Rapid Responses to:

PAPERS: Diana Wellesley, Tracy Boyle, John Barber, and David T Howe Retrospective audit of different antenatal screening policies for Down's syndrome in eight district general hospitals in one health region BMJ 2002; 325: 15 [Abstract] [Full text]

Rapid Responses: Submit a response to this article

Rapid Responses published:

Antenatal Screening can work

David M Lewis   (8 July 2002)

Serum screening for Down's syndrome is much better than age screening

Michael Christiansen, Severin O. Larsen, and Bent Nørgaard-Pedersen   (18 July 2002)

Audit of Down's syndrome screening not valid

Nicholas J Wald, Wayne Huttly, Allan K Hackshaw   (19 July 2002)

Comparative Study on Down screening policies wrongfully discredits risk screening tests

Peter Schielen, Gerard Loeber   (19 July 2002)

Coordination of Pregnancy Screening for Down Syndrome

Philip R. Wyatt, C.Meier, Director, Ontario MSS Database   (8 August 2002)

Conclusions are flawed

Christoph C Lees, Andrew Breeze, Sandy Goodburn   (17 August 2002)

Authors reply to correspondents

David T Howe, Diana Wellesley   (4 September 2002)

Antenatal Screening can work 8 July 2002
David M Lewis, General Practitioner Watford WD24 7PH Send response to journal: Re: Antenatal Screening can work	I am concerned that the Wessex audit for antenatal screening for Downs syndrome does not properly reflect modern approaches. As aninterested GP with many mums requesting information about prenatal screening I try to keep abreast of development sin this area. The article by Wellsley et al does not serve my mothers well: while I cannot comment on the veracity of the audit data, the old references cited to justify some of their claims deserve comment. The following summary of a recent review of antenatal screening is an example that deserves to be read by policy makers so that the baby is not tipped out with the (screening) bath water: "The integration of information on epidemiological risk factors, pregnancy-specific screening markers and ultrasound anomalies can be simply achieved by the calculation of Down's syndrome risk. Statistical modelling predicts that combining different screening modalities will yield very high Down's syndrome detection rates"(1). 1. Cuckle H. Integrating antenatal Down's syndrome screening. Current Opinion in Obstetrics & Gynecology. 13(2):175-81, 2001 Apr
Serum screening for Down's syndrome is much better than age screening 18 July 2002
Michael Christiansen, Consultant Department of Clinical Biochemistry, Statens Serum Institut, DK 2300 S Denmark, Severin O. Larsen, and Bent Nørgaard-Pedersen Send response to journal: Re: Serum screening for Down's syndrome is much better than age screening	Editor- Wellesley et al(1) have performed a retrospective audit of different policies of antenatal Down’s syndrome screening in 8 hospital districts and conclude that serum screening is not significantly better than age-screening. This conclusion contradicts the international experience(2), and cannot be made from the reported data: The actual activities in the different districts do not reflect the policies stated, e.g. a policy, as in district A(1), where serum screening is offered without a routine anomaly scan, result in 11 of 33 detected DS fetuses being found through an anomaly scan! Was serum screening performed in these patients? For each test we lack information on the number of women tested and the detection and screen positive rates for each test. Without this, a comparison between different policies is impossible. The authors make a classical mistake by not correcting for the substantial death rate in utero when calculating the total prevalence and this is important here, where DS cases may be detected as early as in week 12 from CVS and as late as in week 24 from an anomaly scan. Screening performance obviously depends on the age-distribution of the examined population(3). The performance of age-screening can be calculated from the age-distributions of pregnant women and is certainly inferior to the well documented – in prospective settings(2) as well as in audit reports from UK(4) - performance of second trimester serum screening. It is a misunderstanding that modelling studies have presumed that 5% of pregnant women were older than 35 years and it is not surprising to find that the proportion is 15%(5). Serum screening also reduces the rate of invasive procedures in older women(4). Detection rates are partly based on DS cases where the mothers refused prenatal testing! In district A, among women older than 35 years, 9 DS babies out of 31 registered escaped antenatal detection because the mothers refused the test and 2 because the serum screening failed – thus the detection rate is 20 detected cases/ 22 total cases = 91% in the screened population and not 65%. One dare not think of the poor performance of the tests if a greater part of the population had refused to participate in the screening. Hopefully, the study by Wellesley et al(1) will not confuse the issue and delay the abandonment of the obsolete age-screening that is causing an unacceptable number of unnecessary procedure related abortions of healthy foetuses. References: 1. Wellesley D, Boyle T, Barber J, Howe DT. Retrospective audit of different antenatal screening policies for Down’s syndrome in eight district general hospitals in one health region. BMJ 2002; 325: 15. 2. Cuckle H. Established markers in second trimester maternal serum. Early Hum Dev 1996; 47(Suppl): S27-S29. 3. Larsen SO, Christiansen M, Nørgaard-Pedersen B. Calculation of roc curves in multidimensional likelihood ratio based screening with Down’s syndrome as a special case. J Med Screen 1998; 5: 57-62. 4. Edwards VM, Kennedy DM, Worthington DJ. Birmingham Womens’s Hospital Antenatal Screening Unit Audit for Down’s Syndrome and Neural Tube Defects. April 1997 – March 1999. Department of Clinical Biochemistry, Birmingham Women’s Hospital, Birmingham (Vicky.Edwards@bham- womens.thehs.com) 5. Christiansen M, Oxvig C, Wagner J, Qin QP, Nguyen TH, Overgaard MT et al. The proform of eosinophil major basic protein: A new maternal serum marker for Down syndrome. Prenat Diagn 1999; 19: 905-910.
Audit of Down's syndrome screening not valid 19 July 2002
Nicholas J Wald, Professor Wolfson Institute, Barts and the Royal London School of Medicine and Dentistry, London EC1M 6BQ, Wayne Huttly, Allan K Hackshaw Send response to journal: Re: Audit of Down's syndrome screening not valid	Wellesley et al [1] are unjustified in their view that serum screening for Down’s syndrome is not worthwhile. They ignore the wealth of evidence from previous studies that demonstrates the substantial advantage of serum screening over screening based on maternal age alone.[2-4] Their view is based on the results from two Districts, the only ones out of the eight where serum screening was routinely offered (using the Double test). In these two Districts only 24% of affected pregnancies (22 of the 91) were detected antenatally using serum screening even though the Double test has a detection rate of 58% for a 5% false-positive rate.[2] This could, at least in part, be due to a low screening uptake (in which case the focus should be on why this was so), but the uptake of screening cannot be determined from the paper. It could also be due to affected pregnancies that were positive on the serum test and positive on other tests being classified as positive under the other tests. This problem arises because the “detection rate” as used in this paper includes women who declined screening (the performance of screening tests, the uptake of screening and the uptake of amniocentesis should all be reported separately) and tabulating methods of detection (serum screening, scan, maternal age, etc) in mutually exclusive categories which cannot be correct because there will be Down’s syndrome pregnancies positive in two or more of these categories. Another problem is that Wellesley et al compare screening results across different districts. In doing so they cannot (through lack of data) consider the substantial variations between districts in important factors that influence screening such as maternal age, uptake of screening, and use of ultrasound. This problem is solved by comparing different screening methods in the same women – each woman acting as her own control. There is no better controlled study. Had they performed such an analysis, as we suggested in response to their earlier paper,[5] they would have confirmed the substantial advantage of serum screening over age screening. Their call for controlled trials is unjustified. Cont/……… Wald/Editor Page 2 Wellesley et al are incorrect in stating that the mathematical modeling carried out by others to estimate the effect of serum screening assumes that only 5% of pregnant women were aged over 35. Such modeling takes account of the maternal age distribution of the relevant population. The paper is scientifically flawed. It fails to specify the performance of individual screening methods and does not assess the benefits of one over another. It is impossible to draw any valid conclusions about the different methods of Down’s syndrome screening from this paper. Yours faithfully N J Wald FRCP W Huttly BSc A K Hackshaw MSc Professor Screening Manager Statistician References 1. Wellesley D, Boyle T, Barber J, Howe DT. Retrospective audit of different antenatal screening policies for Down’s syndrome in eight district general hospitals in one health region. BMJ 2002;325:15- 2. Haddow JE, Palomaki GE. Prenatal screening for Down syndrome. In: Essentials of Prenatal Diagnosis. JL Simpson and S Elias (Eds). Churchill Livingstone 1993 Edinburgh 3. Wald NJ, Kennard A, Hackshaw A, McGuire A. Antenatal screening for Down's syndrome. J Med Screen 1997; 4:181-247 4. Spencer K. Second trimester prenatal screening for Down’s syndrome using alpha-fetoprotein and free beta hCG: a seven year review. Br J Obstet Gynaecol 1999;106:1287-93 5. Wald NJ, Hackshaw AK, Huttly W. Serum screening programmes are effective and safe (Response to Howe). BMJ 2000;321:763-4
Comparative Study on Down screening policies wrongfully discredits risk screening tests 19 July 2002
Peter Schielen, Porject leader Down syndrome/NBD screening National Institute of Public Health and the environment, Gerard Loeber Send response to journal: Re: Comparative Study on Down screening policies wrongfully discredits risk screening tests	EDITOR, the paper of Wellesley et al. concludes that current additional serum and nuchal translucency screening is less advantageous than previously supposed. They argue along two lines of evidence. They present data to suggest that the age distribution of screened populations differs from the distribution in mathematical model studies. Therefore, they argue, these mathematical studies give an overestimation of the effect of screening of the younger pregnant women. Secondly, they present data of a comparison between screening policies in eight maternity units in various districts in Wessex, suggesting that districts that adopted serum and nuchal translucency screening did not perform better than districts that had a screening policy based on maternal age and an anomaly scan. We dispute that the presented data support the authors’ conclusion. The authors present data of the maternal age distribution of the general population of pregnant women in eight Wessex districts in 1998 , showing that on average 15 % was 35 or older. In this age group, they found 58% of all antenatally detected Down syndrome pregnancies. This led the authors to suggest that maternal age alone would suffice to detect the majority of Down syndrome pregnancies. The mistake the authors make is that not the age distribution in the general population is important in this respect, but the age distribution of the pregnant women that participate in the screening programmes. If the screening programmes would take up a relatively large proportion of older pregnant women, it is only logical that the larger proportion of detected Down syndrome can be found in exactly that age group. Indeed, the uptake of e.g. our own serum screening programme is heavily biased towards older pregnant women, with no selection criteria on maternal age in the counselling. In conclusion, without information on the age distribution of the screened populations, the conclusions of the authors concerning the differences in age distribution are meaningless. The authors present data on the detection rates of various districts suggesting that in those districts, entirely different and strictly defined screening programmes were effective. However, the presented data suggest otherwise. In region A, 11 Down syndrome pregnancies were detected by anomaly scan, while this was not part of the prevailing policy In region B, where nuchal translucency measurement was not part of the policy, 5 Down syndrome pregnancies were detected by nuchal translucency. In regions F and G together, 3 Down syndrome pregnancies were detected by serum screening and 8 by nuchal translucency, while those methods were not part of the local screening policy. In line with the incorrect presumption that the districts under investigation do not represent clearly defined screening policy but rather a mix, to unknown extent, of various screening policies, the authors calculate the detection rates incorrectly. It goes without saying that e.g. the Down syndrome pregnancies detected by nuchal translucency measurement and serum screening should not be included to elevate the detection rates of the districts that represent screening policies based on maternal age and anomaly scans. Finally, the fact that the uptake of the screening programmes is not taken into consideration is another major flaw in the study design. It is evident that any programme with an uptake of e.g. 10% underperformes compared to a programme with an uptake of e.g. 90 % The authors give no data on the uptake of the screening programmes in most of the districts and state that they have no information on the uptake of serum screening and nuchal translucency measurement in district E. In district A and B, the uptake of serum screening dropped to a mere 55-65 % in 1999. The detection rate in districts A and B may well have suffered from this low uptake, especially when one considers that the undetected Down syndrome pregnancies may well have been in the non-screened population of districts A and B. Data on the uptake of the various screening programs might have given important information on possible differences in counselling strategies among the participating districts, that may have influenced the detection rates in the districts immensely. Differences in counselling are yet another reason why performances in various district cannot be compared. In conclusion, while in principle a comparison between the results of districts with different screening policies is useful, the results as presented in this study are meaningless because a) the screening methods are not exclusively confined to one district, b) detection rates are flawed by down syndrome pregnancies detected by other screening methods than the one in order in the particular district and c) because of the absence of uptake figures. While the authors do not present any proof for their assumptions, they discredit the accomplishments of risk-estimation methods for Down syndrome pregnancies in the first and second trimester of pregnancy. There is now a vast body of evidence that first and second trimester serum screening, either or not in combination with nuchal translucency measurement, yields detection rates roughly between 60-90%, depending on the age distribution of the population under investigation, the screening method and the care with which the screening method is performed. Serum screening and nuchal translucency measurement are done in the first or early second trimester of pregnancy, allowing more than enough time for prenatal diagnosis and, if necessary, a termination of the pregnancy. An anomaly scan performed in the 19th or 20th week of pregnancy seriously shortens this period, up until the point where it is almost too late to terminate the pregnancy. Yet, the authors propagate the anomaly scan in combination with maternal age screening as the method of choice. One of the major advantages of serum screening or nuchal transluceny measurement is that it takes off the weight of invasive procedures. Whether both methods functioned as such in e.g. districts A and B or D cannot be extracted from the presented data. The invasive procedure rates as presented by the authors suffer equally from the poorly defined methodology as all other data. It would take data on the uptake of the various screening methods and a one-on-one link between the method of screening and a pregnant women’s decision for invasive testing to judge the effect of screening methods on the uptake of invasive testing. The authors, again, present no such data. P.C.J.I Schielen and J.G. Loeber National Institute of Public Health and Environment Diagnostic Laboratory for Infectious Diseases and Perinatal Screening (LIS) P.O. Box 1 3720 BA BILTHOVEN The Netherlands Telephone: **31-30-2742165 Telefax: **31-30-2744418 E-mail: Peter.Schielen@rivm.nl
Coordination of Pregnancy Screening for Down Syndrome 8 August 2002
Philip R. Wyatt, Chief, Department of Genetics and senior Consultant Ontario Maternal Screening Database North York General Hospital, Toronto, Ontario M2K 1E1, C.Meier, Director, Ontario MSS Database Send response to journal: Re: Coordination of Pregnancy Screening for Down Syndrome	We read, with some disappointment, the recent article by Wellesley et.al. which has the potential of creating the impression that maternal serum screening (MSS) is an ineffectual method of appropriately screening pregnancies. The scientific data strongly confirms that serum screening is far superior to any form of maternal age screening. The information by Wellesley et. al. however does not have sufficent detail to firmly clarify this, owing to what is another example of poorly coordinated screening being provided to a population. In addition, the report does not provide comparions of detection rates and false positive rates in a form that allows effective comparison to well established standards. The province of Ontario has attempt to correct some of the problems associated with maternal screening screening in a population by providing some resources to allow the collection of both the tests results and outcomes in the population of women who elect for MSS during their pregnacy. To help clearly, again, confirm the effectiveness of MSS in population screening ,in response to the recent article, we have been able to examine the first 293,534 women who have had MSS in Ontario where we have outcome information. Previous publications have confirmed that over 95% ascertainment of all cases is present in the database.<1> By having such a complete dataset it is possible, in the same patients, to examine the value and effectiveness of any screening tool on the same population. This data then looks at the outcomes of screening using the same population with all the same variables (a self controlled population comparion). The review of our next 250,000 outcomes should be available soon. The results are listed for comparison: Population screened N=293,534 Screening Method: Maternal Age: 35 years or older at EDD (risk of Downs of 1 in 385 or higher) False positive rate: 16.9% Detection Rate (DR) of Down Syndrome: 45% Triple marker Screening (risk of 1 in 385 or higher) False positive rate: 7.1% Detection rate: 72% Thus using age alone as a screening tool, to obtain a 45% DR almost 17% of the population would have been exposed to aminocentesis risk. Using MSS, for a DR of 72%, only 7.1% of the women would have had to put their at risk of loss. The decrease of FPR means that thousands of women are spared the anxiety of a false positive rate, and potential pregnancy loss, because of the effective availablity of MSS. Prelimary data would suggest that ultrasound, will slightly increase the detection rate but not substantially decrease the false positive rate. This ultrasound data, in a coordinated fashion, is very difficult to review given the absence of effective quality management in linking ultrasound to outcome data. The results however of all these discusions remain the same. The current haphazard methods of providing screening are ineffective and represent poor medical services to patients. What is clear is that an effort must be made to link ulrasound, serum screening, and outcomes to provide services that continue to allow high detection rates with low false positive rates. The use of age and ultrasound can not equal the coordinated linking of all the proven tools (ultrasound, serum markers, age). We hope that Wellesley et al work will be recognized as another report confirming that disorganized screening services should be stopped. 1. Meier C,Huang,T,Wyatt PR, Summers AM. Accuracy of expected risk of Down syndrome using second trimester triple test. Clin Chem Acta 2002;48:653-5
Conclusions are flawed 17 August 2002
Christoph C Lees, Consultant in Obstetrics & Fetal-Maternal Medicine Rosie maternity Hospital, Addenbrooke's NHS Trust, Cambridge CB2 2QQ, Andrew Breeze, Sandy Goodburn Send response to journal: Re: Conclusions are flawed	The results of the retrospective audit of Down syndrome screening strategies employed in the Wessex region(1) suggest that multiple ad hoc screening strategies for Down syndrome ranging from maternal age based invasive testing to nuchal scan, serum screening and detailed second trimester marker scans, are destined to perform poorly. The national adoption of at least the ‘double-test’, though the least effective of the screening strategies available(2) will at least address the current inequality of Down syndrome screening provision that is based more on geographic location that logic. Our own experience of a universally offered triple test from 14 weeks gestation, in the Eastern region has yielded consistent detection rate for Down syndrome of 80 – 85% with a less than 5% invasive procedure rate. This equates to a 2.5-fold increase in the number of Down syndrome pregnancies identified in the serum-screened population, compared to the number of affected pregnancies identified in the group of women who had chosen age-based invasive testing, in the same period (3). In addition, appropriate targeting of amniocentesis in the screened population has led to a great reduction in the number of unnecessary invasive tests that would have been performed in women over the age of 35. First trimester nuchal screening performs similarly well in a non- selected population(4), and the combination of first(5) or second (6) trimester biochemistry and nuchal screening promises to yield detection rates of 90% for even lower false positive rates. We would strongly take issue with the conclusions of Wellesley’s paper. The lamentably low detection rates for Down syndrome quoted underscore the need to offer effective screening to women of all ages whether by serum biochemistry, nuchal screening or a combination of both. Andrew Breeze, Specialist Registrar in Obstetrics and Gynaecology Sandy Goodburn, Prenatal Screening Coordinator Christoph Lees, Consultant in Obstetrics and Materno-Fetal Medicine Rosie Maternity Hospital, Addenbrooke’s NHS Trust, Hills Road, Cambridge CB2 2QQ, UK. Tel: 01223 217 660 Fax: 01223 216 185 Email: cclees@compuserve.com References: 1 Wellesley D, Boyle T, Barber J, Howe D T. Retrospective audit of different antenatal screening policies for Down’s syndrome in eight district general hospitals in one health region. BMJ 2002; 325:15 2 Gilbert RE, Augood C, Gupta R, Ades A E, Logan S, Sculpher M, Van der Meulen J H P. Screening for Down’s syndrome: effects, safety, and cost effectiveness of first and second trimester strategies. BMJ 2001;323:1-6 3 Goodburn SF, Yates JR, Raggatt PR, Carr C, Ferguson-Smith ME, Kershaw AJ, Milton PJ, Ferguson-Smith MA. Second-trimester maternal serum screening using alpha-fetoprotein, human chorionic gonadotrophin, and unconjugated oestriol: experience of a regional programme. Prenat Diagn. 1994 May;14(5):391-402. 4 Snijders RJ, Noble P, Sebire N, Souka A, Nicolaides KH. UK multicentre project on assessment of risk of trisomy 21 by maternal age and fetal nuchal-translucency thickness at 10-14 weeks of gestation. Fetal Medicine Foundation First Trimester Screening Group. Lancet. 1998 Aug 1;352(9125):343-6. 5 Spencer K, Spencer CE, Power M, Moakes A, Nicolaides KH. One stop clinic for assessment of risk for fetal anomalies: a report of the first year of prospective screening for chromosomal anomalies in the first trimester. BJOG 2000 Oct;107(10):1271-5 6 Michailidis GD, Spencer K, Economides DL. The use of nuchal translucency measurement and second trimester biochemical markers in screening for Down’s Syndrome. BJOG 2001 108:1047-1048
Authors reply to correspondents 4 September 2002
David T Howe, Consultant in Fetomaternal medicine Princess Anne Hospital, Southampton, SO16 5YA, Diana Wellesley Send response to journal: Re: Authors reply to correspondents	We have read with interest the rapid responses to our recently published audit of Down screening in units in Wessex. Each of these letters makes a similar point, that modelling of screening programmes using serum screening (first or second trimester), nuchal translucency or various combinations of the two, suggests that they should be more effective than screening by age alone. These methods should result in higher antenatal detection rates for equivalent or lower invasive procedure rates. Their thinking is in line with the RCOG working party on serum screening[1] which stated in 1993 that “pregnancies in women aged over 36 account for approximately 5% of births but include 30% of births with Down syndrome” and went on to list amongst the advantages of serum screening that “it can detect twice as many pregnancies with Down’s syndrome for the same number of women having an amniocentesis.” We have no argument with the authors that this is what would be expected to occur in theory. What this study, and our earlier audit of screening by maternal age in a single health district[2] demonstrate is what happens when the screening policies are used across whole populations and theory is muddied by the reality of day-to-day clinical practice. A reality that includes the use of routine anomaly scanning; in which women attending the early pregnancy unit with bleeding are found to have a fetus with a cystic hygroma; where women decline serum screening; where older women opt for the certainty of an invasive test rather than the uncertain result of a serum screen; where women decline amniocentesis offered on the basis of their age or serum screen result; etc etc. As long as women choose not to behave as expected screening programmes will not perform as predicted by theoretical models. Two other points deserve specific answers. Breeze et al describe the detection rates reported in our study as “lamentable” despite the fact that the method they advocate, serum screening, was used in two of the districts studied and achieved detection rates, amongst women selecting the test, in the range reported from other demonstration projects[3]. Several of the letters make the obvious point that the detection rate could be improved if the uptake of the screening test was higher, implying that low uptake reflects poor counselling of the women. The aim of counselling is not to coerce women into having the test but to provide them with the information to make an informed choice and low uptake rates may reflect better counselling in those areas. Not every mother agrees with these authors that it is desirable to identify and terminate every fetus with Down syndrome. One study of screening for spina bifida showed that women found it difficult to decline because they felt there was an expectation by society that they should participate[4]. Another study of women’s understanding of serum screening found that only 38% who had the blood test knew they were being screened for Down’s[5] and only 36% knew that a negative result did not exclude the possibility. In a study from the West Midlands of serum screening in an ethnically diverse population[6] the uptake of amniocentesis in Asian women following a high risk result was only 35%. This strongly suggests these women did not understand the full implications of the screening test when the blood sample was taken and that better counselling would have lowered the uptake. The many demonstration projects of serum screening provide information for an individual woman who wishes to know in what proportion of pregnancies with Down’s the screen will be positive, and how likely she is to be offered invasive testing. They do not inform those in public health wishing to know whether purchasing serum screening for the local population will achieve the expected advantages and whether the costs involved are justified. Our study shows within Wessex the expected advantages of serum screening are not being achieved and we stand by our conclusion that the theoretical benefits of different screening programmes need to be proven in properly conducted trials of their use in practice. 1. Royal College of Obstetricians and Gynaecologists. Report of the working party on biochemical markers and the detection of Down's syndrome. London: RCOG Press, 1993. 2. Howe DT, Gornall R, Wellesley D, Boyle T, Barber J. Six year survey of screening for Down's syndrome by maternal age and mid-trimester ultrasound scans. British Medical Journal 2000;320:606-10. 3. Wald N, Kennard A, Hackshaw A, McGuire A. Antenatal screening for Down's syndrome. Journal of Medical Screening 1997;4:181-246. 4. Sjogren B, Uddenberg N. Decision making during the prenatal diagnostic procedure. A questionnaire and interview study of 211 women participating in prenatal diagnosis. Prenatal Diagnosis 1988;8:263-73. 5. Smith DK, Shaw RW, Marteau TM. Informed consent to undergo serum screening for Down's syndrome: the gap between policy and practice. British Medical Journal 1994;309:776-777. 6. Ford C, Moore AJ, Jordan PA, Bartlett WA, Wyldes MP, Jones AF, et al. The value of screening for Down's syndrome in a socioeconomically deprived area with a high ethnic population [see comments]. British Journal of Obstetrics & Gynaecology 1998;105(8):855-9.