Letters to the Editor 277
2 Baraitser M, Patton M, Lam s'rs, Brett Although we believe no families decade, 26 042 extended family
EM, Wilson J. The Angelman (happy would possess information on all first members, and a 64-7% chance of one
puppet) syndrome: is it autosomal
recessive? Clin Genet 1987;31:323-30. to fifth degree relatives (that is, our or more families with recurrence on a
3 Hagberg B, Goutieres F, Hanefeld F, estimate of 137 relatives is in that way chance basis. For all types of transverse
Rett A, Wilson J. Rett syndrome: unrealistic), our model of a family with forearm/hand defect the birth pre-
criteria for inclusion and exclusion. valence is much higher (5-8 per 10 000
Brain Dev 1985;7:372-3. only two children for each two parents
4 Pembrey M, Fennell SJ, Van Den underestimates family size. In our from Wynne-Davies and Lamb4), and
Berghe J, et al. The association of clinics we routinely record pedigrees to one would expect 3480 UK cases
Angelman's syndrome with deletions include grandparents, uncles/aunts, within a decade, with 377 614 extended
within 15q11-13. 7 Med Genet 1989;
26:73-7. and first cousins, and examining the family; about 6%YO of these cases would
files of the first 25 families seen in be 'familial'.
Oxford in 1989 (referrals to a general In conditions with a high birth
Recurrence of acheiria in a second genetics clinic) we found a total of 475 prevalence the likelihood of chance
cousin: extremely large pedigrees such relatives of the index cases, that recurrence within an extended family is
may include 'second cases' by chance is, an average of 19 per family (not 13, considerable, and in general one should
as would be expected on the 'simple' beware of arguing from such recur-
Lamont and Salisbury' report unilateral model). This suggests that our estimate rences unless the necessary calculations
absence of the hand (acheiria) in an of 41 relatives per family, if one have been made and a coincidence
index case and in a second cousin. In extends to include second cousins, is appears truly unlikely.
their opinion the possibility that "both also an underestimate. We found 4-68
children could have been affected by uncles/aunts instead of the 'expected' RICHARI) LINDENBAUM,
chance . is too remote to be con- two, and 7-32 first cousins instead of HELEN FIRTH
the 'expected' four, that is, each grand- Department of Medical Genetics,
sidered". They found no common Churchill Hospital,
environmental factor, and proposed a parental pair had on average 3-34 Headington,
"common mutant dominant gene", (4-68/2+1) offspring rather than two, Oxford OX3 7L7.
transmitted from one of the shared and each of these persons had an
grandparents, as "the most cogent average of 1-56 (7-32/4 68) live 1 Lamont M, Salisbury AJ. Unilateral
explanation". While not denying the children at the time the pedigree was absence of the hand in second cousins.
logic of their suggestion, nor the constructed. Assuming that fertility .7 Med (;enet 1989;26:190-2.
2 Birch-Jensen A. Congenital deformities of
evidence from recurrences of acheiria was no less in the previous (great the upper extremities. Copenhagen:
in sibs and in second and third degree grandparental) generation, we calculate Munksgaard, 1949.
relatives, as mentioned in their paper, an average of 9-36 great uncles/aunts,
3 Rogala EJ, Wvnne-Davies R, Littlejohn
A, Gormley J. (Congenital limb
we wish to point out that if one records 31 26 first cousins once removed, and anomalies: frequency and aetiological
pedigrees so as to include second 48-89 second cousins, that is, a total of factors. 7 Med Genet 1974;11:221-33.
cousins (fifth degree relatives) of the 108 51 relatives of the index case if 4 Wynne-Davies R, Lamb DW. C(ongenital
upper limb anomalies: an aetiologic
index case one is potentially including families were to be recorded in this way grouping of clinical, genetic, and
data on a very large number of persons (omitting all generations before the epidemiologic data from 387 patients
indeed, and 'recurrences', on a chance grandparental or after the index case). with "absence" defects, constriction
basis, are not as unlikely as may Lamont and Salisbury give a birth bands, polydactylies, and syndactylies.
initially appear to be the case. prevalence for acheiria of 1 in 65 000. .7 Hand Surg [Ami 1985;10A (part 2):
958-64.
Assuming a simple family structure With about 600 000 live births in the
of two children for every pair of UK per annum one would expect 92 3
parents (that is, a steady state, non- births with acheiria within the last This letter was shown to Drs Lamont
expanding population), and the decade, and these would have a total of and Salisbury, who reply as follows.
recording, in the genetics clinic, of 10 015 relatives up to and including
families to include only sibs, parents, second cousins (92-3 x 108-51). The While not disputing the logic of this
grandparents, uncles/aunts, and first probability of acheiria, by chance, comment by Drs Lindenbaum and
cousins of the index case, one would among these relatives would be 10 015/ Firth on the possibility of recurrence of
record data on 1+2+4+2+4= 13 other 65 000=0 1541, and from the Poisson acheiria within families when consider-
persons per family. Enlarging the distribution the chance of one or more ing the entire United Kingdom popu-
pedigree to include second cousins, such families actually existing would be lation, we would point out that our case
with the more recently born of their 14280!o. report referred to one family with 10
linking relatives, one would record However, Lamont and Salisbury relatives, other than the proband,
four great uncles/aunts, eight first quote an old reference for the birth descended from the common grand-
cousins once removed, and 16 second prevalence of acheiria.2 From the parents. For our quoted prevalence
cousins, that is, a total of 41 relatives of Edinburgh Register of the Newborn, rate of acheiria of 1 in 65 000, the
the index case. Were one to include all Rogala et al- found birth prevalence of probability of recurrence would be
first to fifth degree relatives in previous isolated transverse transcarpal absence 0-00015, surely a low figure. With the
or contemporary generations (eight defects to be 0-4 per 10 000, and this is prevalence rate of I in 25 000 from the
great grandparents, 16 great great in agreement with estimates one can survey by Wynne-Davies and Lamb,
grandparents, etc) one would obtain a make combining information from the probability of recurrence in our
total of 137 relatives per family, while Birch-Jensen2 and from Wynne-Davies family would be 0-0004. From a total
still excluding generations younger and Lamb.4 With this prevalence we of 2833 relatives, including those of
than the index case. would expect 240 UK cases in a third degree, of 102 probands with