Investigating mortality trends in Italy during the COVID-19 pandemic: life expectancy changes within provinces and vaccination campaign impact up to December 2022

OBJECTIVES

We used publicly available population data from 1 January 2019 up to 31 December 2022, to investigate mortality trends in Italy during the COVID-19 pandemic, evaluating changes in life expectancy (LE) at birth within provinces and the impact of COVID-19 vaccinations.

STUDY DESIGN

Aggregate data analysis.

METHODS

Annual period life tables were used to estimate sex-specific LEs within provinces from 2019 to 2022. We used Arriaga decomposition to analyze the contribution of age groups (<60 years and ≥60 years) to annual LE changes. We implemented a Quasi-Poisson regression model to estimate the number of averted deaths by the achieved COVID-19 vaccination rates from January 2021 up to December 2022, simulating a counterfactual scenario where vaccine doses were not administered.

RESULTS

The results revealed geographical heterogeneity in annual LE changes across Italian provinces during the pandemic. By the end of 2022, LE was below the prepandemic levels in 88% of provinces for females and in 76% for males. In addition, we estimated that the achieved vaccination rates averted 460,831 deaths (95% confidence interval: 250,976-707,920), corresponding to a 25% reduction in expected all-cause mortality.

CONCLUSIONS

Overall, the study highlighted the significant role of COVID-19 vaccinations in averting a considerable number of deaths and improving LE. However, by the end of 2022, LE had not fully recovered to prepandemic levels in many provinces. This could be attributed to concurrent factors, including enduring COVID-19 pandemic effects, intense summer heat waves and early onset of seasonal flu. Further research and continuous monitoring are essential to fully comprehend long-term mortality trends and optimize public health strategies.

Berberine phospholipid exerts a positive effect on the glycemic profile of overweight subjects with impaired fasting blood glucose (IFG): a randomized double-blind placebo-controlled clinical trial

OBJECTIVE

Berberine is a plant alkaloid known to exert positive metabolic effects. Human studies have confirmed its ability to improve the lipid and glycemic profile. This study aimed to evaluate the potential benefit of oral supplementation of Berberine PhytosomeTM (2 tablets/day, 550 mg/tablet) on the metabolic profile of subjects with impaired fasting blood glucose (IFG).

PATIENTS AND METHODS

A total of 49 overweight subjects, 28 females and 21 males, were randomly assigned to either the supplemented group (n=24) or placebo (n=25). We considered glycemia as the primary endpoint and total cholesterol, high-density lipoprotein (HDL), total cholesterol/HLD, low-density lipoprotein (LDL), LDL/HDL, triglycerides, insulin, glycated hemoglobin, Homeostasis Model Assessment (HOMA), ApoA, ApoB, ApoB/ApoA, androgen suppression treatment (AST), alternative lengthening of telomeres (ALT), gamma-glutamyl transferase (GGT), creatinine, and body composition by dual-energy X-ray absorptiometry (DXA) as secondary endpoints. These parameters have been assessed at baseline, after 30 days, and after 60 days.

RESULTS

After two months of treatment, through the use of linear mixed effect models, a statistically significant difference between supplemented and placebo groups was observed for glycemia [β=-0.2495% C.I. (-0.47; -0.06), p=0.004], total cholesterol [β=-0.25, 95% C.I. (-0.45; -0.04), p=0.05], total cholesterol/HDL [β=-0.25, 95% C.I. (-0.43; -0.06), p=0.04], triglycerides [β=-0.14, 95% C.I. (-0.25; -0.02), p=0.05], insulin [β=-1.78, 95% C.I. (-2.87; -0.66), p=0.009], ApoB/ApoA [β=-0.08, 95% C.I. (-0.13; -03), p=0.004], Visceral adipose tissue (VAT) [β=-91.50, 95% C.I. (-132.60; -48.19), p<0.0001] and fat mass [β=-945.56, 95% C.I. (-1,424.42; -441.57), p=0.004].

CONCLUSIONS

The use of berberine had no adverse events, supporting its use as a natural alternative to pharmacological therapies in the case of IFG.

Multifactorial analysis of the stochastic epigenetic variability in cord blood confirmed an impact of common behavioral and environmental factors but not of in vitro conception

Background

An increased incidence of imprint-associated disorders has been reported in babies born from assisted reproductive technology (ART). However, previous studies supporting an association between ART and an altered DNA methylation status of the conceived babies have been often conducted on a limited number of methylation sites and without correction for critical potential confounders. Moreover, all the previous studies focused on the identification of methylation changes shared among subjects while an evaluation of stochastic differences has never been conducted. This study aims to evaluate the effect of ART and other common behavioral or environmental factors associated with pregnancy on stochastic epigenetic variability using a multivariate approach.

Results

DNA methylation levels of cord blood from 23 in vitro and 41 naturally conceived children were analyzed using the Infinium HumanMethylation450 BeadChips. After multiple testing correction, no statistically significant difference emerged in the number of cord blood stochastic epigenetic variations or in the methylation levels between in vitro- and in vivo-conceived babies. Conversely, four multiple factor analysis dimensions summarizing common phenotypic, behavioral, or environmental factors (cord blood cell composition, pre or post conception supplementation of folates, birth percentiles, gestational age, cesarean section, pre-gestational mother’s weight, parents’ BMI and obesity status, presence of adverse pregnancy outcomes, mother’s smoking status, and season of birth) were significantly associated with stochastic epigenetic variability. The stochastic epigenetic variation analysis allowed the identification of a rare imprinting defect in the locus GNAS in one of the babies belonging to the control population, which would not have emerged using a classical case-control association analysis.

Conclusions

We confirmed the effect of several common behavioral or environmental factors on the epigenome of newborns and described for the first time an epigenetic effect related to season of birth. Children born after ART did not appear to have an increased risk of genome-wide changes in DNA methylation either at specific loci or randomly scattered throughout the genome. The inability to identify differences between cases and controls suggests that the number of stochastic epigenetic variations potentially induced by ART was not greater than that naturally produced in response to maternal behavior or other common environmental factors.

Electronic supplementary material

The online version of this article (10.1186/s13148-018-0510-3) contains supplementary material, which is available to authorized users.

Enrichment for Northern European-derived multiple sclerosis risk alleles in Sardinia

Background:

The list of genomic loci associated with multiple sclerosis (MS) susceptibility outside the major histocompatibility complex (MHC) in patients of Northern European (NE) ancestry has increased to 103. Despite the extraordinarily high MS prevalence in the isolated Sardinian population, the contribution of genetic risk factors to MS in Sardinia is largely not understood.

Objective:

The objective of this paper is to examine the relevance of non-MHC MS susceptibility variants in Sardinia.

Methods:

We examined a log-additive MS-specific genetic burden score (MSGB) using 110 NE-derived risk alleles in a dataset of 75 Sardinian cases, 346 Sardinian controls and 177 cases and 1967 controls from the United States (US).

Results:

Sardinian cases demonstrate a heavier non-MHC MSGB load than Sardinian controls and US cases ( p = 2E-06, p = 1E-06, respectively). Furthermore, Sardinian controls carry a heavier burden than US controls ( p = 2E-14). Our results confirm the limited ability of the 110-SNP MSGB to predict disease status in Sardinia (AUROC = 0.629).

Conclusions:

Risk alleles discovered in samples of NE ancestry are relevant to MS in Sardinia. Our results suggest a general enrichment of MS susceptibility alleles in Sardinians, encouraging the pursuit of further studies of MS in this population.

Alzheimer's Disease: From Genes to Nutrition

Alzheimer's disease (AD) is widely identified as the most common cause of sporadic dementia. Its aetiology is still debated, as despite several hypotheses, different factors seem to play a role in its establishment and development. Recent studies have proposed a possible preventing role of nutrition. The weight loss typical of earlier phase of disease and the finding of malnutrition as a common trait between patients leads to hypothesize that a supplementation of specific nutrients seems to be useful and effective in terms of improvement of cognitive functions. Malnourished patients show also altered parameters when investigating inflammation markers: for example, hyperhomocysteinemia is a typical finding in elderly affected by dementia, and it can be prevented and corrected by using a proper nutrients supplementation. Pro-inflammatory state can be reduced with supplementation of polyunsaturated fatty acids, vitamins of the group B and phosphatidylserine, that can act reducing IL-1β (pro-inflammatory cytokine) and improving IL-10 (anti-inflammatory cytokine) synthesis. While investigating the role of nutrition, it seems to be deeply linked with genetic; a genetic onset AD-related could be latent and can be influenced by nutritional attitude. AD can be considered a sort of latent clinical condition that would disclose or not, depending also on micro-environment and nutritional parameters. The genetic expression can be influenced by assumptions or not of specific nutrients, with the promotion of different pro- or anti-inflammatory settings. The specific role of each micronutrient (in particular vitamins) and trace elements still needs to be punctuated, as they are involved in a pool of different reactions. Also genes acts not independently but in an interconnected pattern, in which the role of a single gene needs to be cleared, depending on others. This complex system of predisposing conditions and a possible role of nutrition as modulator of the inflammatory state is the object of this review.

Localization and behaviors in null mice suggest that ASIC1 and ASIC2 modulate responses to aversive stimuli

Acid-sensing ion channels (ASICs) generate H(+) -gated Na(+) currents that contribute to neuronal function and animal behavior. Like ASIC1, ASIC2 subunits are expressed in the brain and multimerize with ASIC1 to influence acid-evoked currents and facilitate ASIC1 localization to dendritic spines. To better understand how ASIC2 contributes to brain function, we localized the protein and tested the behavioral consequences of ASIC2 gene disruption. For comparison, we also localized ASIC1 and studied ASIC1(-/-) mice. ASIC2 was prominently expressed in areas of high synaptic density, and with a few exceptions, ASIC1 and ASIC2 localization exhibited substantial overlap. Loss of ASIC1 or ASIC2 decreased freezing behavior in contextual and auditory cue fear conditioning assays, in response to predator odor and in response to CO2 inhalation. In addition, loss of ASIC1 or ASIC2 increased activity in a forced swim assay. These data suggest that ASIC2, like ASIC1, plays a key role in determining the defensive response to aversive stimuli. They also raise the question of whether gene variations in both ASIC1 and ASIC2 might affect fear and panic in humans.

Anti-beta 2 glycoprotein I antibodies and the risk of myocardial infarction in young premenopausal women

BACKGROUND

Contrasting data have been reported on the association between the presence of anti-phospholipid antibodies (aPL) and arterial thrombotic events, particularly those in coronary arteries. This discrepancy is perhaps related to the confounding effect of traditional risk factors. Among them, coronary atherosclerosis appears to be the most important in studies conducted in middle-aged and elderly patients.

OBJECTIVE

To minimize such confounding effects, a multicenter case-control study on the association between aPL and myocardial infarction (MI) was carried out in a rare cohort of young premenopausal women.

METHODS

We evaluated 172 cases hospitalized for a first MI before the age of 45 years and 172 controls individually matched with cases for age, sex and geographical origin. Clinical and laboratory data were collected and levels of anti-cardiolipin (aCL), anti-beta2 glycoprotein I (anti-beta2GPI) and anti-nuclear antibodies (ANA) were measured.

RESULTS

A significant association between MI and IgG/IgM anti-beta2GPI antibodies was observed; the results were confirmed after adjusting for smoking and hypertension (anti-beta2GPI IgG OR = 2.47, 95% CI 1.81-3.38; anti-beta2GPI IgM 4th quartile OR 3.68, 95% CI 1.69-8.02). The association between anti-beta2GPI antibodies and MI was detected in both subgroups with and without coronary artery stenosis. Whereas the association of aCL IgG with MI was modest, ANA showed no significant association with MI. No aPL were found in unselected patients (mainly males) who recently developed acute MI.

CONCLUSIONS

Anti-beta2GPI antibodies are a significant risk factor for MI in young premenopausal women independently of other risk factors, including the degree of coronary artery stenosis.

Bayesian mapping of multiple sclerosis prevalence in the province of Pavia, northern Italy

The geographical analysis of a disease risk is particularly difficult when the disease is non-frequent and the area units are small. The practical use of the Bayesian modelling, instead of the classical frequentist one, is applied to study the geographical variation of multiple sclerosis (MS) across the province of Pavia, Northern Italy. 464 MS-affected individuals resident in the province of Pavia were identified on December 31st 2000. The overall prevalence was 94 per 100,000 inhabitants. This estimate indicates an increasing MS prevalence in the province, in accordance with the vast majority of the Italian areas where prevalence studies have been repeated. We mapped the geographical variation of MS prevalence across the 190 communes of the province both with a classical approach and a Bayesian approach. The frequentist approach produced an extremely dishomogeneous map, while the Bayesian map was much smoother and more interpretable. Our study underlines the usefulness of Bayesian methods to obtain reliable maps of disease prevalence and to identify possible clusters of disease where to carry out further epidemiological investigations.

Tissue plasminogen activator antigen is strongly associated with myocardial infarction in young women

Women who develop acute myocardial infarction (AMI) at a young age have fewer classical risk factors and less coronary stenosis than older women. In this rare population, it is plausible that a heightened hemostatic system may play an important mechanistic role in thrombus formation and in the development of AMI. We chose to investigate whether or not there is an association between premature AMI and the plasma concentrations of five hemostatic measurements that had been previously established as risk factors for AMI, and of the inflammation marker C-reactive protein (CRP). Women who had survived AMI at the age of 45 years or less (n = 141) were drawn from those admitted to 125 Italian coronary care units over a 3-year period. In them, and in an equal number of controls, plasma levels of immunoreactive tissue plasminogen activator (tPA), plasminogen activation inhibitor 1 (PAI-1), von Willebrand factor (VWF), fibrinogen, D-dimer and CRP were measured. Higher levels of VWF, fibrinogen, CRP and tPA were associated with AMI. After adjustment for both classical and hemostatic risk factors, only tPA maintained an independent association with AMI: the odds ratios (taken as an index of relative risk) for tPA values in the middle and higher tertiles were 2.86 (CI 1.63-5.02) and 8.18 (CI 2.66-25.20), respectively. In conclusion, there is a strong association between non-fatal AMI and increased plasma levels of tPA antigen. This finding is thought to be the expression of a reduced rather than enhanced fibrinolytic activity.

Bayesian trio models for association in the presence of genotyping errors

Errors in genotyping can greatly affect family-based association studies. If a mendelian inconsistency is detected, the family is usually removed from the analysis. This reduces power, and may introduce bias. In addition, a large proportion of genotyping errors remain undetected, and these also reduce power. We present a Bayesian framework for performing association studies with SNP data on samples of trios consisting of parents with an affected offspring, while allowing for the presence of both detectable and undetectable genotyping errors. This framework also allows for the inclusion of missing genotypes. Associations between the SNP and disease were modelled in terms of the genotypic relative risks. The performances of the analysis methods were investigated under a variety of models for disease association and genotype error, looking at both power to detect association and precision of genotypic relative risk estimates. As expected, power to detect association decreased as genotyping error probability increased. Importantly, however, analyses allowing for genotyping error had similar power to standard analyses when applied to data without genotyping error. Furthermore, allowing for genotyping error yielded relative risk estimates that were approximately unbiased, together with 95% credible intervals giving approximately correct coverage. The methods were also applied to a real dataset: a sample of schizophrenia cases and their parents genotyped at SNPs in the dysbindin gene. The analysis methods presented here require no prior information on the genotyping error probabilities, and may be fitted in WinBUGS.

An ecologic study of geographical variation in multiple sclerosis risk in central Sardinia, Italy

We carried out an ecological study in the most archaic area of Sardinia to obtain a reliable estimate of the prevalence of multiple sclerosis (MS) and to investigate the geographical variation in the prevalence across the 100 administrative communes. To estimate the area-specific prevalence rate, we adopted a Bayesian approach that makes it possible to filter out the random variation from the estimates and to obtain a map that reflects the true geographical variation in MS prevalence. 428 resident cases were identified by the case register, including 69 multiplex families. The overall prevalence was 157 per 100,000 inhabitants. The Bayesian area-specific prevalence ranged from 143 to 262/100,000. The high prevalence and its moderate geographical variation in a genetically homogeneous population, as well as the high number of multiplex families observed in the communes with the highest prevalence, could be interpreted as representing a high susceptibility of the population to MS.

An immunogenetic map of Lombardy (Northern Italy)

For this study we consulted the Bone Marrow Donors' Registry of Lombardy (Italy) and analyzed 43937 HLA-A,B phenotypes and 13922 HLA-A,B,DR phenotypes. We estimated the HLA-A,B and HLA-A,B,DR haplotype frequencies via the maximum-likelihood method. We analyzed the genetic structure of the 11 provinces of Lombardy by means of Principal Component Analysis and Correspondence Analysis, and estimated the variety of the different haplotypes at provincial level and the percentage of unique phenotypes at village level. We found 11189 different HLA-A,B phenotypes, 661 different HLA-A,B haplotypes and more than 4000 different HLA-A,B,DR haplotypes. We identified 20 villages, in Western Lombardy, very rich in unique/rare phenotypes. Here we report a formula which allows the identification of a putative donor matched for two haplotypes with a recipient. This result may be of great importance for the genetic study of the population of Lombardy and, even more, for bone marrow transplantation programs.

Association between the ancestral haplotype HLA A30B18DR3 and multiple sclerosis in central Sardinia

Association and linkage studies have established the importance of the major histocompatibility complex (MHC) in the susceptibility for multiple sclerosis (MS). We carried out a case-control study to investigate the ancestral haplotype A30B18DR3 and MS in the Nuoro population of Sardinia, which is isolated and genetically distinct from other populations in the Mediterranean basin and characterized by genetic homogeneity, high level of inbreeding, low migration, high prevalence of MS, high frequency of the relevant haplotype, and high past malaria prevalence. Cases and controls were serologically typed for the currently recognized HLA-A, B, and DR antigens. We used a log-linear approach to fit a wide class of models. We tested our hypothesis comparing different models via a likelihood ratio test. We overcame the complication due to unknown gametic phase using expectation-maximization (EM) algorithm as the estimation method. We estimated confidence intervals for odds ratio by using a profile likelihood approach. We found that: (1) the ancestral haplotype A30B18DR3 was associated to MS after allowing for a possible stratification in cases and controls; (2) DR3 allele was conditional independent on disease status, given A30B18 haplotype; (3) there was a tendency for ORs for the high-risk haplotypes to be higher in the high malaria strata; however, this indication did not achieve statistical significance (P = 0.11).

Statistical issues in the analysis of disease mapping data

In this paper we discuss a number of issues that are pertinent to the analysis of disease mapping data. As an illustrative example we consider the mapping of larynx cancer across electoral wards in the North West Thames region of the U.K. Bayesian hierarchical models are now frequently employed to carry out such mapping. In a typical situation, a three-stage hierarchical model is specified in which the data are modelled as a function of area-specific relative risks at stage one; the collection of relative risks across the study region are modelled at stage two; and at stage three prior distributions are assigned to parameters of the stage two distribution. Such models allow area-specific disease relative risks to be 'smoothed' towards global and/or local mean levels across the study region. However, these models contain many structural and functional assumptions at different levels of the hierarchy; we aim to discuss some of these assumptions and illustrate their sensitivity. When relative risks are the endpoint of interest, it is common practice to assume that, for each of the age-sex strata of a particular area, there is a common multiplier (the relative risk) acting upon each of the stratum-specific risks in that area; we will examine this proportionality assumption. We also consider the choices of models and priors at stages two and three of the hierarchy, the effect of outlying areas, and an assessment of the level of smoothing that is being carried out. For inference, we concentrate on the description of the spatial variability in relative risks and on the association between the relative risks of larynx cancer and an area-level measure of socio-economic status.

An Italian epidemiological multicentre study of nocturnal enuresis

OBJECTIVE

To estimate the prevalence of enuresis in schoolchildren in Italy.

SUBJECTS AND METHODS

The Italian Club of Nocturnal Enuresis promoted a prevalence study of nocturnal enuresis using a self-administered questionnaire in seven cities in Northern, Central and Southern Italy. The association between enuresis and potential risk factors, e.g. a family history of enuresis, stress, socio-economic status and abnormal diurnal voiding habits, was investigated. The perceived impact on the child and on the family was also evaluated. A random-cluster sampling scheme was used to obtain a sample of primary and secondary schoolchildren from each city. One primary school and one secondary school for each socio-economic level was sampled in each city, giving a total of 42 schools surveyed; 9086 children were covered by the survey. In a cluster sampling method, the variance of prevalence is divided into two components, binomial and extra-binomial variability. Both the DSM III and DSM IV definitions of enuresis were used because at present, there is no consensus on the diagnostic criteria.

RESULTS

Completed questionnaires were received from 7012 children, an overall response rate of 77.2%. Those aged 6-14 years were analysed, restricting the sample to 6892 children. There were 250 enuretic children using the DSM III definition of enuresis and 112 using the DSM IV definition. The overall prevalence was 3.88% and showed a decreasing trend with increasing age. Bedwetting was more frequent in boys than in girls. The prevalence of enuresis was higher when the child was from a family of low socio-economic status despite the child's age group. The logistic analysis showed that familiality, stress, birthweight, age of attaining diurnal continence, soiling and, for girls, menstruation, were statistically significant variables and thus contributed to predicting the probability of bedwetting, confirming the findings of previous studies. There was a large difference in prevalence using the two DSM definitions; a high percentage of DSM III enuretic children had more than two wet nights per week.

CONCLUSION

It is important that a consensus about the 'working definitions' of enuresis is reached to avoid bias in the recruitment step, to carry out comparable epidemiological studies and to obtain adequate therapeutic responses.

The Sardinian IDDM study: 1. Epidemiology and geographical distribution of IDDM in Sardinia during 1989 to 1994

Analysis of the geographical variation of risk for a disease is a key issue in descriptive epidemiology and may provide useful suggestions for planning further studies to identify the underlying causes. We adopted a Bayesian approach to investigate the geographical distribution of insulin-dependent diabetes mellitus (IDDM) incidence rate across Sardinia. Data on incidence of IDDM in children aged under 15 years (619 IDDM cases) in Sardinia was obtained by the Sardinian Eurodiab ACE register. The overall completeness of ascertainment was: 91.3%. The average yearly standardized incidence rate for the years 1989-1994 was 33.24 per 100000 (95% C.I. 30.60, 35.88), which is the second highest in Europe after Finland. Sex and age-specific risks were higher in males than in females. Considering the variation of IDDM risk according to the age at diagnosis, the risk profile increased up to the 13th year of age for both sexes, being steeper in males. The degree of geographical variation in IDDM risk was small with a slight difference between the highest and the lowest standardized rate across the map. Indeed, even the municipalities at lowest risk in Sardinia showed a risk higher than most European countries. The Sardinian population is genetically atypical, characterized by genetic homogeneity and marked susceptibility to autoimmune diseases. Our finding of a small geographical variation within the island coupled with a marked temporal trend previously observed in data on military conscripts could be interpreted as evidence of a relatively recent environmental aetiological factor that was uniformly distributed across the island and had its effect in a genetically predisposed population.

Disease mapping with errors in covariates

We describe Bayesian hierarchical-spatial models for disease mapping with imprecisely observed ecological covariates. We posit smoothing priors for both the disease submodel and the covariate submodel. We apply the models to an analysis of insulin Dependent Diabetes Mellitus incidence in Sardinia, with malaria prevalence as a covariate.

Bayesian analysis of space-time variation in disease risk

The analysis of variation of risk for a given disease in space and time is a key issue in descriptive epidemiology. When the data are scarce, maximum likelihood estimates of the area-specific risk and of its linear time-trend can be seriously affected by random variation. In this paper, we propose a Bayesian model in which both area-specific intercept and trend are modelled as random effects and correlation between them is allowed for. This model is an extension of that originally proposed for disease mapping. It is illustrated by the analysis of the cumulative prevalence of insulin dependent diabetes mellitus as observed at the military examination of 18-year-old conscripts born in Sardinia during the period 1936-1971. Data concerning the genetic differentiation of the Sardinian population are used to interpret the results.

Bayesian analysis of space-time variation in disease risk

The analysis of variation of risk for a given disease in space and time is a key issue in descriptive epidemiology. When the data are scarce, maximum likelihood estimates of the area-specific risk and of its linear time-trend can be seriously affected by random variation. In this paper, we propose a Bayesian model in which both area-specific intercept and trend are modelled as random effects and correlation between them is allowed for. This model is an extension of that originally proposed for disease mapping. It is illustrated by the analysis of the cumulative prevalence of insulin dependent diabetes mellitus as observed at the military examination of 18-year-old conscripts born in Sardinia during the period 1936-1971. Data concerning the genetic differentiation of the Sardinian population are used to interpret the results.

Bayesian estimates of disease maps: how important are priors?

In the fully Bayesian (FB) approach to disease mapping the choice of the hyperprior distribution of the dispersion parameter is a key issue. In this context we investigated the sensitivity of the rate ratio estimates to the choice of the hyperprior via a simulation study. We also compared the performance of the FB approach to mapping disease risk to the conventional approach of mapping maximum likelihood (ML) estimates and p-values. The study was modelled on the incidence data of insulin dependent diabetes mellitus (IDDM) as observed in the communes of Sardinia.

Childhood Acute Lymphoblastic Leukemia: A Cluster in Southwestern Sardinia (Italy)

In response to public concern about an increase in the incidence of leukemia among children in southwestern Sardinia (Italy), incident cases of childhood cancer (ages 0-14) were ascertained among residents in the province of Cagliari, which comprises all of southern Sardinia, in 1974-89. Completeness of the ascertainment of leukemia cases was validated by comparison with estimates derived from official statistics of mortality and survival curves. A significant excess risk of childhood acute lymphoblastic leukemia (cALL) was found for children residing in the town of Carbonia. The risk was highest in 1983-85, when seven cases occurred versus 0.8 expected. No birth-cohort effect was observed. The cALL incidence rate was significantly higher among children born and residing in Carbonia than among children born in Carbonia but residing elsewhere. However, the cALL cases did not cluster within the town of Carbonia. The proximity of the largest industrial settlement in the region of Sardinia raised the suspicion that environmental pollution was responsible for the observed excess. Information about industrial emissions from this settlement prior to the appearance of the cALL cluster was not sufficient to reject or confirm the hypothesis.

[Bayesian analysis of ecological studies]

Studying the space-time variation of risk for a given disease may give etiological clues and suggestions for planning further studies to investigate the underlying causes. When the observed events are rare, approaches based on maximum likelihood may lead to unstable and largely uninformative estimates of risk and of its time trend due to Poisson sampling variation. In this paper we propose a general Bayesian model for analyzing the variation of risk in space and time. We applied the Bayesian model to the analysis of the geographical variation of breast cancer mortality, to an ecological study on the correlation between lung cancer mortality and degree of urbanization and industrialization and to the analysis of the space-time variation of cumulative prevalence of Insulin Dependent Diabetes Mellitus (IDDM) as observed in military examinations between 1954 and 1989.

Spatial correlation in ecological analysis

This paper presents a statistical approach, originally developed for mapping disease risk, to ecological regression analysis in the presence of spatial autocorrelated extra-Poisson variation. An insight into the effect of allowing for spatial autocorrelation on the relationship between disease rates and explanatory variables is given. Examples based on cancer frequency in Scotland and Sardinia are used to illustrate the interpretation of regression coefficient and further methodological issues.

Childhood. Leukemia in Southwest Sardinia (Italy)

Aim and background

Public concern on an increased incidence of childhood leukemia in SW Sardinia prompted the authors to an epidemiological investigation.

Methods

Incident childhood neoplasms observed in the Cagliari province (Sardinia - Italy) in 1974-1989 were registered. Expected cases of the most frequent childhood cancers were calculated for each town, based on the sex-and age-specific incidence rates in the province.

Results

An excess risk of childhood acute lymphoblastic leukemia (ALL) was observed in Carbonia, a town located in the SW part of the province. The risk was highest in 1983-85, when 7 cases occurred versus 0.8 expected (RR = 8.7; 95 % C.I. = 4.6, 16.3). No spatial clustering of ALL cases was observed within the town.

Conclusions

A significantly higher than expected incidence of childhood ALL was observed in the town of Carbonia in 1983-85. In alternative to chance, possible exposure to environmental pollutants from a near industrial settlement is discussed as the cause of the observed excess, but it is far to be proven. Other hypotheses, including a viral infection in a population with increased susceptibility, as suggested for new urban settlements, cannot be discarded.

Empirical Bayes versus fully Bayesian analysis of geographical variation in disease risk

This paper reviews methods for mapping geographical variation in disease incidence and mortality. Recent results in Bayesian hierarchical modelling of relative risk are discussed. Two approaches to relative risk estimation, along with the related computational procedures, are described and compared. The first is an empirical Bayes approach that uses a technique of penalized log-likelihood maximization; the second approach is fully Bayesian, and uses an innovative stochastic simulation technique called the Gibbs sampler. We chose to map geographical variation in breast cancer and Hodgkin's disease mortality as observed in all the health care districts of Sardinia, to illustrate relevant problems, methods and techniques.

[The use of a multistage model in the analysis of the occupational risks in those exposed to hexavalent chromium]

A brief presentation is made of the theory underlying the multistage model of Armitage and Doll and its implications in assessing the effect of exposure to a carcinogen. The theory foresees different relationships between cancer risk and the dependent time variables (time from beginning of exposure, age at start of employment, duration of exposure and time since end of exposure) according to whether the carcinogen acts in the early or late stages of the process of cell transformation leading to tumour formation. The trend in risk of death from lung cancer was studied in a cohort of workers exposed to hexavalent chromium followed up between 1948 and 1985. The results, referred to the multistage model, indicate that hexavalent chromium probably acts in the later stages of the cell transformation process; the implications of this hypothesis for prevention and/or occupational safety strategies are discussed.

Cancer mortality in an Italian rubber factory

The purpose of the study was to describe the mortality experience of an Italian cohort of rubber workers and an attempt was made to identify any occupational cancer hazards that might currently be affecting men employed in this type of work. A total of 4917 male workers who first started working in a large rubber factory between 1962 and 1972 have been followed up until 31 January 1983. The number of deaths from all causes and from malignant neoplasms was determined and compared with the expected number of deaths calculated from mortality rates for the province in which the population of the plant lived. Mortality from all causes was 85% of that expected. A slight overall excess of deaths from cancer (SMR = 119) was found; this was entirely due to the excess mortality in the 35-44 age group. In order further to evaluate the possible existence of a cancer risk SMRs were analysed by duration of exposure, time since first exposure, and for specific sites of cancer. A trend in SMR with duration of exposure was found for employees with 10 years of follow up or more. A high risk for some tumour sites emerged.