The 97 retreatment patients studied included 35 patients who were cured and 62 patients who were not cured between collection of the initial and last isolate, respectively. Cure was defined as completion of a full course of chemotherapy with negative sputum smears at the end of treatment and at least one preceding follow-up examination.
Previously cured patients
Of the 35 previously cured patients, 29 were men and 6 were women. Their median age was 35 years (range 16–80 years). At collection of initial samples, 20 of these patients were new cases, 4 were treatment failures, 6 were relapses and 5 others were defaulters. The median time interval between the initial and last isolate was 21 months (range 10–83 months). Initial isolates from 7 of the 35 patients were MDR, whereas those from the other 28 patients were non-MDR of which isolates from 18 patients were pan-susceptible. Isolates from the other 10 patients were resistant to at least one drug but non-MDR. A change in drug resistant profiles of isolates from pan-susceptible to MDR was observed in one patient (patient 10) (Table 1).
Table 1. Patient characteristics, treatment history, drug resistance, treatment and fingerprinting results of the five reinfections cases
|Patient number||Age/sex||Treatment history§||Isolate number||Drug resistance||Date of sampling (month/year)||Treatment regimen‡‡||Treatment outcome (month/year)||DNA fingerprinting||Cause of recurrence|
| 5||39/M†||Recurrent||99-2151||MDR¶||06/99||Cat II||Cured (02/00)||Different||Reinfection|
|00-1553 00-1554||MDR||10/00||MDR||Failure (10/01)||Identical|
|01-2676 02-2031||MDR||05/02||MDR||Failure (11/02)||Identical|
|10||45/F‡||Defaulter||96-0666||Susceptible||03/96||Cat II||Cured (11/96)||Different||Reinfection|
|97-1002 97-0990||MDR||07/97||MDR||Cured (04/99)||Identical|
|13||33/M||Failure||95-0413||INH†† resistant||02/95||Cat II||Cured (10/95)||Different||Reinfection|
|97-1160||INH resistant||08/97||Cat II||Defaulted (03/98)||Identical|
|01-1241||INH resistant||05/01||Cat II||Defaulted 08/01||Identical|
|02-0581||INH resistant||01/02||Cat II||Cured (09/02)||Identical|
|17||18/M||New||96-0816||Susceptible||04/96||Cat I||Cured (01/97)||Different||Reinfection|
|99-1744||Susceptible||04/99||Cat II||Cured (12/99)||Different|
|30||30/M||New||95-0796||Susceptible||05/95||Cat I||Cured (07/96)||Different||Reinfection|
|01-2001||Susceptible||07/01||Cat I||Cured (03/01||Different|
For 30 of the 35 patients, spoligotyping, RFLP and/or MIRU-VNTR patterns of the M. tuberculosis strains responsible for the TB recurrence were identical (28 cases) or showed an addition and a deletion of one IS6110 fragment (two cases), but were identical by MIRU-VNTR, indicating endogenous reactivation. The minor changes in RFLP profiles occurred in isolates with a high (>10) IS6110 copy number (data not shown).
For the other 5 (patients 5, 10, 13, 17 and 30) of the 35 patients, DNA fingerprinting patterns of the M. tuberculosis strains responsible for the disease were different for the two episodes, indicating exogenous reinfection or initial mixed infection (Table 1, Figure 2). Reinfection in three of these patients (patients 5, 10, 13) occurred within 8 months after cure from the previous episode. In the other two cases reinfection occurred more than 3 years after the previous episode. We did not find any evidence suggesting that cross contamination was responsible for these five reinfection cases.
Patient 5 was a recurrent case from whom the first sample was collected in June 1999. The patient received Cat II treatment and was declared cured in February 2000. The patient returned with smear-positive positive TB in October of the same year and was put on MDR treatment, but remained culture positive 1 year later. MDR treatment was repeated from May 2002, but because of poor compliance the patient was declared a treatment failure in November 2002 and died shortly afterwards. All samples collected between the initial and last isolates (intervening isolates) yielded M. tuberculosis strains that were genetically different from the initial isolate, but identical to the last isolate. All 6 isolates were MDR.
Patient 10 was a defaulter recruited into the study in March 1996. The patient received Cat II treatment and was declared cured in November of the same year. However, the patient reported with TB again in June 1997, was put on MDR treatment and declared cured in April 1999. The patient remained culture negative after 2 years of follow-up. Isolates taken 14 months after the initial sample were genetically identical to each other, but different from the initial isolate (Figure 2). The initial isolate was pan-susceptible to all drugs tested, whereas the latter isolates were MDR. RIF resistance in the latter isolates was confirmed by the presence of a His526Leu mutation in the rpoB gene and MIC testing (>80 μg/ml). Resistance to INH was confirmed by the presence of a Ser315Asn mutation in the katG gene and a CT mutation at position −15 (designated relative to the mRNA initiation start site) in the inhA gene and also by MIC testing (> 3.2 μg/ml). The sequence of the rpoB, katG and inhA genes of the initial isolate for this patient was of the wild type and the MIC for both drugs was below the cut off for resistance for both INH (0.2 μg/ml) and RIF (40 μg/ml).
Patient 13 was a Cat I treatment failure at collection of the first sample in February 1995. The patient received Cat II treatment and was declared cured afterwards. However, he presented with smear positive TB in August 1997 and again received Cat II treatment, but defaulted 7 months later. In May 2001, Cat II treatment was restarted, but the patient defaulted again after 3 months. Cat II treatment was repeated from January 2002. The patient was declared cured thereafter and no further follow-up samples yielded a positive culture after 2 years. All the isolates from this patient were resistant to INH and had similar spoligotyping patterns. The two latter isolates were identical by MIRU-VNTR typing and had only minor spoligotype differences with the initial and the first intervening isolate. However, by MIRU-VNTR typing, the initial isolate was very different from all the follow-up isolates and only shared the same number of alleles at locus 23 with the first intervening isolate compared with both subsequent isolates. RFLP results, available for the first two isolates showed completely different patterns. In accordance with the interpretation of DNA fingerprinting data in this study, the follow-up isolates were considered to be genetically identical to each other, but different from the initial isolate. This patient was therefore designated a reinfection case.
Patient 17 was a new case from whom the first sample was collected in April 1996. The patient received Cat I treatment and was declared cured 9 months later. The patient was diagnosed with TB again after 3 years and the isolate cultured was genetically different from the first, but were both pan-susceptible. This patient was consequently considered to be a reinfection case (Table 1). This patient received Cat II treatment, was cured and has remained smear-negative.
Patient 30 was also a new TB case diagnosed in May 1995. The patient received Cat I treatment and was declared cured in July 1996. Six years later, the patient presented with TB again. The recurrent isolate was genetically different from the initial isolate, suggesting reinfection (Table 1 and Figure 2). Both isolates were pan-susceptible. The patient was successfully retreated and has remained smear negative.
Patients not cured between isolates
These 62 patients included 42 males and 28 females with a median age of 40 years (range 12–68 years). The median time interval between the initial and last isolate was 8 months (range 3–69 months). The treatment history of these patients included treatment failures (39), relapses (12), defaulters (7) and new cases (4). Initial isolates from 49 patients were MDR, whereas isolates from the other 13 patients were non-MDR, including 6 pan-susceptible and 7 resistant to at least one drug.
A change in DST patterns between the initial and latter isolate(s) among these 62 patients were observed in 14 cases. For 13 of these patients, the change only involved additional resistance to EMB and/or STR (data not shown). Among the polyresistant isolates, amplification of resistance to RIF was observed in one patient by routine DST and confirmed by sequencing of the rpoB gene (wild type for the initial isolate and Asp516Tyr mutation for the last isolate) and MIC testing (<10 μg/ml for the initial and > 80 μg/ml for the last isolate). INH resistance was confirmed by MIC testing (>3.2 μg/ml) for both isolates.
Minor changes in RFLP profiles among serial isolates were observed in 6 cases, which all occurred in high (>10) IS6110 copy number isolates. Spoligotyping and MIRU-VNTR patterns of sets of isolates from each of these patients were identical (data not shown).
Repeat isolates from all, but one of the 62 patients were considered identical, based on identical spoligotyping and RFLP profiles or only minor differences in RFLP (one or two bands), but identical MIRU-VNTR patterns, indicating treatment failure. For one of these patients, DNA fingerprints of the last isolate differed from that of the corresponding initial isolate, suggesting exogenous reinfection with a susceptible isolate, which is unlikely to occur during treatment. This observation was further strengthened by data obtained from investigations for laboratory cross-contamination where it became evident that a sputum sample from the last isolate of this patient had been switched with a specimen from a different patient processed on the same day following DNA fingerprinting. Consequently, disease in all these 62 patients was considered to be as a result of treatment failure.