The main objectives of this study were to assess the factors associated with LTBI among asylum seekers entering Vaud County. A robust score combining six factors (country of origin, travel conditions, age, marital status, cough, prior TB exposure) allowed the identification of AS with the highest risk of LTBI.
The prevalence of LTBI of 24.9% observed in this study among asylum seekers is close to the estimates reported in the literature. Winje and et al. reported a 29%  prevalence while Hardy AB and al, reported 38% . Pareek  demonstrated that the proportion of asylum seekers in UK with a positive IGRA was between 3 and 28%, related to the incidence of tuberculosis in the home country and Mulder  reported a similar result with 20% positive QuantiFERON(®)-TB Gold In-Tube assay among AS.
A striking figure is the detection of 5 subjects with active tuberculosis that passed undetected through the border screening performed several weeks before. We assume that these subjects progressed from a recently acquired infection after border screening according to the natural history of TB. Considering the fact that the majority of cases of tuberculosis among asylum seekers are notified after entry, this is not surprising but underlines the fact that migrants with complaints or health problems should have rapid access to health care and tuberculosis diagnosis. A recent study by Ricks  highlighted the importance of LTBI screening and treatment in order to reduce the burden of TB among foreign born individuals in the US.
The multivariate logistic regression permitted to identify the major factors associated with LTBI. Married individuals from an African or an FSU country that crossed multiple borders to reach Switzerland border and who cough are mostly at risk of being infected. Two minor factors (age and positive history of TB exposure) were also highlighted. Using those six factors we elaborated a predictive model for screening asylum seekers for LTBI resulted in a score with an AUC=81%.
The risk of LTBI increases with age. Indeed the longer a person lives the greater are the risk of being in contact with an individual with active TB. The main feature of using age as LTBI predictor is the presumed time of infection. Due to their young age and to the travel conditions, frequently in very close contact with other persons during prolonged periods, we assume that these individuals have been infected recently.
In our study married individuals also had a higher risk of being infected with an odd ration of 2.0. This is a quite interesting finding since demographic data on disease shows an opposite relation . Although no solid explanation can be given for this finding our collective showed a clear association with LTBI.
The prevalence of tuberculosis in the home country is correlated with a risk of having LTBI. The NICE guidelines suggest LTBI screening for all asylum seekers migrating from countries with a TB prevalence higher than 50/100.000 . Applying this rule to Switzerland would mean screening the majority of asylum seekers and would imply high costs and logistical problems. Due to the limited population of this study the independent evaluation of each country of origin was not possible. Therefore we studied those countries mainly represented in the Swiss asylum seeker population. Bias due to the addition of populations like North Africans (low risk) and sub Saharan Africans (high risk) to the statistical analysis could not be avoided. Due to the absence of asylum seeker from Latin American origin in our collective, we could not assess the risk in this population.
The travel conditions to reach Switzerland were clearly related to the risk of LTBI infection. Individuals travelling directly to destination using airplane meet fewer migrants in their journey and therefore have a lower risk of TB infection. A long and hazardous journey through several borders using ground and/or sea transportation increases the risk of TB contacts and infection. Although the socioeconomic status of the asylum seeker might influence the travel pattern it is difficult to argue that ground/sea transit is less expensive than airplane but it is seems clear that access to airplane is limited to persons with higher socio-economic status and access to official (or fake) documents.
A previous exposure to TB is an obvious factor related with LTBI. A personal history of recent exposure to presumed or confirmed active TB person enhances the risk of LTBI and its reactivation potential mostly during the following two years.
That cough was identified as a risk factor for LTBI is surprising since by definition LTBI is an asymptomatic infection. This could be due to the fact that smokers (who are very prevalent in this population group) have a higher risk of LTBI and tuberculosis than non smokers . Other plausible explanation for this finding could be the congregated way of living, especially during winter months in asylum seeker centers with high exposure to passive smoking as well as the lack of stratification during statistical analysis between chronic and acute coughing due to sample limitations. When present, chronic cough was extensively assessed to rule out disease while acute coughing was usually self-limited.
The limitations of our study are the local setting, the inherent characteristics of this mobile population, the cross sectional design and the voluntary pattern of enrolment. This study provided a realistic description of actual collective of asylum seekers arriving in Vaud county. As the asylum seeker population is randomly allocated in the different regions of Switzerland, we assume that this population group was representative of the demographic details of the whole asylum seeker population in Switzerland. Nevertheless in this study 61% of the recently arrived asylum seeker population was screened using IGRA qualifying this study as representative of the study population. In addition to this, since no selection was applied to the study population, the travel condition and characteristics of the asylum seeker entering Vaud County match those of asylum seekers entering in other western European countries. Some bias could result from the voluntary pattern of enrollment with an over-representation of sick migrants but this setting was essential for the ethical acceptance of the study protocol. As the proportion of migrants with positive IGRAs was similar as in comparable studies, we assume that this was not a bias. Moreover, among AS who entered in the centre, the actual number of eligible persons was lower since many of them left the territory or were rejected before the enrollment procedure could be started. We have decided to include all the asylum seekers that were present on the asylum seekers registry to the study collective to better describe the reality. Finally since this study is time and country specific the collective characteristics are subject to change over time following the shift in immigration pattern.