Smoke or $h!t gets in your eye!

A recent study was brought to my attention by a great reader. I highlight it here, not because it is going to change personalized medicine, but because it illustrates some key points.

The study is entitled: "Responses to Online GSTM1 Genetic Test Results among Smokers Related to Patients with Lung Cancer: A Pilot Study" Cancer Epidemiol Biomarkers Prev 2009;18(7). July 2009

This study interested me for several reasons, the first of these was that it involved a hotly debated environmental detoxification gene which has been tested for by numerous nutrigenetics companies for years now.

The second reason it interested me is because it dealt with a population who knows that there family can get cancer when exposed to smoking. I have always wondered for years why there are these families out there who have multiple members with lung cancer, yet everyone in the family seems to keep on puffing.

Are these families full of ignorant people who can't put 2 and 2 together despite years of public health campaigns? Are the genetically or environmentally predisposed to smoking? Does the family have some weird death wish set of genes?

I have always wondered why they do what they do. This study evaluates precisely these families....

It was published on GenomeWeb and on several blogs as well as several news outlets with titles like:
1. Online Genetic Testing Appears to have Benefits
2. Study Suggests Online Genetic Test May help Smokers Quit
3. Possible Benefit From Online Genetic Testing For Lung Cancer

What they did: They identified relatives of Patients with stage IIIB/IV lung cancer who were receiving care in the Thoracic Oncology Clinic at the H. Lee Moffitt Cancer Center and Research Institute were identified. So in essence, patient who have pretty bad cancer, their relatives....who likely know that it is a bad cancer.

They tested the relatives for GSTM1 genotypes. Which is pretty interesting to me, because as the study admits, the Odds Ratio for Lung cancer in this null population is 1.15 to 1.17. essentially as close to one as you can get......

Which means that this association is pretty freaking weak......

So in essence they are going to clinically LIE to the person tested and tell them they had a higher risk.

Oh and BTW, they only analyzed 44 people.....Hardly a useful sample size, especially when 96% were white.....and 22% were college educated, hardly the 23andMx population......

What they found.
1. Smokers with the at risk profile GSTM1 nulls, remembered at 6 months that they had a "higher" risk.
2. Smokers with the normal risk, forgot that more often.....roughly 50% remembered that they were at "lower" risk
3. Both parties "believed" the test results equally........
4. There were NO SIGNIFICANT DIFFERENCES in uptake of the smoking cessation services between those who received the GSTM1-present and GSTM1-missing test result.......
5. At the 6-month follow-up, the proportion of smokers in each group who reported medication use did not differ significantly.
6. Perceived risk for cancer was the SAME in both groups
7. There were no significant differences over time in confidence in ability to quit smoking between the GSTM1-present and GSTM1-missing relative smokers.


"This observational study was not sufficiently powered for nor was it a study aim to assess smoking cessation as an outcome."

Obviously as it only had 44 participants in it........ So I am freaking left asking myself "Where in the HELL is the benefit?"

There are a ton of limitations which prevent this article from even being a study. It is an observation of a very, very small cohort......I consider this as statistical NOISE.........

But, the authors try to save themselves by stating "However, these limitations need to be balanced against the strengths, which include this study being the first, to our knowledge, to offer genetic testing for a common gene variant online."

Are you serious? Strengths? Maybe Strength, but I think that there is other research currently going on involving online genetic testing.

Maybe not published, but there is ongoing......

Some important points that I take from this paper........
1. DTC Genomics has a huge PR machine that will trick papers into writing bull$hit headlines and false claims....I.E. you cannot trust a single thing published in the lay press about genetic studies EVER!!!!!

2. How does a study with 44 people get noteworthy accolades for proving that essentially a single gene variant, weak association, does nothing in terms of empowering patients.....IFF you accept that this study wasn't just noise......

3. The DTC companies offer THOUSANDS of these little polymorphisms......if the people with normal results can't remember that they are at "REDUCED" risk from ONE, how can we expect people to remember THOUSANDS.....BTW, this is essentially like telling a patient that if they eat a ONE cheeseburger from McDonalds on Tuesday July 16th that they are at risk of a heart attack in the next 5 years......Statistical Bull$hit with ZERO Veracity

4. People will cling to the tiny shred of suspect evidence of the 6 of 44 who quit smoking.......% had the GSTM1-Null genotype, which if you look at statistics is likely pure chance as 50% of the pop is NULL.......Hmm whats the odds of 5/6 being null....I can't believe they even mentioned this in their paper.....What a crock!!!!!

The Sherpa Says: Just because it is published in a journal doesn't make it Good Science. Hell, it doesn't always make it Science at all.......Shame on the press and the bloggers who hyped this piece of garbage.......