Setting: In The Netherlands all immigrants from highly endemic countries undergo obligatory entry screening by X-ray, followed by voluntary half-yearly screening for 2 years.
Objective: To estimate the contribution of screening of immigrants to reductions in tuberculosis transmission.
Design: All bacteriologically confirmed tuberculosis patients diagnosed between 1994 and 1999 with mycobacterial isolates exhibiting identical DNA fingerprints were assigned to clusters. Clusters were assumed to be a consequence of recent transmission and to have had the first patient as their source case. Among patients with pulmonary tuberculosis from highly endemic countries, the characteristics of source patients were compared with those of non-clustered patients.
Results: Of the 1438 selected patients, 187 (13%) were the first in a cluster, 386 (27%) were in a cluster but were not the first case, and 865 (60%) were not clustered. Independent risk factors for being the first in a cluster were young age, nationality, early year of diagnosis, longer duration of treatment and no concurrent extra-pulmonary tuberculosis. In univariate analysis, passively detected patients and patients with a long duration of stay in The Netherlands were more likely to be the first in a cluster than patients detected by screening and recent arrivals, respectively. However, these variables were strongly associated.
Conclusion: Screening of immigrants can reduce tuberculosis transmission. This effect is in part due to confounding by duration of stay.