Evaluation of the eradication program forAmblyomma variegatum (Acari: Ixodidae) on Puerto Rico

A cooperative effort between the United States Department of Agriculture, Animal and Plant Health Inspection Service, Veterinary Services, the Agricultural Research Service, and the Department of Agriculture, Commonwealth Government of Puerto Rico has been successful in eardicatingAmblyomma variegatum (Fabricius) from the islands of Puerto Rico and Vieques (an island municipality of Puerto Rico). Application of acaricides to livestock and dogs at 14-day intervals for an 18-month period eliminated foci of tick infestation on 188 farms in four different locations without additional aerial or ground treatment. Reasons for the success and some of the costs associated with the eradication program on Puerto Rico are presented.     

A spatial micro-simulation analysis of methane emissions from Irish agriculture

Using micro-simulation modelling techniques this paper examines methane emissions across Irish farms. The effects of a carbon equivalent tax on average family farm income are analysed at both the farm and regional level. The spatial micro-simulation model developed uses a technique called simulated annealing to match the Irish Census of Agriculture data to a National Farm Survey. The main advantage of the spatial micro-simulation approach is the fact that it allows one to account for the heterogeneity in the farm population across space. The results of the modelling process are presented using GIS mapping techniques and highlight the fact that there would be significant regional variation in the burden of an agricultural tax that was based on a rate per unit of methane emissions. The results also demonstrate that if the methane tax revenue raised was redistributed in the form of an environmental subsidy to farmers participating in an agri-environmental scheme (the Rural Environment Protection Scheme (REPS)) it would encourage farmers to participate in the scheme and could also have the effect of moving low income farms up the earnings distribution ladder.     

A Modeling Framework to Describe the Transmission of Bluetongue Virus within and between Farms in Great Britain

Background

Recently much attention has been given to developing national-scale micro-simulation models for livestock diseases that can be used to predict spread and assess the impact of control measures. The focus of these models has been on directly transmitted infections with little attention given to vector-borne diseases such as bluetongue, a viral disease of ruminants transmitted by Culicoides biting midges. Yet BT has emerged over the past decade as one of the most important diseases of livestock.

Methodology/Principal Findings

We developed a stochastic, spatially-explicit, farm-level model to describe the spread of bluetongue virus (BTV) within and between farms. Transmission between farms was modeled by a generic kernel, which includes both animal and vector movements. Once a farm acquired infection, the within-farm dynamics were simulated based on the number of cattle and sheep kept on the farm and on local temperatures. Parameter estimates were derived from the published literature and using data from the outbreak of bluetongue in northern Europe in 2006. The model was validated using data on the spread of BTV in Great Britain during 2007. The sensitivity of model predictions to the shape of the transmission kernel was assessed.

Conclusions/Significance

The model is able to replicate the dynamics of BTV in Great Britain. Although uncertainty remains over the precise shape of the transmission kernel and certain aspects of the vector, the modeling approach we develop constitutes an ideal framework in which to incorporate these aspects as more and better data become available. Moreover, the model provides a tool with which to examine scenarios for the spread and control of BTV in Great Britain.     

County-level characteristics associated with incidence,late-stage incidence,and mortality from screenable cancers

BackgroundCancer screening differs by rurality and racial residential segregation, but the relationship between these county-level characteristics is understudied. Understanding this relationship and its implications for cancer outcomes could inform interventions to decrease cancer disparities.MethodsWe linked county-level information from national data sources: 2008–2012 cancer incidence, late-stage incidence, and mortality rates (for breast, cervical, and colorectal cancer) from U.S. Cancer Statistics and the National Death Index; metropolitan status from U.S. Department of Agriculture; residential segregation derived from American Community Survey; and prevalence of cancer screening from National Cancer Institute’s Small Area Estimates. We used multivariable, sparse Poisson generalized linear mixed models to assess cancer incidence, late-stage incidence, and mortality rates by county-level characteristics, controlling for density of physicians and median household income.ResultsCancer incidence, late-stage incidence, and mortality rates were 6–18% lower in metropolitan counties for breast and colorectal cancer, and 2–4% lower in more segregated counties for breast and colorectal cancer. Generally, reductions in cancer associated with residential segregation were limited to non-metropolitan counties. Cancer incidence, late-stage incidence, and mortality rates were associated with screening, with rates for corresponding cancers that were 2–9% higher in areas with more breast and colorectal screening, but 2–15% lower in areas with more cervical screening.DiscussionLower cancer burden was observed in counties that were metropolitan and more segregated. Effect modification was observed by metropolitan status and county-level residential segregation, indicating that residential segregation may impact healthcare access differently in different county types. Additional studies are needed to inform interventions to reduce county-level disparities in cancer incidence, late-stage incidence, and mortality.     

Using calibration weighting to adjust for nonresponse under a plausible model

When we estimate the population total for a survey variableor variables, calibration forces the weighted estimates of certaincovariates to match known or alternatively estimated populationtotals called benchmarks. Calibration can be used to correctfor sample-survey nonresponse, or for coverage error resultingfrom frame undercoverage or unit duplication. The quasi-randomizationtheory supporting its use in nonresponse adjustment treats responseas an additional phase of random sampling. The functional formof a quasi-random response model is assumed to be known, itsparameter values estimated implicitly through the creation ofcalibration weights. Unfortunately, calibration depends uponknown benchmark totals while the covariates in a plausible modelfor survey response may not be the benchmark covariates. Moreover,it may be prudent to keep the number of covariates in a responsemodel small. We use calibration to adjust for nonresponse whenthe benchmark model and covariates may differ, provided thenumber of the former is at least as great as that of the latter.We discuss the estimation of a total for a vector of surveyvariables that do not include the benchmark covariates, butthat may include some of the model covariates. We show how tomeasure both the additional asymptotic variance due to the nonresponsein a calibration-weighted estimator and the full asymptoticvariance of the estimator itself. All variances are determinedwith respect to the randomization mechanism used to select thesample, the response model generating the subset of sample respondents,or both. Data from the U.S. National Agricultural StatisticalService's 2002 Census of Agriculture and simulations are usedto illustrate alternative adjustments for nonresponse. The paperconcludes with some remarks about adjustment for coverage error.     

Presence and Persistence of Coxiella burnetii in the Environments of Goat Farms Associated with a Q Fever Outbreak

Q fever is a zoonotic disease caused by inhalation of the bacterium Coxiella burnetii. Ruminant livestock are common reservoirs for C. burnetii, and bacteria present in aerosols derived from the waste of infected animals can infect humans. The significance of infection from material deposited in the environment versus transmission directly from infected animals is not known. In 2011, an outbreak of Q fever cases on farms in Washington and Montana was associated with infected goats. A study was undertaken to investigate the quantity and spatial distribution of C. burnetii in the environment of these goat farms. Soil, vacuum, and sponge samples collected on seven farms epidemiologically linked to the outbreak were tested for the presence of C. burnetii DNA by quantitative PCR. Overall, 70.1% of the samples were positive for C. burnetii. All farms had positive samples, but the quantity of C. burnetii varied widely between samples and between farms. High quantities of C. burnetii DNA were in goat housing/birthing areas, and only small quantities were found in samples collected more than 50 m from these areas. Follow-up sampling at one of the farms 1 year after the outbreak found small quantities of C. burnetii DNA in air samples and large quantities of C. burnetii persisting in soil and vacuum samples. The results suggest that the highest concentrations of environmental C. burnetii are found in goat birthing areas and that contamination of other areas is mostly associated with human movement.     

Expert Opinion on the Perceived Effectiveness and Importance of On-Farm Biosecurity Measures for Cattle and Swine Farms in Switzerland

Biosecurity is crucial for safeguarding livestock from infectious diseases. Despite the plethora of biosecurity recommendations, published scientific evidence on the effectiveness of individual biosecurity measures is limited. The objective of this study was to assess the perception of Swiss experts about the effectiveness and importance of individual on-farm biosecurity measures for cattle and swine farms (31 and 30 measures, respectively). Using a modified Delphi method, 16 Swiss livestock disease specialists (8 for each species) were interviewed. The experts were asked to rank biosecurity measures that were written on cards, by allocating a score from 0 (lowest) to 5 (highest). Experts ranked biosecurity measures based on their importance related to Swiss legislation, feasibility, as well as the effort required for implementation and the benefit of each biosecurity measure. The experts also ranked biosecurity measures based on their effectiveness in preventing an infectious agent from entering and spreading on a farm, solely based on transmission characteristics of specific pathogens. The pathogens considered by cattle experts were those causing Bluetongue (BT), Bovine Viral Diarrhea (BVD), Foot and Mouth Disease (FMD) and Infectious Bovine Rhinotracheitis (IBR). Swine experts expressed their opinion on the pathogens causing African Swine Fever (ASF), Enzootic Pneumonia (EP), Porcine Reproductive and Respiratory Syndrome (PRRS), as well as FMD. For cattle farms, biosecurity measures that improve disease awareness of farmers were ranked as both most important and most effective. For swine farms, the most important and effective measures identified were those related to animal movements. Among all single measures evaluated, education of farmers was perceived by the experts to be the most important and effective for protecting both Swiss cattle and swine farms from disease. The findings of this study provide an important basis for recommendation to farmers and policy makers.     

Extensive Aggregations of Wild Fish at Coastal Sea-Cage Fish Farms

We present evidence of a largely undocumented environmental effect of coastal sea-cage fish farms on wild fish. We estimated the total abundance and biomass of wild fish aggregated in the immediate vicinity of nine fish farms in the Mediterranean Sea and one farm off the east coast of Australia. Estimates of wild fish aggregations ranged from 2000 to 86000 individuals and from 100 kg to 38.5 tons of fish per farm and were always greater than control locations. Particularly large aggregations (>30000 fish, > 12 tons) occurred at half of the farms. Aggregations were temporally stable for weeks to months and most wild fish associated with farms (88%) were of adult size. Potential effects of such large aggregations of wild fish in the immediate vicinity of fish farms include increased vulnerability to fishing and pathogen transfer between caged and wild fish. We suggest specific legislation should be enacted wherever large aggregations of wild fish occur around fish farms to enhance the positive and reduce the negative effects of association.     

Area-Based Socioeconomic Position and Adult Glioma: A Hierarchical Analysis of Surveillance Epidemiology and End Results Data

Background

Glioma rates vary by demographic factors and geo-political boundaries and this variation suggests higher glioma rates in groups of higher socioeconomic position. The primary goal of this analysis is to investigate the relationship between glioma and county socioeconomic position using U.S. Surveillance Epidemiology and End Results (SEER) data.

Methods

Cases were individuals 25+ years diagnosed with glioma between 2000 and 2006 and residing within the SEER-17 catchment area. County-, sex-, race-, age-specific rates were created in order to investigate individual-level associations (population data from U.S. Census 2000). A Bayesian hierarchical Poisson spatial conditionally autoregressive (CAR) model was utilized to simultaneously estimate individual- and county-level associations while controlling for county spatial dependence.

Results

Those residing in counties of the second, third, and fourth highest quartiles of socioeconomic position have glioma incidence rates that are 1.10 (95% CI: 1.02,1.19), 1.11 (95% CI: 1.02,1.20), 1.14 (95% CI: 1.05,1.23) times that of the first quartile, respectively. A CAR model properly controlled for error spatial dependence. Investigated lag times suggest year 2000 census data yields superior model fit.

Conclusion

Demographically adjusted rates of glioma are elevated in counties of higher socioeconomic position. More well-grounded theory concerning the glioma-socioeconomic position association along with socioeconomic data collected at multiple levels is recommended for future studies investigating this relationship.     

Spatial risk assessment of livestock exposure to pumas in Patagonia, Argentina

Livestock predation and associated human-carnivore conflicts are increasing worldwide and require the development of methods and concepts for risk assessment and conflict management. Here we use knowledge on habitat preference and distribution of pumas and provide a first assessment of the spatial risk of livestock to puma depredation in Patagonian ranches, Argentina. In an initial step, we developed a rule-based habitat model in a Geographic Information System (GIS) to predict the distribution of puma habitat at a regional scale in Patagonia. We then used empirically derived puma occurrence records from Patagonian ranches 1) to test our regional habitat predictions, and 2) to evaluate if paddock characteristics (vegetation cover, topography, and distance to roads) contribute to explain puma occurrences within ranches. Finally, we simulated three livestock management scenarios differing in their spatial and seasonal allocation of livestock to paddocks, and compared the likelihood of livestock exposure to pumas among scenarios. At a regional scale, 22% of the study region was predicted to be suitable for puma home ranges. The greatest uncertainty in these predictions resulted from assumptions on woody vegetation cover requirements at the home range scale. Within ranches, puma occurrences were positively associated with paddock topography, woody vegetation cover on paddocks, and proximity to predicted regional puma habitat. Comparing the risk of predation by puma among simulated livestock management scenarios implied that rotating livestock during seasons may help to reduce the likelihood of livestock exposure to pumas. Our results show the usefulness of rule-based habitat models for describing broad-scale carnivore distributions and for aiding risk assessments to mitigate conflicts between predators and human activities.     

Molecular Characterization of Motile Serovars of Salmonella enterica from Breeder and Commercial Broiler Poultry Farms in Bangladesh

Contaminated poultry and poultry products are a major source of motile Salmonellae for human salmonellosis worldwide. Local circulation of any motile Salmonella serovar in poultry has a wider public health impact beyond its source of origin for being dispersed elsewhere through poultry trades or human travels. To investigate the status of motile Salmonella serovars in breeder farms in Bangladesh, multiple flocks of two breeder farms were observed for a period of six months. In addition, a cross-sectional survey was carried out to determine the prevalence and serovar distribution of motile Salmonella by randomly selecting 100 commercial broiler poultry farms. Five pooled faecal samples representing an entire housed flock of breeders or broilers were screened for presence of motile Salmonella following conventional bacteriological procedures. The Salmonella isolates obtained were subsequently serotyped, and characterized by plasmid profiling and pulsed-field gel electrophoresis (PFGE). The results revealed that both the breeder farms were positive with three Salmonella serovars: S. Virchow, S. Paratyphi B var Java (S. Java) and S. Enteritidis. Eleven of the 100 broiler farms investigated were positive for motile Salmonella, giving a farm-level prevalence of 11% (95% confidence interval 5–17%). S. Virchow and S. Kentucky were the two predominant serovars isolated from the broiler farms. The PFGE genotyping demonstrated that the isolates belonging to the same serovars were closely related due to variation in only 1–4 bands. All the S. Virchow and S. Java isolates, irrespective of breeder or broiler farm origin, were plasmid-free, except for one S. Virchow isolate from a broiler farm that harboured a 9.7 kb-sized plasmid. The S. Kentucky isolates belonged to three plasmid profiles having plasmids of four different sizes, ranging from 2.7 to 109 kb. This is the first report of any motile Salmonella serovars from breeder and commercial broiler poultry farms in Bangladesh.     

Mixed crop–livestock farming systems: a sustainable way to produce beef? Commercial farms results,questions and perspectives

Mixed crop–livestock (MC–L) farming has gained broad consensus as an economically and environmentally sustainable farming system. Working on a Charolais-area suckler cattle farms network, we subdivided the 66 farms of a constant sample, for 2 years (2010 and 2011), into four groups: (i) ‘specialized conventional livestock farms’ (100% grassland-based farms (GF), n=7); (ii) ‘integrated conventional crop–livestock farms’ (specialized farms that only market animal products but that grow cereal crops on-farm for animal feed, n=31); (iii) ‘mixed conventional crop–livestock farms’ (farms that sell beef and cereal crops to market, n=21); and (iv) organic farms (n=7). We analyse the differences in structure and in drivers of technical, economic and environmental performances. The figures for all the farms over 2 years (2010 and 2011) were pooled into a single sample for each group. The farms that sell crops alongside beef miss out on potential economies of scale. These farms are bigger than specialized beef farms (with or without on-farm feed crops) and all types of farms show comparable economic performances. The big MC–L farms make heavier and consequently less efficient use of inputs. This use of less efficient inputs also weakens their environmental performances. This subpopulation of suckler cattle farms appears unable to translate a MC–L strategy into economies of scope. Organic farms most efficiently exploit the diversity of herd feed resources, thus positioning organic agriculture as a prototype MC–L system meeting the core principles of agroecology.     

Farms,pastures and woodlands: the fine‐scale distribution of Palearctic Culicoides spp. biting midges along an agro‐ecological gradient

The spatial epidemiology of Bluetongue virus (BTV) at the landscape level relates to the fine‐scale distribution and dispersal capacities of its vectors, midges belonging to the genus Culicoides Latreille (Diptera: Ceratopogonidae). Although many previous researches have carried out Culicoides sampling on farms, little is known of the fine‐scale distribution of Culicoides in the landscape immediately surrounding farms. The aim of this study was to gain a better understanding of Culicoides populations at increasing distances from typical dairy farms in north‐west Europe, through the use of eight Onderstepoort‐type black‐light traps positioned along linear transects departing from farms, going through pastures and entering woodlands. A total of 16 902 Culicoides were collected in autumn 2008 and spring 2009. The majority were females, of which more than 97% were recognized as potential vectors. In pastures, we found decreasing numbers of female Culicoides as a function of the distance to the farm. This pattern was modelled by leptokurtic models, with parameters depending on season and species. By contrast, the low number of male Culicoides caught were homogeneously distributed along the transects. When transects entered woodlands, we found a higher abundance of Culicoides than expected considering the distance of the sampling sites to the farm, although this varied according to species.     

Obesity and diabetes, the built environment, and the ‘local’ food economy in the United States, 2007

Obesity and diabetes are increasingly attributed to environmental factors, however, little attention has been paid to the influence of the ‘local’ food economy. This paper examines the association of measures relating to the built environment and ‘local’ agriculture with U.S. county-level prevalence of obesity and diabetes. Key indicators of the ‘local’ food economy include the density of farmers’ markets and the presence of farms with direct sales. This paper employs a robust regression estimator to account for non-normality of the data and to accommodate outliers. Overall, the built environment is associated with the prevalence of obesity and diabetes and a strong local’ food economy may play an important role in prevention. Results imply considerable scope for community-level interventions.     

Aflatoxin contamination caused by natural fungal infection of preharvest corn

Summary Aflatoxin contamination of developing corn (Zea mays L.) kernels caused by natural infection byAspergillus flavus Link ex Fries was studied in hybrids developed for the U.S. corn belt and for the southern U.S. and grown at diverse locations in 1977. Planting dates were staggered to examine the effect of crop maturity on infection by the toxin-producing fungus. A broad range of toxin values was observed at harvest; some levels exceeded the highest that had been previously recorded in corn. The highest concentration of aflatoxin B₁ detected was 8030 ppb. Levels of toxin differed significantly among planting dates in Florida and Georgia; the second planting date at these locations contained the highest toxin levels. Elevated concentrations of toxin were characteristic of kernel samples from southern locations and southeast Missouri; at these locations samples from hybrids developed for the south had significantly lower levels of toxin than hybrids developed for the corn belt. Ears with heavy insect damage had higher toxin levels than ears with less evidence of insect attack.Mention of firm names or trade products does not imply that they are endorsed or recommended by the U.S. Department of Agriculture over other firms or similar products not mentioned.     

Disease Prevention versus Data Privacy: Using Landcover Maps to Inform Spatial Epidemic Models

The availability of epidemiological data in the early stages of an outbreak of an infectious disease is vital for modelers to make accurate predictions regarding the likely spread of disease and preferred intervention strategies. However, in some countries, the necessary demographic data are only available at an aggregate scale. We investigated the ability of models of livestock infectious diseases to predict epidemic spread and obtain optimal control policies in the event of imperfect, aggregated data. Taking a geographic information approach, we used land cover data to predict UK farm locations and investigated the influence of using these synthetic location data sets upon epidemiological predictions in the event of an outbreak of foot-and-mouth disease. When broadly classified land cover data were used to create synthetic farm locations, model predictions deviated significantly from those simulated on true data. However, when more resolved subclass land use data were used, moderate to highly accurate predictions of epidemic size, duration and optimal vaccination and ring culling strategies were obtained. This suggests that a geographic information approach may be useful where individual farm-level data are not available, to allow predictive analyses to be carried out regarding the likely spread of disease. This method can also be used for contingency planning in collaboration with policy makers to determine preferred control strategies in the event of a future outbreak of infectious disease in livestock.     

Epidemiology of Rhodococcus equi Strains on Thoroughbred Horse Farms

Pulsed-field gel electrophoresis of restriction endonuclease-digested genomic DNA from a large collection of clinical isolates of Rhodococcus equi, an important pathogen of foals, was used to compare strain distribution between farms and over time. Forty-four strains were found among 209 isolates, with 5 of these accounting for over half the isolates and the 22 strains isolated more than once accounting for 90% of the isolates. The average genotypic diversity on each farm and in each year was found to be less than the genotypic diversity of the isolates taken as a whole, with 5.2% of the total diversity being due to differences between farms and 5.5% to differences between years. A small number of strains on each farm were found to have caused at least half the clinical cases of disease, and these varied between farms and, to a lesser extent, years. Most strains were found on more than one farm, and some very similar restriction patterns were found among isolates from different continents, indicating that strains can be very widespread. Multiple strains were isolated in five of the six cases in which more than one isolate from a single foal was examined, indicating that disease may commonly be caused by simultaneous infection with multiple strains. It was concluded that there are a number of different strains of R. equi which carry the vapA gene, and these strains tend to be widespread, but individual farms tend to have particular strains associated with them.     

Why are rare variants hard to impute? Coalescent models reveal theoretical limits in existing algorithms

Genotype imputation is an indispensable step in human genetic studies. Large reference panels with deeply sequenced genomes now allow interrogating variants with minor allele frequency < 1% without sequencing. Although it is critical to consider limits of this approach, imputation methods for rare variants have only done so empirically; the theoretical basis of their imputation accuracy has not been explored. To provide theoretical consideration of imputation accuracy under the current imputation framework, we develop a coalescent model of imputing rare variants, leveraging the joint genealogy of the sample to be imputed and reference individuals. We show that broadly used imputation algorithms include model misspecifications about this joint genealogy that limit the ability to correctly impute rare variants. We develop closed-form solutions for the probability distribution of this joint genealogy and quantify the inevitable error rate resulting from the model misspecification across a range of allele frequencies and reference sample sizes. We show that the probability of a falsely imputed minor allele decreases with reference sample size, but the proportion of falsely imputed minor alleles mostly depends on the allele count in the reference sample. We summarize the impact of this error on genotype imputation on association tests by calculating the r² between imputed and true genotype and show that even when modeling other sources of error, the impact of the model misspecification has a significant impact on the r² of rare variants. To evaluate these predictions in practice, we compare the imputation of the same dataset across imputation panels of different sizes. Although this empirical imputation accuracy is substantially lower than our theoretical prediction, modeling misspecification seems to further decrease imputation accuracy for variants with low allele counts in the reference. These results provide a framework for developing new imputation algorithms and for interpreting rare variant association analyses.     

Livestock damage and wolf presence

Depredation on livestock and wolf pack Canis lupus distribution was investigated in the province of Arezzo, Tuscany, from 1998 to 2001. Although livestock was uniformly distributed, damage level and distribution were focused in the mountainous area inhabited by wolf packs hence there were huge differences between areas with and without wolves. In the whole province a few farms were persistently affected by predation (6%), and they reported 38% of the total attacks and 37% of the total losses. Surplus killing phenomena involved only sheep and goat farms, affecting 18% of the total. Thirty-five attacks (14% of the total attacks) reported 536 kills (44% of the total livestock killed in the whole province of Arezzo). During the period 1998–2001, compensation costs in the province averaged 86 863 Euros (range: 68 805–99 318 Euros). In the same period, no farmer requested prevention funding from the Tuscan region. During the study period wolf population was stable: wolf packs were distributed on 47% of the whole province (1504 km²), with a density estimated at 2.9±0.7 wolves 100 km⁻².     

Association of arsenic exposure with lung cancer incidence rates in the United States

Background

Although strong exposure to arsenic has been shown to be carcinogenic, its contribution to lung cancer incidence in the United States is not well characterized. We sought to determine if the low-level exposures to arsenic seen in the U.S. are associated with lung cancer incidence after controlling for possible confounders, and to assess the interaction with smoking behavior.

Methodology

Measurements of arsenic stream sediment and soil concentration obtained from the USGS National Geochemical Survey were combined, respectively, with 2008 BRFSS estimates on smoking prevalence and 2000 U.S. Census county level income to determine the effects of these factors on lung cancer incidence, as estimated from respective state-wide cancer registries and the SEER database. Poisson regression was used to determine the association between each variable and age-adjusted county-level lung cancer incidence. ANOVA was used to assess interaction effects between covariates.

Principal Findings

Sediment levels of arsenic were significantly associated with an increase in incident cases of lung cancer (P<0.0001). These effects persisted after controlling for smoking and income (P<0.0001). Across the U.S., exposure to arsenic may contribute to up to 5,297 lung cancer cases per year. There was also a significant interaction between arsenic exposure levels and smoking prevalence (P<0.05).

Conclusions/Significance

Arsenic was significantly associated with lung cancer incidence rates in the U.S. after controlling for smoking and income, indicating that low-level exposure to arsenic is responsible for excess cancer cases in many parts of the U.S. Elevated county smoking prevalence strengthened the association between arsenic exposure and lung cancer incidence rate, an effect previously unseen on a population level.