Tuberculosis screening yield of asylum seekers in Europe

From the authors:

We thank W. Arranzola de Oñate and colleagues for their correspondence on our article “Low yield of screening asylum seekers from countries with a tuberculosis incidence of <50 per 100 000 population” [1]. They report a 4–5 times higher tuberculosis (TB) prevalence among Syrian asylum seekers at entry screening in Belgium (94.2 or 107.6 per 100 000 screened persons, depending on definition used) than we found in the Netherlands. We would like to elaborate on possible explanations for the differences in screening yield, some of which were also mentioned by Arranzola de Oñate and colleagues.

The Dutch data include almost all Syrian asylum seekers, since screening was mandatory in the Netherlands in the study period and done within the first few days after asylum application. In contrast to Belgium, children <5 years of age and pregnant women are also screened in the Netherlands. Children <5 years of age accounted for 9.7% of the Syrian asylum seeker population in our study and none of them had TB at entry screening. Information on the number of pregnant women was not available, but the number is assumed to be small. It is unlikely that these differences in individuals targeted for screening explain the different TB prevalence rates in the two countries.

If we compare radiographic screening results of Syrian asylum seekers in Belgium (TB prevalence 94.2 per 100 000 screened persons; 95% CI 41.2–186.3) and the Netherlands (prevalence 22.2 per 100 000 screened persons; 95% CI 9.7–44.0) then these differences are significant (p=0.01). The Belgium study is based on a smaller number of screened persons (7432 versus 31 470 in our study) and included screening data from 2015 only, while our study covered almost 5 years (January 2011-September 2015). We agree that TB prevalence may gradually change over time in populations affected by war and disaster. Table 1 shows that in the Netherlands most Syrian refugees (94%) were screened in 2014 and 2015 and all Syrian TB cases were identified in these 2 years. It would be interesting to know the Belgian screening results for Syrian refugees in 2014 to see if TB prevalence has changed over time.

In the Netherlands, maximum efforts are made to confirm the TB diagnosis. In our study, 10 (83%) out of 12 asylum seekers with TB were confirmed by a positive culture for Mycobacterium tuberculosis. A recent overview of screening in the main reception centre in the Netherlands reported culture confirmation in 48 (87%) out of 55 identified cases at entry screening [2]. It would be interesting to learn how many of the (Syrian) asylum seekers detected upon entry screening in Belgium had culture-confirmed TB. Empirical treatment of abnormalities possibly compatible with pulmonary TB could explain a higher prevalence of screening in Belgium.

We cannot substantiate whether Syrian asylum seekers arriving in Belgium are a different subgroup using different refugee routes, but this assumption seems unlikely to explain the difference in prevalence of screening. A more obvious explanation could be when clustered cases occur and are included in the study, e.g. a family with several TB cases. In the Netherlands, all M. tuberculosis isolates are subject to DNA fingerprinting and none of the culture-confirmed Syrian TB cases had identical variable number of tandem repeat genotypes or were epidemiologically linked.

We closely monitor the TB situation among asylum seekers, in particular among Syrians after screening was discontinued. We estimated that ∼2000 of the 5259 Syrians applying for asylum in September 2015 were not screened (screening was stopped on September 21, 2015, but was already incomplete before this date when refugees were housed in emergency shelters). In the fourth quarter of 2015 and in the first and second quarters of 2016, respectively, 11 060, 2230 and 1058 Syrians requested asylum in the Netherlands [3], giving a total of 16 348 unscreened Syrian asylum seekers (from mid-September 2015 up until June 30, 2016). Since mid-September 2015, 12 Syrian TB patients were reported to the Netherlands Tuberculosis Register (up until June 30, 2016). Two patients were included in the previous analysis since they had an abnormal chest radiograph at entry screening (one arrived in August 2015 and one in September 2015) and were confirmed by a positive M. tuberculosis culture in the fourth quarter of 2015. Six patients arrived in the Netherlands before discontinuation of screening and were diagnosed between 6–14 months after arrival. Four patients arrived in the Netherlands after discontinuation of screening: three of them were diagnosed with pulmonary TB 0–5 months after arrival, and one with pulmonary TB more than 6 months after arrival. If we assume that the three pulmonary TB patients identified after less than 6 months in the Netherlands potentially could have been identified by radiographic screening this would have resulted in a TB prevalence of 18.4 per 100 000 screened persons (95% CI 4.8–50.0), which is similar to the TB prevalence in our study.

In conclusion, we would like to state that our analysis of entry screening was robust and that close monitoring of the TB situation among Syrian asylum seekers after discontinuation of screening also indicates that their risk for TB is relatively low. The Belgian and Dutch studies on screening (Syrian) asylum seekers show the importance of collecting migrant screening data in a more uniform and systematic way, and reporting on countries' TB screening data and experiences, to better understand screening effectiveness and differences in results between countries. In the recently awarded project, E-Detect TB, by the European Commission several European Union countries are collaborating to pool their screening data together and analyse it [4]. European Union countries not participating in the project will also be invited to join this initiative in a later stage.