[ letters * correspondance Instructions to authors: Only letters submitted by mail, courier or e-mail (not fax) that do not exceed 450 words are considered for publication. Letters referring to recent articles in CMAJ must be received within 2 months of the publication of the article concerned. Letters sent by mail or courier must be submitted in duplicate and printed in letter-quality type without proportional spacing. Letters sent by e-mail should be addressed to
[email protected], indicating in the subject line "letter to the editor of CMAJ " and including the telephone and fax numbers of the author or authors. A signed copy must be sent subsequently by fax or mail. Letters sent by e-mail are published in CMAJ Online (http://wwwvcma.ca/journaIs/cmaj) before they appear in print. Letters must not duplicate material being submitted elsewhere or already published. CMAJ corresponds only with the authors of accepted letters. Rejected letters are destroyed, and accepted letters are subject to editing and abridgement. Directives aux auteurs : La re'daction n'6tudiera pour publication que les lettres d'au plus 450 mots soumises par la poste, par messager ou par courrier electronique (et non par te'lcopieur). Les lettres qui portent sur des articles recents du JAMC doivent etre re,ues dans les 2 mois suivant la parution de l'article en cause. Les lettres envoy6es par la poste ou par messager doivent etre presentees en double et imprim6es qualite' lettre, sans espacement proportionnel. Les lettres envoyees par courrier 6lectronique doivent etre adresse'es a
[email protected]. Leurs auteurs doivent indiquer j la ligne objet lettre a la re'daction du JAMC. et inclure leurs numeros de telephone et de t6l6copieur. Un exemplaire sign6 de la lettre devra suivre par telecopieurou par la poste. Les lettres regues par courrier 6lectronique sont publiees dansJAMC En direct (httpi/Mwwwcma.calournals/cmaj/index f.htm) avant de paraitre dans le journal. 11 faut 6viter de reproduire des documents soumis ailleurs ou d6ja publi6s. Le JAMC ne correspond qu'avec les auteurs des lettres accepte'es, qui peuvent etre corrigees et abreg6es. Les lettres rejetees sont detruites.
TRIPLE-MARKER SCREENING why age is used as a screening crite- tion about maternal serum screening rion at all, not whether maternal before having the test. In addition, FOR DOWN SYNDROME screening is an appropriate alternative. Maternal serum screening can
serum
le
writing in regard to the article "Periodic health examination, 1996 update: 1. Prenatal screening for and diagnosis of Down syndrome" (Can Med Assoc J 1996; are
154:465-79), by Dr. Paul T. Dick with the Canadian Task Force on the Periodic Health Examination. First, the term triple-marker screening"
is
inaccurate, since dou-
ble, triple and quadruple markers can be used. A non-numeric term, such as maternal serum screening, is preferable. Maternal serum screening is part of a continuum of prenatal screening. In fact, such screening has been offered since the 1960s. The article did not discuss this extremely important point. The first screening criterion used for Down syndrome was the mother's age. If she was 35 years or over, she was offered amniocentesis. In the 1980s screening for the maternal serum ct-fetoprotein (MSAFP) level was introduced, and in the 1990s multiple-marker screening was found to be the best screening tool available (Table 1). The article should have questioned *(--
page 331 see page prescribing information see For prescribing 331
provide
before screening can be offered, aderesources must be provided. In Ontario, a pilot maternal-serumscreening program has been in place since July 1, 1993. Experience with this program has shown that widespread use of maternal serum screening requires augmentation of existing genetic-counselling and laboratory quate
of all ages with an of the risk of Down syndrome. This allows mothers to choose whether to have a diagnostic test; it does not replace the diagnostic test. The authors were concerned that 21 % to 3 1 % of services. women with a positive result of a We feel that this article could screening test decide against amnio- have been improved by discussion of centesis. It should be pointed out its contents with those participating that many women who undergo ma- in maternal-serum-screening proternal serum screening are hoping for grams. The introduction of maternal reassurance. serum screening must be consistent We emphasize the need for with the principles of screening and women to receive sufficient informashould be undertaken only when the women
accurate assessment
Screening criterion or technique (risk
cut-off 1/350).
Age
Rate of false-positiv oe results, Y% 11.9
Rate t, of detection :of Down syndrome, %
,ro. NOT armniocentesis procedures performed to detect one case of Down syndrome
39.8
......
202
.......... -
Age plus maternal serum a-fetoprotein level 12.3 49.8 166 serum Triple marker 9.5 73.5 87 *S6urce: Gerry Dimnik,. Mhnage, Ontario Maternal Serum Screening Database, Toronto: personal commnunication, 1996.
ASSOC
J
*
AUG.
1,1996; 155
(3)
CAN MED ASSOC J * AUG. 1, 1996; 155 (3)
211 271
resources required for all aspects of every patient with a headache undergo a computed tomography scan such a program are available. so that the risk of missing a brain tuSandra A. Farrell, MD, FRCPC, mour is reduced. It makes the Ottawa FCCMG ankle rules practically irrelevant! Diane E. Chadwick, PhD I am concerned about the impact Credit Valley Hospital this guideline as published. I beof Mississauga, Ont. it provides false credibility that lieve Anne Marie Summers, MD, FRCPC, for a test that has serious limitations. FCCMG Philip R. Wyatt, MD, PhD North York General Hospital North York, Ont.
I was excited to read that the apparent sensitivity of triple-marker screening had been improved to the point that it had a rate of falsepositive results of only 3.7%. This rate is supported in a table describing the results of four different studies of triple-marker screening. 1-4 Since I have been an opponent of this test, precisely because of its high rate of false-positive results, I decided to critically appraise and present this article to the local journal club. Since the stated rate of falsepositive results contradicted my own experience, I reviewed the four references that served as the basis for the rate cited. In fact, the rates cited were not the results of the triplemarker test alone but of the triplemarker test in conjunction with some form of confirmation of gestational age, most commonly a subsequent ultrasonographic examination. The rate of false-positive results before further evaluation in each of these studies was approximately double the rate after ultrasonographic confirmation. These findings are more consistent with the previously published data. Although the article by Dick and the task force was published under the rubric of clinical practice guidelines, the abstract specifically states that "the economic issues involved are complex and were not considered." How can a practice guideline be considered useful if the economic aspects involved are not considered? This would be like suggesting that 272
CAN MED ASSOC J * ler AOOT 1996; 155 (3)
Gary Viner, BSc, MD, CCFP Assistant professor Department of Family Medicine University of Ottawa Ottawa, Ont. References 1. Haddow JE, Palomaki GE, Knight GJ, et al. Prenatal screening for Down's syndrome with use of maternal serum markers. N Engi J Med
1992;327:588-93. 2. Phillips OP, Elias S, Shulman LP, Andersen RN, Morgan CD, Simpson JL. Maternal serum screening for fetal Down syndrome in women less than 35 years of age using a-fetoprotein, hCG, and unconjugated estriol: a
prospective 2-year study. Obstet Gynecol 1992;80:353-8. 3. Wald NJ, Kennard A, Densem JW, Cuckle HS, Chard T, Butler L. Antenatal maternal serum screening for Down's syndrome: results of a demonstration project. BMJ 1992;305:391-4. 4. Cheung EY, Luthy DA, Zebelman AM, Williams MA, Lieppman RE, Hickok DE. A prospective evaluation of a second-trimester screening test for fetal Down syndrome using maternal serum a-fetoprotein, hCG, and unconjugated estriol. Obstet Gynecol 1993;81:72-7.
[The author responds:] A lthough the letters from Dr. Farrell and associates and Dr. Viner represent different perspectives on screening with the use of maternal serum markers, they both raise questions about the programmatic aspects of prenatal screening for Down syndrome. There also seems to be a misunderstanding of the role of the task force guidelines. Rather than being a practical guide to local prenatal ser-
vices or programs, the recommendations should be viewed as a guide to the evidence supporting and effectiveness of these interventions according to the literature. Viner questions the task force's recommendations for triple-marker screening involving maternal serum levels of a-fetoprotein, B human chorionic gonadotropin and unconjugated estriol. He differentiates between the rate of false-positive results when the markers are used with and without confirmation of gestational age. As he notes, earlier studies of maternal serum markers without confirmation of gestational age reported higher rates of false-positive results. However, the task force used only the four recent studies for estimates of screening effectiveness. 1-4 These studies met the criteria for level 11 evidence and constituted the best available evidence.' The triplemarker screening in these studies included an ultrasonographic examination for confirmation of gestational age. Thus, the estimates cited in the task force recommendations reflect the screening intervention in toto (i.e., maternal serum markers with confirmation of gestational age), as delivered in a comprehensive screening program. Farrell and associates suggest that the use of the term "triple-marker screening" is inaccurate. They correctly point out that screening with the use of maternal serum markers is evolving. However, they miss the point that triple-marker screening is the most effective combination to have undergone widespread evaluation in clinical trials. In keeping with the task force's emphasis on published evidence, the focus on triplemarker screening was deliberate and the use of the term accurate. Farrell and associates believe that age should not be used as screening test. It is unclear, however, whether they are suggesting that there is no role for maternal age in counselling women concerning their options. It
