View report page

167
Later Bradford Hill and Knowelden (3) made a thorough analysis of national data
the results of which may be summarised as follows :
(a) In cases of poliomyelitis the distribution of paralysis by site was abnormal in
children who had had an injection of antigen in the month preceding onset of
illness. In children not so inoculated or without any recent injection paralysis
occurred equally in the right and left arms and paralysis in the legs was between two
or three times as frequent as paralysis in the arms. In those inoculated within the
preceding twenty-eight days paralysis was more frequently observed in the left arm
than in the right, arm and paralysis in the arms was more frequent than paralysis in
the legs.
(b) Comparison of poliomyelitis cases with 'controls showed that more of the
former had been inoculated at some time. Division by time revealed that this excess
lay wholly in inoculations given within the month preceding the onset of the
attack of poliomyelitis, suggesting that inoculation had brought some children into
the paralysed group who would not otherwise have been there.
In reporting the London County Council data (2) which were similar in sort to
those referred to in (b) above it was stated that' these figures suggest that in a few cases
poliomyelitis has been associated with inoculation with the combined (diphtheria and
whooping-cough) antigen, but they do not measure the risk of an inoculated child
being attacked by poliomyelitis. To measure this risk would mean comparing the
incidence of poliomyelitis in inoculated and uninoculated infants and would not be
easy to do because of difficulty in assessing the proportions of these two groups in the
population.' Nevertheless it was considered essential that the difficulties should be faced
and that the risks should be assessed as accurately as possible. It was therefore decided to
survey inoculations carried out during 1949 in London. A total survey was considered
impractical and indeed unnecessary since the numbers involved were so large that a
random sample was likely to be statistically adequate. In fact, a one in five sample was
made of children who had been inoculated.
In Tables 1A, 1B and 1C are shown cases of poliomyelitis aged nine months to two
years in the month in which they occurred related to the month of inoculation and up
to the third month following inoculation. This has been tabulated for A.P.T. and the
combined antigen separately and for all other antigens together (these are a mixture of
P.T.A.P., F.T., T.A.F. and the three proprietary brands of pertussis vaccine). The
month of inoculation is the month in which the last injection was given ; earlier
injections arc ignored. This assumes that if the last injection was given, say seventeen
days before onset and the earlier injection, say, forty days before onset, the earlier
injection did not influence the incidence of poliomyelitis. Thus the cases of poliomyelitis
in these Tables are shown once only in relation to the most recent inoculation ;
there is no question of duplicate entries. The last injection may have been a first, second
or third primary, i.e., they are entered irrespective of order. In Table 2 is a group labelled
'not inoculated group' being the residue of poliomyelitis cases not inoculated within
three months of contraction of poliomyelitis. It is assumed that an injection more than
three months before onset of the disease could not have exerted any influence thereon.
These then form the numerators in the subsequent calculations. The reason for confining
this study to the narrow age limits of 9 months to 1 year 11 months is that the sample
survey demonstrated that it is within these ages that immunisation very largely took
place in 1949. The age of the children throughout this study is reckoned as the age at
the date of inoculation.
The ' exposed to risk ' in these tables are the number of children receiving in any
given month the final inoculation of a series with stated antigen. This number has been
estimated by taking those who completed a course in the stated month together with an
allowance for those who had an inoculation (not the last of a complete course) and
who did not return for further inoculations. There remains the calculation of the
denominators for the ' not inoculated ' group Table 2. In the inoculated groups, cases of