Friday, April 17, 2009

Reading between the Genes: The NY Times Again

There were several articles in this weeks NEJM regarding the attempts to determine the genes that cause various diseases. Since I have been writing about the use of genetic techniques to reduce health care costs and change health care delivery, I thought that a brief note on this issue was worth the while. For the five documents, the 3 NEJM commentaries, the NEJM paper and the NY Times article show what I believe is an example of people talking past each other.

I have been arguing that genetic use for disease susceptibility or screening, staging, treatment and then prevention was to be a major prat of health care change in the next two decades. The Times states "Genes Show Limited Value in Predicting Disease". However the process that the Times is reporting on is the wide band assessment of genes and their propensity to predict a future disease state. Simply the results state that there are oftentimes dozens of genes, SNPs to be precise, which each may have a relative risk of 1.05 towards having the disease at some future time. The many SNPs may add in terms of increasing risk but how exactly they do it is yet known. Thus if say there were 20 genes which gave a propensity to a certain disease and the greatest relative risk was 1.05 and the others reduced that relative risk down to say 1.005, then if one had all 20 the cumulative relative risk is unknown, but greater than 1.05. Following this?

Well what does it mean? It means that many companies like 23 and Me may be selling somewhat useless and in some cases somewhat concerning tests since the results are hardly predictive. The Times kind of says that, but not really. However there are real genetic testing companies like Correlagen.

Correlagen tests for genes which are known to present a clear and present danger from a set of genetically based disorders and for which disorders one can take remedial action. This is meaningful. This is also totally ignored in the papers which appear to address the broad spectrum of all possible cases.

Secondly, it says that we yet fully understand the process. We understand many cancers much better. The now classic Bert Vogelstein model of colon cancer shows how multiple hits on cells in the colon result in cancer. We can use that information in a clinical staging analysis. We cannot screen with that since there is nothing there until the hit occurs. We cannot predict who will get hit by an auto before they are hit.

The multiple SNP model in the paper by Goldstein is of concern however. It addresses the Type 2 Diabetes issue. Again as we have argued from evidence that more than 95% of Type 2 Diabetes is caused by obesity and that when the weight is reduce to below a 22.5 BMI the blood sugar returns to normal and the HbA1c also returns to the 4.5 to 5.5 range. Namely evidence of Type 2 Diabetes goes away. Thus is there a set of genes that correlate with Type 2, and are these genes ones which are more related to eating than to the disease. The studies performed thus far are ambiguous at best.

Thus the Goldstein assertion is questionable. His assertion regarding height may be of merit but we at this point know little of height yet we do know that as childhood and maternal nutrition are improved height increases generation to generation. Is there a similar correlation between nutrition and genes and not between genes and height directly. These are the problems that the geneticists face and oftentimes failed to include.

The paper by Hirschorn is of interest since it does present the issue of pathways. Namely as we have shown in other species, it is the complex interaction of genes, activator and repressor genes, and their control of secondary metabolic pathways that results in may of the effects which we see. No single gene is the sine qua non. It is the network of genes and the secondary pathways that causes what we often see as a disease.

Thus in many ways the Times sensationalizes the story while missing the point. The point is that the problem is more complex than what the researchers had originally set out to resolve. Well that is the nature of research. The answers are most likely hidden in the complex genetic networks that are currently under analysis. Single genes are not the answer. Complex networks are. By stating as the Times does that Genes Show Limited Value is outright wrong. What is shown is that what has been done to date needs to be expanded not abandoned. This is the second time in a month that the Times having reported on the health care area has grossly mis-stated the implications. One wonders how one can rely upon such reporting for assessing other issues.