|  |  |  | | | | Diseases | |  | Childhood Leukaemia - is it increasing? |
| | |
| |  | The Assertion |
It has been suggested that there has been a relentless increase in childhood leukaemia incidence throughout the last century, that this is related to environmental carcinogens, and that there is no understanding of, and inadequate research into, the causes of this increase.
The claims relating to the first half of the 20th Century rely on mortality data, since there were no cancer registries in operation during this period. Assertions that there is a continuing rise in incidence depend on the assumptions that, over the last two decades, there have been no major improvements in diagnosis and classification of childhood leukaemia and that recording of cases has not improved in efficiency and completeness.
| The Evidence |
Most major studies reported from the developed world indicate that the incidence of childhood leukaemia is either stable or, if increasing, is doing so at a very low rate. The rate of increase in the UK is no greater than about 1% per year and this may well be accounted for by improved diagnosis, classification and recording. It is important to realize that a 1% increase on the UK rate represents around 4 to 5 cases per year - this is not to underestimate the importance of every single case of the disease, but it does show that even random variation from year to year could account for part of the changes.
1900 to 1950
Antibiotics did not enter routine clinical use until the second half of the 20th Century; prior to that, management of infectious disease had not greatly improved since the time of the ancients. A patient developing a serious infection would receive nursing care but there were no effective drugs to directly fight the infection. Unsurprisingly, death from infection was common, especially in children.
A child suffering from leukaemia during this pre-antibiotic era would almost certainly have died, usually early in the course of the disease, from overwhelming sepsis. Long-term survival from leukaemia was so rare that attempts have been made to argue that the death rate from childhood leukaemia is an accurate reflection of the incidence rate; this however depends on the assumptin that childhood leukaemia would be accurately diagnosed, either before or at the time of death. The signs and symptoms of acute leukaemia are very non-specific, and it would have been difficult or impossible to tell that the child dying from infection also had leukaemia.
During this period, outside of centres of research excellence, very few hospitals would have had pathology laboratories. This would mean that even where a doctor suspected childhood leukaemia, it would be almost impossible to confirm the diagnosis. Although the survival rate improved little before the 1960s (and introduction of chemotherapy), the capacity to diagnose childhood leukaemia improved steadily throughout this time. It is literally meaningless to compare mortality data from the turn of the century with that from the 1950s.
It should be noted that this period also saw tremendous improvements in standards of living and of communal hygiene, a major impact of vaccination programmes on common childhood infections amd greatly increased mobility and population mixing. These are all factors which would be expected to lead to an increase in childhood leukaemia, particularly in the 2-5 peak age group, on the basis of the infectious exposure theories of Professor Mel Greaves and Professor Leo Kinlen.
1960s to present
Introduction of the first effective treatments in the 1960s meant that for the first time significant numbers of children became long term survivors. This decisively uncoupled incidence data from mortality data - at last childhood leukaemia had ceased to be uniformly fatal. The problems in comparing incidence data, even from the last two decades, let alone from forty years ago lie in consistency of classification and in completeness of recording.
It is generally accepted that Non-Hodgkin's lymphoma (NHL) and lymphocytic/lymphoblastic leukaemia are at extreme ends of a spectrum. One of the most important factors in deciding how to classify the disease is the number of blast cells in the bone marrow. There is strong, albeit anecdotal, evidence to suggest that classification as leukaemia or NHL depended to a large extent on whether diagnosis was made by an oncologist (cancer specialist), a paediatrician, or a haematologist (blood specialist). The United Kingdom Childhood Cancer Study Group (UKCCSG) was set up in 1977 and one of its early achievements was to develop uniform criteria for classifying childhood lymphoid malignancy as NHL or leukaemia.
Evidence of classification errors is not all anecdotal; when the Committe of Medical Aspects of Radiation in the Environment (COMARE), was looking at incidence around Dounreay they included NHL as well as leukaemia because of concerns regarding classification. When they re-examined cases within this area they found that between 1968 and 1984, six cases had been correctly diagnosed as leukaemia, but a further two cases had been incorrectly diagnosed as NHL. In other words, there was a 25% error rate in classification in this area. The introduction of new, very accurate, diagnostic tests has greatly reduced the risk of misclassification, as has the tendency for children to be treated at a limited number of highly specialist centres.
Even where classification is entirely accurate, analysis of trends in incidence depends on the assumption that all, or virtually all, cases are registered locally and that these local registers are providing full data to national registers. Cancer, in adults or children, is not a notificable disease in this country. This means that all reports are voluntary and the completeness of reporting has been very variable over time.
One of the largest and most comprehensive studies of childhood cancer in the world has been carried out by the United Kingdom Childhood Cancer Study (UKCCS). This collected data on all newly diagnosed cases of cancer affecting children - notification systems were set up within all hospitals in Great Britain. Prior to this study the main source of incidence data was the National Registry of Childhood Tumours (NCRT). As a verification of the UKCCS data it was cross-checked with NCRT; this comparision indicated that, through no fault of the NCRT personnel, a number of cases were missing from their figures. This clearly indicates that it is not possible to rely on historical data for incidence of childhood leukaemia in assessing the rate, if any, of increase. A number of local checks have been carried out on childhood leukaemia statistics for the 60s, 70s and 80s and these have confirmed that figures throughout this period are likely to have significantly underestimated the true incidence.
Ken Campbell MSc (Clinical Oncology)
Clinical Information Officer |
|
|
|
