http://informahealthcare.com/jmf ISSN: 1476-7058 (print), 1476-4954 (electronic) J Matern Fetal Neonatal Med, 2014; 27(18): 1839–1844 ! 2014 Informa UK Ltd. DOI: 10.3109/14767058.2014.882306

Diagnostic accuracy of non-invasive fetal RhD genotyping using cell-free fetal DNA: a meta analysis Yu-juan Zhu*, Ying-ru Zheng*, Li Li, Hao Zhou, Xi Liao, Jian-xin Guo, and Ping Yi

J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Fudan University on 05/12/15 For personal use only.

Department of Obstetrics and Gynecology, Research Institute of Surgery, Daping Hospital, Third Military Medical University, Chongqing, PR China

Abstract

Keywords

Objective: To determine the diagnostic accuracy, validity, current limitations of, and possible solutions to, fetal RhD genotyping from maternal blood based on existing studies written in English. Methods: A literature search was conducted that described fetal RhD determination from maternal blood. The number of samples tested, fetal RhD genotype, the source of cell-free fetal DNA, gestational age and fetal Rh type were examined in each study to calculate the accuracy, sensitivity and specificity of fetal RhD genotyping. Results: Forty-one publications, which included 11 129 samples with non-invasive Rh genotyping of cell-free fetal DNA from maternal blood, were selected. After the exclusion of 352 inconclusive samples, the overall diagnostic accuracy was 98.5% (10 611/10 777), and sensitivity and specificity were 99% and 98%, respectively. First trimester diagnosis showed an accuracy of 99%, higher than second and third trimester diagnosis. Thirty studies reported a 100% diagnostic accuracy of fetal RhD genotyping. Conclusion: Non-invasive fetal RhD genotyping from maternal blood has high accuracy, sensitivity and specificity. Methods reducing false results have been explored and applied in research. These achievements indicate that this technique will be widely used in routine clinical care.

Fetal RhD, maternal blood, prenatal diagnosis

Introduction In the Rh blood group system, antigen D is the strongest of the six antigens (C, c; D, d; E, e), and the anti-D isoantibody is the major cause of hemolytic disease in newborns, and with transfusion reactions [1]. RhD-negative people account for 3–5% of the African American population and 15–17% of the Caucasian population in Europe [2,3]; while the incidence is rare in the eastern Asian population. In China, RhD-negative people are mainly distributed in minority areas. For RhD-negative pregnant women, hemolytic disease in a newborn generally happens if the fetus is RhD-positive. Detection of the fetal RhD status of RhD-negative mothers is essential to avoid the occurrence of hemolytic disease. The routine use of antibody screening and anti-D immunoglobulin prophylaxis of RhD-negative mothers has reduced the incidence of hemolytic disease, but may result in over treatment. In 1997, cell-free fetal DNA (cffDNA) circulating in maternal plasma and serum was first detected [4]. Such a discovery made non-invasive, risk-free prenatal diagnosis possible. cffDNA can be detected after 5 weeks of pregnancy [5–8]. Detection of cffDNA is a powerful tool for the diagnosis *Both authors contributed equally to this work. Address for correspondence: Ping Yi, Department of Obstetrics and Gynecology, Research Institute of Surgery, Daping Hospital, Third Military Medical University, 10 Changjiangzhilu, Daping, Yuzhong District, Chongqing 400042, PR China. Tel: +86-23-6875-7925. Fax: +86-23-6881-3806. E-mail: [email protected]

History Received 15 October 2013 Revised 13 December 2013 Accepted 8 January 2014 Published online 10 February 2014

of diseases such as chromosomal aneuploidies [9], single gene disorders [10], fetal RhD determination, and gender determination [11,12]. Fetal DNA enters the maternal blood circulation in two forms: (1) as cell-free fetal DNA; and (2) as fetalderived cells. It is important for RhD-negative pregnant mothers that fetal DNA in their own blood circulation is detected. Exons 4, 5, 7 and 10 can be used to detect RhD-positive samples by real time polymerase chain reaction (PCR). PCR can be used as a molecular biological method for analyzing trace amounts of fetal DNA, but it can produce false positive and false negative results. Fetal RhD genotyping has been applied clinically for RhD-negative pregnant women in some European countries (such as Denmark and the Netherlands) [13,14], but there are limitations to its widespread clinical application. In this study, a review of the literature was undertaken of studies that used a number of maternal blood detection experiments to assess the diagnostic accuracy, validity, current limitations, and possible solutions to problems with fetal RhD genotyping from peripheral maternal blood samples to help reduce the occurrence of hemolytic disease in newborns.

Materials and methods Search strategy A literature search was conducted by combining words from the titles of articles and the article itself. The search was not

1840

Y.-j. Zhu et al.

J Matern Fetal Neonatal Med, 2014; 27(18): 1839–1844

J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Fudan University on 05/12/15 For personal use only.

Table 1. Summary of all identified articles. Author

Year

No. women

No. samples tested

Total correct

Total % correct

total tested

% total correct

TadejaDovc Drnovsek Clausen FB Macher HC Macher HC Gunel Scheffer PG Bombard AT Amaral DR Akolekar R Tynan JA Sedrak M Cardo L Mohammed N Hyland CA Wang XD Atamaniuk J Atamaniuk J Atamaniuk J Grill S Mu¨ller SP Kimura M Finning K Minon JM Rouillac-Le Al-Yatama MK Machado IN Zhou L Clausen FB Clausen FB Brojer E Hromadnikova I Hromadnikova I Hromadnikova I Hromadnikova I Di Simone N Di Simone N Kirstin F I. Randen Sashi C Legler TJ K.M. Finning Xiao Yan Zhong J. Zhang Farideh Z Farideh Z Y.M. Dennis Lo

2013 2012 2011 2011 2011 2011 2011 2011 2011 2011 2011 2010 2010 2009 2009 2009 2009 2009 2009 2008 2008 2008 2008 2007 2007 2006 2005 2005 2005 2005 2005 2005 2005 2005 2004 2004 2004 2003 2003 2002 2002 2001 2000 1999 1999 1998

153 2312 134 1012 40 168 441 88 591 150 90 111 21 140 78 46 17 46 179 1113 13 1997 545 300 54 81 98 38 21 255 21 29 29 45 23 24 283 114 28 27 137 34 58 20 20 57

153 2238 134 1012 40 161 406 88 502 148 90 100 21 135 75 46 17 46 179 1022 12 1805 545 310 54 75 98 38 20 321 21 29 29 45 23 24 226 114 26 27 137 34 58 20 16 57

153 2230 131 1005 37 161 399 88 496 148 90 95 17 135 70 40 12 40 172 1014 12 1788 545 308 49 73 92 38 20 291 21 29 26 45 21 24 223 114 21 26 137 31 57 16 16 55

100.0 99.6 97.8 99.3 92.5 100.0 98.3 100.0 98.8 100.0 100.0 95.0 81.0 100.0 93.3 87.0 70.6 87.0 96.1 99.2 100.0 99.1 100.0 99.4 90.7 97.3 93.9 100.0 100.0 90.7 100.0 100.0 89.7 100.0 91.3 100.0 98.7 100.0 80.8 96.3 100.0 91.2 98.3 80.0 100.0 96.5

153 2312 134 1012 40 168 441 88 586 150 90 111 21 140 78 46 17 46 179 1022 13 1869 545 310 54 81 98 38 21 321 21 29 29 45 23 24 283 114 28 27 137 34 58 20 16 57

100.0 96.5 97.8 99.3 92.5 95.8 90.5 100.0 84.6 98.7 100.0 85.6 81.0 96.4 89.7 87.0 70.6 87.0 96.1 99.2 92.3 95.7 100.0 99.4 90.7 90.1 93.9 100.0 95.2 90.7 100.0 100.0 89.7 100.0 91.3 100.0 78.8 100.0 75.0 96.3 100.0 91.2 98.3 80.0 100.0 96.5

limited by press or publication dates. English literature resources were from the PMC, Highwire database from 1966 to 2013. Documents were retrieved by using keywords separately and together. Keywords were: prenatal diagnosis, fetal Rh, fetal DNA, fetal DNA in maternal blood, and maternal plasma of maternal serum alloimmunization.

Data collection The following article data were input into an Excel sheet for analysis: first author’s name, year, name of journal, number of samples, gestational age, RhD blood group determination, source of fetal DNA and newborn (or fetus) RhD blood group (Table 1).

Literature selection A study was considered eligible for further examination if it met the following inclusion criteria: written in English; the topic was fetal RhD blood group identification using maternal whole blood (or serum, plasma); determination of fetus (or newborn) RhD blood type was provided; and the population was RhD-negative pregnant women. Exclusion criteria were: articles that were summaries or unpublished; and cases that reported on less than 10 pregnant women. If samples were divided into different groups in a report, the groups with fewer than 10 pregnant women were also excluded.

Data processing and analysis Data were analyzed using Meta-DiSc version 1.4. (http:// www.hrc.es/investigacion/metadisc_en.htm) The randomeffects model was calculated with a 95% confidence interval. The overall diagnostic accuracy, sensitivity and specificity for fetal RhD determination from maternal blood were calculated for all studies that met inclusion criteria. Diagnostic accuracy based on trimester was analyzed. Accuracy, ROC curve, sensitivity, specificity, and positive and negative predictive ratios were calculated.

Diagnostic accuracy of non-invasive fetal RhD genotyping

DOI: 10.3109/14767058.2014.882306

Results Analysis of diagnostic accuracy A total of 37 articles, which included 46 protocols and 11 129 samples, were examined in the current study, and 352 inconclusive samples were excluded. Real time PCR was Table 2. Overall and adjusted accuracy of RhD genotyping.

Overall accuracy Accuracy after

n Correct

n Total

% Correct

10 611 10 611

11 129 10 777

95.3 98.5

J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Fudan University on 05/12/15 For personal use only.

Table 3. The positive and negative predictive ratio.

Positive predictive ratio Negative predictive ratio

T*

F*

% Correct

6864 3747

89 77

98.7 98.0

Table 4. Diagnostic accuracy by gestational age (GA) at the time of sampling.

1st 2nd 3rd

used, and five researchers used more than one method in their study, including: taqman RT-PCR, light cyc, thermo and mass spectrum [15–19]. In the studies, similar levels of cffDNA were detected in both maternal plasma and serum [11,20], but plasma was used in most studies; and especially after 2003, only plasma was used. The remaining 10 777 samples were included in the study, of which 10 611 samples were correctly diagnosed, with an overall diagnostic accuracy rate of 95.3% (10 611/11 129) (Table 2). After excluding substandard samples, the accuracy rate was 98.5% (10 611/10 777) (Table 2). Thirty studies reported a diagnostic accuracy of 100% after exclusion (Table 1). The false positive rate was 1.3%, and the false negative rate was 2% (Table 3). A total of 6670 samples that clearly identified the gestational age in articles were collected. Samples were classified based on gestational weeks. The results showed that the diagnostic accuracy in the first trimester was 99%, higher than the second and third trimesters (Table 4). Heterogeneity analysis

*T ¼ True results; F ¼ False results.

Trimester

1841

n Correct

n Total

% Correct

882 3282 2418

898 3322 2450

99.0 98.3 96.4

The ROC curve provided further evidence of high diagnostic accuracy (Figure 1), with a sensitivity of 99% and a specificity of 98% (Figures 2 and 3).

Discussion The results of this meta-analysis showed that the experimental accuracy of non-invasive prenatal diagnosis of RhD blood

Figure 1. Meta-analysis ROC curve. The ROC curve tests the accuracy of the integrated representation. The horizontal coordinate is 1-specificity, the vertical coordinate is sensitivity.

Y.-j. Zhu et al.

J Matern Fetal Neonatal Med, 2014; 27(18): 1839–1844

J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Fudan University on 05/12/15 For personal use only.

1842

Figure 2. Meta-analysis sensitivity. Graphic representation of sensitivity results from meta-analysis of all samples that met inclusion criteria.

group using cffDNA from maternal blood samples reached 98.5%, and also had a high sensitivity and specificity. In most studies, authors used maternal plasma as the source of cffDNA, especially after 2003; although previous studies showed no difference in results between plasma and serum, being 96.5 and 96.1%, respectively [11,20]. Although fetal DNA was detected from the sixth week of pregnancy, and across all gestational ages, both the results from the current study and the results of the Geifman–Holtzman study

[11] showed the highest diagnosis accuracy in the first trimester. Although there is a high accuracy, sensitivity and specificity with this method, false results must be taken into account. The consequences of false positive and false negative results are important. The fetus of an Rh-negative pregnant woman who is misdiagnosed as Rh-positive will be subject to unnecessary Rh immunoglobulin prophylaxis. Conversely, the fetus of a Rh-negative pregnant woman who is misdiagnosed

Diagnostic accuracy of non-invasive fetal RhD genotyping

1843

J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Fudan University on 05/12/15 For personal use only.

DOI: 10.3109/14767058.2014.882306

Figure 3. Meta-analysis specificity. Graphic representation of specificity results from meta-analysis of all samples that met inclusion criteria.

as Rh-negative will not have the necessary treatment. It is therefore very important to explore the causes and solutions for false positive and false negative results. Gene variation, mutation and deletion have been reported as reasons for false positive results. A study by Daniels G. reported that, in a Caucasian population, the RhD-negative phenotype always occurred as a result of RHD gene deletion; and in Africans, there are three Rh haplotypes which produce no D antigen: RHD deletion, RHD pseudo-gene

and RHD-CE-Ds [3,21]. Taking the RHD pseudo-gene, for example, the RHD gene sequence shows little difference to the normal RHD gene, but it will not produce D antigen. Therefore, PCR results will be positive but serology results will be negative. Methods to overcome this problem have been described previously, they include: using the maternal genotype as an internal control for PCR with fetal genotyping, and designing primers for the RHD gene in the studied population.

J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Fudan University on 05/12/15 For personal use only.

1844

Y.-j. Zhu et al.

Reasons for false negative results were mainly because of failing to detect target fetal DNA, including: insufficient DNA or absent fetal DNA in maternal samples, and insensitive experimental methods or conditions. To overcome these problems, researchers used small amounts of different DNA as the internal control for the presence of fetal DNA; for example, tracer mouse DNA and Escherichia coli plasmid [22–24]. It is also important to confirm the origin of amplified fetal DNA in plasma. To improve detection of fetal DNA, timely removal of cells by centrifugation and expedited processing of the sample is necessary. Paternal polymorphic markers can be used as a positive control, especially with negative results. Fetal apparent genetic markers can also be used to verify a fetal origin of amplified DNA by detecting the same gene methylation status between mother and fetus. The first fetal apparent genetic marker was the mammary serine protease inhibitor (Maspin) gene, also known as SERPINB5. Chim et al. reported low methylation in placenta tissue but high methylation in maternal blood cells; therefore, placental SERPINB5 gene can be detected by methylation-specific PCR [25]. Another genetic marker is the RASSF1A gene, a tumor suppressor gene. Chiu et al. reported that RASSF1A showed high methylation in placental tissue but low methylation in maternal blood cells, and hyper-methylated RASSF1A sequences in maternal circulation were derived from the fetus [26,27]. The researchers reported that hypomethylated RASSF1A sequences derived from maternal blood cells could be removed from maternal plasma with the use of methylation-sensitive restriction enzyme digestion. Chan et al. further demonstrated that digestion-resistant RASSF1A DNA can be used as a positive control for noninvasive fetal RhD genotyping [28]. Overall, fetal DNA markers provide reliability for the application of fetal DNA in non-invasive prenatal diagnosis. The results of the current study suggest a high accuracy, sensitivity and specificity for non-invasive fetal RhD genotyping from maternal blood. Large-scale fetal RhD genotyping using maternal blood should be used in clinical practice in the near future.

Acknowledgements The authors acknowledge the support and contribution of all the authors.

Declaration of interest The authors declare that they have no conflicts of interest.

References 1. Chen JC, Lin TM, Chen YL, et al. RHD 1227A is an important genetic marker for RhD(el) individuals. Am J Clin Pathol 2004; 122:193–8. 2. Avent ND. RHD genotyping from maternal plasma: guidelines and technical challenges. Methods Mol Biol 2008;444:185–201. 3. Daniels G. The molecular genetics of blood group polymorphism. Transpl Immunol 2005;14:143–53. 4. Lo YM, Corbetta N, Chamberlain PF, et al. Presence of fetal DNA in maternal plasma and serum. Lancet 1997;350:485–7.

J Matern Fetal Neonatal Med, 2014; 27(18): 1839–1844

5. Farina A, Leshane ES, Romero R, et al. High levels of fetal cell-free DNA in maternal serum: a risk factor for spontaneous preterm delivery. Am J Obstet Gynecol 2005;193:421–5. 6. Galbiati S, Smid M, Gambini D, et al. Fetal DNA detection in maternal plasma throughout gestation. Hum Genet 2005;117: 243–8. 7. Li Y, Holzgreve W, Din E, et al. Cell-free DNA in maternal plasma: is it all a question of size? Ann N Y Acad Sci 2006;1075:81–7. 8. Liu FM, Wang XY, Feng X, et al. Feasibility study of using fetal DNA in maternal plasma for non-invasive prenatal diagnosis. Acta Obstet Gynecol Scand 2007;86:535–41. 9. Puszyk WM, Crea F, Old RW. Noninvasive prenatal diagnosis of aneuploidy using cell-free nucleic acids in maternal blood: promises and unanswered questions. Prenat Diagn 2008;28:1–6. 10. Guo QW, Zhou YL. Noninvasive prenatal diagnosis of single gene disorders through cell-free fetal DNA in maternal blood. Zhonghua Yi Xue Yi Chuan Xue Za Zhi 2009;26:410–13. 11. Geifman-Holtzman O, Grotegut CA, Gaughan JP. Diagnostic accuracy of noninvasive fetal Rh genotyping from maternal blood – a meta-analysis. Am J Obstet Gynecol 2006;195:1163–73. 12. Randen I, Hauge R, Kjeldsen-Kragh J, Fagerhol MK. Prenatal genotyping of RHD and SRY using maternal blood. Vox Sang 2003; 85:300–6. 13. Scheffer PG, Van Der Schoot CE, Page-Christiaens GC, De Haas M. Noninvasive fetal blood group genotyping of rhesus D, c, E and of K in alloimmunised pregnant women: evaluation of a 7-year clinical experience. BJOG 2011;118:1340–8. 14. Clausen FB, Jakobsen TR, Rieneck K, et al. Pre-analytical conditions in non-invasive prenatal testing of cell-free fetal RHD. PLoS One 2013;8:e76990. 15. Macher HC, Noguerol P, Medrano-Campillo P, et al. Standardization non-invasive fetal RHD and SRY determination into clinical routine using a new multiplex RT-PCR assay for fetal cell-free DNA in pregnant women plasma: results in clinical benefits and cost saving. Clin Chim Acta 2011;413:490–4. 16. Bombard AT, Akolekar R, Farkas DH, et al. Fetal RHD genotype detection from circulating cell-free fetal DNA in maternal plasma in non-sensitized RhD negative women. Prenat Diagn 2011;31: 802–8. 17. Tynan JA, Angkachatchai V, Ehrich M, et al. Multiplexed analysis of circulating cell-free fetal nucleic acids for noninvasive prenatal diagnostic RHD testing. Am J Obstet Gynecol 2011;204:251 e1–6. 18. Atamaniuk J, Stuhlmeier KM, Karimi A, Mueller MM. Comparison of PCR methods for detecting fetal RhDin maternal plasma. J Clin Lab Anal 2009;23:24–8. 19. Clausen FB, Krog GR, Rieneck K, et al. Reliable test for prenatal prediction of fetal RhD type using maternal plasma from RhD negative women. Prenat Diagn 2005;25:1040–4. 20. Lo YM, Tein MS, Lau TK, et al. Quantitative analysis of fetal DNA in maternal plasma and serum: implications for noninvasive prenatal diagnosis. Am J Hum Genet 1998;62:768–75. 21. Wagner FF, Flegel WA. RHD gene deletion occurred in the Rhesus box. Blood 2000;95:3662–8. 22. Gautier E, Benachi A, Giovangrandi Y, et al. RhD genotyping by maternal serum analysis: a two-year experience. Am J Obstet Gynecol 2005;192:666–9. 23. Legler TJ, Lynen R, Maas JH, et al. Prediction of fetal Rh D and Rh CcEe phenotype from maternal plasma with real-time polymerase chain reaction. Transfus Apher Sci 2002;27:217–23. 24. Costa JM, Giovangrandi Y, Ernault P, et al. Fetal RHD genotyping in maternal serum during the first trimester of pregnancy. Br J Haematol 2002;119:255–60. 25. Chim SS, Tong YK, Chiu RW, et al. Detection of the placental epigenetic signature of the maspin gene in maternal plasma. Proc Natl Acad Sci U S A 2005;102:14753–8. 26. Chiu RW, Chim SS, Wong IH, et al. Hypermethylation of RASSF1A in human and rhesus placentas. Am J Pathol 2007; 170:941–50. 27. Chiu RW, Lo YM. Non-invasive prenatal diagnosis by fetal nucleic acid analysis in maternal plasma: the coming of age. Semin Fetal Neonatal Med 2011;16:88–93. 28. Chan KC, Ding C, Gerovassili A, et al. Hypermethylated RASSF1A in maternal plasma: a universal fetal DNA marker that improves the reliability of noninvasive prenatal diagnosis. Clin Chem 2006;52: 2211–18.

Diagnostic accuracy of non-invasive fetal RhD genotyping using cell-free fetal DNA: a meta analysis.

To determine the diagnostic accuracy, validity, current limitations of, and possible solutions to, fetal RhD genotyping from maternal blood based on e...
2MB Sizes 0 Downloads 0 Views