Evaluation and Comparison of Sensitivity and Specificity of Ultrasonography in Placenta Accreta Diagnosis in the Second and Third Trimesters

Abstract

Background:

Ultrasound is the selected technique for the detection of placenta accreta spectrum (PAS). This method can detect PAS in 80%-50% of cases. This study aimed to assess and compare the sensitivity and specificity of ultrasonography in the diagnosis of PAS after the first trimester.

Materials and Methods:

In this prospective study that was performed in 2020-2021 on 79 patients at high risk of PAS, all cases underwent ultrasonography in both 18-22 weeks of gestational age (GA) and 32-34 weeks of GA for evaluation of accreta. As per the risk factors, the delivery plan for all mothers was cesarean section. During the cesarean section, the placenta was examined for accreta, and if it was attached to the uterus, a diagnosis of placenta accreta was ascertained and a total abdominal hysterectomy was performed if the patient's bleeding was not controlled during the operation. The final diagnosis of PAS was made based on the pathology report.

Results:

Ultrasound evaluation for PAS in 18-22 weeks of GA had 79.17% specificity, 51.61% sensitivity, 61.54% positive predictive value, and 71.70% negative predictive value. Ultrasound imaging for PAS in 32-34 weeks of GA had 60.8% specificity, 90% sensitivity, 62.52% positive predictive value, and 90.33% negative predictive value.

Conclusion:

It should be concluded that PAS is a critical condition and if the patient is diagnosed in the second or third trimester, special strategies should be applied.

INTRODUCTION

Placenta accreta spectrum (PAS) is a critical condition during pregnancy that results from abnormal placental penetration into the myometrium, leading to heavy bleeding during vaginal delivery.[1]

PAS is the definitive indication for emergency hysterectomy after birth. Severe bleeding when the placenta separates from the uterus (more than 2 liters) leads to diffuse intravascular coagulation disorder, renal failure, and death.[2,3] Hysterectomy is essential in most cases. However, it is associated with complications such as intensive care unit admission (26.2%), cystostomy (15.4%), ureteral injury (2.1%), and, pulmonary embolism (2.1%).[4]

Placenta accretes major risk factors are maternal age (aged more than 35 years) and previous cesarean section (CS). In addition, factors such as previous uterine curettage, extracting the placenta by hand, hysteroscopy, uterine malformations, previous surgery on the uterus, multiparity, uterine fibroids, and maternal sepsis are other predisposing factors.[5] Over the past three decades, the prevalence of placenta accreta in the United States has increased to one per 25,000 births.[6] As per researchers, the reason for the high prevalence of this disease is the increase in the rate of uterine surgeries, including curettage and CS.[7]

With an accurate diagnosis of PAS delivery should be scheduled at a time with optimal availability of necessary personnel and facilities. All of these factors together help reduce maternal morbidity and mortality.[8]

PAS selected detecting technique is ultrasound. This method can detect the PAS in 80%-50% of cases.[9] Despite attempts to diagnose PAS after the first trimester of pregnancy using ultrasound and magnetic resonance imaging (MRI), prenatal and postpartum hemorrhage is the diagnosis signs of many cases of accreta.[10,11] Currently, as per statistics published by the Ministry of Health, the average CS in Iran is more than 3 times the global statistics.[12] About 60% of CS performed in Iran are related to private centers.[13]

As per statistics published in 2017, the prevalence of CS in some parts of Iran was more than 70% of births per year.[14] Following an increase in the rate of CS, its complications, including PAS, will increase significantly. It is very important to have an affordable, accurate, and reliable diagnostic method for timely diagnosis of PAS during the second and third trimesters in high-risk mothers.[15]

So far, studies show the sensitivity and specificity of placenta accreta diagnostic methods. However, very limited studies have been performed to compare the diagnostic value of ultrasound during the second and third trimesters, which, of course, have yielded contradictory results that require further study. Combined ultrasound is still a widely used screening modality because of its low cost, sensitivity, noninvasive, and lower duration. This is a popular technique because it is patient-friendly, available, and quite inexpensive.[16]

The sensitivity of ultrasound findings varies in the different studies and is reported between 33% and 100% and specificity is also widely different.[16,17,18,19,20] Recent studies have made the prenatal diagnosis in the weeks of 18-22 and at the age of 32-34 weeks of pregnancy; they undergo ultrasound again. All cases were followed for outcome till term.

This study aims to assess and compare the sensitivity and specificity of ultrasonography in the placenta accretes diagnose in the second and third trimesters. We decided to find out if it is enough to do one ultrasound to diagnose placenta accreta in the second trimester alone in high-risk pregnant women or if it is necessary to assess the ultrasound in the last trimester again.

MATERIALS AND METHODS

This is a prospective descriptive-analytic study that was performed in 2020-2021 in Beheshti and Al-Zahra hospitals affiliated with Isfahan University of Medical Sciences. Convenience sampling was used for patient selection. The present study was performed on 79 patients at high risk of PAS including the history of CS, curettage, myomectomy, and anterior placenta that referred to a perinatology clinic using easy sampling. The study has been approved by the University Ethics Committee as per code IR.MUI.MED.REC.1399.701 and has been performed by the ethical standards presented in the Declaration of Helsinki and its later amendments.

The inclusion criteria were age more than 18 years, gravid of two or more, and presence of anterior placenta in the current pregnancy associated with having at least one of these risk factors: previous history of CS, history of curettage, myomectomy, and placenta previa. All cases signed a written informed consent to participate in this study. The exclusion criteria were the patient's will to exit the study, unavailability to access the patient's data during follow-up, and maternal death before childbirth.

This study sample size was calculated by sample size formula in receiver operating characteristic curve studies based on the expected sensitivity of 70% (TPR = 0.8) and 70% specificity (FPR = 0.3) and 95% confidence interval equal to 0.18 (W = 0.18) and the patient-to-control ratio was calculated to be 1 (R = 1). Totally, 80 pregnant women were considered as the study population.

Participants’ demographic data, including age, gravid, and previous gynecology information, were collected by a checklist at the beginning of the study.

Because these mothers had risk factors for placenta accreta, they were referred to a perinatologist for ultrasound screening during the second and third trimesters. All ultrasounds were performed by a professor of Perinatology. The patients with suspected placenta accreta in the second and third trimesters underwent transvaginal ultrasonography (with some fluid in the bladder, so that the uterine bladder interface could be estimated well). Color Doppler ultrasonography was not used in any case. The sonography of the placenta was performed using the GE vuluson E6 sonography device and the transducer 3.5–5 MHZ. The placenta was examined using the criteria of the grayscale. All examinations were recorded on videotape.

Diagnostic criteria that suggested placenta accreta, increta, or percreta included one or more of the following situations:

(1) obliteration of the clear space, by definition: the obliteration of any part of the echo-lucent area located between the uterus and placenta; (2) visualization of placental lacunae, defined as multiple linear, irregular vascular spaces within the placenta; and (3) interruption of the posterior bladder wall-uterine interface such that the usual continuous echo-lucent line appears instead as a series of dashes. If the possibility of placenta accreta was suggested in at least 1 scan, that case was considered “positive” by ultrasound, even if this suggestion was revoked in subsequent scans. As per the second and third trimester ultrasonography reports, pregnant mothers fall into two groups: positive PAS and negative PAS. All cases underwent ultrasonography in both 18-22 weeks of gestational age (GA) and 32-34 weeks of GA.

As per the risk factors, the delivery plan for all mothers was CS. During the CS, the placenta was examined for accreta, and if it was attached to the uterus, a diagnosis of placenta accreta was ascertained and a total abdominal hysterectomy (TAH) was performed if the patient's bleeding was not controlled during the operation. The placenta was then sent to the pathology and the final diagnosis of PAS was made based on the pathology report.

The primary outcome of this study was to assess the sensitivity and specificity of the ultrasonographic method in the PASs diagnosis. As per the following formulas, the negative and positive predictive values, sensitivity, and specificity of ultrasonography were calculated for the PASs diagnosis after the first trimesters.

The negative predictive value (NPV) was defined as the ratio of subjects truly diagnosed as negative to all those whose test result was negative. The positive predictive value (PPV) was considered as the ratio of patients truly diagnosed as positive to all those whose test result was negative. The sensitivity was considered as the probability of a positive test, conditioned on truly having the condition, and the specificity was the ability of the test to correctly identify people without the disease.

The obtained data were entered into the Statistical Package for Social Sciences (SPSS) (version 26, SPSS Inc., Chicago, Illinois). Numerical data were present in mean, standard deviation, and descriptive data as frequency distribution (percentage). Chi-square and independent t-tests were used to analyze the data. The P value with an amount of less than. 05 was considered a significant level.

RESULTS

In this study, we evaluated data from 80 patients with the presence of anterior placenta in the current pregnancy associated with having at least one of these risk factors: previous history of CS, history of curettage, myomectomy, and placenta previa that referred to perinatology clinic. One case was excluded due to incomplete data. Data from 79 patients were analyzed. Among them, 33 patients had placenta accreta and 28 of them underwent TAH and five patients had focal accreta and the uterus could be saved. We had 8 cases of percerta, 3 cases of pervia percreta, 4 cases of increta, and 14 cases of accreta. The primary analysis of demographic data showed that the mean age of the patients was 32.76 ± 3.98 years. A comparison of demographics and the history of patients showed significant differences between the two groups. It was observed that patients who underwent TAH had significantly higher age (P <.001), higher gravidity (P =0.003), higher abortion rates (P =0.006), higher curettage rates (P =.029), higher rates of CS history (P =.05), lower GA based on ultrasonography and last menstrual period (P <.001 for both), and higher total risk factors (P <.001). These data are shown in Table 1.

Table 1.

Comparison of different demographic data between two groups

	Outcome			P b
	CS without TAH	CS with TAH	Total
Age	31.54±3.23a	34.65±4.35	32.76±3.98	<0.001
Gravid	2.75±1.04	3.52±1.23	3.05±1.17	0.003
Live	1.46±0.87	1.81±0.95	1.59±0.91	0.055
Abortion	0.13±0.39	0.48±0.72	0.27±0.57	0.006
NVD	0.10±0.42	0.19±0.79	0.14±0.59	0.934
Curettage	0	0.10±0.30	0.04±0.19	0.029
CS	1.48±0.74	1.81±0.91	1.61±0.82	0.050
Death	0.17±0.52	0.23±0.50	0.19±0.50	0.416
GA_LMP_w	36.57±1.63	34.94±1.34	36.09±1.71	<0.001
GA_sono_w	36.75±1.41	35.03±1.56	36.08±1.68	<0.001
Risk factor	2.69±1.03	3.81±1.05	3.12±1.16	<0.001

a: Data are represented by mean ± standard deviation; b: Mann-Whitney U test

It was also shown that 67.08% of patients did not have PAS based on the second-trimester ultrasound evaluations and 32.92% had accreta. Of patients with no accreta in second-trimester ultrasound, 58.5% had PAS based on third-trimester ultrasound evaluations and 41.5% had no accreta. Only 51.6% of patients with positive accreta in third-trimester ultrasound had PAS based on surgical outcomes but 95.45% of patients with negative accreta results based on third-trimester ultrasound did not have accreta based on surgical results. Among 26 patients with positive accreta based on second-trimester ultrasound evaluations, 65.39% had positive results based on the third-trimester ultrasound and 82.4% of them had accreta based on surgery. These data are shown in Figures 1 and 2.

Figure 1.

Outcome of 26 women with positive findings of placenta accrete in the second-trimester ultrasonography

Figure 2.

Outcome of 53 women with negative findings of placenta accrete in the second-trimester ultrasonography

Further analysis showed that ultrasound evaluation for placenta accreta in 18-22 weeks of GA had 79.17% specificity, 51.61% sensitivity, 61.54% PPV, and 71.70% NPV. Ultrasound imaging for placenta accreta in 32-34 weeks of GA had 60.8% specificity, 90% sensitivity, 62.52% PPV, and 90.33% NPV [Table 2]. An important point was that ultrasound evaluation in the 18-22 gestational week had high PPVs and high specificity but the third-trimester ultrasound evaluations had higher NPVs and sensitivity. It could be determined that pregnant women with higher risk factors including higher age, higher gravidity, higher abortion rates, higher curettage rates, higher rates of CS history, and lower GA should be evaluated with ultrasound and negative cases should be re-evaluated in 32-34 weeks of GA for possible PAS.

Table 2.

Sensitivity and specificity of ultrasound for placenta accrete diagnosis according to different pregnancy trimester

Ultasonography	Sensitivity	Specificity	PPV	NPV
Second trimester	51.61%	79.17%	61.54%	71.70%
Third trimester	90%	60.8%	62.53%	90.33%

DISCUSSION

In this study, we evaluated data from 79 patients with high risks of placenta accreta and assessed the sensitivity and specificity of using ultrasound in the diagnosis of the disease. Our data showed that ultrasound evaluation for PAS in 18-22 weeks of GA had 79.17% specificity, 51.61% sensitivity, 61.54% PPV, and 71.70% NPV. Ultrasound imaging for PAS in 32-34 weeks of GA had 60.8% specificity, 90% sensitivity, 62.52% PPV, and 90.33% NPV. These data show that ultrasound imaging of high-risk women both at 18-22 and 32-34 weeks of GA could be a very useful technique in the early diagnosis of placenta accreta.

Furthermore, we observed that patients with placenta accreta who underwent TAH had significantly higher age, higher gravidity, higher abortion rates, higher curettage rates, higher rates of CS history, lower GA based on ultrasonography and last menstrual period, and higher total risk factors compared to patients in CS without TAH group.

Some previous studies on the diagnosis of PAS with ultrasonography. In 2017, Ayati et al. evaluated data from 82 cases with high risks of PAS. In this study, it was reported high-risk women for PAS which have to undergo Doppler ultrasonography at first. When results on Doppler sonography are equivocal for PAS, MRI can be performed due to its high specificity (83%).[21] In another study by Satija and others in 2015, it was indicated that both techniques have relatively good sensitivity for prenatal diagnosis of PAS, although specificity was not remarkable in comparison to other studies. Both methods play a complementary role, and in cases where there are inconclusive findings with one imaging method, another method is useful for diagnosis.[22]

Doppler ultrasonography remains the first primary modality for antenatal diagnosis of placenta accreta, with MRI reserved for cases where ultrasonography is inconclusive. These findings are similar to our study. However, the important point is that we assessed diagnostic values of ultrasound imaging in two different GAs and observed acceptable sensitivity and specificity.

An important point was that ultrasound evaluation in the 18-22 gestational week had high PPVs and high specificity but the third-trimester ultrasound evaluations had higher NPVs and sensitivity. It could be determined that pregnant women with higher risk factors including higher age, higher gravidity, higher abortion rates, higher curettage rates, higher rates of CS history, and lower GA should be evaluated with ultrasound and negative cases should be re-evaluated in 32-34 weeks of GA for possible PAS.

In another study by Rac and others in 2015, it was found that the placenta accreta index by ultrasound may be useful in predicting the risk of placenta accreta from patient to patient.[23] Furthermore, in 2017, Jauniaux and others performed a systematic review and meta-analysis on the use of ultrasound in the diagnosis of PAS. They showed that we can consider ultrasound as a highly sensitive and specific method in the diagnosis of accreta placentation when performed by a skilled clinician. Improving the prenatal screening protocol is important to further better outcomes of this complication.

The reported specificity and sensitivity for ultrasound imaging in this study were 77.41% and 30.22%, respectively.[24] These data are similar to our findings.

In a study by Zalud et al. in 2016, it was stated that the cause of accreta was due to the claim of maternal decidua, which was caused by the invasion of the placenta into the uterine myometrium. The molecular basis for invasive placentation is yet to be elucidated but may involve abnormal paracrine/autocrine signaling between maternal decidua and trophoblast tissue. They showed that women with higher characteristics include age, gravidity, abortion rates, curettage rates, and rates of CS history have a higher possibility of having PAS.[17]

In 2014, Fateme Rahimi-Sharbaf et al. in the Perinatology Department of Tehran University evaluated 323 high-risk women for placenta accreta. They examined a woman in 14-week pregnancy by vaginal and abdominal ultrasound. They were examined in the second and third trimesters. The sensitivity and specificity of the US in the second trimester were 60% and 83.5%, which are the same as our study. The sensitivity and specificity of the US in the third trimester were 71.4% and 88.5% which is different from our study. Rahimi concludes that the US (Ultra Sonography) has good accuracy for PAS in both the second and third trimesters.[18]

In 2021, Ahmed Mohammad Hamda et al. evaluate the sensitivity and specificity of the US in 100 cases. The US sensitivity was 100%, specificity was 85%, NPV was 100%, and PPV was 90.9%. This study did not specify the sensitivity and specificity of ultrasound in each trimester. What was determined separately in our study?[16]

Christin H in 2003 over 12 years of the study period performed 163,855 obstetrics examinations. This study's result was the sensitivity and PPV of the US in the second trimester were 86% and 63%. The sensitivity and PPV of the US in the third trimester were 100% and 48%. In this study, results suggest that a single sonographic marker can predict many cases of accreta as early as 15-20 weeks of gestation. Our finding in this study is different from Chiristin's finding. It can be concluded from our study that the US has higher sensitivity in the third trimester and higher specificity in the second trimester. It means that gynecologists should consider performing the second US evaluation of PAS in high-risk cases in 32-34 weeks.[19]

In 2019, Hui Xia and Shu Chang Ke compare abdominal US detection of placenta accreta after the first trimester. They suggest that high-risk pregnant women for placenta accreta at about 20 weeks should be screened by the US. Data showed that the sensitivity, specificity, PPV, and NPV of abdominal US were 95.65%, 91.78%, 88%, and 97% for the second trimester of pregnancy, respectively. In the third trimester of pregnancy, the sensitivity and NPV of abdominal US significantly enhanced relative to the second trimester. Our finding in this study is the same. It may be the reason that the blood vessel in the placenta of pregnant women in the last trimester was much denser.[20]

Our results are similar to previous data and could have high clinical importance. It was shown that third-trimester ultrasound evaluations have higher NPV compared to second-trimester evaluations. Therefore, conducting the third-trimester ultrasound evaluations could have an acceptable clinical justification. The limitations of our study were the restricted study population and lack of comparison with other placenta disorders. It is recommended that gynecologists should consider performing the third-trimester ultrasound evaluation of placenta accreta in high-risk cases.

CONCLUSION

Evaluation for placenta accreta in 18-22 weeks of GA had 79.17% specificity, 51.61% sensitivity, 61.54% PPV, and 71.70% NPV. Ultrasound imaging for placenta accreta in 32-34 weeks of GA had 60.8% specificity, 90% sensitivity, 62.52% PPV, and 90.33% NPV. It is recommended that gynecologists should consider performing the third-trimester ultrasound evaluation of PAS in high-risk cases including patients with higher age, higher gravidity, higher abortion rates, higher curettage rates, higher rates of CS history, and lower GA. It should be concluded that PAS is a life-threatening condition and if the patient is diagnosed in the second or third trimester, special strategies should be applied. These strategies include referring the patient to more advanced third-degree hospitals with experienced anesthesiology and surgical teams. On the other hand, patients with negative results in both ultrasound evaluations could be managed in more ordinary centers.

Financial support and sponsorship

This study was granted by Isfahan University of Medical Sciences.

Conflicts of interest

There are no conflicts of interest.