuda H, Furuhashi M, Yamamuro O, Kinoshita Y,
Okamoto T, Nakamura H, Matsusawa K, Sakakibara K, Oguchi H, Kawai M, Shimoyama Y, Tamakoshi K, Kikkawa F. Uterine sutures at prior caesarean
section and placenta accreta in subsequent pregnancy: a case–control study. BJOG 2014;121:866–875.

most appropriate remain controversial.1,2 When selecting a

Introduction
caesarean procedure, the priorities must be safe delivery,
Caesarean section is a common operation performed safe completion of the operation, and the minimisation of
worldwide, but the precise details of which procedures are postoperative morbidity. The recent large-scale randomised

866 ª 2014 Royal College of Obstetricians and Gynaecologists

 Uterine suture techniques and placenta accreta

controlled trial, CORONIS, answered these questions by Municipal Hospital). First, we identified patients who had
showing no significant difference in maternal outcomes been diagnosed with placenta praevia between January 1994
within 6 weeks of caesarean section, comparing five ele- and December 2012 from the records of all patients who
ments of operative procedures, including single-layer versus delivered in each hospital. Obstetric histories were then
double-layer uterine closure.3 The next priority should be checked and patients who had undergone one or more prior
to reduce long-term risk and prevent complications in sub- caesarean sections were considered eligible for the study.
sequent pregnancies, including intra-abdominal adhesions, The diagnosis of placenta praevia was reconfirmed from
uterine scar pregnancy, uterine rupture, placenta praevia, images (ultrasonography and/or magnetic resonance imag-
and placenta accreta. ing, MRI) recorded on medical charts and classified as
Placenta accreta is an abnormal condition of invasive total, partial, marginal placenta praevia, low-lying placenta,
placental implantation that causes massive haemorrhage at or normally lying placenta. When the internal os was
delivery and threatens maternal life.4 Several risk factors for dilated, the distinction between total and partial placenta
placenta accreta have been identified, including advanced praevia depended on whether the placenta covered the os
maternal age, multiparity, placenta praevia, a short preg- completely. When the internal os was closed, diagnosis was
nancy interval, a female fetus, submucosal myoma, and made according to the following criteria.13 Total placenta
previous uterine surgery.5–8 Among these factors, placenta praevia was diagnosed when the placenta overlaid the inter-
praevia with previous caesarean section (PPPC) has been nal os and the distance between the placental edge and
suggested to be associated with a higher risk of placenta ac- internal os was ≥2 cm. When the placenta overlaid the
creta. The incidence rate of placenta accreta is one in every internal os and the distance was between 0 and 2 cm, par-
2500 deliveries (0.04%);9 by contrast, this rate markedly tial placenta praevia was diagnosed, and when the distance
elevates to 11–37% in placenta praevia with one previous was 0 cm, marginal placenta praevia was diagnosed. When
caesarean section.10–12 It shows further increases with an the placenta did not overlie the internal os and the placen-
increased number of prior caesarean deliveries (over 60% tal edge was <2 cm away, low-lying placenta was diag-
of placenta praevia with three or more prior caesarean nosed. When the placental edge was ≥2 cm from the os,
deliveries).6,10–12 Thus, PPPC is an important entity for normally lying placenta was diagnosed and the case was
placenta accreta. excluded from the study.
From the aspect of these clinical observations, the caesar- Placenta accreta was diagnosed pathologically based on
ean scar on the uterus is quite likely to influence the inci- uterine and placental specimens, when the decidua basalis
dence of placenta accreta. Uterine low-transverse incision was absent and the placental villous tissue was in direct
has become the most commonly performed technique for contact with the myometrium (placenta accreta in the nar-
caesarean section because of the relatively low risk for uter- row sense). The condition was classified as placenta increta
ine rupture compared with a vertical uterine incision. On when the villi invaded the myometrium, and as placenta
the other hand, which uterine closure techniques would be percreta when the villi penetrated the entire wall.14 We
the most appropriate, including one-layer or two-layer excluded any placenta accreta diagnosed only from clinical
sutures, continuous or interrupted sutures, locking or findings, because it was difficult to confirm the diagnostic
non-locking sutures, remains controversial. As for placenta criteria used by each clinician involved in this retrospective
accreta, there is no evidence regarding the relationship multicentre analysis. Cases were then divided into two
between suture techniques and the occurrence of placenta groups: PPPC with accreta and PPPC without accreta.
accreta for subsequent pregnancies. In this study, we tar- Medical charts were reviewed and the following data
geted women diagnosed with PPPC and investigated the were recorded: maternal age at the time of current delivery;
methods of uterine closure at prior caesarean section to duration of current pregnancy; date of current delivery;
reveal whether different closure methods might influence gravidity; parity; number of vaginal deliveries, caesarean
the incidence of placenta accreta. deliveries, miscarriages, and abortions; sex of the current
neonate; and history of fertility treatment. Histories and
dates of prior uterine operations, including caesarean sec-
Materials and methods
tions, myomectomy, and others, were recorded. We calcu-
Retrospective analysis was performed using data from 11 ter- lated the period between prior caesarean section and the
tiary centres in the central area of Japan (Nagoya University establishment of current pregnancy (2 weeks, 0 days),
Hospital, Japanese Red Cross Nagoya Daiichi Hospital, Nag- rather than the period between prior caesarean section and
oya Daini Red Cross Hospital, Ogaki Municipal Hospital, current delivery, because placentation occurs at the early
Chukyo Hospital, Gifu Prefectural Tajimi Hospital, Anjo stage of pregnancy. To examine the relationship between
Kosei Hospital, Okazaki City Hospital, Toyota Memorial placental location and the incidence of placenta accreta,
Hospital, Toyohashi Municipal Hospital, and Kasugai images from ultrasonography and/or MRI were also

ª 2014 Royal College of Obstetricians and Gynaecologists 867

 Sumigama et al.

reviewed to classify placental location. ‘Posterior location’ vertical uterine incision, the usage of unabsorbable thread,
was classified as the main body of the placenta located on and histories of myomectomy and/or hysteroplasty.
the posterior wall, with the anterior placental edge located The relationship between each factor and the frequency
under the centre of the bladder. Other locations of the pla- of placenta accreta was examined statistically using SPSS 20.0
centa were classified as ‘anterior/centre location’ (Table 1). (Chicago, IL, USA). We examined differences between pla-
The operative notes of the prior caesarean section were centa accreta and no accreta groups using the unpaired
also reviewed to record the following details: maternal age Student’s t–test (for parametric variables), the Mann–Whit-
at the time of prior caesarean section; operation time; ney U–test (for non-parametric variables), contingency
intraoperative blood loss; method of uterine incision (low tables, and the v2 test or Fisher’s exact test (for categorical
transverse incision, classical vertex incision, or other); layer data). For variables showing P < 0.10, multivariable logistic
of uterine myometrial closure (single layer, double layer or regression analysis was used to control for potential con-
more); suture method (continuous, interrupted, or other); founding variables and to obtain adjusted odds ratios, and
suture thread material (absorbable, silk, or other); and P < 0.05 was considered to indicate a statistically signifi-
intra- and postoperative complications. If the patients had cant result.
two or more prior caesarean sections, we reviewed the
operative notes of the immediate prior caesarean section
Results
only. If the prior caesarean section was performed at
another hospital, we requested that the hospital reply to A total of 96 670 deliveries occurred during the study per-
the survey in the form of a questionnaire based on the iod, including 954 patients diagnosed with placenta praevia.
operative notes. In continuous suture cases, we also asked Obstetric history was checked and revealed 820 patients
whether locking or non-locking sutures were used; how- (86.0%) with no history of prior caesarean section (includ-
ever, locking statuses were not written in the operative ing eight pathologically diagnosed placenta accreta: three
notes in several cases. In these cases, we requested the hos- placenta accreta, five increta, and no percreta), and 134
pital to reply in the context of the surgeon’s or hospital’s patients (14.0%) with a history of one or more prior cae-
standard procedure at the time. Exclusion criteria were sarean sections (including 90 cases reconfirmed as total pla-
unknown surgical techniques at prior caesarean section, centa praevia, ten cases as partial placenta praevia, 16 cases
as marginal placenta praevia, and 18 cases as low-lying pla-
Table 1. Characteristics and outcomes centa). Among these 134 PPPC cases, information regard-
ing the method of uterine myometrium incision and
Characteristic Maternal age at the time of current delivery suturing at the prior caesarean section was unavailable in
(years, <35/≥35)
31 cases because of loss or incomplete documentation of
Duration of current pregnancy (weeks)
the operative notes (nine cases), a lack of reply to the ques-
Gravidity (times, <3/≥3)
Parity (times, <2/≥2) tionnaire (five cases), or an unknown hospital had per-
Vaginal deliveries (times, 0/≥1) formed the prior caesarean sections (17 cases). Among the
Caesarean deliveries (times, 1/≥2) 103 cases for which information was obtained, one case of
Miscarriages (times, 0/≥1) vertical uterine incision, two cases of use of unabsorbable
Abortions (times, 0/≥1) thread (silk), and two cases of other uterine operation (one
Sex of the current neonate (male/female)
myomectomy and one Strassmann operation) were
History of fertility treatment (yes/no/unknown)
excluded. Thus, 98 PPPC cases that had a history of prior
Pregnancy interval (from prior caesarean section
to 2 weeks 0 days of current pregnancy) caesarean section treated by uterine low transverse inci-
(months, <12/12–24/24–36/≥36) sions, with the use of absorbable thread only (polyglactin
Placental location (anterior/centre/posterior) 910, polyglycolic acid, polyglycapron 25, or catgut), and
Type of placenta praevia (total/partial/ complete information regarding the method (layers and
marginal/low-lying) sutures) were analysed in the present study (Figure 1).
Maternal age at the time of prior caesarean
Suture methods of the uterine myometrium of these 98
section (years, <35/≥35)
cases were sorted into six groups according to the layer
Operation time at prior caesarean section (min)
Blood loss at prior caesarean section (ml) (single or double) and suture type (interrupted or continu-
Anaesthesia at prior caesarean section (general/local) ous) (Figure 2). In 21 of the 98 cases, a prior caesarean
Myometrial closure at prior caesarean section section had been performed at one of the 11 tertiary hospi-
(method A/B/C/D/E/F, single-layer/double-layer, tals in this study group. In the remaining 77 cases, caesar-
continuous/interrupted) ean sections had been performed at 50 other hospitals. The
Outcome Pathologically diagnosed placenta accreta
98 cases were thus treated at 61 hospitals, with 1–7 cases
treated per hospital.

868 ª 2014 Royal College of Obstetricians and Gynaecologists

 Uterine suture techniques and placenta accreta

Figure 1. Flow of cases in this study.

For the current pregnancy, all 98 PPPC patients delivered differences (P < 0.05) in risk associated with gravidity
by caesarean section, with caesarean hysterectomy in 32 (P = 0.012), parity (P = 0.026), artificial abortion
cases. One patient underwent postpartum hysterectomy the (P = 0.034), prior caesarean section (P = 0.026), type of
day after caesarean section because of postoperative genital placenta praevia (P < 0.001), and placental location
bleeding. Five cases underwent scheduled stepwise opera- (P < 0.001). No significant correlations were seen with
tions (caesarean section without separation of the placenta, maternal age, gestational age at delivery, history of sponta-
arterial embolization the same day or the next day and hys- neous abortion, vaginal delivery, history of fertility treat-
terectomy 1 week later)12. Thus, 38 hysterectomy specimens ment, fetal sex, or the interval between prior caesarean
were obtained, and they were pathologically diagnosed as section and the establishment of the present pregnancy.
placenta accreta (11 accreta in the narrow sense, 16 increta, Table 3 shows the details of prior caesarean section in
and 11 percreta). Among the 60 PPPC cases without hyster- PPPC with and without accreta groups. Univariate analysis
ectomy, no cases were pathologically diagnosed as placenta revealed significant differences in risk associated with
accreta from the placental specimen alone (these 60 cases maternal age at prior caesarean section ≥35 years
included three cases clinically diagnosed as placenta accreta (P = 0.042). No significant correlations were seen in opera-
because the placenta was hard to remove and massive blood tional time and blood loss at prior caesarean section. Intra-
loss occurred). Finally, the 98 PPPC patients were grouped and postoperative complications at prior caesarean section
into 38 patients with pathological placenta accreta (PPPC included bladder injury and repair in one case, uterine
with accreta, 38.8%) and 60 without pathological placenta evacuation because of postoperative genital bleeding in one
accreta (PPPC without accreta, 61.2%). case, severe blood loss (>1500 ml) in eight cases, and blood
Table 2 shows the baseline characteristics as preoperative transfusion in two cases.
predictors of PPPC with and without accreta groups in the Among the uterine closure techniques, method A (sin-
current pregnancy. Univariate analysis revealed significant gle-layer closure with interrupted sutures: SL-Intr) included

ª 2014 Royal College of Obstetricians and Gynaecologists 869

 Sumigama et al.

Figure 2. Suture techniques for the uterine myometrium at prior caesarean section. Cntn, continuous suture; DL, double layer; Intr, interrupted
suture; SL, single layer.

the largest number of cases (30 cases), and showed the the criterion of P < 0.10 in univariate analysis (gravidity,
development of placenta accreta in 40.0% (12/30) of subse- <3 or ≥3; parity, <2 or ≥2; abortion, 0 or ≥1; prior caesar-
quent pregnancies. Group F (double-layer closure with con- ean section, 1 or ≥2; placenta praevia type, total or
tinuous sutures of both layers: DL-Cntn/Cntn) had the non-total; placental location, anterior/centre or posterior;
highest incidence rate of placenta accreta (70.6%, 12/17). maternal age at prior caesarean section, <35 or ≥35 years;
Methods B and E included a small number of patients. The myometrial closure of inner layer, interrupted or continu-
Fisher’s exact test showed a significant difference between ous; Table 4). After adjusting for the effect of each variable,
methods A–F and the incidence of placenta accreta the following factors were still associated with significant
(P = 0.018). We categorised groups into single-layer closure differences in the incidence of placenta accreta: total pla-
(methods A and B; 35 cases) and double-layer closure centa praevia (versus non-total placenta praevia, aOR 18.4,
(methods C–F; 63 cases) to compare the incidence of pla- 95% CI 3.2–107.0, P = 0.001); anterior/centre position of
centa accreta. Placenta accreta occurred in 37.1% (13/35) of placenta (versus posterior placenta, aOR 16.4, 95% CI 3.7–
the single-layer closure group and 39.7% (25/63) of the dou- 72.2, P < 0.001); and continuous suture of the inner layer
ble-layer closure group, which showed no significant differ- (endometrial side) of the uterine myometrium (versus
ence (P = 0.805). Next we examined the suture technique interrupted suture, aOR 6.0, 95% CI 1.4–25.2, P = 0.015).
for the inner layer (endometrial side) of the uterine myome-
trium and categorised interrupted closure at the inner layer
Discussion
(methods A, C, and D; 67 cases) and continuous closure at
the inner layer (methods B, E, and F; 31 cases). Placenta ac- Main findings
creta occurred in 29.9% (20/67) of the interrupted group Using multivariate analysis, we identified total placenta
and in 58.1% (18/31) of the continuous group, which praevia, anterior/centre placenta, and continuous sutures of
showed a significant difference (P = 0.008). In addition, we the inner layer of the uterine myometrium at prior caesar-
analysed whether locking or non-locking sutures were used ean section as independent risk factors for placenta accreta
in the 31 cases of the continuous group, obtaining this in PPPC patients. We also compared single-layer closure of
information in 24 cases. Accreta occurred in 69.2% (4/13) the uterine myometrium with double-layer closure, finding
of the non-locking group and 54.5% (6/11) of the locking no significant difference between the two.
group, which showed no statistical differences (P = 0.532).
Multivariable logistic regression analysis with stepwise Strengths and weaknesses
selection was performed to select independent factors and The strengths of this study were as follows. To reduce selec-
calculate adjusted odds ratios for the eight factors meeting tion bias, we collected all PPPC cases from all deliveries in

870 ª 2014 Royal College of Obstetricians and Gynaecologists

 Uterine suture techniques and placenta accreta

Table 2. Baseline characteristics and placenta accreta in PPPC

Characteristic Total (n = 98) PPPC with accreta PPPC without accreta P

(n = 38) (n = 60)

Maternal age (years) 33.4  3.9 33.1  3.7 33.5  4.0 0.642
Gestational age at delivery (weeks) 35.1  3.8 34.5  4.2 35.5  3.4 0.189
Gravidity
<3 44 11 (25.0%) 33 (75.0%) 0.012
≥3 54 27 (50.0%) 27 (50.0%)
Parity
<2 67 21 (31.3%) 46 (68.7%) 0.026
≥2 31 17 (54.8%) 14 (45.2%)
Miscarriage
0 76 30 (39.5%) 46 (60.5%) 0.792
≥1 22 8 (36.4%) 14 (63.6%)
Abortion
0 86 30 (34.9%) 56 (65.1%) 0.034
≥1 12 8 (66.7%) 4 (33.3%)
Vaginal delivery
0 89 35 (39.3%) 54 (60.7%) 0.725
≥1 9 3 (33.3%) 6 (66.7%)
Caesarean section
1 76 25 (32.9%) 51 (67.1%) 0.026
≥2 22 13 (59.1%) 9 (40.9%)
Fetal sex
Male 45 20 (44.4%) 25 (55.6%) 0.289
Female 53 18 (34.0%) 35 (66.0%)
Type of placenta praevia
Non-total 36 2 (5.6%) 34 (94.4%) <0.001
Total 62 36 (58.1%) 26 (41.9%)
Placental location
Posterior 43 3 (7.0%) 40 (93.0%) <0.001
Anterior/centre 55 35 (63.6%) 20 (36.4%)
Pregnancy interval (months) 27.4 (5.0–125) 24.9 (6.2–125) 29.0 (5.0–93.4) 0.787
<12 10 4 (40.0%) 6 (60.0%) 0.557
12–24 30 13 (43.3%) 17 (56.7%)
24–36 23 6 (26.1%) 17 (73.9%)
≥36 35 15 (42.9%) 20 (57.1%)

Data are presented as means  standard deviations, median (min–max) or n (%).

the 11 hospitals involved in this study and grouped them caesarean sections was reported to increase the risk for
into case (placenta accreta) and control (no accreta). To placenta accreta in previous studies, but showed no signifi-
reduce the ambiguity in the diagnosis, we excluded clinically cant risk in multivariate analysis in this study.10–12 Further-
diagnosed placenta accreta because the criteria were unclear more, if patients had multiple prior caesarean sections, we
in this retrospective study, and only included pathological only researched the techniques for the caesarean section
placenta accreta. immediately prior to this pregnancy. It was unclear whether
This study included some limitations in researching the patients had received different surgical techniques in the
surgical techniques of prior caesarean section. We could other prior caesarean sections than those noted in this
not obtain any information in 31 cases (including ten pla- analysis, and if there were such conditions, we could not
centa accreta) of 134 PPPC. The final sample size was 98 guess what the potential effects might be. Associated infor-
cases, which was rather small. Statistical analysis showed mation about suture methods (locking status and additional
significant findings within these cases; however, other haemostatic sutures) was lacking because this information
parameters that showed no relation might be significant risk was not fully recorded in the operative notes. Information
factors in a larger sample size. For example, multiple prior about the experience of operating surgeons was also lacking,

ª 2014 Royal College of Obstetricians and Gynaecologists 871

 Sumigama et al.

Table 3. Details of prior caesarean section and placenta accreta in PPPC

Characteristic Total (n = 98) PPPC with accreta (n = 38) PPPC without accreta (n = 60) P

Maternal age at prior caesarean section (years) 29.9  3.8 29.5  3.3 30.2  4.1 0.412
<35 84 36 (42.9%) 48 (57.1%) 0.042
≥35 14 2 (14.3%) 12 (85.7%)
Operation time at prior caesarean section (min) 48.7  18.8 49.0  19.3 48.1  18.4 0.843
Blood loss at prior caesarean section (ml) 700 (181–3535) 656 (318–2760) 700 (181–3535) 0.694
Myometrial closure at prior caesarean section*
Method A (SL-Intr) 30 12 (40.0%) 18 (60.0%) 0.018**
Method B (SL-Cntn) 5 1 (20.0%) 4 (80.0%)
Method C (DL-Intr/Intr) 14 3 (21.4%) 11 (78.6%)
Method D (DL-Intr/Cntn) 23 5 (21.7%) 18 (78.3%)
Method E (DL-Cntn/Intr) 9 5 (55.6%) 4 (49.4%)
Method F (DL-Cntn/Cntn) 17 12 (70.6%) 5 (29.4%)
Layer
Single layer (A, B) 35 13 (37.1%) 22 (62.9%) 0.805
Double layer (C–F) 63 25 (39.7%) 38 (60.3%)
Technique for Inner layer
Interrupted (A, C, D) 67 20 (29.9%) 47 (70.1%) 0.008
Continuous (B, E, F) 31 18 (58.1%) 13 (41.9%)

*SL, single layer; DL, double layer; Intr, interrupted suture; Cntn, continuous suture.
**Fisher’s exact test.

were in doubt: no other report exists (consistency/analogy)

Table 4. Adjusted odds ratios with 95% confidence intervals for
placenta accreta in PPPC
and no dose–response effect or experimental verification
was examined (biological gradient/experiment). Specificity
Characteristic aOR 95% CI P and coherence were uncertain in this study.15 We also
applied the grading of recommendations assessment, devel-
Gravidity: ≥3 3.4 0.99–11.6 0.050 opment and evaluation (GRADE) system and obtained the
Placenta praevia type: following assessments; very low quality of evidence and
Total (versus non-total) 18.4 3.16–107 0.001 weak recommendation.16
Placental position:
Anterior/centre (versus posterior) 16.4 3.71–72.2 <0.001
Suture for inner myometrial layer at prior caesarean section:
Interpretation
Continuous (versus interrupted) 6.0 1.42–25.2 0.015 Both continuous sutures and interrupted sutures have been
used routinely for uterine closure. Williams Obstetrics rec-
Multivariable logistic regression analysis with stepwise selection for ommends single- or double-layer continuous sutures,17 and
the eight factors meeting the criterion of P < 0.10 in univariate
Gabbe’s Obstetrics recommends that the first layer should be
analysis was used, and four independent factors were selected for
the table. either interrupted or continuous sutures and that a second
layer should be added with continuous sutures when the
apposition is unsatisfactory.18 The combination of continu-
and differences between surgeons were not assessed. These ous sutures and a single-layer closure provides immediate
factors might affect wound healing and the subsequent closure of the wound, reducing both blood loss and operat-
incidence of placenta accreta, but could not be clarified in ing time compared with that of a double-layer closure.19–21
this study. Conversely, concern has been raised that continuous sutures
We applied Hill’s criteria for consideration of causal can cause excessive pressure and ischaemic effects on the
interference and found at least three relevant elements: (1) myometrium, which might lead to inefficient wound heal-
strength of association – the adjusted odds ratio was rather ing;22 however, little is known regarding the effects of con-
high; (2) temporality – the prior caesarean section pro- tinuous sutures on wound healing in the uterine
ceeded the incidence of placenta accreta; and (3) biological myometrium/endometrium. One non-randomised study
plausibility – the clinical observation that a history of cae- compared continuous with interrupted sutures for a lower
sarean section elevates the incidence of placenta accreta uterine incision, and showed that continuous sutures saved
indicates an association. Other elements were not met or operating time and blood loss: however, no survey for any

872 ª 2014 Royal College of Obstetricians and Gynaecologists

 Uterine suture techniques and placenta accreta

effect on subsequent pregnancy has been reported.23 We showed rather high odds ratios, which was reflected by the
thought that the suture technique for the inner side of the higher possibility that these conditions might cause the pla-
myometrium might be important for healing the endome- centa to overlie the caesarean scar and induce villous inva-
trium, and might influence the protective function against sion to the myometrium.
villous invasion, and showed that continuous suturing of One previous study reported that the ratio of male to
the inner layer represents a risk factor for placenta accreta female infants was decreased in women with abnormally
with an adjusted odds ratio of 6.0 for the next pregnancy, adherent placenta; however, our data showed that fetal sex
regardless of single- or double-layer closure. Bujold et al. did not impact on the frequency of placenta accreta.27
reported that a prior single-layer closure carried more than Advanced maternal age is reported as a risk factor, but this
a two-fold increase in the risk of uterine rupture, compared study found that mothers ≥35 years of age tended to show
with a double-layer closure, which suggests that single-layer a lower risk, although this was not statistically significant
closure adversely impacts uterine scar healing (thinning (P = 0.517).
and/or weakness), leading to inadequate strength to with-
stand the disruptive forces encountered during labour.24,25
Conclusion
In this study, no advantage was seen for double-layer closure
in the prevention of placenta accreta. This result suggests The suture techniques at prior caesarean section showed
that it is not thinning and/or weakness of the myometrium potential to influence the incidence of placenta accreta for
but rather insufficiency of the endometrium that may allow subsequent pregnancies. We suggest interrupted sutures
villous invasion, and once the myometrium has been rather than continuous sutures for the inner layer of the
invaded by trophoblasts, further invasion cannot be pre- uterine myometrium (but with a very low quality of evi-
vented. Because this is an observational study, the biological dence, this is a weak recommendation). Because of the lim-
basis of why continuous sutures had a higher risk of pla- itation of study design, this was a preliminary study and
centa accreta than interrupted sutures remains unexplained; did not have an enough power to influence current clinical
however, we supposed that the more haemostatic ischaemic care; however, there was potential clinical relevance. As
effect of continuous sutures might influence the incomplete only a small number of patients were involved, large-scale
healing of the endometrium. There were other possibilities: studies are needed.
the apposition of the endometrium and myometrium would
also be considerable. The endometrium should be carefully Disclosure of interests
excluded from the wound so as not to be interposed The authors report no conflicts of interest.
between the muscle layers, but our study could not find a
relationship with this. In addition, the use of interrupted Contribution to authorship
sutures must increase residual foreign materials in the scar, This article involved 19 authors and each played an essen-
compared with continuous sutures, which might affect tial part, as follows: SS, supervisor, data collection, analysis
wound healing; however, the influence was difficult to iden- and writing, caesarean section, data collection, and analysis;
tify. We also examined whether the interval between prior TK, AI, and HT, data collection; YM, image analysis; YS,
caesarean section and the establishment of subsequent preg- pathological analysis; KT, statistical analysis; HH, MF, OY,
nancy influenced the incidence of placenta accreta, and YK, TO, HN, KM, KS, HO, MK, and FK, responsible
showed no difference in four groups (<12 months, 12– person at each hospital.
24 months, 24–36 months and ≥36 months). It is likely that
efficient healing of the endometrium occurs within Details of ethics approval
12 months of the caesarean section, and additional healing The Institutional Review Board for Human Research at
would be limited in subsequent periods. Nagoya University approved the collection of patient data
Our study supported previously published data that the for research purposes (approval number 413, initial version
anterior/centre position of the placenta is a risk factor for on 31 August 2006, revised version on 31 July 2008).
placenta accreta.26 Although placenta praevia is a well--
known risk factor,5–8 a PubMed literature search using the Funding
keywords ‘risk factor’, ‘accreta’, and ‘total/complete pla- This work was funded in part by a Grant-in-Aid
centa praevia’, or ‘risk factor’, ‘accreta’, ‘praevia’, and (24791699) to S.S. from the Ministry of Education, Science,
‘review’, revealed no previous reports that showed a differ- Sports, Technology, and Culture of Japan.
ence in the risk for placenta accreta by praevia type. In this
study, total placenta praevia showed an increased risk com- Acknowledgement
pared with non-total placenta praevia. These conditions, We greatly appreciate the cooperation and support of the
i.e. anterior/centre position or total placenta praevia, all the clinicians involved in this study. &

ª 2014 Royal College of Obstetricians and Gynaecologists 873

 Sumigama et al.

References 15 Hill AB. The environment and disease: association or causation? Proc
R Soc Med 1965 May;58:295–300.
1 Dodd JM, Anderson ER, Gates S. Surgical techniques for uterine 16 Schunemann H, Hill S, Guyatt G, Akl EA, Ahmed F. The GRADE
incision and uterine closure at the time of caesarean section. approach and Bradford Hill’s criteria for causation. J Epidemiol
Cochrane Database Syst Rev 2008;16:CD004732. Community Health 2011 May;65:392–5.
2 Levine D, Obstetric MRI. J Magn Reson Imaging 2006 Jul;24:1–15. 17 Cunningham F. Wiliams obstetrics. 21st ed: McGraw-Hill
3 Group CC, Abalos E, Addo V, Brocklehurst P, El Sheikh M, Farrell B, Professional; 2001.
et al. Caesarean section surgical techniques (CORONIS): a fractional, 18 Landon MB. Placental extraction and uterine repair. In: Gabbe SG,
factorial, unmasked, randomised controlled trial. Lancet 2013; editor. Obstetrics: Normal and Problem Pregnancies. 5th edn.
382:234–48. London: Churchill Livingstone; 2007. p. 495.
4 O’Brien JM, Barton JR, Donaldson ES. The management of placenta 19 Hauth JC, Owen J, Davis RO. Transverse uterine incision closure: one
percreta: conservative and operative strategies. Am J Obstet versus two layers. Am J Obstet Gynecol 1992 Oct;167:1108–11.
Gynecol 1996;175:1632–8. 20 Jelsema RD, Wittingen JA, Vander Kolk KJ. Continuous, nonlocking,
5 Belfort MA. Placenta accreta. Am J Obstet Gynecol 2010;203:430–9. single-layer repair of the low transverse uterine incision. J Reprod
6 Miller DA, Chollet JA, Goodwin TM. Clinical risk factors for Med 1993;38:393–6.
placenta previa-placenta accreta. Am J Obstet Gynecol 1997 Jul; 21 Ferrari AG, Frigerio LG, Candotti G, Buscaglia M, Petrone M,
177:210–4. Taglioretti A, et al. Can Joel-Cohen incision and single layer
7 Usta IM, Hobeika EM, Musa AA, Gabriel GE, Nassar AH. Placenta reconstruction reduce cesarean section morbidity? Int J Gynaecol
previa-accreta: risk factors and complications. Am J Obstet Gynecol Obstet 2001 Feb;72:135–43.
2005 Sep;193:1045–9. 22 Hibbard L. Cesarean section and other surgical procedures. In:
8 Oyelese Y, Smulian JC. Placenta previa, placenta accreta, and vasa Gabbe S, editor. Obstetrics. London: Churchill Livingstone; 1986.
previa. Obstet Gynecol 2006 Apr;107:927–41. pp. 517–49.
9 ACOG committee opinion. Placenta accreta. Number 266, January 23 Hohlagschwandtner M, Chalubinski K, Nather A, Husslein P, Joura
2002. American college of obstetricians and gynecologists. Int J EA. Continuous vs interrupted sutures for single-layer closure of
Gynaecol Obstet 2002 Apr;77:77–8. uterine incision at cesarean section. Arch Gynecol Obstet 2003
10 Silver RM, Landon MB, Rouse DJ, Leveno KJ, Spong CY, Thom EA, Apr;268:26–8.
et al. Maternal morbidity associated with multiple repeat cesarean 24 Bujold E, Goyet M, Marcoux S, Brassard N, Cormier B, Hamilton E,
deliveries. Obstet Gynecol 2006 Jun;107:1226–32. et al. The role of uterine closure in the risk of uterine rupture.
11 Clark SL, Koonings PP, Phelan JP. Placenta previa/accreta and prior Obstet Gynecol 2010 Jul;116:43–50.
cesarean section. Obstet Gynecol 1985 Jul;66:89–92. 25 Bujold E, Bujold C, Hamilton EF, Harel F, Gauthier RJ. The impact of
12 Sumigama S, Itakura A, Ota T, Okada M, Kotani T, Hayakawa H, a single-layer or double-layer closure on uterine rupture. Am J
et al. Placenta previa increta/percreta in Japan: a retrospective study Obstet Gynecol 2002 Jun;186:1326–30.
of ultrasound findings, management and clinical course. J Obstet 26 Gielchinsky Y, Rojansky N, Fasouliotis SJ, Ezra Y. Placenta accreta–
Gynaecol Res 2007 Oct;33:606–11. summary of 10 years: a survey of 310 cases. Placenta 2002;23:210–4.
13 Okai T. Ultrasonic diagnosis of placenta previa and placenta accreta. 27 Khong TY, Healy DL, McCloud PI. Pregnancies complicated by
Acta Obst Gynaec Jpn 2008;60:201–9. abnormally adherent placenta and sex ratio at birth. BMJ 1991 Mar
14 Frank HG, Kaufmann P. Pathology of trophoblast invasion. In: 16;302:625–6.
Benirschke K, Kaufmann P, Baergen R, editors. Pathology of the
Human Placenta. 5th edn. Berlin: Springer; 2006. pp. 268–73.

874 ª 2014 Royal College of Obstetricians and Gynaecologists

 Uterine suture techniques and placenta accreta

Surgical techniques at caesarean might modify accreta risk

DJ Lyell
Stanford University Medical Center, Stanford, CA, USA

Mini-commentary on ‘Uterine sutures at prior caesarean section and placenta accreta in subsequent pregnancy:
a case–control study’

Sumigama et al. report a unique and type-2 errors exists. Only the 382:234–48) and CAESAR (CAESAR
association between continuous hys- immediate prior caesarean operative study collaborative group, BJOG
terotomy closure at prior caesarean note was reviewed, yet 22 women 2010;117:1366–76) randomised trials
delivery and placenta accreta. In a underwent more than one prior cae- of common surgical techniques at
multicentre retrospective study the sarean; 34 women also underwent a caesarean, and the Cochrane Review
authors identified 98 cases of pla- prior miscarriage or abortion, but (Dodd et al. Cochrane Database Syst
centa praevia or low-lying placenta, data on dilation and curettage were Rev 2008;3) of the topic, examine the
of whom 38 had pathologically diag- not examined. It is conceivable that a techniques of single- and double-layer
nosed placenta accreta. They exam- surgical technique from an earlier cae- hysterotomy closure, without specify-
ined prior operative notes to identify sarean or scarring from a prior dila- ing whether this involves continuous
suture techniques among women tion and curettage could have or interrupted hysterotomy closure.
with at least one prior caesarean with predisposed these women to accreta As the rates of caesarean delivery
transverse hysterotomy, without per- formation. Furthermore, some miss- and placenta accreta rise worldwide,
manent suture use. Placenta accreta ing surgical technique data from out- much accreta research has been direc-
occurred among 29.9% of women side hospitals was extrapolated from ted towards understanding risk fac-
when the inner layer of the previous the ‘surgeon’s or hospital’s standard tors, diagnosis, and optimal surgical
hysterotomy was closed with inter- procedure at the time’. The authors treatment of placenta accreta, with
rupted sutures, versus 58.1% when acknowledge the limits of their sam- almost no focus on prevention after
continuous sutures were used ple, citing their presumed erroneous the decision to proceed to caesarean
(P = 0.008). When multivariable finding that multiple prior caesarean has been made. Sumigama et al.
logistic regression analysis was per- deliveries, the most common risk fac- uniquely focus on a modifiable tech-
formed to adjust for group differ- tor for accreta development, was not nique that might be used to decrease
ences, history of continuous suture associated with placenta accreta. the development of accreta once the
use remained associated with accreta The authors’ approach is novel in caesarean decision has been made.
(aOR 6.0, 95% CI 1.4–25.2). Of note, that it addresses a potentially modifi- Although admirable in its scope and
no differences in accreta development able surgical technique. They cite a approach, the current study should be
were seen with prior single-layer ver- biologically plausible mechanism for replicated before being adopted into
sus double-layer hysterotomy closure, accreta development, extrapolated widespread practice. Hopefully this
or with use of locking sutures. from Gabbe’s Obstetrics: that continu- study will help to usher in many more
As the authors note, their study has ous sutures may impair wound heal- studies of accreta prevention from the
several limitations. Although a study ing in the myometrium/endometrium operating room.
of this sort is difficult to conduct – as a result of pressure-induced ischae-
the authors obtained operative notes mia. How frequently surgeons per- Disclosure of interests
from 61 hospitals – their numbers are form interrupted hysterotomy closure I have no conflicts of interest related
relatively small, multiple comparisons worldwide is unclear. The large in any way to this commentary. &
are made, and the potential for type-1 CORONIS (Abalos et al. Lancet 2013;

ª 2014 Royal College of Obstetricians and Gynaecologists 875