Increasing incidence of cancer in children and competing risks

Bron: www.thelancet.com/journals/lanonc/article/PIIS1470-2045(18)30498-4/fulltext?dgcid=raven_jbs_etoc_email
Volume 19, ISSUE 9, P1136-1137, September 01, 2018
Open Access
Published:August 08, 2018DOI:https://doi.org/10.1016/S1470-2045(18)30498-4

The Automated Childhood Cancer Information System (ACCIS) has been registering data on cancer occurrence from birth to age 20 years in most European countries since the 1970s. In The Lancet Oncology, Eva Steliarova-Foucher and colleagues1 report the latest ACCIS data analysis trends for the period 1991–2010 for malignant neoplasms in children aged 0–14 years and adolescents aged 15–19 years. Incidence data were collected in 19 countries grouped into east, north, south, and west regions of Europe. The results of this study show two striking features. First, cancer incidence in children aged 0–14 years has gradually increased over the 1991–2010 period.1 Increasing incidence was observed for all cancers (0·54% (95% CI 0·44 to 0·65) per year), leukaemia (0·66% (0·48 to 0·84) per year), lymphoma (0·26% (–0·01 to 0·54) per year), CNS malignant tumours (0·49% (0·20 to 0·77) per year), and other cancers (0·56% (0·40 to 0·72) per year). Second, for all five cancer categories investigated, incidence increases were similar in all four regions.

The ACCIS data reinforce the notion that the slow but steady increase in the incidence of many childhood cancers reported since the 1970s is real, and that it is particularly notable for leukaemia.2, 3, 4 Unfortunately, epidemiological studies so far have not identified clear risk factors for childhood cancer. Therefore, hypotheses for reasons underlying the increasing cancer incidence in children remain largely speculative. However, relating the ACCIS data to the tremendous changes in environmental, living, socioeconomic, and public health conditions in European countries after World War 2 might provide clues as to the nature of these reasons. The timing and magnitude of these changes have been heterogeneous across Europe, and this heterogeneity does not correlate with the similarities in childhood cancer incidence trends observed in the four European regions. For example, before the fall of the Berlin Wall in 1989, access to high-end medical technology was restricted in east Europe. If improvements in medical imaging technology were the reason underlying the increased incidence trends of tumours of the CNS, then a lower incidence of these tumours should have been observed in the early 1990s in east Europe compared with west Europe, with a catch up in incidence afterwards. However, incidence and trends from 1991 to 2010 were about the same in east Europe and west Europe.1

Another contrast across European regions concerns the use of pesticides in agriculture. In 2014, the quantities of pesticide sales per capita were about three times greater in Spain, Italy, and France than in Sweden or the UK.5 If increasing cancer incidence trends were due to pesticides, dissimilarities in incidence trends for leukaemia and lymphoma would be expected between European regions, which was not the case.1

A search for marked changes in health events specific to all European children rapidly identifies the dramatic decreases in mortality of children younger than 15 years after World War 2. For example, infant mortality in Europe ranged from 16·0–77·5 deaths per 1000 newborns in 1960 and dropped to 2·3–9·8 deaths per 1000 newborns in 2010—a decrease of 4·4% per year on average.5 This phenomenon leads to the hypothesis that the death of some children before they reached a certain age precluded the occurrence of cancer by that age. Death would represent a competing risk (or event) in the sense that cancer can no longer occur after death.6 Furthermore, children who would have died could also have been at higher risk of cancer for two possible reasons. First, potentially deadly diseases such as those caused by infectious agents could be involved in cancer occurrence, especially for leukaemia and lymphoma. Second, genetic makeup or a congenital malformation might increase both the susceptibility to be affected by a potentially deadly disease and the risk of cancer. These hypotheses assume that if a child escapes death either because he or she has been protected against potentially deadly diseases (eg, by vaccination or improved living conditions) or has survived its occurrence (eg, because of efficient therapy), then the risk of cancer occurrence would be greater in this child than in other children. This hypothesis could explain why the peak incidence of leukaemia has shifted towards younger ages over time.4, 7 Steadily decreasing mortality in the months following birth would increase the number of younger children at higher risk for leukaemia.

The competing risk hypothesis essentially rests on ecological observations. A way to further examine its likelihood would be to analyse cancer incidence trends using competing risk methods that take into consideration the changes in mortality over time.



Lees verder in de categorie Artikelen | Terug naar homepage | Lees de introductie