Page 134 - 2018_11-Haematologica-web

A. Nikkilä et al.
with increased risk of leukemia.7,8 Furthermore, there is reasonably consistent evidence of a slightly increased risk associated with large birth weight relative to gestational time.9 A higher risk has also been suggested for older parental age, delivery by Cesarean section, and paternal smoking.10–13 However, daycare attendance, allergic dis- eases, maternal folic acid supplementation before birth, and early immune stimulation have been suggested to reduce the risk of leukemia.14–17
Although high doses of ionizing radiation increase the risk of childhood leukemia, the magnitude of any effect from low doses remains uncertain. Some studies have suggested increased risks associated with background radiation and following x-ray examinations in utero and post-natally.18–22 Computed tomography (CT) imaging has been used for almost four decades and its frequency of utilization increased greatly during the 1980s-1990s. The annual number of scans peaked around year 2002; more recently CT scans have been partly replaced by magnetic resonance imaging in pediatric imaging, partly because of the risk of cancer from ionizing radiation.23 In 2015, 5,311 pediatric CT scans were performed in the Finnish popu- lation of 1,024,000 children under 17 years old, which is a low rate compared to that in many other countries.23,24
Four high-quality studies have investigated the associa- tion of pediatric CT scans and childhood leukemia.25–28 The interpretation of the findings must include an evalu- ation of confounding by indication, i.e. underlying condi- tions predisposing children to both CT scans and leukemia.28–30 Nevertheless, the evidence is still limited and the magnitude of the risk needs to be characterized further.
In this study, we examined the magnitude of the risk of childhood leukemia after pediatric CT examinations using a nationwide case-control design with efforts to avoid reverse causation.
Methods
We used a register-based, case-control study with individually matched controls. The key characteristics of the material have been presented previously.10 Briefly, all cases of childhood leukemia (M9800–M9948 in ICD-O-3) diagnosed in Finland dur- ing 1990–2011 (n=1,100) before the age of 15 years were identi- fied from the Finnish Cancer Registry (Figure 1). Three controls were individually matched, by sex and year of birth, for each case from the Population Register Center. In all analyses, a 2-
Figure 1. Flow chart depicting the selection of study subjects. The flow of cases is represented on a white background and the controls on a light gray background. The necessary exclusions are shown in red boxes. Dashed lines represent the linking of the study subjects with the CT scans collected.
1874
haematologica | 2018; 103(11)