Placenta accreta: incidence and risk factors in an area with a particularly high rate of cesarean section

Authors


  • The authors have stated explicitly that there are no conflicts of interest in connection with this article.

Correspondence

Pasquale Martinelli, High Risk Pregnancy Unit, Department of Obstetrics and Gynecology, University Federico II, Via S. Pansini, 5, 80131 Naples, Italy. E-mail: martinel@unina.it

Abstract

Placenta accreta is a rare and potentially life-threatening complication of pregnancy characterized by abnormal adherence of the placenta to the uterine wall. A previously scarred uterus or an abnormal site of placentation in the lower segment is a major risk factor. The aim of this study was to investigate the change in the incidence of placenta accreta and associated risk factors along four decades, from the 1970s to 2000s, in a tertiary south Italian center. We analyzed all cases of placenta accreta in a sample triennium for each decade. The incidence increased from 0.12% during the 1970s, to 0.31% during the 2000s. During the same period, cesarean section rates increased from 17 to 64%. Prior cesarean section was the only risk factor showing a significant concomitant rise. Our results reinforce cesarean section as the most significant predisposing condition for placenta accreta.

Abbreviations
CS

cesarean section

PA

placenta accreta

Introduction

Placenta accreta (PA) is a potentially life-threatening complication of pregnancy characterized by an abnormal adherence of the placenta to the uterine wall [1], secondary to an absence or deficiency of Nitabuch's layer of the decidua [2]. The term abnormal placentation is colloquially used for the three known variants of placenta accreta, increta or percreta [1]. The clinical consequence of abnormal placentation is failure of placental separation leading to massive postpartum hemorrhage with a significant increase in maternal morbidity and mortality [1]. The reported incidence of abnormal placentation is highly variable, ranging from 1:93 000 to 1:111 pregnancies [4]. A deficit in the uterine wall thickness due to a scarred uterus or an abnormal placentation site in the lower segment is a major risk factor [5]. An increasing incidence of abnormal placentation has been considered most likely related to much higher rates of cesarean section (CS) [5, 6]. Countries with a high CS rate, such as Italy [7], are expected to have an increased incidence.

We have investigated the changes in the incidence of PA and associated risk factors along four decades from 1970s in a tertiary south Italian center.

Material and methods

A retrospective study of medical charts to identify all patients with PA was conducted. To evaluate incidence variation from the 1970s to 2000s we analyzed all cases of PA (increta and percreta are included as they could often not be safely distinguished from accreta) in a sample triennium for each decade. Printed copies of the clinical notes were available starting from the 1976. The first triennium sample was considered from January 1976 to December 1978; then in any 10-year interval, an analogue three-year period at was studied, i.e. 1986–1988, 1996–1998 and 2006–2008.

Placenta accreta was defined as any abnormal adherence of the placenta to the uterine wall (“accretism”) [1]. Diagnosis had to be based on clinical and histological findings [4, 6, 8], using (i) histopathologic confirmation on a hysterectomy specimen by absence of the intervening layer of decidua, Nitabuch's layer [2], between placenta and myometrium, (ii) incomplete manual removal of the placenta despite active management of the third stage of labor or (iii) heavy continued bleeding from the implantation site of a well-contracted uterus after difficult removal of the placenta during CS.

Variables included in the analysis were: maternal age, parity, previous abortions and curettages, CS, any other uterine surgery, placenta previa according to third trimester ultrasound examination, in vitro fertilization, uterine artery embolization in a previous pregnancy, female newborn gender [3-5, 8, 9]. Risk factors for PA were analyzed using the chi-squared test for categorical variables, and an ANOVA test for continuous variables. P-values <0.05 were considered significant. The institutional ethical committee approved the study.

Results

During the four triennia there were 30 491 deliveries at our center, from which 50 cases of PA were diagnosed (Table 1). The incidence of PA grew from 0.12% (1/833) during 1976–1978, to 0.31% (1/322). At the same time, CS rate went from 17 to 64% during the last triennium (Figure 1).

Table 1. Risk factor for PA in the four decades and in the total cohort. Prior CS is the only risk factor significantly different between the four decades
 1970s1980s1990s2000sTotal
  1. a

    Statistically significant (p < 0.05).

  2. PA, placenta accreta; IVF, in vitro fertilization.

 n % n % n % n % n %
PA cases110.1240.06150.17200.3150
Prior cesarean sectiona43625032014702346
12180 0 5257 14 
≥221825032094516 32 
Age ≥ 35 years54525032073517 34 
Placenta previa545250320126022 44 
Parity ≥ 2764375320105023 46 
Prior curettage4360 96084021 42 
Other prior uterine surgery2 (18)0 0 0 24
IVF0 0 0 1512
Prior uterine embolization0 0 0 1512
Uterine malformation190 0 0 12
Female sex of newborn98225064011552856
Prior cesarean sectiona, mean ± sd0.6 ± 11.2 ± 1.50.4 ± 0.81.3 ± 10.9 ± 1
Maternal age, mean ± sd32 ± 5.532.7 ± 6.832.8 ± 533.5 ± 532.9 ± 5.1
Parity, mean ± sd2 ± 1.32 ± 1.40.9 ± 0.71.4 ± 1.21.4 ± 1.2
Prior curettage, mean ± sd0.5 ± 0.901 ± 0.90.5 ± 0.60.6 ± 0.8
Figure 1.

Rates (%) of placenta accreta (PA, dashed line) and cesarean section (CS, solid line) at our center in the last four decades (1970s–2000s). Note that different scales for placenta accreta (right axis) and cesarean section (left axis) are used.

Of the PA cases, nine women delivered vaginally. Among them, seven had blood products transfusions, five had dilatation and curettage, one of whom required a hysterectomy. Among women delivering vaginally, there were four hysterectomies due to uncontrollable bleeding.

Forty-one women were delivered by CS: 23 had transfusions, three were successfully treated with curettage and 26 required a hysterectomy. One woman had hypogastric artery ligation as adjuvant treatment to reduce hemorrhage during the cesarean hysterectomy. Twelve women were successfully treated by uterine packing only. There were 30 hysterectomies. In nine cases (30%) the histology confirmed the PA (four increta, five percreta); in 14 the histological result was negative (46%). Seven of the early cases from the first triennia could not be reviewed due to lost or destroyed documents. There were 13 primiparous women, three of whom delivered vaginally, while CS was carried out in 10 for different obstetrical indications. There were no cases of maternal death.

Table 1 shows risk factors in the four decades. No significant differences were seen for any of the most common variables, except previous CS (p < 0.05).

Discussion

This observational study shows an increasing incidence of PA over time from the 1970s to the 2000s. Risk factors did not change to any significant degree over the last four decades except for CS.

Due to the absence of 23% of histology reports, the diagnosis was based mostly on clinical criteria. The literature is controversial on the sensitivity and specificity of the clinical criteria compared with histological diagnosis [3, 4, 8, 10]. The exclusion of the cases with negative histological examination may underestimate the real incidence [3]. The absence of indicative histological features in cases of clinically suspected PA does not exclude the diagnosis [10]. We excluded all cases of simple retained placenta. Most of the cases were discovered at CS and a senior consultant was always involved in the management. It is therefore unlikely that the PA false-positive rate would have influenced the incidence rate, even if some cases of retained placenta were considered PA.

Other authors have reported rising rates of PA in the last decades [4, 6]. To the best of our knowledge, this is the longest observed period reporting the last 40 years of PA frequency variance. The incidence in the last decade is comparable with more recently published studies (0.01–0.9%) [4, 6, 8].

The high CS rate was the only characteristic significantly different from 1970s to 2000s (from 18 to 63%). The possible explanation for this has been investigated previously. The human embryo develops in a relatively hypoxic environment, and data from in vitro studies suggests that oxygen tension determines whether cytotrophoblasts proliferate or invade, thereby regulating placental growth [11]. Embryos may preferentially implant into areas of uterine scarring because of the lower vascularization and lower oxygen tension.

Our study has several limitations, including the retrospective evaluation of case notes where reporting was not consistent with regard to histology and data entry, affecting the reliability of the clinical diagnosis. Moreover, electronic databases were not available from the 1970s and 1980s and this leads us to assume that a considerable amount of missing data cannot be recovered to evaluate the real incidence of PA in the last 40 years. Whether the four sample triennia are representative of the entire decade cannot be verified.

Considering the inevitable worldwide increasing rate of CS, further efforts should be spent on screening and management to prevent the consequent rise in maternal morbidity and mortality due to PA [6].

Acknowledgements

The authors would like to thank Daniela Russo, MD for her research work on anatomopathology records.

Funding

No special funding.

Ancillary