E.Gurbanova1, K.Blondal2, F.Mirzayev3, R.Tahirli3, V.Popova4, R.Mehdiyev1, A.Ismayilov5, A.Altraja6
1Main Medical Department of the Ministry of Justice, Baku, Azerbaijan
2Department of Communicable Disease Prevention and Control, Reykjavik Health Care Services, Reykjavik, Iceland
3WHO, Geneva, Switzerland
4Laboratory of the Specialized Treatment Institution for Detainees with Tuberculosis, Baku, Azerbaijan
5Project HOPE, Dushanbe, Tajikistan
6Department of Pulmonary Medicine, University of Tartu, Tartu, Estonia
Background: The Main Medical Department of Azerbaijan Ministry of Justice implements a comprehensive TB Control programme in the penitentiary system (PS). This programme practices active tuberculosis (TB) case finding among inmates and provides effective treatment for all identified cases of both susceptible and drug-resistant TB in the Special Treatment Institution (STI). The STI also houses a modern reference laboratory that is quality-assured by the Supra-national Reference Laboratory in Borstel, Germany and performs all WHO-recommended diagnostics including Xpert MTB/RIF and culture with following drug-susceptibility testing (DST) on MGIT960. As soon as susceptibility result on R is available from either Xpert MTB/RIF or MGIT960, the patient starts adequate treatment with regimen adjustment after full DST is available.
Since 2010, frozen cultures of all culture-positive samples processed in this laboratory are maintained.
Objectives:to assess the proportion of discordant R resistance results on Xpert MTB/RIF and MGIT960 among patients identified in Azerbaijan PS during 2011-2015.
Design and methods: All pulmonary TB cases undergoing testing for TB in STI were included into this retrospective study. TB cases under treatment monitoring were excluded.Sputum samples from admitted patients were investigated both with Xpert MTB/RIF and MGIT960 culture and DST.To increase the number of strains evaluated, additional strains were revived from culture archive.
Results: Overall 532 strains were included, 286 consecutive sputum samples from admitted patients and 246 frozen strains.
Out of the 532 strains, 33 (6.2%) had discordant results, where 21 (3.9%) were R-resistant according to Xpert MTB/RIF, but R-sensitive according to MGIT960 and 12 (2.3%) were R-sensitive on Xpert MTB/RIF but R-resistant according to MGIT960 (Table 1). Both tests, however, were in a good agreement with each other (Cohen's kappa 0.825, P<0.001). If MGIT results were taken as reference, then sensitivity, specificity, positive predictive value (PPV), negative predictive value (NPV) and test accuracy of Xpert MTB/RIF were 90.8%, 95.2%, 84.9%, 97.2% and 94.2% respectively.
Table 1. Rifampicin susceptibility results by Xpert MTB/RIF and MGIT960.
Xpert MTB/RIF | ||||
Sensitive | Resistant | Total | ||
MGIT960 | Sensitive | 393 | 21 | 414 |
Resistant | 12 | 106 | 118 | |
Total | 405 | 127 | 532 |
Conclusions: Despite the 6.2% discordance between the methods, the results by Xpert MTB/RIF had acceptable sensitivity, specificity, PPV, NPV and accuracy and were in a good agreement with those obtained with MGIT960. Further research and sequencing will be necessary to better understand the genetic profile of strains that exhibit discordance in Rif susceptibility results between these two diagnostic techniques.