The heterozygote carrier screening paradigm is starting to shift from ancestry based screening for carriers of a single or a few genetic diseases to pan-ethnic screening for carriers of a wide range of genetic diseases. New techniques of DNA sequencing make it possible to test a single sample to determine carrier status for dozens of genetic conditions at prices that make carrier screening panels (CSPs) very tempting to healthcare providers and patients. Since carrier screening for one or another genetic disease – cystic fibrosis, Tay-Sachs, hemoglobinopathies – is already offered to essentially all women who are pregnant or planning a pregnancy, why not “screen everyone for everything” at no greater up front cost? And it doesn’t even require a blood sample; a less intimidating buccal sample works just fine.
Part of the understandable justification to move beyond targeted carrier screening programs is the futility of trying to classify people into distinct ancestry/racial/ethnic categories. Gene mutations and genetic diseases have a pesky habit of flowing fluidly between populations, and cultural heritages can be lost through assimilation (See Interesting Digression* below).
So why object to CSPs? After all, don’t people have a right to “know their DNA” and to understand what health and reproductive risks they face?
I am not suggesting that we should stand in the way of anyone’s wishes to “know their DNA.” If someone chooses to acquire that knowledge without the benefit of meeting with a genetic counselor, well, I may disagree with that decision but I respect it. But whether people decide to undergo carrier testing through a genetic counselor or through an online testing company, they need information that is forthright, complete, and transparent; they do not need subtly biased sales pitches. Private companies do not have a vested interest in talking people out of genetic testing.
Before we examine how some labs “objectively” describe carrier screening to patients, we must acknowledge an ethically uncomfortable truth. Carrier screening does not consistently lead to better treatments, encourage greater tolerance of disabilities, stimulate research into cures, or improve psychosocial adaptation to genetic disease. The only compelling reason to devote economic and medical resources to carrier screening is to reduce disease incidence. For better or worse, that is the measure of success of Tay-Sachs screening in Ashkenazi Jews and thalassemia screening on Cyprus.
Three strategies can reduce the incidence of genetic disease. One is mate selection based on carrier status, which is rare except in select populations such as Ultra-orthodox Jews (Hey Single Women out there, when was the last time you met potential Mr. Right and said “Er, you can buy me a drink but do you mind if I take a pedigree and a cheek swab from you before I give you my phone number?”). A second approach is preimplantation genetic diagnosis, but it is available to only a miniscule percentage of the population. The third and only realistic option for most patients is the elephant in the room – prenatal diagnosis with termination of affected fetuses.
Take a look at the web sites of companies such as Counsyl or 23andMe and you get a different narrative. The word “abortion” does not appear. Instead, you read about sperm/ovum donation, preimplantation diagnosis, mental preparation, watchful waiting, and early treatment. No mention is made that early treatment requires testing the baby anyway and that some treatment is available for only a handful of the screened conditions. The websites do not bring up the point that there are no large-scale studies that have shown better familial adaptation to genetic disease when parents have prenatal awareness of their carrier status, so couples really cannot know if testing really will result in mental preparedness. And I am still not sure what watchful waiting is, and how it differs from mental preparedness.
A second concern is that screening for very rare conditions plays on the emotionally vulnerable state of many pregnant women and the difficulty almost anyone has in understanding very, very small numbers in a psychologically meaningful way. Take for example, a condition that has an incidence of 1/100,000 births with a 75% carrier detection rate. Before carrier testing, a couple would have an ~99.9999% of NOT having an affected child; after carrier testing that probability would increase to ~99.99999%. Really, who can tell the difference between those two statistics? It’s difficult just trying to count the number of nines in those numbers. But read about the condition’s severe intellectual disabilities and physical birth defects, and, damn the statistics, give me that test.
A third concern is the lack of complete information about test sensitivity on the information portion of the website. For example, a patient with normal carrier test results might understandably think they would not have to be concerned about having a child with Bardet Biedl syndrome. What the site does not indicate however is that BBS1 and BBS10, the two loci included in the panel, account for less than half of patients with Bardet Biedl syndrome, and that the dozen or so other genes that can cause Bardet Biedl syndrome are not included in the test panel.
A fourth, and maybe the greatest, concern is the ethical difficulty of deciding which conditions to include on a CSP. Tay-Sachs screening among Ashkenazi Jews and thalassemia screening in Cyprus developed with significant input from families, medical professionals, and community and religious leaders. There was widespread agreement in those communities that these were serious diseases and that carrier screening, mate selection, or prenatal diagnosis were ethically acceptable ways of reducing disease incidence. Very little community dialogue has taken place over CSPs. Do we really believe that the world is a better place if we screen for carriers of a common form of hereditary deafness or, God help us, red hair color?
And ruminate on this: a study of 3 million cystic fibrosis carrier tests performed at a single US lab found that 25,000 CF carrier screens needed to be performed to detect one affected fetus. And this is for a relatively common genetic condition with a frequency of about 1/4000 US newborns and a screening program whose success remains debatable. How many carrier screens will need to be performed to detect a fetus or newborn with a rare disorder like isovaleric academia, with a frequency of 1/250,000 births?
It could be that I am just the last of the old wave of genetic counselors who are out of touch with new technologies and changing ethical values, the proverbial last leaf on the tree. Maybe I am a 20th century genetic counselor in a 21st century world in which private industry will become the primary mode for the delivery of medical genetic services. Perhaps when I retire in a decade or so the genetic counseling community will issue a collective sigh of relief. But sometimes Old School cranks have a point.
* – Interesting Digression: I recently learned about the Jews of Acadiana, Jewish merchants who settled among, and who were often culturally and reproductively assimilated into, Louisiana’s Cajuns (although the Cajun Tay-Sachs mutations stem from their French Canadian origins and predates the Ashkenazi admixture). Also, an exploration of why Tay-Sachs screening caught on among Ashkenazi Jews but not among Cajuns would make for an interesting socio-medical-historical study. If a large scale Tay-Sachs screening program were to be introduced among Cajuns, perhaps its motto would be Laissez les bon genes roulez.
The DNA Exchange 
Very well said. Glad to hear I’m not the only “old school – 20th century” genetic counselor out there.
There are more of us 20th century leaves on the tree than you may imagine; you are not alone, Mr. Resta.
I am not yet a genetic counselor (will begin a program this fall), however even I have come to recognize that what you say is true. I think the two most important statements you wrote are:
1. “Private companies do not have a vested interest in talking people out of genetic testing.”
2. “Carrier screening does not consistently lead to better treatments, encourage greater tolerance of disabilities, stimulate research into cures, or improve psychosocial adaptation to genetic disease. The only compelling reason to devote economic and medical resources to carrier screening is to reduce disease incidence.”
Thanks for this post!
Seriously, you went with redheads? That’s hurtful, Bob.
Aside from the random smack at gingers, I agree with all that you say — only there’s no use freaking out over panel testing, as it is just a brief stop along the road to WGS. We won’t be staying here long enough to sort out what is wrong with this paradigm — let’s just transfer all our baggage straight to the next generation location, shall we?
I appreciate hearing your thoughts on this important issue. I will openly admit to being an advocate of screening (working as a genetic counselor in an academic instituion). From my perspective, it’s important that we find new ways to encourage preconception screening on a population scale in the United States. This likely isn’t going to come through a traditional public health mechanism (given the elephant in the room). Therefore, we need to be creative about how to accomplish this so that women and couples have the time they need to think through the issues you raise outside of the pressures of a ticking clock. It doesn’t matter to me whether these ideas come from old or new leaves.
I know this is not the main point of your post but we were very struck by the 1/25,000 number that you quote. We went back and did some digging on the paper and have reason to believe that this is actually a typo or incorrect calculation. We have sent the following e-mail to the corresponding author and will let you know if we get a response:
We were directed to your paper by a recent blog post by Robert Resta at (thednaexchange.com). We were immediately struck by the finding that you would need to perform 25,000 carrier screens to detect one affected fetus. We are trying to understand the derivation of this number and are finding it difficult to do so from the figures presented in the publication.
It would appear that this number is derived from the finding that you detected 119 affected fetuses (according to the abstract) among 2,975,649 carrier screens performed (119/2,975,649= 1/25,005). However, you report in the results section that you actually detected 199 affected fetuses, which would have yielded a result of (199/2,975,649= 1/14,953). It is impossible for us to determine the actual number of affected fetuses you detected, as there is a discrepancy between the abstract and results. As the 25,000 number is being discussed as a point of evidence for why carrier screening may not be as successful as previously anticipated, we feel it is very important to clarify whether this number is correct.
We are similarly confused by the math used in the results section to estimate the percentage of couples that elect for prenatal testing.
First, the two carrier frequencies you quote based on your data are 1/37.6 for all alleles, and 1/38.4 for the ACMG panel. You estimate the number of carrier couples to be 1/1149. How did you calculate this number? 1/37.6^2 is 1/1414. 1/38.4^2 is 1/1475. If you assume the female frequency to be that derived from ACMG, and the male frequency to be that derived from the total alleles then 1/37.6 x 1/38.4= 1/1444.
We agree with the math, in general, to calculate the number of fetuses you would have expected as being approximately 504. We have worked this calculation using all of the different carrier frequencies you give and come up with a range of 492-512 affected fetuses. We agree with the estimation that 199×4=796 carrier couples are predicted to have elected for prenatal diagnosis.
However, the calculations that follow do not seem appropriate. It would seem that as you expected 504 affected fetuses, you would similarly expect 2,016 carrier couples (504×4= 2016) as the basis for your calculations of prenatal diagnosis uptake. We do not understand at all why 1149 would be your expectation of the number of carrier couples. This is your stated carrier couple frequency. While we don’t understand how this is derived, the frequency of carrier couples should be multiplied through by the total number of couples tested to approximate the number of carrier couples in your cohort. Therefore, your estimation that 53% (which we similarly don’t understand the derivation of as 796/1149= 69%) or 2/3 (depending on which sentence you refer to in the results section) of carrier couples elect prenatal diagnosis is incorrect. It seems the estimation based on your data should actually be 796/2016= 39.5%. However, this is likely an underestimate as accurate uptake estimations must be derived from studies initiated at the point of care to avoid the ascertainment bias that you discuss.
We would appreciate your clarification on the derivation of your estimates so that incorrect approximations do not get perpetuated.
Gemma Chandratillake, MPhil, PhD, MS, CGC
Sarah Garcia, PhD, MS, CGC
Looking forward to hearing the response to this inquiry.
In the meantime, I looked at the article. I believe that the 119 figure is correct and the 199 figure is incorrect. My reason is as follows:
The article states in the Results section: “We have performed 1310 prenatal diagnoses for CF since the inception of our program. The genotype results are 832 homozygous wild type, 359 heterozygotes, and 199 affected.”
If you do the addition of 199 + 359 + 832, it does not add up to 1310, whereas 119 + 359 + 832 does equal 1310.
Therefore, I believe (although cannot be certain) that the 1 in 25,000 affected fetus ratio is correct.
Hi Emily,
Good point! But if you use the 119 figure, you then have to further adjust the estimation of the number of couples who elect prenatal diagnosis since the authors use 199 in this calculation (the numerator would then be 119×4=476 rather than 199×4=796). This gives an even lower estimate of the uptake of prenatal diagnosis by carrier couples (476/2016=24%). It seems likely that this is an underestimate since prenatal diagnostic tests may have gone to labs other than Quest (even if the carrier screening tests that identified the carrier couples were sent to Quest). I’m not sure if anyone has specifically looked into the rate of uptake of prenatal diagnosis for CF carrier couples? Since correctly estimating the number of carrier screens needed to detect an affected fetus hinges on knowing the uptake rate for prenatal diagnosis, I would still be cautious about putting too much weight on the 1/25,000 figure since this assumes that all the carrier couples detected by Quest who chose prenatal diagnosis had testing through Quest.
We’ll post any reply we get from the authors here….
Thanks to Sarah and Gemma for their well thought out response. It will be interesting to hear the response of the article’s authors, and I do hope Sarah and Gemma, or the article’s author, will share it here on The DNA Exchange. Your responses (and the others) are what make blogging so exciting, and lead us to think about – and re-think – what we know, what we believe, and what we should be doing clinically.
And, by the way, apologies to all red heads out there. I was not singling you out as reproductively undesirable (though red-headedness is often cited as an example of negative assortative mating in humans). I was just pointing out the ridiculous extremes to which carrier testing could potentially be carried out nowadays. As a severe deuteranope, I can honestly say I have never intentionally acted in a prejudicial way against redheads because of their hair color.
Robert Resta – as a physician, I find it so disappointing to come across posts like this – so you think reducing disease incidence is not an important cause? Not worth it you say? You should spend a day or a week or make it your life’s work to sit in my office and talk to parents who have children born with genetic diseases that could have been screened for using a test like the Counsyl test. I can only hope they continue to add to the number of diseases that are screened for.
Taking a step back, think about the implications to our health care system – proper screening as a pre-conception measure would save our national health care system billions upon billions of dollars. As more people are tested, think about the disease associations a company like Counsyl could uncover – just food for thought. This is the way of the future. You either have your head in the sand, and advise people to do the same (a “cross your fingers and hope for the best” approach), or you show them the future and as a counselor, you EDUCATE them. Anything else would just be lazy and a disservice to the genetic counseling community that does such wonderful work most of the time.
I got a different impression from reading this post.
In response to your first paragraph:
1. The author does not say that reducing disease incidence is not an important cause. Indeed, he says that reducing disease incidence is in fact the “only compelling reason to devote economic and medical resources to carrier screening”.
In response to your second paragraph:
2. There is a difference (ethical and otherwise) between “proper screening as a pre-conception measure”, which you seem to endorse, and prenatal diagnosis/carrier screening in the middle of a pregnancy. In this post, the author seemed to address the latter, not the former.
“…[T]here are no large-scale studies that have shown better familial adaptation to genetic disease when parents have prenatal awareness of their carrier status…” – Are there any “small-scale” studies on this? If so, I’d be really interested to read about them! Any links to publications would be greatly appreciated.
Hmmm….. I am not aware of such studies but I do not claim encyclopedic knowledge. There are a few studies that have looked at prenatal diagnosis of a condition, and compared it to how parents deal with the diagnosis compared to postnatal diagnosis. Not exactly the same thing as what you are getting at, but it might be a start. Here are a few such references:
Clin Nurs Res. 1996 Feb;5(1):81-96.
PMID: 8680341
J Perinatol. 2007 Nov;27(11):687-92. Epub 2007 Aug 23.
PMID: 17717519
Disabil Health J. 2012 Apr;5(2):67-74. doi: 10.1016/j.dhjo.2011.11.001. Epub 2011 Dec 28.
PMID: 22429541
Skreden M, Skari H, Malt UF, Haugen G, Pripp AH, Faugli A, Emblem R.
Am J Med Genet A. 2010 Sep;152A(9):2193-202. doi: 10.1002/ajmg.a.33605.
PMID: 20803642
Bevilacqua F, Palatta S, Mirante N, Cuttini M, Seganti G, Dotta A, Piersigilli F.
J Matern Fetal Neonatal Med. 2013 Aug;26(12):1249-53. doi: 10.3109/14767058.2013.776536. Epub 2013 Mar 13.
PMID: 23419128
Thank you so much for the quick and comprehensive reply! This is exactly what I was hoping for.