Day care in infancy and risk of childhood acute lymphoblastic leukaemia: findings from UK case-control study
BMJ 2005; 330 doi: https://doi.org/10.1136/bmj.38428.521042.8F (Published 02 June 2005) Cite this as: BMJ 2005;330:1294
All rapid responses
Rapid responses are electronic comments to the editor. They enable our users to debate issues raised in articles published on bmj.com. A rapid response is first posted online. If you need the URL (web address) of an individual response, simply click on the response headline and copy the URL from the browser window. A proportion of responses will, after editing, be published online and in the print journal as letters, which are indexed in PubMed. Rapid responses are not indexed in PubMed and they are not journal articles. The BMJ reserves the right to remove responses which are being wilfully misrepresented as published articles or when it is brought to our attention that a response spreads misinformation.
From March 2022, the word limit for rapid responses will be 600 words not including references and author details. We will no longer post responses that exceed this limit.
The word limit for letters selected from posted responses remains 300 words.
Comment on Gilham et al (UKCCS) BMJ 22 April 2005
[reference BMJ, doi:10.1136/bmj.38428.521042.8F (published 22 April
2005)].
The summary box on what was already known, before this study, starts
with "Childhood leukaemia is a biologically diverse disease and is likely
to arise by several aetiological pathways".
This study adds rather striking statistical results to support the
first of the two parts of the infection hypothesis; that is the Greaves
hypothesis that infant deficit of exposure to infection (and consequent
deficit in immune development) increases the risk of ALL. Implicitly this
finding therefore also lends support to the second part: the Kinlen
hypothesis that childhood exposure to infection through population mixing
increases the risk of ALL.
The study reflects the multifactorial nature of the immune system and
how it deals with stresses which may lead to leukaemia. It thereby
reinforces the opening statement from the summary box, that childhood
leukaemia is likely to arise by several aetiological pathways. Among them
may be pathways involving infection in combination with non-infection
exposures which interfere with the immune system.
Thus by strengthening the infection hypothesis, the study adds to the
case for involvement of the several non-infection exposures which have
some evidence of association, such as aspects of diet or the IARC category
2B electromagnetic fields (EMF) among others.
On the other hand, for cases with pre-leukaemic genetic damage, the
study has no bearing on the aetiology of the genetic damage in utero. In
this respect EMF may be further implicated by the results of Draper et al,
BMJ vol. 330, 1290 - 1293, 4 June 2005.
Gilham et al do not show that non-infection mechanisms and causal
factors are not at work, but rather their study enhances the prospects for
them. It is surprising therefore that media responses to this study,
including scientific media, with whatever encouragement, have made such
statements as "there is no association between exposure to electromagnetic
fields and the risk of childhood leukaemia" and generally seemed to
champion the infection hypothesis to the exclusion of other factors, which
might be all the more relevant in combination with infection.
The findings in this paper need to be honed and replicated of course
(and how often has that been demanded of EMF findings!), but the
statistical results are striking, to use a word implied by the authors. I
found the smoothness and consistency of the results slightly surprising
given the rough nature of the proxies for social interaction.
Could the authors have made an error in respect of the counter-
expectation result on groups with and without older siblings? They say
"The odds ratio for formal day care was 0.61 (0.42 to 0.87) for ALL in
children without older siblings and 0.38 (0.26 to 0.54) for those with
older siblings, a non-significant difference in the opposite direction to
that anticipated." The overlap of confidence intervals (as in 0.42 <
0.54) does not necessarily mean the difference was non-significant; it
only suggests less significant than p = .05 squared (p = .0025). As the
odds ratio for each group falls outside the confidence interval for the
other (as in .38 < .42), the result appears significant, though this is
not material to the main conclusions of the study.
M J O’Carroll 19 June 2005
Competing interests:
None declared
Competing interests: No competing interests
Online publication of this paper by Gilman et al on 22nd April was
accompanied by extensive press coverage to the effect that the paper also
showed that electromagnetic fields, such as from powerlines, background
radiation and vitamin K are not causes of childhood leukaemia. We would
like to point out that none of these agents are mentioned in the paper by
Gilman et al, which is confined to cancer rates in infants attending day
care centres.
A pooled analysis by Professor Anders Ahlbom et al (British Journal
of Cancer September 2000), which included the results of a UKCCS
powerlines study, found a doubling of the risk of childhood leukaemia with
magnetic field exposure above 0.4 microtesla, well below the levels seen
near powerlines. The separate results of the UKCCS study alone were
subsequently published in the British Journal of Cancer in November 2000.
This study found 40% more cases than controls near powerlines in the UK.
We also point out that the Dr Gerald Draper et al have recently
published their work (British Medical Journal 3rd June) showing increased
incidence of childhood leukaemia up to 600 metres from 400 and 275 kV
powerlines in the UK where they also cite the above results of the UKCCS
2000 study.
Competing interests:
None declared
Competing interests: No competing interests
Editor,
Firstly let me apologize for not heeding your advice to read
carefully one’s response before clicking the ‘submit’ button. I failed to
do so in my last response and hence Leukaemia became Leukarmia.
Now to answer Dr James E Parker who can see a “red herring” in my
response. I have explained how I arrived at this conclusion[1]. Some of
your readers may recognize my explanation as an example of ‘Ampliative
Induction’[2].
Having declared my position [3] it is up to Dr C Gilham and co-
authors[4] to prove I am wrong.
They can do this very easily by showing there is no significant
statistical association between Acute Lymphoblastic Leukaemia and DPT
vaccine in the children they report.
I await their verdict. After 40 years of waiting I can wait a little
longer.
Michael Innis
References:
1.Innis MD. Rapid Response 26th May 2005
2.Kneale W Probability and Induction 2nd Impression 1963 pp 56-60
Oxford University Press Amen House London E C 4
3.Innis MD Immunization and Childhood Leukaemia Lancet Letter to the
Editor March 13th 1965 i 605
4.Gilham C, Peto J, Simpson J, Roman E et al; Day care in infancy and
risk of childhood acute lymphoblastic leukaemia:findings from UK case-
control
study. BMJ, Apr 2005; 10.1136/bmj.38428.521042.8F.
Competing interests:
As previously declared
Competing interests: No competing interests
With respect I would suggest that Drs Ivanoski (Rapid Response 5 May
2005) and Innis (26 May 2005) miss the point of Gilham et al's article
when they invoke infant immunization in the aetiology of leukaemia by the
introduction of a 'red herring'.
The discovery by Greaves of disease markers on Guthrie cards and cord
blood suggests a period of long latency in childhood leukaemia in relation
to the 'childhood peak'. These markers were present BEFORE the
individual administration of DPT immunization.
James E Parker
Competing interests:
None declared
Competing interests: No competing interests
Editor,
Responding to the report by Gilham et al; [1] I draw attention to a
letter entitled “Immunization and Childhood Leukaemia” [2] in which it was
shown that Leukaemia in children in Brisbane Children’s Hospital from
1958 to 1964 showed a significant statistical association with
immunization against diphtheria, tetanus and whooping cough.
In the letter it was stated, “since antigenic stimulation is known to
cause hyperplasia of mammalian lymphoreticular tissue, and since such
hyperplasia may in some strains of mice, proceed to malignant neoplasia
[3] the possibility that a similar mechanism could operate in children
submitted to the repeated antigenic stimulation inherent in immunization,
seemed worth investigating.”
In view of Dr Ivanovski’s observations that the incidence of
childhood leukaemia increased with the introduction of DPT vaccination [4]
it is virtually certain that, if Gilham et al; investigated the
immunization status of the children in their two groups, they will find
the group with Leukaemia also shows a statistically significant increase
in immunization with DPT vaccine.
This would then call not for rejection, but for a future redefinition
of vaccination schedules, as Dr Ivanovski says.
Michael Innis
References:
1. Gilham, J Peto, J Simpson, E Roman, T O B Eden, M F Greaves, F E
Alexander, BMJ, 1:10.1136/bmj.38428.521042.8F (published 22 April 2005)
2. Innis MD Immunization and Childhood Leukaemia Lancet Letter to
the Editor March 13th 1965 i 605
3. Metcalfe D. Brit. J. Cancer 1961; 15:769
4. Ivanovski P. Prevention of Childhood Leukaemia: Is it possible
BMJ Rapid Response 5th May 2005
Competing interests:
As previously declared
Competing interests: No competing interests
Peter Ivanoski states, with concern, and references:
>>>"Before 1920, acute leukemia among children was a rare
event. A significant peak-age incidence (2-5 years) appeared after 1940.
Since then, the incidence rate of childhood leukemia has been more or less
remarkably stable. This means that some leukemogenic factor must have been
introduced in children's lives some time around 1940.
It is a highly striking coincidence that at the same year the
introduction of immunization against diphtheria was began on a national
scale."<<<
I wonder if our friends at the CDC, NIH, WHO etc. have considered
adding leukemia in addition to diabetes, Guillian-Barre', Autism, SIDS,
Arthritis, Thrombocytopenia, Encephalitis, Death, SBS, Distressed
Breathing, Thimerosal Accumulation in Brain (TAB) (1), delayed speech,
tics, seizures, hallucinations, diziness, Hemorrhagic Vasculomyelinopathy
etc. etc. etc. to "highly coincidental" adverse reactions from the long
list of mass immunizations.
Do you think parents would be informed during their child's well
visit...of any of the above? Well, at least they were "well"...or should
have been prior to the inoculation(s). That is unless you follow most
vacccinators advice. In particular, MMR advice from WHO is to jab unless
the child is in serious risk of dying. And the only reason given not to
jab in this instance is that the death may "incorrectly" be attributed to
the MMR. (2) And we wonder why most all serious adverse vaccine reactions
are attributed to "coincidence". As clearly seen in this WHO advice--take
great lengths to disclaim any adverse vaccine reaction.
Back to the leukemia aspect. If consently informed, would you opt
for the risk of a cough with a grayish film covering the throat and small
risk of death or serious chronic illness sequelae (usually cardiac) from
diphtheria (mostly for the malnourished, immunocompromised or poorly
managed), or the small risk of a "coincidental" vaccine induced leukemia
(and all its associated "goodies")? What about the novel, yet to be seen
vaccine derived, serious illnesses that seem to keep popping up out of the
blue? Do you want to risk those (keeping in mind we're not looking for
the vaccine correlation and if we find it, we'll certainly hide it), or
risk an illness you are hopefully familiar with?
1) http://ehp.niehs.nih.gov/members/2005/7712/7712.pdf Comparison of
Blood and Brain Mercury Levels in Infant Monkeys Exposed to Methylmercury
or Vaccines Containing Thimerosal; Thomas M. Burbacher, Danny D. Shen,
Noelle Liberato, Kimberly S. Grant, Elsa Cernichiari, and Thomas Clarkson
2) Bulletin of the World Health Organisation, 1997; 75 (4) page 371
Competing interests:
None declared
Competing interests: No competing interests
Editor: The recently published study, Day care in infancy and the
risk of childhood acute lymphoblastic leukaemia: findings from the UK case
-control study by Gilham et al(1), has a major problem with bias created
by selective non-response in the control group. Figure 3 of the United
Kingdom Childhood Cancer Study (UKCCS): objectives, materials and
methods(2) shows that of 11,987 eligible controls, only 7,629 (64%) of
control parents were interviewed. A study linking the UKCCS study cases
and controls to the 1991 census of Great Britain(3) showed that case
families lived in more affluent areas than did first choice control
families, but that compared with participating control families, the case
families tended to live in areas that were less affluent than those of
control families. Table 4 of the UKCCS study shows that 57.6% of case
mothers were not gainfully employed compared to 47.3% of the participating
control mothers. I suspect that since a larger proportion of
participating control mothers were gainfully employed than were the case
mothers, their use of day care would be higher. This is reflected in table
2 of the Gilham et al paper. I, therefore, think that the low calculated
odds ratios of childhood leukemia in children attending day care in
infancy is more likely to be due to a selection bias in control
participation than to exposure to infectious agents in day care.
REFERENCES
1. Gilham C, Peto J, Simpson J, Roman E, Eden TOB, Greaves MF, Alexander
FE. Day care in infancy and risk of childhood acute lymphoblastic
leukaemia: findings from the UK case-control study. BMJ,
doi:10.1136/bmj.38428.521042.8F (published 22 April 2005)
2. UK Childhood Cancer Study Investigators. The United Kingdom childhood
cancer study: objectives, materials, methods. British J. Cancer 2000;
82(5), 1073-1102.
3. UK Childhood Cancer Study Investigators (writing committee: Law GR,
Smith AG, Roman E)
The importance of full participation: lessons from a national case-control
study. British J. Cancer 2002; 86(3), 350-355.
Competing interests:
None declared
Competing interests: No competing interests
Speculating about the infective etiology of acute leukemia in
children, in her fascinating paper from 1942 (1) Jean V. Cooke wrote: ‘…it
occurs only in certain persons in whom a constitutional inferiority of the
hematopoietic system makes it vulnerable to injury by infections and
possibly by other types of trauma.’
Over half a century later, Greaves et al. deciphered the meaning of
‘the constitutional inferiority of the hematopoietic system’ (2). Thus,
owing to his ingenious project, today we know that most of childhood
leukemias have prenatal origin. Leukemogenic translocations originating
during fetal life are insufficient for overt leukemia. This has been
additionally confirmed by experiments with transgenic TEL-AML1 mice, which
failed to develop leukemia (3). Thus, it has been concluded that
additional mutations or abnormal expression of another
gen(s) may be necessary to promote leukemogenesis, perhaps through
increased proliferation.
According to Gilham et al’s findings (4) and Greaves’ hypothesis (5),
leukemia occurs in children who lack early immunomodulation, and who later
contract some common infection of low pathogenicity.
Having in mind a very stable incidence of childhood leukemia and the
need of multiple very regular proliferative stresses on B cell precursors
with TEL-AML1 fusion to cause additional mutations (allelic TEL deletions)
for overt leukemia, naturally we must ask the question:
Which is the infection of low pathogenicity that is capable of
causing a proliferative stress, and at the same time to be repetitive and
to occur every year so as to lead to a stable incidence of leukemia?
Before 1920, acute leukemia among children was a rare event. A
significant peak-age incidence (2-5 years) appeared after 1940 (6). Since
then, the incidence rate of childhood leukemia has been more or less
remarkably stable (7). This means that some leukemogenic factor must have
been introduced in children's lives some time around 1940 (6).
It is a highly striking coincidence that at the same year the
introduction of immunization against diphtheria was began on a national
scale (8).
Further experimental work on transgenic TEL-AML1 mice, exposed to
diphtheric anatoxin should be conducted to investigate if there is any
relationship between the introduction of systemic vaccination against
diphtheria and the peak age incidence (at age 2-5 years) appearance of
childhood leukemia.
If some of these mice develop leukemia, we shall make a great step
toward the prevention of childhood leukemia.
This would then call not for rejection, but for a future redefinition
of vaccination schedule (9).
1.Cooke JV, (1942) JAMA 119:547-50.
2.Greaves MF, (1999) Lancet 354:1499-03.
3.Andreasson P, (2001) Cancer Genet Cytogenet 130: 93-104.
4.Gilham C, (2005) BMJ April 22.
5.Greaves MF, (1997) Lancet 349:344-49.
6.Court Brown, WM, (1961) Br Med J 1:981-8.
7.Hjalgrim LL, (2003) J Natl Cancer Inst 95:1539-44.
8.Salisbury DM, Begg NT. Immunisation Against Infectious Disease. HMSO,
Department of Health, London, 1966, p.67.
9.Barrios C, (1996). Eur. J. Immunol. (July) 26 (7): 1489-96.
Competing interests:
None declared
Competing interests: No competing interests
The study by Gilham et al. confirms the hypothesis that reduced
exposure to infection early in life has effects on the maturing immune
system that increase the risk of acute lymphoblastic leukaemia (ALL) and
possibly other malignancies. The immunological basis of this increased
risk is uncertain but it could be the result of the inadequate development
of immune surveillance mechanisms that detect cancer-specific antigenic
determinants. This suggestion raises the possibility of using vaccination
strategies to replace the environmental contacts that would otherwise lead
to immune maturation. Gilham et al. mention Haemophilus influenzae type b
vaccination in this respect but another possible vaccine is Bacille
Calmette Guerin (BCG) which, when given early in life, has been shown in
several studies to afford protection against ALL.1
In this context, a working group of the European Organization for
Research and Treatment of Cancer (EORTC) has shown that vaccination with
BCG early in life, as well as vaccinia vaccination and certain serious but
uncommon infections, reduces the risk of the subsequent development of
melanoma by around 50% and that, in comparison with those not vaccinated,
a history of one or other, or both, of these vaccinations is related to
significant improvement in the survival time of those developing
melanoma.2 Further investigations are required to confirm the protective
effect of BCG against leukaemia and melanoma and to determine whether it
protects against other cancers. If so, a re-introduction of this vaccine
could have significant impacts on human health.
Gilham et al. postulate that the inadequate priming of the immune
system due to a lack of exposure to infection permits subsequent
infections by unknown exogenous agents, probably viruses, to cause immune
dysregulation leading to ALL. This may occur but an alternative
possibility is that the causative agent is not an exogenous one but a
human endogenous retrovirus (HERV). In the case of melanoma, we have
suggested that a failure of immune surveillance permits expression of a
member of the HERV-K family of endogenous retroviruses.3 This expression
leads to intracellular changes resulting in malignant transformation and
also to the presentation of an epitope, HERV-K-MEL, on the cell surface.
Interestingly, BCG, vaccinia and the agents causing the protective
infectious diseases contain homologs of the HERV-K-MEL epitope and could
therefore afford protection by generating a population of cross-reacting
memory T cells that are involved in immune surveillance of pre-malignant
and malignant cells.
If such HERV-related epitopes are involved in ALL and other cancers
their identification could, perhaps, lead to the development of a vaccine
affording protection against a range of cancers.
John M.Grange
Centre for Infectious Diseases and International Health, University
College London. London, UK.
Bernd Krone
Klaus F. Kölmel
Departments of Virology and Dermatology, University of Göttingen,
Göttingen, Germany.
1. Grange JM, Stanford JL. BCG vaccination and cancer. Tubercle 1990;
71: 61-64.
2. Kölmel KF, Grange JM, Krone B, et al. Prior immunisation of patients
with malignant melanoma with vaccinia or BCG is associated with better
survival. An European Organization for Research and Treatment of Cancer
cohort study on 542 patients. European Journal of Cancer 2005, 41: 118-
125.
3. Krone B, Kölmel KF, Henz BM, Grange JM. Protection against melanoma by
vaccination with Bacille Calmette-Guérin (BCG) and/or vaccinia: an
epidemiology-based hypothesis on the nature of a melanoma risk factor and
its immunological control. European Journal of Cancer 2005, 41: 104-117.
Competing interests:
None declared
Competing interests: No competing interests
Cortisol and Stress in Children
The article by Gilham et al raises the question as to whether the
increased
incidence of infections in child care and associated acute leukemia was
linked
with cortisol levels .There is growing evidence that chronic stress is
related to
cortisol failing to lower in the usual diurnal patterns with adverse
outcomes
on health, learning and behaviour. An ongoing West Australian study using
salivary cortisol levels as a biomaker taken morning and afternoon on
three
successive days was linked to Australain standards of quality child care
(QAIS). In unsatisfactory child care as determined by these standards
there
was a failure of the usual decline in afternoon cortisol. However even in
those
child care centres assessed as 'satisfactory' there was a similar failure
of the
afternoon decline of salivary cortisol levels in some of the quality
criteria that
condsidered relationship issues.Could it be that rather than infection
being a
general concern in child care that there is a link with stress in children
and an
adverse cortisol response when child care is unsatisfactory. Might this
have
significant preventive implications both for infection and the reported li
nk
with acute leukemia?
Competing interests:
None declared
Competing interests: No competing interests