*Article* **Comparison of the Results of Prenatal and Postnatal Echocardiography and Postnatal Cardiac MRI in Children with a Congenital Heart Defect**

**Marios Mamalis \*, Tamara Koehler, Ivonne Bedei, Aline Wolter, Johanna Schenk, Ellyda Widriani and Roland Axt-Fliedner \***

> Division of Prenatal Medicine & Fetal Therapy, Department of Obstetrics & Gynecology, Justus-Liebig-University Giessen, 35390 Giessen, Germany

**\*** Correspondence: marios.mamalis@yahoo.com (M.M.); roland.axt-fliedner@gyn.med.uni-giessen.de (R.A.-F.)

**Abstract:** Objective: In fetuses with suspicion of congenital heart disease (CHD), assessment by segmental fetal echocardiography is of great importance. This study sought to examine the concordance of expert fetal echocardiography and postnatal MRI of the heart at a high-volume paediatric heart centre. Methods: The data of two hundred forty-two fetuses have been gathered under the condition of full pre- and postnatal and the presence of a pre- and postnatal diagnosis of CHD. The haemodynamically leading diagnosis was determined for each test person and was then sorted into diagnostic groups. The diagnoses and diagnostic groups were used for the comparison of diagnostic accuracy in fetal echocardiography. Results: All comparisons between the diagnostic methods for detection of congenital heart disease showed an "almost perfect" (Cohen's Kappa > 0.9) strength of agreement for the diagnostic groups. The diagnosis made by prenatal echocardiography showed a sensitivity of 90–100%, a specificity and a negative predictive value of 97–100%, and a positive predictive value of 85–100%. The diagnostic congruence resulted in an "almost perfect" strength of agreement for all evaluated diagnoses (transposition of great arteries, double outlet right ventricle, hypoplastic left heart, tetralogy of Fallot, atrioventricular septal defect). An agreement of Cohen's Kappa > 0.9 was achieved for all groups, with exception of the diagnosis of double outlet right ventricle (0.8) in prenatal echocardiography compared to postnatal echocardiography. This study came to the result of a sensitivity of 88–100%, a specificity and negative predictive value of 97–100%, and a positive predictive value of 84–100%. The performance of cardiac magnetic resonance imaging (MRI) as an additional measure to echocardiography had an added value in the description of the malposition of the great arteries when diagnosed with double outlet right ventricle and in the detailed description of the anatomy of the lung circulation. Conclusions: Prenatal echocardiography could be shown to be a reliable method for detection of congenital heart disease when regarding the slightly lower accuracy of diagnosis for double outlet right ventricle and right heart anomalies. Furthermore, the impact of examiner experience and the consideration of follow-up examinations for further improvement of diagnosis accuracy may not be underestimated. The main advantage of an additional MRI is the possibility to obtain a detailed anatomic description of the blood vessels of the lung and the outflow tract. The conduction of further studies that include false-negative and false-positive cases, and studies that are not set within the high-risk-group, as well as studies in a less specialized setting, would allow the completion and investigation of possible differences and discrepancies when comparing the results that have been obtained in this study.

**Keywords:** congenital heart disease; echocardiography; postnatal cardiac MRI

### **1. Introduction**

Congenital heart defects (CHD) represent a group of congenital anomalies difficult to characterize prenatally. For this reason, an antenatally suspected congenital heart defect

**Citation:** Mamalis, M.; Koehler, T.; Bedei, I.; Wolter, A.; Schenk, J.; Widriani, E.; Axt-Fliedner, R. Comparison of the Results of Prenatal and Postnatal Echocardiography and Postnatal Cardiac MRI in Children with a Congenital Heart Defect. *J. Clin. Med.* **2023**, *12*, 3508. https:// doi.org/10.3390/jcm12103508

Academic Editors: Erich Cosmi and Dinesh K. Kalra

Received: 30 March 2023 Revised: 9 May 2023 Accepted: 15 May 2023 Published: 17 May 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

(CHD) via a fetal echocardiography is postnatally confirmed by a postnatal echocardiography and, on some occasions, postnatal cardiac magnetic resonance imaging (MRI) is adjunctively required.

In this retrospective explorative analysis, we describe our experience in the clinical management of two hundred forty-two fetuses with suspected congenital heart defects (CHD) referred to our unit for diagnosis and management over a five-year period.

The objective of this study is to assess whether and to what extent the prenatal echocardiography is a reliable method for diagnosis of several CHD and if postnatal cardiac MRI provides additional diagnostic value.

### **2. Methods**

During a five-year study period from 2012 to 2017, there have been two hundred fortytwo cases with suspected CHD referred to our fetal medicine unit (FMU) at the University Hospital of Giessen and Marburg (UKGM) for diagnosis and further management. All individuals, in addition to prenatal echocardiography, have underwent postnatal echocardiography and postnatal cardiac MRI examination, when indicated. Prenatal and postnatal echocardiography have been performed in accordance to published guidelines [1–3].

Statistical analysis was performed by the Institute of Biostatistics at Justus-Liebig University Giessen. The interpretation of results has been based on Cohen's Kappa coefficient and the concordance rate of Landis and Koch. The congenital heart defects' spectrum that has been met in our patient's collective is demonstrated in Table 1. The level of significance when testing concordance for all Cohen' s Kappa scores has been a: 5%. Additionally, the Bonferroni correction method and *p*-value have been utilized. Overall results are presented as descriptive data. Sensitivity (SEN) and specificity (SPE), and positive (PPV) and negative predictive values (NPV) were calculated.


**Table 1.** Classification of congenital heart defects in current study.

#### **3. Results**

Two hundred forty-two cases were included in the study. Twenty-four cases were excluded due to termination of pregnancy; seven intrauterine demises occurred. Two hundred eleven cases remained for final analysis.

Table 2 shows the overall prevalence of CHD in our cohort. The strength of agreement for CHD in the different groups has been evaluated (Table 3). Comparing the several CHD in the prenatal-postnatal echocardiography group, the anomalies of right heart and complex CHD have shown the lowest strength of agreement (Cohen' s Kappa: 0.92). For conotruncal and outflow tract anomalies, and septal defects, the strength of agreement has been 0.93 and for left heart anomalies, 0.96.



The lowest strength of agreement has been indicated for septal defects in the prenatal echocardiography-MRI group (Cohen's Kappa: 0.91). The anomalies of right heart and complex CHD have shown a concordance of Cohen's Kappa as high as 0.94 and the conotruncal and outflow tract anomalies' Cohen's Kappa has been 0.95. The left heart anomalies have demonstrated the second highest strength of agreement (Cohen's Kappa: 0.98) below the "Other anomalies" group.


**Table 3.** Strength of agreement; comparison among prenatal echocardiography, postnatal echocardiography and cardiac MRI.

In the postnatal echocardiography-postnatal MRI group, the diagnosis of septal defects has been "perfectly" concordant. The left heart anomalies have shown a concordance rate of 0.98 and the conotruncal and outflow tract anomalies, and right heart anomalies have similarly achieved high concordance (Cohen's Kappa: 0.97).

The comparison of the precision of the diagnosis via different diagnostic methods by utilized Cohen's Kappa has proved to have a significance of *p* < 0.001 in all cases. After implying the Bonferroni correction, all tests for CHD have proved significance.

In Table 4, the predictive parameters for CHD via prenatal echocardiography, compared to postnatal echocardiography, are demonstrated; "Other anomalies" have achieved a positive predictive value (PPV) of 100%. Conotruncal and outflow tract anomalies, and right and left heart anomalies have shown a positive predictive value (PPV) higher than 94%. A low PPV score of 88% and 86% could be reported for septal defects and complex CHD, respectively. In all groups, a PPV higher than 97% has been achieved. The lowest sensitivity (SEN) has been shown for right heart anomalies (90%), while the rest of CHD have demonstrated high sensitivity (SEN) and specificity (SPE) scores (Table 5).

**Table 4.** Sensitivity (SEN), specificity (SPE), positive (PPV) and negative predictive value (NPV) for diagnosis in the prenatal echocardiography. Reference: The postnatal echocardiography. N = 177.


In the prenatal echocardiography-MRI group, the right heart anomalies group has shown the lowest sensitivity (SEN) (90%). Prenatal echocardiography has been proved to provide a high accuracy for diagnosis of left heart anomalies, conotruncal and outflow tract anomalies, and "Other anomalies", as SEN lies as high as 97.5%, 95.8%, and 100%, respectively. Similarly, specificity (SPE) has shown a very high score (>97%) for all CHD. The conotruncal and outflow tract anomalies, septal defects, and complex CHD have

reached a specificity (SPE) of 97.9–99%. For left heart anomalies, right heart anomalies, and "Other anomalies", the SPE has been 100%. Right heart anomalies, despite low SEN, have demonstrated a PPV score of 100%. PPV scores for conotruncal and outflow tract anomalies have been 95.8%, and 100% for left heart anomalies, right heart anomalies, and "Other anomalies". Septal defects have possessed the lowest PPV score of 85.7%. Negative predictive value (NPV) has reached a high score (>97%). Septal defects, complex CHD, and "Other anomalies" have reached a negative predictive value (NPV) of 100%.

**Table 5.** Sensitivity (SEN), specificity (SPE), positive (PPV) and negative predictive value (NPV) for diagnosis in the prenatal echocardiography. Reference: Postnatal cardiac. MRI: 108.


The postnatal echocardiography-MRI group has achieved an "almost perfect" strength of agreement in all CHD. A Cohen's Kappa score of 1 has been shown for septal defects and "Other anomalies", 0.98 for conotruncal and outflow tract anomalies, and 0.97 for right heart abnormalities. The lowest rate has been related to complex CHD with Cohen's Kappa (0.94). For all groups, there is a statistical significance for Cohen's Kappa before and after the Bonferroni correction. For conotruncal and outflow tract anomalies, and left heart anomalies, an SEN score of 100% and an SPE higher than 98% could be achieved. For right heart anomalies, an SEN of 95.5% for postnatal echocardiography has been shown. The lowest SEN of 89% has been related to complex CHD, and it has represented the only one with a score lower than 95%. A high SPE (98%) has been achieved in all groups. For right heart anomalies, septal defects, complex CHD, and "Other anomalies", an SPE of 100% has been achieved. For conotruncal and outflow tract anomalies, and left heart anomalies, an SPE higher than 98% has been shown (Table 6). The right heart anomalies, septal defects and "Other anomalies" have achieved a PPV of 100%. Moreover, the conotruncal and outflow tract anomalies, left heart anomalies, septal defects, and "Other anomalies" have shown an NPV as high as 100%. The right heart anomalies with an NPV of 99% indicate that there is 1% chance of a false-negative result.

**Table 6.** Sensitivity, specificity, positive and negative predictive value for the diagnosis in postnatal echocardiography. Reference: Diagnosis in cardiac MRI. N = 108.


The concordance of transposition of great arteries (TGA), double outlet right ventricle (DORV), hypoplastic left heart (HLH), tetralogy of Fallot (ToF), and atrioventricular septum defect (AVSD) in the prenatal-postnatal echocardiography group has been "almost perfect" achieving Cohen's Kappa higher than 0.8 (Table 7). Double outlet right ventricle (DORV) has been the only entity with a Cohen's Kappa score lower than 0.9. In the prenatal echocardiography-MRI group, double outlet right ventricle (DORV) and atrioventricular septal defect (AVSD) have achieved an "almost perfect" classification with a Cohen's Kappa score lower than 0.9. In case of transposition of great arteries (TGA), hypoplastic left heart (HLH), and tetralogy of Fallot (ToF), the concordance rate has been 1. A same pattern is used to evaluate the concordance of transposition of great arteries (TGA), double outlet right ventricle (DORV), hypoplastic left heart (HLH), tetralogy of Fallot (ToF), and atrioventricular septal defect (AVSD) in the postnatal echocardiography-MRI group. For a concordance with a score higher than 0.9 for all and specifically for AVSD, a rate of 1 has been shown. For all groups, a statistical significance could be shown, which was consistent after Bonferroni correction.

**Table 7.** Strength of agreement; comparison among prenatal echocardiography, postnatal echocardiography and cardiac MRI.


Table 8 shows the quality criteria SEN and SPE for TGA, DORV, HLH, ToF, and AVSD in the prenatal-postnatal echocardiography group. Diagnosis of DORV has achieved the lowest sensitivity of 88.9%. The rest have achieved an SEN higher than 94%. Particularly, TGA and ToF have been as high as 95%. HLH has shown an SEN of 97%, while AVSD, 100%. All groups have presented an SPE higher than 97%, while TGA and HLH have had an SPE as high as 100%. Similarly, ToF has shown an SPE of 99.4%. The lowest SPE has corresponded to DORV (98.1%) and AVSD (97.4%). The PPV has been 85% for DORV and AVSD, while the PPV was 95% for ToF and 100% for TGA and HLH. The NPV for all CHD has achieved a score higher than 98%. The lowest NPV has corresponded to DORV (98.7%). For the rest, the NPV has been as high as almost 100%.

**Table 8.** Sensitivity, specificity, positive (PPV) and negative predictive value (NPV) for the diagnosis in the prenatal echocardiography. Reference: Diagnosis in the postnatal echocardiography. N = 177.


Table 9 shows that the quality criteria used for the prenatal echocardiography-postnatal MRI group have represented a high overall SPE (>97%); in particular for AVSD, as high as 98% and for DORV, almost 99%. The rest have had an SPE of 100%. For TGA, HLH, and ToF, no false-positive diagnoses have been made. DORV has had the lowest SEN score (93.8%), whereas for TGA, HLH, ToF, and AVSD, an SEN score of 100% could be achieved. The NPV has been as high as 99% for DORV, representing the lowest score; the scores for the rest of CHD have achieved as high as 100%, reflecting the high reliability of prenatal echocardiography for these CHD. The PPV for AVSD has been 86% and was the lowest. On the contrary, DORV achieved a PPV score of 94%, and an even higher score has been achieved for TGA, HLH, and ToF, reaching 100%.


**Table 9.** Sensitivity, specificity, positive and negative predictive value for diagnosis in the prenatal echocardiography. Reference: Diagnosis in cardiac MRI. N = 108.

In Table 10, only the CHD with a prenatal number of individuals more than fifteen (N > 15) are demonstrated, in order to minimize a possible misleading effect on the results of smaller groups. The exceptions to these are TGA and AVSD in cardiac MRI, whose numbers in prenatal and postnatal echocardiography have been adequate. The postnatal echocardiography has achieved for different entities an SPE of 98%. It has performed for TGA, DORV, and HLH, an SPE of 99% and for TOF and AVSD, 100%. An SEN for DORV and ToF has been as high as 93%, which is lower compared to TGA, HLH, and AVSD (SEN 100%). On the whole, postnatal echocardiography has performed detection higher than 93% of the Individuals with congenital heart diseases. The postnatal echocardiography has shown an NPV higher than 98%. For DORV and ToF, an NPV has been as high as 100%. A PPV of 86% has been shown for TGA; being the lowest. A total of 14% of individuals with suspicion of TGA have had none. It is important to notice that this low PPV could be in association with the small number of individuals. The rest have shown a PPV higher than 93%. The PPV for DORV has been 94% and for HHL, 97%. In the case of ToF and AVSD, PPV has been 100%.

**Table 10.** Sensitivity, specificity, positive and negative predictive value in the postnatal echocardiography. Reference: Diagnosis in cardiac MRI. N = 108.


### **4. Discussion**

Firstly, we wanted to examine agreement between fetal prenatal and postnatal echocardiography. The results show that prenatal echocardiography is a reliable diagnostic tool for conotruncal and outflow tract anomalies, left heart anomalies, and "Other anomalies"; though, with a higher rate of false-positive diagnosis for septal defects and CHD. Similarly, an additional postnatal echocardiography may be reasonable in the case of right heart anomalies, as this is indicated by the low strength of agreement and high false-negative rate in this group. Regarding the group of diagnoses TGA, DORV, HLH, ToF, and AVSD, it shows an almost perfect agreement, with 6% remaining; though, prenatally undetected and an increased risk of false-negative diagnosis for DORV, which reaches 12%. In addition, DORV shows a higher rate of false-positive diagnosis when compared to rest.

Secondly, we studied the agreement between prenatal echocardiography and postnatal MRI. The rate of prenatally undetected fetuses with CHD was, in general, low, especially for left heart anomalies, excepting right heart anomalies with a higher false-negative rate. However, right heart anomalies have demonstrated the highest PPV score in this group, indicating in this manner, a low risk of false-positive diagnosis. The highest risk of falsenegative diagnosis has been shown in the case of right heart anomalies. The number of cases for MRI for TGA and AVSD has been lower than 15, and we have followed a careful interpretation of these results taking into consideration the reduced reliability of these results (Table 7). An increased rate of false-positive diagnosis has been shown for AVSD.

Thirdly, we found that the agreement in the postnatal echocardiography-postnatal MRI group has been almost perfect (Tables 3 and 7). This proves that the echocardiography in postnatal life is a reliable tool. The complex CHD have shown the lowest strength of agreement and the highest rate of false-negative diagnosis reaching 11%.

This retrospective study leads to the conclusion that the prenatal echocardiography is a reliable diagnostic tool for both making and excluding a diagnosis of CHD. A reliable prenatal echocardiography permits an interdisciplinary approach for parental counselling. In that manner, it is possible for paediatricians and paediatric surgeons to involve, and make a plan for the delivery in, a tertiary hospital [4]. This leads to better outcomes [5–7]. This cohort has proven an "almost perfect" strength of agreement for all CHD in three groups. The left heart anomalies have the highest rate of strength of agreement in prenatal echocardiography and in the postnatal diagnostic methods. The second highest rate of concordance has been achieved by the group of conotruncal and outflow tract anomalies and it is in the same line with Gottliebson et al. [8]. Moreover, Gottliebson et al. have demonstrated a high detection rate for the conotruncal anomalies and for complex CHD such as univentricular heart and heterotaxy syndrome.

Regarding the group of diagnoses TGA, DORV, HLH, ToF, and AVSD, HLHS has demonstrated the highest strength of agreement in all groups (Table 7), and it is in agreement with the results of the left heart anomalies. DORV has demonstrated a high risk of increased false-negative diagnosis in prenatal echocardiography, when compared to postnatal echocardiography, and this is in line with Bensemlali et al. [9], who reported that DORV, and specifically the definition of malposition-type of great arteries, is challenging, and in 80% of cases, it can be correctly detected [9,10]; results which agree with the current study, in which in two cases with correct-detected DORV, the position of great arteries have had to be adjusted (TGA-type or Fallot-type). Regarding right heart anomalies and DORV, the high rate of false-negative diagnosis should be taken into consideration and in case of negative diagnosis, eventually a re-evaluation should be considered. The risk of false-negative diagnosis is related to the lack of re-evaluation or treatment [11]. Right heart anomalies similarly have demonstrated a higher false-negative rate in the prenatal echocardiography when this compared to postnatal MRI and in these cases, individuals are at risk of unnecessary treatments or even termination [11].

Mainly due to a high NPV, only a small number of individuals face a false-negative diagnosis. On the contrary, the risk of a false-positive diagnosis mainly for septal defects, complex CHD, and TGA, DORV, HLH, ToF, and AVSD is higher. The prenatal echocardiography has shown a low SEN score for DORV compared with rest. The false-positive rate in this group varied up to 15% for complex CHD. This may lead to unnecessary tests with consequent parents' disquiet or even misleading decision to termination of pregnancy [12]. The CHD with a lower PPV and positive diagnosis requires, therefore, re-evaluation and adequate counselling of parents [13]. In the present study, due to the small number of

individuals (N < 15) with CHD and "Other anomalies", bias is possible. These results show common features with those of the study about precision of echocardiographic diagnosis in early pregnancy from Pike et al. [14] with an SPE of 97.3%, a PPV of 81.2%, and an NPV as high as 100%. A SEN has reached 100% in the study of Pike et al. [14], and it has been higher than in the present study (90–100%). The difference in the current study lies with the evaluation of findings of the whole pregnancy. Pike et al. [14] recommend a follow-up after 20 weeks to increase the reliability of the diagnosis. Bakiler et al. [15] could demonstrate a high SPE of 98% and predictive values higher than 90%; though, with a SEN of 42%, there is a lower reliability for detection than in the present study. In the current study, there have been generally shown higher scores than in the study of Gottliebson et al. [8]; SPE and NPV of 82–100% and SEN and PPV of 83–100%. However, it is important to take into consideration the time period of the study of Gottliebson et al., which took place from 1998 to 2003, and that it is easier to achieve better results in more recent Cohort than in older ones. Therefore, the results of the present study, that took place from 2012 to 2017 are based on better technical equipment.

The study had some limitations. Firstly, in this retrospective approach the results correspond to a collective with a CHD and they cannot be applied to general population. Our collective has been a "high risk population" that had been referred to our referral centre for final diagnosis and delivery to our maternity unit with direct connection to the paediatric cardiology centre. Ascertainment rate, therefore, was high in our study, however there are CHD cases identified prenatally born outside our centre, which have not undergone MRI examination postnatally. Individuals with milder forms of CHD that have been delivered at local hospitals have been excluded from this study because of unavailable data. Secondly, due to retrospective design of present study and focus on prenatally diagnosed CHD, this study is not able to make a reliable statement about false-negative and false-positive diagnosis. Thirdly, intrauterine demise and termination of pregnancy have been excluded from the present study. Fourthly, it is important to mention that the restricted number of Individuals in some cases could potentially lead to bias. The results in these cases should either compared with other studies with a greater number of Individuals or re-evaluated in future larger studies. Fifthly, the expertise of fetal medicine and fetal cardiology specialists in our unit has contributed to the achieved accurate results [14,16,17].

### **5. Conclusions**

The careful observation and evaluation of the results of the current study leads to the conclusion that the prenatal echocardiography represents a reliable method of detection and exclusion of CHD. It is important to mention that the experience of the examiner and eventually a follow-up to adjust the diagnosis in some cases are essential. The greatest reliability could be shown for the anomalies of the left heart and specifically for hypoplastic left heart in the present study. The anomalies of the right heart, on the contrary, have shown the highest rate of false-negative diagnosis. Septal defects have possessed the highest rate of false-positive diagnosis. The greatest challenge for an accurate diagnosis has been related to DORV due to its extreme anatomical variability and frequent association with complex abnormalities. This study shows that cardiac MRI is an additional diagnostic tool for the detailed study of the vascular anatomy and pulmonary supply. In general, it is recommended to re-evaluate the results of the present study in future studies with "low risk collective" and examiners with varying experience. Further studies with a greater number of individuals are essential for comparing and verifying our results. Additionally, it would be interesting to assess in future studies the concordance in relation to gestational week and evaluate the best time to detect progressive CHD. Lastly in case of a prospective study it would be meaningful to include false-positive and false-negative results.

**Author Contributions:** M.M. was responsible for writing original draft preparation. M.M. and R.A.-F. was responsible for writing—review and editing, while the rest of Authors (T.K., I.B., A.W., J.S. and E.W.) for validation and investigation. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted in accordance with the Declaration of Helsinki, and approved by the Institutional Review Board (Ethics approval Committee Justus Giessen University) (protocol number 169/17).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** Data is unavailable due to privacy or ethical restrictions.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

## *Article* **Amniocentesis—When It Is Clear That It Is Not Clear**

**Razvan Ciortea 1, Andrei Mihai Malutan 1, Carmen Elena Bucuri 1, Costin Berceanu 2, Maria Patricia Rada 1, Cristina Mihaela Ormindean 1,\* and Dan Mihu <sup>1</sup>**


**Abstract:** A fetus identified to be at risk for chromosomal abnormalities may benefit from identification of genetic defects through amniocentesis. Although the risks associated with amniocentesis are considered to be minimal, being an invasive procedure it is not completely without complications. Background and Objectives: The current study aims to identify correlations between blood contamination of samples collected during amniocentesis and certain factors dependent on the instruments used (thickness of the needle used to aspirate the fluid), the location of the placenta, and uterine vascularity (more pronounced in multiparous patients). Materials and Methods: The study included 190 patients in the second trimester of pregnancy who met one of the criteria for invasive prenatal diagnosis (age over 35 years, high risk in first trimester screening, history of pregnancies with genetic abnormalities, etc.). The amniotic fluid samples collected from these patients were analyzed in terms of blood contamination of the amniotic fluid aspirated with maternal cells Results: Of the patients in whom the procedure was performed using 21 G size needles, 16 samples (13.33% of the total) were contaminated. None of the samples collected from patients where a 20 G needle was used were contaminated. There was a statistically significant association between the lack of contamination and the use of Doppler ultrasound in multiparous patients with anterior placenta in whom a 21-gauge needle was used for amniocentesis. Conclusions: There is an increased rate of sample contamination (statistically significant) when using 21 G needle sizes and a significant difference in contamination between primiparous and multiparous patients, with contamination being more frequent in multiparous patients. The use of Doppler ultrasonography may benefit the procedure, as the contamination rate was significantly reduced when used during amniocentesis.

**Keywords:** amniocentesis; contamination; technique; needle gauge; doppler; pregnancy; risk

### **1. Introduction**

Amniocentesis is an invasive procedure performed primarily in the second trimester of pregnancy to establish a prenatal diagnosis [1,2]. It was first performed in 1967, and by the mid-1970s it was accepted as a tool used for prenatal diagnosis [3]. It involves obtaining fetal cells from the amniotic fluid by ultrasonographically guided puncture of the amniotic sac. A fetal ultrasound is performed prior to amniocentesis to confirm fetal viability, gestational age, number of fetuses, placental location, amniotic fluid volume, fetal anatomical survey, uterine cavity abnormalities, or presence of fibroids. The procedure is performed to identify any fetal chromosomal anomalies [2,4]. A fetus identified to be at risk for chromosomal abnormalities may benefit from identification of genetic defects through amniocentesis. On the one hand this helps the family in making an informed decision on whether or not to continue the course of pregnancy, preparation for delivery, and neonatal prognosis, and on the other hand it helps the physician in making a diagnosis [1,5].

The indications for amniocentesis may vary depending on the geographical region, medical centre, personal, or family history of the patient or risk factors identified for

**Citation:** Ciortea, R.; Malutan, A.M.; Bucuri, C.E.; Berceanu, C.; Rada, M.P.; Ormindean, C.M.; Mihu, D. Amniocentesis—When It Is Clear That It Is Not Clear. *J. Clin. Med.* **2023**, *12*, 454. https://doi.org/10.3390/ jcm12020454

Academic Editor: Roland Axt-Fliedner

Received: 27 November 2022 Revised: 26 December 2022 Accepted: 1 January 2023 Published: 6 January 2023

**Copyright:** © 2023 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

each individual patient. Patients over 35 years of age, those with a family history of genetic abnormalities, patients with a history of children with different genetic syndromes, pregnancies in which abnormalities are ultrasonographically identified, and pregnancies with positive screening for abnormalities may benefit from this diagnostic procedure [1,2,6].

Since 2015, the number of pregnant patients undergoing this type of invasive investigation has started to progressively and steadily decrease due to the development of non-invasive methods of prenatal testing (NIPT) that involve the identification of fetal DNA from maternal blood. Based on sequencing of cell-free fetal DNA (cff DNA) from maternal plasma (NIPT), professional societies have issued their present positions on noninvasive testing regarding Down syndrome (trisomy 21), in addition to other autosomal aneuploidies (trisomy 18 and 13 [7–10]). All existing statements emphasize that NIPT should not be offered as a diagnostic test for fetal aneuploidy. Many declarations also insist that there is lacking evidence for NIPT to be used as a screening test in a wide population, even though of late there have been some studies that validate good performance in women at average risk. DNA sequenced represents a combination of maternal and fetal cell-free DNA, with the latter actually originating from the placenta, thus rendering NIPT for common autosomal aneuploidies less than fully accurate [11]. A positive result (beckoning a supposed aneuploidy) may be created by factors other than an aneuploid fetal karyotype, therefore including placental mosaicism, a vanishing twin or a maternal tumor; false alarms are foreseeable [12].

NIPT is a much more accurate examination for common autosomal aneuploidies than cFTS. Nevertheless, a positive NIPT result should not be looked upon as a concluding diagnosis, as the placenta contributes to the cff DNA. Women should consequently be advised to have a positive result confirmed through invasive testing, if possible, by amniocentesis [13].

Therefore, amniocentesis and chorionic villus sampling remain the preferred methods of diagnosis [5,14]. Although the risks associated with amniocentesis are considered to be minimal, being an invasive procedure, it is not completely without complications. These include loss of amniotic fluid, which can occur both during and after the test. Most often this disappears within a week, but in rare cases it may continue throughout pregnancy increasing the maternal and fetal risk for infection, fetal compromise (cardiac compression) and preterm birth [2,4,5]. Another complication associated with amniocentesis, described in the literature, is fetal injury during maneuvering, which is why amniotic sac puncture is performed under ultrasonographic guidance. One of the most feared complications is pregnancy loss, but the rate of pregnancy loss after amniocentesis is less than 1% [2,3,5].

A recent study shows that the complication rate associated with amniocentesis may be related to the thickness of the needle used to perform the procedure, the number of punctures performed, the puncture performed transplacentally and, last but not least, the experience of the operator. Literature data show a need for at least 30 procedures performed annually to maintain manuality in performing the diagnostic maneuver, as well as to minimise risks related to the experience of the performer [2,5].

Contamination of amniotic fluid collected during amniocentesis with maternal cells has been identified to be responsible for diagnostic errors as early as 1976 [15]. The maternal cells are thought to be artificially introduced into the amniotic fluid sample, as a result of placental bleeding during amniocentesis [16], but they also can be maternal cells from all tissues that are punctured during the procedure. In 1983 Benn et al. [17] reported a decrease in the contamination rate of samples by removing the first millilitres of amniotic fluid collected, with a 2.5-fold lower contamination rate [15,18].

The current study aims to identify correlations between blood contamination of samples collected during amniocentesis and certain factors dependent on the instruments used (thickness of the needle used to aspirate the fluid), the location of the placenta, and uterine vascularity (more pronounced in multiparous patients).

### **2. Materials and Methods**

The present study is a prospective cohort study, carried out in the period 2016–2021 in the Obstetrics-Gynecology Clinic "Dominic Stanca" Cluj Napoca, Romania. The study included 190 patients in the second trimester of pregnancy (between 16 and 19 weeks of gestation) who met one of the criteria for invasive prenatal diagnosis (age over 35 years, high risk in prenatal screening, history of pregnancies with genetic abnormalities, family history, pregnancies with ultrasonographic anomalies). Patients where the procedure was performed using needles with other sizes (18 or 22 G), patients with increased risk of amniotic fluid contamination (uterine fibroid located in the anterior uterine wall, patients with haematological pathologies, and those with changes in coagulation parameters), and also patients who required multiple punctures for amniotic fluid collection, or the ones in which the procedure needed to be repeated, were not included in the study Figure 1.

**Figure 1.** Inclusion and exclusion criteria for patients.

This prospective cohort research was conducted after obtaining the approval of the Ethics Commission of the University of Medicine and Pharmacy "Iuliu Hat,ieganu" Cluj Napoca (73/22.03.2016) and the informed consent of the patients included in the study group.

The amniotic fluid samples collected from these patients were analyzed in terms of blood contamination of the amniotic fluid aspirated with maternal cells and an attempt was made to make associations according to the following parameters: thickness of the needle used to perform amniocentesis (20–21 G), location of the placenta (at the level of the anterior/posterior uterine wall), uterine vascularization (parity-primipara/multipara).

The analysis was performed on patients who met one of the three criteria, two of the three criteria, or all three criteria. For the latter, additionally data were also analysed according to whether or not Doppler ultrasound was performed.

The technique of performing amniocentesis involves preparation of the patient's tegument and its disinfection with antiseptic solution, followed by puncture of the tegument, under ultrasonographic guidance, Doppler mode was employed only during the insertion of the needle in order to choose a less vascularized area of the uterus or placenta, using needles with sizes between 20–21 G until the amniotic sac is punctured. Once penetrated into the cavity, 15–30 mL of amniotic fluid is aspirated, none of the aspirated fluid is disposed. The collected sample is then sent to the laboratory for fetal karyotype analysis. For establishing the contamination of the collected samples the test method used involves the numerical analysis of short repetitive sequences specific to each individual, called STRs, through QF-PCR (Quantitative Fluorescence—PCR) technique, which is based on multiplex-PCR amplification using fluorescent labelling. DNA extraction was performed

from the processed sample, followed by PCR reaction using the Devyser Compact v3 kit. PCR products were migrated through capillary electrophoresis using the ABI3500 Genetic Analyzer. Interpretation was performed using GeneMapper analysis software. Fetal and maternal STR markers were analysed in parallel and contamination of the amniotic fluid sample with maternal cells was analyzed. Prenatal sampling procedures carry a risk of including maternal cells alongside the envisioned fetal sample. This risk has been empirically calculated as approximately 0.5% in amniotic fluid sampling, 1–2% in chorionic villi sampling, and can possibly be higher, reliant on the method used for invasive sampling. The markers present in the putative fetal DNA sample (derived from cultured or uncultured cells from amniotic fluid, cord blood, chorionic villi, or products of conception) are compared to those from a maternal DNA sample (from a maternal blood specimen). By inspecting these samples in parallel and associating the relative ratios of alleles at each individual marker, it is possible to estimate the presence and level of contaminating maternal cells in the fetal sample. The finding of MCC does not automatically dictate that a repeat invasive sampling procedure be recommended or performed. The mere presence of MCC does not impede diagnostic testing, while the test employed is robust to the level of contamination observed. Common cutoffs include 10% for SNP oligonucleotide analysis (SOMA) and 15% for direct sequencing-based tests [19]. These common cutoffs are the ones we used in our assessments.

We have excluded visual analyzing of samples because data from the literature show that even in macroscopically clear samples maternal cells can be identified. Amniotic fluid samples contamination with maternal cells, depending on the proportion of maternal cells identified, can reduce the accuracy of the result, and may also lead to the impossibility of interpreting the sample.

For amniocentesis, 20 G needles were used in 70 patients who met the inclusion criteria between 2016 and 2018 and 21 G needles in 120 patients who met the inclusion criteria for our study between 2019 and 2021. Amniocentesis procedures were performed either with a 120 mm 20-gauge (G) needle specific for amniocentesis and chorionic villous sampling (Egemen International Amniocentesis Needle, ˙ Izmir, Turkey), or with a 120 mm 21 G amniocentesis needle (Wallace Amniocentesis Needles, Cooper Surgical, Trumbull, CT, USA). Both needles are provided with an adjustable stopper to determine insertion depth, and with a female luer lock connector to collect the amniotic fluid through a syringe.

Statistical processing of the data was performed using the χ<sup>2</sup> (chi-square) statistical test. The significance threshold for the statistical test used was α = 0.05 (5%) and risks were also calculated (RE—risk in the exposed, RN—risk in the unexposed and RR—relative risk or risk ratio).

### **3. Results**

From May 2016 to July 2021, more than 200 diagnostic amniocenteses were performed in our institution, and of these a significant sample, 190 patients, consented for enrollment in the study.

The location of the placenta was in the anterior uterine wall in 55% of the study participants, the rest (45% of the patients) had the placenta located in the posterior uterine wall. In terms of uterine vascularity during pregnancy, a variable directly influenced by the parity of the patients, 38% of those included in the study group were primiparous, and the remaining 62% of patients were multiparous.

Samples were analyzed according to the fulfillment of one to all three criteria (thickness of the needle used to perform amniocentesis, location of the placenta, uterine vascularization).

### *3.1. One of Three Criteria Used*

Of the total 190 patients included in the study, in 120 (63.16%) amniocentesis was performed using a 21 G needle and in 70 (36.84%) the procedure was performed using a 20 G needle. Of the patients in whom the procedure was performed using 21 G size needles, 16 samples (13.33% of the total) were contaminated. None of the samples collected from patients where a 20 G needle was used were contaminated. There was a statistically significant association between the use of a 21 G needle for amniotic sac puncture and contamination of collected samples (*p* = 0.0035) Figure 2.

**Figure 2.** Number of contaminated versus uncontaminated samples according to the size of needle used for amniocentesis.

In the patients included in the study, the contamination of the samples was analyzed taking into account the location of the placenta in the anterior or posterior uterine wall; thus from the group of 190 patients, in 105 of them (55.26%) the placenta was located anteriorly, out of which when performing amniocentesis in a group of 34 patients the sample was collected by punctioning across the placental tissue, and in 85 (44.74%) the placenta was located posteriorly. Of the patients with anterior placenta, in 13 patients (6.84% of the total) the sample collected was contaminated. Amongst patients with posterior placenta, only three patients (1.58% of the total) had contaminated samples. No statistically significant association could be observed between placental position and contamination of samples collected for prenatal diagnosis (*p* = 0.0546, RE = 12.38, RN = 3.53, RR = 3.51).

The last criterion for data analysis considered was the number of previous pregnancies of the patients, thus out of the group of 190 patients, 116 of them (61.05%) were multiparous and 74 (38.95%) were primiparous. Of the multiparous patients, in 14 patients (7.37% of the total) the samples collected showed contamination, and of the primiparous patients, in 2 of them (1.05% of the total) the samples were classified as contaminated. A statistically significant association was observed for sample contamination in multiparous versus primiparous patients (*p* = 0.0456) Figure 3.

#### *3.2. Two of the Three Criteria Used*

A total of 105 patients had placenta located in the anterior uterine wall; in 80 of them (76.19%) amniocentesis was performed using 21 G needles, and in 27 patients the sample was collected by punctioning across the placenta. In the other study participants, a number of 25 patients (23.81%), out of which in 7 cases the sample was collected through placental punctioning, the operation was performed with 20 G needles. Of the patients in whom 21 G needles were used for amniocentesis, in 13 patients (12.38% of all patients with anteriour placenta) the samples were contaminated. In none of the patients in whom amniotic sac puncture was performed using 20 G needles was there contamination of the samples. There was a statistically significant association between the use of a 21 G needle to perform amniocentesis and contamination of samples obtained from patients with previous placenta (*p* = 0.0226) Figure 4.

**Figure 3.** Number of contaminated versus uncontaminated samples according to vascularity during pregnancy (multiparous versus primiparous) regardless of needle gauge used.

**Figure 4.** Sample contamination in patients with anteriorly located placenta correlated with the size of the needle used for sampling.

The correlation between patient parity and the size of the needle used to perform the maneuver was also analyzed. Out of the total 116 multiparous patients, in 74 of them (63.79%) amniocentesis was performed using a 21 G needle and in 42 of them (36.21%) a 20 G needle was used for the maneuver. Among the patients in whom 21 G needle was used for amniotic fluid collection, samples collected from 14 patients (12.07% of all multiparous patients) were contaminated, but there were no contaminated samples in patients in whom 20 G needle was used. There was a statistically significant association between the use of a 21 G needle for amniocentesis in multiparous patients and contamination of amniotic fluid samples (*p* = 0.007) Figure 5.

A correlation was also sought between the use of a 21 G needle and multiparity for amniocentesis. Thus, out of the total of 120 patients in whom a 21 G needle was used for amniocentesis, 74 (61.67%) were multiparous and 46 (38.33%) were primiparous. Of the multiparous patients, 14 patients (11.67% of the total number of patients who used a 21-gauge needle for amniocentesis) had contaminated samples, and of the primiparous patients, only two samples had maternal cell contamination (1.67% of the total number of patients who used a 21-gauge needle for amniocentesis). A statistically significant association was observed between the use of a 21 G needle for amniocentesis and contamination of samples collected from multiparous patients (*p* = 0.0448) Figure 6.

**Figure 5.** Contamination of samples based on multiparity and different gauge of needle used for collection.

**Figure 6.** Sample contamination based on parity with 21 G needle.

In patients who met two of the three criteria established, correlations were also sought between anterior placental location and patient parity, placental location at the posterior uterine wall and multiparity or primiparity, multiparity and anterior or posterior placental location, and primiparity and anterior or posterior placental location, posterior location of the placenta and size of needle used for amniocentesis (20 G or 21 G), use of 21 G needle and location of the placenta in the anterior or posterior uterine wall, and primiparity of the patient and thickness of needle used (20 G or 21 G), with no statistically significant associations between these criteria.

#### *3.3. All of the Three Criteria Used*

The presence of a statistically significant association between the use of a 21 G needle in multiparous patients and the location of the placenta in the anterior uterine wall compared to the use of 20 G needles in primiparous patients and a location of the placenta in the posterior uterine wall was analyzed.

Out of the total of 49 multiparous patients with anterior placenta in whom amniotic sac puncture was performed using 21 G needles, in 23 patients the punction was realised across the placenta, and in 11 of them (22.45% of them) maternal cells were identified present in the collected samples. None of the samples collected from primiparous patients in whom the placenta was identified in the posterior uterine wall and a 20 G needle was used to puncture the amniotic sac showed contamination.

There was a statistically significant association between the use of a 21 G needle for amniocentesis in multiparous patients with anteriorly located placenta and contamination of amniotic puncture samples (*p* = 0.0454) Figure 7.

**Figure 7.** Contamination of samples according to the use of a 21 G needle in multiparous patients with placenta located at the anterior uterine wall compared to the use of 20 G needles in primiparous patients with placenta located at the posterior uterine wall.

For patients who met all three criteria, we also looked for correlations between the size of the needle used (20 G or 21 G), the anterior location of the placenta in multiparous patients, the size of the needle used (20 G or 21 G), the anterior location of the placenta in primiparous patients, the use of 21 G needles, the anterior location of the placenta and the parity of the patients, and the use of 21 G needles. Posterior location of the placenta and parity of the patients and correlations were also sought between needle size used for amniocentesis (21 G), parity (primiparous and multiparous) and anterior or posterior location of the placenta, but no statistically significant associations could be found to correlate these criteria with contamination of amniocentesis samples (*p* > 0.05). Using these conditions it was however observed that 21 G needles were used for all samples collected that showed contamination.

Further research was carried out to find ways in which the amniocentesis procedure could be improved to avoid contamination. Thus, the use of Doppler ultrasound during the maneuver was attempted and the data were then analyzed using the same statistical tests and reference intervals.

We looked for a difference between performing or not performing Doppler ultrasound and contamination of samples obtained from multiparous patients with placenta located in the anterior uterine wall and in whom 21 G needles were used for amniocentesis.

Of the total 49 multiparous patients with anterior placenta in whom 21 G needles were used for amniotic fluid collection, in 39 of them (79.59%) the collection was performed under Doppler ultrasonographic guidance and in 10 of them (20.41%) using only grayscale (2 D) ultrasonography. Of the patients in whom Doppler ultrasonography was used, in two patients (4.08% of the total multiparous patients with anterior placenta in whom 21 needle was used for amniocentesis) maternal cells were identified in the samples collected. Of the patients in whom Doppler ultrasound was not used, nine patients (18.37% of all multiparous patients with previous placenta in whom a 21-gauge needle was used for amniocentesis) showed contamination in the samples collected.

There was a statistically significant association between the lack of contamination and the use of Doppler ultrasound in multiparous patients with anterior placenta in whom a 21-gauge needle was used for amniocentesis (*p* < 0.0001) Figure 8.

**Figure 8.** Use of Doppler ultrasound and reduction in contamination of samples collected using 21 G needles in multiparous patients with anteriorly located placenta.

This method has also been tested to improve amniocentesis performance using 21 G needles in primiparous patients with placenta located at the anterior uterine wall was also tested. Thus, out of a total of 31 primiparous patients with anterior placenta in whom 21 G needles were used for amniocentesis, Doppler ultrasound was performed in 27 of them (87.10%) and not in 4 of them (12.90%). None of the patients in whom Doppler ultrasound was used had contaminated samples, and of the patients in whom Doppler ultrasound was not used, two patients (6.45% of all primiparous patients with previous placenta in whom 21 G needle was used for amniocentesis) had contaminated samples.

There was a statistically significant association between lack of contamination and the use of Doppler ultrasound in primiparous patients with previous placenta in whom a 21-gauge needle was used for amniocentesis (*p* = 0.0146) Figure 9.

Comparative analysis of the group of patients with placenta located in the anterior uterine wall who underwent amniocentesis using 21 G size needles and Doppler ultrasound to perform amniocentesis showed that out of the total 66 patients, 39 of them (59.09%) were multiparous and 27 (40.91%) were primiparous. Among the multiparous patients, two patients (3.03% of those in whom previous placental location was detected in which 21 G needle was used for amniocentesis and Doppler ultrasound was used) had samples contaminated with maternal cells. None of the samples collected from primiparous patients were contaminated. No statistical significance was found (*p* = 0.6).

The reduction in contamination of samples collected during amniocentesis using 21 G needles and Doppler ultrasound was also analyzed for posterior placental location in both primiparous and multiparous patients.

**Figure 9.** Association between lack of contamination of samples by performing amniocentesis using Doppler ultrasound.

Of the total 25 multiparous patients with posterior placenta in whom 21 G needles were used for amniocentesis, 18 of them (72.00%) benefited from the use of Doppler ultrasonography and in seven (28.00%) it was not used. None of the patients in whom Doppler ultrasound was performed had contaminated samples. Of the patients in whom Doppler ultrasound was not performed, in three patients (12.00% of the total of multiparous patients with posterior placenta in whom 21 needles were used for amniocentesis) the samples showed maternal cell contamination.

There was a statistically significant association between the lack of contamination of amniotic fluid samples collected and the use of Doppler ultrasound in multiparous patients with posterior placenta in whom a 21 G needle was used for amniocentesis (*p* = 0.0152) Figure 10.

**Figure 10.** Reduction in contamination of amniocentesis samples using Doppler ultrasound.

The existence of statistically significant associations between the use of 21 G needles, placenta with posterior uterine wall insertion, primiparous patients, the existence of a difference between primiparous and multiparous patients, and the non-use of Doppler ultrasound during the maneuver was also investigated, without identifying the existence of statistically significant correlations.

Data and results from the present study are listed in Table 1.


**Table 1.** Study results. RE—Exposed risk, RN—unexposed risk, RR—relative risk.

### **4. Discussion**

Amniocentesis is currently the most widely used antenatal diagnostic method. Over time, the way of performing the maneuver has been improved in order to reduce the risks associated with amniotic sac puncture, but these risks, although considered minimal, must be known by the couple undergoing such a procedure [3,6]. Amniocentesis is a minor surgical procedure, usually performed in the second trimester of pregnancy to ensure that an optimal number of fetal cells are extracted from the amniotic fluid [1,15]. Amniocentesis is associated with higher rates of successful, clear taps, and lower rates of bloody taps (reduced from 2.4% to 0.8%) when performed under direct ultrasound control with continuous needle tip visualization. Best practice is that ultrasound scanning during the procedure be performed by the person inserting the needle [18]. When we also utilize the Doppler function, we enhance the safety of the maneuver, taking into account the added possibility of visualization of increased flow through some parts of the placenta, which can then be avoided. This is one of the major advantages of the Doppler feature. Additional advantage consists of avoiding a major complication, specifically puncturing the umbilical cord, which is certainly better visualized through Doppler examination, precisely useful in oligohydramnios situations where the cord has less freedom of motion. One of the ways to evaluate the success of this procedure is by assessing the postprocedural abortion rate which reaches 0.6% for ultrasound-guided maneuvers. Another criterion used to assess the success of the procedure is the quality of the specimen extracted, as contamination with maternal cells can lead to diagnostic errors [15,16].

Most studies in the literature have focused on how to perform amniocentesis in order to reduce the risks associated with the maneuver and very few studies have focused on the association between the size of needles used to perform amniotic sac puncture and the occurrence of complications. In this direction, the aim of our study was to identify the increased risk of specimen contamination taking into account both maternal factors (placental location, parity) and extrinsic factors (instruments used to perform the procedure, more specifically the size of needles used to perform amniocentesis). At the same time, a method to reduce the risk of contamination of the samples by using Doppler examination during the procedure was analyzed. Current research results show a significant association between maternal cell contamination of amniotic fluid samples collected during amniocentesis and the use of fine needles of smaller diameter (21 G). This finding is supported by two studies: the first one was published by Peter A. Benn and Lillian Y.F. Hsu [17] and identified a higher rate of contamination of samples obtained after puncture of the amniotic sac using needles larger than 20 G, and the second one was conducted by Apostolos P., Athanasiadis et al. [20] who demonstrate that the use of smaller needle diameters (22 G) causes greater trauma to tissues during amniocentesis, primarily through the longer duration of the amniotic fluid aspiration process. At the same time, the present study identified a higher risk of contamination associated with the use of small puncture needles (21 G), regardless of the patient's parity (primiparous or multiparous). The use of larger gauge needles (20 G) shortens the time to perform the procedure, but is associated with a higher rate of amniotic fluid leakage postprocedurally [21]. This hypothesis is demonstrated through an experimental study design by Devlieger R et al. who tested amniotic fluid leakage in patients undergoing Caesarean section at term [22].

There are conflicting data in the literature regarding the increased risk of contamination associated with placental location at the uterine cavity walls. The findings of the study led by Hockstein et al. [23] show a lack of association between the risk of contamination of collected amniotic fluid samples and placental location in the uterine cavity. In contrast, research by Nub et al. [16] found a significantly higher rate of maternal cells in samples collected from patients with placenta located in the anterior uterine wall. This hypothesis is also supported by the studies performed by Pergamentet et al. [23] and Giorlandino C et al. [24]. They concluded that the frequency of contaminated samples was higher in three different situations: when the collection was performed transplacentally, when more punctures of the amniotic sac were required for collection, and when the maneuver was performed by a less experienced physician [25]. The present study shows a significant association between the amniotic fluid contamination and the anterior location of the placenta, thus an increased rate of contamination in association with transplacental amniotic fluid collection. The risk of contamination is further increased by two factors: when the placenta is located on the anterior wall and when the amniotic fluid extraction is performed using small (21 G) needles. Apostolos P. Athanasiadis et al. [20] and Uludag S [26] supported the hypothesis in which the use of 20 G needles for amniocentesis reduces intrauterine bleeding and therefore amniotic fluid contamination.

Another risk factor for samples contamination is the parity of patients, showing a significant association between multiparity and frequency of contamination.

In this research the usefulness of Doppler ultrasound during amniocentesis was investigated, demonstrating a reduction in contamination through its use. This hypothesis is in addition to Jennifer Weida et al. [15] and Benn et al. [17] studies that showed a significant 2.5-fold reduction in the frequency of contamination after removal of the first 1–2 mL of aspirated amniotic fluid.

One of the strengths of the present study is the large number of patients included in the study, with 190 patients undergoing invasive genetic testing. Also, the analysis of the contaminated samples with maternal cells taking into account all the criteria, considering the fact that the literature data analyses the factors separately, is another strength of the current research. Data from the present study show a significant reduction in the risk of contamination while using Doppler ultrasound for identifying uteroplacental circulation during amniocentesis, regardless of the size of the needle used to perform the puncture.

The present study has some limitations, among which we mention the impossibility to quantify the degree of contamination of the samples in order to assess the risk of false positive or negative results. Also, the existence of a correlation between the size of the needle used to perform the maneuver and the number of maternal cells identified in the sample analyzed was not possible.

### **5. Conclusions**

There is an increased rate of sample contamination (statistically significant) when using 21 G needle sizes and a significant difference in contamination between primiparous and multiparous patients, with contamination being more frequent in multiparous patients. There was a significantly higher rate of contamination of samples associated with amniotic sac puncture using 21 G needles, as opposed to the contamination rate observed when larger gauge needles (20 G) were used. The contamination rate was significantly higher in multiparous patients when 21 G needles were used. In multiparous patients, contamination of collected samples is more frequent when 21 G needles are used for maneuvering.

A higher rate of contamination is observed in multiparous patients with a previously located placenta in whom 21 G needles were used to collect amniotic fluid for prenatal diagnosis.

The use of Doppler ultrasonography clearly benefits the procedure, as the contamination rate was significantly reduced when used during amniocentesis regardless of the factors favoring contamination present. Thus, it is our strong belief that only Doppler ultrasound should accompany amniocentesis procedures. This hypothesis should open opportunities for future research into the usefulness of Doppler studies during invasive diagnostic maneuvers.

**Author Contributions:** Conceptualization, R.C. and D.M.; methodology, A.M.M. and C.B.; software, C.M.O.; validation, R.C., C.E.B. and C.M.O.; formal analysis, M.P.R.; investigation, R.C.; resources, R.C.; data curation, C.E.B.; writing—original draft preparation, C.M.O.; writing—review and editing, C.M.O.; visualization, D.M. and C.B.; supervision, R.C.; project administration, R.C.; funding acquisition, R.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

**Institutional Review Board Statement:** The study was conducted in accordance with the Declaration of Helsinki, and approved by Ethics Commission of the University of Medicine and Pharmacy "Iuliu Hat,ieganu" Cluj Napoca (73/22.03.2016).

**Informed Consent Statement:** Informed consent was obtained from all subjects involved in the study.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

### **References**


**Disclaimer/Publisher's Note:** The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.
