Long-term relapses after 12-month treatment for Mycobacterium kansasii lung disease

Study subjects

Patients were identified by reviewing the register of positive cultures at the HUB from January 1, 1990 to December 31, 2005. M. kansasii lung disease was diagnosed according to the ATS criteria 4. Data were retrieved from the individual registers available at the tuberculosis control unit (TCU) and from clinical charts of each patient. Patients with known HIV infection were excluded.

Study design and setting

The present study was a retrospective observational study performed at the TCU of the HUB, a 900-bed teaching hospital in the metropolitan area of Barcelona, which admits only adults. Approval was obtained from the ethics committee of the HUB (CEIC number 177/07). Patients diagnosed with either tuberculosis or NTM infection were referred to the TCU for treatment, unless they were HIV positive. The treatment schedule for M. kansasii lung disease consisted of oral rifampin (600 mg·day⁻¹), isoniazid (300 mg·day⁻¹) and ethambutol (25 mg·kg⁻¹·day⁻¹ during the first 2 months, 15 mg·kg⁻¹·day⁻¹ thereafter) for a 12-month period, with i.m. streptomycin (750–1,000 mg two to three times per week) during the first 3 months. Patients were treated and followed-up according to an established protocol, which included regular nurse and medical visits for assessment of clinical and microbiological evolution, treatment compliance and adverse effects of drugs. Nonattendance at visits was registered and attempts were made to recover missing patients. Microbiological response was monitored by means of sputum smear and culture (whenever sputum could be obtained) every 2 months throughout the treatment period. Chest radiographs were taken at month 2 of treatment and then every 4–6 months if clinical and microbiological evolution was satisfactory. All patients underwent periodic ophthalmological examination, including visual acuity, colour vision and visual field. Post-treatment follow-up with sputum examination and chest radiograph was maintained for a minimum of 3 yrs. However, follow-up could be lengthened or even continued indefinitely due to the concomitant chronic pulmonary disease. Adherence to treatment was initially assessed by asking whether the prescribed treatment had been followed or not and by attendance at scheduled visits. Moreover, since the year 2000, isoniazid metabolites in urine and/or urine colour were tested at every visit 11. Directly observed therapy was not regularly offered to patients with NTM infection. Specific information on the diagnosis, clinical picture, radiology, treatment, adherence and follow-up of the mycobacterial disease for each patient was entered into a database at the TCU.

Microbiological studies

Clinical specimens were processed following the conventional digestion-decontamination procedure 12. Smears were stained by auramine-rhodamine fluorochrome. Positive slides were confirmed by the Ziehl–Neelsen stain 13. Samples (0.5 mL) of specimens were inoculated into BACTEC 12B vials and onto Löwenstein–Jensen medium slants. The liquid cultures were performed, from 1990 to 1998, using the BACTEC 460 system (Becton Dickinson, Sparks, MD, USA) and, from 1999 onwards, using the MB/BacT system (BioMérieux SA, Marcy l’Etoile, France). Isolates were identified by conventional biochemical and culture tests, PCR-restriction fragment length polymorphism analysis of the hsp65 gene and DNA probes (AccuProbe; GenProbe Inc., San Diego, CA, USA). Susceptibility testing was performed with the radiometric (BACTEC) susceptibility test (Becton Dickinson), based on the interpretative criteria used with Mycobacterium tuberculosis 14–16. Susceptibility breakpoints adopted were as follows: 1 μg·mL⁻¹ for isoniazid and rifampin, 5 μg·mL⁻¹ for ethambutol, and 2 μg·mL⁻¹ for streptomycin.

Definitions

Patients were considered to be cured if, at the end of the treatment, they presented resolution of symptoms, disappearance or improvement of radiological alterations and sputum negativity, when available. If sputum could not be obtained at the end of treatment, in addition to the clinical and radiological improvement, at least one negative sputum culture during treatment was mandatory to consider a patient as cured. Relapse was defined as the occurrence, after conversion to negative and completion of a full course of treatment, of two or more positive sputum cultures during a period of >2 weeks, regardless of radiographic appearance. Alternatively, the occurrence of two or more positive sputum cultures with positive smear ≥1 day apart and radiographic changes was also accepted. Treatment was considered to begin on the date when at least rifampin and ethambutol were started. Compliance was considered appropriate whenever attendance at visits was regular and the urine examination (colour and/or isoniazid metabolites) was positive in >80% of visits.

Data analysis

The main end-point of evaluation was the proportion of patients who relapsed. Relapses were expressed as cumulative incidence and rates per person-yr, to adjust for differences in follow-up (incidence density). Since the chance of relapse did not remain stable indefinitely (most relapses occurring in the first years after stopping treatment), all cases were censored after completing the fifth year of follow-up.

Baseline characteristics and treatment regimens

Of the 111 patients with M. kansasii lung disease, 92 (82.8%) completed a full course of treatment. Of these, 79 (85.9%) patients were treated for 12 months with a combination of rifampin, isoniazid and ethambutol. In 66 (83.5%) of these patients, this was supplemented with streptomycin for the first 3 months. In the remaining 13 patients, streptomycin was skipped because of toxicity concerns (n = 5), known allergy (n = 1) or other reasons (n = 7). Four patients were not included in the analysis: one of them was retreated because of positive culture (however, relapse criteria were not fulfilled); and the other three had no follow-up to be assessed. In total, 75 patients were finally included for the analysis. Of these, 37 (49.3%) patients were HIV negative, while in the other 38 (50.7%), HIV status was unknown. No significant difference was found between patients who underwent HIV testing and those who did not, except that chronic pulmonary disease was more frequent in those who did not undergo testing, and there was a lower mean age in those who were tested (76.0% versus 44.4% (p = 0.04) and 53.0 versus 42.6 yrs (p = 0.002) for those not tested versus those who underwent testing, respectively). The main baseline characteristics of the 75 patients are shown in table 1⇓. Most patients were males in the sixth decade of life with some comorbidity, mainly chronic pulmonary disease. Upper-lobe cavitated infiltrates were the predominant radiological finding. Sputum smear was positive in >65% of cases. Strain sensitivity was available in 48 (64.0%) cases. Isolates from 38 (79.2%) of these patients were susceptible to all the drugs. Of the nine strains with some degree of resistance, eight had a minimum inhibitory concentration (MIC) of 5 μg·mL⁻¹ for isoniazid and one had an MIC of 2 μg·mL⁻¹ for rifampin.

Outcome

During follow-up, five (6.6%) out of 75 patients relapsed after a median (range) of 41.5 (11–48) months. The incidence density rate was 2.19 (95% confidence interval 0.71–5.12) per 100 patient-yrs (table 1⇑). All five relapses occurred among patients treated with the combination that included streptomycin. There were no relevant differences regarding main characteristics between patients who relapsed and those who did not, apart from a higher proportion of patients with positive sputum smear in the nonrelapse group and a trend towards higher comorbidity in the relapse group (0 versus 65.7% positive sputum smear (p = 0.007) and 100 versus 70.0% with comorbities (p = 0.18) for relapsed versus nonrelapsed, respectively; table 2⇓). Although not statistically significant, there was a trend towards a relationship between age and relapse: none of the 17 patients ≤40 yrs of age relapsed, compared with five out of the 58 patients >40 yrs of age (p = 0.27). Treatment compliance was considered to be appropriate in the five relapse patients, and the susceptibility test, which was available for all five cases, showed susceptibility to the four drugs tested (rifampin, isoniazid, ethambutol and streptomycin) before the treatment and at the time of relapse. In two out of the five patients, streptomycin was discontinued after 2 and 4 weeks, respectively, because of dizziness. Table 3⇓ provides detailed information on the five relapses. The five patients were retreated with the same drugs used in the first episode, for 18 months. After a median (range) of follow-up of 5 (3.5–8) yrs, they did not have further relapses.

Regarding safety, 11 (14.7%) of the 75 patients suffered some kind of adverse event, which resulted in discontinuation of some drug in seven (9.3%) cases. Dizziness, presumably caused by the streptomycin, was the most frequent cause of treatment discontinuation (four (6.3%) out of the 63 treated with streptomycin). Isoniazid was discontinued in three (4.0%) cases, ethambutol in two (2.7%), and all treatment drugs in one (1.3%) case.

In total, 28 (25.2%) of the 111 initial patients died, 13 of them before finishing treatment. Death was not related to M. kansasii disease or to treatment in any of these cases. Malignancy, mainly lung and throat neoplasm, was the cause of death in 14 (50%) cases.