Understanding tuberculosis epidemiology using structured statistical models

Molecular epidemiological studies can provide novel insights into the transmission of infectious diseases such as tuberculosis. Typically, risk factors for transmission are identified using traditional hypothesis-driven statistical methods such as logistic regression. However, limitations become apparent in these approaches as the scope of these studies expand to include additional epidemiological and bacterial genomic data. Here we examine the use of Bayesian models to analyze tuberculosis epidemiology. We begin by exploring the use of Bayesian networks (BNs) to identify the distribution of tuberculosis patient attributes (including demographic and clinical attributes). Using existing algorithms for constructing BNs from observational data, we learned a BN from data about tuberculosis patients collected in San Francisco from 1991 to 1999. We verified that the resulting probabilistic models did in fact capture known statistical relationships. Next, we examine the use of newly introduced methods for representing and automatically constructing probabilistic models in structured domains. We use statistical relational models (SRMs) to model distributions over relational domains. SRMs are ideally suited to richly structured epidemiological data. We use a data-driven method to construct a statistical relational model directly from data stored in a relational database. The resulting model reveals the relationships between variables in the data and describes their distribution. We applied this procedure to the data on tuberculosis patients in San Francisco from 1991 to 1999, their Mycobacterium tuberculosis strains, and data on contact investigations. The resulting statistical relational model corroborated previously reported findings and revealed several novel associations. These models illustrate the potential for this approach to reveal relationships within richly structured data that may not be apparent using conventional statistical approaches. We show that Bayesian methods, in particular statistical relational models, are an important tool for understanding infectious disease epidemiology.     

Ixodes ricinus density, and distribution and prevalence of Borrelia burgdorferi sensu lato infection along an altitudinal gradient

In this study, we measured the phenology of Ixodes ricinus ticks and their infection with Borrelia burgdorferi sensu lato (sl) simultaneously along an altitudinal gradient to assess the impact of climate on the phenology of ticks and on their infection with B. burgdorferi sl. From 1999 to 2001, free-living I. ricinus ticks were collected monthly by flagging vegetation at three different altitudes (620, 740, and 900 m above sea level) on the slope of a mountain in Chaumont (Neuchatel, Switzerland). I. ricinus ticks were examined for the presence of B. burgdorferi sl by using direct fluorescent antibody assay and isolation of spirochetes. Borrelia species were characterized by polymerase chain reaction followed by restriction fragment-length polymorphism. Tick density and tick phenology varied with altitude. Although the peak tick density decreased and the onset of ticks was delayed with altitude, the phenology was much more stable among years at the highest altitudes than at the lowest. The prevalence of B. burgdorferi infection in nymphs and adults decreased with altitude. The prevalence of infection differed significantly among years, and it was significantly higher in adults (30%) than in nymphs (21%). B. burgdorferi infection in adults was positively related with adult density, but this was not observed for nymphs. Five B. burgdorferi sl genospecies were successfully isolated: B. garinii, B. burgdorferi sensu stricto, B. afzelii, B. valaisiana, and B. lusitaniae. Mixed infections were obtained from five of 140 infected ticks. The greatest diversity in Borrelia species was observed at the lowest altitude where all five Borrelia species were present, whereas at the two highest altitudes, B. lusitaniae was not observed.     

Partial NK cell tolerance induced by radioresistant host cells in rats transplanted with MHC-mismatched bone marrow

We have studied the effect of radioresistant host cells in inducing tolerance and adaptation of the MHC recognition repertoire of donor-derived NK cells in stem cell allotransplanted (allo-SCT) rats. Sub-lethally irradiated PVG.1AV1 rats (RT1(av1)) were transplanted with bone marrow from fully MHC-mismatched allotype-marked PVG.7B (RT1(c)) rats; MHC-identical PVG (RT1(c)) controls were transplanted in parallel. In the PVG.7B → PVG.1AV1 allogeneic chimeras, NK cells were donor derived and showed partial tolerance toward host cells. Allogeneic chimeras failed to efficiently reject PVG.1AV1 cells by an NK-mediated mechanism in vivo (allogeneic lymphocyte cytotoxicity), and IL-2-cultured NK cells derived from these chimeras showed diminished cytolytic activity against PVG.1AV1 cells in vitro. There were corresponding changes in the phenotype and function of the highly alloreactive Ly49i2(+) NK cells, which are specifically inhibited by a donor MHC class I ligand, RT1-A1(c). The ligand-negative host MHC haplotype apparently induced expression of a second uncharacterized inhibitory MHC receptor responsible for the partial tolerance toward host-derived cells, along with a modest increase in Ly49i2 receptor levels. The host MHC haplotype did not induce a general hyporesponsiveness in Ly49i2(+) NK cells, which showed normal activation responses in a panel of MHC congenic strains. The data suggest that the MHC constitution of radiation-resistant host cells can have permanent, albeit not fully tolerogenic, effects on the development of a functional NK repertoire following allo-SCT.     

Immigrant status, antenatal depressive symptoms, and frequency and source of violence: what's the relationship?

This study describes the prevalence of violence during pregnancy and examines the association between the experience of violence since the beginning of pregnancy and the prevalence of antenatal depressive symptoms while taking into account immigrant status. Cross-sectional study including 5,162 pregnant women attending Montreal hospitals for antenatal care was conducted, with 1,400 being born outside of Canada. CES-D scale was used to evaluate depression at 24-26?weeks of pregnancy. The Abuse Assessment Screen scale was used to determine the frequency and severity of violence since the beginning of pregnancy. Relationship with abuser was also considered. All modeling was done using logistic regressions. Threats were the most frequent type of violence, with 63?% happening more than once. Long-term immigrant women reported the highest prevalence of all types of violence (7.7?%). Intimate partner violence (IPV) (15?%) was most frequently reported among the poorest pregnant women. Strong associations exist between more than one episode of abuse and depression (POR?=?5.21 [3.73; 7.23], and IPV and depression [POR?=?5.81 [4.19; 8.08]. Immigrant status did not change the associations between violence and depression. Violence against pregnant women is not rare in Canada, and it is associated with antenatal depressive symptoms. These findings support future development of perinatal screening for violence, follow-up, and a culturally sensitive referral system.     

Protein expression levels of carcinoembryonic antigen (CEA) in Danish ovarian cancer patients: from the Danish 'MALOVA'ovarian cancer study

AIMS: To determine the variation in expression of carcinoembryonic antigen (CEA) in 760 epithelial ovarian tumours from Denmark, and to correlate expression with clinicopathological parameters and prognosis for the disease. METHODS: Using tissue arrays (TA), we analysed CEA expression in tissues from 189 women diagnosed with low malignant potential ovarian tumours (LMP, borderline ovarian tumours) and 571 women diagnosed with ovarian cancer (OC). RESULTS: Using 30% as the cut-off level for CEA over-expression, 18% of LMPs and 4% of OCs were positive. A higher proportion of mucinous tumours were positive compared with other histological subtypes (p<0.00001). A univariate survival analysis suggested a shorter disease specific survival for patients with 30% or higher CEA expression in the tumour tissue (p = 0.004). In a Cox survival analysis, which included 569 OC cases subgrouped by stage (I to IV), the highest CEA expression compared with no expression was found to be a prognostic factor (level 3 versus negative: HR = 2.12, 95%CI 1.11-4.05). FIGO stage, residual tumour after primary surgery, age at diagnosis, other histological types versus serous adenocarcinoma and low versus high histological grade tumours were also prognostic factors. CONCLUSION: These data suggest that CEA expression is an independent prognostic factor for mucinous OC in Danish patients.     

Clostridium difficile infection after allogeneic hematopoietic stem cell transplantation: incidence, risk factors, and outcome

Clostridium difficile (C. difficile) infection was observed in 13% of recipients after hematopoietic stem cell transplantation (HSCT), mainly in the first month posttransplantation. Risk factors were cord blood as the source of stem cells, acute graft-versus-host disease (GVHD), and total body irradiation (TBI). No association was found with an increased risk of mortality. The purpose of this study was to evaluate the incidence, risk factors, and outcome of C. difficile infection (CDI) after HSCT. We conducted a single-center, retrospective, cohort study on all patients who received an allogeneic HSCT from January 2004 to December 2007. All patients with diarrhea in the first year after HSCT were tested for the presence of C. difficile in stools. Among the 407 assessable patients, 53 presented at least 1 CDI in the first year post-HSCT. The total incidence rate?was 5.6 cases of CDI per 10,000 patient-days. Fifty percent of cases were diagnosed in the first month after HSCT, and 95% occurred during the first 6 months. Fewer than 5% of patients with CDI had severe diarrhea and severe complications were never observed. TBI in the conditioning regimen, cord blood as the source of stem cells, and acute graft-versus-host disease (aGVHD) were independently associated with CDI. Six patients (11%) had a recurrence of CDI. Four patients required second-line treatment with vancomycin. With a median follow-up of 22 months, the 2-year overall survival rates were similar between patients who presented a CDI and those who did not. CDI was observed in approximately 13% of recipients after HSCT, mainly in the first month posttransplantation and was associated with CB, aGVHD, and TBI. CDI was not associated either with severe complications or with an increased risk of mortality in this large cohort of patients.     

Functional polymorphisms in the LTF gene and risk of coronary artery stenosis

Plasma lactoferrin concentrations are increased in patients with coronary artery stenosis. We investigated the effects of LTF gene polymorphisms in 305 healthy blood donors and their associations with coronary artery stenosis in 236 patients admitted for coronary angiography. Lactoferrin concentrations were determined by enzyme immunoassay. Genotyping was performed by polymerase chain reaction and DNA sequencing of LTF exons 2 and 4. In the blood donors, the deletion variant of rs10662431 and the G allele of rs1126478 were associated with higher plasma lactoferrin concentrations. The G allele of rs1126478 was more frequent in patients with significant coronary artery stenosis (p = 0.018, p value limit for significance by permutation = 0.030). The association remained significant in logistic regression with adjustment for clinical risk factors (odds ratio 2.485 [95% confidence interval 1.116-5.536], p = 0.026), but was weakened upon the inclusion of plasma lactoferrin (odds ratio 2.295 [0.949-5.550], p = 0.064). Current evidence indicates that rs1126478 affects the antibacterial effect of lactoferrin and that lactoferrin is involved in lipid metabolism. The relationships among lactoferrin genotypes, lactoferrin concentrations, and clinical factors on the risk for atherosclerosis are not fully understood, but the G allele of rs1126478 seems to have a detrimental effect in a European population.     

Immunoglobulin G antibody profiles against Anopheles salivary proteins in domestic animals in Senegal

Although domestic animals may not be permissive for Plasmodium, they could nevertheless play a role in the epidemiology of malaria by attracting Anopheles away from humans. To investigate interactions between domestic animals and mosquitoes, we assayed immunoglobulin G (IgG) antibodies directed against the salivary proteins of Anopheles gambiae in domestic animals living in Senegalese villages where malaria is endemic. By Western blotting, sera from bovines (n=6), ovines (n=36), and caprines (n=36) did not react with Anopheles whole saliva. In contrast, equine sera recognized proteins in both saliva and salivary gland extracts. Two of the major immunogens (32 and 72 kDa) were also reactive in extracts from other major mosquito genera (Aedes and Culex), but reactions toAnopheles-specific antigens were detected in 12 of 17 horses. These data suggest that horses strongly react to Anopheles bites, and further experiments on horses are warranted to investigate the impact of this domestic animal species on the transmission of human malaria.     

Birthweight,Childhood Body Mass Index,Height and Growth,and Risk of Polycystic Ovary Syndrome

IntroductionAdult obesity is linked with polycystic ovary syndrome (PCOS), but the importance of body size at ages before PCOS is diagnosed is unknown.ObjectiveTo investigate associations between a woman''s own birthweight, childhood body mass index (BMI), height and growth patterns in relation to her risk of PCOS.MethodsWe included 65,665 girls from the Copenhagen School Health Records Register, born in the period 1960–1996, with information on birthweight and measured weight and height at the ages of 7–13 years. Overweight was defined using International Obesity Task Force (IOTF) criteria. From the Danish National Patient Register, 606 women aged 15–50 years were identified. Hazard ratios (HRs) and 95% confidence intervals (CIs) were estimated by Cox regression analysis.ResultsBirthweight was not associated with PCOS. At the age of 7–13 years, girls with overweight had a higher risk of developing PCOS than girls without overweight; HR 2.83 (95% CI 2.34–3.42) at age 7 years and 2.99 (95% CI 2.38–3.76) at age 13 years. Furthermore, girls with overweight at both 7 and 13 years had a higher risk of developing PCOS than girls without overweight or overweight at only one age. Height was positively associated with PCOS risk at all ages. Girls who were persistently tall or changed from tall to average height had a higher risk of developing PCOS than girls with average height growth.ConclusionOverweight and tall stature in childhood are positively associated with PCOS risk, but birthweight is not.