Non-parametric estimation of the age-at-onset distribution from a cross-sectional sample.

We propose and study a simple and innovative non-parametric approach to estimate the age-of-onset distribution for a disease from a cross-sectional sample of the population that includes individuals with prevalent disease. First, we estimate the joint distribution of two event times, the age of disease onset and the survival time after disease onset. We accommodate that individuals had to be alive at the time of the study by conditioning on their survival until the age at sampling. We propose a computationally efficient expectation-maximization (EM) algorithm and derive the asymptotic properties of the resulting estimates. From these joint probabilities we then obtain non-parametric estimates of the age-at-onset distribution by marginalizing over the survival time after disease onset to death. The method accommodates categorical covariates and can be used to obtain unbiased estimates of the covariate distribution in the source population. We show in simulations that our method performs well in finite samples even under large amounts of truncation for prevalent cases. We apply the proposed method to data from female participants in the Washington Ashkenazi Study to estimate the age-at-onset distribution of breast cancer associated with carrying BRCA1 or BRCA2 mutations.

A population-based study of the prevalence and risk factors of low-grade Plasmodium falciparum malaria infection in children aged 0-15 years old in northern Tanzania.

OBJECTIVES: Northern Tanzania experiences significant malaria-related morbidity and mortality, but accurate data are scarce. We update the data on patterns of low-grade Plasmodium falciparum malaria infection among children in northern Tanzania. METHODS: Plasmodium falciparum malaria prevalence (pfPR) was assessed in a representative sample of 819 children enrolled in 94 villages in northern Tanzania between October 2015 and August 2016, using a complex survey design. Individual- and household-level risk factors for pfPR were elicited using structured questionnaires. pfPR was assessed using rapid diagnostic tests (RDTs) and thick film microscopy (TFM). Associations with pfPR, based on RDT, were assessed using adjusted odds ratios (aOR) and confidence intervals (CI) from weighted survey logistic regression models. RESULTS: Plasmodium falciparum malaria prevalence (pfPR) was 39.5% (95% CI: 31.5, 47.5) by RDT and 33.4% (26.0, 40.6) by TFM. pfPR by RDT was inversely associated with higher-education parents, especially mothers (5-7 years of education: aOR 0.55; 95% CI: 0.31, 0.96, senior secondary education: aOR 0.10; 95% CI: 0.02, 0.55), living in a house near the main road (aOR 0.34; 95% CI: 0.15, 0.76), in a larger household (two rooms: aOR 0.40; 95% CI: 0.21, 0.79, more than two rooms OR 0.35; 95% CI: 0.20, 0.62). Keeping a dog near or inside the house was positively associated with pfPR (aOR 2.01; 95% CI: 1.26, 3.21). pfPR was not associated with bed-net use or indoor residual spraying. CONCLUSIONS: Nearly 40% of children in northern Tanzania had low-grade malaria antigenaemia. Higher parental education and household metrics but not mosquito bed-net use were inversely associated with pfPR.

OBJECTIFS: La Tanzanie connaît une morbidité et une mortalité importantes liées au paludisme, mais les données précises sont rares. Nous mettons à jour les données sur les profils en matière d'infection par le paludisme à Plasmodium falciparum de faible grade chez les enfants dans le nord de la Tanzanie. MÉTHODES: La prévalence du paludisme à P. falciparum (pfPR) a été évaluée sur un échantillon représentatif de 819 enfants inscrits dans 94 villages dans le nord de la Tanzanie entre octobre 2015 et août 2016, à l'aide d'un plan d'enquête complexe. Des facteurs de risque de pfPR au niveau individuel et au niveau du ménage ont été déterminés à l'aide de questionnaires structurés. La pfPR a été évaluée à l'aide de tests de diagnostic rapides (TDR) et de microscopie à film épais (TFM). Les associations avec la pfPR, sur la base des TDR, ont été évaluées à l'aide des rapports de cotes ajustés (aOR) et des intervalles de confiance (IC) de modèles de régression logistique de surveillances pondérées. RÉSULTATS: La pfPR était de 39,5% (IC95%: 31,5-47,5) avec les TDR et de 33,4% (26,0-40,6) avec la TFM. La pfPR par les TDR était inversement associée aux parents avec un niveau d'éducation plus élevé, en particulier les mères (5-7 ans d'études: aOR: 0,55; IC95%: 0,31-0,96, enseignement secondaire supérieur: aOR: 0,10; IC95%: 0,02-0,55), vivre dans une maison proche de la route principale (aOR: 0,34; IC95%: 0,15-0,76), dans un ménage plus grand (2 chambres: aOR: 0,40; IC95%: 0,21-0,79, plus de 2 pièces aOR: 0,35; IC95%: 0,20-0,62). Garder un chien près ou à l'intérieur de la maison était positivement associé à la pfPR (aOR: 2,01; IC95%: 1,26-3,21). La pfPR n'était pas associée à l'utilisation de moustiquaire ou à la pulvérisation de résidus à l'intérieur. CONCLUSIONS: Près de 40% des enfants dans nord de la Tanzanie présentaient une antigénémie paludéenne de faible grade. Un niveau d'éducation parentale plus élevé et les indicateurs du ménage, mais pas l'utilisation de moustiquaires, étaient inversement associés à la pfPR.

Sex steroid hormones in relation to Barrett's esophagus: an analysis of the FINBAR Study.

Previously, we observed strong positive associations between circulating concentrations of free testosterone and free dihydrotestosterone (DHT) in relation to Barrett's esophagus in a US male military population. To replicate these findings, we conducted a second study of sex steroid hormones and Barrett's esophagus in the Factors Influencing the Barrett/Adenocarcinoma Relationship (FINBAR) Study based in Northern Ireland and Ireland. We used mass spectrometry to quantitate EDTA plasma concentrations of nine sex steroid hormones and ELISA to quantitate sex hormone-binding globulin in 177 male Barrett's esophagus cases and 185 male general population controls within the FINBAR Study. Free testosterone, free DHT, and free estradiol were estimated using standard formulas. Multivariable logistic regression estimated odds ratios (OR) and 95% confidence intervals (95%CI) of associations between exposures and Barrett's esophagus. While plasma hormone and sex hormone-binding globulin concentrations were not associated with all cases of Barrett's esophagus, we did observe positive associations with estrogens in younger men (e.g. estrone + estradiol ORcontinuous per ½IQR  = 2.92, 95%CI:1.08, 7.89), and free androgens in men with higher waist-to-hip ratios (e.g. free testosterone ORcontinuous per ½IQR  = 2.71, 95%CI:1.06, 6.92). Stratification by body mass index, antireflux medications, and geographic location did not materially affect the results. This study found evidence for associations between circulating sex steroid hormones and Barrett's esophagus in younger men and men with higher waist-to-hip ratios. Further studies are necessary to elucidate whether sex steroid hormones are consistently associated with esophageal adenocarcinogenesis.

Risk of Renal Cell Carcinoma Among Kidney Transplant Recipients in the United States.

Renal cell carcinoma (RCC) is a common malignancy following kidney transplantation. We describe RCC risk and examine RCC risk factors among US kidney recipients (1987-2010). The Transplant Cancer Match Study links the US transplant registry with 15 cancer registries. Standardized incidence ratios (SIRs) were used to compare RCC risk (overall and for clear cell [ccRCC] and papillary subtypes) to the general population. Associations with risk factors were assessed using Cox models. We identified 683 RCCs among 116 208 kidney recipients. RCC risk was substantially elevated compared with the general population (SIR 5.68, 95% confidence interval 5.27-6.13), especially for papillary RCC (SIR 13.3 versus 3.98 for ccRCC). Among kidney recipients, RCC risk was significantly elevated for blacks compared to whites (hazard ratio [HR] 1.50) and lower in females than males (HR 0.56). RCC risk increased with prolonged dialysis preceding transplantation (p-trend < 0.0001). Risk was variably associated for RCC subtypes with some medical conditions that were indications for transplantation: ccRCC risk was reduced with polycystic kidney disease (HR 0.54), and papillary RCC was increased with hypertensive nephrosclerosis (HR 2.02) and vascular diseases (HR 1.86). In conclusion, kidney recipients experience substantially elevated risk of RCC, especially for papillary RCC, and multiple factors contribute to these cancers.

Comparison of Cancer Diagnoses Between the US Solid Organ Transplant Registry and Linked Central Cancer Registries.

US transplant centers are required to report cancers in transplant recipients to the transplant network. The accuracy and completeness of these data, collected in the Scientific Registry of Transplant Recipients (SRTR), are unknown. We compared diagnoses in the SRTR and 15 linked cancer registries for colorectal, liver, lung, breast, prostate and kidney cancers; melanoma; and non-Hodgkin lymphoma (NHL). Among 187 384 transplants, 9323 cancers were documented in the SRTR or cancer registries. Only 36.8% of cancers were in both, with 47.5% and 15.7% of cases additionally documented solely in cancer registries or the SRTR, respectively. Agreement between the SRTR and cancer registries varied (kappa = 0.28 for liver cancer and kappa = 0.52-0.66 for lung, prostate, kidney, colorectum, and breast cancers). Upon evaluation, some NHLs documented only in cancer registries were identified in the SRTR as another type of posttransplant lymphoproliferative disorder. Some SRTR-only cases were explained by miscoding (colorectal cancer instead of anal cancer, metastases as lung or liver cancers) or missed matches with cancer registries, partly due to recipients' outmigration from catchment areas. Estimated sensitivity for identifying cancer was 52.5% for the SRTR and 84.3% for cancer registries. In conclusion, SRTR cancer data are substantially incomplete, limiting their usefulness for surveillance and research.

Brain tumors in patients with myotonic dystrophy: a population-based study.

BACKGROUND AND PURPOSE: Patients with myotonic dystrophy (DM) are at high risk of brain cancer. This study describes the spectrum of brain neoplasms in DM patients. METHODS: Data from 1119 DM patients identified from the National Swedish Patient Register between 1987 and 2007 were linked to the National Cancer and the Cause of Death Registers. Standardized incidence ratios (SIRs) and cumulative incidence to quantify the relative and absolute risks of brain neoplasms were calculated and the Kaplan-Meier estimator was used for survival analysis. Patient follow-up started at birth or the age at the start of Swedish cancer registration (1 January 1958) and ended at the age of brain neoplasm diagnosis, death or on 31 December 2007. RESULTS: Twenty patients developed brain neoplasm during follow-up {median age 53, range 2-76 years, accounting for a five-fold excess risk of brain tumors during the patient lifetime [SIR = 5.4, 95% confidence interval (CI) 3.4-8.1, P = 1 × 10(-5) ]}. Astrocytoma was the most common histological subtype (n = 16, 80%), and almost all cases (n = 19) developed after age 20. No statistically significant differences in gender-specific risks (SIR in men 6.3 and in women 3.8, P-heterogeneity 0.46) were observed. After accounting for competing mortality related to DM, the cumulative incidence of brain neoplasms reached 2.9% (95% CI 1.8%-4.7%) by age 70. Five-year survival after brain tumor diagnosis was 52% (95%CI 29%-75%) overall (number at risk 8) and 34% (95% CI 26%-47%) for malignant neoplasms (number at risk 5). CONCLUSION: Despite the high relative risk of DM-related brain tumors, the absolute risk is modest. Nonetheless, careful evaluation of DM patients with new central nervous system symptoms is warranted.

On the use and misuse of scalar scores of confounders in design and analysis of observational studies.

We assess the asymptotic bias of estimates of exposure effects conditional on covariates when summary scores of confounders, instead of the confounders themselves, are used to analyze observational data. First, we study regression models for cohort data that are adjusted for summary scores. Second, we derive the asymptotic bias for case-control studies when cases and controls are matched on a summary score, and then analyzed either using conditional logistic regression or by unconditional logistic regression adjusted for the summary score. Two scores, the propensity score (PS) and the disease risk score (DRS) are studied in detail. For cohort analysis, when regression models are adjusted for the PS, the estimated conditional treatment effect is unbiased only for linear models, or at the null for non-linear models. Adjustment of cohort data for DRS yields unbiased estimates only for linear regression; all other estimates of exposure effects are biased. Matching cases and controls on DRS and analyzing them using conditional logistic regression yields unbiased estimates of exposure effect, whereas adjusting for the DRS in unconditional logistic regression yields biased estimates, even under the null hypothesis of no association. Matching cases and controls on the PS yield unbiased estimates only under the null for both conditional and unconditional logistic regression, adjusted for the PS. We study the bias for various confounding scenarios and compare our asymptotic results with those from simulations with limited sample sizes. To create realistic correlations among multiple confounders, we also based simulations on a real dataset.

Elevated risk of venous but not arterial thrombosis in Waldenström macroglobulinemia/lymphoplasmacytic lymphoma.

BACKGROUND: Many malignancies, including multiple myeloma and its precursor, monoclonal gammopathy of unknown significant, are associated with an elevated risk of thromboembolism. There is limited information on the risk of thrombosis in patients with Waldenström macroglobulinemia (WM) and lymphoplasmacytic lymphoma (LPL). OBJECTIVES: To assess the risk of venous and arterial thrombosis in WM/LPL patients in a large population-based cohort study in Sweden. PATIENTS/METHODS: A total of 2190 patients with WM/LPL and 8086 matched controls were identified through Swedish registers between 1987 and 2005. Information on occurrence of venous and arterial thrombosis after the diagnosis of WM/LPL was obtained through the centralized Swedish Patient Register, with follow-up to 2006. Cox regression models were used to calculate hazard ratios (HRs) and 95% confidence intervals (CIs). RESULTS: Patients with WM/LPL had a significantly increased risk of venous thrombosis and the highest risk was observed during the first year following diagnosis (HR = 4.0, 95% CI 2.5-6.4). The risk was significantly elevated 5 (HR = 2.3, 95% CI 1.7-3.0) and 10 years after diagnosis (HR = 2.0, 95% CI 1.6-2.5). There was no increased risk of arterial thrombosis during any period of follow-up time (10-year HR = 1.0, 95% CI 0.9-1.1). CONCLUSIONS: Venous thrombosis is a significant cause of morbidity in patients with WM/LPL. The potential role of thromboprophylaxis in WM/LPL, especially during the first year after diagnosis and in patients treated with thrombogenic agents, needs to be assessed to further improve outcome in WM/LPL patients.

Low to moderate alcohol consumption on serum vitamin D and other indicators of bone health in postmenopausal women in a controlled feeding study.

The objective of this study was to evaluate the effect of 8 weeks of no alcohol, low (1 drink or 15 g/day) and moderate (2 drinks or 30 g/day) alcohol consumption on markers of bone health: fasting serum 25-hydroxy vitamin D (25(OH)D), osteocalcin (OC), bone-specific alkaline phosphatase (BSAP), urine deoxypyridinoline (DPD) and helical peptide (HP) in postmenopausal women (n=51). Compared with no alcohol, 1 or 2 drinks/day for 8 weeks had no significant impact on any of the bone markers. Within each alcohol group, obese women had significantly lower serum 25(OH)D and higher DPD concentrations than normal weight women. Season significantly affected the concentrations of serum 25(OH)D, but there was no significant interaction between alcohol and season on serum 25(OH)D concentrations. Low or moderate alcohol consumption for 8 weeks had no significant impact on markers of bone health in postmenopausal women.

Cancers among US organ donors: a comparison of transplant and cancer registry diagnoses.

Transmission of cancer is a life-threatening complication of transplantation. Monitoring transplantation practice requires complete recording of donor cancers. The US Scientific Registry of Transplant Recipients (SRTR) captures cancers in deceased donors (beginning in 1994) and living donors (2004). We linked the SRTR (52,599 donors, 110,762 transplants) with state cancer registries. Cancer registries identified cancers in 519 donors: 373 deceased donors (0.9%) and 146 living donors (1.2%). Among deceased donors, 50.7% of cancers were brain tumors. Among living donors, 54.0% were diagnosed after donation; most were cancers common in the general population (e.g. breast, prostate). There were 1063 deceased donors with cancer diagnosed in the SRTR or cancer registry, and the SRTR lacked a cancer diagnosis for 107 (10.1%) of these. There were 103 living donors with cancer before or at donation, diagnosed in the SRTR or cancer registry, and the SRTR did not have a cancer diagnosis for 43 (41.7%) of these. The SRTR does not record cancers after donation in living donors and so missed 81 cancers documented in cancer registries. In conclusion, donor cancers are uncommon, but lack of documentation of some cases highlights a need for improved ascertainment and reporting by organ procurement organizations and transplant programs.

Iodine-131 dose-dependent gene expression: alterations in both normal and tumour thyroid tissues of post-Chernobyl thyroid cancers.

BACKGROUND: A strong, consistent association between childhood irradiation and subsequent thyroid cancer provides an excellent model for studying radiation carcinogenesis. METHODS: We evaluated gene expression in 63 paired RNA specimens from frozen normal and tumour thyroid tissues with individual iodine-131 (I-131) doses (0.008-8.6 Gy, no unirradiated controls) received from Chernobyl fallout during childhood (Ukrainian-American cohort). Approximately half of these randomly selected samples (32 tumour/normal tissue RNA specimens) were hybridised on 64 whole-genome microarrays (Agilent, 4 × 44 K). Associations between I-131 dose and gene expression were assessed separately in normal and tumour tissues using Kruskal-Wallis and linear trend tests. Of 155 genes significantly associated with I-131 after Bonferroni correction and with ≥2-fold increase per dose category, we selected 95 genes. On the remaining 31 RNA samples these genes were used for validation purposes using qRT-PCR. RESULTS: Expression of eight genes (ABCC3, C1orf9, C6orf62, FGFR1OP2, HEY2, NDOR1, STAT3, and UCP3) in normal tissue and six genes (ANKRD46, CD47, HNRNPH1, NDOR1, SCEL, and SERPINA1) in tumour tissue was significantly associated with I-131. PANTHER/DAVID pathway analyses demonstrated significant over-representation of genes coding for nucleic acid binding in normal and tumour tissues, and for p53, EGF, and FGF signalling pathways in tumour tissue. CONCLUSION: The multistep process of radiation carcinogenesis begins in histologically normal thyroid tissue and may involve dose-dependent gene expression changes.

Hormonal risk factors and invasive epithelial ovarian cancer risk by parity.

BACKGROUND: Recent studies have suggested that several ovarian cancer risk factors differ by parity status, but these findings have not been confirmed. We evaluated whether known risk factors of ovarian cancer differ between nulliparous and parous women using data from two large prospective cohorts. METHODS: Data from the National Institutes of Health-AARP Diet and Health Study and the Prostate, Lung, Colorectal and Ovarian Cancer Screening Trial were combined for this analysis. Cox regression models were used to estimate associations with ovarian cancer risk. Risk heterogeneity by parity status was assessed using likelihood-ratio tests. RESULTS: Among the 125 437 women included in the analysis, there were 16 589 (13%) nulliparous women and 108 848 (87%) parous women. Of the 623 women diagnosed with invasive epithelial ovarian cancer, 102 (16%) were nulliparous and 521 (84%) were parous. While parity reduced ovarian cancer risk, no differences were found for other risk factors by parity. Among ever users of hormone therapy, body mass index suggestively increased the risk of ovarian cancer by 1.5-fold in nulliparous but not parous women (P-heterogeneity=0.08). CONCLUSION: While nulliparous women have higher ovarian cancer risk than parous women, our findings suggest that the relative effects of most other risk factors do not differ by parity.

Risk of lymphoma subtypes after solid organ transplantation in the United States.

BACKGROUND: Solid organ transplant recipients have high risk of lymphomas, including non-Hodgkin lymphoma (NHL) and Hodgkin lymphoma (HL). A gap in our understanding of post-transplant lymphomas involves the spectrum and associated risks of their many histologic subtypes. METHODS: We linked nationwide data on solid organ transplants from the US Scientific Registry of Transplant Recipients (1987-2008) to 14 state and regional cancer registries, yielding 791 281 person-years of follow-up for 19 distinct NHL subtypes and HL. We calculated standardised incidence ratios (SIRs) and used Poisson regression to compare SIRs by recipient age, transplanted organ, and time since transplantation. RESULTS: The risk varied widely across subtypes, with strong elevations (SIRs 10-100) for hepatosplenic T-cell lymphoma, Burkitt's lymphoma, NK/T-cell lymphoma, diffuse large B-cell lymphoma, and anaplastic large-cell lymphoma (both systemic and primary cutaneous forms). Moderate elevations (SIRs 2-4) were observed for HL and lymphoplasmacytic, peripheral T-cell, and marginal zone lymphomas, but SIRs for indolent lymphoma subtypes were not elevated. Generally, SIRs were highest for younger recipients (<20 years) and those receiving organs other than kidneys. CONCLUSION: Transplant recipients experience markedly elevated risk of a distinct spectrum of lymphoma subtypes. These findings support the aetiologic relevance of immunosuppression for certain subtypes and underscore the importance of detailed haematopathologic workup for transplant recipients with suspected lymphoma.

Common genetic variants in the 9p21 region and their associations with multiple tumours.

BACKGROUND: The chromosome 9p21.3 region has been implicated in the pathogenesis of multiple cancers. METHODS: We systematically examined up to 203 tagging SNPs of 22 genes on 9p21.3 (19.9-32.8 Mb) in eight case-control studies: thyroid cancer, endometrial cancer (EC), renal cell carcinoma, colorectal cancer (CRC), colorectal adenoma (CA), oesophageal squamous cell carcinoma (ESCC), gastric cardia adenocarcinoma and osteosarcoma (OS). We used logistic regression to perform single SNP analyses for each study separately, adjusting for study-specific covariates. We combined SNP results across studies by fixed-effect meta-analyses and a newly developed subset-based statistical approach (ASSET). Gene-based P-values were obtained by the minP method using the Adaptive Rank Truncated Product program. We adjusted for multiple comparisons by Bonferroni correction. RESULTS: Rs3731239 in cyclin-dependent kinase inhibitors 2A (CDKN2A) was significantly associated with ESCC (P=7 × 10(-6)). The CDKN2A-ESCC association was further supported by gene-based analyses (Pgene=0.0001). In the meta-analyses by ASSET, four SNPs (rs3731239 in CDKN2A, rs615552 and rs573687 in CDKN2B and rs564398 in CDKN2BAS) showed significant associations with ESCC and EC (P<2.46 × 10(-4)). One SNP in MTAP (methylthioadenosine phosphorylase) (rs7023329) that was previously associated with melanoma and nevi in multiple genome-wide association studies was associated with CRC, CA and OS by ASSET (P=0.007). CONCLUSION: Our data indicate that genetic variants in CDKN2A, and possibly nearby genes, may be associated with ESCC and several other tumours, further highlighting the importance of 9p21.3 genetic variants in carcinogenesis.

On estimating average effects for multiple treatment groups.

We propose to estimate average exposure (or treatment) effects from observational data for multiple exposure groups by fitting an approximation of the marginal sample distribution of the response variable in each exposure group to the data. The marginal sample distribution is a function of the true distribution of the response variable in the population and the assignment rule governing the allocation of the subjects to different exposure groups. The assignment rule can depend on the response variable in addition to measured covariates, and hence the method is appropriate even when the assumption of ignorable treatment assignment is not justified. We estimate the exposure effects are estimated based on the population expectation (PE) of the outcome variable. We compare the PE approach with an instrumental variable method and with several other methods including propensity score based approaches that assume ignorable assignment mechanisms. We evaluate the robustness of the PE method under model misspecifications and illustrate it using data from a study of the impact of soy consumption on urinary concentrations of estrogen and estrogen metabolites in Asian American women. Published 2012. This article is a US Government work and is in the public domain in the USA.

Variance computations for functional of absolute risk estimates.

We present a simple influence function based approach to compute the variances of estimates of absolute risk and functions of absolute risk. We apply this approach to criteria that assess the impact of changes in the risk factor distribution on absolute risk for an individual and at the population level. As an illustration we use an absolute risk prediction model for breast cancer that includes modifiable risk factors in addition to standard breast cancer risk factors. Influence function based variance estimates for absolute risk and the criteria are compared to bootstrap variance estimates.

Determinants of Epstein-Barr virus-positive gastric cancer: an international pooled analysis.

BACKGROUND: Meta-analyses of the published literature indicate that about 9% of gastric cancers contain Epstein-Barr virus (EBV), with consistent and significant differences by sex and anatomic subsite. This study aimed to identify additional determinants of EBV positivity and their joint effects. METHODS: From 15 international populations with consistent laboratory testing for EBV, we pooled individual-level data for 5081 gastric cancer cases including information on age, sex, subsite, histologic type, diagnostic stage, geographic region, and period of diagnosis. First, we combined population-specific EBV prevalence estimates using random effects meta-analysis. We then aggregated individual-level data to estimate odds ratios of EBV positivity in relation to all variables, accounting for within-population clustering. RESULTS: In unadjusted analyses, EBV positivity was significantly higher in males, young subjects, non-antral subsites, diffuse-type histology, and in studies from the Americas. Multivariable analyses confirmed significant associations with histology and region. Sex interacted with age (P=0.003) and subsite (P=0.002) such that male predominance decreased with age for both subsites. The positivity of EBV was not significantly associated with either stage or time period. CONCLUSION: Aggregating individual-level data provides additional information over meta-analyses. Distinguishing histologic and geographic features as well as interactions among age, sex, and subsite further support classification of EBV-associated gastric cancer as a distinct aetiologic entity.

Risk alleles for chronic hepatitis B are associated with decreased mRNA expression of HLA-DPA1 and HLA-DPB1 in normal human liver.

A genome-wide association study identified single nucleotide polymorphisms (SNPs) rs3077 and rs9277535 located in the 3' untranslated regions of human leukocyte antigen (HLA) class II genes HLA-DPA1 and HLA-DPB1, respectively, as the independent variants most strongly associated with chronic hepatitis B. We examined whether these SNPs are associated with mRNA expression of HLA-DPA1 and HLA-DPB1. We identified gene expression-associated SNPs (eSNPs) in normal liver samples obtained from 651 individuals of European ancestry by integrating genotype (~650 000 SNPs) and gene expression (>39 000 transcripts) data from each sample. We used the Kruskal-Wallis test to determine associations between gene expression and genotype. To confirm findings, we measured allelic expression imbalance (AEI) of complementary DNA compared with DNA in liver specimens from subjects who were heterozygous for rs3077 and rs9277535. On a genome-wide basis, rs3077 was the SNP most strongly associated with HLA-DPA1 expression (p=10(-48)), and rs9277535 was strongly associated with HLA-DPB1 expression (p=10(-15)). Consistent with these gene expression associations, we observed AEI for both rs3077 (p=3.0 × 10(-7); 17 samples) and rs9277535 (p=0.001; 17 samples). We conclude that the variants previously associated with chronic hepatitis B are also strongly associated with mRNA expression of HLA-DPA1 and HLA-DPB1, suggesting that expression of these genes is important in control of HBV.

Two criteria for evaluating risk prediction models.

We propose and study two criteria to assess the usefulness of models that predict risk of disease incidence for screening and prevention, or the usefulness of prognostic models for management following disease diagnosis. The first criterion, the proportion of cases followed PCF (q), is the proportion of individuals who will develop disease who are included in the proportion q of individuals in the population at highest risk. The second criterion is the proportion needed to follow-up, PNF (p), namely the proportion of the general population at highest risk that one needs to follow in order that a proportion p of those destined to become cases will be followed. PCF (q) assesses the effectiveness of a program that follows 100q% of the population at highest risk. PNF (p) assess the feasibility of covering 100p% of cases by indicating how much of the population at highest risk must be followed. We show the relationship of those two criteria to the Lorenz curve and its inverse, and present distribution theory for estimates of PCF and PNF. We develop new methods, based on influence functions, for inference for a single risk model, and also for comparing the PCFs and PNFs of two risk models, both of which were evaluated in the same validation data.

Prolactin serum levels and breast cancer: relationships with risk factors and tumour characteristics among pre- and postmenopausal women in a population-based case-control study from Poland.

BACKGROUND: Previous prospective studies have found an association between prolactin (PRL) levels and increased risk of breast cancer. Using data from a population-based breast cancer case-control study conducted in two cities in Poland (2000-2003), we examined the association of PRL levels with breast cancer risk factors among controls and with tumour characteristics among the cases. METHODS: We analysed PRL serum levels among 773 controls without breast cancer matched on age and residence to 776 invasive breast cancer cases with available pretreatment serum. Tumours were centrally reviewed and prepared as tissue microarrays for immunohistochemical analysis. Breast cancer risk factors, assessed by interview, were related to serum PRL levels among controls using analysis of variance. Mean serum PRL levels by tumour characteristics are reported. These associations also were evaluated using polytomous logistic regression. RESULTS: Prolactin levels were associated with nulliparity in premenopausal (P=0.05) but not in postmenopausal women. Associations in postmenopausal women included an inverse association with increasing body mass index (P=0.0008) and direct association with use of recent/current hormone therapy (P=0.0006). In case-only analyses, higher PRL levels were more strongly associated with lobular compared with ductal carcinoma among postmenopausal women (P=0.02). Levels were not different by tumour size, grade, node involvement or oestrogen receptor, progesterone receptor, or human epidermal growth factor receptor 2 status. CONCLUSIONS: Our analysis demonstrates that PRL levels are higher among premenopausal nulliparous as compared with parous women. Among postmenopausal women, levels were higher among hormone users and lower among obese women. These results may have value in understanding the mechanisms underlying several breast cancer risk factor associations.