Sunday, June 12, 2011

Genetic testing and potential harm: DTC or trust me I’m a doctor? Recently at a couple of conferences (European Human Genetics conference and Consumer Genetics Conf) there have been various speakers questioning DTC genetics and calling for all health related personal genetics to be delivered through medical practitioners. I argued in the past that unregulated tests delivered through practitioners actually have the potential for more harm, not less. By coincidence last week some discrepancies in a DTC and a via MD test were pointed out to me – and they seem topical.
Breast feeding has many benefits one of which appears to be increased IQ scores – however not all studies agree, some indicating that results may be confounded by maternal intelligence (see Wikipedia). Sometimes inconsistencies in associating an action with an outcome can be resolved by looking at genetic variation (which tends to increase the error bars when not accounted for).
So in 2007 some headlines were made when a study was published by Caspi’s group (PNAS, open access) reporting that there is an increase in IQ but only in infants who were not GG for the FADS2 rs174575 SNP. It seemed like a good study, 2 cohorts whose location could not have been farther apart and nearly 3,000 children tested. The numbers begin to get much smaller though when they are dissected as only approx 8% were homozygous GG:

Interesting results – but just one study and there were some doubts expressed at the time (e.g. “they have a measure of maternal IQ but don't directly include it in the published multiple regression suggests that they tried it, but didn't like the results”)
These are preliminary results but a genetic test is was being offered to medical practitioners to enable “informed decisions” to be made (they don’t cite the study but it must be Caspi). They say:
Genetic testing performed by Existence Genetics enables parents to have access to this extremely useful information about their newborn. Although it may be ideal for most mothers to breastfeed, many are in situations that make the decision not so simple. For example, some women take medications that prevent them from breastfeeding, and others experience pain, poor milk production, and other difficulties.
If the baby doesn’t have this gene then breastfeeding appears to have no effect upon IQ.
Empowered by this groundbreaking research, Existence Genetics now provides women, pediatricians, wet nurses, and lactation consultants with exclusive genetic testing services that will allow them to make a more informed decision about breastfeeding.
Infants who have a specific version of the FADS2 gene will have an increase of approximately 7 IQ points, whereas infants without the variant won’t experience any IQ boost from breastfeeding.
Clearly the only influence that this result could have would be to make it easier for mothers not to breastfeed if there is no IQ benefit. Mothers of FADS2 GG babies will be told that breastfeeding will not increase IQ, there is no difference.
Rule No.1: do not base a genetic test on one study, even if it does look compelling. Break this rule at your peril.
Heres why:
Steer CD, Davey Smith G, Emmett PM, Hibbeln JR, Golding J (2010) FADS2 Polymorphisms Modify the Effect of Breastfeeding on Child IQ. PLoS ONE 5(7): e11570.
In a larger study Steer et al attempted replication in 5,394 children. They were not able to reproduce the results, in fact they reported that breastfed GG children performed better by 5.8 IQ points. GG children actually showed the greatest difference.
And another study:

Martin NW, Benyamin B, Hansell NK, Montgomery GW, Martin NG, Wright MJ, Bates TC. Cognitive function in adolescence: testing for interactions between breast-feeding and FADS2 polymorphisms. J Am Acad Child Adolesc Psychiatry. 2011 Jan;50(1):55-62.e4. Epub 2010 Dec 3.
They also were unable to repeat the Caspi study – unadjusted data showed a benefit if breastfeeding across all genotypes although they also say that when adjusted for sex, SES, paternal and maternal education, and birth weight the significance disappeared – they support the hypothesis that the breastfeeding/IQ association “reflects variation in parental cognitive ability” – in any case there was no FADS2 GG effect.
Based on the follow-up studies there is no support for any personalised advice based on the FADS2 gene – it cannot be used as a reason not to breastfeed and the largest study actually showed increased benefits for IQ in the group that would be advised that not breastfeeding makes no difference
This same doubtful information is also delivered by 23andme: just one study and no mention of the two larger studies. It is though announced as a “preliminary research report”.
I am disappointed that 23andme have not caught the other 2 papers – do they not have an army of researchers keeping up to date? It’s an enormously emotive subject and thus very important, it suggests that they need to improve their literature update methods and I hope they will – try using the “alert me when this article is cited” feature common to almost all journals.
This is does not particularly look good for either of the companies, but there is a big difference – one is delivered by the healthcare practitioner as “useful information” to help you decide to not breastfeed, the other is presented as results of a study. Neither are justified but I would say that the potential for a harmful impact is increased in one of the situations rather than the other. Which has the potential for more harm? Even if you are of the opinion that both are equally harmful, it certainly does not support the claims that because DTC could be harmful it should be banned and all delivered through medical practitioners because that would protect the patient. My view is that DTC or MD is not the problem – transparency is.
I suppose FADS2 is an appropriate name at least
Caspi A, Williams B, Kim-Cohen J, Craig IW, Milne BJ, Poulton R, Schalkwyk LC, Taylor A, Werts H, & Moffitt TE (2007). Moderation of breastfeeding effects on the IQ by genetic variation in fatty acid metabolism. Proceedings of the National Academy of Sciences of the United States of America, 104 (47), 18860-5 PMID: 17984066

Monday, April 25, 2011

If You Want to Practice Evidence Based Medicine then… Use the Evidence

Travelling on a train to Scotland, all for the first time:

Clinical researcher says:  "Look the cows in Scotland are black!".
Physicist replies: "No. Some of the cows in Scotland are black"
Mathematician says:  "No, actually in Scotland there is at least one cow black on one side"
I’ve written about folic acid, homocysteine and various clinical trials of nutrients in the past. I just want to revisit it briefly because I think there are some seriously wrong messages emerging from the conclusions. For an excellent comment on evidence based recommendations see Colby Vorland.

Homocysteine (Hcy) is associated with CVD, may or may not be causal but the evidence is good (Wald et al, BMJ, 2002). Folic acid (Vitamin B9) in reasonable amounts will lower homocysteine  but clinical trials have been carried out, mostly showing not much benefit as far a CVD is concerned. A recent meta-analysis of 8 trials (total 37,485 individuals) seems to confirm this: “Dietary supplementation with folic acid to lower homocysteine levels had no significant effects within 5 years on cardiovascular events”.

BUT… all the studies were in ill people (mean age 65yrs), 7/8 with heart disease or stroke and 1 study on patients with kidney disease:

So a fair conclusion is that people who have already had a heart attack, stroke or some other sort of vascular disease may not benefit from homocysteine lowering – e.g. there was no reduction in second heart attacks, at least over the short term (mean 5 yrs).
Pity, but not so unexpected, after all vitamins are not medicines to be used when one becomes ill at 65 years old – vitamins are for life.
There are also several possibilities why Hcy lowering had apparently no effect on the patients, and also Wald et al point out that aspirin use may mask the effect of homocysteine lowering – it was interesting to see that they felt the need to underline what should be obvious:
“The fact that most of the trials included patients with pre-existing cardiovascular disease leaves open the possibility that folic acid has a useful role in the primary prevention of ischaemic heart disease, where aspirin is generally not used, but not in secondary prevention, where it is routine.”

So what do Clarke et al conclude? It’s starts off OK:
“The present meta-analysis has demonstrated that lowering homocysteine levels by an average of 25% (about 3 µmol/L) for an average of 5 years has no significant effect on the incidence of major vascular events during the scheduled treatment period.”
(They just missed out the aged ill patients bit)

They go on to discuss how observational and epidemiological studies associate homocysteine with cardiovascular disease and stroke (but do not point out that these studies were in the general population, not a diseased subset) and finally:
“One-third of adults in the United States and one-quarter of those in the United Kingdom report taking daily multivitamin supplements containing folic acid. The present meta-analysis, however, found no evidence that routine use of folic acid for 5 years has any material effect on cardiovascular or non cardiovascular events in the North American and European populations studied.”

These are two sentences from different contexts entirely – the results of the meta-analysis apply only to the study group and have NO bearing on the folic acid and homocysteine as far as primary prevention is concerned. The reviewers really should not have let that pass.

There is also a commentary on this article in the same issue which goes even further in the wrong direction. It begins discussing the role of vitamins in health and prevention then discusses the impact of the meta-analysis:
  • Clarke and colleagues definitively answers most of the questions about the efficacy of B-vitamin therapy.
  • There were no trends toward benefit  in any of the primary outcomes overall or in the prespecified subgroups
  • This is a cautionary tale for all who are searching for therapies to prevent disease in the general population.
But as the commentary says there is a “strong scientific rationale” for Hcy lowering and “consistent observational data from high quality cohort studies with more than 10 years of prospective follow-up confirmed the outcomes suggested by hypotheses arising from laboratory and animal data.”
This all remains true and has not been touched by any of the clinical trials but the author goes on:
“However, randomized trials of vitamin therapy with folate, vitamin B6, vitamin B12…demonstrated that none of these therapies are effective for preventing cardiovascular disease…in the general population.”
This is simply not true, the trials did not study the general population and have NO relevance at all for primary prevention. There’s more:
“As trusted counselors with an ethical mandate to first do no harm, we physicians should keep this history in mind as we advise our patients about the risks and benefits of dietary supplements… Vitamins taken in excess of the dose required to prevent deficiency states have not improved  not improved our patients’ health and may harm them. We should recommend therapies to prevent disease in healthy patients only when randomized trials unequivocally demonstrate that net benefits outweigh net harms, and we should continue to emphasize the importance of a nutritious diet, regular physical activity, and no smoking as the best ways to optimize health.”

Couple of things: a) to not recommend is also a recommendation and b) how does the author know about the benefits of physical activity & nutritious diet and possible harms of smoking? All the evidence is from epidemiology and observation – they have not been confirmed by randomized trials!
Maybe the cows in Scotland are black.

Clarke R, Halsey J, Lewington S, Lonn E, Armitage J, Manson JE, Bønaa KH, Spence JD, Nygård O, Jamison R, Gaziano JM, Guarino P, Bennett D, Mir F, Peto R, Collins R, & B-Vitamin Treatment Trialists' Collaboration (2010). Effects of lowering homocysteine levels with B vitamins on cardiovascular disease, cancer, and cause-specific mortality: Meta-analysis of 8 randomized trials involving 37 485 individuals. Archives of internal medicine, 170 (18), 1622-31 PMID: 20937919

Monday, January 17, 2011

Breast feeding: Personal genetics is in it’s infancy…

There is a lot in the new today about breast feeding – should it be exclusive for at least 6 months (WHO) or should it include some earlier introduction of solid foods (BMJ article)? A lot of the reporting is confusing the issue, claiming or implying that the BMJ authors are saying the 6 months is too long (and pointing out that they have been consultants to the baby food industry). The BMJ authors are not saying that, they simply say that it might be prudent to introduce gradually some solid foods, while still breast feeding before 6 months.

There is a useful account at the Marion Nestle blog and I left a comment there which I have expanded upon here. This could be an area where genetic testing can help – blanket advice is never the best for all but confusion (as we have now) is even worse. I blogged about celiac disease and genetic testing a while ago and part of the post addressed exactly the issue of breastfeeding and weaning.

...babies differ! Probably, very likely, one type of advice is not applicable to every situation, but we don't yet have much idea of how to select.

The choice here is not breastfeeding vs. infant formula, it should be made very clear that the evidence suggesting adding solid foods before 6 months does NOT also suggest stopping breastfeeding. Precisely the opposite. The ESPGHAN (European Society for Pediatric Gastroenterology, Hepatology, and Nutrition) issued some precise advice targeted especially at infants with possible celiac disease risk (i.e. a genetic predisposition): For further details see

It is prudent to avoid both early (<4 months) and late (>or=7 months) introduction of gluten, and to introduce gluten gradually while the infant is still breast-fed, inasmuch as this may reduce the risk of celiac disease, type 1 diabetes mellitus, and wheat allergy.

Using genetics to give targeted advice is in it's infancy (excuse me!) - but could be useful in this case. If the universal advice were to suggest introducing solid foods between 4-6 months, there is a danger that it could lead to more frequent introduction actually before 4 months, which according to the evidence is even more risky for allergies and autoimmune diseases. So in this perspective the WHO advice could be considered to be wise, but there are exceptions.

I’m not so sure I understand the SACN logic though:

Currently available evidence on the timing of introduction of gluten into the infant diet and subsequent risk of coeliac disease and T1DM is insufficient to support recommendations about the appropriate timing of introduction of gluten into the infant diet for either the general population or high-risk sub-populations. However, there is evidence suggesting that not being breastfed at the time when gluten is introduced into the diet is associated with an increased risk of subsequently developing coeliac disease.

Source: Draft SACN/COT Statement on the Timing of Introduction of Gluten into the Infant’s Diet

I understand why earlier introduction of gluten may not be the best advice for the general population, but for high-risk sub-populations, I’m not so sure. The evidence is not water tight but that which does exist has been reproducible and indicates that 4-6 month introduction of gluten (while still breastfeeding) is the lower risk option. There are no studies that contradict this and none that suggest that there are more benefits than risks by waiting until after 6 months in high risk subpopulations. There are some clinical trials underway (e.g. PreventCD) which may help, but until then, the benefits and risks calculation suggests following the ESPGHAN advice, see above.