A total of 405 smear-positive pulmonary TB patients were diagnosed in the Gulripsch hospital between March 2003 and September 2005. Out of these, 366 (90.4%) met the inclusion criteria. Data from 326 patients were included in the final analysis. The 40 exclusions were: due either to shipment problems with sputum specimen for MTB culture for 11 cases, 3 cases with culture contamination and 26 cases with baseline negative culture results.
shows patients' baseline characteristics. A total of 195 patients (59.8%) had strains resistant to at least one 1st
line drug and 135 (41.4%) patients were at least resistant to isoniazid. Among the 195 patients, 57 (29.3%) had strains resistant to at least one 2nd
line drug. A total of 69 patients were PDR-TB (21.2%) and 68 were MDR-TB (20.9%). Among PDR-TB, 2 (2.9%) were resistance to R. Among MDR-TB cases, 3 were XDR-TB (4.5%). The median number of drugs (1st
line), to which MDR-TB patients were resistant to was 4.5 drugs, interquartile range (IQR) 
. One fourth harboured a strain belonging to the Beijing family.
The overall treatment success rate was 72.7% (237/326), defaulting rate 16.2% (53/326), death rate 7.7% (25/326), and failure rate 3.4% (11/326). shows treatment outcomes according to drug resistance patterns. Success rate was 85.2% for patients with fully drug susceptible patients, and 78.3% for patients with PDR-TB strains. There was no statistically significant difference in success rate between patients with full susceptible, H+-S, or HE+-S resistant strains (p
0.67). The small number of PDR-TB resistant to R didn't allow comparison of outcome with other PDR-TB.
Tuberculosis treatment outcomes according to baseline drug resistance patterns.
Almost one fourth of MDR-TB patients died after a median period of treatment of 1 month, IQR [0–7.6]. One third discontinued treatment after a median period of 6.6 months IQR [4–12.9].
MDR-TB treatment success rate was 41.9% (13/31) for patients without resistance to 2nd
line drugs and 27.8% (10/36) for patients with resistance to any 2nd
line drug. However, this difference was non-significant (p
0.22). Out of the 3 XDR-TB patients 2 died and 1 defaulted.
Out of 68 MDR-TB patients, 37 (54.4%) converted to culture negative. After exclusion of 21 patients who died or defaulted during the first 6 months of treatment, the culture conversion rate was 78.7% (37/47). Median time to culture conversion was 6 months IQR 
. Among the 37 patients who converted, 4 (10.8%) reverted to culture positive during treatment at 3 (n
2), 6 (n
1) and 8 (n
1) months after culture conversion. Of these 4 patients, 1 failed treatment, 2 died and 1 defaulted.
None of the baseline MDR-TB patients' characteristics were significantly associated with a negative outcome ().
shows patterns of drug resistance amplification and re-infection during treatment. Three patients with fully susceptible MTB strains were re-infected with a MDR-TB strain and had negative outcomes despite treatment adaptation. Cluster analysis showed that for one of them, the index case was a fellow patient hospitalised during the same period. This cluster included the strains from these 2 patients. Another patient was infected one month after starting treatment with a strain belonging to a cluster of MDR-TB strains identified in 3 other patients. They were all hospitalised for MDR-TB treatment during the same period as the contact case. The last patient was infected with a strain belonging to a cluster of MDR-TB strains identified in 15 patients. Three patients were hospitalised for MDR-TB treatment when the contact was regularly visiting the hospital to receive ambulatory treatment.
Evolution of drug resistance during treatment: amplification and re-infection.
Among 58 PDR-TB strains of patients with complete follow-up culture and finger printing results, 2 amplified to MDR-TB during treatment (3.4%). 1/7 (14.3%) and 1/48 (2.1%) PDR-TB patients were infected with a Beijing strains and with a non-Beijing strain who amplified to MDR-TB, respectively ().
Of the 47 MDR-TB patients susceptible to Ofx at baseline with follow-up culture results, 8 developed Ofx resistance (). DNA fingerprinting was possible for 7 of these (1 culture was contaminated) showing re-infection with a XDR-TB strain in one patient and amplification to Ofx in 6 patients. The amplification rate to Ofx was 13.3% (6/45). It was 3.1% (1/32) and 2.2% (1/46) for Km and Cm, respectively. The Ofx resistance amplification rate was higher in MDR-TB patients infected with a Beijing strain 5/29 (17.2%) compared with those infected with a non-Beijing strain 1/16 (6.2%). However, this does not reach statistical significance (p
0.56). Finally, of the 46 MDR-TB patients who were not XDR-TB at baseline and had valid follow-up culture results and RFLP results, 3 amplified to XDR TB during treatment (6.7%) and one was re-infected with a Beijing XDR-TB strain. The increase in Ofx resistance during treatment was significantly associated with a negative outcome (p<0.001).