Theme 5 Epidemiology.

INSERM UMR1094, Tropical Neuroepidemiology, University of Limoges, Institute of Neuroepidemiology and Tropical Neurology, CNRS FR 3503 GEIST, Limoges, France, CHU de Nouakchott, Nouakchott, Mauritania, Laboratoire de recherché en Neurosciences, CHU Mustapha, Algiers, Algeria, Department of Neurology Razi Hospital, La Manouba, Tunis, Tunisia, Division of Neurology, Centre for Research in Neurodegenerative Disease, Stellenbosch University, South Africa, Department of Neurology, CHNU Fann, UCAD, Dakar, Senegal, Service de neurologie CHU, Campus Université de Lomé, Université de Lomé, Togo, Laboratory of Chronic and Neurologic Diseases Epidemiology, University of Abomey-Calavi, Neurology Unit, CNHU Cotonou, Cotonou, Benin, Service de Neurologie, CHU Sourô Sanou, Bobo-Dioulasso, Burkina Faso, CHU Limoges, Service de Neurologie, Centre expert ALS, Limoges, France

Methods: A hospital-based multi-centre cohort study was undertaken in African countries. ALS patients diagnosed from 2005 to 2017 in the participant centres were included. An ALS expert neurologist confirmed the diagnosis through medical records. Patients were categorized at diagnosis according to El Escorial revised criteria and follow-up assessments were performed. Descriptive and analytical statistics were conducted. Subgroup analyses were performed according to subcontinent, based on the United Nations Statistics Division. Survival analyses were carried out with the Kaplan-Meier method.
Results: 10 centres from 89 African countries participated. An overall 185 patients were included: 114 in Northern Africa, 30 in Southern Africa and 41 in Western Africa. A male predominance was found with a sex ratio (male/female) of 2.9 overall. Median age at onset was 54 years (interquartile range (IQR) 464). Western African patients had younger age at onset (47 years) compared with Northern and Southern Africa, p¼0.0001. The median diagnostic delay was 12 months. Onset was bulbar in 22.7% of cases. Median ALS Functional Rating Scale Revised (ALSFRS-R) was 38 at diagnosis. 59 cases presented with atypical signs (32%), of which nearly 40% had cognitive impairment. Concerning the management, 80% had a medical or traditional treatment. However, only 47 cases had access to riluzole, mostly in Northern Africa. Overall, the median survival since diagnosis was 14 months. Longer survival was observed in Northern Africa with a median of 19 months compared with Southern Africa (11 months) and Western Africa (6 months), log rank p¼0.001.
Conclusions: This is the first ALS multi-centre cohort study in African countries. Our research contributes to improving the clinical and survival understanding of ALS in Africa. However, we are aware that hospital-based studies are limited by selection bias. Further studies are needed to clarify variability in clinical features. Genetic research should also be included for a more comprehensive approach on ALS in African populations.
Objectives: To determine the presence of spatial correlation of mortality cases of ALS among 29 Health Services areas (HAs) in Chile, over 24 years. To explore the significance of environmental variables that could explain this correlation.
Methods: A data base of 696 strata of fatal cases of ALS was built, including sex, age (4 strata), education (3 strata) and 29 geographical areas, over 24 years (19902013). Deaths were extracted from the national registry from the Ministry of Health. Contextual variables were estimated using a national representative socioeconomic characterization survey that has been carried out biennially since 1990. For each stratum and year the percentage of population exposed to: 1) living in a rural zone, 2) agricultural occupation, 3) mining occupation and 4) belonging to an ethnic group, was estimated. Denominators of each stratum were extracted from the inter-census estimation of the National Institute of Statistics. We estimated the ratio between observed and expected mortality rates of ALS for each geographic area, each year, and for the total period of study. Random effect and spatial dependence of neighboring areas were modeled separately through Poisson models using Markov Chain Monte Carlo procedures. Residuals of adjusted and non-adjusted models were explored. Magnitude of effect of variables are expressed as relative rate ratios (RR).
Results and discussion: Combining the information of all years, a higher risk of ALS mortality was observed in three central-north Health Services areas (range: RR 1.22 to 1.41), while a lower risk was concentrated in the very north (range: RR 0.53 to 0.57) and central-south Health Services areas (range: RR 0.55 to 0.77). Observing each year separately, the clustering was not evident. The strata that were exposed to a greater level of rurality presented a higher risk for death by ALS (RR 1.11(95% CI 1.031.21), by each additional 10% of rurality), independent of other co-variables. Initially, exposure to mining occupation showed lower risk (RR 0.74 (0.630.86) but this association was confounded mainly by the age of these strata (full adjusted RR 1.14 (0.971.34)). Although raw analysis showed that agricultural occupation was associated to a higher risk of ALS mortality (RR 1.34 (1.351.40)), the significance disappears after adjusting (RR 0.97 (0.911.04)). Ethnicity did not show significant association in the full adjusted model, while females presented a protective risk (RR 0.66 (0.600.74)). Lowest education stratum showed higher risk of death by ALS, independent of other factors.
Conclusion: In the long term, ALS mortality in Chile presents spatial correlation. At an ecological level, only rurality was associated to higher mortality by ALS independent of other variables. Background: In general, the number of persons with amyotrophic lateral sclerosis (ALS) was derived from the patient databases, and the data were normally collected through the National ALS Registry. This non-probabilistic sampling method requires the patient to register actively, while there were few studies on the method of estimating the patient population. In addition, to date no studies on the incidence of ALS in the Beijing municipality of China have been conducted. Beijing is a city with a population of nearly 21,000,000, where ALS epidemiological studies are crucial.
Objectives: Estimate the number of patients and determine the incidence and influencing factors of ALS in Beijing from 2010 to 2015.
Methods: The number of patients diagnosed with ALS between 2010 and 2015 were created from two aspects, namely, the patient visit record (including basic information and diagnostic results) of the Peking University No.3 Hospital (PUH3) and the census report (including basic information, hospital diagnosed at, treatment methods and current conditions) of Oriental Rain ALS care center (ORALS). PUH3 is the main ALS diagnostic hospital in Beijing, where patients diagnosed will choose a care center for adjuvant therapy, eg ORALS. By examining the consistency of age and gender distribution, it was demonstrated that the number of patients diagnosed at PUH3 in the visit record corresponded to the proportion of the total number diagnosed by PUH3 in the census report. Further, the number of ALS patients in Beijing's other hospitals was estimated. Finally, combined with demographic data, the incidence and influencing factors of ALS were analyzed.
Results: A total of 3381 patients diagnosed with ALS within the 6-year period were estimated. Average yearly incidence was 2.708/100,000, where the male incidence was 3.308/100,000, and female incidence was 2.069/ 100,000. Male:female ratio was 1.698:1. The age distribution obeyed the 3-parameter Weibull distribution. The average age of males was 53.89 years and females was 52.42 years. There were differences in the incidence of different age groups (Chi-square¼74.360; p50.05). The number of patients above the age of 20 years began to rise slowly, to reach a peak at 5560 years, and then began to fall.
Discussion: The incidence estimates of ALS in Beijing are consistent with findings from long established ALS registries in Europe and from some studies on incidence in Asia. Males are more susceptible to ALS than females. The ALS incidence is related to age. The incidence increased with age before the age of 60 years, and gradually decreased above 60 years. The patient age range is mainly 3065 years. Later, the risk of the disease will decline and the cause may be related to the pathogenesis of ALS.
Background: Establishing mechanisms of disease causation in neurodegenerative diseases had long seemed to be beyond the pale of traditional epidemiological tools. Over the past 30 years it has become increasingly more plausible that many cases of ALS may be triggered by cumulative changes, such as accrued somatic mutations (nucleic acid changes), occurring in a single cell.
Objective: To demonstrate how this conclusion regarding the likely mechanisms by which many cases of ALS are triggered was derived by the systematic application of classical epidemiological principles.
Design/Methods: The following epidemiological approaches were applied to understanding ALS initiation. First, John Snow's approach to seeking out patterns actively. Non-random patterns of disease occurrence are keys to understanding its cause. Evidence-based methods were applied to selecting the most credible associations. Next, AB Hill's 9 viewpoints to inferring causation from associations were utilized. Of these the most critical was that cause must precede (biological) disease onset. Confidence in conclusions about causality increases if they are supported by findings that are consistent, coherent, and biologically plausible, and if a biological gradient can be demonstrated. Clinical and epidemiological features and characteristics of sporadic, familial, and Western Pacific ALS were considered.
Results: Biological plausibility that ALS may be triggered by changes in one cell is supported by clinical, electrophysiological, and immunohistopathological evidence that most cases of the disease have focal onset and spread by contiguity within the motor super-network. Most cases of ALS are now understood to be corticallygenerated, part of the ALS-FTD spectrum. Greater network complexity and cellularity confers greater risk that the disease will start in the corresponding region. The age-dependent increase in incidence of sporadic ALS supports accrual of changes in a multi-step process, affecting cell components that do not turn over rapidly (DNA rather than protein). Smoking, the only established risk factor for the disease, has a known mechanism of causing disease by inducing DNA changes, as demonstrated in carcinogenesis. The best interpretation of age-dependent penetrance of familial ALS is that there is need for additional nucleic acid changes to occur so that potential disease will become active disease. However, lower mean age of onset in many cases of familial ALS is consistent with an explanation that fewer, time-and age-dependent, acquired steps are needed if the first has been inherited. Western Pacific ALS is now understood by Spencer to be a 'genotoxic' disease, rather than a 'neurotoxic' disease, to account for the long latency from exposure to a putative environmental cause and disease onset, and for the inexorable progression, once disease has started, even if the toxic exposure is long gone.
Conclusions: Application of classical epidemiological methods supports a likely mechanism of biological onset of disease in many cases of ALS: accrued somatic mutations. Background: Community concerns relative to unusual spatial and temporal patterns in amyotrophic lateral sclerosis (ALS) prevalence, along with a largely unknown disease etiology, led to the establishment of a statewide registry for ALS in Massachusetts in 2003. Following pilot studies and methods development, the Argeo Paul Cellucci ALS Registry of Massachusetts began tracking ALS cases across the state in 2007. Designated a reportable disease in Massachusetts, ALS cases are reported annually to the Massachusetts Department of Public Health (MDPH) by hospitals, ALS clinics, and neurologists in accordance with federal and state privacy protection regulations. This first-in-the-nation registry allows for a comprehensive capture, population-based surveillance system for monitoring the occurrence of ALS and provides a unique opportunity to explore possible environmental causes of the disease. Here, we present results from 2008 to 2012.
Objectives: To determine patterns of incidence and prevalence of ALS in Massachusetts and to provide a resource for ALS researchers.
Methods: Medical records are obtained and abstracted for all patients who have been diagnosed with or evaluated for ALS in Massachusetts. All eligible cases are reviewed by consulting ALS specialists to confirm diagnosis based on El Escorial criteria. The MDPH Registry of Vital Records and the National Death Index are used to confirm any deaths. Clinical and demographic information is recorded and used by the MDPH to investigate spatial and temporal patterns of ALS in Massachusetts.
Results: From 2008 to 2012, 833 new ALS cases were reported to MDPH. The age-adjusted incidence rate was 2.3 people per 100,000 per year. The incidence of ALS increased by age until approximately age 80 years with the highest incidence rate observed for those aged 7079 years. The male to female case ratio was 1.2. ALS prevalence and incidence will be presented by geography and stratified by other demographic variables including family history. Survival and disease progression statistics will also be presented.

Conclusion:
The Massachusetts ALS Registry is the only comprehensive statewide system of reporting medically diagnosed cases of ALS in the United States. Difficulty in the diagnosis of ALS presents challenges for estimating disease rates, but systematic data collection and physician verification protocols enable consistent reporting. The Massachusetts ALS Registry is the only state government mandated registry in the United States. It has established a baseline for surveillance of a disease that is not widely tracked, particularly in the United States. Data are available for epidemiologic studies and to ALS researchers to better understand the causes of ALS. Data are available to inform patient service needs and epidemiologic studies. Researchers can apply for access with the MDPH's IRB (Institutional Review Board). Methods: Eligible pilot study participants must have enrolled in the Registry and consented to be contacted about research projects. Pilot study participants could contribute to one or both components: 1) in-home collection of blood, urine, hair, and nails, on two occasions approximately six months apart or saliva (blood draw failure); and 2) post-mortem collection of brain, spinal cord, CSF, bone, muscle, and skin. Blood and urine specimens were processed into smaller aliquots. Brain and spinal cord were processed into fixed and frozen sections. The National ALS Biorepository continues collecting blood, urine, and/or saliva from ALS patients enrolled in the Registry in their homes on one occasion and postmortem collection tissue types were unchanged. The processing of specimens remains the same, however RNA and DNA are extracted in batches as the specimens arrive.
Results: 339 people consented to donate biological specimens during the pilot study from all 50 states and 330 provided specimens at least once. Approximately 18% of participants did not complete the second in-home donation due mostly to death and illness. Participants' ages ranged from 31 to 87 years and 60% were male. 54% of participants lived 50 or more miles from an ALS specialty/referral center. 30 people consented to postmortem tissue donation from 18 states. Post-mortem participants were 43 to 75 years of age at the time of consent and 50% were male. 21 post-mortem donations have been completed. The first 20 cases examined showed neuropathological-confirmed ALS with TDP-43 inclusions (ALS-TDP) with a mean age at death of 63 years (range 4376), 40% male, and mean brain weight of 1269 g. 65% of subjects had comorbid pathological diseases, including the neuropathological changes of Alzheimer's disease, primary age-related tauopathy, Lewy body disease, vascular disease with infarcts, encephalitis, and tumor. The DNA from 317 pilot study participants is being evaluated for mutations in known ALS genes using the NeuroChip array. 9.5% were positive for C9orf72. The National ALS Biorepository has consented approximately 100 ALS patients enrolled in the Registry to date to provide blood, urine, and/or saliva specimens and three have consented to post-mortem tissue collection.
Discussion and conclusions: Creating a geographically diverse biorepository has unique challenges. However, these specimens will be a valuable resource for researchers who can request specimens for ALS studies. We are conducting a web-based study (entitled 'ALS Quest') to look for risk factors for ALS/MND. The study consists of an online questionnaire, which is filled out by people both with and without ALS/MND. Community members and friends and relatives of ALS patients are encouraged to participate as controls. The questionnaire, which uses Qualtrics software, is available online at www.alsquest.org. The ALS Quest survey questions were initially based on experience from administering an Australia-wide paper-based ALS risk factor questionnaire which had over 2,000 respondents. New and modified questions were added from literature searches that implicated new risk factors and from validated ALS questionnaires generously supplied by investigators in the USA and Europe.
Survey questions include details of: occupational activities, chemical and pesticide exposures, physical activity, smoking, family sizes and medical histories, the ALS functional rating scale, index and ring finger lengths, personality types, and life event stressors. The questionnaire employs skip logic, which creates a custom path through the survey that varies based on a respondent's answers. The questionnaire takes about 90 minutes to complete, and can be conducted in multiple sessions. Computers, tablets, and smart phones can be used to take the survey. No personally-identifiable information is sought, so all responses are anonymous.
The questionnaire has been translated into 23 languages, so it is now available in at least one official language of the majority of countries around the world. More languages are planned to be added in the near future.
The ALS Quest questionnaire was launched in 2015 and will be active for at least another 10 years, so that enough responses can be collected for between-nation comparisons. To date, just under 1,000 responses have been collected. Respondents have been from 32 different countries. The most common sources of information about the questionnaire cited by respondents have been: ALS Associations, the internet, friends, ALS patients, health professionals, community groups, Facebook, the USA CDC National ALS Registry, the Canadian Neuromuscular Disease Registry, and ALS researchers.
Once sufficient numbers of responses have been received we plan to make the survey answers open source so that other investigations can mine the data. Objective: The objective of both studies is to examine the associations between OS (operationalized using questionnaire-based risk factor assessment and selected biomarkers of OS) and ALS disease progression (recording function, with the ALS functional rating scale-revised (ALSFRS-R) and forced vital capacity, and survival). To achieve this objective, we 1) obtained longitudinal data regarding environmental, lifestyle, dietary, and psychological risk factors, 2) measured clinical disease progression, and 3) obtained biospecimens to assess markers of OS.
Methods: Data for ARREST are obtained via telephone over a 24-month period. Patients are recruited from the National ALS Registry and included if their symptoms began within 24 months of screening. To date, 85 patients have been enrolled.
COSMOS was a multicenter collaborative study conducted at 16 study sites. COSMOS obtained data via inperson assessments and telephone interviews. We recruited 355 patients who had symptoms for less than 18 months at baseline.
Results: Compared to COSMOS, patients in ARREST are more likely to be white (97.4% vs. 87.9%) and have higher educational attainment (BA/BS or higher: 59.2% vs. 45.3%). Patients in both studies were similar in age, gender, and ethnicity. ARREST patients were more likely to use Medicare as their primary insurance (44.9% vs. 31.4%).
At baseline, ARREST patients have lower ALSFRS-R scores compared to COSMOS patients (34.2 AE 7.0 vs. 36.0 AE 6.7). They report longer duration of symptoms at baseline (13.6 AE 6.2 vs. 11.7 AE 4.5), but this is perhaps due to study design. A greater percentage of patients with progressive muscular atrophy enrolled in ARREST (13.3% vs. 7.9%) (ALS is most common, but other motor neuron diseases were included in both studies). Respiratory function, body mass index, region of onset, and percentage of patients that tested positive for the C9orf72 gene did not differ between the studies.
At baseline, ARREST patients scored better on word generation and exhibited less emotional lability than COSMOS patients; other cognitive and behavioral measures did not differ between the studies.
Discussion: ARREST was intended to add to the rich data collected from COSMOS. While there are some differences in demographic and baseline ALS functional parameters, these can be controlled statistically in the analytical phase. To our knowledge, this telephone-based study design is the first of its kind in ALS research and will be useful in future large-scale, population-based epidemiological studies. Background: Amyotrophic lateral sclerosis (ALS) is estimated to affect 5000 people in the UK at any one time, but the true figure is not known as there is no single record of who is affected. There are five ALS registers in England, Wales and Northern Ireland: the South East ALS Register (SEALS), Northern Ireland ALS Register, Sheffield Teaching Hospital Register, Peninsula Network ALS Register and South Wales ALS Register. The postcode catchment areas these population registers recruit from cover approximately 32% of the population of England, Wales and Northern Ireland.
Aim: To identify all individuals with ALS in England, Wales and Northern Ireland, to estimate incidence and prevalence, and to act as a resource for matching services to need.
Methods: The catchment areas of MND Care and Research Centres were mapped, along with catchments for MND Association Regional Care Development Advisor staff and community teams. A central database was designed to allow transfer of data from multiple sites in different formats in an anonymized, legally and ethically compliant way. A web-based tool for direct capture of patient information was developed. All captured information is compliant with similar European registers and the Scottish register.
Results: MND Care and Research Centres with established population registers were recruited. Automated data transfer protocols were successfully generated, as well as manual transfer of flat files from existing clinic databases and our donated template database. The 5 initial registers and additional sites at Oxford and Brighton have approval to submit data to the National Register. The register contains a mixture of prospectively collected records with complete or nearly complete data to date and approximately 2500 records incorporated from the South East ALS Register.
Background: In Sweden several hospitals care for amyotrophic lateral sclerosis (ALS) patients through specialized ALS teams, although some of the clinics only provide care for a handful of patients. As a result, the diagnostic criteria and treatment guidelines at these centers may not always be based on the most recently updated guidelines. Additionally, the research activity among the ALS teams might not be prominent and the patients from smaller hospitals are therefore not as often included in research projects or randomized clinical trials.
Objective: Our primary aim was to construct a national motor neuron disease (MND) Quality Registry in Sweden, which will assure high quality health care for all MND patients, including ALS. Our secondary aim is to create a research base by prospectively following the entire ALS population in Sweden.
Methods: The variables included in the MND Registry were jointly decided by a steering group. The layout of the registry, with a screen displaying the current and the past variables, gives the clinicians and researchers an instant overview of the temporal trend of important clinical variables. An internet based patient own reporting portal (PER) was also implemented to actively involve patients in the registration of symptoms and health status.
Results: To date around 80% of all prevalent ALS patients in Sweden (n¼666), including all ALS patients of the Stockholm area (n¼194), are registered in the MND Registry. ALS was the most common diagnosis registered (n¼153, 78.9%), followed by primary lateral sclerosis (PLS) (n¼13, 6.7%), and progressive spinal muscular atrophy (PSMA) (n¼8, 4.1%). A slightly higher proportion of these patients were females (n¼100, 52%), and females and males had a similar age at ALS diagnosis (61 years). Females seemed to be overrepresented in both the youngest and oldest age groups, ie 40 years, 4150 years, and !71 years at diagnosis, although these gender differences were not statistically significant (all p values 40.05).
The vast majority of patients (n¼78 males, 83.0%; n¼87 females, 87.0%) had been treated with riluzole. Around half of the patients had been treated with enteral nutrition (n¼98, 50.5%), and around 30% of the patients had been treated with non-invasive ventilation (n¼ 58, 29.9%). A minority of the patients had been treated with invasive ventilation or tracheostomy (n¼16, 8.2%).

Discussion and conclusion:
The two main strengths of the Swedish MND Registry are the easy accessibility, via the internet and the possibility for the patients to participate through the self-reporting tool. These set-ups are not evident in the other international MND quality registries and will ensure an important contribution of the registry in both patient care and research. Email address for correspondence: rooneyj4@tcd.ie Keywords: incidence, structured data, survival Introduction: Total correlation explanation (CorEx) is a new approach to discovering structure in data. The method is based on information as entropy theory and has been shown to outperform a number of alternative clustering methods (1).
Methods: All cases on the Irish ALS register from 2006 to 2014 were included. For each case, variables included were: age at diagnosis, diagnostic delay, site of onset, gender, revised El Escorial criteria, survival in months and C9orf72 status. The CorEx algorithm was applied to these data allowing for one to six latent variables with one to six different levels possible per variable. For each setting, the model was run 40 times. The overall best fit model was that which produced the highest overall total correlation.
Results: 944 cases met the inclusion criteria. The model that best explained correlation in the data from the Irish ALS register indicated four latent variables with a maximum of 4 dimensions. The four latent variables clustered as: 1) Diagnostic delay and survival, 2) Sex and site of onset, 3) Age of diagnosis and C9orf72, 4) El Escorial criteria.
Discussion: The CorEx algorithm detected structure in the data that confirms the known association of variables in the Irish ALS datasets. Further work is ongoing to explore the dimensionality of the latent variables. EPI-12 Epidemiological survey of SBMA in Italian north-east regions 28th International Symposium on ALS/MND onset, X-linked neuromuscular disease. It is caused by a CAG repeat expansion localized in the first exon of the androgen receptor (AR) gene. The disease manifests clearly only in male patients, while female carriers of the mutation are usually asymptomatic, and it is characterized by slowly progressive, mostly proximal, limbs and bulbar muscle weakness and atrophy, which is associated with wide systemic involvement. The diagnosis is made through a genetic test for AR-CAG expansion. Different SBMA founder haplotypes have been described in patients from around the world and no expansion-prone haplotypes were found. SBMA is considered a very rare disease and its real epidemiology is not known. This is mainly due to the presence of minimally symptomatic patients who may not search for medical advice.
Objective: Aims of this study were to perform a survey of SBMA epidemiology in three north-east Italian regions and to clarify the AR-haplotype diversity in our cohort.
Methods: 95 patients came from the three north-east Italian regions (Veneto, Emilia-Romagna, and Friuli-Venezia Giulia) and were referred to the Neuromuscular Centre of the University of Padova between January 2006 and May 2017. For each family, pedigree structure was traced back to at least 3 generations. In search of relationships between apparently unrelated patients, we genotyped 46 SBMA families and reconstructed the haplotypes associated with the CAG repeat, defined by eight SNPs and six flanking STRs.
Results: 46 families for a total of 95 genetically confirmed patients were considered. Prevalence was 0.89/100,000 for the total population (95/10,584,487) and 1.84/100,000 for the male population (95/514,325). Mean incidence per year for the period 2006 through 2017 was 0.09/100,000 for the total population and 0.1/100,000 for the male population. Mean age of onset, defined as first appearance of muscular weakness, was 48.08 years +/10.48. The yearly mortality rate was 0.02/100,000 for the general population and 0.04/100,000 for the male population, with average age at death 75.8 +/5.98 years. Mean disease duration was 20.83 years +/ 11.72. We found a strong founder effect with 21% of the patients coming from the Verona district and 12% coming from the Reggio Emilia district, confirmed by molecular haplotype analysis.
Conclusions: This is the first study describing the epidemiology of SBMA considering such a wide population of patients. Our data confirm previous epidemiological studies conducted on small populations of patients. Nevertheless, estimation of the real distribution of the disease remains difficult since asymptomatic patients are frequent and not always undergoing genetic testing, so the disease is probably underdiagnosed and further studies are warranted to obtain complete epidemiology data. Objectives: To describe a clustering of ALS cases in a small hamlet in the French Alps and discuss potential exposures in the patients' environment.
Cases: Between 1991 and 2013, five ALS cases were described in the small hamlet of Montchavin, Savoie, France. Two of them were spouses. They have all lived in the hamlet since early childhood or birth. Patients were all neighbours and the hamlet comprised 200 persons living here on a regular basis.
Between 2010 and 2015, seven other cases of ALS were identified in individuals that did not originate from the village but spent a part of their life in the area. One had lived here since 2009, another spent half of his lifetime in Montchavin, two were present during holidays for more than 20 years, and two others lived in close villages and had an activity, professional or leisure, in the hamlet.
Results: The study of the cases did not identify any genetic/familial relationships between the patients and, when possible, a genetic analysis was performed and was negative.
For the 5 ALS patients originating from the area, given the interval of observation, 22 years, the population of 200 and the incidence of ALS, 2.5/100,000, the standardized incidence ratio¼45 (p50.00001). It does not seem possible to estimate an incidence ratio for the whole group of 12 patients in the cluster area, as the population during winter and summertime greatly increases, and for seven of them, exposure was not continuous.
Exposures: Common exposures were found between the 12 patients: all but one practiced sport intensively, eight had a vegetable garden, six used to eat local mushrooms, five were ski teachers, and four had a restaurant. The water used for the vegetable gardens was collected immediately downstream from the ski area where snow guns were intensively using snomax for artificial snow, with a mix containing bacteria. Among the five ALS cases originating from the area, four of them were chronically exposed to polycyclic aromatic hydrocarbons, and in two houses an abnormally high level of radon was measured (4400 Beq/m 3 , with a maximum at 1116).
Discussion: We present a highly significant cluster of ALS in a small hamlet in the French Alps. Patients are neighbours and share many exposures, some of them being known for their toxicity regarding neuronal/cellular metabolism. Others factors are more exploratory, such as vegetable gardens where heavy metals may be implicated. The investigation of BMAA is underway in the area, as well as in the patients' brains that have been collected. four patients are still alive and these brains may be collected in the future. Reoccurring seasonal cyanobacterial harmful algal blooms (CHABs) persist in many waters, and recent work has shown links between CHAB and elevated risk of amyotrophic lateral sclerosis (ALS). Quantifying the exposure levels of CHAB as a potential risk factor for ALS is complicated by human mobility, potential pathways, and data availability. In this work, we develop phycocyanin concentration (ie CHAB exposure) maps using satellite remote sensing, across northern New England, to assess relationships with ALS cases, using a spatial epidemiological approach. Strategic semi-analytical regression models integrated Landsat and in situ observations to map phycocyanin concentration (PC) for all lakes greater than 8 ha (n¼4117) across the region. Then, systematic versions of a Bayesian Poisson log-linear model were fitted to assess the mapped PC as a risk factor for ALS while accounting for model uncertainty and modifiable area unit problems. The satellite remote sensing of PC had strong overall ability to map conditions (adj. R2, 0.86; RMSE, 11.92) and spatial variability across the region. PC tended to be positively associated with ALS risk with the level of significance depending on fixed model components. Meta-analysis shows that when average PC exposure is 100 mg/L, an all model average odds ratio is 1.48, meaning there is about a 48% increase in average ALS risk. This research generated the first regionally comprehensive map of PC for thousands of lakes and integrated robust spatial uncertainty. The outcomes support the hypothesis that cyanotoxins increase the risk of ALS, which helps our understanding of the etiology of ALS. EPI-15 Geospatial association between amyotrophic lateral sclerosis and water quality in northern New England X Shi 1 , N Torbick 2 , A Codamon 1 , P Henegan 3 , B Guetti 4 , A Andrew 3 , E Stommel 3 , W Bradley 5 Background: Less than 10% of ALS cases are familial with a known gene mutation. Several environmental factors have been linked to ALS, and recently exposure to cyanotoxins derived from cyanobacteria have been added to this list (15).
Objective: By estimating, through a geospatial analytical approach, exposure of ALS patients to cyanotoxins sourced from waterbodies, we intend to provide new evidence to the discussion of the association between cyanotoxins and ALS.
Case study: This study uses a database of the residences of ALS patients and controls in New Hampshire and Vermont to compare residential exposures, of ALS cases and controls, to cyanobacteria in waterbodies.

Methods:
We used remote sensing data and ground samples to map phycocyanin (PCN) and chlorophyl-a (chl-a) in waterbodies of 48 hectares. The spread of cyanotoxins from waterbodies to the residences of subjects was modeled using the kernel density estimation (KDE) with 1 km, 5 km, and 10 km as the bandwidth. Background: Electrical occupations have been linked to an increased risk of amyotrophic lateral sclerosis (ALS) in multiple studies. The underlying mechanism is however not clear. We explored the association between occupational exposure to electric shocks and extremely lowfrequency magnetic fields (ELF-MF) and the risk of ALS in 3 population-based case-control studies.

Results
Methods: ALS patients and population-based controls were recruited in the Netherlands, Ireland and Italy, between 2010 and 2015, as part of the Euro-MOTOR project. Lifetime occupational and lifestyle histories were obtained using structured questionnaires. We applied 2 previously developed job exposure matrices assigning low (i. background), medium or high risk of electric injuries due to shocks and exposure to ELF-MF. Odds ratios (OR) and 95% confidence intervals (CI) for ALS were estimated by logistic regression for both occupational exposures in the same model, adjusted for age, sex, cohort, smoking status and alcohol drinking.
Results: Complete occupational histories and information on confounding variables was available for 1423 confirmed ALS cases and 2851 controls. Among cases, 20% had ever been high exposed to electric shocks, and 7% had ever been high exposed to ELF-MF. Among controls, the proportions were 16% and 5%, respectively. Exposure to electric shocks and ELF-MF were only weakly correlated (Pearson correlation R¼0.37). Both ever exposure to electric shocks (OR¼1.23, 95% CI 1.051.44) and ELF-MF (OR¼1.15, 95% CI 0.991.32) were associated with ALS. No difference between the level of exposure or a trend with duration of exposure was observed for either exposure.
Conclusion: Our findings support a possible independent association between occupational exposure to electric shocks and ELF-MF and the risk of ALS. Background: ALS has been found to occur in conjugal couples and geospatial clusters. Both of these characteristics are suggestive of environmental contributors (1). Identified clusters of cases were determined to exceed expected numbers and were both statistically and clinically significant. They concluded that the importance of scale should not be underestimated and recommended that environmental sources be explored in further studies. This investigation began to explore the recommendation, specifically the possible contribution of living near certain types of transportation facilities.
Objectives: To determine the spatial relationship, if any, that highways have on ALS patients compared to clinic controls. Using GIS analysis of case and control locations with road network data, proximity analyses are being conducted to calculate the mean distances between these locations and the road network for demographic groups. The mean distances for cases are then being compared to the mean distances for controls to see if the cases are different to the controls. If the cases are closer to the roads than the controls, the factors associated with the differences are being explored.
Method: Using case-control questionnaire data, the addresses of ALS patients and controls were geocoded. Once tested for locational accuracy, the distance to the closest highway was measured. Using TIGER street centerlines (www.arcgis.com), highways were extracted using MTFCC codes S1100 (primary) and S1200 (secondary). The analyses were done using ArcGIS on the data, that is in a NAD 1983 datum/spheroid, and projected to a UTM Zone 18 North coordinate system. There are currently 103 ALS cases and 138 control records. An additional source of control data has been created from the US Census demographic data. Population locations for two million people have been created using Census Block centroids. Control locations are then extracted from the total population using the 46 age and sex categories used by the US Census. Available census controls were assigned to known residential addresses. Mean distances and their standard deviations were generated from the individual distances for groupings of cases and controls using the US Census age and sex categories. All datasets were processed to extract the approximate DHMC ( shown that people who smoke at onset of ALS have shorter survival than former or never smokers, and earlier age of disease onset. A population based study on the risk of smoking and ALS, also reported that lower alcohol consumption is a risk factor of ALS.
We aimed to discover whether smoking was associated with a risk of ALS in a case controlled study of people from the UK.
Methods: This was a retrospective case controlled study including people with a diagnosis of MND confirmed by a consultant neurologist. The data were collected from three centres in the UK. People diagnosed with ALS were asked about their current smoking habits and smoking history as part of a telephone interview with a trained healthcare professional. We performed a multivariate logistic regression analysis to compute odds ratios for the risk of never, former and current smoking and risk of ALS. The odds ratios were adjusted for gender, age at survey, educational attainment, cigarette pack years and alcohol consumption.

Results:
The analysis was performed on complete data from 201 cases and 199 controls. The gender ratio was 56:44 male-female in the control group and 42:58 in the in the patient group. The average age was 64.5 years for controls and 63.1 years for cases. We found there was an increased risk of ALS in current smokers (adjusted OR 3.46, 95% CI 1.2410.56, p-value 0.0003), but not for former smokers. We found there was a significant association of ALS with being a former drinker (adjusted OR 4.60, 95% CI 1.9910.94, p-value 0.0004), but current or never drinking had no effect. Most survey responders who had given up drinking (60%) gave symptoms associated with the disease as the reason.
Discussion: In our study population smoking carries an increased risk of ALS. The effect of alcohol consumption on disease risk has not been confirmed by our analysis. Objectives: The ratio of the length of the index finger (2D) to the ring finger (4D) (2D:4D) has been reported to be lower (ie 2D54D) in people with ALS than non-ALS controls. This has led to suggestions that exposure to increased prenatal testosterone, which also lowers this ratio, could be a risk factor for ALS. In an attempt to test this hypothesis, we examined 2D:4Ds from large numbers of ALS patients and controls.
Setting: An online multilingual questionnaire enabled respondents to measure their own index and ring finger lengths.
Participants: Of the initial 949 respondents, 572 remained for analysis after elimination for inability to straighten fingers, not answering the question, statistical outliers, and age540 years. Respondents remaining for analysis were 202 ALS patients (125 males, 77 females) and 370 non-ALS controls (112 males, 258 females).
Results: Unpaired t-tests with 95% confidence intervals were used to assess differences in mean 2D:4Ds. Males had significantly lower mean 2D:4Ds than females, in both ALS and control groups, for both left and right hands. However, no significant differences were found in 2D:4Ds between ALS and control groups, in either males or females, for either left or right hands. Receiver operating characteristic (ROC) curves showed no power for 2D:4Ds to predict ALS status in either males or females.
Conclusions: 2D:4Ds did not differ between ALS patients and controls in this study. This was despite the dataset being large enough to confirm the established finding of lower 2D:4Ds in males compared to females. These findings do not therefore support the hypothesis that exposure to increased prenatal testosterone is a risk factor for ALS. A putative lower 2D:4D has been proposed to explain the link between ALS and exercise, but our results indicate that other exercise-related factors need to be found to explain this association. Conclusions: Based on real-world database, we provided up-to-date prevalence, treatment status, and medical costs for ALS in Japan. Our results demonstrated efficacy of real-world database analysis to obtain the latest national trend of rare diseases including ALS, which may provide important information for clinicians and policymakers. Objective: Clinicians who see people with ALS often mention that they present as being particularly 'nice', with agreeable and outgoing personalities. These pleasant traits stand out particularly when considering ALS patients face a devastating disease. In an attempt to see if any evidence could be found to support these anecdotal reports of niceness, we conducted a case-control study looking for personality differences between ALS patients and non-ALS controls.
Methods: The web-based international ALS risk factor questionnaire 'ALS Quest' includes the Big Five Inventory of personality traits, ie agreeableness, conscientiousness, extraversion, neuroticism, and openness. ALS patients were asked to respond to the Big Five inventory in a manner reflective of their pre-diagnostic state. The Big Five factors were compared between ALS and non-ALS groups using logistic regression with odds ratios and 95% confidence intervals, controlled for age and gender.
Results: 929 respondents (355 ALS and 574 controls) met the criteria for inclusion. Odds ratios were significantly greater than one in ALS patients for agreeableness, conscientiousness, and extraversion, indicating these traits were more likely to be found in patients than controls.
Conclusion: People with ALS tend to be more agreeable, more conscientious, and more extraverted than control subjects. These factors are likely to be found more often in people with an outgoing, risk-taking personality type who may be prone to some risk factors associated with ALS, such as athleticism, head injury, and smoking. Another possibility is that genetic polymorphisms associated with these personality factors co-localise with polymorphisms that increase the likelihood of getting ALS. With the greater numbers of respondents expected for the ALS Quest survey in the future, international comparisons of personality types in ALS and controls will be able to be made. Introduction: Evidence shows that ALS progression is curvilinear and heterogenous. However, the routinely used progression rate (PR¼(48-ALSFRS-R/disease duration)) only reflects progression at a circumscribed time-point rather than across the disease. It is crucial to develop indices that can describe individual progression over the whole disease course.
Objectives: The study aimed to develop a model that uses regularly collected ALSFRS-R scores to describe and relate to individual events within the ALS disease course and reduce noise associated with the ALSFRS-R.
Methods: We used a sigmoidal decay function to describe the transition from full health to maximum disease for patients in the PRO-ACT database (1). The model yields three parameters: D50¼time taken for ALSFRS-R to drop to 24, dx¼slope of ALSFRS-R decrease, relative D50 (rD50)¼calculated value describing individual disease covered in reference to D50, where 0 would signify onset and 0.5 would indicate the time-point of halved functionality. D50 was used to mathematically derive disease phases: Phase I (early semi-stable phase), Phase II (early progressive phase), Phase III (late progressive phase) and Phase IV (late semi-stable phase).
Results: Using ALSFRS-R scores and disease duration from onset to ALSFRS-R date, we determined D50, dx, and rD50 for 4838 patients. The relationship between D50 and dx was highly linear (r¼0.889), indicating that the entire disease course can be described using D50 alone. A significant correlation between D50 and survival was observed (r¼0.654). Although significant correlations were observed between D50 and transformed (1/PR) first (r¼0.601) and last (r¼0.772) recorded PRs, the difference in the correlations indicates how profoundly PR can vary over time. When selecting for patients for whom model data were best fitted to actual data, we showed that riluzole treatment significantly prolonged mean D50 (43.6 vs. 40.7 months). We were also able to use rD50 to demonstrate progressive worsening of both forced and slow vital capacity for most patients. Statistical significance was set at p50.01.
Conclusion: D50, rD50, and dx enable comparisons between vastly different time-scales of disease courses and also enable staging of individual events, ie sampling at any given time-point can be correlated to any of the three parameters. This may aid the discovery of early prognostic