www.AJOG.org Letters to the Editors

sample size.6 The authors concluded that the outcome of thistrial did not answer which treatment, amnioreduction or laser,was the superior therapeutic option. Despite the inclusionof the second trial, the overall rate of death approached statis-tical significance (95% CI, 0.65e1.01) in the most recentCochrane review. Additionally, another metaanalysis thatincluded 4 studies and >600 patients has been previouslypublished in the same journal as the current committeeguideline.7 The authors found that fetuses undergoing laserweremore likely to survive (RR, 2.04; CI, 1.52e2.76), less likelyto die in the neonatal period (RR, 0.20; CI, 0.15e0.40), and lesslikely to experience neurologic morbidity (RR, 0.20; CI,0.12e0.33). Finally, a recent metaanalysis of cerebral injuryfollowing laser therapy vs amnioreduction was undertaken byvan Klink et al.8 Four studies involving 357 children in theamnioreduction group and 269 children in the laser groupwere analyzed. Cerebral injury in liveborn infants in theamnioreduction group was 7.69-fold increased over that inliveborns in the laser group (95% CI, 2.78e20.0). Whenneonatal deaths were excluded from the analysis, infants frompregnancies treated with amnioreduction were still noted toexhibit a marked increase in neurologic injury (RR, 3.23; 95%CI, 1.45e7.14) when compared to the laser group.

It would therefore appear that multiple metaanalysessupport the beneficial role of laser treatment for advancedcases of twin-twin transfusion. -

Kenneth J. Moise Jr, MDDepartment of Obstetrics, Gynecology, and Reproductive MedicineDivision of Maternal-Fetal MedicineUniversity of Texas School of Medicine at HoustonHouston, TXKenneth.J.Moise@uth.tmc.edu

Michael W. Bebbington, MDAnthony Johnson, DODepartment of Obstetrics, Gynecology, and Reproductive MedicineDivision of Maternal-Fetal MedicineUniversity of Texas School of Medicine at HoustonHouston, TXTexas Fetal CenterChildren’s Memorial Hermann HospitalHouston, TX

Martin Walker, MDDepartment of Obstetrics and GynecologySection of Fetal TherapyEvergreen HospitalSeattle, WA

Mark Johnson, MDThe Center for Fetal Diagnosis and TreatmentThe Children’s Hospital of PhiladelphiaPhiladelphia, PA

The authors report no conflict of interest.

REFERENCES

1. Society for Maternal-Fetal Medicine, Simpson LL. Twin-twin trans-fusion syndrome. Am J Obstet Gynecol 2013;208:3-18.

2. Roberts D, Gates S, Kilby M, Neilson JP. Interventions for twin-twintransfusion syndrome: a Cochrane review. Ultrasound Obstet Gynecol2008;31:701-11.3. Roberts D, Neilson JP, Kilby M, Gates S. Interventions for the treat-ment of twin-twin transfusion syndrome [review]. Cochrane DatabaseSyst Rev 2008;1:CD002073.4. Senat MV, Deprest J, Boulvain M, Paupe A, Winer N, Ville Y. Endo-scopic laser surgery versus serial amnioreduction for severe twin-to-twintransfusion syndrome. N Engl J Med 2004;351:136-44.5. Crombleholme TM, Shera D, Lee H, et al. A prospective, randomized,multicenter trial of amnioreduction vs selective fetoscopic laser photo-coagulation for the treatment of severe twin-twin transfusion syndrome.Am J Obstet Gynecol 2007;197:396.e1-9.6. Johnson A, Moise KJ Jr. A randomized trial for the treatment of TTTS, toolittle to answer the question. Am J Obstet Gynecol 2008;198:608-9; authorreply 9.7. Rossi AC, D’Addario V. Laser therapy and serial amnioreduction astreatment for twin-twin transfusion syndrome: a metaanalysis and reviewof literature. Am J Obstet Gynecol 2008;198:147-52.8. van Klink JM, Koopman HM, van Zwet EW, Oepkes D, Walther FJ,Lopriore E. Cerebral injury and neurodevelopmental impairmentafter amnioreduction versus laser surgery in twin-twin transfusionsyndrome: a systematic review and meta-analysis. Fetal Diagn Ther2013;33:81-9.

ª 2013 Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajog.2013.03.020

REPLY

We thank Dr Moise et al for their interest and congratulationsfor our Society for Maternal-Fetal Medicine (SMFM) clinicalguideline on twin-twin transfusion syndrome. Dr Moise et alhave a concern that not all data in the metaanalyses on thistopic were reported in the SMFM clinical guideline. Weprovided details from both of the 2 randomized trials ontwin-twin transfusion syndrome treated by laser, includingtables representing the primary and several secondary out-comes.1,2 We also reported the results of the metaanalysisbased on these 2 randomized controlled trials (RCTs). Webelieve that the recommendations provided in the SMFMclinical guideline primarily reflect data available from level-Isources, and are consistent with the data available from thepublished RCTs.

Regarding the metaanalyses, Roberts et al3 reported as themain result a nonsignificantly different rate of overall death(48% vs 59%; relative risk, 0.81; 95% confidence interval,0.65e1.01) in the laser vs amnioreduction groups, respec-tively. Of course, there are many other outcomes reported,some of which are significantly better for laser (as Dr Moiseet al point out) and some of which instead found no dif-ferences between groups (eg, babies alive at 6 months whoneeded treatment for major neurologic abnormality, 4% vs7%; relative risk, 0.58; 95% confidence interval, 0.18e1.86).The metaanalysis by Rossi and D’Addario4 included only 1of the RCTs2 and 9 nonrandomized series. Similarly, themetaanalysis by van Klink et al5 also included data mostlyfrom non-RCT sources. These do not represent level-I datametaanalyses and are superseded by the available level-Idata.

AUGUST 2013 American Journal of Obstetrics & Gynecology 159

Letters to the Editors www.AJOG.org

Society for Maternal-Fetal Medicine Publications Committee409 12th St. S.W.Washington, DC 20024Contact: Beth Steele (esteele@smfm.org)

Each member of the Society for Maternal-Fetal MedicinePublications Committee has submitted a conflict of interestdisclosure delineating personal, professional, and/or businessinterest that might be perceived as a real or potential conflict ofinterest in relation to this publication andnone have a conflict.-

REFERENCES

1. Crombleholme TM, Shera D, Lee H, et al. A prospective, randomized,multicenter trial of amnioreduction vs selective fetoscopic laser

160 American Journal of Obstetrics & Gynecology AUGUST 2013

photocoagulation for the treatment of severe twin-twin transfusion syn-drome. Am J Obstet Gynecol 2007;197:396.e1-9.2. Senat MV, Deprest J, Boulvain M, Paupe A, Winer N, Ville Y. Endo-scopic laser surgery versus serial amnioreduction for severe twin-to-twintransfusion syndrome. N Engl J Med 2004;351:136-44.3. Roberts D, Gates S, Kilby M, Neilson JP. Interventions for twin-twintransfusion syndrome: a Cochrane review. Ultrasound Obstet Gynecol2008;31:701-11.4. Rossi AC, D’Addario V. Umbilical cord occlusion for selective feticide incomplicated monochorionic twins: a systematic review of literature. Am JObstet Gynecol 2009;200:123-9.5. van Klink JM, Koopman HM, van Zwet EW, Oepkes D, Walther FJ,Lopriore E. Cerebral injury and neurodevelopmental impairment afteramnioreduction versus laser surgery in twin-twin transfusion syndrome: asystematic review and meta-analysis. Fetal Diagn Ther 2013;33:81-9.

ª 2013 Mosby, Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajog.2013.03.021

Still a screening test: more attention needed to noninvasiveprenatal test false-positive rates

TO THE EDITORS: The study of Nicolaides et al,1

“Noninvasive prenatal testing for fetal trisomies in aroutinely screened first-trimester population,” demonstratesthat techniques analyzing cell free fetal DNA in the maternalcirculation can perform well to detect aneuploidy in a generalobstetric population. However, we must consider the impactof false-positive results before promoting this technique as aprimary screening tool. The proportion of false- to true-positive results will be much higher among pregnancies inlow-risk women. The study of Nicolaides et al1 confirms thisproblem, but does not highlight it.

Noninvasive prenatal tests (NIPT) are being aggressivelymarketed directly to the public and to obstetric clinicians.Companies prominently cite >99% sensitivity and specificityfor trisomy 21 and trisomy 18 on their reports and promo-tional materials. While these figures are based on results frompublished trials,2,3 they may be misleading to both patientsand physicians, due to the common confusion of high spec-ificity with high positive predictive value. With the perceptionthat NIPT is highly accurate, there is a real danger thatpositive test results will be interpreted as diagnostic, leadingcouples to terminate pregnancies without waiting for aconfirmatory karyotype.

Since NIPT became commercially available in the UnitedStates we are aware of 13 fetal samples received at 6 US cytoge-netics laboratories (ACL Laboratories, Park Ridge, IL; Brighamand Women’s Hospital, Boston, MA; Case Western Reserve,Cleveland, OH; Greenwood Genetic Center, Greenwood, SC;Johns Hopkins, Baltimore, MD; Stanford, Palo Alto, CA) thatwere positive for trisomy using NIPT but not confirmed byconventional chromosome analysis. SevenNIPTwere positive fortrisomy 18, 5 positive for trisomy 21, and 1 for trisomy 13.Normal karyotypes for 11 were obtained on amniocentesis and

for 2 through chorionic villus sampling. Although we cannotcalculate a false-positive rate without knowing the denominatorfrom which these samples were drawn, the study of Nicolaideset al1 itself provides some information about the rate of false-positive findings. They identified 2 cases of euploid fetuses with“risk scores” for trisomy 18 above their screening cutoff, repre-senting a false-positive rate of 2 of 1949 (0.10%). The prevalenceof confirmed trisomy 18 in first trimester in their unselectedpopulation was only 3 of 2049 women (0.15%), and will be stilllower in a low-risk population or in pregnancies screened later ingestation. Thus for rarer trisomies, the false-positive rate mayexceed the true-positive rate.

The authors correctly suggest that confined placental mosa-icism may explain some apparent false-positive results. Otherpossibilities include low-frequency fetal or maternal mosaicism.However, as with any screening test, some false-positive resultsare inherent in need to choose cutoffs tominimize false-negativefindings. Systematic evaluation of cases with discordance be-tween positive NIPT results and cytogenetic results should beundertaken before these tests are applied generally, or analysisis extended to more chromosomes. Postnatal evaluation ofthe newborns, their placentas, and possibly the mother’s kar-yotype from cases with discordance is needed. We recommendinitiating a registry for discordant cases to allow systematicassessment of false-positive results. Meanwhile physicians andmidwives offering NIPT must themselves understand andconvey to their patients that positive results must be confirmedby diagnostic tests before pregnancies are terminated. -

Rosemary E. Reiss, MDDepartment of Obstetrics and GynecologyBrigham and Women’s HospitalBoston, MA 02115rreiss@partners.org