Genetics, race, ethnicity, and healthBMJ 2004; 328 doi: https://doi.org/10.1136/bmj.328.7447.1070 (Published 29 April 2004) Cite this as: BMJ 2004;328:1070
All rapid responses
In this article, Pearce et al draw confuse two separate issues. They
state genetic variation is predominantly intra-racial rather than inter-
racial. This is intuitive given common ancestry and the extent of
population admixture over centuries.
They then summarise “genetic factors are important for health but are
a small part of a large and complex picture” and state “benefits are a
long way off and require large investments with potential benefits for a
few high risk individuals (and researchers)”. This unsupported stance
ignores available evidence.
For example, smoking behaviour clearly has strong environmental
influences, but it is heavily dependent on genetics. A meta-analysis of
six large twin studies estimates the heritability (h2) of smoking
initiation at 55%. These findings are reproducible regardless of the sex,
age, or ethnicity of twins, or whether they were reared together or apart.
This evidence is reinforced by that from animal studies: Strains of inbred
mice differ in their sensitivity to the behavioural and physiological
effects of nicotine, including development of tolerance. Public Health
measures to curb smoking are of limited efficacy – should we be content
with this or strive to better understand why people smoke?
All complementary aspects of research should be encouraged – a
disapproving three page article on the value of Professors of Maori
Studies would rightly never have reached your pages.
1) Li M. et al A meta-analysis of estimated genetic and environmental
effects on smoking behavior in male and female adult twins. Addiction
2) Carmelli D. SGE, Robinette D., Fabsitz R.,. Genetic influence on
smoking--a study of male twins. N Engl J Med 1992;327(12):829-833.
3). Hannah M. C., Hopper JL, Mathews JD. Twin concordance for a binary
trait. II. Nested analysis of ever-smoking and ex-smoking traits and
unnested analysis of a "committed-smoking" trait. Am J Hum Genet
4) True WR, Heath AC, Scherrer JF, Waterman B, Goldberg J, Lin N, et al.
Genetic and environmental contributions to smoking. Addiction
5) Kendler K. S., Thornton LM, Pedersen NL. Tobacco consumption in Swedish
twins reared apart and reared together. Arch Gen Psychiatry 2000;57(9):886
6) Collins AC, Miner LL, Marks MJ. Genetic influences on acute responses
to nicotine and nicotine tolerance in the mouse. Pharmacol Biochem Behav
7) Robinson SF, Marks MJ, Collins AC. Inbred mouse strains vary in oral
self-selection of nicotine. Psychopharmacology (Berl) 1996;124(4):332-9.
8) Marks MJ, Stitzel JA, Collins AC. Genetic influences on nicotine
responses. Pharmacol Biochem Behav 1989;33(3):667-78.
9) Batra V., Patkar AA, Berrettini WH, Weinstein SP, Leone FT. The genetic
determinants of smoking. Chest 2003;123(5):1730-9.
Competing interests: No competing interests
A Response to Pearce et al. (2004): BMJ 328: 1070-2
Pearce et al. (2004)1 provide a valuable commentary on “Genetics,
race, ethnicity and health” in the 1 May edition of BMJ. Their arguments
hinge on the observations of Dick Lewontin2 and others regarding the
fractional distribution of genetic variation between individuals and
between ethnic groups, being much greater in the former than the latter
comparison. They go on to provide many instructive examples of
environmental influences on the health of Maori and Pacific Island people
living in New Zealand. However, they stereotype other researchers as
genetic chauvinists quite unfairly by implying they are amongst those who
believe “that the genotype [exclusively] determines the phenotype” and
further of their being guilty of extending concept this to explain
differences in disease prevalence between ethnic groups.
Rather, in my own experience human molecular geneticists take the
balanced view that genes and environment and the various interactions
between them are all important factors. Thus particular genetic variants
with roles in disease susceptibility may well differ in frequency between
ethnic groups, indeed that they are expected to do so. This “confusion of
questions “ and the tricky task of finding an adequate definition of human
races has been very well explained recently by Feldman et al. (2003)3.
Pearce et al. (2004) quote our preliminary report4 on ADH 2*2
frequencies in Maori and European subjects as an example of the failure of
genetics to explain comparative rates of alcoholism in New Zealand. It
seems a pity that they did not consult our later full report, Chambers et
al. (2002)5 before writing their commentary. There we show clearly that
both Maori and Europeans enjoy the same protective effect of ADH 2*2, but
that Maori benefit more since this allele is more common in Polynesians
(frequency = 0.46) than Caucasians (frequency 0.03). My own inference is
that, although these alleles may in part determine relative susceptibility
of individuals within ethnic groups, they do not predict relative
susceptibility between members of different ethnic groups.
Indeed, quite the opposite is the case. Maori and Polynesians
subjects in New Zealand show a higher frequency of alcohol abuse behaviour
despite the prevalence of a protective genetic factor. It is widely
acknowledged that rates of alcoholism are set by the socio-cultural
milieu. For Maori these include higher overall rates of alcohol
consumption per drinking session set against lower frequency of
consumption and a higher prevalence of abstainers6. Thus it is my well-
established view that all human behavioural disorders are complex
phenomena with some greater or lesser genetic basis. This is quite
different from the rather disappointingly prejudiced picture drawn by
Pearce et al. (2004).
1. Pearce, N., Foliaki, S., Sporle, A. and Cunnigham, C. (2004)
Genetics, race, ethnicity and health. Brit. Med. J. 328: 1070-1072.
2. Lewontin R. C. (1991) The doctrine of DNA Penguin Books, London.
3. Feldman M. W., Lewontin R. C. and King M.-C. (2003) A genetic melting
melting-pot. Nature 424: 374..
4. Marshall, S.J., O'Brien, D., Wheeler, K., Hermans, I.F., Chambers,
G.K., Robinson, G.M. and Stace, N. (1994). Genes and alcoholism: A
preliminary report. NZ Med. J. 107: 106-107.
5. Chambers, G.K., Marshall, S.J., Robinson, G.M., Maguire, S.M., Newton-
Howes, J. and Chong, N.J. (2002). The genetics of alcoholism in
Polynesians: Alcohol and aldehyde dehydrogenase genotypes in young men.
Alcoholism Clin. Exp. Res. 26: 949-955.
6. Bramley, D., Broad, J. Harris, R. Reid, P. and Jackson, R (2003)
Differences in patterns of alcohol consumption between Maori and non-Maori
in Aotearoa (New Zealand). New Zealand Med. J. 116: 645-651.
Competing interests: No competing interests
Pearce and associates (2004) have best attempted to explain racial
differences, ethnic diversities and health perspectives more by
environmental operating factors than by genetic mechanisms. They may be
correct but we would like to emphasize that the genetic make-up of an
determines the environmental changes and likewise environment modifies the
genetic endowment of an individual across life cycle. By and large, there
is a constant interaction between genes of an indivdual and the
environment. The resultant effect may definitely impact race, ethnicity
and the health of the human beings. Evidently, we know very little about
genetic complexities but we know a lot about environmental factors
affecting race, ethnicity and health. However, we are beginning to know
more about genetics, which will possibly change the explanatory model of
health, race and ethnicity from environment to genetic one in future.
Surely, time will tell the truth whether envornmental model will continue
to dominate the scene or genetic model will surpass environmental model in
substantially explaining racial, ethnic and health differences across
national and international boundries.
Neil Pearce, Sunia Foliaki, Andrew Sporle, and Chris Cunningham.
Genetics, race, ethnicity, and health
BMJ 2004; 328: 1070-1072
Competing interests: No competing interests
The sequencing of human genome has brought with it, hopes of complete
cure. Though extremely complex and mostly un-understood, more than 99% of
the human genome is same in all human beings (Science 1997; 278:1580). The
less than 1% variation in the three gigabases of DNA sequence is due to
about one and a half million single nucleotide polymorphisms (Nature 2001;
409:928) many of which are known to be involved in the causation of
different diseases, while the effect of others is undiscovered and so is
the etiology of a number of diseases.
It is encouraging to see that the genetic basis of different
disorders that follow Mendelian inheritance has been elucidated and their
cure can be a possibility, when gene therapy comes into play. However, the
largest chunk of human morbidity and mortality is accounted for by complex
disorders like hypertension, diabetes and atherosclerosis that do not have
a clear genetic basis.
These disorders have a relatively weak genetic component that
involves multiple genes with none of the loci having a strong penetrance.
In the last few decades, linkage analysis and association studies have
shown that the involvement of different parts of the genome and different
SNPs in the etiology of such disorders varies from one population to
Association studies have had variable results in different
populations, with each population having a different proportion of allele
frequencies, which shows that complex disorders have a diverse genetic
component. There are more than 700 alleles of LDL receptor gene that have
been associated with increased risk of coronary artery disease and about
480 alleles of BRCA1 gene that are associated with a higher risk of
familial breast and ovarian cancer, just to name a few (Online Mendelian
Inheritance in Man). Our team has worked on the genotype-phenotype
association studies of the candidate genes of hypertension. It has been a
unique experience to compare the variety of results seen in the Caucasian,
Middle Eastern and South Asian populations. As seen in the past, we had
mixed results in all three populations but interestingly, all those
candidate gene loci that had a positive association with high blood
pressure in the Middle Eastern population showed a highly significant
level of association. We believe that Middle Eastern and South Asian
populations are genetically distinct and conserved groups. High degree of
consanguinity and the availability of large, up to three generation
pedigrees make these groups ideal for genetic studies. Linkage analysis
followed by sib-pair linkage can be helpful in identifying putative causal
relationship of involved genetic loci.
Gene therapy has not taken any practical clinical form yet, but when
it does, it is going to rely on genetic data that is specific for every
population. It might be feasible to treat a complex disorder by targeting
therapy on a particular focus of genetic change that is more common or has
higher penetrance in a population. Baseline knowledge of the genetic
architecture of a population will have a definite importance in the
determination of pharmacological interventions as well. Hence to set the
stage for gene therapy, it is important to know the differences in genetic
structure that exist in different ethnic or geographical groups.
Competing interests: No competing interests
It is my hypothesis that changes in testosterone levels are directly
involved in human evolution and may be identified in modern populations as
the “secular trend.” (“Androgens in Human Evolution,” Rivista di Biologia
/ Biology Forum 2001; 94: 345-362). Changes in testosterone occur more
rapidly than changes in genes. Therefore, the “races” may reflect
different testosterone levels that affect growth and development as well
as maintenance of the adult form. While the data are not complete, as an
example that may produce useful applications, consider that black women
produce more testosterone than white women and, in one study, black males
produce significantly more testosterone than white males. Testosterone
levels may produce differences that significantly impact health. I
suggest the effects of maternal testosterone levels on the fetus are the
driving force of testosterone on populations.
Pearce, et al., considered possible explanations of race and
essentially dismissed race as a factor affecting health. Therefore, race
is “on the table.” Lets examine the most often cited, and most often
dismissed, difference, skin color. It is my hypothesis that skin color is
affected by levels of testosterone. This is also a trait that originates
in utero. It is known that testosterone and ultraviolet light work
together in stimulating melanocyte structure and function. "Cultured skin
receiving both UVL [ultraviolet light] and testosterone illustrates a
synergistic effect." (J Exp Zoo 1978; 204: 229). This could account for
the findings of Jablonski and Chaplin that "In all populations for which
skin reflectance data were available for males and females, females were
found to be lighter skinned than males." (J Human Evolution 2000; 39: 57-
106) Testosterone is definitely involved in melanocyte function.
Healthy black males produce significantly more testosterone than
healthy white males (J Nat Cancer Instit 1986; 76: 421). Melanocyes from
black males grow differently from melanocytes from white males, in
culture. In one study, melanocytes, derived from neonatal foreskins, an
area of skin directly affected by testosterone that has not been
significantly exposed to sunlight exhibit differences. "At the
ultrastructural level, cultured melanocytes derived from black (negroid)
neonatal skin (B-M) had numerous mature rod-shaped stage IV melanosomes,
while white (caucasoid) skin-derived melanocytes (W-M) in culture
contained no mature melanosomes. Growth rate, cell yield, and in vitro
lifespan for B-M were more than twice that for W-M in pure melanocyte
cultures in the presence of MGF [melanocyte growth factor]. Our results
suggest that MGF-dependent growth of B-M differs from that of W-M." (J
Cell Physiol 1988; 135: 262-8). Melanocytes grown in culture, without
testosterone added to the culture media, inherently express a difference
in growth potential between black and white males.
Melanocytes from neonates already exhibit differences in growth rate
according to race. "Differences in size and number of melanosomes
attributable to race of the tissue donor were readily apparent, and
pigment content of melanocytes from both black and Caucasian donors
appeared to increase with time in culture. Newborn melanocytes
proliferated more rapidly and survived longer than did adult melanocytes,
but there were no consistent morphologic differences as a function of
donor age." (J Invest Dermatol 1984; 83: 370-6).
I think the effects of testosterone on melanocytes first occurs in
utero. In utero, black fetuses are exposed to higher levels of
testosterone. “Black mothers had higher androstenedione and testosterone
concentrations than white mothers.” (Cancer Causes Control 2003; 14: 347-
55) "Serum testosterone was modestly, but significantly, greater in the
black than in the white women." (J Clin Endocrinol Metab 1996; 81: 1023-6.
I suggest melanocytes are stimulated by increased testosterone during
gestation in blacks.
According to my explanation of human evolution, lower testosterone
groups migrated away from the equator. Therefore, lower levels of
testosterone contribute to lighter skin in groups living away from the
equator. This racial difference, skin pigmentation, may be directly
influenced by testosterone levels.
Testosterone may be connected with numerous pathologies.
Testosterone is known to reduce the response of the immune system. In
fact, this may explain the increased levels of HIV infection in blacks,
especially males, compared to whites. Testosterone, in rats, reduces CD4+
T cells and reduces the CD4+/CD8+ ratio as well stimulates as an increase
in CD8+ T cells, which is diagnostic of AIDS in humans (Int
Immunopharmacol 2003; 3: 1853-60). The same pattern exists in black
compared to white children in “infected and uninfected children born to
HIV-infected women.” The pattern exists whether or not HIV infection has
occurred. It is inherent; the HIV more readily infects those with a more
vulnerable immune system and then exaggerates the difference.
“This study investigated whether age-related patterns of immunologic
markers in 1488 uninfected (9789 measurements) and 186 infected (3414
measurements) children differed by gender and race. CD4+, CD8+, and
absolute lymphocytes by HIV infection status, gender, and race were
assessed using linear mixed-effects natural cubic spline models, allowing
for prematurity and maternal CD4+ cell count. In uninfected children,
levels of all 3 markers peaked twice in the first few months of life,
declining to adult levels by around 8 years of age; uninfected boys and
uninfected black children had significantly reduced CD4+ and absolute
lymphocyte counts; the gender difference was especially pronounced in
black children. Infected children had substantially lower levels and
distinctly different patterns; with, e.g., by age 6 months CD4+ cell
counts nearly 1200 per mm3 lower than in uninfected infants. Levels also
significantly differed by gender and race for infected children, although
for gender in the opposite direction. The gender and race differences in
CD4+ levels were not explained by a general lymphocytosis nor were they
confounded by treatment. These substantial differences in immunologic
markers may reflect underlying genetic influence on the cellular immune
system and may have implications for clinical decisions about therapeutic
management.” (J Acquir Immune Defic Syndr 2003; 33: 635-41)
Preeclampsia is definitely connected with testosterone and is
increasing within our populations. (I suggest the secular trend is
actually an increase in the percentage of individuals of higher
testosterone. As they increase, we see their characteristic also
increase. This is human evolution identifiable within modern
populations.) A number of reports connect high levels of testosterone
directly to preeclampsia: "Levels of the potent androgen testosterone were
significantly higher in primigravid women with preeclampsia than in
normotensive women with similar gestational and maternal ages. This
difference may indicate a role for testosterone in the pathogenesis of
preeclampsia." (Acromite, M.T., et al., "Androgens in preeclampsia,"
American Journal of Obstetrics and Gynecology 1999; 180: 60-3) and. "A
history of preeclampsia an average of 17 yr earlier thus appears to be
associated with elevated levels of testosterone, which may contribute to
the increased risk of vascular morbidity in such women." (Laivuori, H., et
al., "Evidence of high circulating testosterone in women with prior
preeclampsia," Journal of Clinical Endocrinology and Metabolism 1998; 83:
Prematurity is also a very serious problem that is increasing.
Exposure to testosterone, during pregnancy, results in increased
probability of low birth weight. This has been tested. Prenatal
testosterone exposure reduced body weight of fetuses and newborn rats
(Arzneimittelforschung 1984; 34: 780). Another study in sheep found that
both fetal survival and growth were markedly impaired by prenatal
testosterone administration (Metabolism 1978; 27: 253). Even more ominous
is the finding that alcohol and testosterone, combined, induced low birth
weight in rats (Tetratology 1989; 40: 335).
Pearce, et al., cite the plight of North American Pima Indians as an
example of ill effects of “new white settlers” who imposed negative
environmental stresses upon the Pima population. While it is possible
that the diet imposed on this group “by the United States Government” may
have resulted “in one of the highest rates of prevalence of diabetes in
the world,” Pimas exhibit high testosterone levels (J Endocrinol Invest
1993; 16: 403-6)
Female breast cancer is also an increasing problem. While there is
not a lot of difference in incidence rates between white and black women,
black women present with much larger tumors. While this is attributed to
later exposure to diagnostic situations, it may also be due to the
increased testosterone in black women. It was reported in 2002 that
testosterone may be more directly connected to breast cancer risk than
estradiol. "…androstenedione and testosterone might be more strongly
associated with [breast cancer] risk than estradiol." (J Natl Cancer Inst
2002; 94: 606-616).
There are a number of other disorders that may also be linked to
testosterone levels. I suggest “race” may be directly connected to
testosterone levels. It is my hypothesis that testosterone levels are
involved in race and differences in health between races. History is full
of human malignancy; it occurs every day. We must attempt to identify
racial abuse and try to stop it when it occurs. We must also be careful
that we not find guilt where it does not exist.
Competing interests: No competing interests