Serial IGRA testing of trainees in the healthcare sector in a country with low incidence for tuberculosis - a prospective cohort study

Background: Healthcare workers (HCW) are a risk group for tuberculosis (TB). That is why interferon-gamma release assay (IGRA) serial testing is performed on HCWs repeatedly exposed to infectious patients or materials. However, the variability of IGRA in serial testing is not yet well understood. We therefore analysed the prevalence of positive IGRA results as well as conversion and reversion rates in the serial testing of healthcare trainees in a low-incidence country. Methods: In a prospective cohort study, all trainees (n=194) who began training as a nurse or healthcare worker at the Vivantes Healthcare Training Institute in Berlin on 1 October 2008 or 1 April 2009 were IGRA-tested at three different times during the three years of training. Socio-demographic data and possible risk factors (e.g., TB contacts, time spent abroad, area of work) were recorded by means of a standardised questionnaire. The QuantiFERON Gold In-Tube (QFT) was used as an IGRA. Results: At the beginning of the training the cohort comprised 194 trainees. 70% were female. Their average age was 23. The prevalence of positive QFT was 2.1% (4/194). In the first follow-up test, 2 out of 154 (1.3%) tested IGRA-positive, 151 (98%) had constantly negative results. One IGRA was constantly positive (0.6%) and there was one conversion and one reversion (0.6% respectively). In the second follow-up (n=142) there was again one conversion (0.7%), one reversion and the one constantly positive test result in all three QFT. This trainee had active TB in 2002. All other test results were constantly negative (n=139; 98%). No case of active tuberculosis was diagnosed over the three-year observation period. Contact with TB patients was reported by 42 (29.6%) trainees during the follow-up. The two trainees with a conversion in QFT had no known contact with TB patients. Discordant results in the three consecutive QFT were observed in three trainees (2.1%). Using a borderline zone from 0.2-0.7 IU/mL reduced the number of trainees with discordant results from three to one - a reversion. Conclusion: The prevalence rate of latent TB infection is low in healthcare trainees without known risk factors for TB infection in their history. The infection risk seems to be low in this population even though contacts with TB patients during the training were reported. Introducing a borderline zone for the interpretation of reversions and conversions in this cohort appears to be safe and reduces the number of discordant results and helps to avoid unnecessary chest X-rays and preventive treatment.