Tuesday, June 14, 2005

Racist Drug Development.

UPDATE (22 June 2005): Nature has a news item on this drug with some good insights from Francis Collins. Collins suggests, "We should move without delay from blurry and potentially misleading surrogates for drug response, such as race, to the more specific causes."

An important point that needs to be brought up is that race is a somewhat subjective classification. Someone who identifies himself as African-American may be of less than 25% African ancestry. While there is real genetics behind the differences between Africans, Europeans, and Asians, sometimes ethnicity seems to overpower those differences in a cultural setting. Do we prescribe a drug that only seems to work in people of African ancestry to a patient who identifies himself as African-American even if he's mostly of European ancestry?

I know razib will bring this up, so I will point out that the majority of African-Americans are mostly of African ancestry.



African-Americans are the result of admixture between Africans, Europeans, and Native Americans. Furthermore, Africans are a paraphyletic group containing more genetic diversity than all other human populations combined. That's why this bothers me a bit:

This Thursday, eight years after the drug [BiDil] was rejected for use in the general public, an F.D.A. panel will consider whether BiDil should become the first drug intended for one racial group, in this case, African-Americans.

A study of 1,050 African-American heart failure patients showed that BiDil significantly reduced death and hospitalization, prompting the American Heart Association to call BiDil one of the top developments of 2004. BiDil increases levels of nitric oxide, which widens blood vessels.

The drug's maker, NitroMed Inc., says its decision to test and market BiDil as a drug for African-Americans is based on solid science. But BiDil's application has engendered controversy, with many scientists convinced that race is too broad and ill-defined a category to be relevant in determining a drug's approval, especially since geneticists have failed to identify a biological divide separating one race from another.

. . .

"My criticism of the African-American Heart Failure Study is that they only studied African-Americans," he [Dr. Joshua Hare, a cardiologist at the Johns Hopkins University Medical Center] said. "To really test the hypothesis is to study both populations and then show, aha, the African-Americans did respond better. They didn't do that."

. . .

Jonathan Kahn, a medical ethicist at Hamline University law school in St. Paul, said BiDil's approval as a black-only drug would give an official ring to the discredited idea that race is a biological category.

The scientist in me says that there is very little reason to suspect a drug would work differently in African-Americans than in European-Americans because of genetic difference between the two ethnic groups. There may be some environmental differences, however, that cause the drug to perform differently in the two populations.

The commie pinko bastard in me notes the "historical inequities in medical treatment for African-Americans" and wonders if this is an example of social justice.

5 Comments:

At 11:13 PM, Blogger Razib said...

1) african americans are admixed, but the evidence from various sourcies (both scientific and historical) seems to suggest that ~80% of their alleles derive from african sources (and one must remember that these alleles are not equally distributed by skewed toward higher and lower concentrations depending on local subpopulations, so that blacks in the islands off the coast of georgia are almost "pure" while those in the pacific northwest maybe mostly white genetically), and more specifically, usually west african sources (though a minority likely derive from the kingdom of kongo region and perhaps even mozambique).

2) the diversity of africans as regards many regions of their genome (NRY or mtDNA for example) does not imply that the loci which might have relevance to the medication in question are particularly diverse. as an example, the MC1R loci are functionally constrained in africans and far less diverse than in europeans or east asians.

3) a lot of this is probably politics at work.

 
At 3:20 AM, Blogger Peter Ellis said...

Further to razib's comment, it's extremely unlikely that the african-american rootstock was a random sampling of the african genetic stock.

In broad terms, I guess one good indicator is that if the disease itself appears to be ethnically skewed once you've standardised for environmental factors (don't know if that's the case for heart disease, but consider Tay-Sachs, Gaucher, sickle cell...), then it's eminently plausible that treatment response will also be skewed.

 
At 10:25 PM, Anonymous Jason Malloy said...

The scientist in me says that there is very little reason to suspect a drug would work differently in African-Americans than in European-Americans because of genetic difference between the two ethnic groups.

If interpreted as this statement is clearly intended to be interpreted you are claiming it as some sort of scientific fact that African Americans and European Americans (and all ethnically defined groups) are genetically undifferentiated. This is not credible.

The reason you should suspect some drug-responses to differ between ethnic groups is the very same reason you should expect drug responses to differ between individuals - because genetic differences have phenotypic consequences. This is not just an effluvial bit of bar trivia, much less a baseless speculation, but an essential fact of evolution.

Please don't claim to be speaking as a scientist if you are going to make erroneous assertions.

 
At 10:00 AM, Blogger RPM said...

Please don't claim to be speaking as a scientist if you are going to make erroneous assertions.

As I mentioned at the beginning of my post, African-Americans are an admixed group. This means they are are a mix of European, African, and Native American ancestry. Altough razib pointed out that African Americans may not be as admixed as I originally thought, I still think there are limitations in basing drug development on ethnic classifications rather than phylogenetic differences.

There are genetic differences between ethnic groups, but truly admixed groups are not genetically unique because they have very few unique alleles. Admixed groups do have a unique combination of alleles, which may be important.

 
At 5:49 PM, Blogger Razib said...

I still think there are limitations in basing drug development on ethnic classifications rather than phylogenetic differences.

i think it depends on what you mean by "basing" drug development. i would be skeptical of drugs developed a priori for X ethnic group. but, i think it is cool double check if a drug ends up having different responses in different groups.

 

Post a Comment

<< Home