pathtalk.org is a weblog about pathology and laboratory medicine.

Testing-testing: Two SNPs for $200

The ASCP daily newsletter dated 10/12/07 had a short paragraph about home kit DNA diagnostics. Risk assessment for tailored treatment plans is a marketing strategy supporting the use of DIY (do-it-yourself) DNA swabs fishing for SNPs (single nucleotide polymorphisms) deemed as high risk in some common diseases such as atrial fibrillation. The current prevalence of atrial fibrillation is estimated to be 2.3 million with a projected increase by 2.5 fold to more than 5.6 million by 2050 due to the growing proportion of patients 65 years and older¹. Possible contributing factors or causes for atrial fibrillation are as follows²:

• High blood pressure
• Heart attacks
• Abnormal heart valves
• Congenital heart defects
• Stress due to pneumonia
• Sick sinus syndrome
• Etc

When atrial fibrillation occurs in patients without these risk factors, lone atrial fibrillation is in the consideration. The stroke risk for patients with atrial fibrillation increases per year and is as follows (stroke rate percent per patient year)³:

• 1.3 in those aged 50 to 59 years
• 2.2 in those aged 60 to 69 years
• 4.2 for those aged 70 to 79 years
• 5.1 for those aged 80 to 89 years

One company offers a home do it yourself DNA diagnostics kit for risk assessment for myocardial infarction, atrial fibrillation, and diabetes. The kit costs $200 and uses DNA derived from buccal swabs to detect 2 single nucleotide polymorphisms on chromosome 4q25 that the company has identified as major risk factors for atrial fibrillation⁴. The company conducted a genome wide analysis of 300,000 Icelandic patients with atrial fibrillation, and found 2 SNPs associated with increased atrial fibrillation on the PITX2 gene on chromosome 4q25⁴. The results were published in Nature⁵. The test was also validated against Swedish patients and patients from MGH and is also found to be prevalent in the Han Chinese population⁴. If identified, the SNPs are associated with a 1.8 fold increase in atrial fibrillation. The value of the test as reported by the company is identifying patients who are at risk so that diagnostic tests may be utilized more effectively, and the onus is on the physician directing care for the patient to determine how to utilize this information in the treatment plan. Several polymorphisms have been reported as being associated with atrial fibrillation^6,7,8.

Therefore, the significance of the results of this type of DIY test is baffling to me. The original studies were conducted and validated on a relatively homogeneous population. Currently, it is still believed that atrial fibrillation is a heterogenous disease with several contributing factors. Genetics is probably a major factor. So would a negative test result put the average non-Swedish, non-Icelandic, Non-Han Chinese citizen and the doctor treating that patient at ease? Would a positive test result at some point dictate insurance eligibility and premiums?

If you are “lucky” enough to read Bloomberg. com on the right day⁹, you would find out that a company can offer you a DNA test for myocardial infarction risk for a mere $200. If positive, the risk of MI is 1.6 (the risk for smokers is at least 2-fold higher). The population for this study was slightly more diverse (Ottawa, Texas, Denmark, Great Britain).

References
1. JAMA. 2001 May 9;285(18):2370-5
2. http://www.mayoclinic.com/health/ atrial-fibrillation/DS00291/DSECTION=3
3. http://patients.uptodate.com/topic.asp?file=carrhyth/50724
4. http://www.decodediagnostics.com/AF-genetics.php
5. Nature Volume 448(7151), 19 July 2007, pp 353-3576.
6. Heart Rhythm. 2007 Apr;4(4):469-75. Epub 2006 Dec 157.
7. Circulation. 2003 Jun 17;107(23):2880-3. Epub 2003 Jun 28.
8. N Engl J Med. 1997 Mar 27;336(13):905-119.
9. http://www.bloomberg.com/apps/news? pid=newsarchive&sid=aBpFp6D4Q9BE

A big problem in the near future is vast proliferation in the number of SNP association ‘diagnostic’ or ‘prognostic’ tests. This is a fine example. What does one do with a 1.8 fold increased risk. Is this meant to be a screening test to save on the cost of actually having people go to the doctor and get a physical exam (not that!)? If so, what is the positive and negative predictive value? Sensitivity? Specificity? Would using this test impact patient outcome in any way? Let’s face it this is the bottom line (aside from the actual bottom line).

This story reminds me of the FDAs recent recommendation to test certain polymorphisms in VKORC and CYP2c9 in patients who are to be put on warfarin. Unfortunately, although there is loads of data to show that people with certain polymorphisms are more prone to negative events and require either more or less warfarin to achieve a steady state dose, there is no data to prove that testing for these polymorphisms actually impacts patient outcome. An editorial for another day…

Yes this is exactly the focus of the post– i am wondering if this will be a shift for us all not only for common medical disease but extending to solid tumors etc. we already see this in some manner in leukemia/lymphomas where some places have focused FISH panels for prognostication for LPD/MPD

this warfarin sensitivity is a very real step into an era of personalized pharmacogenomics —

-

Gretchen – I view pathology as the last line of defense against inappropriate testing, and the champions of best practices when it comes to laboratory medicine. Raising awareness (publishing) and educating your clinician colleagues on best practices when it comes to new technologies and discoveries is our responsibility. Unfortunately, we are often up against very persuasive marking forces… Certainly, there are a wealth of very good molecular tests that have dramatically impacted patient care. We just need to keep the bar high.

Concerning warfarin – a prospective study showing that a patient population that is dosed based on CYP2c9 and VKORC polymorphisms does better than a patient population dosed by best clinical judgement would go a long way towards appeasing the skeptics.

Several key pieces of data were not discussed in the current studies which would also add weight to the utility of SNP testing in this context. For example, what was the dose of warfarin when individuals had their negative event, what was their INR at that time, what was their INR and dose at their previous visit. etc.