ARTICLE
In the opening of his paper on male pattern balding [1], Harry Harris
commented "although it is well known that baldness tends to run in families,
the exact manner of its inheritance has never been satisfactorily determined".
This state of affairs remains largely unaltered over 50 years later. In
essence, there have been two hypotheses to explain the inheritance of
male pattern balding. The first, proposed by Osborn in 1916 [2], is that
balding is a dominant trait determined by a single autosomal gene with
two alleles B (balding) and b (non-balding). She
suggested that balding would develop in both men and women who are homozygous
for B, and in men, but not women, who are heterozygous (Bb).
This view was broadly supported by Harris in an analysis of family histories
in 120 men in whom balding had commenced during their twenties [1]. However,
he was unable to ascribe late-onset balding to the same gene. The second
hypothesis is that balding is a polygenic trait. This has been suggested
by a number of authors, most prominently by Küster and Happle, in
a critique of the previously published data [3]. They proposed that, as
a polygenic trait, the predisposition to baldness is normally distributed
in the population and that the clinical expression of balding is a threshold
effect. As balding is androgen dependent the threshold will be higher
in women and they will therefore need more or stronger balding genes than
men.
There are several reasons why it is difficult to study the inheritance
of male balding:
1. Male pattern balding is very common in Caucasian men. This makes
conventional pedigree analysis difficult to interpret.
2. The frequency of balding in the population increases with age. Thus,
studies that rely on the presence or absence of balding in young men are
flawed by the fact that young non-bald men may later develop balding.
Some investigators have used the concept of "premature" balding to get
round this difficulty but the definitions are arbitrary and not based
on any clear evidence that premature balding exists as an entity separate
from balding in general.
3. The clinical pattern of male pattern balding is highly variable.
For example, some men show deep temporal recession with normal hair density
in the mid-frontal region of the scalp whereas others have a diffuse pattern
of hair loss with little or no recession. Indirect evidence that different
patterns of hair loss in male pattern balding are not necessarily under
the same genetic control comes from a recent study which showed an association
of coronary artery disease with vertex balding but not with fronto-temporal
recession [4].
4. There is uncertainty about the phenotype in women. Female pattern
hair loss is often regarded as the same condition as male balding but
this may not necessarily be the case [5, 6].
As part of a study aimed primarily at elucidating the familial basis
of female pattern hair loss we examined scalp hair status in a cohort
of men. We also took detailed family histories of hair status in first
degree family members. Although the results do not provide a conclusive
answer to understanding the inheritance of male pattern balding they do
show that genetic factors also influence non-balding and, from this observation,
suggest new lines of investigation.
Methods
The subjects were men attending a general dermatology clinic during
a 6 month period with complaints unrelated to their hair. Men with medical
conditions or taking drugs likely to affect hair growth were excluded.
The subjects were examined by one of the investigators (AGM) and scalp
hair status was recorded using the Norwood-Hamilton scale [7]. The Ludwig
classification was used in a small number of cases presenting with a female
pattern of hair loss [8]. In subjects with hair loss the age of onset
was recorded. A detailed family history was taken in each subject. Subjects
were asked to indicate hair status in their relatives from diagrams of
Norwood-Hamilton and Ludwig scales.
The Norwood-Hamilton scale is non-linear and some categories are relatively
uncommon, e.g. IIIa, IVa. We therefore reclassified hair status
in our subjects as follows:
Men with a female (Ludwig) pattern of hair loss were placed in the mild
category.
Results
Frequency of balding and the effect of age
We examined 572 men with ages evenly distributed between 16 and 91.
The frequency of balding of all degrees of severity increased progressively
with advancing age (Fig. 1)
as did the frequency of severe balding. By 70 years of age 80% of men
were balding and half of these had severe balding. The frequency of balding
reported by the subjects in their fathers and brothers was lower at all
ages than that observed in the subjects themselves (Fig.
2). The observed frequency of balding in the subjects appeared to
be still increasing at age 80 whereas the reported frequency in fathers
and brothers had stabilised at about 60% by the age of 60. In both groups
of men, therefore, there was a small proportion who appeared resistant
to balding even in old age. Preliminary analysis indicated that these
elderly non-bald men had a high frequency of non-balding in their male
relatives and we went on to examine this possible genetic influence on
non-balding in more detail.
Paternal influence on balding and non-balding
Almost all men who developed balding before the age of 30 had a balding
father. With increasing age, however, the frequency of balding in the
fathers of balding men fell so that by the age of 60 it was only slightly
greater than that recorded in the fathers of all subjects (Fig.
3). Conversely, the frequency of non-balding in fathers of non-bald
subjects aged under 30 was similar to that in all fathers of all men aged
under 30. However, with increasing subject age the proportion of non-bald
men with a non-bald father increased so that by the age of 60 it exceeded
the population frequency by about 75% (Fig.
4).
To examine these relationships in more detail we calculated the relative
risk of balding in subjects with a balding father, and the relative risk
of non-balding in subjects with a non-balding father. There was a significant
increase in the relative risk of balding in young men with a balding father
(RR 5.5, 95% CI 1.26-23.99) which fell with increasing subject age to
approach unity in elderly men. The opposite trend was seen in non-bald
men. Here the relative risk of non-balding in young men with a non-bald
father was low but increased with age to 3.2 (95% CI 1.82-5.58) in subjects
aged 70 and over (Fig. 5).
Is the predisposition
to balding or non-balding determined by a single gene?
Most previous studies have analysed the genetics of balding in young
men. As mentioned above this approach is limited by the fact that the
destiny of hair status in young non-bald men is unknown some will
inevitably develop balding in later life. We are all aware that severe
balding does occur in young men and that these men often have a strong
family history of early balding. However, in reality, severe balding in
young men is relatively uncommon and in the majority of the population
balding does not occur in this setting. Indeed, only 2 of the 94 men aged
under 30 in this study showed balding of moderate degree and none had
severe balding. To analyse our data we therefore elected to study elderly
men where hair status is more likely to be fully expressed. Our own observations
showed that the frequency of balding is still increasing at age 70 suggesting
that all men will ultimately go bald, were they to live long enough. If
correct, this interpretation is strong evidence that balding is a polygenic
trait. On the other hand, the reported frequency of balding in the subjects'
male relatives did not increase beyond the age of 60. This implies a dichotomy
between balding and non-balding suggesting the influence of a single gene,
although in those possessing the putative balding allele other genes may
influence such characteristics as age of onset and severity. To test whether
this latter concept is valid we used the method described by Harris [1]
to analyse our data in elderly men and their male relatives. Hair status
in female relatives was ignored in view of the uncertainty surrounding
the identity of male and female balding. We also discounted X-linked inheritance
due to the clear evidence of a paternal influence on male balding (although
this does not rule out a possible role of a gene situated in the pseudo-autosomal
region of the X chromosome).
Details of the mathematics of the analysis can be found in Harris's
paper. In brief, we assume that balding/non-balding is determined by a
single gene with two alleles, balding (B) and non-balding (b)
which occur in the population with frequencies p and q respectively. If
b is recessive, and there is full penetrance in heterozygotes,
q2 is given by the frequency of non-balding (bb) in
the population sample. From this we can calculate p (as 1-q) and thence
the frequencies of BB, Bb and bb in all possible
matings. A similar calculation can be performed if B is recessive
where the frequency of BB (p2) is given by the frequency
of balding. For any given frequency of balding in the population we can
then derive the expected frequencies for dominant or recessive inheritance
of balding in the progeny of the following sets of matings:
1. All matings capable of producing some bald individuals.
2. All matings capable of producing some bald individuals in which the
father is also bald.
3. All matings capable of producing some bald individuals in which the
father is not bald.
4. And, the frequency of balding in fathers of bald individuals.
The expected frequencies for dominant and recessive inheritance can
then be compared with the observed frequencies. The calculations are performed
on the brothers of bald subjects i.e. they are derived from matings
which can produce at least one bald individual. The subject is not included
in the count to avoid the non-inclusion of families which, although capable
of producing bald individuals, did not in fact do so. All men included
in the analyses
subjects, fathers and brothers were aged 60 and over, and
balding was defined as any degree of balding more than Norwood-Hamilton
II. There were 174 brothers of balding men available for analysis. The
numbers in each category were distributed as shown in Table
I. The expected numbers for both autosomal dominant and autosomal
recessive inheritance, based on the frequency distributions derived as
above, and assuming a frequency of balding in the population of 0.6, are
also shown. The results showed that the observed numbers differed significantly
from expected numbers for both dominant and recessive inheritance. This
held for any frequency of balding in the population above 0.6.
The same analysis was then performed on the brothers (n = 34) of elderly
non-bald men. The results were compared with the expected frequencies
for dominant and recessive inheritance if non-balding is due to a single
autosomal gene if the frequency of non-balding in men aged 60 and over
is 0.4. Once again, the observed numbers in each category differed significantly
from expected numbers for both types of inheritance at this, and all other,
population frequencies of non-balding (Table
II).
Discussion
Balding is a common and complex androgen-dependent trait in adult humans
[9]. Age is an important factor as the frequency and severity of balding
in the population increase with age. Whether age acts independently of
androgens to cause a reduction in scalp hair density in elderly men is
difficult to test conclusively without access to a large number of elderly
eunuchs although Hamilton observed that these individuals have "luxuriant"
hair [10]. The inter-individual differences in age of onset, rate of progression,
severity and familial predisposition also point to a genetic component
to balding. Although it has to be admitted that the genetic basis of male
balding has not been the subject of extensive research it has, for the
reasons outlined earlier, proved difficult to study and remains a controversial
subject. Our study has also failed to give conclusive answers and illustrates
some of the difficulties inherent in this field of research. In particular,
reliance on family histories may be misleading. For example, our results
suggest that men underestimate the frequency of balding in elderly male
relatives casting doubt on the validity of detailed genetic analysis on
historical data. Nevertheless, the trends are probably reliable and allow
general conclusions to be made.
Balding before the age of 30 is uncommon of 94 men in this age
group only 14 showed any degree of balding and none showed severe balding.
Thirteen of these young bald men had a balding father, a frequency far
higher than in the population at large. This observation implies a pronounced
genetic influence on early balding and supports the validity of this group
as a target for genetic analysis. In older men, however, the paternal
influence on balding declined so that by the age of 60 the frequency of
balding in the fathers of balding men was similar to the population frequency.
In contrast, whether or not the frequency of balding in the population
reaches a zenith or continues to increase with age, there is clearly a
small group of men who are resistant to balding into old age. As a group,
these non-bald elderly men have strong family histories of non-balding,
an observation made by Hamilton although he did not pursue it [10]. It
is to be expected that, in an autosomal genetic trait which is infrequent
in the population, the risk of that trait being expressed by the sons
of men with the trait will be increased over that in sons of men who do
not show the trait, whatever the mode of inheritance. Our results are
therefore consistent with a genetic influence on balding in young men,
in whom balding is relatively uncommon and also a genetic influence on
non-balding in elderly men. We were unable to show that balding and non-balding
in elderly men is due to the action of a single gene. This may be taken
as further evidence that balding and non-balding have a multifactorial
causation although this conclusion must be viewed with some scepticism
because of the doubts about the reliability of the historical data. Confining
the calculations to subjects, brothers and fathers all aged 60 and over
also limited the numbers of brothers available for analysis, particularly
in the brothers of non-bald subjects.
CONCLUSION
In summary, our results indicate that there is a paternal influence on
balding, especially in young men, and on non-balding in elderly men. The
nature of the genetic predispositions to balding and non-balding remains
unclear although we were unable to show that either balding or non-balding
is due to the action of a single gene. From a practical viewpoint there
are certain advantages to studying the genetics of non-balding in elderly
men. Non-balding is an easier endpoint to define than balding, the expression
of balding and non-balding is more fully developed than in young men and
uncertainty about the destiny of hair status in siblings is minimised.
Also, if we are to develop new therapies for male balding it may be fruitful
to understand how natural resistance to balding is conferred. On the other
hand, family studies on elderly non-balding men are complicated by the
fact that the parents are often not available for testing. Finally, although
the same genes may be responsible for determining the predisposition to
balding and to non-balding we cannot assume that this the case and it
is possible that resistance to balding is conferred by different genes
from those which predispose to the development of balding.
Acknowledgements
We are grateful to the Special Trustees for the Former United Sheffield
Hospitals who funded this study and to Dr. Rosie Taylor for statistical
advice.
REFERENCES
1. Harris H. The inheritance of premature baldness in men. Annals
of Eugenics 1946; 13: 172-81.
2. Osborn D. Inheritance of baldness. J Hered 1916; 7:
347-55.
3. Küster W, Happle R. The inheritance of common baldness:
two B or not two B? J Am Acad Dermatol 1984; 11: 921-6.
4. Lotufo PA, Chae CU, Ajani UA, Hennekens CH, Manson JE. Male
pattern baldness and coronary heart disease: the Physicians' Health Study.
Arch Intern Med 2000; 160: 165-71.
5. Orme S, Cullen DE, Messenger AG. Diffuse female hair loss:
are androgens necessary? Br J Dermatol 1999; 141: 521-3.
6. Norwood OT, Lehr B. Female androgenetic alopecia: a separate
entity. Dermatol Surg 2000; 26: 679-82.
7. Norwood O. Male pattern baldness: classification and incidence.
South Med J 1975; 68: 1359-65.
8. Ludwig E. Classification of the types of androgenetic alopecia
(common baldness) occurring in the female sex. Br J Dermatol 1977;
97: 247-54.
9. Hamilton JB. Male hormone is pre-requisite and an incitant
in common baldness. Am J Anat 1942; 71: 451-80.
10. Hamilton JB. Patterned hair loss in man: types and incidence.
Ann NY Acad Sci 1951; 53: 708-28.
|
Printable version
