Maria A. Karalexi

Childhood central nervous system tumours: Incidence and time trends in 13 Southern and Eastern European cancer registries

AIM Following completion of the first 5-year nationwide childhood (0-14 years) registration in Greece, central nervous system (CNS) tumour incidence rates are compared with those of 12 registries operating in 10 Southern-Eastern European countries. METHODS All CNS tumours, as defined by the International Classification of Childhood Cancer (ICCC-3) and registered in any period between 1983 and 2014 were collected from the collaborating cancer registries. Data were evaluated using standard International Agency for Research on Cancer (IARC) criteria. Crude and age-adjusted incidence rates (AIR) by age/gender/diagnostic subgroup were calculated, whereas time trends were assessed through Poisson and Joinpoint regression models. RESULTS 6062 CNS tumours were retrieved with non-malignant CNS tumours recorded in eight registries; therefore, the analyses were performed on 5191 malignant tumours. Proportion of death certificate only cases was low and morphologic verification overall high; yet five registries presented >10% unspecified neoplasms. The male/female ratio was 1.3 and incidence decreased gradually with age, apart from Turkey and Ukraine. Overall AIR for malignant tumours was 23/10(6) children, with the highest rates noted in Croatia and Serbia. A statistically significant AIR increase was noted in Bulgaria, whereas significant decreases were noted in Belarus, Croatia, Cyprus and Serbia. Although astrocytomas were overall the most common subgroup (30%) followed by embryonal tumours (26%), the latter was the predominant subgroup in six registries. CONCLUSION Childhood cancer registration is expanding in Southern-Eastern Europe. The heterogeneity in registration practices and incidence patterns of CNS tumours necessitates further investigation aiming to provide clues in aetiology and direct investments into surveillance and early tumour detection.

Prelabor cesarean delivery and early-onset acute childhood leukemia risk.

The long-term impact of cesarean delivery (CD) on the health of the offspring is being explored methodically. We sought to investigate the effect of birth by (a) prelabor and (b) during-labor CD on the risk of early-onset (⩽3 years) acute lymphoblastic leukemia (ALL), specifically of its prevailing precursor B (B-ALL) subtype. A total of 1099 incident cases of ALL (957 B-ALL), 131 of acute myeloid leukemia (AML), and their 1 : 1 age-matched and sex-matched controls, derived from the Nationwide Registry for Childhood Hematological Malignancies (1996–2013), were analyzed using multivariate regression models. A null association was found between prelabor and/or during labor CD and either ALL (B-ALL) or AML in the 0–14 age range. By contrast, birth by CD increased significantly the risk of early-onset ALL [odds ratioCD (ORCD)=1.57, 95% confidence interval (CI): 1.10–2.24] mainly on account of prelabor CD (ORprelaborCD=1.66, 95% CI: 1.13–2.43). The respective figures were even higher for the early-onset precursor B-ALL (ORCD=1.66, 95% CI: 1.15–2.40 and ORprelaborCD=1.79, 95% CI: 1.21–2.66), whereas no association emerged for early-onset AML. Prelabor CD, which deprives exposure of the fetus/infant to the presumably beneficial effect of stress hormones released in both vaginal labor and during labor CD, was associated exclusively with an increased risk of early-onset ALL, particularly the precursor B-ALL subtype. If confirmed, these adverse long-term outcomes of CD may point to re-evaluation of prelabor CD practices and prompt scientific discussion on the best ways to simulate the effects of vaginal delivery, such as a precesarean induction of labor.

Albuminuria in Association with Cognitive Function and Dementia: A Systematic Review and Meta-Analysis

Cerebral microvascular disease is considered to contribute to cognitive dysfunction. We opted to explore whether albuminuria, a marker of systemic microangiopathy, is associated with cognitive impairment, dementia, and cognitive function.

Childhood central nervous system tumour mortality and survival in Southern and Eastern Europe (1983-2014 ): Gaps persist across 14 cancer registries.

AIM Childhood central nervous system (CNS) tumour registration and control programs in Southern and Eastern Europe remain thin, despite the lethal nature of the disease. Mortality/survival data were assembled to estimate the burden of malignant CNS tumours, as well as the potential role of sociodemographic survival determinants across 14 cancer registries of this region. METHODS Average age-adjusted mortality rates were calculated, whereas time trends were quantified through Poisson and Joinpoint regressions. Kaplan-Meier curves were derived for the maximum and the more recent (10 and 5 year) registration periods. Multivariate Cox regression models were used to assess demographic and disease-related determinants. RESULTS Variations in mortality (8-16 per million) and survival (5-year: 35-69%) were substantial among the participating registries; in most registries mortality trend was stable, whereas Bulgaria, having the highest starting rate, experienced decreasing annual mortality (-2.4%, p=0.001). A steep decrease in survival rates was evident before the second year of follow-up. After controlling for diagnostic subgroup, age, gender and diagnostic year, Greece seemed to present higher survival compared with the other contributing registries, although the follow-up period was short. Irrespective of country, however, rural residence was found to impose substantial adverse repercussions on survival (hazard ratio (HR): 1.2, 95% confidence interval (CI): 1.1-1.4). CONCLUSION Cross-country mortality and survival variations possibly reflect suboptimal levels of health care delivery and cancer control in some regions of Southern and Eastern Europe, notwithstanding questionable death certification patterns or follow-up procedures. Continuous childhood cancer registration and linkage with clinical data are prerequisite for the reduction of survival inequalities across Europe.

Circulating adiponectin levels in relation to carotid atherosclerotic plaque presence, ischemic stroke risk, and mortality: A systematic review and meta-analyses

BACKGROUND Low circulating levels of adiponectin, an anti-inflammatory and vasculoprotective adipokine, are associated with obesity, type 2 diabetes, and atherosclerotic disease. Presence of unstable plaques in the carotid artery is a known etiological factor causing ischemic strokes. Herein, we systematically reviewed the association between circulating adiponectin and progression of carotid atherosclerotic disease, particularly evaluating the occurrence of (1) carotid atherosclerotic plaques, (2) ischemic stroke, and (3) mortality in subjects who suffered a previous ischemic stroke. METHODS Medline, Embase, Biosis, Scopus, Web of Science, and Pubmed were searched for published studies and conference abstracts. The effect size and 95% confidence intervals (CIs) of the individual studies were pooled using fixed-effect or random-effect models. The quality of the eligible studies was evaluated using the Newcastle-Ottawa quality assessment scale. Sensitivity, subgroup, and meta-regression analyses were performed to address the impact of various risk factors on the association between adiponectin and ischemic stroke risk. RESULTS Twelve studies fulfilled the inclusion criteria for 3 independent meta-analyses. The association of increasing circulating adiponectin levels (5μg/mL-increment) with presence of carotid plaque was not conclusive (n=327; OR: 1.07; 95% CI: 0.85-1.35; 2 studies), whereas high adiponectin levels showed a significant 8% increase in risk of ischemic stroke (n=13,683; 7 studies), with a more sizable association observed among men compared to women. HDL was observed to have a marginal effect on the association between adiponectin and ischemic stroke, while other evaluated parameters were not found to be effect modifiers. A non-significant association of adiponectin with mortality was yielded (n=663; OR: 2.58; 95% CI: 0.69-9.62; 3 studies). Although no publication bias was evident, there was significant between-study heterogeneity in most analyses. CONCLUSION It appears that the direction of the relationship between adiponectin and carotid atherosclerotic plaque presence is dependent on the duration, severity, and nature of the underlying disease, while increased adiponectin levels were associated with an increase in risk for ischemic stroke. Lastly, the results from the mortality meta-analysis remain inconclusive. Future properly designed studies are necessary to further elucidate the role of adiponectin on atherosclerotic plaque development, and its related outcomes.

Patterns of cellular phone use among young people in 12 countries: Implications for RF exposure

Characterizing exposure to radiofrequency (RF) fields from wireless telecommunications technologies during childhood and adolescence is a research priority in investigating the health effects of RF. The Mobi-Expo study aimed to describe characteristics and determinants of cellular phone use in 534 young people (10-24years) in 12 countries. The study used a specifically designed software application installed on smartphones to collect data on the use of wireless telecommunications devices within this age group. The role of gender, age, maternal education, calendar period, and country was evaluated through multivariate models mutually adjusting for all variables. Call number and duration were higher among females compared to males (geometric mean (GM) ratio 1.17 and 1.42, respectively), among 20-24year olds compared to 10-14year olds (GM ratio 2.09 and 4.40, respectively), and among lowest compared to highest social classes (GM ratio 1.52 and 1.58, respectively). The number of SMS was higher in females (GM ratio 1.46) and the middle age group (15-19year olds: GM ratio 2.21 compared to 10-14year olds) and decreased over time. Data use was highest in the oldest age group, whereas Wi-Fi use was highest in the middle age group. Both data and Wi-Fi use increased over time. Large differences in the number and duration of calls, SMS, and data/Wi-Fi use were seen by country, with country and age accounting for up to 50% of the variance. Hands-free and laterality of use did not show significant differences by sex, age, education, study period, or country. Although limited by a convenience sample, these results provide valuable insights to the design, analysis, and interpretation of future epidemiological studies concerning the health effects of exposure resulting from cellular phone use in young people. In addition, the information provided by this research may be used to design strategies to minimize RF exposure.

Parental alcohol consumption and risk of leukemia in the offspring: a systematic review and meta-analysis

Parental alcohol consumption before and during pregnancy has been linked to adverse outcomes in the offspring including leukemogenesis. We, therefore, aimed to systematically assess and quantitatively synthesize published data on the association of paternal consumption during preconception and maternal consumption during pregnancy with leukemia risk in childhood (0–14 years). Following Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, we searched PubMed (until February 2016) and the reference lists of the relevant studies. Observational studies examining the association between parental alcohol consumption and childhood leukemia were considered eligible. Data extracted from 39 case–control studies (over 16 000 leukemia cases and 30 000 controls) were pooled and summary-effect estimates were calculated. Subgroup analyses were carried out by main acute leukemia type [lymphoblastic or myeloid), cytogenetics/genetic polymorphisms, and specific alcohol beverages. We found a statistically significant dose–response association of any level of maternal alcohol consumption compared with nondrinking during pregnancy exclusively with acute myeloid leukemia (AML) [odds ratio (OR)moderate consumption: 1.64, 95% confidence intervals (CIs): 1.23–2.17 and ORhigh consumption: 2.36, 95% CI: 1.60–3.49]. In contrast, no association of paternal preconception consumption with any leukemia type was noted. In beverage-specific analyses, only a positive association of maternal wine drinking with childhood AML was found, which was more pronounced in analyses including only studies on infant leukemia (ORwine: 2.12, 95% CI: 1.16–3.90). The largest ever meta-analysis shows a sizeable, statistically significant dose–response association of maternal alcohol consumption during index pregnancy with AML risk. Future research exploring the role of genetic polymorphisms is anticipated to shed light on the underlying pathophysiology.

Mortality and survival patterns of childhood lymphomas: geographic and age-specific patterns in Southern-Eastern European and SEER/US registration data.

Childhood (0‐14 years) lymphomas, nowadays, present a highly curable malignancy compared with other types of cancer. We used readily available cancer registration data to assess mortality and survival disparities among children residing in Southern‐Eastern European (SEE) countries and those in the United States. Average age‐standardized mortality rates and time trends of Hodgkin (HL) and non‐Hodgkin (NHL; including Burkitt [BL]) lymphomas in 14 SEE cancer registries (1990‐2014) and the Surveillance, Epidemiology, and End Results Program (SEER, United States; 1990‐2012) were calculated. Survival patterns in a total of 8918 cases distinguishing also BL were assessed through Kaplan‐Meier curves and multivariate Cox regression models. Variable, rather decreasing, mortality trends were noted among SEE. Rates were overall higher than that in SEER (1.02/106), which presented a sizeable (−4.8%, P = .0001) annual change. Additionally, remarkable survival improvements were manifested in SEER (10 years: 96%, 86%, and 90% for HL, NHL, and BL, respectively), whereas diverse, still lower, rates were noted in SEE. Non‐HL was associated with a poorer outcome and an amphi‐directional age‐specific pattern; specifically, prognosis was inferior in children younger than 5 years than in those who are 10 to 14 years old from SEE (hazard ratio 1.58, 95% confidence interval 1.28‐1.96) and superior in children who are 5 to 9 years old from SEER/United States (hazard ratio 0.63, 95% confidence interval 0.46‐0.88) than in those who are 10 to 14 years old. In conclusion, higher SEE lymphoma mortality rates than those in SEER, but overall decreasing trends, were found. Despite significant survival gains among developed countries, there are still substantial geographic, disease subtype‐specific, and age‐specific outcome disparities pointing to persisting gaps in the implementation of new treatment modalities and indicating further research needs.

Methodological Remarks Regarding the Meta-Analysis on Possible Associations of Maternal Factors During Pregnancy With the Risk of Childhood Acute Lymphoblastic Leukemia.

To the Editor: We read with great interest the recent metaanalysis by Yan et al.[1] on the associations of maternal factors during pregnancy and the risk of childhood acute lymphoblastic leukemia (ALL). Yet, certain methodological shortcomings, overlooked in this meta-analysis, need to be addressed. Exclusion of overlapping studies, namely, studies that have been performed on the same population during time intervals overlapping each other, is a prerequisite in the quantitative synthesis of studies included in a meta-analysis, as to avoid introducing replicate datasets in the synthesized scientific evidence.[2] The authors stated that the issue of overlap was addressed by retaining only the largest or themost recent study; nevertheless, in the forest plots provided by Yan et al. several publications springing from the same population during coinciding time periods were included as independent studies. Most strikingly, in the forest plot for birth order (Fig. 2), the AusALL study, a case–control study conducted during 2003–2007[3] was represented by five studies [Bailey et al. (2011), Milne et al. (2009), Milne et al. (2010), Milne et al. (2012), and Milne et al. (2013)] along with twoGreek studies [Diamantaras et al. (2013) and Lariou et al. (2013)] reporting on the same study population of the Nationwide Registry of Childhood Hematological Malignancies (NARECHEM) during the period 1999–2003;[4] likewise, in both included Californian studies [Ma et al. (2005) and Oksuzyan et al. (2013)], the Californian Cancer Registrywas the source of cases and the study periods coincided.[5] Similarly, in the forest plot examining maternal age (Supplementary Fig. S2), six studies [(Bartley et al. (2005), Chang et al. (2006), Kwan et al. (2005), Kwan et al. (2009), Ma et al. (2005a), and Ma et al. (2005b)], overlapped completely with the study by Oksuzyan et al.,[5] three Greek studies [(Diamantaras et al. (2013), Lariou et al. (2013), and Petridou et al. (2005)] were based on overlapping populations during the same time period and five studies [Bailey et al. (2010), Bailey et al. (2011), Milne et al. (2009), Milne et al. (2013), and Reid et al. (2011)] were actually different publications of the Aus-ALL study for the period 2003– 2007.[3] Moreover, as studies published up to 2014 were included, it is indirectly inferred that the end of search date, not disclosed in the paper, was 2014. If this is indeed the case, multiple potentially eligible studies were not included; not even, a large US multicenter study by Johnson et al. comprising over 4,000 childhood ALL cases, [6] published back in 2009. In conclusion, repeated inclusion of the same study populations via multiple publications as well as non-identification of eligible studies during the search and selection of studies process, may have considerably distorted the pooled effect estimates in the recent meta-analysis addressing exposures, as maternal age, birth order, and maternal education in association with childhood ALL. An elaborate strategy employing snowball procedures seems required in this context, as the respective variables are often used in multivariate models, aiming to investigate other exposures.

History of Maternal Fetal Loss and Childhood Leukaemia Risk in Subsequent Offspring: Differentials by Miscarriage or Stillbirth History and Disease Subtype

BACKGROUND Despite the putative intrauterine origins of childhood (0-14 years) leukaemia, it is complex to assess the impact of perinatal factors on disease onset. Results on the association of maternal history of fetal loss (miscarriage/stillbirth) with specific disease subtypes in the subsequent offspring are in conflict. We sought to investigate whether miscarriage and stillbirth may have different impacts on the risk of acute lymphoblastic leukaemia (ALL) and of its main immunophenotypes (B-cell and T-cell ALL), as contrasted to acute myeloid leukaemia (AML). METHODS One thousand ninety-nine ALL incidents (957 B-ALL) and 131 AML cases along with 1:1 age and gender-matched controls derived from the Nationwide Registry for Childhood Hematological Malignancies and Brain Tumors (1996-2013) were studied. Multivariable regression models were used to assess the roles of previous miscarriage(s) and stillbirth(s) on ALL (overall, B-, T-ALL) and AML, controlling for potential confounders. RESULTS Statistically significant exposure and disease subtype-specific associations of previous miscarriage(s) exclusively with AML [odds ratio (OR) 1.67, 95% confidence interval (CI) 1.00, 2.81] and stillbirth(s) with ALL [OR 4.82, 95% CI 1.63, 14.24] and B-ALL particularly, emerged. CONCLUSION Differential pathophysiological pathways pertaining to genetic polymorphisms or cytogenetic aberrations are likely to create hostile environments leading either to fetal loss or the development of specific leukaemia subtypes in subsequent offspring, notably distinct associations of maternal miscarriage history confined to AML and stillbirth history confined to ALL (specifically B-ALL). If confirmed and further supported by studies revealing underlying mechanisms, these results may shed light on the divergent leukemogenesis processes.